Brian C. Smith

We continue our survey of the infrared (IR) spectroscopy of the carbonyl bond with a look at carboxylic acids. These molecules have C=O, C-O, and O-H bonds, giving them a number of diagnostically useful group wavenumbers and making this functional group easy to spot. We also see how the spectra of saturated and aromatic carboxylic acids differ from each other. In the "IR Spectral Interpretation Review" sidebar, we look at how the strength of intermolecular interactions affect IR peak widths, which is germane to carboxylic acids because they engage strongly in hydrogen bonding.

	This is the third installment in our examination of the infrared (IR) spectroscopy of the carbonyl or C=O bond, yet we have barely scratched the surface of this versatile functional group. So, you can look forward to several more articles on this ubiquitous and important chemical bond. In this column, we examine the spectra of carboxylic acids. Future installments will focus on carboxylic acid derivatives such as carboxylates, acid anhydrides, and esters.


	Infrared Spectral Interpretation Review: Infrared Spectral Peak Widths

	The carboxylic acid functional group contains a carbonyl bond flanked by a carbon, called the 

, and an -OH or hydroxyl group as shown in Figure 1.


	Figure 1: The structural framework of the carboxylic acid functional group. Note the locations of the alpha carbon and the hydroxyl group.

	If the alpha carbon is saturated, the carboxylic acid is said to be saturated. If the alpha carbon is part of an aromatic ring, then we have an aromatic acid.

	Carboxylic acids are considered acids because in aqueous solution they can ionize to give a proton and a negatively charged carboxylate (CO

) ion. However, they are weak acids, easily illustrated by the fact that a carboxylic acid, called 

 is found in vinegar and we eat it regularly, whereas if would be deadly to eat the much stronger mineral acids such as sulfuric, hydrochloric, or nitric.

	In the first installment of this column published in January 2015 (1), I discussed that the one characteristic that determines peak widths in IR spectra is the strength of intermolecular bonding. Briefly, weak intermolecular interactions give narrow peaks while strong intermolecular interactions give broad peaks. This is explained further in the "Infrared Spectral Interpretation Review" sidebar. Any functional group that contains a hydroxyl or -OH bond will generally engage in strong hydrogen bonding. Carboxylic acids are no exception to this. In fact, the hydrogen bonding in these molecules is so strong that the solid and liquid states of carboxylic acid are not found as individual molecules but are bound up in what are called 


	Figure 2: An example of a carboxylic acid dimer, a result of the strong hydrogen bonding that takes place in the solid and liquid forms of these molecules.

IR Spectra of Saturated and Aromatic Carboxylic Acids

	Thus, when the IR spectra of solid and liquid carboxylic acids are measured it is the spectrum of the dimer that is actually seen. Not surprisingly then, the IR peaks of carboxylic acids are unusually broad as shown in the spectrum of 2-methyl butyric acid in Figure 3.


	Figure 3: The IR spectrum of 2-methylbutyric acid, C

	The O-H stretch, labeled A in Figure 3, is so broad that there isn't really a peak position for it, just a range in which this wide envelope falls, from 3500 to 2500 cm

 (going forward all peak positions quoted will be in cm

). Note that the C-H stretches from the methyl and methylene groups in this molecule are narrow peaks that fall on top of the much broader O-H stretching envelope. It's not unusual in IR spectra for narrow peaks to show up on top of or as a shoulder on broader peaks when both happen to fall in the same wavenumber range such as this. This spectral region is another example of how peak widths can be used to distinguish functional groups from each other, with the nonpolar C-H bond stretches being much narrower than the polar O-H stretch.

	In my experience, spectral regions like the one around 3000 in Figure 3 can be tricky to interpret. As trained spectral interpreters our eyes tend to look for C-H stretching peaks because they are so common, and in Figure 3 these peaks are clearly visible. What is tricky is that by focusing on these narrow peaks you may miss that they are riding on a broad spectral envelope and that there is an O-H bond present. I point this out because it's a mistake I have made myself.

	Note in Figure 3 that there are some wiggles and bumps on the low wavenumber side of the OH stretching peak from around 2800 to 2500, which are labeled B. These features are common in the spectra of carboxylic acids, and are caused by overtone and combination bands. As discussed last time (5) these features are generally weak; however, recall that peak intensity in IR spectra is related to dµ/d

, the dipole moment change with respect to bond length during a vibration (1). For highly polar molecules, such as carboxylic acids, dµ/d

 for the fundamental bands is large, and it is relatively large for overtone and combination bands, hence the intensity of the peaks from 2800 to 2500.


	Quiz Section: Answer to the November Quiz and a New Interpretation Challenge

	Like alcohols (3), which have an OH stretching and wagging peak, carboxylic acids also have an O-H wagging peak. In Figure 3 it is at 942, labeled F, and is generally found from 960 to 900. Note how broad this peak is because of hydrogen bonding. There are very few other functional groups that have broad peaks around 900, making this peak diagnostically useful.

	Besides OH stretching and bending peaks, carboxylic acids of course have a carbonyl stretch. In Figure 3 it is seen at 1709, and in general for saturated carboxylic acids this peak falls from 1730 to 1700. Note that the C=O stretches of saturated ketones and aldehydes also fall in this range. Thus, as has been pointed out before (4), the position of carbonyl stretches by themselves is not enough to identify specific C=O containing functional groups, other peaks will be needed. In the case of carboxylic acids, it is the combination of a broad -OH stretch and a C=O stretching peak that gives us a diagnostically useful pattern. Other molecules, such as water and alcohols, have broad OH stretching peaks but no C=O peak. Many functional groups have C=O bonds, but most of them do not have OH stretching peaks, which is why this diagnostic pattern is useful.

	A close look at Figure 1 shows that carboxylic acids also have a C-O bond. We know from previous experience (4) that C-O stretching peaks generally fall from 1300 to 1000. Carboxylic acids are no exception to this general rule. The C-O stretching peak in Figure 3 falls at 1230 and is labeled E. In general, for carboxylic acids this peak appears between 1320 and 1210. Carboxylic acids are unusual in that they give so many intense diagnostically useful peaks. Specifically, the OH stretch, C=O stretch, C-O stretch, and OH wag.

	The IR spectrum of an aromatic carboxylic acid, benzoic acid, is seen in Figure 4.


	Figure 4: The IR spectrum of benzoic acid, C

	I will be the first to admit that this spectrum is scary-it contains many intense peaks and on the surface appears difficult to interpret. This problem arises because as we study molecules with more and more functional groups in them their spectra will get more complex. Also, the highly polar nature of the carboxylic acid groups causes many peaks to appear with more intensity than they would in a less polar molecule because of large values of dµ/d

. However, using the 12-step process already taught to you on how to systematically interpret a spectrum (6), the analysis of Figure 4 can be tackled.

	The first thing that should stand out in Figure 4 is the broad envelope labeled A that extends from 3500 to 2000; it makes up almost half the spectrum and is one of the broadest IR peaks you will ever see! This is of course the OH stretch of the carboxylic acid functional group. The C-H stretching peak of the aromatic ring is a sharp shoulder at 3071. As in Figure 3, there are overtone and combination bands in the 2800 to 2500 neighborhood labeled B in Figure 4. The C=O stretch of benzoic acid is at 1685, and in general for aromatic carboxylic acids this peak falls from 1710 to 1680, lower than for saturated acids, because of conjugation (3). The C-O stretch of benzoic acid is at 1292, and the O-H wag at 934, which is in line with the peak positions quoted above. Note that the only peak position here sensitive to whether a carboxylic acid is saturated or aromatic is the C=O stretch. The diagnostic group wavenumbers for all carboxylic acids are listed in Table I.

	We have seen that carboxylic acids have broad OH stretching and bending peaks because of hydrogen bonding. The OH stretch is a broad envelope from 3500 to 2500, while the OH wag falls from 960 to 900. The carbonyl stretch for carboxylic acids falls from 1730 to 1700 for saturated acids and 1710 to 1680 for aromatic acids because of conjugation. The C-O stretch of these molecules is found between 1320 and 1210. The diagnostically useful pattern worth remembering to detect a carboxylic acid in a spectrum is a broad OH stretch combined with a C=O stretch. The broad OH wag around 900 is also unique and diagnostically useful.

Infrared Spectral Interpretation: A Systematic Approach

 has more than three decades of experience as an infrared spectroscopist. He has published numerous peer reviewed papers and has written three books on the subject: 

Quantitative Spectroscopy: Theory and Practice

 published by Elsevier. As a spectroscopic trainer, he has helped thousands of people around the world improve their infrared analyses. He earned his PhD in physical chemistry from Dartmouth College. He can be reached at: 

Articles

How to spot carboxylic acids in your IR spectra

The C=O Bond, Part III: Carboxylic Acids

The carbonyl or C=O group is the perfect functional group for detection by infrared (IR) spectroscopy because its stretching vibration peak is intense and is located in a unique wavenumber range. In this introduction to the IR spectroscopy of the carbonyl group we explore why the peak is intense, and see how to apply that knowledge to the analysis of the spectra of ketones.

	Across all the installments I have written so far there has been an overarching structure: Each large section has been devoted to the infrared (IR) spectroscopy of a specific chemical bond. We started with the C-H bond, moved on to the C-O bond, and now it is time to discuss the spectroscopy of the C=O or carbonyl group. These bonds are very common and are found in ketones, aldehydes, esters, and carboxylic acids, among others. The types of materials where you will find carbonyl groups include polymers, proteins, fats, solvents, and pharmaceuticals.

	The carbon in a C=O bond is referred to as the “carbonyl carbon” as shown in Figure 1.

	Carbonyl bonds are highly polar because of the large electronegativity difference between carbon and oxygen. As a result, the carbonyl carbon has a large partial positive charge and the oxygen has a large partial negative charge as denoted in Figure 1. Recall that in a chemical bond when there are two charges separated by a distance it forms what is called a 

 (1). Additionally, remember that one of the characteristics that determines the intensity of infrared peaks is the change in dipole moment with respect to bond length, dµ/d

	Since the carbonyl group has a large dipole moment, when this group stretches and contracts the value of dµ/d

 is large, thereby giving an intense peak. This vibration is illustrated in Figure 2.

	Carbonyl stretching peaks generally fall between 1900 and 1600 cm

 (assume all peak positions hereafter are in wavenumber units), a relatively unique part of the IR spectrum. This area is sometimes referred to as the 

 as a result. The carbonyl stretching peak is perhaps the perfect example of a group wavenumber, a peak that shows up intensely in a unique and relatively narrow wavenumber range. This attribute makes C=O stretches one of the easiest IR peaks to spot and assign. Because of this trait, an IR spectrometer is, in some respects, the perfect carbonyl detector.

	Carbonyl bonds can be divided into two classes depending on what type of carbons are attached to the carbonyl carbon. A 

 has two saturated carbons attached to the carbonyl carbon. An 

 has one or more aromatic carbons attached to the carbonyl carbon. An example of an aromatic carbonyl group is shown in Figure 3.

	Saturated and aromatic carbonyl groups can be distinguished from each other using IR spectroscopy. In general, aromatic C=O stretching peaks fall ~30 cm

 lower than saturated C=O stretching peaks. This is because aromatic rings, such as benzene, contain p-type orbitals with electron density that sticks up out of the plane of the molecule (2) as illustrated to the left in Figure 3. The carbonyl group also contains a p-type orbital that points through space toward the orbitals on the aromatic ring, Figure 3. The orbitals are close enough that they somewhat overlap, allowing some of the electron density from the carbonyl bond to be pulled off into the benzene ring in a phenomenon known as 

, which is illustrated at right in Figure 3. Conjugation weakens the C=O bond, lowers its force constant, and hence its peak position is lowered, on average about 30 cm

 (1). Thus, for almost every carbonyl containing functional group we discuss I will quote two carbonyl stretching regions, one for the saturated and one for the aromatic versions of the functional group.

	Perhaps one of the most common chemical structures to contain a carbonyl group are ketone molecules. Ketones consist of a C=O with two carbons attached, one to the left and one to the right as illustrated in Figure 4.

	Note that the carbons attached to the carbonyl carbon are denoted “alpha carbons.”

	A common example of a ketone-containing molecule is acetone (dimethyl ketone), whose IR spectrum is shown in Figure 5.

	The carbonyl stretch of acetone falls at 1716, and is labeled A in Figure 5. In general, for saturated ketones this peak appears at 1715 ± 10. Note how large this peak is, easily the biggest in the spectrum. This is common for the carbonyl stretches of many functional groups.

	It is tempting to assume that the carbonyl stretching peak for ketones uniquely identifies them, but this is not the case. Saturated esters, aldehydes, and carboxylic acids, amongst others, have their C=O stretching peaks around 1715. Thus, we will see that the C=O stretching peak is diagnostic for carbonyl groups but not for anything else. We will be dependent upon other peaks in IR spectra besides the C=O stretch to distinguish the different types of carbonyl containing functional groups from each other.

	For ketones, the peak that can help distinguish them from other functional groups is the C-C-C stretch vibration, which is illustrated in Figure 6.

	Note that this vibration involves the two alpha carbons stretching asymmetrically about the carbonyl carbon. This vibration gives rise to an intense peak between 1230 and 1100 for saturated ketones. This peak appears in the spectrum of acetone at 1222 and is labeled B. Note it is the second largest peak in the spectrum.

	Normally a C-C stretching vibration peak is small because the electronegativity difference between two carbon atoms is often negligible, giving nonpolar bonds and small values of dµ/d

. However, the carbonyl carbon in the C=O bond has a large positive partial charge on it, as shown in Figure 1. This charge polarizes the bonds to the two alpha carbons, and when these C-C bonds stretch dµ/d

 is large, giving the typically large ketone C-C-C stretching peak labeled B in Figure 5.

	The chemical structure of an aromatic ketone, acetophenone, is shown in Figure 7. Note that although there is an aromatic and a saturated carbon attached to the carbonyl carbon in acetophenone, it is considered an aromatic ketone. This is because it only takes the presence of a single aromatic substituent to engage in conjugation and lower the carbonyl stretching peak position. The IR spectrum of acetophenone (phenyl methyl ketone) is also shown in Figure 7.

	The carbonyl stretch, labeled A, is at 1686, while the C-C-C stretch, labeled B, is at 1266. For aromatic ketones, generally the C-C-C stretch falls between 1300 and 1230. Note that these are the two most intense peaks in the spectrum. Structurally acetophenone differs from acetone because a methyl group has been replaced by a phenyl group. Thus, the C=O stretching peak in acetophenone is 30 cm

 lower than in acetone because of conjugation. A summary of the group wavenumbers for ketones is shown in Table I.

	Note that both the C=O stretch and the C-C-C stretch are sensitive as to whether a ketone is saturated or aromatic, so either one of these peaks can be used to distinguish saturated and aromatic ketones from each other. The C=O and C-C-C stretching peaks are the best peaks to use to determine if there is a ketone in a sample.

	A methyl ketone is a structure where one or more of the substituents on the carbonyl carbon is a methyl group. Both acetone and acetophenone are examples of methyl ketones. Previously we learned (2) that methyl groups in alkyl chains have an umbrella mode around 1375, and we also observed that this peak was medium to weak in intensity. The methyl groups in acetone and acetophenone also exhibit an umbrella mode at 1362 and 1360 respectively, labeled C in Figures 5 and 7. Note how intense these two umbrella mode peaks are: In both cases they are the third biggest peak in the spectrum, and in the case of acetone the umbrella mode is almost as big as the C-C-C stretch!

 umbrella mode is indicative of a methyl ketone. Thus, if your analysis of a spectrum indicates that a ketone and methyl group are present, and the CH

 umbrella mode is unusually large, it is a good bet you have a methyl ketone. This peak is diagnostically very useful. After a ketone is identified the next step is to determine the two substituents attached to the carbonyl carbon. The large methyl ketone umbrella mode tells us that one of the substituents is a CH

	The carbonyl group, C=O, has a large dipole moment and thus an intense stretching vibration whose peak generally appears from 1900 to 1600. When an aromatic ring is attached to a carbonyl carbon, the C=O stretching peak position is lowered by about 30 cm

 Ketones are a type of carbonyl containing functional group with two carbons attached to the carbonyl carbon. They exhibit diagnostically useful C=O and C-C-C stretches, both of whose peak positions are sensitive as to whether the ketone is saturated or aromatic. Methyl ketones display a unique high-intensity CH

, is the West Coast Business Development Manager for CAMO Software, a company that sells chemometric, multivariate analysis, and process control software. Before joining CAMO, Dr. Smith ran his own FT-IR training and consulting business for more than 20 years. Dr. Smith has written three books on infrared spectroscopy: 

 published by Academic Press. He can be reached at: 

An introduction to the IR spectroscopy of the carbonyl group, exploring why the peak is intense and showing how to apply that knowledge to the analysis of the spectra of ketones

The Carbonyl Group, Part I: Introduction

We finish our examination of the infrared (IR) spectroscopy of the C-O bond with a discussion of carbohydrates. These important biological molecules contain both alcohol and ether linkages, making their C-O stretching regions particularly rich and challenging to interpret. To round out this potpourri I’ve added alkynes, a type of unsaturated functional group for which there wasn’t room in earlier columns. 

 is “a mixture, especially of unrelated objects, subjects etc.” One of the challenges of writing multiple column installments for a magazine is making the topics fit into the space allotted in each issue. More than once in the past there has been a topic that was too short to stand by itself as an installment, but too long to combine with any related topics. Hence the need for the occasional catch-all column like this one, where these orphaned topics can find a home. The reason that this installment is a potpourri is that the topics included here are unrelated to each other. There will be more potpourri columns in the future.

	This is also the 14th installment of this workshop, and in the last two and a half years we have covered a lot of territory. The topics covered in potpourri installments will be new to you but are related to things we have discussed before. Therefore, you will learn something new from these columns while exercising some of what you have already learned, hopefully aiding in your learning process.

	Carbohydrates are an important class of biological compounds including sugars and cellulose. The term was originally coined because their molecular formulas suggested they were “hydrates of carbon” (2), although we now know this is not true. Structurally speaking, carbohydrates contain saturated rings with ether linkages in them, and multiple alcohol groups as shown in Figure 1.

	So far in these columns most of the compounds whose spectra have been examined have been synthetic compounds of industrial importance, but of course biochemicals are important too because they allow life on planet earth to exist! In many of the infrared (IR) spectral interpretation training courses I have taught, one of the most frequent questions has been, “Can IR tell the difference between synthetic and natural materials?” The short answer is no. Many of the most important classes of biochemicals, including proteins, fats, carbohydrates, and nucleic acids, are structurally similar to organic compounds and contain the same functional groups as synthetic chemicals. The spectrum of any functional group will be the same regardless of whether it was made in a living cell or a test tube. For example, the spectrum of a methyl group will be pretty much the same whether it is part of an alkane or a protein. Thus, IR spectroscopy is not a good “natural versus synthetic” discriminator. However, since biochemicals do contain organic functional groups, we can determine their structure by IR spectroscopy the same way we do for synthetic materials.

	The chemical structure of an important and well known carbohydrate, glucose, is shown in Figure 1. 

	Glucose is the product of photosynthesis, and is the molecule many animals burn for energy. Note that it is a cyclic six-membered ring with the shape of a distorted hexagon. Glucose is an example of a monosaccharide, meaning it contains one cyclic sugar unit. Five of the vertices of this saturated ring represents a carbon atom, and the ring also contains an oxygen atom attached to two saturated carbons giving a saturated ether linkage. Additionally, there are also five -OH groups, four of which are secondary alcohols and one of which is a primary alcohol. We have studied the spectra of all these functional groups before (3–5). The plethora of hydroxyl groups means that carbohydrates engage extensively in hydrogen bonding and that there will be strong, broad O-H stretches. The seven separate saturated C-O bonds means there will be a lot of C-O stretching peaks between 1200 and 1000 cm

 (going forward assume all peak positions are in cm

 even if not noted). The IR spectrum of glucose is shown in Figure 2.

	There is a broad, strong O-H stretching envelope in Figure 2 with two peaks at 3402 and 3318. This is the first time we have seen an O-H stretching peak like this. It means there are two chemically distinct alcohol groups in the molecule with different bond strengths, force constants, and hence peak positions. My guess is that one of the peaks is from the secondary alcohols present, and the other from the primary alcohol present. There is also an O-H wagging peak at 606.

	There is a peak at 1622 from adsorbed water, although the peak is small and water’s contribution to the -OH stretching peak intensity would also be small. It is no surprise that glucose and many other carbohydrates absorb water from the atmosphere since they contain a lot of -OH groups, are highly polar, and hence are hygroscopic.

	Four of the carbons in the glucose ring are structurally similar-each being attached to two other carbons, one oxygen, and one hydrogen, with the oxygens each being part of a secondary alcohol. We will call the situation where there is only one hydrogen attached to a carbon a 

. The lone hydrogens in glucose are all saturated and hence will have C-H stretches below 3000 (2). However, these peaks will not be as intense as typical C-H stretches since there is only one bond contributing to the peak intensity instead of two or three as is the case for methylene and methyl groups. Saturated lone C-H groups tend to have their stretching peaks around 2900 (7). The oddball fifth carbon in the glucose ring has a primary alcohol group hanging off it that contains a CH

 moieties that are responsible for the small C-H stretching peaks shown in Figure 2 around 2900.

	Recall that saturated ethers have a C-O stretching peak from 1140 to 1070, secondary alcohols have one from 1150 to 1075, and in primary alcohols this peak appears between 1075 and 1000 (3–5), making for a potentially complex C-O stretching region. As can be seen in Figure 2 there are four peaks in this region. This C-O stretching peak region complexity is typical of carbohydrates, and makes it difficult to assign the C-O stretching peaks of these molecules exactly.

	The IR spectrum of a disaccharide, sucrose, otherwise known as table sugar, is shown in Figure 3.

	Sucrose is a disaccharide because it contains two sugar rings. It consists of a glucose ring attached via a saturated ether linkage to a five-membered fructose ring as noted in Figure 3. Fructose gets its name from the fact that it is a sugar found in fruit (2). Note in the spectrum that there are multiple O-H and C-O stretching peaks, even more so than in the spectrum of glucose. Many carbohydrates consist of multiple sugar rings connected by ether linkages. As these molecules get bigger there are many more types of O-H and C-O bonds, hence the increase in spectral complexity. Also note in the spectrum of sucrose that the C-H stretch and O-H wag peaks have turned into shapeless blobs thanks to spectral complexity and hydrogen bonding.

 are polymeric carbohydrates consisting of many sugar units strung together. Perhaps the most common polysaccharide is cellulose, whose IR spectrum is shown in Figure 4.

	Cellulose is found in the cell walls of plants and is found in our every day lives as wood, paper, and cotton. In fact, the spectrum in Figure 4 is of a paper towel. The cellulose molecule is a series of glucose units stitched together by saturated ether linkages. There is so much hydrogen bonding in cellulose that the complex O-H stretching peak patterns we saw in glucose and fructose are gone, replaced by a broad blob. The C-H stretches are similarly blobby, and the C-O stretching region is best characterized as an envelope with any number of shoulder peaks on top of it.

	Notice then as we proceeded from mono- to di- to polysaccharides the spectra got more complex because of the many types of O-H and C-O bonds and eventually became so complex that the peaks overlapped enough to form blobs. Hydrogen bonding also helped increase spectral band broadening. The spectra of carbohydrates then are characterized by multiple O-H stretches and multiple C-O stretches. This information is summarized in Table I.

	In past columns (10,11) we discussed unsaturated carbons and looked at the spectra of two families of unsaturated hydrocarbons, substituted benzene rings and alkenes. I left out one last family from the discussion, alkynes. Alkynes contain the C≡C structural unit, properly called a carbon-carbon triple bond. These bonds are unsaturated because they can be hydrogenated to form saturated compounds just like aromatic rings and alkenes (2). The two varieties of alkynes, monosubstituted and disubstituted, are shown in Figure 5. The “R-” symbols in Figure 5 represent nonhydrogen substituent atoms.

	A monosubstituted alkyne has one hydrogen and one R-group attached, and a disubstituted alkyne has no hydrogens and two R-groups attached. The simplest alkyne with the chemical formula C

, or H-C≡C-H, is called acetylene. Sometimes alkynes are named as acetylene derivatives. We can distinguish the two types of alkynes from each other using IR spectroscopy. The IR spectrum of a monosubstituted alkyne, 3,3-dimethyl-1-butyne (or 

-butyl acetylene) is shown in Figure 6, and its relevant peak assignments are shown in Table II.

	As we read the spectrum from left to right the first thing we encounter is an intense, relatively sharp peak labeled A at 3311. We have seen in the past (3,4) that O-H stretches fall in this same region. However, recall that these peaks are widened by hydrogen bonding. There is no hydrogen bonding in alkynes, hence the peak at 3311 is narrow and cannot be from an O-H bond. It is in fact the C-H stretch of the hydrogen attached directly to the C≡C group, and this peak generally falls from 3350 to 3250 for monosubstituted alkynes. This is far and away the highest wavenumber C-H stretch we have seen yet and may ever see. Its unique position and sharpness set it apart and make it a great monosubstituted alkyne group wavenumber. The H-C≡C group also has a C-H wag, shown as a big peak labeled C at 631 in Figure 6. This is also an excellent group wavenumber because of its intensity, sharpness, and low peak position. This peak can be found generally from 700 to 600 for monosubstituted alkynes. O-H wags also show up in this region (3,4), but again the sharpness of this peak means it is not from an OH group. The spectra of monosubstituted alkynes then have two sharp, intense peaks at high and low wavenumber, in a sense bookending the spectrum.

	The C≡C linkage in monosubstituted alkynes has a stretching peak between 2140 and 2100, which is shown in Figure 6 labeled B at 2106. This is a very unusual peak position, which should make this peak a useful group wavenumber. However, like the C=C stretch of alkenes (11) this peak has variable intensity and hence limited utility because it may be intense, weak, or nonexistent depending on the molecule. Recall that one of the things that determines peak intensity is the change in dipole moment with respect to bond length (dµ/d

) for a vibration (8). For a monosubstituted alkyne with a polar substituent dµ/d

 will be large and the C≡C stretching peak will be large. For a monosubstituted alkyne with a nonpolar substituent like 

-butyl acetylene, dµ/dx will be small and hence the peak intensity will be small as seen in Figure 6. For a symmetrically substituted alkyne such as acetylene (H-C≡C-H) dµ/d

 for the triple bond stretch is zero and there is no IR peak due to C≡C stretching. The unreliable appearance of this peak makes it a poor group wavenumber. Again, the best peaks to identify a monosubstituted alkyne are the H-C≡C stretch and wag.

	A disubstituted alkyne has two non-hydrogen substituents attached, such as dimethyl acetylene, CH

. Since there are no hydrogens attached to the triple bond, disubstituted alkynes do not have a H-C≡C stretching or bending vibration. Disubstituted alkynes do have a C≡C stretching peak that falls between 2260 and 2190, and this peak can be used to distinguish disubstituted from monosubstituted alkynes. However, just like for monosubstituted alkynes, the C≡C stretch is also an unreliable group wavenumber for the same reasons. For a symmetrically substituted disubstituted alkyne like dimethyl acetylene there are no H-C≡C peaks, and since dµ/d

 for the C≡C stretch is zero there is no triple bond stretching peak. Symmetrically substituted alkynes then are one of the rare organic functional groups that have no useful IR group wavenumbers, and their presence in a sample must be detected by means other than IR spectroscopy.

	In this first IR spectral interpretation potpourri we have looked at the spectra of two dissimilar molecules, carbohydrates and alkynes. Carbohydrates contain many different types of O-H and C-O bonds, and hence have complex O-H stretching and C-O stretching regions from 3500 to 3300 and 1200 to 1000, respectively.

	Alkynes contain C≡C bonds and come in mono- and disubstituted forms. Mono-substituted alkynes contain a unique H-C≡C group with a sharp, intense high wavenumber C-H stretch, and a sharp, intense, low wavenumber C-H wag. C≡C stretching peaks appear uniquely around 2200 but suffer from variable intensity, making them a questionable group wavenumber. This is the only diagnostic peak that disubstituted alkynes possess, sometimes making them difficult to detect.

 (G.C. Merriam, Cambridge, Massachusetts, 1946).

		Beilstein Electronic Organic Chemical Database, contents as of 6/2017.

, PhD, is the West Coast Business Development Manager for CAMO Software, a company that sells chemometric, multivariate analysis, and process control software. Before joining CAMO, Dr. Smith ran his own FT-IR training and consulting business for more than 20 years. Dr. Smith has written three books on infrared spectroscopy: 

Two unrelated discussions are presented: carbohydrates and alkynes.

An IR Spectral Interpretation Potpourri: Carbohydrates and Alkynes

Our survey of the spectroscopy of the C-O bond continues where we complete our discussion of alcohols and discuss how to distinguish pure alcohols, alcohol–water mixtures, and pure water from each other.

	In the last column (1) I introduced you to the spectroscopy of the C-O bond, and we covered how to interpret the spectra of primary alcohols. As I discussed then, in addition to primary alcohols, secondary alcohols, tertiary alcohols, and phenols can be analyzed by infrared spectroscopy as well. The chemical structures of primary, secondary, and tertiary alcohols are shown in Figure 1.

	Remember that in general alcohols have a broad, strong O-H stretch at 3350 ± 50 cm

 (assume all peak positions stated after this are in cm

 units), an in-plane -OH bend at 1350 ± 50, and an O-H wag at 650 ± 50 (1). Lastly, primary alcohols typically have a C-C-O asymmetric stretch (hereinafter called the C-O stretch) between 1000 and 1075. As we will learn in this installment, the key to distinguishing these alcohols from each other is the position of the C-O stretching peak.

	The Spectra of Secondary and Tertiary Alcohols

	A spectrum of a secondary alcohol, called isopropyl or “rubbing” alcohol, is shown in Figure 2, and its peak assignments are shown in Table I. As we have learned (1), the infrared peaks of alcohols are broadened because of hydrogen bending and are hence easy to spot. In Figure 2 the OH stretch is labeled A at 3349, the in-plane bend labeled C appears at 1309, and the -OH wag labeled G appears at 655. Recall that the C-O stretch is frequently the largest peak between 1300 and 1000 (1). Following this rule we can assign the peak labeled D at 1129 as the C-O stretch of isopropyl alcohol. Note that this is higher in wavenumber than the range quoted for primary alcohols of 1075 to 1000. For secondary alcohols generally the C-O stretch falls between 1150 and 1075. Thus, a C-O stretch below 1075 can be assigned as a primary alcohol, and a C-O stretch above 1075 can be assigned as a secondary alcohol. The C-C-O symmetric stretch of isopropyl alcohol is at 817 and is labled F in Figure 2. Peak B is a “split” umbrella mode due to the branch point in isopropyl alcohol. More on this type of peak and branch points in a future column. Peak E is from a C-C stretching vibration. Carbon-carbon stretching vibration peaks are typically not this intense. However, the oxygen attached to the carbon atoms polarizes these bonds, increasing the change in dipole moment with respect to distance during the vibration, resulting in an increase in peak intensity.

	The spectrum of a tertiary alcohol, tert-butanol, is seen in Figure 3. The OH peaks fall as expected, with the stretch at 3371, the in-plane bend at 1366, and wag at 648. Note that 

-butanol also has a split umbrella mode with peaks at 1379 and 1366 because of its branch point. The reasonably intense peak at 914 is the symmetric C-C-O stretch. The biggest peak between 1300 and 1000 is at 1202 and is assigned as the C-O stretch. In general for tertiary alcohols this peak falls between 1210 and 1100. Unfortunately, this range overlaps significantly with the range for secondary alcohols. Using the C-O stretching peak position as noted in Table II, primary and secondary alcohols can be distinguished, primary and tertiary alcohols can be distinguished, but secondary and tertiary alcohols may or may not be distinguished depending upon where their C-O stretches fall. Unfortunately, there are no other useful IR peaks to solve this problem. Other information, such as a 

C nuclear magnetic resonance (NMR) spectrum, may be needed to distinguish secondary and tertiary alcohols from each other.

	As stated in the last column (1), phenols contain an OH group attached to an aromatic ring. The spectrum of the namesake molecule of this class, phenol, is shown in Figure 4.

	This spectrum is one of the busiest we have seen so far, and I included it not to confuse you, but to be real. As the molecules we study include more and more functional groups, their spectra will become more complicated and unfortunately more challenging to interpret. Consider studying this spectrum good preparation for what is to come in future columns.

	The OH stretch at 3345 and in-plane bend at 1367 can be easily seen in Figure 4. This latter peak is in the same wavenumber range as methyl group umbrella modes; however, note that there are no methyl C-H stretches below 3000, and the 1367 peak is too broad to be from a CH

 group. The largest peak between 1300 and 1000 is at 1231, and is assigned as the C-O stretch. In general, for phenols this peak falls between 1260 and 1200. Usually phenols can be distinguished from primary and secondary alcohols based on the position of their C-O stretch, and can often be distinguished from tertiary alcohols unless the C-O stretch falls from 1210 to 1200. In Figure 4 this is the case, but we know this is not a saturated alcohol because of the lack of C-H stretches between 3000 and 2800.

	We would expect the spectrum of phenol to have an OH wag at 650 ± 50, but it is not obvious on an initial examination of Figure 4. However, if you look closely there is a broad envelope in this region with sharper benzene ring peaks on top of it. So the wag is there-it’s just hard to assign a peak position to it. This is often a problem for phenols, so be on the lookout for broad envelopes like this one at low wavenumber if you suspect a phenol is present in a sample.

	Table II summarizes the peak positions for the various alcohols.

	The infrared spectrum of liquid water is shown in Figure 5. Both water and alcohols have an OH bond and thus have OH stretching peaks that are intense, broad, and fall in the same place around 3300. Because of this shared spectral feature, the presence of an OH stretch in a spectrum is not diagnostic for the presence of an alcohol, only for the presence of an OH group. If you see an OH stretch then how can you distinguish the spectra of a pure alcohol, water, or a mixture of the two? The solution to this problem is Table III.

	Note in Figure 5 that water has a scissors bending peak near 1630. This peak occurs because in water the oxygen has two hydrogens attached to it. By definition alcohols only have one O-H bond, and hence their spectra will not exhibit this peak. Also remember that water is inorganic and does not have a C-O bond, thus it will not have the C-O stretching peak that the spectra of all alcohols exhibit. This is the information that was used to construct Table III, where the presence or absence of these peaks can be used to distinguish these three types of samples from each other. For example, pure water has an OH stretch and a scissors peak, but no C-O stretch. A pure alcohol has an OH stretch and no scissors peak, but does have a C-O stretching peak. Lastly, a mixture of an alcohol and water has an OH stretch, a scissors peak from the water present, and a C-O stretch because of the alcohol present. Water–alcohol mixtures are relatively common because some alcohols are hygroscopic and are readily soluble in water. A mnemonic for remembering Table III is that for pure water the pattern is yes yes no, for pure alcohols it is yes no yes, and for a mixture of the two it is yes yes yes. This table should allow you to distinguish these types of samples from each other.

	The spectra of alcohols and phenols typically exhibit a large peak between 1300 and 1000 that can be assigned as the asymmetric C-C-O stretch (“C-O stretch”). For primary alcohols this peak falls from 1075 to 1000, for secondary alcohols from 1150 to 1075, for tertiary alcohols from 1210 to 1100, and for phenols from 1260 to 1200. Thus the peak range for primary alcohols is unique and they are easily distinguished from all other secondary alcohols can be distinguished from primary alcohols but their C-O stretching has some overlap with tertiary alcohols, and phenols can usually be distinguished unless their C-O stretching peak falls between 1210 and 1200. Information about alcohol C-O stretching peak positions is summarized in Table II. The spectra of water, pure alcohols, and their mixtures can be distinguished using the presence or absence of the water scissors peak at 1630 and the alcohol C-O stretch between 1300 and 1000, as detailed in Table II.

 is the West Coast Business Development Manager for CAMO Software, a company that sells chemometric, multivariate analysis, and process control software. Before joining CAMO, Dr. Smith ran his own FT-IR training and consulting business for more than 20 years. Dr. Smith has written three books on infrared spectroscopy: 

In addition to primary alcohols there exist secondary and tertiary alcohols.

Brian C. Smith

Articles

The C=O Bond, Part III: Carboxylic Acids

The Carbonyl Group, Part I: Introduction

An IR Spectral Interpretation Potpourri: Carbohydrates and Alkynes

Alcohols—The Rest of the Story