Intro to FTIR Microscopy
The first Infrared spectra were generated using gratings to scan the infrared spectral region, slits to isolate spectral lines and thermopiles for the detection of infrared light. Fourier Transform Infrared (FTIR) spectrometers do not use gratings, but rather spectra are generated in the time domain, following the position of a moving mirror and the occurrence of constructive and destructive interference. A Fast Fourier Transform (FFT) then converts the signal from time to frequency domain. Since the FFT calculation takes time to compute, commercial FTIR instrument use and design closely followed the trend of the miniaturization of computers. By the 1980s gratings-based instruments were phased out for the more desirable FT ones. FTIR has three main advantages over gratings-based measurements. Firstly, since the full spectrum reaches the detector all at once (instead of a small band of radiation as in grating-based spectrometers), the spectrum takes much less time to collect (Fellgett’s Advantage). Secondly, since there are fewer optics and no slits, the total power of each data point is much higher (Jaquinot Advantage). Thirdly, the resolution is much greater than with a grating, enabling spectral subtraction, concentration curves, and library searches among other functions (Conne’s Advantage).
The electromagnetic spectrum consists of different regions corresponding to different energy (E), frequency (ѵ), and wavelength (λ) ranges as seen in Figure 1. The unit for Near-, mid-, and far-infrared, the wavenumber (cm-1), is derived from the inverse relationship between wavelength and frequency.
IR light changes dipole moments (Figure 2) in molecules which correspond to specific vibrational energy. Vibrational energy corresponds to two variables: reduced mass (μ) and bond spring constant (k) (Equation 1). For k constant we can look at C-C, C=C, and C≡C showing an increase of 800 cm-1 across the series (Table 1). Substituting atoms in a C-C bond with nitrogen and oxygen causes a shift of 100 cm-1 (Table 2). By looking at the two series, it can be seen that bond strength alters the wavenumbers more than mass.
Since every functional group is composed of different atoms and bond strengths, vibrations are unique to functional groups and classes of functional groups (e.g. O-H and C-H stretches appear around 3200 and 2900, respectively). A correlation chart with various functional group vibrations can be seen in Figure 3. Since the collection of vibrational energy bands for all of the functional groups a molecule has is unique to every molecule, these peaks can be used for identification in library searches.
The three major parts of an FTIR are the source, the interferometer, and the detector (Figure 4).
The source is a broad band emitting filament such as a mid-IR ceramic source (50-7,800 cm-1), near-IR halogen lamp (2,200 – 25,000 cm-1), or a far-IR mercury lamp (10-700 cm-1). The interferometer is the heart of FTIR and consists of a beamsplitter, a stationary mirror, a moving mirror, and a timing laser (box in figure 4). The beamsplitter splits the light from a source into two paths with half the light going to a stationary mirror and the other half going to a moving mirror. In most FTIR systems the beamsplitter is placed at 45 degrees to the source, but for high throughput applications a 28-degree interferometer is used due to polarization effects close to the Brewster’s Angle. Common beamsplitter materials are KBr (375 – 12,000 cm-1) for mid-IR, Quartz (4,000 – 25,000 cm-1) for Near-IR, and Mylar (30 – 680 cm-1) for Far-IR. The beams from the moving and stationary mirrors are recombined back at the beamsplitter and steered toward the sample. The difference in path of the mirrors causes constructive and destructive interference over the course of time it takes for the moving mirror to make a pass. The signal versus mirror position (and, thus, time) is called an interferogram. A laser is used to determine the position of the moving mirror using the known wavelength of the laser (Figure 5). HeNe lasers are the industry norm due their excellent stability compared to solid state or diode lasers. This laser stability allows for spectral additions, library searches and other functions that need high wavenumber accuracy (Connes Advantage).
The light then is steered through the sample and into a detector whose time domain signal is converted to frequency domain via Fast Fourier Transform by a computer. The power of the beam (Po) is attenuated by the sample by absorbance by the sample (Po), Figure 6. The relationship between power, transmittance, and absorbance can be seen in Equation 3.
Detectors convert photons into measurable electric signals to be sent to the computer. Common detectors include room temperature DLATGS (220 to 15,000 cm-1) for routine analysis, liquid nitrogen cooled MCT (450 to 12,000 cm-1) for high sensitivity applications, Si-photodiodes (10,000 to 25,000 cm-1) for near-IR studies, and silicon bolometers (10 to 650 cm-1) for far-IR studies. A list of sources, windows, beamsplitters, and detectors may be found in Table 3.
The Near-IR portion of the electromagnetic spectrum falls between 4,000 to 12,800 cm-1. This region consists of overtones (two of the same vibrational modes happening simultaneously) and combination (two different vibrational modes happening simultaneously). Since these modes are not strictly quantum mechanically allowed, the intensity of the modes is often quite low. These spectra are often complex, and chemometric techniques, such as multivariate analysis, are used. In spite of the drawbacks, there are clear advantages to Near-IR spectroscopy. Firstly, the path length of the light is such that bulk samples can be analyzed with little to no sample preparation. Secondly, water does not affect signal as it does in IR. These two benefits have been of great value to process chemistry and bulk analysis of incoming/outgoing goods.
The Far Infrared region lies between 10 cm-1 and 700 cm-1. The bonds that show in this region are 3+ atom functional groups, such as -C-C-C- bending, and lattice vibrations in crystalline materials. Since these are highly dependent on conformation or crystal structure, materials with the same chemical structure but different crystal structure may be distinguished using Far-IR. There are two disadvantages to Far IR. Firstly, water absorbs strongly in this region making a purged or evacuated system absolutely necessary. Secondly, the intensity of these modes is weak, so sensitive detectors and high-powered sources are needed.
Attenuated Total Reflectance (ATR) spectroscopy has arisen as the primary sampling method for FTIR spectroscopy. Its major advantage is the lack of sample preparation for liquid and solid samples. When light reflects off of certain materials (diamond, ZnSe, etc.) at a critical angle, the light undergoes total reflectance with a small amount of light being absorbed into material (sample) in contact with the crystal surface, Figure 7. The penetration depth is dependent on the refractive index of both the sample (generally ~1.5) and the crystal itself. Since the refractive index is defendant on wavelength, spectra taken with ATR have slightly different intensity ratios across the spectrum and may need to be corrected to compare to transmission spectra.
The three most common crystals are diamond, zinc selenide, and germanium, each having advantages and disadvantages. Diamond crystals are rugged, have a penetration depth of 1.5 microns, low wavenumber cutoff (200 cm-1) but have poor throughput in the 2200 cm-1 region. ZnSe has exceptional throughput but a high cutoff (650 cm-1). Germanium has a very low penetration depth (0.8 micron) and is useful for highly absorbing substances. The properties and care of common ATR crystals can be found in Table 3.
Transmission is the most straightforward technique, but requires the most sample preparation. The light from the interferometer passes through a sample (Figure 8), with desired beam path through the sample dependent on state (solid, liquid, or gas). For solids, either a KBr pellet containing the sample may be pressed or a diamond anvil cell may be used. For liquids, the sample may be injected into a liquid cell or applied to an IR transparent window or card. Gases may be monitored after being introduced into a gas cell. Often times gas cells are heated to avoid condensation inside the cell.
Specular reflectance makes use of reflected light and is useful for flat, shiny, or samples or coatings. Little or no sample preparation is needed. Sometimes spectra obtained from this technique includes both the imaginary and real components of the light, giving rise to derivative peaks or other distortions. Kramers-Kronig correction may be used to fix distortions.
Diffuse reflectance (DRIFT) makes use of scattered light and is thus useful for rough samples such as powders. Sample preparation can be a solid powder (neat), powder mixed with KBr, or abraded sample on special abrasive paper. To properly compare relative peak intensities from a transmission spectrum, a Kubelka-Munk correction is required.
Functional Group ID – Since every functional group has a specific wavenumber, FTIR may be used to verify the presence of a functional group in a compound.
Unknown ID – Since every molecule has a unique structure that gives a unique spectrum, library databases may be used to match a spectrum of an unknown compound to a compound or mixture.
QA/QC – FTIR is commonly used to inspect incoming and outgoing goods to ensure that products meet purity requirements or to match the correct compound with a process.
Reaction Monitoring – FTIR may be used to monitor reactions such as curing by observing bond formation at specific functional group peaks (e.g. C=O at 1700cm-1).
Quantitative Analysis – FTIR can be used to quantify amount of a compound in a solution or mixture by creating a calibration curve.