Spectroscopic Methods in Organic Chemistry PDF

Ian Fleming · Dudley Williams Spectroscopic Methods in Organic Chemistry Seventh Edition Spectroscopic Methods in Organic Chemistry Ian Fleming Dudley Williams Spectroscopic Methods in Organic Chemistry Seventh Edition Ian Fleming Dudley Will...

Ian Fleming · Dudley Williams Spectroscopic Methods in Organic Chemistry Seventh Edition Spectroscopic Methods in Organic Chemistry Ian Fleming Dudley Williams Spectroscopic Methods in Organic Chemistry Seventh Edition Ian Fleming Dudley Williams Department of Chemistry Department of Chemistry University of Cambridge University of Cambridge Cambridge Cambridge UK UK ISBN 978-3-030-18251-9    ISBN 978-3-030-18252-6 (eBook) https://doi.org/10.1007/978-3-030-18252-6 © Springer Nature Switzerland AG 2019 1st-6th edition: © McGraw-Hill Education 1966, 1973, 1980, 1989, 1995, 2007 This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part of the ma- terial is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter de- veloped. The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use. The publisher, the authors, and the editors are safe to assume that the advice and information in this book are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or the editors give a warranty, expressed or implied, with respect to the material contained herein or for any errors or omissions that may have been made. The publisher remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. This Springer imprint is published by the registered company Springer Nature Switzerland AG The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland Preface This book is the seventh edition of a well-established introductory guide to the interpre- tation of the mass, ultraviolet, infrared and nuclear magnetic resonance spectra of organic compounds. It is a textbook suitable for a course in the application of these techniques to structure determination and as a handbook for organic chemists to keep on their desks throughout their career. These four spectroscopic methods are used routinely to determine the structure of orga- nic compounds. Every organic chemist needs to be skilled in how to apply them and to know which method works for which problem. In outline, the mass spectrum gives the molecular formula, the ultraviolet spectrum identifies conjugated systems, the infrared or Raman spectrum identifies functional groups and the nuclear magnetic resonance spectra identify how the atoms are connected. One or more of these techniques are frequently enough to identify the complete chemical structure of an unknown compound. If they are not enough on their own, there are other methods that the organic chemist can turn to: X-ray diffraction, microwave absorption, electron spin resonance, atomic force spectros- copy and circular dichroism, among others. Powerful though they are, these techniques are all more specialised and less part of the everyday practice of most organic chemists. In preparing this edition, I have moved the chapter on mass spectrometry to be the first, in order to reflect its importance in giving a molecular formula—a vitally important first step in a structure determination. Along with the following chapters on UV and IR spectra, it remains concise, to reflect the lesser importance of all these techniques, with the discus- sion of the theoretical background kept to a minimum, since application of spectroscopic methods is possible without a detailed command of the theory behind them. Benefitting from my experience teaching the subject over the last 7 years, I have com- pletely rewritten the chapter on NMR spectroscopy, dividing it into two to separate the discussion of 1D spectra from the 2D spectra. In addition, I have greatly expanded the text with explanations of the spin physics that the earlier edition entirely lacked. I have intro- duced each of the pulse sequences, and the vector description of the resultant magnetisa- CRY page is updated for previous edition information. v vi Preface tion, as each is needed in order to understand its effect on the spectra. It is nowhere near a rigorous account but simply provides the minimum level of understanding an organic ­chemist using NMR needs. The augmented tables of data at the ends of the chapters on MS, IR and 1D-NMR ­remain an important resource for reference by any chemist practising organic synthesis. Finally, in Chap. 6, there are 11 worked examples illustrating the ways in which the four spectroscopic methods can be brought together to solve simple structural problems, and in Chap. 7, there are 34 problem sets with increasing difficulty for you to work through, many more with 2D NMR spectra than in the earlier editions. I have been helped by Drs. Richard Horan, Chris Jones and Ed Anderson for their work on the sixth edition that has survived into this; Duncan Howe in Cambridge and Dr. Dean Olson and Lingyang Zhu in Urbana Champaign took the many new NMR spectra; my colleague Jeremy Sanders has given me much useful advice, both about this book and more generally; several colleagues (Chris Urch, Scott Denmark, Alain Krief, Mike Alders- ley and Rob Britton) have been generous with samples or spectra that have found their way into the book; I am most grateful to them all and finally to Manuel Perez for the gift of yearly licences to use the Mestrenova program for processing NMR spectra; I used ­Photoshop and ChemDraw to prepare all the figures. Cambridge, UK Ian Fleming Contents 1 Mass Spectra   1 1.1 Introduction   1 1.2 Ion Production   2 1.2.1 Electron Impact (EI)    2 1.2.2 Chemical Ionisation (CI)   3 1.2.3 Electrospray Ionisation (ESI)   5 1.2.4 Fast Ion Bombardment (FIB or LSIMS)   6 1.2.5 Laser Desorption (LD) and Matrix-Assisted Laser Desorption (MALDI)   7 1.3 Ion Analysis   8 1.3.1 Magnetic Analysers   8 1.3.2 Time-Of–Flight (TOF) Analysers    8 1.3.3 Quadrupole Analysers   9 1.3.4 Ion Cyclotron Resonance (ICR) Analysers 10 1.3.5 Ion-Trap Analysers 10 1.4 Structural Information from EI Mass Spectra 11 1.4.1 The Features of an EI Spectrum 11 1.4.2 The Molecular Ion 12 1.4.3 Isotopic Abundances  13 1.4.4 Identifying the Molecular Ion  14 1.4.5 Fragmentation in EI Spectra  15 1.5 Fragmentation in CI and FIB Spectra  27 1.5.1 Fragmentation in CI Spectra  27 1.5.2 Fragmentation in FIB (LSMIS) Spectra  28 1.6 Some Examples of Mass Spectrometry in Action 30 1.6.1 San Joaquin Oil  31 1.6.2 Oleic Acid  32 1.6.3 The Oviposition Pheromone  33 1.6.4 Identifying Antibodies 34 1.6.5 The ESI Spectra of Melittin and the Human Parathyroid Hormone 35 1.6.6 ESI-FT-ICR and ESI-FT-Orbitrap Spectra 37 vii viii Contents 1.7 Separation Coupled to Mass Spectrometry 37 1.7.1 GC/MS and LC/MS 37 1.7.2 MS/MS 39 1.8 Interpreting the Spectrum of an Unknown 40 1.9 Internet 41 1.10 Further Reading 41 1.11 Problems  43 1.12 Tables of Data  45 References 53 2 Ultraviolet and Visible Spectra  55 2.1 Introduction 55 2.2 Chromophores 56 2.3 The Absorption Laws 57 2.4 Measurement of the Spectrum 57 2.5 Vibrational Fine Structure 58 2.6 Selection Rules and Intensity 58 2.7 Solvent Effects 60 2.8 Searching for a Chromophore 60 2.9 Definitions 61 2.10 Conjugated Dienes 62 2.11 Polyenes and Poly-Ynes  64 2.12 Ketones and Aldehydes; π→π∗ Transitions 66 2.13 Ketones and Aldehydes; n→π∗ Transitions 69 2.14 α,β-Unsaturated Acids, Esters, Nitriles and Amides 70 2.15 Aromatic Compounds  70 2.16 Quinones  77 2.17 Corroles, Chlorins and Porphyrins 77 2.18 Non-conjugated Interacting Chromophores 78 2.19 The Effect of Steric Hindrance to Coplanarity  79 2.20 Internet 80 2.21 Further Readings 81 2.22 Problems  81 References 83 3 Infrared and Raman Spectra 85 3.1 Introduction 85 3.2 Preparation of Samples and Examination in an Infrared Spectrometer 86 3.3 Selection Rules 88 3.4 The Infrared Spectrum 88 3.5 The Use of the Tables of Characteristic Group Frequencies 90 3.6 Stretching Frequencies of Single Bonds to Hydrogen 91 3.7 Stretching Frequencies of Triple and Cumulated Double Bonds 94 Contents ix 3.8 Stretching Frequencies in the Double-Bond Region 96 3.9 Characteristic Vibrations of Aromatic Rings 100 3.10 Groups Absorbing in the Fingerprint Region 101 3.11 Raman Spectra 101 3.12 Internet 105 3.13 Further Reading 106 3.14 Problems  107 3.15 Correlation Charts 109 3.16 Tables of Data  111 4 1D Nuclear Magnetic Resonance Spectra 123 4.1 Nuclear Spin and Resonance 123 4.2 Taking a Spectrum 125 4.3 The Chemical Shift  134 4.4 Factors Affecting the Chemical Shift 139 4.4.1 The Inductive Effect 139 4.4.2 Anisotropy of Chemical Bonds 140 4.4.3 Polar Effects of Conjugation 143 4.4.4 Van der Waals Forces 145 4.4.5 Isotope Effects 146 4.4.6 Estimating a Chemical Shift  147 4.4.7 Hydrogen Bonds  149 4.4.8 Solvent Effects and Temperature 151 4.5 Spin-Spin Coupling to 13C 151 4.5.1 13C–2H Coupling  152 4.5.2 13C–1H Coupling  153 4.5.3 13C–13C Coupling 159 4.6  H–1H Coupling: Multiplicity and Coupling Patterns 161 1 4.6.1 1H–1H Vicinal Coupling (3JHH) 162 4.6.2 AB Systems 169 4.6.3 1H–1H Geminal Coupling (2JHH) 174 4.6.4 1H–1H Long-Range Coupling (4JHH and 5JHH)  178 4.6.5 Deviations From First-Order Coupling 184 4.7 1H–1H Coupling: The Magnitude of Coupling Constants 190 4.7.1 The Sign of Coupling Constants 190 4.7.2 Vicinal Coupling (3JHH)  192 4.7.3 Geminal Coupling (2JHH) 199 4.7.4 Long-Range Coupling (4JHH and 5JHH) 201 4.7.5 C–H Coupling (1JCH, 2JCH and 3JCH) 201 4.8 Coupling From 1H and 13C to 19F and 31P 202 4.8.1 13C NMR Spectra of Compounds Containing 19F and 31P 202 4.8.2 1H NMR Spectra of Compounds Containing 19F and 31P 205 x Contents 4.9 Relaxation and Its Consequences 206 4.9.1 Longitudinal Relaxation 207 4.9.2 Transverse Relaxation and Exchange  209 4.10 Improving the NMR Spectrum  215 4.10.1 The Effect of Changing the Magnetic Field 215 4.10.2 Solvent Effects 217 4.10.3 Shift Reagents  220 4.11 Spin Decoupling 221 4.11.1 Simple Spin Decoupling  221 4.11.2 Difference Decoupling 222 4.12 Identifying Spin Systems: 1D-TOCSY 223 4.13 The Nuclear Overhauser Effect 227 4.13.1 Origins 227 4.13.2 NOE-Difference Spectra  229 4.14 The Rotating Frame of Reference 232 4.15 Assignment of CH3, CH2, CH and Fully Substituted Carbons in 13C NMR 237 4.15.1 The Attached Proton Test (APT) 237 4.15.2 DEPT 243 4.16 Hints for Structure Determination Using 1D-NMR 245 4.16.1 Carbon Spectra 245 4.16.2 Proton Spectra 246 4.17 Further Information 247 4.17.1 The Internet 247 4.17.2 Further Reading 248 4.18 Tables of Data  251 References 276 5 2D-Nuclear Magnetic Resonance Spectra  277 5.1 The Basic Pulse Sequence 277 5.2 COSY 281 5.2.1 Cross Peaks from Scalar Coupling  281 5.2.2 Polarisation Transfer  282 5.2.3 The Origin of Cross Peaks 285 5.2.4 Displaying COSY Spectra 286 5.2.5 Interpreting COSY Spectra  287 5.2.6 Axial Signals 291 5.2.7 Gradient Pulses  292 5.2.8 DQF-COSY 293 5.2.9 Phase Structure in COSY Spectra 294 5.3 2D-TOCSY  296 5.4 NOESY  299 5.5 Cross-Correlated 2D Spectra Identifying 1-Bond Connections 302 Contents xi 5.5.1 Heteronuclear Multiple Quantum Coherence (HMQC) Spectra 302 5.5.2 Heteronuclear Single Quantum Coherence (HSQC) Spectra  303 5.5.3 Examples of HSQC Spectra 304 5.5.4 Non-Uniform Sampling (NUS) 306 5.5.5 Cross-Peak Detail: Determining the Sign of Coupling Constants 309 5.5.6 Clip-HSQC 310 5.5.7 Deconvoluting a 1H-Spectrum Using the HSQC Spectrum 311 5.5.8 HSQC-TOCSY  313 5.5.9 HETCOR 313 5.6 Cross-Correlated 2D Spectra Identifying 2- and 3-Bond Connections 314 5.6.1 The HMBC Pulse Sequence 314 5.6.2 HMBC Spectra 315 5.7 Some Specialised NMR Techniques  318 5.7.1 ADEQUATE: Identifying 13C–13C Connections  318 5.7.2 INADEQUATE: Identifying 13C–13C Connections  320 5.7.3 HSQC-HECADE: Measuring the Sign and Magnitude of 13C–1H Coupling Constants 325 5.8 Three- and Four-Dimensional NMR 329 5.9 Hints for Structure Determination Using 2D-NMR 331 5.10 Further Reading 332 5.11 Table of Information  332 Reference  333 6 Worked Examples 335 6.1 General Approach  335 6.2 Simple Worked Examples Using 13C NMR Alone 336 6.3 Simple Worked Examples Using 1H NMR Alone 337 6.4 Simple Worked Examples Using the Combined Application of MS, UV, IR and 1D-NMR Spectroscopic Methods 339 6.5 Worked Examples Using the Combined Application of MS, UV, IR, 1D-NMR and 2D-NMR Spectroscopic Methods 348 7 Problem Sets  361 7.1 Chemical Shift Problems 361 7.2 1D-NMR Chemical Shift and Coupling Problems  364 7.3 Problems Using all the Spectroscopic Methods 370 Index 415 Mass Spectra 1 1.1 Introduction A mass spectrometer is a device for producing and weighing ions from a compound for which we wish to obtain molecular weight and structural information. All mass spectrome- ters use three basic steps: molecules M are taken into the gas phase; ions, such as the cations M +, MH+ or MNa+, are produced from them (unless the molecules are already charged); and the ions are separated according to their mass-to-charge ratios (m/ze). The value of z is normally one, and since e is a constant (the charge on one electron), m/z gives the mass of the ion. Some of the devices that are used to produce gas-phase ions put enough vibrational energy into the ions to cause them to fragment in various ways to produce new ions with smaller m/z ratios. Through this fragmentation, structural information can be obtained. Because the molecular weight and the molecular formula of an unknown are usually the first pieces of information to be sought in the investigation of a chemical structure, mass spectrometry is often the first of the spectroscopic techniques to be called upon. For that reason this chapter comes first. A mass spectrometer detects only the charged components (e.g. MH+, and its associa- ted fragments A+, B+, C+, etc.), because only they are contained, accelerated or deflected by the electromagnetic or electrostatic fields used in the various analytical systems, and only they give an electric signal when they hit the collector plate. When the array of ions has been separated and recorded, the output is known as a mass spectrum. It is a record of the abundance of each ion reaching the detector (plotted vertically) against its m/z value (plotted horizontally). The mass spectrum is a result of a series of competing and consecu- tive unimolecular reactions, and what it looks like is determined by the chemical structure and reactivity of the sample molecules. It is not a spectroscopic method based on electro- magnetic radiation, but since it complements information provided to the organic chemist by the UV, IR and NMR spectra, it is conveniently considered alongside them. © Springer Nature Switzerland AG 2019 1 I. Fleming, D. Williams, Spectroscopic Methods in Organic Chemistry, https://doi.org/10.1007/978-3-030-18252-6_1 2 1 Mass Spectra Mass spectrometry is the most sensitive of all these methods. It can be carried out rou- tinely with a few micrograms of sample, and in favourable cases even with picograms (10−12 g), making it especially important in solving problems where only a very small sample is available, as in the detection and analysis of such trace materials as pheromones, atmospheric pollutants, pesticide residues, and of drugs and their metabolites, especially in forensic science and in medical research. It is even possible to use mass spectrometry to analyse individual slices of tissue to detect the change from cancerous to non-cancerous as an operation is in progress. There is a variety of instruments available for taking a mass spectrum—they differ in the methods by which molecules are taken into the gas phase, they differ by the way in which molecules are induced to produce ions, and they differ by the method used to ana- lyse the ions. Perhaps most significantly, as far as this book is concerned, they also differ in the degree to which they induce fragmentation. In this chapter, we shall see how these instruments provide molecular weight and structural information for molecules with a re- latively low molecular weight, say 1) 61 CH2CH2SH+ Aliphatic thiol 66 H2S2+ Dialkyl disulfide 68 CH2CH2CH2CN+ 69 CF3+, C5H9+ 70 C5H10+ 1.12 Tables of Data 51 Table 1.15 (continued) Groups commonly m/z associated with the mass Possible inference 71 C5H11+ C5H11 —X 71 C3H7CO + Propyl ketone, butyrate ester 72 CH2=C(OH)C 2H5+ Some ethyl alkyl ketones 72 C3H7CH=NH2+ and isomers Some amines 73 C4H9O+ 73 CO 2C2H5+ Ethyl ester 73 (CH3)3Si+ (CH3)3Si —X 74 CH2=C(OH)OCH 3+ Some methyl esters 75 C2H5CO(OH 2)+ C2H5CO 2CnH2n+1 (n>1) 75 (CH3)2Si=OH + (CH3)3SiO —X 76 C6H4+ C6H5 —X, X—C6H4 —Y 77 C6H5+ 78 C6H6+ 79 C6H7+ 79/81 (1:1) Br+ 80/82 (1:1) HBr+ X 80 C5H6N+ N N H X 81 C5H5 O+ X O 83/85/87 HCCl 3+ CHCl 3 (9:6:1) 85 C 6 H 13 + C 6 H 13 —X 85 C 4 H 9 CO + C 4 H 9 CO—X 85 O O X 85 O O X O O 86 CH 2 =CHC(OH)C 3 H 7 + Some propyl alkyl ketones 86 C 4 H 9 CH=NH 2 & isomers Some amines 87 CH 2 =CHC(=OH +)OCH 3 X—CH 2 CH 2 CO 2 CH 3 91 C 7 H 7+ C 6 H 5 CH 2 —X 92 C 7 H 8+ C 6 H 5 CH 2 —alkyl 52 1 Mass Spectra Table 1.15 (continued) Groups commonly associated m/z with the mass Possible inference 92 C 7 H 8+ X N 91/93 (3:1) n-Alkyl chloride Cl 93/95 (1:1) CH 2 Br + R—CH 2 Br C6H6O+ C 6 H 7 O—alkyl (alkyl> CH 3 O O 94 X N N H H O O 95 X O O O Me 95 C6H7O+ X O 97 C 5H5S+ X S O O 97 O O 99 X O O O O 105 C6H5 CO + C 6 H 5 CO—X 105 C 8 H 9+ CH 3 C 6 H 4 CH 2 —X 106 C7H8N+ Me X N 107 C7H7O+ HO X 107/109 (1:1) C 2 H 4 Br + O O 111 X S S References 53 Table 1.15 (continued) Groups commonly associated m/z with the mass Possible inference 121 C8H9O+ MeO X 122 C 6 H 5 CO 2 H + Alkyl benzoates 123 C 6 H 5 CO 2 H 2 + Alkyl benzoates 127 I+ 128 HI + 135/137 (1:1) n-Alkyl bromide (>pentyl) Br X 130 C9H8N+ N H 141 CH 3I+ 147 (CH 3)3 Si=O - Si(CH 3)3+ O 149 O R Dialkyl phthalates O MeO X 160 C 10 H 10 NO + N H MeO X 190 C 11 H 12 NO 2+ MeO N H References 1. Smith RD, Loo JA, Edmonds CG, Baringa CJ (1990). Anal Chem 62:882–899 2. Loo JA, Quinn JP, Ryu SI, Henry KD, Senko MW, McLafferty FW (1992). Proc Natl Acad Sci U S A 89:286–289 3. Covey TR, Huang EC, Henion JD (1991). Anal Chem 63:1193–1200 Ultraviolet and Visible Spectra 2 2.1 Introduction The ultraviolet and visible spectra of organic compounds are associated with transitions between electronic energy levels in which an electron from a low-energy orbital in the ground state is promoted into a higher-energy orbital. Normally, the transition occurs from a filled to a formerly empty orbital (Fig. 2.1) to create a singlet excited state. The wa- velength of the absorption is a measure of the separation of the energies of the ground and excited states, related to but not the same as the energy separation E in Fig. 2.1 of the energy levels of the orbitals concerned. Energy is related to wavelength by Eq. (2.1). 1.19 ´ 10 5 ( E kJmol -1 =) l ( nm ) (2.1) Thus, 297 nm, for example, is equivalent to 400 kJ (~96 kcal)—enough energy to initi- ate many interesting reactions; compounds should not, therefore, be left in the ultraviolet beam any longer than is necessary. Fig. 2.1 Electronic transitions between occupied and unoccupied molecular orbitals hn E Ground state Excited state © Springer Nature Switzerland AG 2019 55 I. Fleming, D. Williams, Spectroscopic Methods in Organic Chemistry, https://doi.org/10.1007/978-3-030-18252-6_2 56 2 Ultraviolet and Visible Spectra 2.2 Chromophores The word chromophore is used to describe the system containing the electrons respon- sible for the absorption in question. Chromophores leading to the shortest wavelength absorption, in other words the highest energy separation, are found when electrons in σ bonds are excited, giving rise to absorption in the 120–150 nm (1 nm = 10−7 cm = 10 Å = 1 mμ) range, corresponding to the transition x in Fig. 2.2. Isolated double bonds like that in ethene give rise to a strong absorption maximum at 162 nm, corresponding to the transition y in Fig. 2.2. Since the air is full of σ and π bonds, it strongly absorbs UV light below 200 nm, and this range is known as the va- cuum ultraviolet, since air must be excluded from the instrument in order to detect the absorption. The absorption at these short wavelengths is difficult to measure and of little use in structure determination. Above 200 nm, however, excitation of electrons from conjugated π-orbitals, gives rise to readily measured and informative spectra. When two double bonds are conjuga- ted, the energy level of the highest occupied molecular orbital ψ2 (the HOMO) is raised in energy relative to the π orbital of the isolated double bonds, and that of the lowest unoccupied molecular orbital ψ3∗ (the LUMO) is lowered relative to π∗. The transition from ψ2 to ψ3∗ is now associated with the even smaller value z in Fig. 2.2. This transi- tion appears in the spectrum of butadiene as a strong, easily detected, and easily mea- sured maximum at 217 nm. The same principle governs the energy levels when unlike chromophores, for example those of an α,β-unsaturated ketone, are conjugated to- gether. Thus, methyl vinyl ketone has an absorption maximum at 225 nm, while neither a carbonyl group nor an isolated C=C double bond has a strong maximum above 200 nm. Since the longest wavelength absorption is usually that caused by promotion of an electron from the HOMO to the LUMO, it measures how far apart in energy those important orbitals are. If yet another π bond is brought into conjugation, the separation of the HOMO and LUMO is further reduced, and absorption occurs at a longer wavelength, with hexatri- ene absorbing at 267 nm. Each successive addition of a double bond reduces the energy gap, and moves the longest wavelength maximum further towards the visible. The long conjugated polyene lycopene, with 11 conjugated double bonds, has its longest wa- velength absorption maximum at 504 nm (ε 158,000) with a tail reaching far into the visible region (absorbing the light in the blue to the orange range). Lycopene is respon- sible for the red colour of tomatoes. The most important point to be made is that, in general: The longer the conjugated system, the longer the wavelength of the absorption maxi- mum. 2.4 Measurement of the Spectrum 57 y4* s* p* p* p* y3* pC y z x pC sp3C sp3C y2 p p p s y1 ethane ethene butadiene Fig. 2.2 The molecular orbitals of single, double and conjugated double bonds 2.3 The Absorption Laws Two empirical laws have been formulated about the absorption intensity. Lambert’s law states that the fraction of the incident light absorbed is independent of the intensity of the source. Beer’s law states that the absorption is proportional to the number of absorbing molecules. From these laws, the remaining variables give the Eq. (2.2). I0 log10 = e.l.c (2.2) I I0 and I are the intensities of the incident and transmitted light respectively, l is the path length of the absorbing solution in centimetres, and c is the concentration in moles per litre. Log10(I0/I) is called the absorbance or optical density; ε is known as the molar extinction coefficient and has units of 1000 cm2 mol−1 but the units are, by convention, not normally expressed. 2.4 Measurement of the Spectrum The ultraviolet or visible spectrum is usually taken using a dilute solution. An appropriate quantity of the compound (often about 1 mg when the compound has a molecular weight of 100–400) is weighed accurately, dissolved in the solvent of choice, and made up to, for instance, 100 mL. Common solvents are 95% ethanol (commercial absolute ethanol cont- ains residual benzene, which absorbs in the ultraviolet), hexane and cyclohexane. A ­portion of this solution is transferred to a silica cell 1 cm from front to back internally 58 2 Ultraviolet and Visible Spectra (the value l in Eq. (2.2)), and the pure solvent is transferred into an accurately matched cell. Two equal beams of ultraviolet or visible light are passed, one through the solution of the sample, and one through the pure solvent. The intensities of the transmitted beams are then compared over the whole wavelength range of the instrument. The spectrum is plotted automatically on most instruments as a log10(I0/I) ordinate and λ abscissa and might look something like Fig. 2.3, which is the spectrum of styrene 1 (molecular weight 104) as a solution of 0.535 mg in 100 mL of hexane and a path length of 1 cm. For publication and comparisons the optical density is converted to an ε versus λ or log ε versus λ plot using Eq. (2.2). The unit of λ is almost always nm. The intensity of a tran- sition is better measured by the area under an absorption peak (when plotted as ε against frequency), but for convenience, and because of the difficulty of dealing with overlapping bands, εmax, the maximum intensity of the absorption, is adopted in everyday use. Spectra are quoted, therefore, in terms of λmax, the wavelength at the maximum of the absorption peak read directly off the plot like that in Fig. 2.3, where it is 250 nm, and the ε value of 14,700 calculated from the value of log10(I0/I), which is 0.756 on the plot in Fig. 2.3. Typically, this spectrum would be reported thus: λmax 250 (hexane) (ε 14 700) and 282 (ε 750) nm with the second of the peaks, barely resolved in Fig. 2.3, explained in Sect. 2.15. 2.5 Vibrational Fine Structure The excitation of electrons is accompanied by changes in the vibrational and rotational quantum numbers so that what would otherwise be an absorption line becomes a broad peak containing all the vibrational and rotational fine structure. Because of interactions of solute with solvent molecules this is usually not resolved, and a smooth curve is obser- ved like that illustrated in Fig. 2.3. In the vapour phase, in non-polar solvents, and with certain peaks (e.g. benzene with the 260 nm band), vibrational fine structure is sometimes resolved. 2.6 Selection Rules and Intensity The irradiation of organic compounds does not always give rise to excitation of electrons from any filled orbital to any unfilled orbital, because there are selection rules based on symmetry governing which transitions are allowed. The intensity of the absorption is the- refore a function of the ‘allowedness’, or otherwise, of the electronic transition and of the 2.6 Selection Rules and Intensity 59 emax is derived from the ordinate using Eq. 2.2 log10(I0/I) 0.5 1 lmax 0 210 250 l 300 nm Fig. 2.3 Ultraviolet spectrum of styrene target area able to capture the light. In practice, a chromophore with two double bonds conjugated together gives rise to absorption by a fully allowed transition with ε values of about 10,000, while many forbidden transitions (which do nevertheless occur) have ε values below 1000. The important point is that, in general: The longer the conjugated system, the more intense the absorption. The selection rules are a function of the symmetry and multiplicity of the ground state and excited state orbitals concerned, and whether the molecular orbitals of each overlap. A theoretical picture is given in the books by Jaffe and Orchin and by Murrell, listed in the section on further reading, but a simple knowledge of which of the commonly encountered transitions are allowed and which are forbidden is adequate for anyone using UV spectra simply to determine organic structures or to follow reaction kinetics. Thus, the important promotion of an electron from the HOMO of a linear conjugated system to the LUMO of the same system is allowed, and always leads to intense absorption. In contrast, two im- portant forbidden transitions are the n→π∗ band near 300 nm of ketones, with ε values of the order of 10–100, and the benzene 260 nm band, and its equivalent in more complicated systems, with ε values from 100 upwards. ‘Forbidden’ transitions like these, with εmax ty- pically 210 nm lmax 227 nm (e 5500) This observation provides an opportunity to stress that changes between the ultraviolet spectrum of a starting material and a product make it one of the easiest tools to use for following the kinetics of a chemical reaction, and that ultraviolet spectroscopy is possibly used more for this purpose than in structure determination. Nevertheless, the immediate and highly sensitive detection of conjugated systems is still a powerful application for this the oldest of the spectroscopic methods. 2.20 Internet The Internet is a continuously evolving system, with links and protocols changing fre- quently. The following information is inevitably incomplete and may no longer apply, but it gives you a guide to what you can expect. Some websites require particular operating systems and may only work with a limited range of browsers, some require payment, and some require you to register and to download programs before you can use them. Ultraviolet spectroscopy is not as well served on the Internet as the other spectroscopic methods. The set of books, Organic Electronic Spectral Data, is still the best source for UV and visible spectroscopic data. There is a database of 1600 compounds with UV data on the NIST website belon- ging to the United States Secretary of Commerce: http://webbook.nist.gov/chemistry/ name-ser.html Type in the name of the compound you want, check the box for UV/Vis spectrum, and click on Search, and if the ultraviolet spectrum is available it will show it to you. ACD (Advanced Chemistry Development) Spectroscopy sell proprietary software cal- led ACD/SpecManager that handles all four spectroscopic methods, as well as other ana- lytical tools: http://www.acdlabs.com/products/spec_lab/exp_spectra/ It is able to process and store the output of the instruments that take spectra, and can be used to catalogue, share and present your own data. It also gives access to a few free data- bases for UV spectra. Wiley-VCH keep an up-to-date website on their spectroscopic books and provide links. The URL giving access to information about spectroscopy, including UV, is: 2.22 Problems 81 http://www.spectroscopynow.com/Spy/basehtml/SpyH/1,1181,7-4-773-0-773- directories%2D%2D0,00.html 2.21 Further Readings Data (1960–96) Organic electronic spectral data, vols. 1–31. Wiley, New York (1979) Sadtler handbook of ultraviolet spectra. Sadtler Research Laboratories Kirschenbaum DM (ed) (1972) Atlas of protein spectra in the ultraviolet and visible region. Plenum, New York Textbooks Jaffe HH, Orchin M (1962) Theory and applications of ultraviolet spectroscopy. Wiley, New York Murrell JN (1963) The theory of the electronic spectra of organic molecules. Methuen, London Barrow GR (1964) Introduction to molecular spectroscopy. McGraw-Hill, New York Brittain EFH, George WO, Wells CHJ (1970) Introduction to molecular spectroscopy. Academic, London Scott AI (1964) Interpretation of the ultraviolet spectra of natural products. Pergamon Press, Oxford Mason SF (1963) Chapter 7. The electronic absorption spectra of heterocyclic com- pounds. In: Physical methods in heterocyclic chemistry, vol II. Academic, New York Rao CNR (1975) Ultraviolet and visible spectroscopy, 3rd edn. Butterworths, London Thomas MJK (1996) Ultraviolet and visible spectroscopy, 2nd edn. Wiley, Chichester 2.22 Problems 1. The reaction of cyclohexanone with isopropenyl Grignard followed by treating the product with acid gave a hydrocarbon, C9H14 which was incorrectly formulated as the structure 51. Suggest an alternative structure in better agreement with its observed UV spectrum: λmax 242 nm. BrMg OH O H2SO4 51 82 2 Ultraviolet and Visible Spectra 2. The UV spectra for the three compounds C7H8O in Problems 1–3 in Ch. 1, are given below. Assign which UV spectrum A-C is derived from which isomer in Problems 1–3. A: λmax 219 (ε 6900), 265sh (ε 1200), 271 (ε 1800) and 278 (ε 1500) nm, unaffected by adding sodium hydroxide. B: λmax 210 (ε 7600) and 252 (ε 250) nm unaffected by adding sodium hydroxide. C: λmax 228 (ε 5800) and 279 (ε 1900) nm changed to λmax 295 (ε 2630) nm on ad- ding sodium hydroxide 3. The five isomers 52–56 of formula C8H8O2 can be distinguished by each of the four spectroscopic methods. The 13C-NMR spectra show that they all have six different car- bon atoms. The UV spectra are the least decisive but it is possible to be reasonably confident of all of the assignments. The UV spectra are printed below, labelled with the letters D-H. Assign the letters D-H to the numbers 52–56. OAc CHO CO2H CO2Me CO2H MeO Me 52 53 54 55 56 e e 10,000 10,000 D E 5,000 5,000 250 l 300 250 l 300 e e e 10,000 10,000 10,000 F G H 5,000 5,000 5,000 250 l 300 250 l 300 250 l 300 4. The cyclopropyl ketone 57 on treatment with base gave an isomer to which the stereo- structure 58 was first assigned. Later, Ranganathan and Shechter realised that this could not possibly be true, because, unlike the starting material, the isomer was coloured bright yellow. Suggest a better structure for the product, which must have more conju- gation in order to be yellow. References 83 COPh COPh 1. KOH, MeOH 2. HCl, cold COPh COPh 57 58 5. The intensely black solution of heptafulvalene 59 in dichloromethane was rendered colourless by treatment with one equivalent of dichlorocarbene. On adding water, the aqueous layer became deep blue, with λmax at 600 nm. Suggest structures for the inter- mediate I and the product J, compatible with these dramatic changes in electronic ab- sorption. :CHCl H 2O I J C15H13Cl C15H13 59 References 1. Kamlet MJ (ed) (1960–1996) Organic electronic spectral data. Wiley, New York 2. Scott AI (1964) Interpretation of the ultraviolet spectra of natural products. Pergamon Press, Oxford 3. Nayler P, Whiting MC (1955). J Chem Soc 1955:3037–3047 4. Sörensen JS, Bruun T, Holme D, Sörensen NA (1954). Acta Chem Scand 8:26–33 5. Bohlmann F, Mannhardt H-J, Viehe H-G (1955). Chem Ber 88:361–370 6. Jaffe HH, Orchin M (1962) Theory and applications of ultraviolet spectroscopy. Wiley, New York Infrared and Raman Spectra 3 3.1 Introduction The infrared spectra of organic compounds are associated with transitions between vibrational energy levels. Molecular vibrations may be detected and measured either in an infrared spectrum or indirectly in a Raman spectrum. The most useful vibrations, from the point of view of the organic chemist, occur in the narrower range of 3.5–16 μm (1 μm = 10−6 m). The position of an absorption band in the spectrum may be expressed in microns (μm), but standard practice uses a frequency scale in the form of wavenumbers, which are the reciprocals of the wavelength, cm−1. The useful range of the infrared for an organic chemist is between 4000 cm−1 at the high-frequency end and 600 cm−1 at the low-frequency end. Many functional groups have vibration frequencies, characteristic of that functional group, within well-defined regions of this range; these are summarised in Charts 3.1, 3.2, 3.3, and 3.4 at the end of this chapter, with more detail in the tables of data that follow. Because many functional groups can be identified by their characteristic vibration frequencies, the infrared spectrum is the simplest, most rapid, and often most reliable means for identifying the functional groups. Equations (3.1) and (3.2), which are derived from the model of a mass m vibrating at a frequency ν on the end of a spring, are useful in understanding the range of values of the vibrational frequencies of various kinds of bonds. k v= (3.1) m 1 1 1 = + (3.2) m m1 m2 © Springer Nature Switzerland AG 2019 85 I. Fleming, D. Williams, Spectroscopic Methods in Organic Chemistry, https://doi.org/10.1007/978-3-030-18252-6_3 86 3 Infrared and Raman Spectra The value k is a measure of the strength of the spring and m, m1 and m2 the masses of the objects at each end. The value of m in Eq. (3.1) is determined by the relationship in Eq. (3.2). If one of the masses (say, m1) is much larger than the other, 1/m1 is correspondingly smaller, and the relevant mass m for Eq. (3.1) is close to that of m2—making it analogous to the case where one end of the ‘spring’ is fixed. This is the situation when one of the elements is hydrogen and the other element is any of the common elements C, N, O, P or S. Other things being equal: (1) C–H bonds will have higher stretching frequencies than C–C bonds, which in turn are likely to be higher than C–halogen bonds; (2) O–H bonds will have higher stretching frequencies than O–D bonds; and (3), since k increases with increasing bond order, the relative stretching frequencies of carbon-carbon bonds lie in the order: C≡C > C=C > C–C. 3.2  reparation of Samples and Examination in an Infrared P Spectrometer Older spectrometers used a source of infrared light which had been split into two beams of equal intensity. Only one of these was passed through the sample, and the difference in intensities of the two beams was then plotted as a function of wavenumber. Using this technology, a scan typically took about 10 min as each wavelength was examined. Most spectrometers in use today use a Fourier transform method, and the spectra are called Fourier transform infrared (FT-IR) spectra. These are taken on a spectrometer with the features simplified in Fig. 3.1. A source of infrared light emitting radiation throughout the whole frequency range of the instrument, typically 4600–400 cm−1 is again divided into two beams of equal intensity. One beam, reflected from the fixed mirror, is passed through the sample typically placed onto a diamond on the outside of the instrument with a reflecting surface above it. The other beam is reflected from a moving mirror causing the beam to traverses a range of path lengths before passing through the sample. Recombination of the two beams produces an interference pattern. The sum of the interference patterns in the time domain is known as an interferogram, which looks nothing like a spectrum, but which contains information about all the frequencies absorbed by the sample. Fourier transformation of the interfero- gram, using a computer built into the instrument, converts it into a plot of absorption against wavenumber in the frequency domain. There is more about Fourier transformation in Sect. 3.11 on Raman spectra and in Sect. 4.2 on NMR spectra. There are several advantages to FT-IR over the older method, and few disadvantages. Because it is not necessary to scan each wavenumber successively, the whole spectrum is measured in a few seconds, making it easier to use to extract the spectrum of com- pounds produced only for a short period in the outflow of a chromatograph. Because it is not dependent upon a slit and a prism or grating, the resolution is higher. FT-IR is better for examining small samples, because several scans can be added together. 3.2 Preparation of Samples and Examination in an Infrared Spectrometer 87 Fig. 3.1 Schematic layout of Sample IR source an FT-IR spectrometer cell Detector Fixed mirror Computer Moving Spectrum mirror 1560 1550 820 720 CCl4 3030 2990 1260 1180 940 920 CHCl3 2210 2180 940 860 CDCl3 2400 2200 1600 1400 CS2 4000 3000 2000 1500 1000 cm–1 Fig. 3.2 Solvent absorption in the IR; dark areas are the regions in which the solvent interferes Finally, the digital form in which the data are handled in the computer allows for adjust- ment and refinement by subtracting the background absorption of the medium in which the spectrum was taken, by subtracting the unavoidable signal from the carbon dioxide in the air, or by subtracting the spectrum of a known impurity from that of a mixture to reveal the spectrum of the pure component. However, the way in which infrared spectra are taken does not significantly affect their appearance. The older spectra and FT-IR spectra look very similar, and older spectra in the literature are still valuable for com- parison. A drop of a liquid (the sample or a solution) is simply placed on the diamond on the surface of the spectrometer. Common solvents are carbon tetrachloride, chloroform, deu- terochloroform and carbon disulfide, which between them have transparency over the whole of the range (Fig. 3.2). Solids, 1 mg is plenty, may be deposited like the liquids described above, by evapora- tion from solution dropped onto the diamond, or directly as a powder, with a sacrifice, usually small, from scattering off the solid surface. Older spectrometers used sodium chlo- ride plates, between which the sample, liquid or solid, was squeezed, and through which 88 3 Infrared and Raman Spectra the beam was passed. A hydrocarbon mulling agent (Nujol, Kaydol) was used to reduce light scattering from solids. Alternatively, the solid was ground with potassium bromide and the mixture pressed into a transparent disc. You will come across spectra taken in all these ways in the older literature. Because of intermolecular interactions, band positions in solid state spectra are often different from those of the corresponding solution spectra. This is particularly true of those functional groups which take part in hydrogen bonding. The number of resolved lines is usually greater in solid state spectra, so that comparison of the spectra of, for example, synthetic and natural samples in order to determine identity is best done in the solid state. This is only true, of course, when the same crystalline modification is in use; racemic, synthetic material, for example, should be compared with enantiomerically pure, natural material in solution. 3.3 Selection Rules Infrared radiation is only absorbed by the sample when the oscillating dipole moment (from a molecular vibration) interacts with the oscillating electric vector of the infrared beam. A simple rule for deciding if this interaction (and hence absorption of light) occurs is that the dipole moment at one extreme of a vibration must be different from the dipole moment at the other extreme of the vibration. In the Raman effect a corresponding inter- action occurs only when the polarisability of the molecule changes in the course of the vibration. The different selection rules lead to the different applications of infrared and Raman spectra. The most important consequence of these selection rules is that in a molecule with a centre of symmetry those vibrations symmetrical about the centre of symmetry are in- active in the infrared but active in the Raman (see Sect. 3.11); those vibrations which are not centrosymmetric are usually active in the infrared but inactive in the Raman. This is doubly useful, for it means that the two types of spectra are complementary. Furthermore, the more easily obtained, the infrared, is the more useful, because most functional groups are not centrosymmetric. The symmetry properties of a molecule in a solid can be different from those of an isolated molecule. This can lead to the appea- rance of infrared absorption bands in a solid state spectrum which would be forbidden in solution. 3.4 The Infrared Spectrum A complex molecule has many vibrational modes which involve the whole molecule. To a good approximation, however, many of these molecular vibrations are largely associa- ted with the vibrations of individual bonds and are called localised vibrations. These localised vibrations are useful for the identification of functional groups, especially the 3.4 The Infrared Spectrum 89 stretching vibrations of O–H and N–H single bonds and all kinds of triple and double bonds, almost all of which occur with frequencies greater than 1500 cm−1. The stretching vibrations of other single bonds, most bending vibrations and the soggier vibrations of the molecule as a whole give rise to a series of absorption bands at lower energy, below 1500 cm−1, the positions of which are characteristic of that molecule. The net result is a region above 1500 cm−1 showing absorption bands assignable to a number of functional groups, and a region containing many bands, characteristic of the compound in question and no other compound, below 1500 cm−1. For obvious reasons, this is called the fingerprint region. Figure 3.3 shows a representative infrared spectrum, that of cortisone acetate 1. It shows the strong absorption from the stretching vibrations above 1500 cm−1 demonst- rating the presence of each of the functional groups: the O–H group, three different C=O groups and the weaker absorption of the C=C double bond, as well as displaying a characteristic fingerprint below 1500 cm−1. By convention absorbance is plotted downwards, opposite to the convention for ultraviolet spectra, but the maxima are still called peaks or bands. Rotational fine structure is smoothed out, and the intensity is frequently not recorded. When intensity is recorded, it is usually expressed subjectively as strong (s), medium (m) or weak (w). To obtain a high-quality spectrum, the quantity of substance is adjusted so that the strongest peaks absorb something close to 90% of the light. The scale on the abscissa is linear in frequency, but most instruments change the scale, either at 2200 cm−1 or at 2000 cm−1 to double the scale at the low-frequency end. The ordinate is linear in percent transmittance, with 100% at the top and 0% at the bottom. The regions in which the different functional groups absorb are summarised below Fig. 3.3. The stretching vibrations of single bonds to hydrogen give rise to the absorption at the high-frequency end of the spectrum as a result of the low mass of the hydrogen atom, making it easy to detect the presence of O–H and N–H groups. Since most organic ­compounds have C–H bonds, their absorption close to 3000 cm−1 is rarely useful, although C–H bonds attached to double and triple bonds can be usefully identified. Thereafter, the order of stretching frequencies follows the order: triple bonds at higher frequency than double bonds and double bonds higher than single bonds—on the whole the greater the strength of the bond between two similar atoms the higher the frequency of the vibration. Bending vibrations are of lower frequency and usually appear in the fingerprint region below 1500 cm−1, but one exception is the N–H bending vibration, which can appear in the 1600–1500 cm−1 region. Although many absorption bands are effectively associated with the vibrations of indi- vidual bonds, many vibrations are coupled vibrations of two or more components of the whole molecule. Whether localised or not, stretching vibrations are given the symbol ν, and the various bending vibrations are given the symbol δ. Coupled vibrations may be subdivided into asymmetric and symmetric stretching, and the various bending modes into scissoring, rocking, wagging and twisting, as defined for a methylene group in Fig. 3.4. A coupled asymmetric and symmetric stretching pair is found with many other groups, like 90 3 Infrared and Raman Spectra C— H C=C O— H 3 x C=O 'fingerprint' 4000 3000 2000 1500 1000 O— H C C C=O Other stretching, bending N— H C N C=N and combination bands. C— H X=Y=Z C=C The fingerprint region stretching stretching N=O stretching N— H ester C=O 1755 cm–1 bending O ketone C=O 1710 cm–1 O O O O H 3440 cm–1 broad H conjugated C=C 1615 cm–1 H H α,β-unsaturated ketone C=O 1665 cm–1 O 1 Fig. 3.3 The infrared spectrum of cortisone acetate 1 primary amines, carboxylic anhydrides, carboxylate ions and nitro groups, each of which has two equal bonds close together. 3.5 The Use of the Tables of Characteristic Group Frequencies Reference charts and tables of data are collected together at the end of this chapter for ready reference. Each of the three frequency ranges above 1500 cm−1 shown in Fig. 3.3 is expanded to give more detail in Charts 3.1, 3.2, 3.3, and 3.4 in Sec. 3.15. These charts summarise the narrower ranges within which each of the functional groups absorbs. The absorption bands which are found in the fingerprint region and which are assignable to functional groups are occasionally useful, either because they are sometimes strong bands in otherwise featureless regions or because their absence may rule out incorrect structures, but such identifications should be regarded as helpful rather than as definitive, since there 3.6 Stretching Frequencies of Single Bonds to Hydrogen 91 + + + H H H H H H H H H H H H Asymmetric Symmetric Scissoring Rocking Wagging Twisting stretching stretching δs(CH2) 720 cm–1 1305 cm–1 1300 cm–1 νas(CH2) νs(CH2) 1470 cm–1 2930 cm–1 2850 cm–1 Fig. 3.4 Coupled vibrations of a methylene group are usually many bands in this area. Tables of detailed information can be found in Sec. 3.16 at the end of this chapter, arranged by functional groups roughly in descending order of their stretching frequencies. One could deal with the spectrum of an unknown as follows. Examine each of the three main regions of the spectrum above the fingerprint region, as identified on Fig. 3.3; at this stage certain combinations of structures can be ruled out—the absence of O–H or C=O, for example—and some tentative conclusions reached. Where there is still ambiguity— which kind of carbonyl group, for example—the tables corresponding to those groups that might be present should be consulted. It is important to be sure that the bands under con- sideration are of the appropriate intensity for the structure suspected. A weak signal in the carbonyl region, for example, is not good evidence that a carbonyl group is present, since carbonyl stretching is always strong—it is more likely to be an overtone or to have been produced by an impurity. The text following this section amplifies some of the detail for each of the main functi- onal groups, and shows the appearance, sometimes characteristic, of several of the bands. Cross-reference to the tables at the end is inevitable and will need to be frequent. 3.6 Stretching Frequencies of Single Bonds to Hydrogen C–H Bonds C–H bonds do not take part in hydrogen bonding and so their position is little affected by the state of measurement or their chemical environment. C–C vibrations, which absorb in the fingerprint region, are generally weak and not practically useful. Since many organic molecules possess saturated C–H bonds, their absorption bands, stretching in the 3000–2800 cm−1 region and bending in the fingerprint region, are of little diagnostic value, but a few special structural features in saturated C–H groupings do give rise to char- acteristic absorption bands. Thus, methyl and methylene groups usually show two sharp bands from the symmetric and asymmetric stretching (Fig. 3.4), which can sometimes be picked out, but the general appearance of the accumulation of all the saturated C–H stretching vibrations often leads to broader and not fully resolved bands like those illus- trated in many of the spectra above and below. The absence of saturated C–H absorption 92 3 Infrared and Raman Spectra overtone of C=O overtone 3410 C CH of C=O H 2120 3420 3060 O N—H H 3310 2720 C H CH2 & CH3 2940-2860 Bohlmann CH2 & CH3 3000-2850 bands CH2 & CH3 3000-2780 2740 4000 3000 cm–1 4000 3000 2200 cm–1 4000 3000 cm–1 4000 3000 cm–1 H O H H O 2 OEt 3 4 H 5 NH Fig. 3.5 Some characteristic C–H absorptions in the infrared in a spectrum is, of course, diagnostic evidence for the absence of such a part structure, as in the spectrum of benzonitrile 14 (Fig. 3.8). Unsaturated and aromatic C–H stretching frequencies can be distinguished from the saturated C–H absorption, since they occur a little above 3000 cm−1 and are relatively weak, as in the spectrum of ethyl benzoate 2 (Fig. 3.5) and benzonitrile 14 (Fig. 3.8). Terminal acetylenes give rise to a characteristic strong, sharp line close to 3300 cm−1 from ≡C–H stretching, as in the spectrum of hexyne 3 (Fig. 3.5). A C–H bond antiperiplanar to a lone pair on oxygen or nitrogen is weakened, and the stretching frequency is lowered. Thus, the C–H bond of aldehydes gives rise to a relatively sharp band close to 2760 cm−1, as seen in the spectrum of heptanal 4 (Fig. 3.5). Similarly, ethers and amines show low-frequency bands in the region 2850–2700 cm−1 when the lone pair is antiperiplanar to a C–H bond, as it is in six-membered rings like that of tetrahydroisoquinoline 5, where the peak or peaks in this range are known as Bohlmann bands. O–H Bonds The value of the O–H stretching frequency has been used for many years as a test for, and measure of, the strength of hydrogen bonds. The stronger the hydrogen bond the longer the O–H bond, the lower the vibration frequency, and the broader and the more intense the absorption band. O–H bonds not involved in hydrogen bonding have a sharp band in the 3650–3590 cm−1 range, typically observed for samples in the vapour phase, in very dilute solution, or when such factors as steric hindrance limit hydrogen bonding. Frequently solution phase spectra show two bands, one sharp at high frequency for the OH groups not involved in H-bonding and another broader and at somewhat lower frequency for those that are involved in H-bonding, as seen in the somewhat hindered alcohol 6 (Fig. 3.6). Pure liquids, solids, and many solutions, on the other hand, show only the broad strong band in the 3600–3200 cm−1 range, because of exchange and because of the differ- ent degrees of hydrogen bonding present within the sample, as seen in the spectrum of the primary alcohol 7 (Fig. 3.6) and oleol 26 (Fig. 3.18). Weak intramolecular hydrogen bonds, like those in 1,2-diols for example, show a sharp band in the range 3570–3450 cm−1, the precise position being a measure of the 3.6 Stretching Frequencies of Single Bonds to Hydrogen 93 3620 free O—H 3300-2400 3330 3460 intermolecularly O H O intermolecularly O H O H-bonded H-bonded O H O O—H O H O O—H CH2 and CH3 1710 4000 3000 cm –1 4000 3000 cm –1 4000 3000 2000 1700 cm–1 O H O 6 7 O H 8 H O O Fig. 3.6 Some characteristic O–H absorptions in the infrared strength of the hydrogen bond. Strong intramolecular hydrogen bonds usually give rise to broad and strong absorption in the 3200–2500 cm−1 range. When a carbonyl group is the hydrogen bond acceptor, its characteristic stretching frequency is lowered, as seen for example in the dimeric association of most carboxylic acids. A broad absorption in the 3200–2500 cm−1 range, usually seen under and surrounding the C–H absorption, accom- panied by carbonyl absorption in the 1710–1650 cm−1 region, is highly characteristic of carboxylic acids, like hexanoic acid 8 (Fig. 3.6), or of their vinylogues, β-dicarbonyl compounds in the enol form. Distinctions can be made among the various hydrogen-bonding possibilities by testing the effect of dilution: intramolecular hydrogen bonds are unaffected, while intermolecular hydrogen bonds are broken, leading to an increase in—or the appearance of—free O–H absorption. Spectra taken of samples in the solid state almost always show only the broad strong band in the range 3400–3200 cm−1. Replacing an H with a D atom leads to absorption at a frequency 0.73 times lower. Note that the weak bands in the O–H region of the spectra of the ester 2 and the aldehyde 4 are clearly not O–H stretching, because they are too weak. S–H bond stretching is significantly lower in frequency, typically appearing as a weak and slightly broadened band near 2600 cm−1. N–H Bonds The stretching frequencies of the N–H bonds of amines are typically in the range 3500–3300 cm−1. They are less intense than those of O–H bonds but can easily be confused with hydrogen-bonded O–H stretching. Because an N–H has a weaker tendency to form a hydrogen bond, its absorption is often sharper, and can be very sharp, as in the N–H band shown by the indole N–H from tryptophan 9 (Fig. 3.7). Even in very dilute solutions N–H bonds never give rise to absorption as high in fre- quency as the free O–H near 3600 cm−1. Primary amines and amides like the amide 10 (Fig. 3.7) give rise to two bands, typically one at 3350 and the other at 3250 cm−1, because there are two stretching vibrations, one symmetrical and the other unsymmetrical, similar to those of a methylene group (Fig. 3.4). Secondary amines like morpholine 11 (and te- trahydroisoquinoline 5) only have the one band. Secondary amides in an s-trans configu- 94 3 Infrared and Raman Spectra free and CH & CH2 H H H N H N NH 3330 H-bonded indole N—H N H 3400 —NH + 3050-2100 3340 3140 3295, 3210 3 CH and CH2 CH and CH2 and 3080 4000 3000 2200 cm 4000 –1 3000 cm–1 4000 3000 cm–1 4000 3000 cm–1 CO2– O O NH3+ H 9 N 10 N 11 12 H N N O H H H Fig. 3.7 Some characteristic N–H absorptions in the infrared ration have only one band, but lactams, in which the amide group is in an s-cis configura- tion, often show several bands from various hydrogen-bonded associations, especially in the solid state, as in the spectrum of caprolactam 12 (Fig. 3.7). Amine salts and the zwit- terions of amino acids give rise to several N–H stretching bands on the low-frequency side of any C–H absorption, and sometimes reaching as low as 2000 cm−1, as in the spectrum of tryptophan 9 (Fig. 3.7). 3.7  tretching Frequencies of Triple and Cumulated S Double Bonds Terminal acetylenes absorb in the narrow range 2140–2100 cm−1, as in the spectrum of hexyne 3 in Fig. 3.5. Internal acetylenes absorb in the range 2300–2150 cm−1, with conju- gated triple bonds and enynes at the lower end of the range. Remembering Eqs. (3.1) and (3.2), we can understand how an internal acetylene will behave as though both carbons have somewhat larger masses than will the carbon of an acetylenic C–H group. The con- sequence is that internal acetylenes absorb at higher frequency, but the absorptions are often weak, because of the small change in dipole moment on stretching. A symmetrical acetylene shows no triple-bond stretch at all. Conjugation with carbonyl groups usually has little effect on the position of C≡C absorption, but substantially increases the intensity. When more than one acetylenic linkage is present, and sometimes when there is only one, there can be more absorption bands in this region than there are triple bonds to account for them. Nitriles absorb in the range 2260–2200 cm−1, but are often weak, as in the spectrum of the nitrile 13, for example, which was originally assigned the wrong structure because the weak nitrile absorption had been overlooked. Cyanohydrins are notorious in this respect, sometimes showing no nitrile stretching band at all. As usual, conjugation lowers the fre- quency and increases the intensity, as in the spectrum of benzonitrile 14. Isonitriles, nitrile 3.7 Stretching Frequencies of Triple and Cumulated Double Bonds 95 oxides, diazonium salts and thiocyanates such as 15 absorb strongly in the region 2300– 2100 cm−1 (Fig. 3.8). The stretching vibrations of the two double bonds in cumulated double-bonded systems X=Y=Z, such as those found in carbon dioxide, isocyanates, isothiocyanates, diazoalka- nes, ketenes and allenes, are coupled, with an unsymmetrical and a symmetrical pair. The former gives rise to a strong band in the range 2350–1930 cm−1 and the latter to a band in the fingerprint region, except for symmetrical systems, for which it is IR-forbidden. Carbon dioxide has a sharp band at 2349 cm−1. Allenes such as dimethylallene 17 give a sharp absorption in the characteristically nar- row range 1960–1930 cm−1. The other cumulated double-bond systems come in between, with isocyanates and isothiocyanates, such as cyclohexylisothiocyanate 16, giving rise to a broad band in contrast to the narrow band seen for thiocyanates 15 and allenes 17 (Fig. 3.9). H2O in C N KBr disk 2250 C N SC N 2225 2150 4000 3000 2000 cm–1 4000 3000 2000 cm–1 4000 3000 2000 cm–1 OMe CN 13 14 15 SCN O CN O Fig. 3.8 Unusually weak saturated and strong conjugated nitrile absorption NCS 17 16 C=C=C N=C=S 1960 2110 4000 3000 2000 cm–1 4000 3000 2000 cm–1 Fig. 3.9 Characteristic absorptions of an isothiocyanate and an allene 96 3 Infrared and Raman Spectra 3.8 Stretching Frequencies in the Double-Bond Region C=O Double Bonds Identifying which of the several kinds of carbonyl group is present in a molecule is one of the most important uses of infrared spectroscopy. Carbonyl bands are always strong, with carboxylic acids generally stronger than esters, and esters stronger than ketones or aldehydes. Amide absorption is usually similar in intensity to that of ketones but is subject to greater variations. The precise position of carbonyl absorption is governed by the electronic structure and the extent to which the carbonyl group is involved in hydrogen bonding. The general trends of structural variation on the position of C=O stretching frequencies are summarised in Fig. 3.10. It is helpful to use the stretching vibration of a saturated ketone at 1710 cm−1 as a reference point, and to compare the consequence of all the variations in structure with that number. 1. The more electronegative the group X in the system R–C(=O)–X, the higher is the frequency, except that this trend from the inductive effect of X is offset by the effect in the π system of any lone pairs on X. Thus, the inductive effect raises the frequency of the absorption. The C=O stretching frequency falls in the order: acid fluorides, chlorides, bromides, esters and amides, as the electronegativity of the attached atom falls. In the opposite direction, an electropositive substituent like a silyl group lowers the frequency. However, ketones have their C=O stretching fre- quencies between those of comparable esters and amides, and so the electronegativity of X is not the only factor. 2. The overlap of the lone pair of electrons on X with the C=O bond, illustrated by the curly arrows in the amide in Fig. 3.10, reduces the double-bond character of the C=O bond, while increasing the double-bond character of the C–N bond. In molecular orbital terms, the total π bonding is increased by the overlap, lowering the π energy, but the extra bonding is shared over three atoms, instead of being isolated on the two atoms of the C=O group. This overlap will have most effect when the lone pair is relatively high in energy. The less electronegative X is, the higher is the energy of its lone-pair orbitals, and the more effective the overlap in reducing the π-bonding in the C=O bond. The net result is to lower the frequency of the C=O stretching vibration in amides below that of ketones, even though N is more electronegative than C. On the other hand, the higher electronegativity of O, coupled with the less-effective π-overlap of its lone pairs, keeps the stretching O O O O O O O O O O O Cl OR H R OH NR2 SiR3 O– anhydrides acid chlorides esters aldehydes ketones acids amides acylsilanes carboxylate ions 1820 & 1760 1800 1740 1730 1710 1710 1660 1640 1580 Fig. 3.10 Representative stretching frequencies (in cm−1) of C=O groups 3.8 Stretching Frequencies in the Double-Bond Region 97 frequency of esters above that of ketones. Similarly, the hyperconjugative overlap of the neighbouring C–H (or C–C) bonds in a ketone reduces the C=O double-bond character, and lowers its stretching frequency relative to that of an al- dehyde. On the other hand, if the oxygen lone pair in an ester overlaps with another double bond, it is less effective in lowering the frequency of the carbonyl vibration, with the result that vinyl and phenyl esters absorb at higher frequency than alkyl esters (see vinyl acetate 21 in Fig. 3.13), and carboxylic anhydrides even more so. Anhydrides also show two bands rather than one, because the two carbonyl groups have an unsymmetrical and a symmetrical vibration, with the former pushed a little above that of a comparable acid chloride and the latter rather more below it. 3. Hydrogen bonding to a carbonyl group causes a shift to lower frequency of 40–60 cm−1. Acids, amides with an N–H, enolised β-dicarbonyl systems, and o-hydroxy- and o-aminophenyl carbonyl compounds show this effect, as illustrated by the carboxylic acid’s coming in the list in Fig. 3.10 between the ester and the amide, instead of being the same as the ester. All carbonyl compounds tend to give slightly lower values for the carbonyl stretching frequency in the solid state compared with the value for dilute solutions. 4. Ring strain in cyclic compounds causes a relatively large shift to higher frequency (Fig. 3.11). This phenomenon provides a remarkably reliable test of ring size, dis- tinguishing clearly between four-, five- and larger-membered ring ketones, lactones and lactams. Six-ring and larger ketones, lactones and lactams, show the normal frequency found for the open-chain compounds. 5. α,β-Unsaturation causes a lowering of frequency of 15–40 cm−1 because overlap of the π molecular orbitals over four atoms, while lowering the energy overall, re- duces the double-bond character of the C=O bond itself. The unsaturated ketone group in cortisone acetate 1 can be seen as having the lowest frequency of the three carbonyl peaks in Fig. 3.3, with the saturated ketone just above it and the ester just above that. The same effect can be seen in the spectrum of ethyl crotonate 20 in Fig. 3.13. The effect of α,β-unsaturation is noticeably less in amides, where little shift is observed, and that sometimes even to higher frequency. O O O O O X X X X etc. 1813 X=CH2 1775 X=CH2 1750 X=CH2 1715 X=CH2 1710 X=O 1841 X=O 1774 X=O 1750 X=O 1727 X=NH 1750 X=NH 1717 X=NH 1673 X=NH 1669 Fig. 3.11 The effect of ring size on C=O stretching frequencies (in cm−1) 98 3 Infrared and Raman Spectra amide I amide II amide I amide II C=O C=O 1690 1610 1660 1560 1645 1660 2000 1500 cm–1 2000 1500 cm–1 2000 1500 cm–1 2000 1500 cm–1 O O 10 18 19 O 12 NH2 N N N O H H Fig. 3.12 Amide C=O stretching vibrations 6. Where more than one of the structural influences on a particular carbonyl group is operating, the net effect is usually close to additive, with α,β-unsaturation and ring strain having opposite effects. 7. Amides are special in showing greater variation in the number and position of the bands in the carbonyl region. In particular, they show extra bands in addition to the localised stretching of the C=O bond. Thus, primary and secondary amides, which also have an N–H bond, show at least two bands. The one at higher fre- quency is called the amide I band, and the other, at lower frequency is called the amide II band. The spectra of the primary amide 10 and of N-methylacetamide 18 in Fig. 3.12 show the amide I and II bands, whereas the spectra of N,N-dimethylacetamide 19 and caprolactam 12 show a single peak. Both bands are affected by hydrogen bonding and are therefore significantly different when the spectra are taken of solutions or of the solid amide. For the full range of all types of carbonyl absorption, see the tables at the end of this chapter. C=N Double Bonds Imines, oximes and other C=N double bonds absorb in the range 1690–1630 cm−1. They are weaker than carbonyl absorption, and are difficult to identify because they absorb in the same region as C=C double bonds. C=C Double Bond Unconjugated alkenes absorb in the range 1680–1620 cm−1. The more substituted C=C double bonds absorb at the high-frequency end of the range, the less substituted at the low-frequency end. The intensity is generally less than that for the C=O stretching of carbonyl groups. The absorption may be very weak when the double bond is more or less symmetrically substituted, and absent when the double bond is symmetrically substituted. Oleol 29 has a cis double bond near the middle of a long chain. Since it is lo- cally almost symmetrically substituted, its stretching is only just detectable, too weakly to be diagnostic in the infrared spectrum (Fig. 3.19). 3.8 Stretching Frequencies in the Double-Bond Region 99 O O OEt H O C=C C=C 20 C=O 1720 1665 975 21 C=O 1762 1645 H 4000 3000 2000 1500 1000 cm–1 4000 3000 2000 1500 1000 cm–1 Fig. 3.13 Conjugated C=C and C=O stretching vibrations Conjugation lowers the stretching frequency of C=C double bonds only a little, as seen in the spectrum of ethyl crotonate 20 in Fig. 3.13, where it gives rise to the moderately strong peak at 1665 cm−1, which is nevertheless weaker than the carbonyl peak at 1720 cm−1. Vinyl esters like vinyl acetate 21, where the C=C double bond is conjugated to an oxygen lone pair, also give rise to a strong band at the low end of the C=C range. The C=C stretching vibration can also be seen as the small peak at 1615 cm−1 on the low-fre- quency side of the three carbonyl peaks in Fig. 3.3, and as the small peak at 1640 cm−1 between the amide I and amide II bands of the amide 10 in Fig. 3.12. If the double bond is exocyclic to a ring the frequency rises as the ring size decreases. A double bond within a ring shows the opposite trend: the frequency falls as the ring size decreases. N=O Double Bonds Nitro groups have asymmetrical and symmetrical N=O stretching vibrations, the former a strong band in the range 1570–1540 cm−1 and the latter a strong band in the range 1390–1340 cm−1 in the fingerprint region. They can be seen easily as the strongest peaks when they are conjugated with the benzene ring as they are in o-­nitrobenzyl alcohol 22, but they are weaker, although still evident, when they are not conjugated, as in ω-nitroacetophenone 23 in Fig. 3.14. The relative weakness of the bands when they are not Ar-H conjugated 1595 & 1580 C=O 1705 – – – – O O O O O O O O OH +N +N +N +N 1525 1350 1570 1380 4000 3000 2000 1500 1000 cm–1 4000 3000 2000 1500 1000 cm–1 O O O N N 22 O 23 O OH Fig. 3.14 Asymmetrical and symmetrical stretching absorption of nitro groups 100 3 Infrared and Raman Spectra conjugated makes it uncertain in 23 whether the symmetric stretching is the band at 1380 or the stronger band at 1330 cm−1. If the assignment in Fig. 3.14 is correct, conjugation with the benzene ring has lowered the N=O stretching bands in the nitrobenzene 22 by 45 and 30 cm−1, as well as increasing their intensity. Nitrates, nitramines, nitrites and nitroso compounds also absorb in the 1650–1400 cm−1 region. 3.9 Characteristic Vibrations of Aromatic Rings There are several features in infrared spectra that combine to give evidence of the presence of an aromatic ring. The C–H stretch in aromatic rings can almost always be detected just above 3000 cm−1 as a weak peak or peaks, but this does not distinguish them from the corresponding alkene signals. More characteristic are the two or three bands in the 1600– 1500 cm−1 region shown by most six-membered aromatic rings such as benzenes, polycy- clic aromatic rings and pyridines. Typically, a benzene ring conjugated to a double bond has three bands like those marked in the spectrum of ethyl benzoate 2 in Fig. 3.15, usually close to 1600, 1580 and 1500 cm−1. The two bands at higher frequency are usually the more intense when the benzene ring is further conjugated, but they are usually weak when the benzene ring is not conjugated, as seen in the infrared spectrum of the aromatic ring in tetrahydroisoquinoline 5 in Fig. 3.15. The relative intensities vary in other less predictable ways, and sometimes the two higher frequency bands appear as one. There are bands in the fingerprint region in the 1225–950 cm−1 range which are of little use in providing structural information. The shape and number of the two to six overtone and combination bands which appear in the 2000–1600 cm−1 region is a function of the substitution pattern of the benzene ring, but the use of this region for this purpose has been overtaken by NMR spectroscopy, which does the job much better. A fourth group of strong bands below 900 cm−1 produced by the out-of-plane C–H bending vibrations is affected by the number of adjacent hydrogen atoms on the ring. It too can be used, but rarely is nowadays, to identify the substitution pattern. aryl C-H NH benzene ring aryl C-H 1605, 1590 and 1495 benzene ring C=O 1720 1595, 1580 & 1485 4000 3000 2000 1500 1000 cm–1 4000 3000 2000 1500 1000 cm–1 O 2 OEt 5 NH Fig. 3.15 Characteristic vibrations of aromatic rings 3.11 Raman Spectra 101 3.10 Groups Absorbing in the Fingerprint Region Strong bands in the fingerprint region arise from the stretching vibrations of a few other doubly bonded functional groups like sulfonyl, thiocarbonyl and phosphoryl, but they are easily confused with the stretching vibrations of single bonds like C–O and C–halogen, which are always strong, and the bending vibrations of C–H bonds. They are rarely diagnostically useful, unless their absence is informative. The last few tables at the end of this chapter give an indication where some of these absorptions commonly occur, and include some of the less common and more specialised functional groups, like those of boron, silicon, phosphorus and sulfur. A few of these bands are noteworthy. The methine C–H bending and methyl CH3 and methylene CH2 symmetrical bending vibrations give rise in many aliphatic compounds to two broad bands close to 1450 and 1380 cm−1, which can clearly be seen in the spectrum of oleol 29 in Fig. 3.19. The out-of-plane vibration of trans –CH=CH– double bonds is one of the more usefully diagnostic bending vibrations. It occurs in a narrow range 970–960 cm−1, or at slightly higher frequency if conjugated, and it is always strong. In contrast, the corresponding vibration of the cis isomer is of lower intensity and at lower frequency, typically in the range 730–675 cm−1. The band at 975 cm−1 in the fingerprint of ethyl trans-crotonate 20 (Fig. 3.13) clearly shows that such a feature may be present; if there were no band there, it would be diagnostic of the absence of this feature, as in the spectrum of the cis-alkene oleol 26 in Fig. 3.18. The N–H bending vibration of primary and secondary amides can appear just above the fingerprint region, and may be responsible for what is called the amide-II band. 3.11 Raman Spectra Raman spectra are generally taken on instruments using laser sources, and the quantity of material needed is of the order of a few mg. The sample may be a gas, liquid or solid, and may be in solution if the solvent does not interfere with the signals of interest. The sample is irradiated with a laser beam of monochromatic light, and the scattered light is examined using photoelectric detection. The sample does not even need to be taken out of its bottle or ampoule if the container is