17.1: Mid-Infrared Absorption Spectometry

Mid-infrared spectrometry is used for the routine qualitative analysis and, to a lesser extent, the quantitative analysis of organic molecules. In this section we consider absorption spectrometry in which we measure the absorbance of IR light as it passes through a gas, solution, liquid, or solid sample. In Section 17.2 we consider reflectance spectrometry in which we measure the absorbance of IR light as it reflects off the surface of a solid sample or a thin film of a liquid sample.

Sample Handling

Infrared spectroscopy is routinely used to analyze gas, liquid, and solid samples. We know from Beer's law, \(A = \epsilon b C\), that absorbance is a linear function of the analyte's concentration, \(C\), and the distance, \(b\), the light travels through the sample. The challenge with obtaining an IR spectrum, is rarely the analyte's concentration or path length; instead it is finding materials and solvents that are transparent to IR radiation. The optical windows in IR cells are made from materials, such as NaCl and KBr, that are transparent to infrared radiation.

Gas Phase Samples

The cell for analyzing a sample in the gas phase generally is a 5–10 cm glass cylinder fitted with optically transparent windows. For an analyte with a particularly small concentration, the sample cell is designed with reflective surfaces that allow the infrared radiation to make several passes through the cell before it exits the sample cell, increasing the pathlength and, therefore, the absorbance.

Solution Solutions

The analysis of a sample in solution is limited by the solvent’s IR absorbing properties, with carbon tetrachloride, CCl₄, carbon disulfide, CS₂, and chloroform, CHCl₃, being common solvents. A typical solution cell is shown in Figure \(\PageIndex{1}\). It is fashioned with two NaCl windows separated by a spacer. By changing the spacer, pathlengths from 0.015–1.0 mm are obtained. The sample is introduced into the cell using a syringe and the sample inlet port.

Figure \(\PageIndex{1}\): IR cell for analytes that are in solution. (a) View from above showing the sample inlet, the sample outlet, and the NaCl window. (b) View from the side showing the two NaCl plates.

Liquid Phase Samples

A sample that is a volatile liquid may be analyzed using the solution cell in Figure \(\PageIndex{1}\). For a non-volatile liquid sample, however, a suitable sample for qualitative work can be prepared by placing a drop of the liquid between the two NaCl plates shown in Figure \(\PageIndex{2}a\), forming a thin film that typically is less than 0.01 mm thick. An alternative approach is to place a drop of the sample on a disposable card equipped with a polyethylene "window" that IR transparent with the exception of strong absorption bands at 2918 cm^–1 and 2849 cm^–1 (Figure \(\PageIndex{2}b\)).

Figure \(\PageIndex{2}\): Two examples of IR sample cells for liquid samples: (a) NaCl salts plates; (b) disposable card with a polyethylene window that is IR transparent with the exception of strong absorption bands at 2918 cm^–1 and 2849 cm^–1.

Solid Phase Samples

Transparent solid samples are analyzed by placing them directly in the IR beam. Most solid samples, however, are opaque, and are first dispersed in a more transparent medium before recording the IR spectrum. If a suitable solvent is available, then the solid is analyzed by preparing a solution and analyzing as described above. When a suitable solvent is not available, solid samples are analyzed by preparing a mull of the finely powdered sample with a suitable oil and then smearing it on a NaCl salt plate or a disposable IR card (Figure \(\PageIndex{2}\)). Alternatively, the powdered sample is mixed with KBr and pressed, under high pressure, into a thin, optically transparent pellet, as shown in Figure \(\PageIndex{3}\).

Figure \(\PageIndex{3}\): KBr press for preparing a solid sample for an IR analysis. The photo on the left shows the press. A threaded bolt is screwed into the back of the press (hidden from view). A portion of the sample, mixed with KBr, is added to the press and the second threaded bolt is screwed into the press. The two bolts are tightened by turning against each other, forming an optically transparent pellet, as shown by the photo on the right that looks through the pellet. With the threaded bolts removed, the press sits in a holder and placed in the spectrometer's optical path.

Qualitative Analysis

The most important application of mid-infrared spectroscopy is in the qualitative identification of organic molecules. Figure \(\PageIndex{4}\) shows mid-IR solution spectra for four simple alcohols: methanol, CH₃OH, ethanol, CH₃CH₂OH, propanol, CH₃CH₂CH₂OH, and isopropanol, (CH₃)₂CHOH. Clearly there are similarities and differences in these four spectra: similarities that might lead us to expect that each molecule contains the same functional groups and differences that appear as features unique to a particular molecule. The similarities in these four spectra appear at the higher wavenumber end of the x-axis scale; we call the peaks we find there group frequencies. The differences in these four spectra occur below approximately 1500 cm^–1 in what we call the fingerprint region.

Note

The fingerprint region is defined here as beginning at 1500 cm^–1, extending to the lowest wavenumber shown on the x-axis. If you do some searching on the fingerprint region you will see that there is no broad agreement on where it begins. In my searching, I found sources that place the beginning of the fingerprint region as 1500 cm^–1, 1450 cm^–1, 1300 cm^–1, 1200 cm^–1, and 1000 cm^–1.

Figure \(\PageIndex{4}\): IR spectra for four simple alcohols: methanol, CH₃OH, ethanol, CH₃CH₂OH, propanol, CH₃CH₂CH₂OH, and iso-propanol, (CH₃)₂CHOH. Samples were prepared using carbon tetrachloride as a solvent. The vertical dashed red line at 1500 cm^–1 marks the beginning of the fingerprint region. The original data for these spectra are from NIST.

Group Frequencies

All four of the spectra in Figure \(\PageIndex{4}\) share a small intensity, sharp peak at approximately 3650 cm^–1, a strong intensity, broad peak at approximately 3350 cm^–1, and two medium intensity, sharp peaks at 2950 cm^–1 and 3850 cm^–1. By comparing spectra for these and other compounds, we know that the presence of a broad peak between approximately 3200 cm^–1 and 3600 cm^–1 is good evidence that the compound contains a hydrogen-bonded –OH functional group. The sharp peak at approximately 3650 cm^–1 also is evidence of an –OH functional group, but one that is not hydrogen-bonded. The two sharp peaks at 2950 cm^–1 and 3850 cm^–1 are consistent with C–H bonds. All four of these peaks are for stretching vibrations. Tables of group frequencies are routinely available.

The "Fingerprint" Region

Figure \(\PageIndex{5}\) shows a close-up of the fingerprint region for the alcohol samples in Figure \(\PageIndex{4}\). Of particular interest with this set of samples is the increasing complexity of the spectra as we move from the simplest of these alcohols (methanol), to the most complex of these alcohols (propanol and isopropanol). Also of interest is that each spectrum is unique in a way that allows us to confirm a sample by matching it against a library of recorded spectra. There are a number of accessible collections of spectra that are available for this purpose. One such collection of spectra is the NIST Webbook—NIST is the National Institute of Standards and Technology—which is the source of the data used to display the spectra included in this section's figures and which includes spectra for over 16,000 compounds.

Figure \(\PageIndex{5}\): IR spectra in the fingerprint region for four simple alcohols: methanol, CH₃OH, ethanol, CH₃CH₂OH, propanol, CH₃CH₂CH₂OH, and iso-propanol, (CH₃)₂CHOH. Samples were prepared using carbon tetrachloride as a solvent. The original data for these spectra are from NIST.

Computer Search Systems

With the availability of computerized data acquisition and storage it is possible to build digital libraries of standard reference spectra. The identity of an a unknown compound often can be determined by comparing its spectrum against a library of reference spectra, a process known as spectral searching. Comparisons are made using an algorithm that calculates the cumulative difference between the sample’s spectrum and a reference spectrum. For example, one simple algorithm uses the following equation

\[D = \sum_{i = 1}^n | (A_{sample})_i - (A_{reference})_i | \label{spec_sub} \]

where D is the cumulative difference, A_sample is the sample’s absorbance at wavelength or wavenumber i, A_reference is the absorbance of the reference compound at the same wavelength or wavenumber, and n is the number of digitized points in the spectra. Note that the spectra are defined here by absobrance instead of transmittance as absorbance is directly proportional to concentration. The cumulative absolute difference is calculated for each reference spectrum. The reference compound with the smallest value of D is the closest match to the unknown compound. The accuracy of spectral searching is limited by the number and type of compounds included in the library, and by the effect of the sample’s matrix on the spectrum.

Another advantage of computerized data acquisition is the ability to subtract one spectrum from another. When coupled with spectral searching it is possible to determine the identity of several components in a sample without the need of a prior separation step by repeatedly searching and subtracting reference spectra. An example is shown in Figure \(\PageIndex{6}\) in which the composition of a two-component mixture is determined by successive searching and subtraction. Figure \(\PageIndex{6}a\) shows the spectrum of the mixture. A search of the spectral library selects cocaine•HCl (Figure \(\PageIndex{6}b\)) as a likely component of the mixture. Subtracting the reference spectrum for cocaine•HCl from the mixture’s spectrum leaves a result (Figure \(\PageIndex{6}c\)) that closely matches mannitol’s reference spectrum (Figure \(\PageIndex{6}d\)). Subtracting the reference spectrum for mannitol leaves a small residual signal (Figure \(\PageIndex{6}e\)).

Figure \(\PageIndex{6}\): Identifying the components of a mixture by spectral searching and subtracting. (a) IR spectrum of the mixture; (b) Reference IR spectrum of cocaine•HCl; (c) Result of subtracting the spectrum of cocaine•HCl from the mixture’s spectrum; (d) Reference IR spectrum of mannitol; and (e) The residual spectrum after removing mannitol’s contribution to the mixture’s spectrum. IR spectra traditionally are displayed using percent transmittance, %T, along the y-axis. Because absorbance—not percent transmittance—is a linear function of concentration, spectral searching and spectral subtraction, is easier to do when displaying absorbance on the y-axis.

Quantitative Applications

A quantitative analysis based on the absorption of infrared radiation, although important, is encountered less frequently than with UV/Vis absorption, primarily due to the three issues raised here.

Deviations from Beer's Law

One challenge for quantitative IR is the greater tendency for instrumental deviations from Beer’s law when using infrared radiation. Because an infrared absorption band is relatively narrow, any deviation due to the lack of monochromatic radiation is more pronounced. In addition, infrared sources are less intense than UV/Vis sources, which makes stray radiation more of a problem. Differences between the path lengths for samples and for standards when using thin liquid films or KBr pellets are a problem, although an internal standard can correct for any difference in pathlength; alternatively, we can use the cell shown in Figure \(\PageIndex{1}\) to maintain a constant path length.

Background Correction

The water and carbon dioxide in air have strong absorbances in the mid-IR. A double-beam dispersive instrument corrects for the contributions of CO₂ and H₂O vapor because they are present in both pathways through the instrument. An FT-IR, however, includes only a single optical path, so it is necessary to collect a separate spectrum to compensate for the absorbance of atmospheric CO₂ and H₂O vapor. This is done by collecting a background spectrum without the sample and storing the result in the instrument’s computer memory. The background spectrum is removed from the sample’s spectrum by taking the ratio the two signals. Another approach is to flush the sample compartment with nitrogren.

Measuring Absorbance

Another challenge for quantitative IR is that establishing a 100% T (A = 0) baseline often is difficult because the optical properties of NaCl sample cells may change significantly with wavelength due to contamination and degradation. We can minimize this problem by measuring absorbance relative to a baseline established for the absorption band. Figure \(\PageIndex{7}\) shows how this is accomplished.

Figure \(\PageIndex{7}\): Method for determining absorbance from an IR spectrum.

Typical Applications

A recent review paper [Fahelelbom, K. M.; Saleh, A.; Al-Tabakha, M. A.; Ashames, A. A. Rev. Anal. Chem. 2022, 41, 21–33] summarizes the rich literature in quantitative mid-infrared spectrometry. Among the areas covered are the analysis of pharmaceuticals, including antibiotics, antihypertensives, antivirals, and counterfeit drugs. Mid-infrared spectrometry also finds use for the analysis of environmentally significant gases, such as methane, CH₄, hydrogen chloride, HCl, sulfur dioxide, SO₂, and nitric oxide, NO.