Raman spectroscopic probes withstand hostile environments
When they are used to make Raman-spectroscopic measurements of chemical processes, fiberoptic probes must be carefully designed to withstand reactive chemicals and high temperatures.
A variety of spectroscopic methods can be used to monitor chemical reactions and processes. Applications include the manufacturing of polymers, the refining of fossil fuels, and numerous other reactions of both organic and inorganic species. More recently, on-line spectroscopy has become focused on the pharmaceutical industry. The Food and Drug Administration’s Process Analytical Technology initiative obliges drug manufacturers to monitor processes in real time rather than relying solely on tests of the final product. The processes to be monitored include synthetic reactions, blending processes, and the distribution of active ingredients in finished tablets. A Raman spectroscopic technique for these purposes involves measurements taken via a fiberoptic probe; when used in harsh environments, such a probe requires careful design and choice of optical and optomechanical materials.
Raman-probe optics
Raman spectroscopy has advantages over other techniques for monitoring chemical changes. Comparisons are usually made with mid-IR (usually Fourier-transform-IR, or FT-IR) and near-IR, which are both absorption methods. As a light-scattering technique, Raman does not have the sample-thickness problems of FT-IR (see Laser Focus World, September 2003, p. 89). Raman measurements are usually made in the visible or near-IR regions (450 to 1100 nm), allowing the use of silica optical fibers. From a chemist’s point of view, Raman spectra contain fundamental vibrational modes rather than overtones and combinations observed in near-IR spectroscopy, making the spectral data easier to correlate to molecular structure.
The primary disadvantage of Raman spectroscopy is lack of sensitivity; Raman is typically reserved for applications where analytes are at least 1% to 10% in concentration. Fiberoptic probes enable the measurement of Raman spectra over long lengths. In industrial plants, the distance between a measurement point and control room (which usually houses the spectrometer) is generally less than 150 m. Silica optical fibers are relatively inexpensive, robust, and readily available to bridge the gap between sample and spectrometer. At the distal end of the fibers, filtering and focusing optics are required to effectively excite and collect Raman scattering while blocking undesirable photons.1
The earliest Raman fiberoptic probes consisted of fiber bundles, usually with a central fiber for sample excitation, and one or more fibers to collect the backscattered light. These bundle probes were compact and inexpensive; however, they often suffered from poor collection efficiency and inadequate optical filtering. Today, the vast majority of commercial Raman fiberoptic probes are coaxial designs in which optical components are positioned in the optical path between the excitation and collection fibers to filter the strong Rayleigh line (elastic scattering) and to reduce silica background. These probes use a single excitation fiber that attaches to the laser and a single collection fiber to couple the Raman scattering to a spectrograph. One lens serves as both the focusing optic for the excitation beam as well as the collection optic for the backscattered light.
FIGURE 1. A coaxial fiberoptic Raman probe has integrated filtering of extraneous radiation (left). The probe is equipped with standard ANSI flanges or compression fittings and must be designed to withstand chemical baths and high temperatures and pressures (right).
The two-fiber coaxial probe was originally designed as a multipurpose sampling tool for a variety of environmental and military applications (see Fig. 1).2 The key optical advantages of the probe are high attenuation of the Rayleigh line (optical filtering to optical densities of greater than 8), efficient collection of Raman scatter, and small physical size (12.5-mm diameter). Each probe is designed for a single excitation wavelength and incorporates three types of optical filters (bandpass, dichroic, and longpass). Together, the filters serve to eliminate silica background and spurious radiation from the excitation beam prior to sample excitation, and filter the Rayleigh line from the collected scattered light. Hard-coated metal-oxide filters are used throughout the probe for resistance to temperature and humidity.
A variety of fiber sizes can be used with the Raman probe to optically match the laser output and spectrograph input. For highest Raman signal, the collection fiber is usually at least double the core size of the excitation fiber. As a result, the measurement volume of the probe is elongated along the optical axis.
Withstanding high temperatures and pressures
In recent years, the most important development in Raman fiberoptic probes has been optomechanical engineering to withstand immersion in chemical media at elevated temperatures and pressures. Two different approaches have been pursued to address these criteria. The simplest and most optically efficient method is to place a protective sleeve over the entire probe. Sleeves can be made of stainless steel, Hastelloy, or even ceramics and polymers that are resistant to chemical attack. An optical window is installed at the tip to transmit the excitation and collection beams. Probes usually use elastomeric O-rings, brazes, or gold gaskets to form the window seal; compatibility of the seal material with the physical and chemical environment is critical.
Synthetic fused silica (SFS) is the best optical window for these Raman measurements. But SFS is not particularly strong, necessitating very thick windows for high-pressure applications. Strong acids and other aggressive chemicals can also attack SFS. Sapphire is more commonly used for in-situ applications. Sapphire is extremely robust, making it ideal for high-pressure conditions. It is generally inert to chemical attack, including resistance to very acidic environments. The combination of Hastelloy C and sapphire is typically used in corrosive media. Optically, sapphire is less ideal as it has a higher refractive index, resulting in lower transmission; the inside window face can be antireflection-coated to reduce reflection losses. The sapphire grade must also be chosen carefully to avoid strong emission bands that can obscure the Raman spectrum. In the most aggressive chemical environments or extreme pressure conditions, diamond windows have been used. Type IIA diamonds are usually selected to avoid fluorescence caused by the laser excitation. Because diamond itself has a very strong Raman band, and the material is relatively expensive, it is used only in the limited number of applications in which sapphire would be inappropriate.
These industrial Raman probes can withstand environmental extremes of greater than 200°C and 1500 psi. Because the optics are placed to the tip of the immersion sleeve, minimal optical throughput is lost, and the probe can be physically made to any length-a probe greater than 3.5 m long was recently designed to monitor a combustion process. The separate probe/sleeve construction enables the user to set the working distance by adjusting the distance of the focal point with respect to the window (see Fig. 2).
FIGURE 2. Variants of immersion optics for the measurement of transparent and opaque samples include a long-working-distance version (left), a short-working-distance version (center), and a biconvex spherical lens developed at InPhotonics that is loaded into a press-fit seal (right). The optical properties of the lens are not degraded by the press-fit mounting.
For transparent solutions, the working distance is not critical. For opaque samples (either scattering or absorbing), a short working distance is required for optimum Raman signal.
Collimation extends capabilities
Another approach to making immersible Raman fiberoptic probes is to keep the probe head outside of the chemical process by collimating the optical beam through a tube where the final focusing optic is installed. This design ensures that the optical filters are maintained at a safe distance from a heat source, enabling measurements of processes at much higher temperatures. It also allows the use of narrower immersion tubes with smaller windows that withstand even higher pressures (up to 5000 psi).
On this type of probe, various focusing optics can be used at the probe tip. With a flat window forming the seal to the metal tube, a lens can be positioned to achieve an optimum working distance for a particular solution. Another option is to use a single optic as both the window and lens; a spherical (ball) lens has been shown to provide satisfactory focusing on very opaque samples.3, 4 Advantages of the ball lens include reproducible measurement volume, a focal point close to the window face, and less stringent optical alignment. A 6.35-mm-diameter sapphire ball lens has a back focal length of 0.475 mm, which is ideal for opaque solutions. With a 4.5-mm collimated beam size, the numerical aperture (NA) of the lens is 0.62. This is more optically efficient than mounting an f/1 lens inside a flat window, where the NA is 0.22 (the same as the fiber NA).
Loaded press-fit seal
With the ball lens, displacement of the collimated beam away from the optical axis moves the focal point on the opposite side of the sphere, but it does not affect the focus volume. Thus, the spherical lens is an excellent optic for in-situ Raman spectroscopy. The mounting of a complete spherical optic is not as straightforward as mounting a flat optic, especially when elastomeric O-rings cannot be used for reasons of chemical compatibility. As a solution, a biconvex lens developed at InPhotonics has been designed to fit into a loaded, press-fit seal (patent pending) while maintaining the optical properties of the spherical lens.❑
REFERENCES
1. I. R. Lewis and P. R. Griffiths, Appl. Spectrosc. 50(10), 12A (1996).
2. M. M. Carrabba and R. D. Rauh, “Apparatus for Measuring Raman Spectra over Optical Fibers,” U.S. Patent 5,112,127 (1992).
3. B. J. Marquardt et al., Proc. SPIE 4469, 62 (2001).
4. B. J. Marquardt and L. W. Burgess, “Optical Immersion Probe Incorporating a Spherical Lens,” U.S. Patent 6,831,745 (2004).
NANCY KAWAI is marketing manager and ROBERT FORNEY is a design engineer at InPhotonics, 111 Downey St., Norwood, MA 02062; e-mail: [email protected].