Interpretation of Infrared Spectra, A Practical...

Interpretation of Infrared Spectra, A Practical Approach

John Coates

inEncyclopedia of Analytical Chemistry

R.A. Meyers (Ed.)pp. 10815–10837

John Wiley & Sons Ltd, Chichester, 2000

INTERPRETATION OF INFRARED SPECTRA, A PRACTICAL APPROACH 1

Interpretation of InfraredSpectra, A Practical Approach

John CoatesCoates Consulting, Newtown, USA

1 Introduction 1

2 The Origins of the Infrared Spectrum 2

3 Spectral Interpretation by Application ofVibrational Group Frequencies 63.1 The Hydrocarbon Species and

Molecular Backbone 63.2 Simple Functional Groups 93.3 The Carbonyl Group 123.4 Other Functional Groups Associated

with Heteroatoms 133.5 Simple Inorganics 14

4 The Practical Situation – Obtaining theSpectrum and Interpreting the Results 154.1 Sample History 164.2 Physical Characteristics of the

Sample 174.3 The Chemistry of the Sample 174.4 The Infrared Sampling Method 18

5 An Overview to Infrared SpectralInterpretation – Some Simple Rules andGuidelines 195.1 A Quick Diagnostic Assessment of

an Infrared Spectrum 20

Abbreviations and Acronyms 22

Related Articles 22

References 23

The vibrational spectrum of a molecule is considered tobe a unique physical property and is characteristic of themolecule. As such, the infrared spectrum can be used asa fingerprint for identification by the comparison of thespectrum from an ‘‘unknown’’ with previously recordedreference spectra. This is the basis of computer-basedspectral searching. In the absence of a suitable referencedatabase, it is possible to effect a basic interpretation of thespectrum from first principles, leading to characterization,and possibly even identification of an unknown sample.This first principles approach is based on the fact thatstructural features of the molecule, whether they are thebackbone of the molecule or the functional groups attachedto the molecule, produce characteristic and reproducible

absorptions in the spectrum. This information can indicatewhether there is backbone to the structure and, if so,whether the backbone consists of linear or branched chains.Next it is possible to determine if there is unsaturationand/or aromatic rings in the structure. Finally, it is possibleto deduce whether specific functional groups are present.If detected, one is also able to determine local orientationof the group and its local environment and/or location inthe structure. The origins of the sample, its prehistory, andthe manner in which the sample is handled all have impacton the final result. Basic rules of interpretation exist and,if followed, a simple, first-pass interpretation leading tomaterial characterization is possible. This article addressesthese issues in a simple, logical fashion. Practical examplesare included to help guide the reader through the basicconcepts of infrared spectral interpretation.

1 INTRODUCTION

The qualitative aspects of infrared spectroscopy are oneof the most powerful attributes of this diverse andversatile analytical technique. Over the years, much hasbeen published in terms of the fundamental absorptionfrequencies (also known as group frequencies) which arethe key to unlocking the structure–spectral relationshipsof the associated molecular vibrations. Applying thisknowledge at the practical routine level tends to bea mixture of art and science. While many purists willargue against this statement, this author believes that itis not possible to teach a person to become proficient asan interpretive spectroscopist by merely presenting theknown relationships between structure and the observedspectra. Instead, the practical approach, which has beenadopted in this text, is to help the reader appreciate thevisual aspects of the spectroscopy and how to interpretthese relative to the structure and chemistry of the sample.This is achieved by recognizing characteristic shapesand patterns within the spectrum, and by applying theinformation obtained from published group frequencydata, along with other chemical and physical data fromthe sample.

Included in the text is a discussion of the interrelation-ships that exist between the practical side of acquiringthe spectrum, the chemistry and physics of the sampleunder study, the physical interactions of the sample withits environment, and the impact of the structure on thespectrum. In essence, the interpretation of infrared spec-tra is much more than simply assigning group frequencies.The spectrum is rich in information, and this article isintended to help the reader to extract the maximumusing the knowledge available for the sample and theacquired spectral data. One important factor to bear in

Encyclopedia of Analytical ChemistryR.A. Meyers (Ed.) Copyright John Wiley & Sons Ltd

2 INFRARED SPECTROSCOPY

mind is that a successful interpretation is based not onlyon the presence of particular bands within the spectrum,but also the absence of other important bands. Completeclasses of compounds can be rapidly excluded during theinterpretation by the use of no-band information.

It must be understood that this article addressesthe issue of infrared spectral interpretation from theperspective of the average operator of an infraredinstrument. It is not a detailed treatise on the theoryof infrared spectroscopy where the modes of vibrationare discussed in terms of group theory, and wheremathematical models are used to compare theoreticaland observed values for the fundamental vibrations of amolecule. There are many excellent texts that cover thissubject..1 – 4/ Instead, this article focuses on the day-to-day problems associated with characterizing a material orattempting to perform some form of identification. Oneof the main challenges in presenting a text on spectralinterpretation is to form a balance between the theorythat is needed to appreciate the links between molecularstructure and the observed spectrum and the practice.For this reason, a minimum amount of relevant theoryis included in the next section, which provides a basicunderstanding of why the spectrum exists, how it isformed, and what factors contribute to the complexityof observed spectra. It has been assumed that the readerhas a fundamental knowledge of molecular theory andbonding, and that there is an understanding of basicstructures, in particular for organic compounds.

Infrared spectral interpretation may be applied toboth organic and inorganic compounds, and there aremany specialized texts dealing with these compounds, incombination and as individual specialized texts. Thereare too many to reference comprehensively, and thereader is directed to a publication that provides abibliography of the most important reference texts..5/

However, the most informative general reference textsare included,.6 – 14/ with books by Socrates.10/ and Lin-Vien.11/ being recommended for general organics, andby Nakamoto.13/ and Nyquist et al..14/ for inorganics(salts and coordination compounds). There are numerousspecialized texts dealing with specific classes of materials,and undoubtedly polymers and plastics form the largestindividual class..15 – 17/ In this particular case, texts byHummel and Scholl.16/ and Koenig.17/ provide a goodbasic understanding.

The following comments are made relative to the con-ventions used within this article. The term frequencyis used for band/peak position throughout, and this isexpressed in the commonly used units of wavenum-ber (cm�1). The average modern infrared instrumentrecords spectra from an upper limit of around 4000 cm�1

(by convention) down to 400 cm�1 as defined by theoptics of the instrument (commonly based on potassium

bromide, KBr). For this reason, when a spectral region isquoted in the text, the higher value will be quoted first,consistent with the normal left-to-right (high to low cm�1)representation of spectra. Also, the terms infrared band,peak and absorption will be used interchangeably withinthe text to refer to a characteristic spectral feature.

The spectral group frequencies provided in this textwere obtained from various literature sources publishedover the past 30 years, and most of these are includedin the cited literature. Every attempt to ensure accuracyhas been taken; however, there will be instances whenindividual functional groups may fall outside the quotedranges. This is to be expected for several reasons: theinfluences of other functional groups within a molecule,the impact of preferred spatial orientations, and environ-mental effects (chemical and physical interactions) on themolecule.

The preferred format for presenting spectral data forqualitative analysis is in the percentage transmittanceformat, which has a logarithmic relationship (� log10) withrespect to the linear concentration format (absorbance).This format, which is the natural output of mostinstruments (after background ratio), provides the bestdynamic range for both weak and intense bands. In thiscase, the peak maximum is actually represented as aminimum, and is the point of lowest transmittance for aparticular band.

2 THE ORIGINS OF THE INFRAREDSPECTRUM

In the most basic terms, the infrared spectrum is formedas a consequence of the absorption of electromagneticradiation at frequencies that correlate to the vibration ofspecific sets of chemical bonds from within a molecule.First, it is important to reflect on the distribution of energypossessed by a molecule at any given moment, defined asthe sum of the contributing energy terms (Equation 1):

Etotal D Eelectronic C Evibrational C Erotational C Etranslational

.1/The translational energy relates to the displacement ofmolecules in space as a function of the normal thermalmotions of matter. Rotational energy, which gives riseto its own form of spectroscopy, is observed as thetumbling motion of a molecule, which is the result ofthe absorption of energy within the microwave region.The vibrational energy component is a higher energyterm and corresponds to the absorption of energy by amolecule as the component atoms vibrate about the meancenter of their chemical bonds. The electronic componentis linked to the energy transitions of electrons as they


are distributed throughout the molecule, either localizedwithin specific bonds, or delocalized over structures, suchas an aromatic ring. In order to observe such electronictransitions, it is necessary to apply energy in the form ofvisible and ultraviolet radiation (Equation 2):

E D hn frequency/energy .2/

The fundamental requirement for infrared activity, lead-ing to absorption of infrared radiation, is that there mustbe a net change in dipole moment during the vibrationfor the molecule or the functional group under study.Another important form of vibrational spectroscopy isRaman spectroscopy, which is complementary to infraredspectroscopy. The selection rules for Raman spectroscopyare different to those for infrared spectroscopy, and inthis case a net change in bond polarizability must beobserved for a transition to be Raman active. The remain-ing theoretical discussion in this article will be limited to avery simple model for the infrared spectrum. The readeris encouraged to refer to more complete texts.2 – 4/ fordetailed discussion of the fundamentals.

While it was stated that the fundamental infraredabsorption frequencies are not the only component to beevaluated in a spectral interpretation, they are the essenceand foundation of the art. For the most part, the basicmodel of the simple harmonic oscillator and its modifica-tion to account for anharmonicity suffice to explain theorigin of many of the characteristic frequencies that canbe assigned to particular combinations of atoms withina molecule. From a simple statement of Hooke’s law wecan express the fundamental vibrational frequency of amolecular ensemble according to Equation (3):

n D 12pc

√k

µ.3/

where n D fundamental vibration frequency, k D forceconstant, and µ D reduced mass. The reduced mass, µ Dm1m2/.m1 Cm2/, where m1 and m2 are the componentmasses for the chemical bond under consideration.

This simple equation provides a link between thestrength (or springiness) of the covalent bond betweentwo atoms (or molecular fragments), the mass of the inter-acting atoms (molecular fragments) and the frequency ofvibration. Although simple in concept, there is a rea-sonably good fit between the bond stretching vibrationspredicted and the values observed for the fundamentals.

This simple model does not account for repulsion andattraction of the electron cloud at the extremes of thevibration, and does not accommodate the concept ofbond dissociation at high levels of absorbed energy. Amodel incorporating anharmonicity terms is commonlyused to interpret the deviations from ideality and theoverall energy–spatial relationship during the vibration

of a bond between two atomic centers. The fundamental,which involves an energy transition between the groundstate and the first vibrational quantum level, is essen-tially unaffected by the anharmonicity terms. However,transitions that extend beyond the first quantum level(to the second, third, fourth, etc.), which give rise toweaker absorptions, known as overtones, are influencedby anharmonicity, which must be taken into accountwhen assessing the frequency of these higher frequencyvibrations.

Having defined the basis for the simple vibration ofan atomic bond, it is necessary to look at the moleculeas a whole. It is very easy to imagine that there is aninfinite number of vibrations, which in reality wouldlead to a totally disorganized model for interpretation.Instead, we describe the model in terms of a minimumset of fundamental vibrations, based on a threefold set ofcoordinate axes, which are known as the normal modesof vibration. All the possible variants of the vibrationalmotions of the molecule can be reduced to this minimumset by projection on to the threefold axes. It can be shownthat the number of normal modes of vibration for a givenmolecule can be determined from Equations (4) and (5):

number of normal modes D 3N � 6 (nonlinear) .4/

D 3N � 5 (linear) .5/

where N is the number of component atoms in themolecule.

In practice, apart from the simplest of compounds, mostmolecules have nonlinear structures, except where a spe-cific functional group or groups generate a predominantlinear arrangement to the component atoms. If we calcu-late the number of modes for a simple hydrocarbon, suchas methane (nonlinear, tetrahedral structure), a value ofnine is obtained. This would imply that nine sets of absorp-tion frequencies would be observed in the spectrum ofmethane gas. In reality, the number observed is far less,corresponding to the asymmetric and symmetric stretch-ing and bending of the C�H bonds about the centralcarbon atom. The reason for the smaller than expectednumber is that several of the vibrations are redundant ordegenerate, that is, the same amount of energy is requiredfor these vibrations. Note that although a small numberof vibrational modes is predicted, and in fact observed,the appearance of the methane spectrum at first glanceis far more complex than expected, especially at higherspectral resolutions (<1 cm�1). At relatively high resolu-tions, a fine structure is superimposed, originating fromrotational bands, which involve significantly lower energytransitions. Each of the sets of vibrational–rotationalabsorptions manifest this superimposed fine structurefor low-molecular-weight gaseous compounds, methanebeing a good example. Several medium-molecular-weight


compounds may also show evidence of some fine struc-ture when studied in the vapor state. For example, it iscommon to observe the sharp feature (or spike) assignedto the Q-branch of the vibrational–rotational spectrum,as indicated by the vapor spectrum of acetone (Figure 1).

If we proceed up the homologous series from methane(CH4) to n-hexane (C6H14), there are 20 componentatoms, which would imply 54 normal modes. In this casethe picture is slightly more complex. Methane is a uniquemolecule, and only contains one type of C�H group – noother types of bond exist in this molecule. In hexane thereare several types of bond and functionality. For reference,a simple two-dimensional representation of the structureis provided in Figure 2(a).

As we can see, there are two terminal methyl groups(CH3) and four connecting methylene groups (CH2).Each of these groups has its corresponding C�Hstretching and bending vibrations (see later text for theactual absorption frequencies). Also, the methyl groupsare linked to a neighboring methylene group, which isin turn linked to neighboring methylene groups, and soon. These linkages feature carbon–carbon bonds. Forinterpretation, we view the C�H groups as functionalgroups, giving rise to the common group frequencies,and the C�C linkages as the backbone, producing theskeletal vibrations. As a rule, a group frequency maybe applied generally to most compounds featuring thecorresponding functional group. In contrast, the skeletalvibrations are unique to a specific molecule. The groupfrequencies help to characterize a compound, and the

100

80

60

40

20

0

4000

3600

3200

2800

2400

2000

1800

1600

1400

1200

1000 80

0

600

400

Tran

smitt

ance

(%

)

Q

Q

Q

Q

Wavenumber (cm−1)

Figure 1 Vapor spectrum of acetone with characteristicQ-branch slitting, denoted by Q. Copyright Coates Consulting.

H CH

H CC

CC

CH

HHHH

HH H HH

H

(a) (b)

CC

CC H

H

HHH

HH

CH

HH

H

CHH

H

Figure 2 Structures for hexane isomers: (a) n-hexane and(b) isohexane (2-methylpentane). Copyright Coates Consulting.

combination of the bands associated with these groupfrequencies and the skeletal frequencies are used toidentify a specific compound. The latter forms the basis ofthe use of reference spectra for spectral matching by visualcomparison or by computer-based searching, for theidentification of an unknown from its infrared spectrum.

The group frequencies may be viewed quantitatively,as well as qualitatively. A given absorption band assignedto a functional group increases proportionately with thenumber times that functional group occurs within themolecule. For example, in the spectrum of n-hexane, theintensities measured for the group frequency absorptionsassigned to methyl and methylene correspond to fourmethylene groups and two methyl groups on a rel-ative basis, when compared with other hydrocarboncompounds within a homologous series. For example,if we examine the C�H stretching (or bending) bandintensities for CH3 and CH2, we will observe that therelative intensities of CH3 to CH2 decrease with increasein chain length. Restated, there is less methyl contribu-tion and more methylene contribution with increase inchain length/molecular weight. The reverse holds true ifwe examine the spectra of linear hydrocarbons with chainlengths shorter than that of hexane.

If we apply these ideas to a different hexane iso-mer, such as isohexane (2-methylpentane), we wouldsee significant differences in the spectrum. These can beexplained by evaluating the structure (Figure 2b), whichcontains three methyl groups, two methylene groups, anda group that contains a single hydrogen attached to car-bon (the methyne group). This adds a new complexityto the spectrum: the main absorptions show differencesin appearance, caused by the changes in relative bandintensities, splittings of absorptions occur (originatingfrom spatial/mechanical interaction of adjacent methylgroups), and changes are observed in the distributions ofthe C�C skeletal vibrations, in part due to the splitting bythe methyl side chain. Further discussions concerning theimpact of chain branching are covered later in this article.Comparison of Figures 3 and 4 provides a graphical repre-sentation of the aspects discussed for the hexanes of struc-turally similar compounds, i.e. n-heptane and isooctane.

From a first-order perspective, the idea of the quan-titative aspects of the group frequencies carries throughfor most functional groups, and the overall spectrum isessentially a composite of the group frequencies, withband intensities in part related to the contribution of eachfunctional group in the molecule. This assumes that thefunctional group does give rise to infrared absorptionfrequencies (most do), and it is understood that eachgroup has its own unique contribution based on its extinc-tion coefficient (or infrared absorption cross-section).

Returning to the fundamental model, we should nowlook at the larger picture. In reality, we assign the


00

3600

3200

2800

2400

2000

1800

1600

1400

1200

1000 80

0

600

100

80

60

40

20

0

Wavenumber (cm−1)

Tran

smitt

ance

(%

)

CH3 (CH2)5 CH3

Figure 3 Attenuated total reflectance (ATR) spectrum ofn-heptane. Copyright Coates Consulting.

CH3 C

CH3

CH3

CH2 CH CH3

CH3

4000

3600

3200

2800

2400

2000

1800

1600

1400

1200

1000 80

0

600

100

80

60

40

20

0

Wavenumber (cm−1)

Tran

smitt

ance

(%

)

Figure 4 ATR spectrum of 2,2,4-trimethylpentane (isooctane).Copyright Coates Consulting.

observed absorption frequencies in the infrared spectrumto much more that just simple harmonic (or anharmonic)stretching vibrations. In practice, we find that variousother deformation motions (angular changes), such asbending and twisting about certain centers within amolecule, also have impact, and contribute to the overallabsorption spectrum. By rationalizing the effort needed tomove the atoms relative to each other, one can appreciatethat it takes less energy to bend a bond than to stretchit. Consequently, we can readily accept the notion thatthe stretching absorptions of a vibrating chemical bondoccur at higher frequencies (wavenumbers) than thecorresponding bending or bond deformation vibrations,with the understanding, of course, that energy andfrequency are proportionally related. A good example isthe C�H set of vibrations, observed in the hydrocarbonspectra, and in virtually all organic compounds. Here, thesimple C�H stretching vibrations for saturated aliphaticspecies occur between 3000 and 2800 cm�1, and thecorresponding simple bending vibrations nominally occurbetween 1500 and 1300 cm�1.

Next in our understanding is that it can take slightlymore energy to excite a molecule to an asymmetric than a

symmetric vibration. While this might be less intuitive, itis still a rational concept, and therefore easy to understandand accept. Again, we see a good example with the C�Hstretch of an aliphatic compound (or fragment), wherewe observe the asymmetric C�H stretch of the methyland methylene groups (2960 and 2930 cm�1, respectively)occurring at slightly higher frequency than symmetricvibrations (2875 and 2855 cm�1, respectively for methyland methylene). For the most part, this simple ruleholds true for most common sets of vibrations. Naturallythere are always exceptions, and a breakdown of therationale may occur when other effects come into play,such as induced electronic, spatial or entropy-relatedeffects.

There are many other spatially related scenarios thattend to follow well-orchestrated patterns, examples beingin-plane and out-of-plane vibrations, the differencesbetween cis and trans spatial relationships, and a varietyof multicentered vibrations that are defined as twistingor rocking modes. Many of these are exhibited with theC�H vibrations that occur in saturated, unsaturated andaromatic compounds. Molecular symmetry of the staticor the dynamic (during vibration) molecule has a largeimpact on the spectrum, in addition to factors such asrelative electronegativity, bond order and relative massof the participating atoms.

Finally, while discussing the vibrational origins ofinfrared spectra, it is worth commenting that furthercomplexity may be noted in the spectrum, beyondwhat is expected based on the fundamentals. As noted,transitions to higher energy levels, although theoreti-cally not allowed, can occur and these give rise toovertone bands, which in the mid-infrared region occurat approximately twice the fundamental frequency forthe first overtone. Higher overtones exist, typically thesecond (3ð fundamental) and third (4ð fundamental),and sometimes higher, and these are observed, withextremely low intensity, relative to the fundamental in thenear-infrared spectral regions, between 800 and 2500 nm(12 500 and 4000 cm�1). Other types of bands that can addcomplexity to a spectrum are combination bands (sum anddifference), bands due to transitions from energy stateshigher than the ground state or ‘‘hot bands’’, and bandsdue to interactions between a weaker overtone or com-bination band and a fundamental of the same or similarfrequency, known as Fermi resonance bands. In the lat-ter case, two relatively strong absorptions are observed,where normally only a single absorption is expected forthe fundamental..3/

As additional functional groups are added to a basicbackbone structure, forming a more complex molecule,additional bands are observed, either directly associatedwith the fundamental vibrations of the functional groups,or indirectly related to interactions between component


functional groups or the basic substructure. Such interac-tions can be severe, and result in overwhelming distortionsin the appearance of the spectrum, a good examplebeing hydrogen bonding. This will be dealt with in depthlater.

3 SPECTRAL INTERPRETATION BYAPPLICATION OF VIBRATIONALGROUP FREQUENCIES

This section includes tabulated data relative to themost significant group frequencies for the most commonfunctional groups and structural components found inorganic compounds. Brief reference is also made to simpleinorganic compounds, in the form of simple ionic species.More detailed listings can be found in published literature,and the reader is encouraged to acquire one or more ofthese reference texts..9 – 13/ As already indicated, the useof tabulated data is only a part of the interpretationprocess, and other facets of the spectrum must be takeninto account.

To help gain an understanding of infrared spectralinterpretation, it is instructive to start at the root of mostorganic compounds, namely the fundamental backboneor the parent hydrocarbon structure. We shall start withthe simple, aliphatic hydrocarbon, which is at the rootof most aliphatic compounds. Aliphatic hydrocarbonsexist in simple linear chains, branched chains and incyclic structures – examples of the linear and branchedchain scenarios were provided earlier for hexane isomers.Any one molecule may exist with one or more of thesecomponent structures. The infrared spectrum can provideinformation on the existence of most of these structures,either directly or by inference.

The introduction of unsaturation in the form of adouble or triple bond has a profound impact on thechemistry of the molecule, and likewise it has a significantinfluence on the infrared spectrum. Similarly, the same isobserved when an aromatic structure is present withina molecule. Infrared spectroscopy is a powerful toolfor identifying the presence of these functionalities. Itprovides information specific to the group itself, andalso on the interaction of the group with other partsof the molecule and on the spatial properties of thegroup. Examples of these include conjugation between adouble bond and another unsaturated center, an aromaticring or a group, such as a carbonyl (CDO), and theorientation or location of the double bond within themolecule, such as cis or trans and medial or terminal. Itshould be noted that cis/trans relationships are not specificto unsaturated hydrocarbons, and the terminology isreferenced elsewhere, such as with secondary amidestructures. Again, the associated changes in the spatial

arrangement of the groups involved is reflected inthe infrared spectrum as additional bands and addedcomplexity.

As we move on to simple organic compounds, whereone or more functional groups or heteroatoms are addedto the molecule, we see many changes occurring in thespectrum. These result from the bonding associated withthe functional group, and also local disturbances to thebasic backbone spectrum that relate again to spatialchanges and also to local and neighboring electroniceffects. Examples of such functionalities are halogens,simple oxygen species, such as hydroxy and ether groups,and amino compounds. Carbonyl compounds, where theadded functional group includes the CDO bond, alsoprovide very profound contributions to the spectrum,and because of the wide diversity of these compoundsthey are best dealt with as a separate class.

A very characteristic group of compounds, froma spectral point of view, are the multiple-bondednitrogen compounds, such as cyanides and cyanates.These typically have very characteristic absorptions,which are easy to assign, and are free from spectralinterferences. The same can be said for some ofthe hydrides of heteroatoms, such as sulfides (thiols),silanes, and phosphines. Finally, there are other, oxygen-containing functional groups, as encountered in thenitrogen-oxy (NOx), phosphorus-oxy (POx), silicon-oxy(SiOx), and sulfur-oxy (SOx) compounds. These aresometimes more difficult to identify from first principles,and a knowledge of the presence of the heteroatomis helpful. The spectra are characteristic, but many ofthe oxy absorptions occur within a crowded and highlyoverlapped region of the spectrum, mainly between1350 and 950 cm�1. Also, many of these compoundsfeature C�O bonding, which is common in otherfrequently encountered functionalities such as ethersand esters.

3.1 The Hydrocarbon Species and Molecular Backbone

In this section we include the characteristic absorptionfrequencies encountered for the parent hydrocarbonspecies and the associated backbone or substituent group.This includes aliphatic and aromatic structures. Thespectral contributions are characterized, as previouslynoted, as C�H stretching and bending vibrations andC�C vibrations (stretching and bending), which forthe most part are unique for each molecule, and aregenerally described as skeletal vibrations. In the caseof aromatic compounds, ring CDC�C stretching andbending vibrations are highly characteristic, and arediagnostic. Likewise, the same can be said for theunsaturated carbon–carbon multiple bonding in alkeneand alkyne structures.


Table 1 Saturated aliphatic (alkane/alkyl) group frequencies

Group frequency (cm�1) Functional group/assignment

Methyl (−CH3)2970–2950/2880–2860 Methyl C�H asym./sym. stretch1470–1430/1380–1370 Methyl C�H asym./sym. bend1385–1380/1370–1365 gem-Dimethyl or ‘‘iso’’- (doublet)

1395–1385/1365 Trimethyl or ‘‘tert-butyl’’ (multiplet)Methylene ( CH2)

2935–2915/2865–2845 Methylene C�H asym./sym. stretch1485–1445 Methylene C�H bend750–720 Methylene �(CH2)n� rocking

(n ½ 3)1055–1000/1005–925 Cyclohexane ring vibrations

Methyne ( CH−)2900–2880 Methyne C�H stretch1350–1330 Methyne C�H bend1300–700 Skeletal C�C vibrations

Special methyl (−CH3) frequencies2850–2815 Methoxy, methyl ether O�CH3,

C�H stretch2820–2780 Methylamino, N�CH3, C�H stretch

3.1.1 Saturated Aliphatic and Alicyclic Compounds

See Table 1. The C�H stretch vibrations for methyland methylene are the most characteristic in termsof recognizing the compound as an organic compoundcontaining at least one aliphatic fragment or center. Thebending vibrations help to tell more about the basicstructure. For example, a strong methylene/methyl band(1470 cm�1) and a weak methyl band (1380 cm�1), plusa band at 725–720 cm�1 (methylene rocking vibration)is indicative of a long-chain linear aliphatic structure(note that splitting may be observed for the 1470 and720 cm�1 bands, which is indicative of a long-chaincompound, and is attributed to a crystallinity and a highdegree of regularity for the linear backbone structure).In contrast, strong methyl bands, showing significantsplitting, and a comparatively weaker methylene/methylband indicate chain branching, and the possibility ofisopropyl or tert-butyl substituents (depending on theamount of splitting, and the relative band intensities).A comparison between linear and branched chainhydrocarbons can be seen in Figures 3 and 4, where in thecase of isooctane, both isopropyl and tert-butyl groupsare present.

3.1.2 Unsaturated Compounds

See Table 2. As already commented upon, the saturatedhydrocarbon C�H stretching absorptions all occur below3000 cm�1. Any band structures observed between 3150and 3000 cm�1 are almost exclusively indicative ofunsaturation (CDC�H) and/or aromatic rings. The

Table 2 Olefinic (alkene) group frequencies

Origin Group frequency, Assignmentwavenumber (cm�1)

CDC 1680–1620 Alkenyl CDC stretch1625 Aryl-substituted CDC1600 Conjugated CDC

C�H 3095–3075 Terminal (vinyl) C�H stretchC3040–3010

3095–3075 Pendant (vinylidene) C�Hstretch

3040–3010 Medial, cis- or trans-C�Hstretch

C�H 1420–1410 Vinyl C�H in-plane bend1310–1290 Vinylidene C�H in-plane

bendC�H 995–985C 915–890 Vinyl C�H out-of-plane bend

895–885 Vinylidene C�H out-of-planebend

C�H 970–960 trans-C�H out-of-plane bend700 (broad) cis-C�H out-of-plane bend40

00

3600

3200

2800

2400

2000

1800

1600

1400

1200

1000 80

0

600

100

80

60

40

20

0

Wavenumber (cm−1)

Tran

smitt

ance

(%

)

C CHH

H(CH2)3 CH3

Figure 5 ATR spectrum of 1-hexene. Copyright Coates Con-sulting.

unsaturated hydrocarbons featuring CDC, with attachedhydrogens, usually occur as either a single or a pairof absorptions, in the ranges indicated in Table 2. Asnoted, the number of bands and their positions areindicative of the double bond location and the spatialarrangement around the double bond. The position of theCDC stretching frequency does vary slightly as a functionof orientation around the double bond, but it is lessinformative than the C�H information. The C�H out-of-plane bending is typically the most informative relativeto the location and spatial geometry of the double bond,where terminal and medial double bonds may be clearlydifferentiated. Figure 5 provides a good example with thespectrum of 1-hexene, which contains the terminal vinylgroup. Note that a fully substituted, medial double bondhas only the CDC as the sole indicator of the presenceof the double bond, unless the bond is conjugated with asecond unsaturated site.


Table 3 Aromatic ring (aryl) group frequencies


CDC�Ca 1615–1580 Aromatic ring stretchCDC�Ca 1510–1450 Aromatic ring stretchC�H 3130–3070 Aromatic C�H stretchC�H 1225–950 (several) Aromatic C�H in-plane

bendC�H 900–670 (several) Aromatic C�H

out-of-plane bend770–730C 710–690 Monosubstitution (phenyl)

770–735 1,2-Disubstitution (ortho)810–750C 900–860 1,3-Disubstitution (meta)

860–800 1,4-Disubstitution (para)‘‘Combi’’b 2000–1660 (several) Aromatic combination

bands

a CDC�C used as an approximation of the unique aromatic ringbonding.

b ‘‘Combi’’ denotes assignment to combination bands.

3.1.3 Aromatic Compounds

See Table 3. The existence of one or more aromaticrings in a structure is normally readily determinedfrom the C�H and CDC�C ring-related vibrations. TheC�H stretching occurs above 3000 cm�1 and is typicallyexhibited as a multiplicity of weak-to-moderate bands,compared with the aliphatic C�H stretch. The structureof the bands is defined by the number and positions of theC�H bonds around the ring, which in turn are related tothe nature and number of other substituents on the ring.Note that the same applies to the C�H out-of-plane bend-ing vibrations, which are frequently used to determine thedegree and nature of substitution on the ring – examplesare provided in Figure 6(a–c), with the comparison ofthe three xylene isomers. This picture often becomesmore complex if multiple- or fused-ring structures existin a compound. The other most important set of bandsare the aromatic ring vibrations centered around 1600and 1500 cm�1, which usually appear as a pair of bandstructures, often with some splitting. The appearance andratio of these band structures is strongly dependent onthe position and nature of substituents on the ring.

3.1.4 Acetylenic Compounds

See Table 4. Although acetylenic compounds are not verycommon, the spectrum associated with the C�C structurecan be characteristic. It is instructive to note the impacton the carbon–carbon bond stretching as a function ofincrease in bond order for the series of single-, double-,and triple-bonded carbon:

C�C stretch:¾1350–1000 cm�1 (skeletal vibrations)CDC stretch: 1680–1620 cm�1

C�C stretch: 2260–2100 cm�1

4000

3600

3200

2800

2400

2000

1800

1600

1400

1200

1000 80

0

100

80

60

40

20

0

CH3

H

H H

CH3

H

(a)

4000

3600

3200

2800

2400

2000

1800

1600

1400

1200

1000 80

0

100

80

60

40

20

0Tran

smitt

ance

(%

)

CH3

H

H CH3

H

H

(b)

4000

3600

3200

2800

2400

2000

1800

1600

1400

1200

1000 80

0

100

80

60

40

20

0

Wavenumber (cm−1)

CH3

CH3

H

H

H

H

(c)

Figure 6 ATR spectra of xylene isomers: (a) o-xylene, 1,2-dimethylbenzene; (b) m-xylene, 1,3-dimethylbenzene; (c) p-xylene, 1,4-dimethylbenzene. Copyright Coates Consulting.

Table 4 Acetylenic (alkyne) group frequencies


C�C 2140–2100 Terminal alkyne(monosubstituted)

C�C 2260–2190 Medial alkyne(disubstituted)

C�H 3320–3310 Alkyne C�H stretchC�H 680–610 Alkyne C�H bendC�H 630 (typical) Alkyne C�H bend

This increase in bond order produces a correspondingincrease in bond strength, which in turn increases the force


constant, k (see Equation 3), supporting the Hooke’s lawmodel described earlier.

As noted in Table 4, the position of the C�C bond isinfluenced by whether the group is terminal or medial.The single hydrogen of the terminal acetylene itself isvery characteristic, reflecting the labile nature of theacetylenic C�H.

3.2 Simple Functional Groups

Obviously, there is a potentially broad number ofmolecular fragments that can be considered to befunctional groups attached to an organic structure orbackbone. This section features the most simple andmost common of the functional groups, �C�X, i.e. thehalogens (X D F, Cl, Br and I), hydroxy (X D OH),oxy or ether (X D OR, where R D alkyl), and amino(X D NH2, DNH or �N). With the exception of thecarbonyl functionality, these three basic functional groupscover most of the common occurrences in simple organiccompounds. Note that for the oxy/hydroxy and aminofunctionalities, these are molecular fragments, and theycontribute their own set of characteristic absorptions tothe spectrum of the compound. In fact, the bondingbetween the functional group and the backbone is onlyone part of the overall picture used for the spectralinterpretation.

3.2.1 Halogenated Compounds

See Table 5. In principle, the interpretation of the spectraof molecules containing one or more halogens wouldseem to be straightforward. The functionality is simple,with just a single atom linked to carbon to form the group.With the polar nature of this group, one would expect thespectral contribution to be distinctive. In reality, this isnot always the case.

In aliphatic compounds, the C�X bond typicallypossesses a unique group frequency, which may be

Table 5 Aliphatic organohalogen compound groupfrequencies

Origin Group frequency, Assignmentwavenumber (cm�1)a

C�F 1150–1000 Aliphatic fluoro compounds,C�F stretch

C�Cl 800–700 Aliphatic chloro compounds,C�Cl stretch

C�Br 700–600 Aliphatic bromo compounds,C�Br stretch

C�I 600–500 Aliphatic iodo compounds,C�I stretch

a Note that the ranges quoted serve as a guide only; the actual rangesare influenced by carbon chain length, the actual number of halogensubstituents, and the molecular conformations present.

assigned to the halogen–carbon stretching. When a singlehalogen is present, the determination of this group isstraightforward. However, if more than one halogen ispresent, the interpretation is usually more complex. Insuch cases, the result varies depending on whether thehalogens are on the same or different carbon atoms, and, ifon different atoms, whether the atoms are close neighbors.This is particularly the case with small molecules, andthe resultant spectral complexity arises from the factthat there is restricted rotation about the carbon–carbonbond.

Single bonds usually exhibit free rotation, which wouldnormally mean that there are no preferred spatialorientations for the molecules. However, owing to thesize of the halogen atom, relative to the carbon andhydrogen that form the backbone, the molecules tendto exhibit certain specific conformations, where thespatial interaction between neighboring halogen atomsis minimized, and each conformation provides its owncontribution to the overall spectrum. It is importantto appreciate that this issue of spatial orientation hasan impact even on high-molecular-weight compounds,such as the polyhalogenated polymers, e.g. poly(vinylchloride). Here preferred orientations have an impact onthe crystallinity of the polymer, and this in turn has asignificant impact on both the spectrum and the physicalproperties of the material.

Another important issue to consider with halogensubstituents is the high electronegativity of the halogenatom. This can have a noticeable impact on the spectrumof neighboring group frequencies, including adjacenthydrogen atoms. In such cases, significant shifting ofthe C�H frequencies can occur – the direction of theshift being dependent on the location of the C�H,and whether the halogen adds or extracts electrondensity from the C�H bond – adding strengthens (higherfrequency) and extracting weakens (lower frequency).The same influences can be observed with halogen-substituted carbonyl compounds, such as acyl halidesand a-substituted acids, where the bond strength of thecarbonyl group is increased (see section 3.3). In mostcases, both a shift to higher frequency and an increase inabsorption strength for the band are observed.

Table 5 only presents the group frequencies for thealiphatic compounds, because no well-defined C�Xabsorptions are observed for halogen-substituted aro-matic compounds. The presence of a halogen on anaromatic ring can be detected indirectly from its elec-tronic impact on the in-plane C�H bending vibrations.Normally, we do not consider the in-plane bending bandsto be of use because, as pointed out earlier, these occur ina spectral region that is crowded by other important groupfrequencies. However, in the case of a halogen-substitutedring, the intensity of these vibrations is enhanced relative


to other absorptions by as much as three to four times.For reference, it is informative to compare the intensitiesfor these bands, between 1150 and 1000 cm�1, for thespectra of toluene and chlorobenzene (Figure 7a and b).

3.2.2 Hydroxy and Ether Compounds

See Table 6 for alcohols and hydroxy compounds.The hydroxy function is probably one of the mostdominant and characteristic of all of the infrared groupfrequencies. In most chemical environments, the hydroxygroup does not exist in isolation, and a high degreeof association is experienced as a result of extensivehydrogen bonding with other hydroxy groups. Thesehydroxy groups may be within the same molecule(intramolecular hydrogen bonding) or they most likelyexist between neighboring molecules (intermolecularhydrogen bonding). The impact of hydrogen bondingis to produce significant band broadening and to lowerthe mean absorption frequency. The lowering of thefrequency tends to be a function of the degree andstrength of the hydrogen bonding. In compounds such ascarboxylic acids, which exhibit extremely strong hydrogenbonding, forming a stable dimeric structure, a highlycharacteristic, large shift to lower frequencies is observed.

4000

3600

3200

2800

2400

2000

1800

1600

1400

1200

1000 80

0

600

100

80

60

40

20

0

Wavenumber (cm−1)

Tran

smitt

ance

(%

)

(a)

ClH

H H

H

H

4000

3600

3200

2800

2400

2000

1800

1600

1400

1200

1000 80

0

100

80

60

40

20

0

Wavenumber (cm−1)

Tran

smitt

ance

(%

)

(b)

CH3

H

H

H

HH

Figure 7 Comparison of ATR spectra of (a) chlorobenzeneand (b) toluene. Copyright Coates Consulting.

Table 6 Alcohol and hydroxy compound group frequencies


O�H 3570–3200 (broad) Hydroxy group, H-bondedOH stretch

3400–3200 Normal ‘‘polymeric’’ OHstretch

3550–3450 Dimeric OH stretch3570–3540 Internally bonded OH stretch

O�H 3645–3600 (narrow) Nonbonded hydroxy group,OH stretch

3645–3630 Primary alcohol, OH stretch3635–3620 Secondary alcohol, OH stretch3620–3540 Tertiary alcohol, OH stretch3640–3530a Phenols, OH stretch

O�H 1350–1260 Primary or secondary, OHin-plane bend

1410–1310 Phenol or tertiary alcohol, OHbend

720–590 Alcohol, OH out-of-planebend

C�O ¾1050b Primary alcohol, C�O stretch¾1100b Secondary alcohol, C�O

stretch¾1150b Tertiary alcohol, C�O stretch¾1200b Phenol, C�O stretch

a Frequency influenced by nature and position of other ring substituents.b Approximate center of range for the group frequency.

In special circumstances, where the hydroxy groupis isolated – either because of steric hindrance effectsor because the sample is in the vapor state or ina dilute solution of a nonpolar solvent – the band ischaracteristically narrow, and is observed at the natural,higher frequency. This absorption is important for thecharacterization of certain hindered phenol antioxidants,a commercially important class of compounds in the food,polymer, and formulated oil industries.

It must be appreciated that while the hydroxy absorp-tion is singly one of the most important bands in theinfrared spectrum, other vibrations are also importantfor the actual characterization of the compound. Alco-hols exist as three distinct classes – primary, secondaryand tertiary – distinguished by the degree of carbon sub-stitution on the central hydroxy-substituted carbon, asingle substitution being primary, double substitutionbeing secondary, and triple substitution being tertiary.This is an important fact, because the chemistry andoxidation stability of the alcohol are strongly influencedby the degree of substitution. Whether an alcohol is pri-mary (1°), secondary (2°) or tertiary (3°), may be reflectedin the position of the OH stretch absorption, but typicallythis is determined by the other absorptions, in particu-lar the C�O� stretching frequency. Another absorptionof lower importance, but often characteristic, is assigned


to another form of bending vibration, the out-of-planebend or wagging vibration of the O�H. The OH bend-ing vibrations are broadened by hydrogen bonding as isthe stretching absorption, but often to a lesser extent.The differences between primary and secondary alcoholscan be appreciated from Figure 8(a) and (b), where thespectra of 1- and 2-octanol are presented.

See Table 7 for ethers and oxy compounds. In somerespects, ethers are related to alcohol and hydroxy

4000

3600

3200

2800

2400

2000

1800

1600

1400

1200

1000 80

0

600

100

50

0

Wavenumber (cm−1)

Tran

smitt

ance

(%

)

(a)

4000

3600

3200

2800

2400

2000

1800

1600

1400

1200

1000 80

0

600

100

50

0

Wavenumber (cm−1)

Tran

smitt

ance

(%

)

(b)

Figure 8 ATR spectra of (a) primary (1-octanol) and (b) sec-ondary (2-octanol) alcohols. Copyright Coates Consulting.

Table 7 Ether and oxy compound group frequencies


C�H 2820–2810 Methoxy, C�H stretch(CH3�O�)

C�O�C 1150–1050 Alkyl-substituted ether,C�O stretch

C�O�C 1140–1070 Cyclic ethers, largerings, C�O stretch

f�O�H 1270–1230 Aromatic ethers,aryl�O stretch

C�O� ¾1250C 890–800a Epoxy and oxirane ringsC�O�O�C 890–820a Peroxides, C�O�O�

stretch

a Typically very weak, and not very characteristic in the infrared. Tendto be more characteristic in the Raman spectrum.

compounds, where the hydrogen of the hydroxy groupis replaced by an aliphatic (alkyl) or aromatic (aryl)molecular fragment. Having stated that, the overallappearance of an ether spectrum is drastically differentfrom that of a related alcohol. This is due to theoverwhelming effect of hydrogen bonding on the hydroxygroup. However, many of the relationships that exist forthe C�O component of the alcohol carry over to thecorresponding ether. The relationships that pertain toprimary, secondary, and tertiary structures remain intact.The main difference is that one now considers the bondingon both sides of the oxygen, because if carbon is on bothsides, then two ether bonds exist. Ethers can exist assimple ethers (same group both sides) and mixed ethers(different groups both sides). Infrared spectroscopy isfairly sensitive for differentiating these ether functions,especially when the structures are mixed aliphatic oraliphatic/aromatic.

3.2.3 Amino Compounds

See Table 8. In some respects, the infrared spectra andthe characteristic group frequencies of amines tend to

Table 8 Amine and amino compound group frequencies


Primary aminoN�H 3400–3380 Aliphatic primary amine,

C3345–3325 NH stretchN�H 3510–3460 Aromatic primary amine,

C3415–3380 NH stretchN�H 1650–1590 Primary amine, NH bendC�N 1090–1020 Primary amine, CN stretch

Secondary aminoN�H 3360–3310 Aliphatic secondary amine,

NH stretchN�H ¾3450 Aromatic secondary amine,

NH stretchN�H 3490–3430 Heterocyclic amine,

NH stretchDN�H 3350–3320 Imino compounds,

NH stretchN�H 1650–1550 Secondary amine, NH bend

C�N 1190–1130 Secondary amine,CN stretch

Tertiary aminoC�N 1210–1150 Tertiary amine, CN stretch

Aromatic aminoC�N 1340–1250 Aromatic primary amine,

CN stretchC�N 1350–1280 Aromatic secondary amine,

CN stretchC�N 1360–1310 Aromatic tertiary amine,

CN stretch


parallel those of alcohols and ethers. Before proceeding,there is an important distinction to be made betweenthe nomenclature used to describe alcohol structures,compared with that used for amines. The terms primary,secondary, and tertiary are used to describe amines, butthe substitution relates to the nitrogen, not the adjoiningcarbon (as with alcohols). The terminology used is�NH2

primary, NH secondary and N� tertiary. As before,these structural differences are important, and theystrongly influence the chemistry and the reactivity of thenitrogen and the N�H group (primary and secondary).Again, the infrared spectrum is very diagnostic, and itis possible to differentiate readily the functional groupstructures, especially for the primary and secondaryamino compounds (see Figure 9a–c for an examplecomparison of primary, secondary, and tertiary aromaticamines).

As with hydroxy compounds, hydrogen bonding isimportant, but the effect tends to be weaker thanfor the hydroxy group, and the overall effect on thespectrum is slightly less pronounced. This situation altersin the related ammonium and amino salts, where stronghydrogen bonding is experienced, and a correspondingbroadening of the associated NH absorptions is observed.

Note that only the primary and secondary aminesexhibit the most characteristic group frequencies, whichare associated with the N�H bond. Tertiary aminesare very comparable to ethers, and the main diagnosticinformation is gathered from the C�N vibrations only. Animportant exception is the methylamino (and dimethyl-amino) group, which like its counterpart (methoxy inethers) has a characteristic C�H stretching vibration atlower than the normal C�H stretching frequency formethyl.

3.3 The Carbonyl Group

Carbonyl compounds are not only chemically important,but are also important in the interpretation of infraredspectra. The CDO absorption is almost always one of themost characteristic in the entire spectrum, and it is alsomost likely to be the most intense spectral feature. Table 9provides an example listing of some of the commoncarbonyl frequencies as a function of the particular typeof carbonyl group. In essence, a ketone is consideredthe root compound, with an aldehyde being a specialcase, where the carbonyl group is terminal, and only hasone substituent, the other being a single hydrogen atom.All of the other carbonyl compounds, in a way, can beconsidered to be derived from the base ketone structure,where one or both alkyl (or aryl) substituents are replacedby another functionality, e.g. from a single hydroxy group,in the case of carboxylic acids, to two ether groups, as inthe case of an organic carbonate.

NH2

H

H H

H

H

4000

3600

3200

2800

2400

2000

1800

1600

1400

1200

1000 80

0

600

100

80

60

40

20

0

(a)

4000

3600

3200

2800

2400

2000

1800

1600

1400

1200

1000 80

0

600

100

80

60

40

20

0Tran

smitt

ance

(%

)

(b)

NHH

H H

H

H

CH3

4000

3600

3200

2800

2400

2000

1800

1600

1400

1200

1000 80

0

600

100

80

60

40

20

0

Wavenumber (cm−1)(c)

NH

H H

H

H

CH3 CH3

Figure 9 ATR spectra of (a) primary (aniline), (b) secondary(N-methylaniline) and (c) tertiary (N,N-dimethylaniline) aro-matic amines. Copyright Coates Consulting.

The actual diagnostic carbonyl absorption frequencyis dependent on the electronic characteristics of thesubstituent group, which in turn define very much thechemical characteristics and reactivity of the specificcarbonyl compound. Spatial and structural factors canbe important, in particular ring stress, as in the caseof lactones (cyclic esters) and lactams (cyclic amides).In this case, the greater the ring stress, the higher isthe carbonyl absorption frequency. In common withprevious observations for a double-bonded functionality,conjugation plays an important role in the observedcarbonyl frequency. This includes connection to anaromatic ring or conjugation to a CDC or another CDO.


Table 9 Example carbonyl compound group frequencies

Group frequency (cm�1) Functional group

1610–1550/1420–1300 Carboxylate (carboxylic acidsalt)

1680–1630 Amide1690–1675/(1650–1600)a Quinone or conjugated ketone

1725–1700 Carboxylic acid1725–1705 Ketone

1740–1725/(2800–2700)b Aldehyde1750–1725 Ester

1735 Six-membered ring lactone1760–1740 Alkyl carbonate1815–1770 Acid (acyl) halide1820–1775 Aryl carbonate

1850–1800/1790–1740 Open-chain acid anhydride1870–1820/1800–1775 Five-membered ring

anhydride2100–1800 Transition metal carbonyls

a Lower frequency band is from the conjugated double bond.b Higher frequency band characteristic of aldehydes, associated with

the terminal aldehydic C�H stretch.

As previously discussed, a lowering of the parent groupfrequency is observed. This effect is important for thedifferentiation of certain types of carbonyl compound,in the determination of whether the carbonyl group isdirectly or indirectly attached an aromatic ring, e.g. theability to differentiate aryl acetates from alkyl benzoates.In the case of the acetate, the ring is joined to the ‘‘etheroxygen’’ of the ester group, and is not conjugated with thecarbonyl, whereas with the benzoate, the ring is directlyconjugated with the group, and the carbonyl absorptionfrequency is correspondingly lowered.

Often, the frequency ranges for the different classes ofcarbonyl compound overlap, and the carbonyl frequencyalone is not sufficient to characterize the functionalgroup. In most cases, spectral information from theother component of the functional group is used for thecharacterization. Carboxylic acids are a good example,where the C�O, C�O�H and O�H vibrations are highlycharacteristic. Similarly, esters (C�O�C) and amides(C�N and N�H) are other frequently encounteredexamples. The frequencies provided earlier for theseadditional functionalities (C�O, C�N, and N�H) ingeneral apply, although the actual observed frequenciesmay differ slightly, being modified by the carbonyl group.It is worthwhile returning momentarily to carboxylicacids, because they are unique, in so far as the hydroxygroup has direct interaction with the carbonyl group,by the formation of a stable dimeric hydrogen-bondedstructure in the condensed phase (solid and liquid).Note that this structure disappears in the vapor state. Acharacteristic broad feature in the range 3300–2500 cm�1,that overlaps the C�H stretching region, and with asecondary absorption close to 2600 cm�1, is observed for

the hydrogen-bonded O�H of most carboxylic acids. Theimpact of this overlap can be appreciated by referenceto the spectrum of butyric acid, provided in Figure 10.Other bands that are associated with the C�O andO�H components tend to be less pronounced, andsometimes may be overlapped with other fingerprintabsorptions of the molecule. These are located in theranges 1320–1210 cm�1 (C�O stretch) and 960–850 cm�1

(hydrogen-bonded O�H out-of-plane bending).

3.4 Other Functional Groups Associated withHeteroatoms

Potentially there are very large numbers of differentorganic-based compounds that are associated with one ormore heteroatoms. These are in addition to the simplehalogen- and amino-based compounds that have alreadybeen covered. A detailed discussion of such compoundsis beyond the scope of this article. A few illustrativeexamples are included here in order to provide the readerwith a feel for the spectral contributions of commonlyencountered compounds that feature heteroatoms. Theseinclude triple-bonded and cumulated double-bondednitrogen compounds, such as cyano and cyanato com-pounds, hetero-oxy compounds, such as nitro, sulfoxy,phosphoroxy and silicoxy compounds, and thiols andthio-substituted compounds. Many of these belong toimportant compound classes, and complete texts existthat focus on the spectral attributes of such compounds.

3.4.1 Multiple-bonded and Cumulated Double-bondedNitrogen Compounds

Nitrogen compounds featuring triple or cumulated doublebonds, such as cyanides or nitriles (�C�N) and cyanates(�O�C�N), all provide a unique spectrum, typically witha single, normally intense absorption at 2280–2200 cm�1

100

80

60

40

20

0

4000

3600

3200

2800

2400

2000

1600

1200 80

0

400

Tra

nsm

ittan

ce (

%)

Wavenumber (cm−1)

CH3 (CH2)2 CO

OH

Figure 10 Transmission spectrum of n-butyric acid. CopyrightCoates Consulting.


(for cyano compounds) and 2285–1990 cm�1 (cyanates,isocyanates, thiocyanates, etc.). The band is assigned tothe stretching vibration, which is the asymmetric stretch inthe case of the cumulated double-bonded compounds (thesymmetric stretch is typically weak, and is not diagnosticof the group).

Double-bonded nitrogen groups, such as imino groups( CDN�) and azo groups (�NDN�), exhibit absorptionsclose to the carbonyl (CDO) and alkene (CDC) doublebond stretching region. While they are characteristic forthe functional group, they are sometimes difficult to assignfrom first principles because of the overlap with othercommon functional groups in the region. For example,dependent on substitution or location, the CDN� groupcan occur in the same spectral region as the CDO of anamide. Examples of the group frequencies for a fewcommon multiple-bonded and cumulated double-bondedcompounds are provided in Table 10.

3.4.2 Hetero-oxy Compounds

The group of compounds covered here feature X�O(where X D nitrogen, sulfur, phosphorus, and silicon) andXDO vibrations. In general, the characteristic absorp-tions span a similar range to that covered by thecorresponding carbon–oxygen compounds, with mostof the stretching frequencies observed within the mainfingerprint spectral region (1500–400 cm�1). Many ofthe compounds can be considered to be analogs ofethers, especially when an alkoxy group is present,featuring the X�O�C linkage. Interestingly, the nitrogroup (�NO2) is isoelectronic with the carboxylateion group (�CO2

�), and both provide very similarspectra for the main functional group. See Table 11for example group frequencies for hetero-oxy com-pounds.

3.4.3 Thiols and Thio-substituted Compounds

Thiols and thio-substituted compounds, by definition, canbe considered to be the direct analogs of the equivalent

Table 10 Examples of nitrogen multiple and cumulateddouble bond compound group frequencies


2280–2240 Aliphatic cyanide/nitrile2240–2220 Aromatic cyanide/nitrile

2260–2240/1190–1080 Cyanate (�OCN and C�OCNstretch)

2276–2240 Isocyanate (�NDCDO asym.stretch)

2175–2140 Thiocyanate (�SCN)2150–1990 Isothiocyanate (�NCS)1690–1590 Open-chain imino (�CDN�)1630–1575 Open-chain azo (�NDN�)

Table 11 Example group frequencies for simple hetero-oxycompounds


Nitrogen-oxy compounds1560–1540/1380–1350a Aliphatic nitro compounds1555–1485/1355–1320a Aromatic nitro compounds1640–1620/1285–1270a Organic nitrates

Phosphorus-oxy compounds1350–1250 Organic phosphates (PDO

stretch)1050–990 Aliphatic phosphates

(P�O�C stretch)1240–1190/995–850 Aromatic phosphates

(P�O�C stretch)Sulfur-oxy compounds

1335–1300/1170–1135a Dialkyl/aryl sulfones1420–1370/1200–1180a Organic sulfates1365–1340/1200–1100a Sulfonates

Silicon-oxy compounds1095–1075/1055–1020 Organic siloxane or silicone

(Si�O�Si)1110–1080 Organic siloxane or silicone

(Si�O�C)

a Asymmetric/symmetric XO2 stretch (NO2 and SO2).

oxygenated compounds such as alcohols and ethers.Unlike the oxygen-containing analogs, the equivalentC�S and C�S�H stretching vibrations tend to giverise to very weak absorptions in the infrared spectrum.The C�S and S�H bonds are highly polarizable,and hence produce stronger spectral activity in theRaman spectrum than the infrared spectrum. The highermass of sulfur, compared with oxygen, results in thecharacteristic group frequencies occurring at noticeablylower frequencies than the oxygen-containing analogs,as noted in Table 12 for compounds containing S�Hand C�S bonds. The thiol S�H is probably the onlyinfrared absorption that can be considered to be of usefor the general characterization of these compounds,and most of the others are obtained by inference, orfrom knowledge of the compound type. Catenationis a unique chemical characteristic of sulfur, wherethe formation of S�S bonds in extended chains iscommon, hence the inclusion here of the S�S stretchingfrequencies.

3.5 Simple Inorganics

Characterization of compounds via infrared spectroscopyis not limited to organic compounds. Any inorganiccompound that forms bonds of a covalent naturewithin a molecular ion fragment, cation or anion,will produce a characteristic absorption spectrum, withassociated group frequencies. In a manner, certain aspects


Table 12 Common group frequencies for thiols andthio-substituted compounds


2600–2550 Thiols (S�H stretch)710–685 Thiol or thioether, CH2�S�

(C�S stretch)660–630 Thioethers, CH3�S� (C�S

stretch)715–670 Aryl thioethers, f�S (C�S

stretch)705–570 Disulfides (C�S stretch)620–600 Disulfides (S�S stretch)500–430 Aryl disulfides (S�S stretch)500–470 Polysulfides (S�S stretch)

Table 13 Example group frequencies for common inorganicions


1490–1410/880–860a Carbonate ion1130–1080/680–610a Sulfate ion1380–1350/840–815a Nitrate ion

1100–1000 Phosphate ion1100–900 Silicate ion

3300–3030/1430–1390a Ammonium ion2200–2000 Cyanide ion, thiocyanate ion,

and related ions

a Typically, the first absorption is intense and broad, and the secondhas weak to medium intensity and is narrow. Both often exist asmultiple band structures, and this may be used to characterize individualcompounds.

have already been covered for inorganic compounds inthe form of salts of carboxylic acids and amino andammonium compounds, which can be extended to metalcomplexes, chemical fragments associated with hetero-oxy groups (nitrates, sulfates, phosphates, silicates, etc.),and transition metal carbonyl compounds. All complexionic compounds (containing more than one atom) andcoordination compounds produce characteristic spectra.Many of the associated group frequencies can beused diagnostically for characterization. The structureand orientation of the ion or complex, both as anisolated entity or within a crystal lattice, are importantfactors that affect the appearance and nature of theinfrared spectrum. Hydration of compounds (water ofcrystallization) also has a large effect on the spectrum,and often adds a lot of complexity, in the form ofadditional absorption bands and structure to existingbands. The subject is far too broad to be covered inthis article, and only a few example group frequenciesare included here (Table 13). The reader is directedto the recommended standard texts for more detailedinformation..13,14/

4 THE PRACTICAL SITUATION –OBTAINING THE SPECTRUM ANDINTERPRETING THE RESULTS

Up to this point, the fundamentals of interpretation havebeen discussed from the most basic concepts of infraredabsorption by a molecular species and the impact ofchemical functionality on the resultant spectrum. In manyways, this discussion has treated the molecule as a more orless isolated species, with no consideration of the physicalstate of the actual sample or the environment in whichthe molecule exists. Physical state and the molecular,chemical and physical environments have a profoundeffect on the infrared spectrum. As a result, it is just asimportant to understand and interpret these effects asit is to perform the fundamental interpretation of thefunctional groups from first principles.

This particular section may be one of the mostimportant for many of the readers, because it reflectsthe real world, that is taking a sample, preparing it foranalysis, and making use of all of the available informationon the sample and the spectrum or spectra obtained.Note that sometimes there are several options in theway that a sample may be handled for infrared analysis.It is often beneficial to consider the different options.Before we start to examine the situation it is important tounderstand the importance of the interpretation and todetermine the real requirements. Here are some examplescenarios:

1. The sample (or spectrum) is a ‘‘total unknown’’and an identification is required – examples includeforensic samples, environmental waste samples, ornew discovery samples, where a new material hasbeen synthesized or discovered.

2. The sample (or spectrum) is an unknown and itneeds to be characterized or classified – examplesinclude commercial applications where new additivesor components are included in a material to pro-vide a specific property; in such cases this could beconsidered the basis of competitive product analy-sis.

3. The sample generally is known but the existence ofa specific chemical class needs to be determined –examples include contaminant analysis, analysisfor toxicology or environmental reasons, materialadditives, etc.

4. The sample is a complete known and the inter-pretation is required to confirm the material com-position and/or quality – examples include productquality control and the confirmation of a struc-ture or functionality of a newly synthesized mate-rial.


There are clearly many other scenarios that can beconsidered, and the examples presented may cross overinto more than one of these scenarios. However, the ideahere is to help define a strategy for the interpretation.For the most part, this text will focus on scenarios 1and 2, although a passing reference may be made to theothers. Note also that there are two ways to address theinterpretation:

ž one is to attempt to assign as many bands as possibleusing group frequency tables;

ž the other is to attempt to build up a picture usinginformation from key spectral bands (with a knowl-edge of the group frequencies) and rationalizing thespectrum relative to the sample as known, and anyknown history, including the sample preparation andpresentation technique.

The latter approach is the best way to go, andis typically far more rewarding. Also, it is seldomthat all bands in the infrared spectrum of a samplecan be fully or accurately assigned. There are toomany ambiguities and coincidences for a definitiveinterpretation to be obtained from the spectra of mostmaterials – remembering, of course, that the sample maybe a mixture.

To begin, let us start from either the point wherea spectrum has been generated or where a sample ispresented. In practice, these are very different startingpoints. The latter is always preferred because with thesample in hand it is possible to gain a lot of first-handinformation about nature which will ultimately help in thefinal interpretation. The term ‘‘unknown’’ is used becausethis is how people often view the interpretation process;however, in reality there are seldom true unknowns – inother words, the reason for the interpretation itselfoften provides implied information about the sample.Even if a person presents a prerecorded spectrum, itis usually accompanied by some additional information.This at minimum should include a knowledge of howthe sample was prepared, whether it is a solid, liquid,or gas, or some mixed state, and maybe some physicalcharacteristics (color, appearance, melting point/boilingpoint, etc.). For the benefit of presenting a completepicture, we shall assume that the sample is availableand that we have the opportunity to perform additionaltests and measurements on it. Further, more detaileddiscussions of this practical approach can be found in theliterature..18 – 20/

4.1 Sample History

As noted, we are seldom dealing with a true unknown,and typical situations include the following:

ž a sample of fiber taken or extracted from a particularenvironment;

ž a contaminant removed from a material;ž a suspicious liquid found leaking out of some drums;ž a residue extracted from a surface;ž a residual liquid or solid remaining after storage or

treatment of a product;ž extracted additives or components from a commercial

product;ž a mass that was produced after a specific chemical

reaction;ž or simply a manufactured product.

Most people can equate with these different sets ofcircumstances. Each one brings information about thesample.

In the case of a fiber, if it is organic, then it typicallyfits within certain classes of polymeric materials. Itmay have orientation properties, which will influencethe appearance of the infrared spectrum, depending onhow it is sampled, and if there is more than one fiber,then the original material may have been a blend. Theenvironment from where the sample was taken will beimportant because there may be surface contaminantsthat may produce spurious bands in the spectrum,which may be removed by some sample pretreatment,such as washing with a solvent. Alternatively, thefiber may be coated, or degraded, thermally or byoxidation.

The characterization of contaminants is a commonapplication for infrared spectroscopy. For this type ofexercise it is necessary to know the importance of thecontaminant and why it is believed to be a contaminant.Was it extracted from a liquid or solid matrix? if so,what type of matrix, and what was the composition of thematrix? Is it possible that the contaminant originatedfrom the matrix itself, as a breakdown product? – ifso, then there should be some chemical similarities orcommon heritage. Was it embedded within a matrix, suchas a plastic/polymer film or a sheet of paper? – if so, isthe contaminant itself potentially contaminated by thematrix? This is especially important for interpretation,because the matrix may have a very intense infraredspectrum of its own.

For extracted materials, such as additives removed froma polymer, it is important to know something about thepolymer matrix – few extractions are 100% efficient, andit is common to have carry-over. In many such cases, thespectrum obtained reflects that of a mixture rather than apure compound. Things to bear in mind here are what theoriginal base polymers were, what other additives couldbe present, and what material, usually a solvent, was usedfor the extraction.


Looking at residues and residual products of a reactionis similar to the concept of characterizing a contaminant.Residues formed within products are often the resultof a breakdown reaction or an unexpected interactionbetween components. If derived from one or more of thecomponents, then a knowledge of all of the ingredientsis essential – again we would be looking for chemicallyrelated species in the residue, and reference spectra ofeach of the ingredients would be important in attemptingto characterize the material. In the case of a reactionproduct, spectra for the reactants and any solvents orsubstrates (catalysts, etc.) used would be helpful.

Finally, the identification of an end product from itsinfrared spectrum is common practice. By definition,unless the final product is a pure chemical, mostmanufactured products are mixtures, often featuringblended solid and/or liquid components from percentagerange concentrations to trace amounts. Often at issue ishow much can be identified from the interpretation ofa single spectrum. It is often necessary to perform someform of preconcentration on one or more ingredients,or to perform a chemical modification to change thespectral response of one or more of the major ingredients.This latter operation may help to remove the spectralinterferences from a major component for determiningthe presence of minor components.

4.2 Physical Characteristics of the Sample

From considerations based on the discussions in thesection above, a good working knowledge of the sampleallows one to determine basic information about thesample. This information can help in the determination ofthe best method of sampling. It can also help to indicate ifany special treatments to the sample are necessary beforeproceeding, such as the removal of residual solvents,or the need for selective extraction, in the event thatthe sample has more than one phase, or is known tobe a mixture. Beyond this there are other importantphysical characteristics of the sample that can assist theinterpretation process. As indicated, physical state canbe very important. For example, it has been noted thatnitro compounds are isoelectronic with organic acid salts.Many simple nitro compounds are liquids, whereas simplecarboxylates are usually solids.

There are, however, many other physical characteristicsand attributes that can be linked to the sample and itschemistry. This in turn can be linked to important spectralfeatures in the infrared spectrum. Two important charac-teristics are color and odor. Nitrogen compounds oftenhave color associated with them, e.g. nitro-substitutedcompounds often have a characteristic yellow or orangecolor, especially substituted phenols and amines. Sim-ilarly, unsaturated nitrogen compounds, such as azo

compounds, and highly conjugated olefinic compounds,such as carotenes, are similarly colored. This can be usedas a guide to look for the specific functionality in thespectrum.

Likewise, odor can be a useful guide. This author isnot advocating that a person should inhale the vaporof a sample directly, for obvious reasons. However, ifunder cautious sampling conditions a characteristic odoris detected, then this can provide a valuable insight regard-ing the presence of particular functional groups. Commonexamples are alcohols, esters, and ketones, which oftenpossess relatively pleasant fragrances, especially whenassociated with some degree of unsaturation. An almond-like odor might indicate a nitro compound, benzaldehyde,or a cyano compound (caution!), a fish-like odor is oftenassociated with amino or amido compounds, and a putridor ‘‘bad-cheese’’ odor is associated with certain carboxylicacids. With time, one can build up a mental library of com-mon odors and fragrances, and these can be very valuablefor initial material screening, by helping one look for thepresence of specific functional groups.

Volatility and flammability can be useful guides.Most volatile compounds tend to have a relatively lowmolecular weight, and they usually are unassociated. Thislatter point helps to rule out certain hydrogen-bondedspecies, such as carboxylic acids, amines, and multiplehydroxy-substituted compounds (such as glycols). If asample is submitted for some form of combustion analysis,it is worthwhile monitoring how it burns. If it burns witha clean, bluish flame, that might indicate that the sampleis oxygenated (alcohol or ether). If it has a yellow and/orsmoky flame, it indicates an excess of carbon, possibleunsaturation or an aromatic compound. The latter canproduce a very smoky flame.

4.3 The Chemistry of the Sample

First, there are issues relating to solubility. The old adage,like dissolves like, is very appropriate. Nonpolar materialssuch as hydrocarbons tend to be more soluble in nonpolarsolvents, such as hydrocarbons or chlorinated solvents,whereas polar materials favor solvents such as methanoland acetone, and in extreme cases their solubility maybe limited to the most polar solvents, such as water.If a material is potentially identified from its infraredspectrum as a carboxylic acid or an amino compound, thenit is worthwhile to check the pH. Also, it is worthwhileto check its ability to form a salt – carboxylic acids reactwith sodium carbonate or sodium hydroxide to form acarboxylate salt, and amines react with dilute mineralacids, such as hydrochloric acid, to form amine salts(amine hydrochlorides with hydrochloric acid).

Certain compounds react with water; examples areacid halides and acid anhydrides. Alternatively, these


compounds will react with alcohols and amines to formesters and amides. The formation of such derivatives iscommon practice in the identification of organic com-pounds, and can be extremely informative in a definitiveidentification by infrared spectroscopy. Other chemicalmodifications that can be equally informative are for theidentification of olefinic or unsaturated compounds. Asimple reaction with bromine (caution – use in solution,or prepare in situ) will often confirm the presence ofa double bond. The reaction product can be helpful incharacterizing the nature of the unsaturation, especiallyin the case of a cis configuration, which is often spectrallymasked by other absorptions in the region.

One similar modification, which is strictly not chemicalin nature, is deuterium exchange. In this case, a compoundwith either a labile hydroxy group or an amino group isplaced in equilibrium with a deuterated solvent, such asdeuterium oxide (heavy water). The replacement of theO�H in the molecule by O�D, in the case of hydroxycompounds, can help in the confirmation of the hydroxygroup, and it can shift an overlapping hydroxy functionfrom the C�H stretching and bending regions, therebyassisting in the determination of the backbone structure.Note that the shift induced by the replacement of O�Hby O�D is nearly as much as 1000 cm�1.

Finally, an important chemical characteristic to beunderstood is hydrogen bonding. The infrared spectrumis a wonderful tool for helping to monitor the presenceof this phenomenon. However, one must be aware ofthe fact that hydrogen bonding is usually concentrationdependent. If a hydroxy compound is examined ina nonpolar solvent, one should be aware that thehydroxy group spectrum, in particular bands assignedto the stretching vibration, will change shape, position,and intensity as a function of concentration. Also,diagnostically, if the hydroxy group bands do notchange significantly with concentration, then this is astrong indication of intramolecular (internal) hydrogenbonding. This latter information is important in helpingto determine the spatial relationships of hydroxy groupswithin a molecule.

The last two sections might seem to be obvious basicknowledge in the chemistry and physical properties of amaterial. However, just simple observations of the typedescribed can be invaluable in determining the presenceof specific functional groups when used in combinationwith the infrared spectrum. Likewise, the use of negativespectral information, viz. the absence of a characteristicgroup frequency, is equally important. For example, ifa sample has an almond-like odor, but does not havea strong band between 2200 and 2000 cm�1, then thecompound is more likely to be a nitro compound orbenzaldehyde than a cyano compound. If an aldehydic

carbonyl group is absent, then the compound probablycontains a nitro group.

4.4 The Infrared Sampling Method

The final appearance of the infrared spectrum is veryimportant in the interpretation process. Where possible,it always helps to be able to study the sample as it occursnaturally, without any form of physical modification. Thiseliminates the possibility of any interactions, or evenchemical modifications. A common method for handlingsolid samples is known as the compressed alkali metalhalide pellet method (KBr pellet or disk method). Thismethod of sample preparation is prone to difficulties, andit often requires a skilled operator to produce good-quality spectra on a routine basis. A poor samplingtechnique with this method can lead to spectral artifactsand distortions, all of which can cause difficulties in theinterpretation. Furthermore, some compounds may reactwith KBr, causing either a liberation of the halogen(oxidants) or halogen exchange (halide salts). The actof applying pressure can also change the appearanceof the spectrum, especially when a material exists inmore than one polymorphic form. The pressure maycause one polymorph to dominate, thereby changingthe entire absorption band distribution of the spectrum.This is especially a problem with certain pharmaceuticalproducts.

Other pretreatment methods, such as melts and castfilms for solid samples, can cause similar problems. Thecooling process (from melts) or the casting process caninduce preferential crystallization of a particular form ofthe material, and sometimes preferred orientations canresult. All of these effects can give rise to nonrepresenta-tive spectra, which can cause interpretation problems.

With liquids, it is always beneficial to examine thesample as it is received. Liquid transmission cells havebeen used successfully in the past, but care and attentionmust be paid to the nature of the window material andto the cell pathlength. Low-molecular-weight materialsusually require very short pathlengths in order to obtain agood-quality spectrum with most of the absorption bandson-scale. If too many bands bottom out, it is difficult toascertain the relative intensity of certain key functionalgroups, and sometimes key factors, such as band splitting,may be missed. Also, if the sample is of low molecularweight, and volatile, it is important to ensure that the cellseals adequately.

In recent years, focus has been placed on reflectancemethods of measurement, primarily from the surfaceof a sample. The most common methods are diffusereflectance, specular reflectance and/or transflectance,and ATR. Generally, these are nondestructive meth-ods, and require only a minimum amount of sample


preparation. The specular and transflectance methodshave found greatest use with microscope-based methodsof spectral measurement. Diffuse reflectance and ATRare important laboratory methods, and both lend them-selves to rapid and reproducible sampling. However, bothmethods have an impact on the spectrum, and neithermethod necessarily provides a true spectrum of the sam-ple free from some form of distortion or aberration. ATRis the simplest to handle, and its main limitation is that itprovides a spectrum where the spectrum intensity acrossthe spectrum increases as a function of wavelength. Bothdiffuse reflectance and specular reflectance can be themost difficult to handle, because both methods can causesevere band distortions, linked to surface and refrac-tive index phenomena. In all three cases, the nature andthe origin of the distortions are understood, and can beassigned to known physical properties. If necessary, soft-ware algorithms can be applied to correct for most of thedistortions – a wavelength correction function in the caseof ATR (for qualitative use), and a spectrum transfor-mation, known as the Kramers–Kronig transform, in thecase of surface reflectance distortions.

One other technique that is not so common but isimportant for certain ‘‘difficult-to-handle’’ samples isphotoacoustic spectroscopy. The accessory, which alsofunctions as the detector, provides a true absorptionspectrum from a sample directly, with no samplemodification. It is particularly good for samples thatnaturally have high absorption, such as carbon-filledmaterials, and for materials that are a good thermalconductor. As with ATR spectra, there is a wavelengthdependence of the peak intensity across the spectrum.

The entire subject of sample handling is very broad,and generally falls outside the scope of this article.However, its impact on the final infrared spectrum andits resultant impact on the interpretation process canbe very profound. Therefore, the reader is encouragedto read specific texts that address the subject in greaterdetail..21 – 23/

5 AN OVERVIEW TO INFRAREDSPECTRAL INTERPRETATION – SOMESIMPLE RULES AND GUIDELINES

This final section is a review of the ideas presented so faras to how to go about the interpretation of an infraredspectrum. With it, there is the implied understanding thatthe infrared spectrum is only one source of information,and additional information is most likely required for asatisfactory analysis of the spectrum. The following area set of simple guidelines to help a person through theearly stages of rationalizing an infrared spectrum and to

get to the first stages of an interpretation. However, notethat this is not a definitive set of rules.

First, it is important to appreciate that there are oftenrecurring patterns in the spectrum that are character-istic of certain classes or compounds and/or functionalgroups. Such patterns can be far more diagnostic inthe interpretation process than merely isolated groupfrequencies. Examples include the overall appearanceof the hydrogen-bonded hydroxy functionality, the pat-terns associated with unsaturation and aromaticity, andthe combinations of bands used to characterize differ-ent carbonyl compounds. To start, it is worthwhile todetermine if the compound is organic (evidence of C�Hfunctionality) and if there is a characteristic chain orring associated with the compound. Having determinedthis, it is worthwhile to reflect on the complexity of thespectrum, and to see if one or two characteristic function-alities, associated with simple functional groups (�OH,�NH, CDO, etc.) are present. Often, the like-to-like ruleapplies – that is, a simple spectrum is usually indicative ofa simple molecular structure, and vice versa.

If the spectrum clearly indicates multiple functionality,then be aware that mutual interaction may be occurring,and this may be causing some deviations from ideality,relative to the common group frequency assignments.Also, remember that multiple functionality might reflect acomplex structure, but it might also reflect that the sampleis a mixture. In both cases, an appreciation of the existenceof spectral coincidences is very important. For example,a strong absorption between 1250 and 1150 cm�1 can beassigned to numerous types of chemical bonds, from theC�F group to the Si�O group. A careful assessment of theother spectral features, and most likely an understandingof the nature of the sample, will usually help in resolvingsome of these spectral coincidence situations. The use ofknown relative band intensities for certain functionalgroups can also be helpful in determining whetherband overlap is occurring, and in particular whetherthe bands from a group are being obscured by othermore intense bands. A good example is the overlap ofthe cis C�H out-of-plane bending vibration, which istypically overlapped by the �(CH2)n� rocking vibrationof long-chain aliphatic compounds.

Also, if multiple functionality exists, some form ofseparation scheme may be in order to help determineif the sample is a mixture and, if so, to help separatesome of the components. While this may not alwaysbe necessary – there are many commercial mixturesidentifiable from the original spectrum – it is oftenrequired for a more definitive evaluation of a samplefrom its infrared spectrum. An example of such a mixtureis provided in Figure 11. The spectrum of this four-component mixture can be interpreted from a knowledgeof the components – acetone, methanol, dichloromethane


4000

3600

3200

2800

2400

2000

1800

1600

1400

1200

1000 80

0

100

80

60

40

20

0

Wavenumber (cm−1)

Tran

smitt

ance

(%

)

Figure 11 ATR spectrum of a commercial solvent mixturecontaining methanol, dichloromethane, toluene and acetone.Copyright Coates Consulting.

and toluene. However, the use of separation techniques,such as water-based solvent extraction, and partialdistillation can help in determining the identity of thecomponents.

Finally, in terms of using the first-assessment rulesoutlined in the remainder of this section, it is important toincorporate all known chemical and physical information,including pertinent test data, such as the presence ofunsaturation, which may be used to corroborate thespectral data for a suspected double bond. Rememberthat simple information, such as color and odor, canbe equally important – highly conjugated compoundsand nitro compounds often give rise to significantcolor, and numerous functional groups, such as esters,ketones, unsaturation, and nitro groups, have highlycharacteristic fragrances or odors, even in the solidstate.

The following first assessment guide is based on theexamination of critical spectral regions for evidence ofhighly characteristic molecular backbones or fragmentsand functional groups. An important point to rememberhere is that negative information, in the form ofno-band regions, is extremely important, and helpsto eliminate major classes of chemical compounds,thereby reducing the number of possibilities in the finalinterpretation.

5.1 A Quick Diagnostic Assessment of an InfraredSpectrum

5.1.1 Step 1: Overall Spectrum Appearance

Is the spectrum simple? Are there only a few maincharacteristic absorption bands, say less than five?

If the spectrum is simple, as defined, then thecompound may be a low-molecular-weight organic orinorganic compound, such as a simple salt of a common

molecular ion (carbonate, sulfate, nitrate, ammonium,etc.) or a covalent species (chloroform, dichloromethane,methanol, water, etc.). In this case, the physical propertiesare important – if the material is a crystalline solid, thenit is most likely a salt; if it is a mobile, possibly volatile,liquid, then the latter probably applies. Note that there arealways exceptions, e.g. bromoform and iodoform, simplecovalent compounds, are crystalline solids (both melt atlow temperatures, whereas salts do not). One other classof compounds worth mentioning here is simple polymers,such as polyethylene and polytetrafluoroethylene, bothof which exhibit very simple spectra when obtained fromthin films.

Are some of the bands broad? Hydrogen bond-ing possibly exists and can be exhibited in hydrates,water solutions, alcohols, ammonium compounds, aminocompounds, etc. Note that some inorganics may exhibita combination of both broad and very narrow bands,even for certain hydrated species. These are oftenrelated to the crystalline structure of the compound andthe symmetry of certain aspects of the molecular/ionicstructure.

5.1.2 Diagnostic Step 2: Testing for Organics andHydrocarbons – Absorptions in the Region3200–2700 cm�1

Well-defined and characteristic absorptions in this regionare normally characteristic of carbon- and hydrogen-containing species, and are assigned to various forms ofC�H stretching.

Are there absorptions above 3000 cm�1? If yes, thenthe compound is likely to be unsaturated (contains CDC)or aromatic. If isolated absorptions occur at 3010 and/or3040 cm�1, then the absorbing species is mostly simpleolefinic unsaturation.

Is the main absorption below 3000 cm�1? If so, thecompound is probably aliphatic. If the main absorptionsare approximately 2935 and 2860 cm�1, and there are alsoabsorptions at 1470 and 720 cm�1, then the compoundprobably contains a long linear aliphatic chain.

5.1.3 Diagnostic Step 3: Testing for Hydroxy or AminoGroups – Absorptions in the Region3650–3250 cm�1

Hydroxy or amino groups mainly dominate this region,both giving rise to very characteristic band profiles. Thepresence or absence of hydrogen bonding is well delin-eated in this region. One other characteristic vibrationthat is observed in this region, which is unlikely tobe confused with any other functional group, is theC�H stretch of a terminal alkyne (acetylenic com-pound). This exhibits a relatively narrow absorption at


3300 cm�1, and is associated with a second (C�C) bandat 2200 cm�1.

Is the main absorption band in the area broad? Ifso, the compound is probably a hydroxy or aminogroup. Knowledge of the presence of nitrogen or anitrogenous base (pH or titration) helps to support anamino compound.

Are there additional moderate to intense bands inthe ranges 1600–1300, 1200–1000 and 800–600 cm�1?If so, the compound is likely to be a simple hydroxycompound. Remember that the simple hydrogen-bondedOH absorption of a hydroxy (alcohol) function has a verycharacteristic shape.

If the feature is relatively sharp and occurs between3670 and 3550 cm�1, the compound probably contains anon-hydrogen-bonded hydroxy group, often an alcoholor phenol with a sterically hindered OH group. Notealso that this spectral feature is also exhibited by certaininorganics and minerals, and is indicative of a ‘‘free’’OH group, either on the surface, or embedded within acrystal lattice, and free from interactions with other ionsor groups.

5.1.4 Diagnostic Step 4: Testing for CarbonylCompounds – Absorptions in the Region1850–1650 cm�1

A major band in this region usually indicates the presenceof a CDO group (carbonyl compound). Note that cer-tain CDN (imino) and NDN (azo) groups may interfere,example compounds being purines.

Is the absorption at the high end of the range, e.g.1775 cm�1 or above? If yes, the compound probablycontains a reactive carbonyl, such as an anhydride,acid halide (acyl halide) or a-halogenated carbonyl, ora strained-ring carbonyl, such as a lactone, or an organiccarbonate.

Is the absorption in the middle of the range,1750–1700 cm�1? If yes, the compound is probably a sim-ple carbonyl compound, such as a ketone, an aldehyde,an ester, or a carboxylic acid.

Is the absorption at the low end of the range, e.g. below1700 cm�1? If yes, the compound is probably an amide ora carboxylate (carboxylic acid salt). Note that conjugationwith another carbonyl group, or a double bond oraromatic ring, will lower the carbonyl absorption by30–50 cm�1. Therefore, conjugated aldehydes, ketones,esters, and carboxylic acids may fall into this lower endcategory. If conjugation with a double bond is present,then a second strong absorption should be observednearby, between 1650 and 1600 cm�1. In the case of theinvolvement of an aromatic ring, it is important to lookfor evidence of the characteristic aromatic absorptions(see Diagnostic Step 6).

5.1.5 Diagnostic Step 5: Testing for Unsaturation – Weakto Moderate Absorption in the Region1670–1620 cm�1

Note that this is not a particularly definitive diagnosticwithout the support of other corroborating absorptions.

A relatively narrow, weak-to-moderate absorption,normally centered around 1650 cm�1, is indicative ofolefinic unsaturation (CDC). A lowering of this fre-quency, accompanied by intensification of the band, ischaracteristic of conjugation with another double bondstructure, such as CDC, CDO or an aromatic ring (seecomments in Diagnostic Step 4).

The test for unsaturation normally involves a checkfor C�H peaks above 3000 cm�1, and typically as asingle, or possibly a pair, of absorptions (dependent onconfiguration of the double bond) at ¾3085 cm�1 and/or¾3025 cm�1 – not to be confused with the multiple bandsituation for aromatics (see Diagnostic Step 6).

Are there single or multiple bands between 1000 and880 cm�1 (C�H out-of-plane bending)? If so, this, incombination with the 1650 and 3010 and/or 3040 cm�1

bands, can be used as an assignment of unsaturation.This is either terminal (or vinyl �CHDCH2), with apair of bands, normally located about 990 cm�1 andclose to 900 cm�1, or trans unsaturation (CHDCH),with a band centered around 965–960 cm�1. A singleband around 890 cm�1, in conjunction with the otherunsaturation absorptions, can indicate a vinylidene (orpendant, CDCH2) type of olefinic double bond.

Note that a cis configuration (CHDCH) is sometimeshard to characterize, because the normally characteristicC�H out-of-plane bending is broad, with only a moderateabsorption, and occurs in a region (around 700 cm�1)where it can overlap with absorptions from otherstructural components, such as long-chain aliphatics,aromatics, and halogen compounds.

5.1.6 Diagnostic Step 6: Testing forAromatics – Well-defined Absorptions in theRegion 1615–1495 cm�1

One, but typically two, sets of bands in the region, with oneset around 1600 cm�1 and the other around 1500 cm�1,are consistent with an aromatic compound. Sometimesthe absorptions occur as single bands, at other timesone or more of the band sets may appear split, as anasymmetric doublet. This assignment is usually confirmedby the presence of other bands in the spectrum.

Are there a series of weak-to-moderate absorptionsin the region 3150–3000 cm�1, which can be assignedto aromatic C�H stretching? These usually support thepresence of the 1600/1500 cm�1 aromatic ring bands (seeDiagnostic Step 2).


For simple aromatic compounds, a series of weakabsorptions may be observed between 2000 and1700 cm�1. These are multiple combination bands thatare noticeably weak, and are typically not to be confusedwith an isolated weak carbonyl absorption.

The C�H bending vibrations can also be used tosupport the presence of an aromatic structure. If thereare medium-to-strong absorptions, sometimes more thanone, between 850 and 670 cm�1, these can be assigned toC�H out-of-plane bending on an aromatic ring. Thesesupport the assignment of the 1600/1500 cm�1 bands, andthe locations of the bands are often indicative of thenature of the substitution of the aromatic ring. For simplestructures, it is possible to differentiate mono-, di- (ortho,meta and para), and certain poly-substitutions. Note thata single, strong band may support simple ortho (around750 cm�1) or para (around 830 cm�1) substitution.

Many organic compounds have multiple band struc-tures in the region 1150–950 cm�1. The in-plane C�Hbending vibrations of aromatic compounds typically occurin this region, and can exist as complex band struc-tures (multiple, well-defined bands). They tend not to bediagnostic for many compounds because of conflict andoverlap with other functional group absorptions, includ-ing some skeletal (backbone) vibrations. However, theyare clearly observable in simple hydrocarbons as weak-to-moderate absorptions, and in halogen-substitutedaromatic compounds (such as chlorobenzenes) as charac-teristic strong absorptions.

5.1.7 Diagnostic Step 7: Testing for Multiple Bonding(Often with a Bond Order of 2 orHigher) – Absorption in the Region2300–1990 cm�1

Multiply bonded nitrogen compounds and cumulatedmultiple bond compounds, such as cyanides (nitriles),cyanates, isocyanates, thiocyanates, and diazo com-pounds, exhibit characteristic absorptions in the region2300–1990 cm�1. Cyanide absorptions vary from weak tomoderate to strong, depending on the nature of othersubstituents in the molecule. Conjugation, including sub-stitution on an aromatic ring, modifies the intensity of thisabsorption.

If the absorption band is extremely strong, the groupis more likely to be a three-center functionality, such as acyanate, an isocyanate, or a thiocyanate.

Acetylenes and alkyne substituents also exhibit a weak-to-moderate absorption within the same spectral region.If the acetylenic function is internal (medial), then theabsorption can be extremely weak. However, a terminalalkyne group (C�C) band may have moderate intensity,and this is supported by a characteristic C�H stretchingabsorption around 3310 cm�1. A relatively low-frequency

C�H bending vibration around 630 cm�1 also confirmsthe terminal alkyne vibration.

Metal carbonyls have unique bonding, where thebond order between the carbon and oxygen is betweentwo and three, and where the additional electrons areprovided by dp back-bonding from the accompanyingtransition metal atom. The multiply bonded CO groupprovides an extremely intense absorption band closeto 2000 cm�1 (typically between 2100 and 1900 cm�1),the actual position of the band(s) and the complexityof the bands being dependent on the structure of thecompound.

Note that an isolated weak-to-moderate intensityabsorption above the specified region, and typicallybetween 2700 and 2400 cm�1, is not normally associatedwith a multiply bonded compound. These absorptionsare usually from hydride vibrations, such as from silanes(Si�H), thiols and sulfides (S�H), phosphines (P�H),arsines (As�H), boranes (B�H), etc. Knowledge ofthe existence of the accompanying heteroatom helpsin the assignment of the X�H bonds in these com-pounds.

This concludes the summary data for a first-passassessment of an infrared spectrum, based on themost common and characteristic group frequencies.The technique is not foolproof, but for many simplecompounds, if one systematically applies the diagnostictests on the spectrum from a sample, as presentedabove, it is possible to gain some understanding of thechemical functionality. It is always hoped that moreinformation can be gained from the sample, eitherby first-hand knowledge or by asking questions, or byperforming additional tests. Once a basic interpretationis accomplished, and the sample is broadly characterized,it is recommended that reference spectra are usedto try to obtain a more exact match to the sample.Always bear in mind that the sample could be amixture, such as a commercially formulated/compoundedmaterial. In such cases, an exact match may not bepossible, unless the reference spectrum exists for theformulation.

ABBREVIATIONS AND ACRONYMS

ATR Attenuated Total Reflectance

RELATED ARTICLES

Infrared Spectroscopy (Volume 12)Quantitative Analysis, Infrared ž Theory of InfraredSpectroscopy


REFERENCES

1. H.A. Szymanski, R.E. Erickson, Infrared Band Hand-book, Plenum Press, New York, 2 Volumes, 1970.

2. P. Gans, Vibrating Molecules: an Introduction to theInterpretation of Infrared and Raman Spectra, Chapman& Hall, London, 1975.

3. N.B. Colthrup, L.H. Daly, S.E. Wiberley, Introduction toInfrared and Raman Spectroscopy, Academic Press, SanDiego, CA, 1–73, 1990.

4. J.M. Hollas, Modern Spectroscopy, 3rd edition, JohnWiley & Sons, New York, 1996.

5. J.P. Coates, ‘The Interpretation of Infrared Spectra:Published Reference Sources’, Appl. Spectrosc. Rev.,31(1–2), 179–192 (1996).

6. L.J. Bellamy, Infrared Spectra of Complex Molecules,Chapman & Hall, New York, Vol. 1, 1975.

7. L.J. Bellamy, Advances in Infrared Group Frequencies,Infrared Spectra of Complex Molecules, Chapman & Hall,New York, Vol. 2, 1980.

8. D. Steele, Interpretation of Vibrational Spectra, Chapman& Hall, London, 1971.

9. N.B. Colthrup, L.H. Daly, S.E. Wiberley, Introduction toInfrared and Raman Spectroscopy, Academic Press, SanDiego, CA, 1990.

10. G. Socrates, Infrared Characteristic Group Frequencies,John Wiley & Sons, New York, 1994.

11. D. Lin-Vien, N.B. Colthup, W.G. Fateley, J.G. Grasselli,Infrared and Raman Characteristic Frequencies of OrganicMolecules, Academic Press, San Diego, CA, 1991.

12. B. Smith, Infrared Spectral Interpretation, a SystematicApproach, CRC Press, Boca Raton, FL, 1999.

13. K. Nakamoto, Infrared and Raman Spectra of Inorganicand Coordination Compounds, Parts A and B, John Wiley& Sons, New York, 1997.

14. R.A. Nyquist, C.L. Putzig, M.A. Leugers, Handbook ofInfrared and Raman Spectra of Inorganic Compoundsand Organic Salts, Academic Press, San Diego, CA, 1997.

15. J. Haslam, H.A. Willis, D.C. Squirrel, Identification andAnalysis of Plastics, John Wiley & Sons, New York, 1979;originally published by Iliffe, London, 1972.

16. D.O. Hummel, F.K. Scholl, Atlas of Polymer and PlasticsAnalysis, Verlag Chemie, Weinheim, Vol. 1–3, 1980–81.

17. J.L. Koenig, Spectroscopy of Polymers, American Chem-ical Society, Washington, DC, 1992.

18. J.P. Coates, ‘Questions and Answers: What to Do Whenthe Sample Arrives, Part II’, Spectroscopy, 11(5), 24–32(1996).

19. J.P. Coates, ‘Into the Unknown: the First in the Series’,Spectroscopy, 11(5), 32–39 (1996).

20. J.P. Coates, ‘Once More into the Unknown’, Spec-troscopy, 11(6), 14–23 (1996).

21. J.P. Coates, ‘A Review of Sampling Methods for InfraredSpectroscopy’, in Applied Spectroscopy: a Compact Refer-ence for Practitioners, eds. J. Workman, A.W. Sprinsteen,Academic Press, New York, 49–91, 1998.

22. P.B. Coleman (ed.), Practical Sampling Techniques forInfrared Analysis, CRC Press, Boca Raton, FL, 1993.

23. G. Dent, J.M. Chalmers, Industrial Analysis with Vibra-tional Spectroscopy, Royal Society of Chemistry, Cam-bridge, 1997.

Interpretation of Infrared Spectra, A Practical...

Documents

Transcript of Interpretation of Infrared Spectra, A Practical...