Raman spectroscopy

From Wikipedia, the free encyclopedia

https://en.wikipedia.org/wiki/Raman_spectroscopy 

 

Energy-level diagram showing the states involved in Raman spectra.

Raman spectroscopy (/ˈrɑːmən/); (named after Indian physicist C. V. Raman) is a spectroscopic technique typically used to determine vibrational modes of molecules, although rotational and other low-frequency modes of systems may also be observed. Raman spectroscopy is commonly used in chemistry to provide a structural fingerprint by which molecules can be identified.

Raman spectroscopy relies upon inelastic scattering of photons, known as Raman scattering. A source of monochromatic light, usually from a laser in the visible, near infrared, or near ultraviolet range is used, although X-rays can also be used. The laser light interacts with molecular vibrations, phonons or other excitations in the system, resulting in the energy of the laser photons being shifted up or down. The shift in energy gives information about the vibrational modes in the system. Infrared spectroscopy typically yields similar, complementary, information.

Typically, a sample is illuminated with a laser beam. Electromagnetic radiation from the illuminated spot is collected with a lens and sent through a monochromator. Elastic scattered radiation at the wavelength corresponding to the laser line (Rayleigh scattering) is filtered out by either a notch filter, edge pass filter, or a band pass filter, while the rest of the collected light is dispersed onto a detector.

Spontaneous Raman scattering is typically very weak; as a result, for many years the main difficulty in collecting Raman spectra was separating the weak inelastically scattered light from the intense Rayleigh scattered laser light (referred to as "laser rejection"). Historically, Raman spectrometers used holographic gratings and multiple dispersion stages to achieve a high degree of laser rejection. In the past, photomultipliers were the detectors of choice for dispersive Raman setups, which resulted in long acquisition times. However, modern instrumentation almost universally employs notch or edge filters for laser rejection. Dispersive single-stage spectrographs (axial transmissive (AT) or Czerny–Turner (CT) monochromators) paired with CCD detectors are most common although Fourier transform (FT) spectrometers are also common for use with NIR lasers.

The name "Raman spectroscopy" typically refers to vibrational Raman using laser wavelengths which are not absorbed by the sample. There are many other variations of Raman spectroscopy including surface-enhanced Raman, resonance Raman, tip-enhanced Raman, polarized Raman, stimulated Raman, transmission Raman, spatially-offset Raman, and hyper Raman.

Theory

The magnitude of the Raman effect correlates with polarizability of the electrons in a molecule. It is a form of inelastic light scattering, where a photon excites the sample. This excitation puts the molecule into a virtual energy state for a short time before the photon is emitted. Inelastic scattering means that the energy of the emitted photon is of either lower or higher energy than the incident photon. After the scattering event, the sample is in a different rotational or vibrational state.

For the total energy of the system to remain constant after the molecule moves to a new rovibronic (rotational-vibrational-electronic) state, the scattered photon shifts to a different energy, and therefore a different frequency. This energy difference is equal to that between the initial and final rovibronic states of the molecule. If the final state is higher in energy than the initial state, the scattered photon will be shifted to a lower frequency (lower energy) so that the total energy remains the same. This shift in frequency is called a Stokes shift, or downshift. If the final state is lower in energy, the scattered photon will be shifted to a higher frequency, which is called an anti-Stokes shift, or upshift.

For a molecule to exhibit a Raman effect, there must be a change in its electric dipole-electric dipole polarizability with respect to the vibrational coordinate corresponding to the rovibronic state. The intensity of the Raman scattering is proportional to this polarizability change. Therefore, the Raman spectrum (scattering intensity as a function of the frequency shifts) depends on the rovibronic states of the molecule.

The Raman effect is based on the interaction between the electron cloud of a sample and the external electric field of the monochromatic light, which can create an induced dipole moment within the molecule based on its polarizability. Because the laser light does not excite the molecule there can be no real transition between energy levels. The Raman effect should not be confused with emission (fluorescence or phosphorescence), where a molecule in an excited electronic state emits a photon and returns to the ground electronic state, in many cases to a vibrationally excited state on the ground electronic state potential energy surface. Raman scattering also contrasts with infrared (IR) absorption, where the energy of the absorbed photon matches the difference in energy between the initial and final rovibronic states. The dependence of Raman on the electric dipole-electric dipole polarizability derivative also differs from IR spectroscopy, which depends on the electric dipole moment derivative, the atomic polar tensor (APT). This contrasting feature allows rovibronic transitions that might not be active in IR to be analyzed using Raman spectroscopy, as exemplified by the rule of mutual exclusion in centrosymmetric molecules. Transitions which have large Raman intensities often have weak IR intensities and vice versa. If a bond is strongly polarized, a small change in its length such as that which occurs during a vibration has only a small resultant effect on polarization. Vibrations involving polar bonds (e.g. C-O , N-O , O-H) are therefore, comparatively weak Raman scatterers. Such polarized bonds, however, carry their electrical charges during the vibrational motion, (unless neutralized by symmetry factors), and this results in a larger net dipole moment change during the vibration, producing a strong IR absorption band. Conversely, relatively neutral bonds (e.g. C-C , C-H , C=C) suffer large changes in polarizability during a vibration. However, the dipole moment is not similarly affected such that while vibrations involving predominantly this type of bond are strong Raman scatterers, they are weak in the IR. A third vibrational spectroscopy technique, inelastic incoherent neutron scattering (IINS), can be used to determine the frequencies of vibrations in highly symmetric molecules that may be both IR and Raman inactive. The IINS selection rules, or allowed transitions, differ from those of IR and Raman, so the three techniques are complementary. They all give the same frequency for a given vibrational transition, but the relative intensities provide different information due to the different types of interaction between the molecule and the incoming particles, photons for IR and Raman, and neutrons for IINS.

History

Although the inelastic scattering of light was predicted by Adolf Smekal in 1923, it was not observed in practice until 1928. The Raman effect was named after one of its discoverers, the Indian scientist C. V. Raman, who observed the effect in organic liquids in 1928 together with K. S. Krishnan, and independently by Grigory Landsberg and Leonid Mandelstam in inorganic crystals. Raman won the Nobel Prize in Physics in 1930 for this discovery. The first observation of Raman spectra in gases was in 1929 by Franco Rasetti.

Systematic pioneering theory of the Raman effect was developed by Czechoslovak physicist George Placzek between 1930 and 1934.^[5] The mercury arc became the principal light source, first with photographic detection and then with spectrophotometric detection.

In the years following its discovery, Raman spectroscopy was used to provide the first catalog of molecular vibrational frequencies. Typically, the sample was held in a long tube and illuminated along its length with a beam of filtered monochromatic light generated by a gas discharge lamp. The photons that were scattered by the sample were collected through an optical flat at the end of the tube. To maximize the sensitivity, the sample was highly concentrated (1 M or more) and relatively large volumes (5 mL or more) were used.

Raman shift

Raman shifts are typically reported in wavenumbers, which have units of inverse length, as this value is directly related to energy. In order to convert between spectral wavelength and wavenumbers of shift in the Raman spectrum, the following formula can be used:

${\displaystyle \Delta {\tilde {\nu }}=\left({\frac {1}{\lambda _{0}}}-{\frac {1}{\lambda _{1}}}\right)\ ,}$

where ${\displaystyle \Delta {\tilde {\nu }}}$ is the Raman shift expressed in wavenumber, ${\displaystyle \lambda _{0}}$ is the excitation wavelength, and ${\displaystyle \lambda _{1}}$ is the Raman spectrum wavelength. Most commonly, the unit chosen for expressing wavenumber in Raman spectra is inverse centimeters (cm⁻¹). Since wavelength is often expressed in units of nanometers (nm), the formula above can scale for this unit conversion explicitly, giving

${\displaystyle \Delta {\tilde {\nu }}({\text{cm}}^{-1})=\left({\frac {1}{\lambda _{0}({\text{nm}})}}-{\frac {1}{\lambda _{1}({\text{nm}})}}\right)\times {\frac {(10^{7}{\text{nm}})}{({\text{cm}})}}.}$

Instrumentation

An early Raman spectrum of benzene published by Raman and Krishnan.

 

Schematic of one possible dispersive Raman spectroscopy setup.

Modern Raman spectroscopy nearly always involves the use of lasers as excitation light sources. Because lasers were not available until more than three decades after the discovery of the effect, Raman and Krishnan used a mercury lamp and photographic plates to record spectra. Early spectra took hours or even days to acquire due to weak light sources, poor sensitivity of the detectors and the weak Raman scattering cross-sections of most materials. Various colored filters and chemical solutions were used to select certain wavelength regions for excitation and detection but the photographic spectra were still dominated by a broad center line corresponding to Rayleigh scattering of the excitation source.

Technological advances have made Raman spectroscopy much more sensitive, particularly since the 1980s. The most common modern detectors are now charge-coupled devices (CCDs). Photodiode arrays and photomultiplier tubes were common prior to the adoption of CCDs. The advent of reliable, stable, inexpensive lasers with narrow bandwidths has also had an impact.

Lasers

Raman spectroscopy requires a light source such as a laser. The resolution of the spectrum relies on the bandwidth of the laser source used. Generally shorter wavelength lasers give stronger Raman scattering due to the ν⁴ increase in Raman scattering cross-sections, but issues with sample degradation or fluorescence may result.

Continuous wave lasers are most common for normal Raman spectroscopy, but pulsed lasers may also be used. These often have wider bandwidths than their CW counterparts but are very useful for other forms of Raman spectroscopy such as transient, time-resolved and resonance Raman.

Detectors

Raman scattered light is typically collected and either dispersed by a spectrograph or used with an interferometer for detection by Fourier Transform (FT) methods. In many cases commercially available FT-IR spectrometers can be modified to become FT-Raman spectrometers.

Detectors for dispersive Raman

In most cases, modern Raman spectrometers use array detectors such as CCDs. Various types of CCDs exist which are optimized for different wavelength ranges. Intensified CCDs can be used for very weak signals and/or pulsed lasers. The spectral range depends on the size of the CCD and the focal length of spectrograph used.

It was once common to use monochromators coupled to photomultiplier tubes. In this case the monochromator would need to be moved in order to scan through a spectral range.

Detectors for FT–Raman

FT–Raman is almost always used with NIR lasers and appropriate detectors must be used depending on the exciting wavelength. Germanium or Indium gallium arsenide (InGaAs) detectors are commonly used.

Filters

It is usually necessary to separate the Raman scattered light from the Rayleigh signal and reflected laser signal in order to collect high quality Raman spectra using a laser rejection filter. Notch or long-pass optical filters are typically used for this purpose. Before the advent of holographic filters it was common to use a triple-grating monochromator in subtractive mode to isolate the desired signal. This may still be used to record very small Raman shifts as holographic filters typically reflect some of the low frequency bands in addition to the unshifted laser light. However, Volume hologram filters are becoming more common which allow shifts as low as 5 cm⁻¹ to be observed.

Applications

Raman spectroscopy is used in chemistry to identify molecules and study chemical bonding and intramolecular bonds. Because vibrational frequencies are specific to a molecule's chemical bonds and symmetry (the fingerprint region of organic molecules is in the wavenumber range 500–1500 cm⁻¹), Raman provides a fingerprint to identify molecules. For instance, Raman and IR spectra were used to determine the vibrational frequencies of SiO, Si₂O₂, and Si₃O₃ on the basis of normal coordinate analyses. Raman is also used to study the addition of a substrate to an enzyme.

In solid-state physics, Raman spectroscopy is used to characterize materials, measure temperature, and find the crystallographic orientation of a sample. As with single molecules, a solid material can be identified by characteristic phonon modes. Information on the population of a phonon mode is given by the ratio of the Stokes and anti-Stokes intensity of the spontaneous Raman signal. Raman spectroscopy can also be used to observe other low frequency excitations of a solid, such as plasmons, magnons, and superconducting gap excitations. Distributed temperature sensing (DTS) uses the Raman-shifted backscatter from laser pulses to determine the temperature along optical fibers. The orientation of an anisotropic crystal can be found from the polarization of Raman-scattered light with respect to the crystal and the polarization of the laser light, if the crystal structure’s point group is known.

In nanotechnology, a Raman microscope can be used to analyze nanowires to better understand their structures, and the radial breathing mode of carbon nanotubes is commonly used to evaluate their diameter.

Raman active fibers, such as aramid and carbon, have vibrational modes that show a shift in Raman frequency with applied stress. Polypropylene fibers exhibit similar shifts.

In solid state chemistry and the bio-pharmaceutical industry, Raman spectroscopy can be used to not only identify active pharmaceutical ingredients (APIs), but to identify their polymorphic forms, if more than one exist. For example, the drug Cayston (aztreonam), marketed by Gilead Sciences for cystic fibrosis, can be identified and characterized by IR and Raman spectroscopy. Using the correct polymorphic form in bio-pharmaceutical formulations is critical, since different forms have different physical properties, like solubility and melting point.

Raman spectroscopy has a wide variety of applications in biology and medicine. It has helped confirm the existence of low-frequency phonons in proteins and DNA, promoting studies of low-frequency collective motion in proteins and DNA and their biological functions. Raman reporter molecules with olefin or alkyne moieties are being developed for tissue imaging with SERS-labeled antibodies. Raman spectroscopy has also been used as a noninvasive technique for real-time, in situ biochemical characterization of wounds. Multivariate analysis of Raman spectra has enabled development of a quantitative measure for wound healing progress. Spatially offset Raman spectroscopy (SORS), which is less sensitive to surface layers than conventional Raman, can be used to discover counterfeit drugs without opening their packaging, and to non-invasively study biological tissue. A huge reason why Raman spectroscopy is so useful in biological applications is because its results often do not face interference from water molecules, due to the fact that they have permanent dipole moments, and as a result, the Raman scattering cannot be picked up on. This is a large advantage, specifically in biological applications. Raman spectroscopy also has a wide usage for studying biominerals. Lastly, Raman gas analyzers have many practical applications, including real-time monitoring of anesthetic and respiratory gas mixtures during surgery.

Raman spectroscopy has been used in several research projects as a means to detect explosives from a safe distance using laser beams.

Raman Spectroscopy is being further developed so it could be used in the clinical setting. Raman4Clinic is a European organization that is working on incorporating Raman Spectroscopy techniques in the medical field. They are currently working on different projects, one of them being monitoring cancer using bodily fluids such as urine and blood samples which are easily accessible. This technique would be less stressful on the patients than constantly having to take biopsies which are not always risk free.

Art and cultural heritage

Raman spectroscopy is an efficient and non-destructive way to investigate works of art and cultural heritage artifacts, in part because it is a non-invasive process which can be applied in situ. It can be used to analyze the corrosion products on the surfaces of artifacts (statues, pottery, etc.), which can lend insight into the corrosive environments experienced by the artifacts. The resulting spectra can also be compared to the spectra of surfaces that are cleaned or intentionally corroded, which can aid in determining the authenticity of valuable historical artifacts.

It is capable of identifying individual pigments in paintings and their degradation products, which can provide insight into the working method of an artist in addition to aiding in authentication of paintings. It also gives information about the original state of the painting in cases where the pigments have degraded with age.

In addition to paintings and artifacts, Raman spectroscopy can be used to investigate the chemical composition of historical documents (such as the Book of Kells), which can provide insight about the social and economic conditions when they were created. It also offers a noninvasive way to determine the best method of preservation or conservation of such cultural heritage artifacts, by providing insight into the causes behind deterioration.

The IRUG (Infrared and Raman Users Group) Spectral Database is a rigorously peer-reviewed online database of IR and Raman reference spectra for cultural heritage materials such as works of art, architecture, and archaeological artifacts. The database is open for the general public to peruse, and includes interactive spectra for over a hundred different types of pigments and paints.

Microspectroscopy

Hyperspectral Raman imaging can provide distribution maps of chemical compounds and material properties: Example of an unhydrated clinker remnant in a 19th-century cement mortar (cement chemist's nomenclature: C ≙ CaO, A ≙ Al₂O₃, S ≙ SiO₂, F ≙ Fe₂O₃).

Raman spectroscopy offers several advantages for microscopic analysis. Since it is a light scattering technique, specimens do not need to be fixed or sectioned. Raman spectra can be collected from a very small volume (< 1 µm in diameter, < 10 µm in depth); these spectra allow the identification of species present in that volume. Water does not generally interfere with Raman spectral analysis. Thus, Raman spectroscopy is suitable for the microscopic examination of minerals, materials such as polymers and ceramics, cells, proteins and forensic trace evidence. A Raman microscope begins with a standard optical microscope, and adds an excitation laser, a monochromator or polychromator, and a sensitive detector (such as a charge-coupled device (CCD), or photomultiplier tube (PMT)). FT-Raman has also been used with microscopes, typically in combination with near-infrared (NIR) laser excitation. Ultraviolet microscopes and UV enhanced optics must be used when a UV laser source is used for Raman microspectroscopy.

In direct imaging (also termed global imaging or wide-field illumination), the whole field of view is examined for light scattering integrated over a small range of wavenumbers (Raman shifts). For instance, a wavenumber characteristic for cholesterol could be used to record the distribution of cholesterol within a cell culture. This technique is being used for the characterization of large scale devices, mapping of different compounds and dynamics study. It has already been use for the characterization of graphene layers, J-aggregated dyes inside carbon nanotubes and multiple other 2D materials such as MoS₂ and WSe₂. Since the excitation beam is dispersed over the whole field of view, those measurements can be done without damaging the sample.

The most common approach is hyperspectral imaging or chemical imaging, in which thousands of Raman spectra are acquired from all over the field of view by, for example, raster scanning of a focused laser beam through a sample. The data can be used to generate images showing the location and amount of different components. Having the full spectroscopic information available in every measurement spot has the advantage that several components can be mapped at the same time, including chemically similar and even polymorphic forms, which cannot be distinguished by detecting only one single wavenumber. Furthermore, material properties such as stress and strain, crystal orientation, crystallinity and incorporation of foreign ions into crystal lattices (e.g., doping, solid solution series) can be determined from hyperspectral maps. Taking the cell culture example, a hyperspectral image could show the distribution of cholesterol, as well as proteins, nucleic acids, and fatty acids. Sophisticated signal- and image-processing techniques can be used to ignore the presence of water, culture media, buffers, and other interferences.

Because a Raman microscope is a diffraction-limited system, its spatial resolution depends on the wavelength of light, the numerical aperture of the focusing element, and — in the case of confocal microscopy — on the diameter of the confocal aperture. When operated in the visible to near-infrared range, a Raman microscope can achieve lateral resolutions of approx. 1 µm down to 250 nm, depending on the wavelength and type of objective lens (e.g., air vs. water or oil immersion lenses). The depth resolution (if not limited by the optical penetration depth of the sample) can range from 1-6 µm with the smallest confocal pinhole aperture to 10s of micrometers when operated without a confocal pinhole. Depending on the sample, the high laser power density due to microscopic focussing can have the benefit of enhanced photobleaching of molecules emitting interfering fluorescence. However, the laser wavelength and laser power have to be carefully selected for each type of sample to avoid its degradation.

Applications of Raman imaging range from materials sciences to biological studies. For each type of sample, the measurement parameters have to be individually optimized. For that reason, modern Raman microscopes are often equipped with several lasers offering different wavelengths, a set of objective lenses, and neutral density filters for tuning of the laser power reaching the sample. Selection of the laser wavelength mainly depends on optical properties of the sample and on the aim of the investigation. For example, Raman microscopy of biological and medical specimens is often performed using red to near-infrared excitation (e.g., 785 nm, or 1064 nm wavelength). Due to typically low absorbances of biological samples in this spectral range, the risk of damaging the specimen as well as autofluorescence emission are reduced, and high penetration depths into tissues can be achieved. However, the intensity of Raman scattering at long wavelengths is low (owing to the ω⁴ dependence of Raman scattering intensity), leading to long acquisition times. On the other hand, resonance Raman imaging of single-cell algae at 532 nm (green) can specifically probe the carotenoid distribution within a cell by a using low laser power of ~5 µW and only 100 ms acquisition time.

Raman scattering, specifically tip-enhanced Raman spectroscopy, produces high resolution hyperspectral images of single molecules atoms, and DNA.

Polarization dependence of Raman scattering

Raman scattering is polarization sensitive and can provide detailed information on symmetry of Raman active modes. While conventional Raman spectroscopy identifies chemical composition, polarization effects on Raman spectra can reveal information on the orientation of molecules in single crystals and anisotropic materials, e.g. strained plastic sheets, as well as the symmetry of vibrational modes.

Polarization–dependent Raman spectroscopy uses (plane) polarized laser excitation from a polarizer. The Raman scattered light collected is passed through a second polarizer (called the analyzer) before entering the detector. The analyzer is oriented either parallel or perpendicular to the polarization of the laser. Spectra acquired with the analyzer set at both perpendicular and parallel to the excitation plane can be used to calculate the depolarization ratio. Typically a polarization scrambler is placed between the analyzer and detector also. It is convenient in polarized Raman spectroscopy to describe the propagation and polarization directions using Porto's notation, described by and named after Brazilian physicist Sergio Pereira da Silva Porto.

For isotropic solutions, the Raman scattering from each mode either retains the polarization of the laser or becomes partly or fully depolarized. If the vibrational mode involved in the Raman scattering process is totally symmetric then the polarization of the Raman scattering will be the same as that of the incoming laser beam. In the case that the vibrational mode is not totally symmetric then the polarization will be lost (scrambled) partially or totally, which is referred to as depolarization. Hence polarized Raman spectroscopy can provide detailed information as to the symmetry labels of vibrational modes.

In the solid state, polarized Raman spectroscopy can be useful in the study of oriented samples such as single crystals. The polarizability of a vibrational mode is not equal along and across the bond. Therefore the intensity of the Raman scattering will be different when the laser's polarization is along and orthogonal to a particular bond axis. This effect can provide information on the orientation of molecules with a single crystal or material. The spectral information arising from this analysis is often used to understand macro-molecular orientation in crystal lattices, liquid crystals or polymer samples.

Characterization of the symmetry of a vibrational mode

The polarization technique is useful in understanding the connections between molecular symmetry, Raman activity, and peaks in the corresponding Raman spectra. Polarized light in one direction only gives access to some Raman–active modes, but rotating the polarization gives access to other modes. Each mode is separated according to its symmetry.

The symmetry of a vibrational mode is deduced from the depolarization ratio ρ, which is the ratio of the Raman scattering with polarization orthogonal to the incident laser and the Raman scattering with the same polarization as the incident laser: ${\displaystyle \rho ={\frac {I_{r}}{I_{u}}}}$ Here ${\displaystyle I_{r}}$ is the intensity of Raman scattering when the analyzer is rotated 90 degrees with respect to the incident light's polarization axis, and ${\displaystyle I_{u}}$ the intensity of Raman scattering when the analyzer is aligned with the polarization of the incident laser. When polarized light interacts with a molecule, it distorts the molecule which induces an equal and opposite effect in the plane-wave, causing it to be rotated by the difference between the orientation of the molecule and the angle of polarization of the light wave. If ρ ≥ ${\displaystyle {\frac {3}{4}}}$, then the vibrations at that frequency are depolarized; meaning they are not totally symmetric.

Variants

At least 25 variations of Raman spectroscopy have been developed. The usual purpose is to enhance the sensitivity (e.g., surface-enhanced Raman), to improve the spatial resolution (Raman microscopy), or to acquire very specific information (resonance Raman).

Spontaneous (or far-field) Raman spectroscopy

Correlative Raman imaging: Comparison of topographical (AFM, top) and Raman images of GaSe. Scale bar is 5 μm.

Terms such as spontaneous Raman spectroscopy or normal Raman spectroscopy summarize Raman spectroscopy techniques based on Raman scattering by using normal far-field optics as described above. Variants of normal Raman spectroscopy exist with respect to excitation-detection geometries, combination with other techniques, use of special (polarizing) optics and specific choice of excitation wavelengths for resonance enhancement.

• Correlative Raman imaging – Raman microscopy can be combined with complementary imaging methods, such as atomic force microscopy (Raman-AFM) and scanning electron microscopy (Raman-SEM) to compare Raman distribution maps with (or overlay them onto) topographical or morphological images, and to correlate Raman spectra with complementary physical or chemical information (e.g., gained by SEM-EDX).
• Resonance Raman spectroscopy – The excitation wavelength is matched to an electronic transition of the molecule or crystal, so that vibrational modes associated with the excited electronic state are greatly enhanced. This is useful for studying large molecules such as polypeptides, which might show hundreds of bands in "conventional" Raman spectra. It is also useful for associating normal modes with their observed frequency shifts.
• Angle-resolved Raman spectroscopy – Not only are standard Raman results recorded but also the angle with respect to the incident laser. If the orientation of the sample is known then detailed information about the phonon dispersion relation can also be gleaned from a single test.
• Optical tweezers Raman spectroscopy (OTRS) – Used to study individual particles, and even biochemical processes in single cells trapped by optical tweezers.
• Spatially offset Raman spectroscopy (SORS) – The Raman scattering beneath an obscuring surface is retrieved from a scaled subtraction of two spectra taken at two spatially offset points.
• Raman optical activity (ROA) – Measures vibrational optical activity by means of a small difference in the intensity of Raman scattering from chiral molecules in right- and left-circularly polarized incident light or, equivalently, a small circularly polarized component in the scattered light.
• Transmission Raman – Allows probing of a significant bulk of a turbid material, such as powders, capsules, living tissue, etc. It was largely ignored following investigations in the late 1960s (Schrader and Bergmann, 1967) but was rediscovered in 2006 as a means of rapid assay of pharmaceutical dosage forms. There are medical diagnostic applications particularly in the detection of cancer.
• Micro-cavity substrates – A method that improves the detection limit of conventional Raman spectra using micro-Raman in a micro-cavity coated with reflective Au or Ag. The micro-cavity has a radius of several micrometers and enhances the entire Raman signal by providing multiple excitations of the sample and couples the forward-scattered Raman photons toward the collection optics in the back-scattered Raman geometry.
• Stand-off remote Raman. – In standoff Raman, the sample is measured at a distance from the Raman spectrometer, usually by using a telescope for light collection. Remote Raman spectroscopy was proposed in the 1960s and initially developed for the measurement of atmospheric gases. The technique was extended In 1992 by Angel et al. for standoff Raman detection of hazardous inorganic and organic compounds.
• X-ray Raman scattering – Measures electronic transitions rather than vibrations.

Enhanced (or near-field) Raman spectroscopy

Enhancement of Raman scattering is achieved by local electric-field enhancement by optical near-field effects (e.g. localized surface plasmons).

• Surface-enhanced Raman spectroscopy (SERS) – Normally done in a silver or gold colloid or a substrate containing silver or gold. Surface plasmons of silver and gold are excited by the laser, resulting in an increase in the electric fields surrounding the metal. Given that Raman intensities are proportional to the electric field, there is large increase in the measured signal (by up to 10¹¹). This effect was originally observed by Martin Fleischmann but the prevailing explanation was proposed by Van Duyne in 1977. A comprehensive theory of the effect was given by Lombardi and Birke.
• Surface-enhanced resonance Raman spectroscopy (SERRS) – A combination of SERS and resonance Raman spectroscopy that uses proximity to a surface to increase Raman intensity, and excitation wavelength matched to the maximum absorbance of the molecule being analysed.
• Tip-enhanced Raman spectroscopy (TERS) – Uses a metallic (usually silver-/gold-coated AFM or STM) tip to enhance the Raman signals of molecules situated in its vicinity. The spatial resolution is approximately the size of the tip apex (20–30 nm). TERS has been shown to have sensitivity down to the single molecule level and holds some promise for bioanalysis applications and DNA sequencing. TERS was used to image the vibrational normal modes of single molecules.
• Surface plasmon polariton enhanced Raman scattering (SPPERS) – This approach exploits apertureless metallic conical tips for near field excitation of molecules. This technique differs from the TERS approach due to its inherent capability of suppressing the background field. In fact, when an appropriate laser source impinges on the base of the cone, a TM0 mode (polaritonic mode) can be locally created, namely far away from the excitation spot (apex of the tip). The mode can propagate along the tip without producing any radiation field up to the tip apex where it interacts with the molecule. In this way, the focal plane is separated from the excitation plane by a distance given by the tip length, and no background plays any role in the Raman excitation of the molecule.

Non-linear Raman spectroscopy

Raman signal enhancements are achieved through non-linear optical effects, typically realized by mixing two or more wavelengths emitted by spatially and temporally synchronized pulsed lasers.

• Hyper Raman – A non-linear effect in which the vibrational modes interact with the second harmonic of the excitation beam. This requires very high power, but allows the observation of vibrational modes that are normally "silent". It frequently relies on SERS-type enhancement to boost the sensitivity.
• Stimulated Raman spectroscopy (SRS) – A pump-probe technique, where a spatially coincident, two color pulse (with polarization either parallel or perpendicular) transfers the population from ground to a rovibrationally excited state. If the difference in energy corresponds to an allowed Raman transition, scattered light will correspond to loss or gain in the pump beam.
• Inverse Raman spectroscopy – A synonym for stimulated Raman loss spectroscopy.
• Coherent anti-Stokes Raman spectroscopy (CARS) – Two laser beams are used to generate a coherent anti-Stokes frequency beam, which can be enhanced by resonance.

Chemical imaging

From Wikipedia, the free encyclopedia

https://en.wikipedia.org/wiki/Chemical_imaging 

 

Chemical imaging (as quantitative – chemical mapping) is the analytical capability to create a visual image of components distribution from simultaneous measurement of spectra and spatial, time information. Hyperspectral imaging measures contiguous spectral bands, as opposed to multispectral imaging which measures spaced spectral bands.

The main idea - for chemical imaging, the analyst may choose to take as many data spectrum measured at a particular chemical component in spatial location at time; this is useful for chemical identification and quantification. Alternatively, selecting an image plane at a particular data spectrum (PCA - multivariable data of wavelength, spatial location at time) can map the spatial distribution of sample components, provided that their spectral signatures are different at the selected data spectrum.

Software for chemical imaging is most specific and distinguished from chemical methods such as chemometrics.

Imaging instrumentation has three components: a radiation source to illuminate the sample, a spectrally selective element, and usually a detector array (the camera) to collect the images. The data format is called a hypercube. The data set may be visualized as a data cube, a three-dimensional block of data spanning two spatial dimensions (x and y), with a series of wavelengths (lambda) making up the third (spectral) axis. The hypercube can be visually and mathematically treated as a series of spectrally resolved images (each image plane corresponding to the image at one wavelength) or a series of spatially resolved spectra.

History

Commercially available laboratory-based chemical imaging systems emerged in the early 1990s (ref. 1-5). In addition to economic factors, such as the need for sophisticated electronics and extremely high-end computers, a significant barrier to commercialization of infrared imaging was that the focal plane array (FPA) needed to read IR images were not readily available as commercial items. As high-speed electronics and sophisticated computers became more commonplace, and infrared cameras became readily commercially available, laboratory chemical imaging systems were introduced.

Initially used for novel research in specialized laboratories, chemical imaging became a more commonplace analytical technique used for general R&D, quality assurance (QA) and quality control (QC) in less than a decade. The rapid acceptance of the technology in a variety of industries (pharmaceutical, polymers, semiconductors, security, forensics and agriculture) rests in the wealth of information characterizing both chemical composition and morphology. The parallel nature of chemical imaging data makes it possible to analyze multiple samples simultaneously for applications that require high throughput analysis in addition to characterizing a single sample.

Applications

Hyperspectral imaging is most often applied to either solid or gel samples, and has applications in chemistry, biology, medicine, pharmacy (see also for example: food science, biotechnology, agriculture and industry. NIR, IR and Raman chemical imaging is also referred to as hyperspectral, spectroscopic, spectral or multispectral imaging (also see microspectroscopy). However, other ultra-sensitive and selective imaging techniques are also in use that involve either UV-visible or fluorescence microspectroscopy. Many imaging techniques can be used to analyze samples of all sizes, from the single molecule to the cellular level in biology and medicine, and to images of planetary systems in astronomy, but different instrumentation is employed for making observations on such widely different systems.

Any material that depends on chemical gradients for functionality may be amenable to study by an analytical technique that couples spatial and chemical characterization. To efficiently and effectively design and manufacture such materials, the ‘what’ and the ‘where’ must both be measured. The demand for this type of analysis is increasing as manufactured materials become more complex. Chemical imaging techniques are critical to understanding modern manufactured products and in some cases is a non-destructive technique so that samples are preserved for further testing.

Many materials, both manufactured and naturally occurring, derive their functionality from the spatial distribution of sample components. For example, extended release pharmaceutical formulations can be achieved by using a coating that acts as a barrier layer. The release of active ingredient is controlled by the presence of this barrier, and imperfections in the coating, such as discontinuities, may result in altered performance. In the semi-conductor industry, irregularities or contaminants in silicon wafers or printed micro-circuits can lead to failure of these components. The functionality of biological systems is also dependent upon chemical gradients – a single cell, tissue, and even whole organs function because of the very specific arrangement of components. It has been shown that even small changes in chemical composition and distribution may be an early indicator of disease.

Principles

Chemical imaging shares the fundamentals of vibrational spectroscopic techniques, but provides additional information by way of the simultaneous acquisition of spatially resolved spectra. It combines the advantages of digital imaging with the attributes of spectroscopic measurements. Briefly, vibrational spectroscopy measures the interaction of light with matter. Photons that interact with a sample are either absorbed or scattered; photons of specific energy are absorbed, and the pattern of absorption provides information, or a fingerprint, on the molecules that are present in the sample.

On the other hand, in terms of the observation setup, chemical imaging can be carried out in one of the following modes: (optical) absorption, emission (fluorescence), (optical) transmission or scattering (Raman). A consensus currently exists that the fluorescence (emission) and Raman scattering modes are the most sensitive and powerful, but also the most expensive.

In a transmission measurement, the radiation goes through a sample and is measured by a detector placed on the far side of the sample. The energy transferred from the incoming radiation to the molecule(s) can be calculated as the difference between the quantity of photons that were emitted by the source and the quantity that is measured by the detector. In a diffuse reflectance measurement, the same energy difference measurement is made, but the source and detector are located on the same side of the sample, and the photons that are measured have re-emerged from the illuminated side of the sample rather than passed through it. The energy may be measured at one or multiple wavelengths; when a series of measurements are made, the response curve is called a spectrum.

A key element in acquiring spectra is that the radiation must somehow be energy selected – either before or after interacting with the sample. Wavelength selection can be accomplished with a fixed filter, tunable filter, spectrograph, an interferometer, or other devices. For a fixed filter approach, it is not efficient to collect a significant number of wavelengths, and multispectral data are usually collected. Interferometer-based chemical imaging requires that entire spectral ranges be collected, and therefore results in hyperspectral data. Tunable filters have the flexibility to provide either multi- or hyperspectral data, depending on analytical requirements.

Spectra are typically measured with an imaging spectrometer, based on a Focal Plane Array.

Terminology

Some words common in spectroscopy, optical microscopy and photography have been adapted or their scope modified for their use in chemical imaging. They include: resolution, field of view and magnification. There are two types of resolution in chemical imaging. The spectral resolution refers to the ability to resolve small energy differences; it applies to the spectral axis. The spatial resolution is the minimum distance between two objects that is required for them to be detected as distinct objects. The spatial resolution is influenced by the field of view, a physical measure of the size of the area probed by the analysis. In imaging, the field of view is a product of the magnification and the number of pixels in the detector array. The magnification is a ratio of the physical area of the detector array divided by the area of the sample field of view. Higher magnifications for the same detector image a smaller area of the sample.

Types of vibrational chemical imaging instruments

Chemical imaging has been implemented for mid-infrared, near-infrared spectroscopy and Raman spectroscopy. As with their bulk spectroscopy counterparts, each imaging technique has particular strengths and weaknesses, and are best suited to fulfill different needs.

Mid-infrared chemical imaging

A set of stones scanned with a Specim LWIR-C hyperspectral imager in the thermal infrared range from 7.7 μm to 12.4 μm. Minerals such as quartz and feldspar spectra are clearly recognizable.

Mid-infrared (MIR) spectroscopy probes fundamental molecular vibrations, which arise in the spectral range 2,500-25,000 nm. Commercial imaging implementations in the MIR region employ hyperspectral imagers or Fourier Transform Infrared (FT-IR) interferometers, depending on the application. The MIR absorption bands tend to be relatively narrow and well-resolved; direct spectral interpretation is often possible by an experienced spectroscopist. MIR spectroscopy can distinguish subtle changes in chemistry and structure, and is often used for the identification of unknown materials. The absorptions in this spectral range are relatively strong; for this reason, sample presentation is important to limit the amount of material interacting with the incoming radiation in the MIR region. Data can be collected in reflectance, transmission, or emission mode. Water is a very strong absorber of MIR radiation and wet samples often require advanced sampling procedures (such as attenuated total reflectance). Commercial instruments include point and line mapping, and imaging. Mid-infrared chemical imaging can also be performed with nanometer level spatial resolution using atomic force microscope based infrared spectroscopy (AFM-IR).

Remote chemical imaging of a simultaneous release of SF₆ and NH₃ at 1.5km using the Telops Hyper-Cam imaging spectrometer

Atmospheric windows in the infrared spectrum are also employed to perform chemical imaging remotely. In these spectral regions the atmospheric gases (mainly water and CO₂) present low absorption and allow infrared viewing over kilometer distances. Target molecules can then be viewed using the selective absorption/emission processes described above. An example of the chemical imaging of a simultaneous release of SF₆ and NH₃ is shown in the image.

Near-infrared chemical imaging

The analytical near infrared (NIR) region spans the range from 780 nm to 2,500 nm. The absorption bands seen in this spectral range arise from overtones and combination bands of O-H, N-H, C-H and S-H stretching and bending vibrations. Absorption is one to two orders of magnitude smaller in the NIR compared to the MIR; this phenomenon eliminates the need for extensive sample preparation. Thick and thin samples can be analyzed without any sample preparation, it is possible to acquire NIR chemical images through some packaging materials, and the technique can be used to examine hydrated samples, within limits. Intact samples can be imaged in transmittance or diffuse reflectance.

The lineshapes for overtone and combination bands tend to be much broader and more overlapped than for the fundamental bands seen in the MIR. Often, multivariate methods are used to separate spectral signatures of sample components. NIR chemical imaging is particularly useful for performing rapid, reproducible and non-destructive analyses of known materials. NIR imaging instruments are typically based on a hyperspectral camera, a tunable filter or an FT-IR interferometer. External light source is always needed, such as sun (outdoor scans, remote sensing) or a halogen lamp (laboratory, industrial measurements).

Raman chemical imaging

The Raman shift chemical imaging spectral range spans from approximately 50 to 4,000 cm⁻¹; the actual spectral range over which a particular Raman measurement is made is a function of the laser excitation frequency. The basic principle behind Raman spectroscopy differs from the MIR and NIR in that the x-axis of the Raman spectrum is measured as a function of energy shift (in cm⁻¹) relative to the frequency of the laser used as the source of radiation. Briefly, the Raman spectrum arises from inelastic scattering of incident photons, which requires a change in polarizability with vibration, as opposed to infrared absorption, which requires a change in dipole moment with vibration. The end result is spectral information that is similar and in many cases complementary to the MIR. The Raman effect is weak - only about one in 10⁷ photons incident to the sample undergoes Raman scattering. Both organic and inorganic materials possess a Raman spectrum; they generally produce sharp bands that are chemically specific. Fluorescence is a competing phenomenon and, depending on the sample, can overwhelm the Raman signal, for both bulk spectroscopy and imaging implementations.

Raman chemical imaging requires little or no sample preparation. However, physical sample sectioning may be used to expose the surface of interest, with care taken to obtain a surface that is as flat as possible. The conditions required for a particular measurement dictate the level of invasiveness of the technique, and samples that are sensitive to high power laser radiation may be damaged during analysis. It is relatively insensitive to the presence of water in the sample and is therefore useful for imaging samples that contain water such as biological material.

Fluorescence Imaging (Ultraviolet, visible and near infrared regions)

Emission microspectroscopy is a sensitive technique with excitation and emission ranging from the ultraviolet, visible and NIR regions. As such, it has numerous biomedical, biotechnological and agricultural applications. There are several powerful, highly specific and sensitive fluorescence techniques that are currently in use, or still being developed; among the former are FLIM, FRAP, FRET and FLIM-FRET; among the latter are NIR fluorescence and probe-sensitivity enhanced NIR fluorescence microspectroscopy and nanospectroscopy techniques (see "Further reading" section). Fluorescence emission microspectroscopy and imaging are also commonly used to locate protein crystals in solution, for the characterization of metamaterials and biotechnology devices.

Sampling and samples

The value of imaging lies in the ability to resolve spatial heterogeneities in solid-state or gel/gel-like samples. Imaging a liquid or even a suspension has limited use as constant sample motion serves to average spatial information, unless ultra-fast recording techniques are employed as in fluorescence correlation microspectroscopy or FLIM observations where a single molecule may be monitored at extremely high (photon) detection speed. High-throughput experiments (such as imaging multi-well plates) of liquid samples can however provide valuable information. In this case, the parallel acquisition of thousands of spectra can be used to compare differences between samples, rather than the more common implementation of exploring spatial heterogeneity within a single sample.

Similarly, there is no benefit in imaging a truly homogeneous sample, as a single point spectrometer will generate the same spectral information. Of course the definition of homogeneity is dependent on the spatial resolution of the imaging system employed. For MIR imaging, where wavelengths span from 3-10 micrometres, objects on the order of 5 micrometres may theoretically be resolved. The sampled areas are limited by current experimental implementations because illumination is provided by the interferometer. Raman imaging may be able to resolve particles less than 1 micrometre in size, but the sample area that can be illuminated is severely limited. With Raman imaging, it is considered impractical to image large areas and, consequently, large samples. FT-NIR chemical/hyperspectral imaging usually resolves only larger objects (>10 micrometres), and is better suited for large samples because illumination sources are readily available. However, FT-NIR microspectroscopy was recently reported to be capable of about 1.2 micron (micrometer) resolution in biological samples Furthermore, two-photon excitation FCS experiments were reported to have attained 15 nanometer resolution on biomembrane thin films with a special coincidence photon-counting setup.

Detection limit

The concept of the detection limit for chemical imaging is quite different from for bulk spectroscopy, as it depends on the sample itself. Because a bulk spectrum represents an average of the materials present, the spectral signatures of trace components are simply overwhelmed by dilution. In imaging however, each pixel has a corresponding spectrum. If the physical size of the trace contaminant is on the order of the pixel size imaged on the sample, its spectral signature will likely be detectable. If however, the trace component is dispersed homogeneously (relative to pixel image size) throughout a sample, it will not be detectable. Therefore, detection limits of chemical imaging techniques are strongly influenced by particle size, the chemical and spatial heterogeneity of the sample, and the spatial resolution of the image.

Data analysis

Data analysis methods for chemical imaging data sets typically employ mathematical algorithms common to single point spectroscopy or to image analysis. The reasoning is that the spectrum acquired by each detector is equivalent to a single point spectrum; therefore pre-processing, chemometrics and pattern recognition techniques are utilized with the similar goal to separate chemical and physical effects and perform a qualitative or quantitative characterization of individual sample components. In the spatial dimension, each chemical image is equivalent to a digital image and standard image analysis and robust statistical analysis can be used for feature extraction.

Software

• FECOM Object Learning Software (OLS), industrial in-line hyperspectral feature processing 
• UmBio Evince Image, multivariate hyperspectral image analysis 
• Perception System; in-line hyperspectral imaging for industry 

Multispectral image

From Wikipedia, the free encyclopedia

 

 

Play media

Video by SDO simultaneously showing sections of the Sun at various wavelengths

A multispectral image is one that captures image data within specific wavelength ranges across the electromagnetic spectrum. The wavelengths may be separated by filters or detected via the use of instruments that are sensitive to particular wavelengths, including light from frequencies beyond the visible light range, i.e. infrared and ultra-violet. Spectral imaging can allow extraction of additional information the human eye fails to capture with its visible receptors for red, green and blue. It was originally developed for military target identification and reconnaissance. Early space-based imaging platforms incorporated multispectral imaging technology to map details of the Earth related to coastal boundaries, vegetation, and landforms. Multispectral imaging has also found use in document and painting analysis.

Multispectral imaging measures light in a small number (typically 3 to 15) of spectral bands.

 Hyperspectral imaging is a special case of spectral imaging where often hundreds of contiguous spectral bands are available.

Applications

Military target tracking

Multispectral imaging measures light emission and is often used in detecting or tracking military targets. In 2003, researchers at the United States Army Research Laboratory and the Federal Laboratory Collaborative Technology Alliance reported a dual band multispectral imaging focal plane array (FPA). This FPA allowed researchers to look at two infrared (IR) planes at the same time. Because mid-wave infrared (MWIR) and long wave infrared (LWIR) technologies measure radiation inherent to the object and require no external light source, they also are referred to as thermal imaging methods.

The brightness of the image produced by a thermal imager depends on the objects emissivity and temperature.  Every material has an infrared signature that aids in the identification of the object. These signatures are less pronounced in hyperspectral systems (which image in many more bands than multispectral systems) and when exposed to wind and, more dramatically, to rain. Sometimes the surface of the target may reflect infrared energy. This reflection may misconstrue the true reading of the objects’ inherent radiation. Imaging systems that use MWIR technology function better with solar reflections on the target's surface and produce more definitive images of hot objects, such as engines, compared to LWIR technology. However, LWIR operates better in hazy environments like smoke or fog because less scattering occurs in the longer wavelengths. Researchers claim that dual-band technologies combine these advantages to provide more information from an image, particularly in the realm of target tracking.

For nighttime target detection, thermal imaging outperformed single-band multispectral imaging. Citation. Dual band MWIR and LWIR technology resulted in better visualization during the nighttime than MWIR alone. Citation Citation. The US Army reports that its dual band LWIR/MWIR FPA demonstrated better visualizing of tactical vehicles than MWIR alone after tracking them through both day and night.  

Land mine detection

By analyzing the emissivity of ground surfaces, multispectral imaging can detect the presence of underground missiles. Surface and sub-surface soil possess different physical and chemical properties that appear in spectral analysis. Disturbed soil has increased emissivity in the wavelength range of 8.5 to 9.5 micrometers while demonstrating no change in wavelengths greater than 10 micrometers. The US Army Research Laboratory's dual MWIR/LWIR FPA used "red" and "blue" detectors to search for areas with enhanced emissivity. The red detector acts as a backdrop, verifying realms of undisturbed soil areas, as it is sensitive to the 10.4 micrometer wavelength. The blue detector is sensitive to wavelengths of 9.3 micrometers. If the intensity of the blue image changes when scanning, that region is likely disturbed. The scientists reported that fusing these two images increased detection capabilities.

Ballistic missile detection

Intercepting an intercontinental ballistic missile (ICBM) in its boost phase requires imaging of the hard body as well as the rocket plumes. MWIR presents a strong signal from highly heated objects including rocket plumes, while LWIR produces emissions from the missile's body material. The US Army Research Laboratory reported that with their dual-band MWIR/LWIR technology, tracking of the Atlas 5 Evolved Expendable Launch Vehicles, similar in design to ICBMs, picked up both the missile body and plumage.

Space-based imaging

Most radiometers for remote sensing (RS) acquire multispectral images. Dividing the spectrum into many bands, multispectral is the opposite of panchromatic, which records only the total intensity of radiation falling on each pixel. Usually, Earth observation satellites have three or more radiometers. Each acquires one digital image (in remote sensing, called a 'scene') in a small spectral band. The bands are grouped into wavelength regions based on the origin of the light and the interests of the researchers.

Weather Forecasting

Modern weather satellites produce imagery in a variety of spectra. 

Multispectral imaging combines two to five spectral imaging bands of relatively large bandwidth into a single optical system. A multispectral system usually provides a combination of visible (0.4 to 0.7 µm), near infrared (NIR; 0.7 to 1 µm), short-wave infrared (SWIR; 1 to 1.7 µm), mid-wave infrared (MWIR; 3.5 to 5 µm) or long-wave infrared (LWIR; 8 to 12 µm) bands into a single system. — Valerie C. Coffey

In the case of Landsat satellites, several different band designations have been used, with as many as 11 bands (Landsat 8) comprising a multispectral image. Spectral imaging with a higher radiometric resolution (involving hundreds or thousands of bands), finer spectral resolution (involving smaller bands), or wider spectral coverage may be called hyperspectral or ultraspectral.

Documents and artworks

The technology has also assisted in the interpretation of ancient papyri, such as those found at Herculaneum, by imaging the fragments in the infrared range (1000 nm). Often, the text on the documents appears to the naked eye as black ink on black paper. At 1000 nm, the difference in how paper and ink reflect infrared light makes the text clearly readable. It has also been used to image the Archimedes palimpsest by imaging the parchment leaves in bandwidths from 365–870 nm, and then using advanced digital image processing techniques to reveal the undertext with Archimedes' work. Multispectral imaging has been used in a Mellon Foundation project at Yale University to compare inks in medieval English manuscripts.

Multispectral imaging can be employed for investigation of paintings and other works of art. The painting is irradiated by ultraviolet, visible and infrared rays and the reflected radiation is recorded in a camera sensitive in this regions of the spectrum. The image can also be registered using the transmitted instead of reflected radiation. In special cases the painting can be irradiated by UV, VIS or IR rays and the fluorescence of pigments or varnishes can be registered.

Multispectral imaging has also been used to examine discolorations and stains on old books and manuscripts. Comparing the "spectral fingerprint" of a stain to the characteristics of known chemical substances can make it possible to identify the stain. This technique has been used to examine medical and alchemical texts, seeking hints about the activities of early chemists and the possible chemical substances they may have used in their experiments. Like a cook spilling flour or vinegar on a cookbook, an early chemist might have left tangible evidence on the pages of the ingredients used to make medicines.

Spectral bands

The wavelengths are approximate; exact values depend on the particular satellite's instruments:

• Blue, 450–515..520 nm, is used for atmosphere and deep water imaging, and can reach depths up to 150 feet (50 m) in clear water.
• Green, 515..520–590..600 nm, is used for imaging vegetation and deep water structures, up to 90 feet (30 m) in clear water.
• Red, 600..630–680..690 nm, is used for imaging man-made objects, in water up to 30 feet (9 m) deep, soil, and vegetation.
• Near infrared (NIR), 750–900 nm, is used primarily for imaging vegetation.
• Mid-infrared (MIR), 1550–1750 nm, is used for imaging vegetation, soil moisture content, and some forest fires.
• Far-infrared (FIR), 2080–2350 nm, is used for imaging soil, moisture, geological features, silicates, clays, and fires.
• Thermal infrared, 10400-12500 nm, uses emitted instead of reflected radiation to image geological structures, thermal differences in water currents, fires, and for night studies.
• Radar and related technologies are useful for mapping terrain and for detecting various objects.

Spectral band usage

For different purposes, different combinations of spectral bands can be used. They are usually represented with red, green, and blue channels. Mapping of bands to colors depends on the purpose of the image and the personal preferences of the analysts. Thermal infrared is often omitted from consideration due to poor spatial resolution, except for special purposes.

• True-color uses only red, green, and blue channels, mapped to their respective colors. As a plain color photograph, it is good for analyzing man-made objects, and is easy to understand for beginner analysts.
• Green-red-infrared, where the blue channel is replaced with near infrared, is used for vegetation, which is highly reflective in near IR; it then shows as blue. This combination is often used to detect vegetation and camouflage.
• Blue-NIR-MIR, where the blue channel uses visible blue, green uses NIR (so vegetation stays green), and MIR is shown as red. Such images allow the water depth, vegetation coverage, soil moisture content, and the presence of fires to be seen, all in a single image.

Many other combinations are in use. NIR is often shown as red, causing vegetation-covered areas to appear red.

Classification

Unlike other Aerial photographic and satellite image interpretation work, these multispectral images do not make it easy to identify directly the feature type by visual inspection. Hence the remote sensing data has to be classified first, followed by processing by various data enhancement techniques so as to help the user to understand the features that are present in the image.

Such classification is a complex task which involves rigorous validation of the training samples depending on the classification algorithm used. The techniques can be grouped mainly into two types.

• Supervised classification techniques
• Unsupervised classification techniques

Supervised classification makes use of training samples. Training samples are areas on the ground for which there is Ground truth, that is, what is there is known. The spectral signatures of the training areas are used to search for similar signatures in the remaining pixels of the image, and we will classify accordingly. This use of training samples for classification is called supervised classification. Expert knowledge is very important in this method since the selection of the training samples and a biased selection can badly affect the accuracy of classification. Popular techniques include the Maximum likelihood principle and Convolutional neural network. The Maximum likelihood principle calculates the probability of a pixel belonging to a class (i.e. feature) and allots the pixel to its most probable class. Newer Convolutional neural network based methods account for both spatial proximity and entire spectra to determine the most likely class.

In case of unsupervised classification no prior knowledge is required for classifying the features of the image. The natural clustering or grouping of the pixel values, i.e. the gray levels of the pixels, are observed. Then a threshold is defined for adopting the number of classes in the image. The finer the threshold value, the more classes there will be. However, beyond a certain limit the same class will be represented in different classes in the sense that variation in the class is represented. After forming the clusters, ground truth validation is done to identify the class the image pixel belongs to. Thus in this unsupervised classification apriori information about the classes is not required. One of the popular methods in unsupervised classification is k-means clustering.

Multispectral data analysis software

• MicroMSI is endorsed by the NGA.
• Opticks is an open-source remote sensing application.
• Multispec is freeware multispectral analysis software.
• Gerbil is open source multispectral visualization and analysis software.