Comment on this article
The Heart of the Flame
Why is a diesel dirty? How can rockets get a cleaner launch? What makes fiber-optic cables strong? The answers all lie in getting a better burn.
April 1994
by Bruce Fellman
When spring came to the Montana ranch where Marshall Long '80PhD grew up, the youngster would grab a propane torch, tie it to the end of a long pole, and drive the family pickup out to the fields. There, he'd set ablaze all the bundles of wheat straw that the sheep had not eaten over the winter and then watch in awe as the plains burned in preparation for plowing.
“I was a firebug from way back,” quips Long, adding that his youthful fascination with combustion eventually led to a career dedicated to learning how to see what goes on in the heart of a flame.
At Yale, the 40-year-old scientist, now professor of mechanical engineering, has plenty of company. The University is home to an ambitious effort by members of the engineering faculty to understand precisely what happens when something—be it pasture land, a log in the fireplace, or an exotic rocket propellant—goes up in smoke. Some of the scientists are looking inside the cylinders of engines to watch, millisecond by millisecond, the evolution of ignition and exhaust; others are trapping soot particles to learn how to manufacturer better paint pigments and purer fiber-optic cables. Still other Yale researchers have been successful in capturing fuel droplets in the process of combustion, figuring out how certain chemicals can reduce pollution, building believable models of turbulence—the process that leads to the mixing critical to keeping every fire going—even creating a flame that “burns” inside a computer.
Surprisingly perhaps, although humankind first domesticated fire more than a half-million years ago, and combustion now literally powers civilization, scientists knew very little about anything beyond combustion fundamentals until quite recently. The basic chemical reaction—add together a fuel and what’s called an oxidizer, heat them to the proper temperature, and you get water vapor and carbon dioxide—was worked out more than 200 years ago by Antoine Lavoisier. Later, by examining the gases and solid particles forged in the flame’s crucible, scientists discovered that combustion involved infinitely more than the simple chemistry that Lavoisier had outlined. But they lacked the tools to determine how these “intermediates,” many of them pollutants that haunt modern society, were made.
Not that ignorance stopped progress. Ingenious engineers sidestepped this lack of basic knowledge by using a time-honored procedure known as “cut and try.” Unfortunately, say scientists and industry officials alike, we’re rapidly reaching a point where tinkering of even the most sophisticated variety will no longer suffice.
And so, armed with a variety of high-tech tools, Yale scientists are carefully dissecting flames atom by atom in an attempt to understand combustion well enough to reduce the process to a series of equations constituting a mathematical model. Engineers can then plug this replica into their plans for, say, a new diesel engine or a trash incinerator and be reasonably certain of how the device will actually work before anyone goes through the difficult and expensive process of building it. “Reliable prediction—that’s the holy grail,” says Long.
The laser is the chief tool in this quest. About a quarter-century ago, scientists began skewering flames with laser beams to see what light-scattering—as researchers call the interaction between the laser and the atoms involved in combustion—might reveal about the cast of characters that make fleeting appearances on fire’s stage. This approach provided a gold mine of information.
Long explains that when light encounters an atom or molecule in the flame, any of several things can happen. “One we term Rayleigh, or elastic, scattering,” he says, noting that in this case, the light bounces off the molecule, and neither is changed in any fundamental way. (Rayleigh scattering, which was named in honor of a 19th-century English physicist, accounts for the sky’s blue color.)
With inelastic, or Raman, scattering, the color of the laser beam changes in a predictable—and measurable—way when it meets a particular molecule. There is also the possibility that one of the chemical players in the fiery drama will react with the light and fluoresce, that is, give off a characteristic color signal that reveals its identity.
Flame detectives, working in conjunction with Yale’s Center for Laser Diagnostics, have modified the basic light-scattering laser techniques in a variety of ways. Long, for example, shines broad sheets of laser light through flames. “I’d like to measure the entire flow field of a flame at once,” he explains. “If you focus a laser on a single point in the flow, you can only measure one point at a time. A laser sheet gives you a two-dimensional picture, but even that’s not enough. Every real-world flame is turbulent and three-dimensional. To understand how it behaves, we’re developing a system that uses television cameras and image intensifiers that are sensitive to very low light levels.”
At present, turbulence is proving an elusive quarry, but to see a reasonable facsimile of a simple flame, you need look no further than Mitchell Smooke’s computer screen. There, the professor of mechanical engineering displays a study in vibrant red, blue, green, and yellow—a portrait of a two-dimensional, or laminar, flame similar to that of a candle or a Bunsen burner.
The ease with which the drawing is accomplished belies the complicated scientific and computational work Smooke’s art requires. The model of how a basic fuel like propane burns, says the engineer, is the result of solving from “half a million to several million equations simultaneously.”
That math is the shorthand researchers use to describe the physical and chemical properties that govern a flame’s life, and only recently have computers become powerful enough to handle the requisite number-crunching with any kind of speed. “Five years ago, this problem—depicting what we term an axisymmetric diffusion flame—took 150 hours of supercomputer time,” notes Smooke. “Today, we can solve it in 8 to 10 hours on a desktop supercomputer.”
Modeling what happens when a more complex hydrocarbon like gasoline ignites will have to await improvements in technology, particularly the ability to link networks of powerful computers together to work toward a common goal (Smooke developed his model using the recently inaugurated parallel computer system at the National Supercomputing Center at Cornell). However, despite this inability to model the complicated fuels important in everyday life, Smooke says that his computerized fire has already led to some exciting discoveries that may have real-world payoffs.
The researcher explains that with a model, one can easily change key conditions of a burn to learn how they affect some of the important intermediate and end products. In a just-completed set of experiments, Smooke examined how the temperature of a flame can lead to the production of nitrogen oxides, important components of smog. There are two primary mechanisms, the scientist says. One is supposed to dominate at temperatures above 2,600 degrees Fahrenheit; the other leads to nitrogen oxide pollution at lower temperatures.
But when Smooke ran the model, he discovered something unexpected. Above 2,600 degrees, the primary avenue to nitrogen oxide production appears to be via the lower temperature route. “So we could be seeing something we didn’t know,” he says. “Experiments using laser diagnostics in real flames will tell us whether this finding makes sense.”
Lasers anchor the model to reality; the model suggests new places in a flame to probe. Together, they and other diagnostic techniques in use at Yale can lead to important improvements in the broad range of industrial processes that make use of combustion. For example, Marshall Long and his colleagues are working with Texaco—using an engine equipped with windows through which lasers can shine—on a project to help develop the diagnostics technology that the company hopes to use to characterize the way its gasoline additives perform.
Assistant professor of mechanical engineering Alessandro Gomez studies the way sprays of various kinds of fuels ignite, research of interest to NASA, as well as to anyone who drives or makes a car with fuel injectors or a diesel engine (understanding how to characterize and deliver sprays of a consistent droplet size is also critical to manufacturers of ink-jet printers and certain medicines). Gomez has designed a testing apparatus that can serve as an engine surrogate.
“A diesel often produces a tremendous amount of soot,” notes the scientist. “This is caused by the incomplete mixing of fuel and oxidizer, and by improving the atomization mechanism, you can lessen one important pollutant.”
But not everyone wants to decrease particulate matter, as soot and other of combustion’s solid leftovers are more properly known. “Sometimes, these particles are useful products, not nuisances,” notes Daniel Rosner, a professor of chemical engineering and director of Yale’s High Temperature Chemical Reaction Engineering Laboratory. “We’re interested in learning how to control the formation and fate of particulates.”
So are numerous federal agencies, most prominently the U.S. Department of Energy and the Air Force, and private industries, among them Shell, DuPont, and General Motors, each of which has helped fund Rosner’s research. The scientist, whose investigations have encompassed such topics as ash formation in coal and the development of jet fuels, is particularly interested these days in the “road-building” technology required for the proposed information superhighway. The ability to move vast quantities of information depends, says Rosner, on the availability of “optical wave guides”—fiber-optic cables, in common parlance. These transmit the light pulses that are used for communications, and they have to be very pure to avoid light losses.
The cables are made by a combustion process in which powders are burned at high temperatures to create particles of glass that can then be harvested and crafted into cable. What looks like a sooty procedure in desperate need of cleanup is, in fact, “the synthesis of a very valuable material,” says Rosner.
Using flames to manufacture substances is actually an ancient technique—and a rather imprecise one. But the production of the fiber-optic cables of the future requires a knowledge of exactly what’s happening in the fiery crucible and the ability to “tweak the flame” accordingly. Equipped with lasers and an electron microscope to examine particles as they form, Rosner is out to custom-tailor soot.
On the other hand, Lisa Pfefferle, an associate professor of chemical engineering, is examining ways to eliminate particulates and pollutants like nitrogen oxide in diesels and gas turbines. “Basically, these substances are the result of incomplete combustion,” she explains, adding that many of them cause health problems.
Often, manufacturers who produce noxious molecules try to eliminate them after the fact, using such devices as catalytic converters and smokestack scrubbers designed to purge exhaust of at least some of its bad breath. But Pfefferle is attempting to tackle the problem at its source and come up with “a way to have more complete combustion and not produce the pollutants in the first place.” The researcher’s line of attack involves taking “snapshots” of the relative concentrations of stable and unstable molecules as they form, thus helping to ensure that only the desirable substances are created. Pfefferle has been experimenting with what is known technically as catalytic combustion. In essence, the burn takes place in the presence of a catalyst such as platinum, the main ingredient in automobile catalytic converters; the catalyst alters the fundamental combustion chemistry in a way that may enable humanity to breathe easier.
Pfefferle’s research has obvious commercial possibilities, and in the late-1980s, she, her father, and her husband formed Precision Combustion, Inc., a company based in New Haven’s Science Park that is exploring the application of catalysts in engine designs. Indeed, much of Yale’s combustion work could result in patents and spin-off technologies. There is, however, an important caveat. Before scientists can say that they understand combustion and have tamed it, they have to come to grips with turbulence. If you look at a normal flame—be it a campfire or the fiery exhaust from the space shuttle’s main engines—it’s obvious that the burn is anything but simple.
Turbulence has been almost impossible to characterize in mathematical shorthand. The problem is bigger than the biggest supercomputers we possess, says Katepalli Sreenivasan. A professor of mechanical engineering, he has tried to break the process into its component parts and deal only with what seem to be important and universal features rather than with everything at once.
“Each turbulent flow is different in its own way, but there do appear to be common elements,” says Sreenivasan, who frequently uses lasers and dye tanks to find order in what at first glance looks like chaos.
The researcher has recently come up with a computerized version that closely resembles the real thing. “If I’m correct,” Sreenivasan explains, “then it means that I have a reasonable understanding of the physics that creates turbulence.”
That would be an important achievement, for being able to model just the physical side of the process could, for example, help airline designers reduce the turbulence-induced drag on jet aircraft. “A 1-percent reduction could save a billion dollars a year,” says Sreenivasan.
But for all its potential utility, his model has serious limitations. The turbulent combustion taking place in the heart of a flame involves both physics and chemistry, and at this point, Sreenivasan’s computer fire takes only physics into account. Clearly, combustion researchers have plenty of work left before the domestication of fire is complete.
“We’re getting closer,” says Marshall Long. “There’s still a lot of 'cut and try' in the design of devices that use combustion, and there’s much that we don’t know. But our approach is starting to have an impact.”  |
