A long path is needed for gas measurement

Long path cells are necessary for gas measurements mainly because of the low density of the gaseous samples. A gas consists of fast-moving molecules that occupy a volume about one thousand times greater than the volume of a comparable mass of condensed phase material. Furthermore, a gas that is to be measured is frequently only a minor component in air, nitrogen, or other gaseous matrix material. The matrix gas may be totally transparent to infrared radiation, as in the case of dry air, or at least it is likely to be partially transparent. The minor components are thus measurable by infrared absorption, if the optical path is long enough.

One may consider the contrast between measuring pollutants in water and measuring pollutants in air. A millimeter thickness of water transmits almost no infrared radiation, while a kilometer thickness of air is highly transmissive in the infrared. Water pollutants are thus not detectable by infrared radiation until they are separated from the water, while pollutants in air are easily detected without separation.

gascel~1.gifFigure 1. Three ways to record the spectrum of the atmosphere.

The study of the atmosphere

The study of the atmosphere is the principal example of the efficacy of the long path infrared method of trace gas measurement. Looking at the sun with infrared instrumentation has been the main way that people have come to understand what molecules are in the atmosphere, how they are distributed with altitude and what atmospheric changes have been taking place due to human activities. Observations from balloons and satellites have been especially productive.

There have been three main methods of measuring the absorption of infrared radiation by the atmospheric gases, as illustrated in Figure 1. Figure 1a shows the case or looking at the sun from an observing point on the earth’s surface. For measurements not concerned with the lower troposphere, an observation point at high altitude is chosen.(Migeotte, et al. 1956) The higher the observatory, the less is the interference from water vapor. To supplement the mountaintop observatories, high flying aircraft have been used as the observing platform. (Mankin, et al., 1979)

The second method of recording atmospheric spectra, shown in Figure 1b, is to observe the setting or rising sun from a balloon or satellite that is above most of the atmosphere. (Farmer, et al., 1980), (Murcray, et al., 1969), (Williams, et al., 1976). This is called atmospheric limb spectroscopy. In these cases, the optical path- concentration product can be hundreds of kilometer- atmospheres. (1 km.-atm. optical path is equivalent to a one kilometer path at sea level.) The lowest point in the optical path, called the tangent altitude, depends on how far down the optical system is pointed. For a viewing direction 94 degrees off the vertical, for example, from an instrument at 30 km. altitude, the tangent altitude is about 15 km., and the equivalent optical path is about 100 km.-atm. A major advantage of sighting along a high altitude path to the sun is that very little water vapor exists in the stratosphere, and the water lines do not dominate the spectrum as at lower altitudes. A disadvantage of this method is that observations can only be made at sunrise and sunset.

The observation method depicted in Figure 1c is that which has been used in the urban smog. (Scott, et al., 1957), (Hanst, et al., 1982), (Herget, 1979). In this case the path is horizontal, between an infrared radiation source and the detection system. The source may be at one end of the optical path and the detector at the other end, as shown in Figure 1c, or the source and detector may be at one end, with a reflector at the other end. A single reflection or many reflections may be used. In recent years, the method of figure 1c has also been called Optical Remote Sensing (a misnomer because it is in fact in-situ sensing). See the review articles by Grant, et al., (1992) and Simonds, et al., (1994).

Long path cells for laboratory measurement of gases (White’s cell)

pfundtcell.gifFigure 2 Pfundt Cell
whitec~1.gifFigure 3 White Cell - Basic set of 4 passes.

Prior to the second world war, long pipes, with or without a reflective end-piece were used in gas studies. In addition, a triple-pass cell with large optical aperture was used by Pfund for the study of the atmosphere.(Pfundt, 1939) This type of cell had two spherical mirrors facing each other at a separation equal to the mirror focal length. Each mirror had a central hole. Light from the source was passed through a focus at one of the holes and then to the facing mirror. This sent the beam in a state of collimation to the opposite mirror. From there the radiation was focused out through the second hole. This is drawn in figure 2. This three-pass configuration lengthened the optical path somewhat, while conserving energy, but it soon was eclipsed by the optical design described by J. U. White, (1942), known ever since as the “White” cell. White’s cell can be thought of as a descendant of the Pfund cell in that it uses a spherical mirror to collect a diverging beam of radiation, but that is about where the similarity ends. White’s design was a tremendous stride forward in that it permitted many reflections in a compact vo lume while conserving most of the energy that was collected at the first reflection. The White cell has not really been improved upon in the sixty years since White’s original paper was published. This cell uses three spherical mirrors and operates with the input and output both at the same end, as shown in figure 3. In this case, the mirror separations are at the radius of curvature, which is twice the focal length. The figure shows the basic set of four passes, but there can be multiple-passing in increments of four passes. After the initial collection of light, the only energy lost in the cell is that which is absorbed by the mirror surfaces. The large aperture is maintained by the continual re-focusing.

2rowcell.jpgFigure 4 Location of images on the field mirror.

As shown in figure 3, the light from the source is focused initially into a real image at the entrance aperture of the cell. In figure 4, that is designated as the zeroth image. After that, the beam diverges and is collected by the first of the mirrors at the opposite end, called the “objective” mirrors. From there the energy is refocused onto the lower part of the opposite mirror, called the “field” mirror. This first image is marked 2--for two passes. The field mirror is aimed so that the reflected diverging beam falls entirely on the second objective mirror. This is then aimed to form another image (marked 4) above the center line of the field mirror alongside the zeroth image. If this image falls symmetrically opposite the first image (numbered 2), the beam will be returned to the first objective so that all the energy is again collected and returned. Then there will be at least four more passes.

The number of images allowed in the row depends on the placement of the first image in the lower part of the mirror. If it falls exactly on the vertical center line, no more than four passes are possible. The farther to the right it falls, the greater the number of passes allowed. In practice, the number of passes is four times the number of images in the bottom row.

In order to have as much energy as possible going through the system, an enlarged image of the source is usually projected into the entrance aperture, and the exit image is then demagnified when the transmitted radiation is fed to the detection system.


Frequently, the only major source of energy loss in a cell is the absorption at the mirror surfaces. If R is the reflection coefficient of the mirror surface, R to the nth power is the fractional amount of energy transmitted by the cell after n reflections. Some sample calculations show the amounts transmitted. A traditional choice for mirror coating has been gold, for which the value of R may be 0.98. Then in a 52-pass cell with gold-coated mirrors, the fraction of energy transmitted will be 0.98 to the 52nd power, or 35%. With 104 passes, the fraction of energy transmitted will be 0.98 to the 104th power, or 12%. If the mirrors can have R equal to 0.995, which is possible with protected silver, then the 52-pass cell will transmit 77% and the 104-pass cell will transmit 59%.

Some Applications

smog.gifFigure 5 SMOG

During the past 50 years, White cells have been used in gas studies in all parts of the world. Notable examples are the work done at the Ohio State University,(Howard and Chapman, 1952) at the National Research Council of Canada, (Bernstein and Herzberg (1948) and at the Pennsylvania State University.(Rank, et al., 1962). In the middle 1950's, a White cell was first applied in the laboratory study of the chemical reactions of the pollutants found in the urban air, leading to a number of significant findings on the creation of urban ozone, ozonides, peroxy nitrates and other toxic pollutants. (Stephens, et al., 1956). In the 1970's, White cells were used in many laboratory studies of the photochemistry of air contaminants, and in the direct study of the atmosphere. (Hanst, et al., 1975), (Hanst, et al., 1977), (Pitts, et al., 1977). Present-day applications of the White cell include the following:

l5.gifFigure 6 100 meter cell in sample compartment

Ambient air pollution measurements

Stack gas analysis

Engine exhaust analysis

Release of pollutants in painting.

Studies in photochemistry and kinetics

Analyzing gases for semi-conductor industry

Monitoring industrial chemical processes

Measuring air quality in the work place

Analyzing effluents in chromatography

Analyzing gases from materials pyrolysis

Measuring trace impurities in reagent gas

Measuring volatile pollutants in soil and water

Studies on greenhouse and ozone effects



For most of these present-day applications, the long path cell is mounted in the sample compartment of a Fourier transform spectrometer, as shown by the 100-meter cell in Figure 6. At the other end of the scale of cell size are small cells with 3.2 meters of optical path folded in a volume of 100 cc.

Variations on White’s cell

Two Rows of Images White’s original paper showed all the images on the field mirror in a single row. However, in adjusting the mirrors, one soon finds that it is beneficial to split the image array into two rows. This then requires the field mirror to be cut in a “mushroom” or “T” shape, as shown in figure 4.

Four Rows of Images A modification of the White system that increases its energy throughput was described by Horn and Pimentel.(1971). They created two additional rows of images and doubled the number of passes by means of a retro-reflecting pair of mirrors at the normal exit port. This increased the amount of available mirror surface at the in-focus end of the cell. That allowed the use of a more collimated beam, thus increasing the energy throughput without enlarging the main mirrors. Using this system with special high-quality mirror coatings, Horn and Pimentel worked at 254 reflections along a 10 meter base path, for a total path of 2.54 Km.

Six Rows of Images Infrared Analysis, Inc. has carried the Horn and Pimentel concept one step further by putting a second retro-reflecting pair of mirrors on the input side of the field mirror. This creates six rows of images. In this case the field mirror shape can be the “inverted mushroom”, as is used in the simple 2-row White cell. The six-row cell may be called a “tripled-up” White cell. Additional improvements in the design of large cells for outdoor use have been described by Ritz, et al., (1992).

Cell Construction Techniques and Materials.


White cell construction can involve exquisite mechanisms, such as differential screws, cam-driven mirror adjusters, and mechanisms to move both objective mirrors simultaneously; but in practice it is found that such mechanisms are not really necessary. Simple spring mountings of the mirrors, with screws that can be turned to change the aim of the mirrors, are all that is required.

Before the advent of the laser, a White cell had to be lined up with the aid of visible light, which is easy enough to do in a darkened room. The brightness and convenience of the laser is so great, however, that when the laser became available, the use of visible light for alignment was soon given up. Either a helium-neon laser or a red diode laser may be used.

A permanently-aligned White cell is easily fabricated by cementing the three mirrors into a glass pipe. The laser is the main tool used in the alignment process. Such a cell may be termed “user-friendly” because there is never any question as to whether or not the cell is in alignment. An adjustable White cell can be made by cementing the mirrors into glass ball joints.(Hanst, 1978).

The best overall choice of cell wall material is glass. This choice is based mainly on the chemical fact that most trace gases do not interact with the glass surface. Glass is, after all, one of the most corrosion resistant materials known. Traces of acidic vapors, such as the pollutants NO, NO2, SO2, HCl, HNO3, CO2, HCN, and many others, will remain constant in concentration in a glass cell for a relatively long time. In a metal-bodied cell, such as a stainless steel cell, however, the acidic vapors will be eaten up by the wall material rather quickly. The oxide surface of anodized aluminum has also been found to be unreactive toward acidic trace gases. A cell with inner surfaces of glass and anodized aluminum is the best choice one can make. It is an added bonus that glass is transparent and easy to clean.

Mirror coatings can be gold, silver, aluminum, or other metal. These metals may be protected with an overcoating of silicon monoxide. For work in the ultraviolet, it is necessary to use the “enhanced aluminum”, which is aluminum metal overcoated with Magnesium Fluoride. Multi-layer dielectric coatings may be put on the mirrors to create spectral regions in the ultraviolet or visible that have very high reflection coefficients, near 100%. If the multi-layer coating is transparent in the visible, it then becomes difficult to align the cell. In that case, a layer of aluminum may be put down first, and then the multi-layer coating is put on top.

Popular White cell window materials for infrared use are potassium bromide, potassium chloride, barium fluoride, and zinc selenide. Some people prefer zinc selenide because it is non-hygroscopic (not attacked by water vapor). However, zinc selenide has a high index of refraction, which causes it to reject about 30% of the infrared energy that impinges on it. Then with two zinc selenide windows on a cell, about half of the energy is lost. Another disadvantage of zinc selenide is that the high index causes interference fringes to appear in the infrared spectrum. To minimize these problems, one can have the zinc selenide window manufactured with non-parallel faces (wedged) and given an anti-reflection coating on each side. This makes it very expensive. A better idea would be to use the inexpensive crystal potassium chloride which, if attacked by water, can be re-polished or replaced.

Sometimes it is suggested that a White cell be built with optical windows between the mirrors. Unfortunately, this is not practical because the windows will reject a significant portion of the energy on each pass through the cell. Possibly, Brewster-angled windows could be used, as in lasers, but this would make the cell complicated and expensive.

 With regard to the heating of White cells

With appropriate heating jackets, White cells can be operated at elevated temperatures. All components, including valves and connectors should be at the elevated temperature, because if there are any cold spots on the assembly, vapors may condense there. It is best, however, to avoid using a multiple-pass cell at elevated temperature. When studying products of combustion, for example, one may think it necessary to operate a White cell at high temperature in order to prevent water from condensing in the cell. This is not necessarily the case, however, because the cell may be operated at room temperature and the sample may be diluted with dry air or nitrogen while it is being admitted to the cell. The dilution should be great enough to reduce the water vapor concentration below the room temperature equilibrium value.

One may ask: If the dilution is 5-to-1, isn’t that going to reduce the measurement sensitivity by a factor 5. The answer is no. The dilution will reduce all absorbances by a factor of 5, but this includes the interfering water lines. Measuring compounds like NO, NO2 and SO2, whose bands are badly mixed in with the water lines, will be made easier by the dilution. For these compounds, a 5-to-1 dilution may only reduce their measurement sensitivity by a factor of two. Anyway, in analyzing combustion gases, measurement sensitivity is not a problem. With a good spectrometer, an optical path of a few meters will give measurement sensitivities below one part per million for most toxic gases. If you want to go still lower, just lengthen the path.

Will it be necessary to monitor the dilution with flow meters? Again the answer is no. You will only need to know the water or CO2 concentration in the undiluted sample (something that is almost always known), and then the measured concentration of water or CO2 will tell the dilution factor. Most of the water and CO2 lines and bands will have high absorbance, even in the diluted sample. One must be careful therefore to avoid measurement error due to line saturation. Just as with any other gaseous molecule of low molecular weight, the quantitative measurement of H2O and CO2 should be made only on spectral features of low absorbance.

Another reason for avoiding high temperature measurements is the matter of fast corrosion of the mirrors and other cell parts. Corrosion reactions proceed much faster at high temperature than they do at room temperature.


Cell Maintenance and Restoration.

The interior of a White cell can become soiled. With a glass cell body this is easy to see. When the mirrors are soiled, the laser spots on the field mirror become very dim. The user may then dismantle, clean and re-assemble the cell. The cell body and mirrors can be washed with detergent and water. The parts are then rinsed with distilled water and blown dry with nitrogen. If the mirror surfaces have been corroded by acid vapors, the mirrors may be returned to the manufacturer for re-coating.

Return to Table of Contents