Wind measurements: Trimethyl aluminum (TMA) chemical release technique

An Introduction to Space Instrumentation, Edited by K. Oyama and C. Z. Cheng, 47 51. Wind measurements: Trimethyl aluminum (TMA) chemical release technique M. F. Larsen Deparment of Physics and Astronomy, Clemson University, USA Chemical release techniques have been used since the earliest days of sounding rocket measurements in the upper atmosphere. The measurements require no telemetry or other complicated electronics and the mechanical systems are relatively simple and very robust. The technique was therefore used early on when more complicated measurement techniques were difficult to implement. Chemical release techniques have remained important, however, even as rocket payloads have become more complex. In particular, the tracer measurements remain important as one of the primary techniques for obtaining neutral wind measurements in the upper mesosphere and thermosphere. Various metal vapor tracers have been used for wind measurements in the upper atmosphere, but the most widely used tracer has become trimethyl aluminum (TMA), which has been used in hundreds of rocket launches since the 1960 s. The advantages of TMA include the chemiluminescence produced by the tracer, which makes wind measurements possible at any time during the night, and the fact that the tracer is a liquid, which makes it possible to control both the timing and duration of the release more accurately. An overview of the TMA wind measurement technique is presented, including the design of TMA release hardware and recent results from TMA measurements. Key words: Sounding rockets, thermospheric winds, measurement techniques. 1. Introduction The International Geophysical Year (IGY) in 1958 represents the first coordinated effort to obtain measurements of basic thermodynamic and dynamic parameters in the upper atmosphere. The development of sounding rocket vehicles at the same time that could carry instrumented payload into the thermosphere/ionosphere region made it possible for the first time to make in situ observations of the properties of the region. The chemical tracer technique was quickly developed during the same period as a technique for probing the region, especially the wind profiles, in the near-space region. Some of the first such measurements used sodium as a tracer (see, e.g., Bedinger et al., 1958). When vaporized and released as a gas in the upper atmosphere, sodium produces a bright green resonant emission in response to solar illumination. The technique therefore required that the trails be released in twilight conditions when the trails in the upper atmosphere are illuminated by the sun but observing sites on the ground are in darkness. The star background in the photographs of the trails taken from the ground-based sites were used to determine the look angle to each trail segment. Combining photographs taken simultaneously from several different spatially-separated camera sites made it possible to use triangulation to determine the location of a trail as a function of altitude. Trails were observed over periods of afew minutes to many tens of minutes, and wind profiles were obtained from the change in the position coordinates with time. A description of the results from an early series of twelve launches can be found in the article by Edwards et al. (1963). Sodium tracers were used to a limited extent through the 1960 s and into the early 1970 s, but already in the early Copyright c TERRAPUB, 2013. part of the 1960 s interest was developing in the use of aluminum oxide as a tracer. Vaporization of metal sodium requires a thermite burner to develop sufficiently high temperatures to vaporize the sodium. Such a system is mechanically rugged, but the high temperatures can adversely affect electronics and other payload components. Other drawbacks were the fact that the releases could not be stopped or otherwise modulated after they were initiated, and the requirement for twilight conditions limited the time when the experiments could be carried out to a short period near dusk or dawn. The details of the initial development of aluminum tracer techniques is not clear, but interest in the metal was likely the result of aluminum grenade deployments in the upper atmosphere that were designed to provide information about atmospheric temperature profiles from the propagation of sound waves generated by the detonation of the grenades. The aluminum dust that was a by-product of the detonations was found to be visible from the ground. A number of solid rocket propellants also contain aluminum powder, and the rocket exhaust trails are often visible for a period of minutes due to the chemiluminescence associated with the oxidation of the aluminum. The aluminum grenade technique was first used in the early 1960 s as a way to obtain neutral drift measurements in the upper atmosphere (e.g., Stroud et al., 1960). Soon after, the first trimethyl aluminum wind measurements were made (Rosenberg et al., 1963). Trimethyl aluminum (TMA) is a pyrophoric liquid that burns on contact with oxygen. The liquid itself is clear and has a density and viscosity similar to kerosene. A pressurized container can be used to eject the TMA in a continuous stream. The spontaneous reaction that occurs between the liquid and atmospheric oxygen leads to the desired aluminum deposition without the requirement for explosives or high-temperature burners. An additional benefit is that 47

48 M. F. LARSEN: TMA WIND MEASUREMENT TECHNIQUE Fig. 1. Drawing shows a cutaway section of a typical TMA release canister. The internal section of the canister has two sealed sections separated by a free-floating piston. The upper section holds the liquid trimethyl aluminum, and the lower section holds nitrogen gas pressurized to approximately 100 p.s.i. The exit path for the TMA is through a pyrotechnic valve and a solenoid valve used to modulate the flow. Fig. 2. Figure shows a drawing of a TMA canister in a 14-inch payload skin section. The plumbing on the TMA section deck plate is connected directly to the payload skin surface where the TMA is ejected through a nozzle. the flow of the liquid can be modulated with a solenoid valve, for example. For these reasons, TMA quickly became the preferred chemical tracer for sounding rocket neutral wind measurements. A large number of such measurements were made in the 1960 s and 1970 s when relatively little was still known about the structure of the upper atmosphere and few alternative techniques were available. An extensive summary of the chemical release wind measurements that have been carried out since 1958 are given in the article by Larsen (2002), including references to many of the published results. Somewhat surprising is the fact that the TMA technique is still the preferred method for in situ wind profile measurements in the upper atmosphere. A few wind determinations on rockets with electronic instruments have been attempted, but with limited success. The TMA technique has a number of drawbacks, but the technique is robust and produces high-resolutions and accurate wind profiles in a straightforward way. 2. Release Technique Description Trimethyl aluminum is a liquid with physical properties similar to clear kerosene. Figure 1 shows the cross-section of a typical payload canister used to produce a chemical trail. The interior of the canister has two sections, one for the TMA and one for the pressurized nitrogen gas, which is compatible with TMA. The two sections are separated by a free-floating piston with several O-ring seals to separate the liquid section from the gas. Typical quantities are afew kilograms of TMA for a 100-km long trail, and nitrogen pressurized to 100 pounds per square inch. The top deck plate has the plumbing, including several valves and a nozzle, that allows the TMA to exit the payload. The first valve in the exit path is a safety valve that is closed when the canister is transported or when critical operations are being carried out on the payload. The safety valve is operated manually and is opened prior to launch. Next is the normally-closed pyrotechnic valve that is actuated in flight by payload electronics to initiate the TMA flow. Finally, there is a normally-open solenoid valve operated by a programmable timer to produce the desired trail modulation. The particular design shown in the figure was developed by the author, but nearly all the TMA payloads that have been flown have had a similar design. The typical sequence of events is that the pyrotechnic valve is fired at a predetermined time after launch. The force of the nitrogen gas on the internal piston starts the TMA flow. The solenoid valve can create structure in the trail by turning the liquid flow on and off in a predetermined sequence to produce a sequence of puffs or blobs. The same valve can also be used to turn off the flow in altitude ranges where the TMA is not a useful tracer. For example, the flow can be turned off at higher altitudes near apogee where the TMA diffuses too rapidly to be useful, thus creating separate trails on the up-leg and down-leg portion of the trajectory. An example of the configuration of the payload canister inside the payload skin is shown in Fig. 2. The plumbing is connected directly to a nozzle vented through the payload skin. The flow rate is controlled by the nozzle orifice size. Since the system is sealed, the nitrogen gas pressure decreases with time over the period that the release occurs. The total pressure decrease will clearly depend on the volume occupied initially by the liquid, and the total volume of the canister. The design parameters are usually chosen so that the final nitrogen pressure is approximately half of the initial value. Since the exit nozzle size is fixed, the liquid flow rate will decrease with time from the start to the end of the release. TMA is a pyrophoric liquid that ignites on contact with oxygen and reacts explosively on contact with water. It is therefore hazardous to the personnel loading the payload canisters. Protective fire suits are required for the loading operation. The payload canister is connected to a sealed system that is also connected to nitrogen, kerosene, and TMA tanks for loading. The system is flushed with nitrogen gas and kerosene to reduce the residual levels of oxygen and water vapor inside the system to acceptable levels before introducing the TMA liquid. In the upper atmosphere where the ambient oxygen concentrations are small, the reaction between TMA and oxy-

M. F. LARSEN: TMA WIND MEASUREMENT TECHNIQUE 49 Fig. 4. The lefthand panel shows a standard monochrome image of the trail. The center panel shows a spectrograph image of the same trail in which the emission wavelengths are separated horizontally, with shorter wavelengths on the left and longer wavelengths on the right. The righthand panel shows the spectrrograph information for two altitudes in graphical form. The upper curve corresponds to an altitude that is in sunlight and shows the characteristics blue-wavelength emissions associated with aluminum. The lower curve is for an altitude in darkness and shows a broad-band white emission spectrum associated with the chemiluminescence of the TMA. Fig. 3. Photograph of a TMA trail released in nighttime conditions from a rocket launched from Kagoshima Space Center in Japan during the Sporadi E Experiment over Kyushu (SEEK). The image was published on the cover of the Japanese astronomy publication Tenmon and shows a chemiluminescent trail observed from a ground site with standard camera equipment. gen is slow and produces a long-lasting chemilumiscence that can be observed from the ground with both the naked eye and white-light cameras for periods of a few minutes to many tens of minutes. An example of a TMA release from the Sporadic E Experiment over Kyushu launch (Larsen et al., 1996) is shown in Fig. 3, taken from the cover photograph of the Japanese astronomy journal Tenmon. The trail is initially a nearly straight line since it is released along the rocket trajectory. The rotational and speed shears in the upper atmospher distort the trail. The effect of a high-speed flow in a narrow height range is evident in the lower part of the trail. The visible light from the trail was associated with the chemiluminescence produced by the TMA/oxygen reaction. There was no sunlight at the time of the release. TMA that is released in twilight conditions produces both the white light associated with the chemiluminescence and a strong blue light associated with an aluminum resonace emission line. Both contributions are shown in the example in Fig. 4, which is from one of the first TMA releases carried out in the upper atmosphere (Rosenberg et al., 1964). A photograph of the trail is shown in the lefthand panel in the figure. The center panel shows the spectrograph image of the same trail. In the upper half of the spectrograph image several resonance emission lines can be seen. In the lower part, the emissions are nearly uniform over a broad part of the spectral range. In this case, the lower part of the trail was in darkness and the upper part was in sunlight. The righthand panel shows the graphical representation of the trail emissions at two selected altitudes. The strongest emission line occurs at 484.2 nm. The color image in Fig. 5 shows the effect clearly. The photograph was taken from Arctic Village, Alaska, of a trail released in evening twilight conditions from a rocket launched from Poker Flat Research Range in Alaska. The experiment, including the chemical release results, were described in the article by Mikkelsen et al. (1981). The lower part of the trail, which is below the Earth s shadow, shows the weaker white color associated with chemiluminescence. The upper part, which is in sunlight, shows the strong blue color associated with the resonance emission line. 3. Measurement Technique Obtaining wind profiles from the TMA trails requires observations of the trails from at least two spatially-separated, ground-based camera sites so that a triangulation can be carried out to determine the position of each trail segment. The change in the horizontal position with time then gives the velocity at each altitude where the triangulation is carried out. The background star field is used to determine the right ascension and declination of points within the image. The astronomical coordinates can, in turn, be converted to horizon coordinates, i.e., azimuth and elevation. The latter information is then used to obtain the intersections of linesof-sight from each of the observing sites. The determination of of the astronomical coordinates is usually accomplished by a least-squares fitting of the image star positions to the corresponding star positions in a star catalog. In the earliest experiments, the intersection points were obtained by projecting the film images with a system of projectors that were set up with a scaled-down geometry and look directions that corresponded to those of the original camera sites. Adjusting the focus for the projectors made it possible to bring

50 M. F. LARSEN: TMA WIND MEASUREMENT TECHNIQUE Fig. 6. The figure shows photographs of a trail obtained simultaneously from two different ground-based sites. The triangulation is carried out as a computer-assisted manual procedure for identifying and matching trail features. The red dots show the trail centers. The dashed yellow lines show the corresponding lines-of-sight from the other site. Fig. 5. Photograph of a TMA trail released in twilight conditions at dusk from a rocket launched from Poker Flat Research Range in Alaska on Feb. 28. 1978. The image was taken from Arctic Village, Alaska. The trail shows the broad-band white emissions from the lower part of the trail, which is in darkness, and the resonant blue emission associated with aluminum in the upper part, which is in sunlight. successive trail sections into focus and thus to determine the scaled intersection points. Figure 6 shows two images of the same trail released over Alaska during an auroral event from two different camera site locations. After the fitting of the star field is completed, the next task is to determine the location of the center of the trail. The analyst inputs initial guesses, and the computer program refines the estimate of the location for each trail segment. The result is shown as the red dots in the two images. Each dot has a corresponding line-of-sight direction from the other site, shown as the dashed yellow lines in the images. Matching the corresponding points produces an altitude profile of the geographical position of the points along the trail. An example of the result of the triangulation is shown in Fig. 7 (Bishop et al., 2004). The lefthand image shows a trail released in New Mexico photographed from one of the ground-based camera sites. The altitude information obtained from the triangulation procedure has been used to label key features in the trail. The velocity profile obtained from the time sequence of images is shown in the righthand panel in the figure. Since 1958 more than 500 chemical release wind profile measurements have been made from sounding rockets. The article by Larsen (2002) summarized much of the available Fig. 7. Lefthand panel shows an image of TMA trail released from a rocket launched from White Sands, New Mexico. The image was obtained from a site at Sunspot, New Mexico. The altitudes for key sections of the trail are labeled. The corresponding wind profile obtained by triangulation is shown in the righthand panel. The distortions of the trail created by the speed and rotational shears in the wind profile are evident in the photograph of the trail. data and gave an extensive list of citations to publications, both those in the refereed literature and those only available in various contractor or government agency reports that presented wind profile results. At the time, approximately 400 profiles were available. Some of the measurements in the first two decades used sodium and lithium trails, but the majority of the releases used TMA. The statistics of the wind profile data are described in detail by Larsen (2002), but Fig. 8 shows one of the key findings of the study. Specifically, overlaying all the profiles in the same plot shows large winds in the lower thermosphere occurring most frequently in the altitude range between 100 and 110 km. The wind speeds are frequently in the range between 100 and 150 m s 1, i.e., much larger than would be expected from tidal theory, for example. There are also large wind shears above and below the wind maxima, and

M. F. LARSEN: TMA WIND MEASUREMENT TECHNIQUE 51 Fig. 8. The figure shows the superposition of wind profiles obtained from more than 400 individual rocket launches. Nearly all the profiles were obtained from TMA trail releases, although the data set includes afew sodium or lithium release wind profiles. The overall shape of the superposed curves show the characteristic wind maximum in the 100 to 110-km altitude range. the shears in the altitude range between 90 and 100 km are often large enough to be unstable or close to instability. 4. Future Improvements It appears to be difficult to significantly improve the mechanical design of the system for deploying the TMA chemical. A well-known problem has been the occurrence of a so-called re-entry bag on the down-leg portion of the trail. Specifically, a significant fraction of the TMA that is released, perhaps as much as 50% does not react with oxygen initially and continues along the rocket trajectory. As the particles enter the turbulent region below an altitude near 100 km on the down leg, the reaction rate increases substantially, resulting in a strong increase in the brighness and width of the trail at those heights. The re-entry bag is often visible for 30 or 40 minutes. Attempts were made in the late 1960 s and early 1970 s to increase the reaction rate as the trail was released. Heaters were attached to the payload skin and to the exit nozzle area. Mixed releases of water and TMA were also attempted since TMA reacts more violently with water, either in liquid or vapor form, as an attempt to increase the trail brightness. None of these techniques produced a significant improvement in the brightness, uniformity, or duration of the releases. The basic triangulation procedure used to obtain the trail positions has not changed significantly since the early experiments in the 1960 s. The star-field fitting and trail segment matching has benefitted from the availability of computer techniques as an aid in accomplishing the fitting and matching, but the techniques remain the same, at least in principle. Efforts to automate the process further are underway. Advances in image feature or pattern recognition will likely be helpful in this respect. A more critical need is for automation of techniques that can be used to obtain the trail expansion rates as a function of altitude and horizontal position. The expansion rates of the trails provide one of the few techniques available for studies of turbulence in the critical region near the turbopause. The first high-resolution, quantitative measurements of this type were made by Justus (1967) and Rees et al. (1972). The latter study, in particular, provided a very careful and detailed analysis of the relevant processes and found strong evidence for enhanced diffusion rates above the apparent turbopause altitude, suggesting that the turbulent transport effects were important even in part of the altitude range where the trail appeared to be laminar. Only one other similar analysis has been carried out, namely the study by Bishop et al. (2004) of the TMA trails released from a launch in New Mexico. The results were similar to those of Rees et al. (1972). The small number of such analyses is due to the labor intensive nature of the image analysis required to extract turbulent trail expansion measurements. The rapid advances made in image processing and image analysis techniques offer the potential for automating the analysis and thus significantly increasing the available data related to turbulence in this critical altitude range of the atmosphere. Acknowledgments. The author gratefully acknowledges partial support by NASA Grants NNX09A120G, NNX10AL26G, and NNX11AE17G while writing the manuscript. References Bedinger, J. F., E. R. Manring, and S. N. Ghosh, Study of sodium vapor ejected into the upper atmosphere, J. Geophys. Res., 63, 19 29, 1958. Bishop, R. L., M. F. Larsen, J. H. Hecht, A. Z. Liu, and C. S. Gardner, TOMEX: Mesospheric and lower thermospheric diffusivities and instability layers, J. Geophys. Res., 109, doi:10.1029/2002jd003079, D02S03, 2004. Edwards, H. D., M. M. Cooksey, C. G. Justus, R. N. Fuller, and D. L. Albritton, Upper-atmosphere wind measurements determined from twelve rocket experiments, J. Geophys. Res., 68, 3021 3032, 1963. Justus, C. G., The eddy diffusivities, energy balance parameters, and heating rate of upper atmospheric turbulence, J. Geophys. Res., 72, 1035 1039, 1967. Larsen, M. F., Winds and shears in the mesosphere and lower thermosphere: Results from four decades of chemical release wind measurements, J. Geophys. Res., 107, doi:10.1029/2001ja000218, 2002. Larsen, M. F., S. Fukao, M. Yamamoto, R. Tsunoda, K. Igarashi, and T. Ono, The SEEK chemical release experiment: Observed neutral wind profile in a region of sporadic E, Geophys. Res. Lett., 25, 1789 1792, 1998. Mikkelsen, I. S., T. S. Jorgensen, M. C. Kelley, M. F. Larsen, E. Pereira, and J. Vickrey, Neutral winds and electric fields in the dusk auroral oval, 1. Measurements, J. Geophys. Res., 86, 1513 1524, 1981. Rees, D., R. G. Roper, K. H. Lloyd, and C. H. Low, Determination of the structure of the atmosphere between 90 and 250 km by means of contaminant releases at Woomera, May 1968, Philos. Trans. R. Soc. London A, 271, 631 663, 1972. Rosenberg, N. W., D. Golomb, and E. F. Allen, Chemiluminescence of trimethyl aluminum released into the upper atmosphere, J. Geophys. Res., 68, 5895 5898, 1963. Rosenberg, N. W., D. Golomb, E. F. Allen, Resonance radiation of A10 from trimethyl aluminum released into the upper atmosphere, J. Geophys. Res., 69, 1451 1454, 1964. Stroud, W. G., W. Nordberg, W. R. Bandeen, F. L. Bartman, and P. Titus, Rocket-grenade measurements of temperatures and winds in the mesosphere over Churchill, Canada, J. Geophys. Res., 65, 2307 2323, 1960. M. F. Larsen (e-mail: mlarsen@clemson.edu)