MIT Research News' Journal
[Most Recent Entries]
[Calendar View]
Wednesday, July 23rd, 2014
| Time |
Event |
| 12:00a |
Making the cut Diode lasers — used in laser pointers, barcode scanners, DVD players, and other low-power applications — are perhaps the most efficient, compact, and low-cost lasers available.
Attempts have been made over the years to amplify the brightness of these valuable lasers for industrial applications, such as welding and cutting metal. But boosting power usually means decreasing beam quality, or focus. And the beam never gets intense enough to melt metal.
Now MIT Lincoln Laboratory spinout TeraDiode is commercializing a multikilowatt diode laser system that’s bright enough to cut and weld — even through a half-inch of steel — at greater efficiencies than today’s industrial lasers.
The 4-kilowatt TeraBlade runs on a novel power-scaling technique developed at MIT that manipulates individual diode laser beams into a single output ray. This allows for boosting power of a diode laser, while preserving a very focused beam.
“[The TeraBlade] has comparable beam quality as compared with traditional manufacturing lasers, such as carbon dioxide, disk, and fiber,” says TeraDiode co-founder and vice president Robin Huang, a former Lincoln Laboratory researcher and TeraBlade co-inventor. “However, because the TeraBlade is a direct-diode laser, it has the highest efficiency and lowest cost of ownership as compared with these other lasers.”
Huang says TeraBlade represents a “third generation” of industrial lasers. The first generation, which evolved a few decades ago, was carbon dioxide lasers, in which electricity runs through a gas to produce light. These are very bright, but can be as large as trucks and operate at about 20 percent efficiency.
Then came diode-pumped solid-state (DPSS) lasers — including disk and fiber — that first transfer energy from diode lasers into a medium, usually a crystal, before converting it into a laser beam. These operate only up to about 30 percent efficiency.
But the TeraBlade, aptly called a “direct-diode” laser, uses light directly from the diodes, skipping the DPSS conversion step and saving energy, Huang says. This means the TeraBlade operates with just as much power and brightness as all other industrial lasers — about 2,600 megawatts per square centimeter per steradian — at roughly 40 percent efficiency.
Wavelength beam combining
At the core of the TeraBlade is a power-scaling technique known as wavelength beam combining (WBC), or incoherent beam combining, developed by Huang and former Lincoln Laboratory researcher and TeraDiode co-founder Bien Chann, who is now the company’s vice president and chief technology officer.
Diode lasers are tiny semiconductor devices that, when electrically charged, cause electrons to create photons of the same wavelength, or color, traveling in the same direction. When fed through an output collimation lens, this creates a ray of laser light.
An individual diode laser — in, say, a laser pointer — can emit a beam, in infrared and near-infrared wavelengths, that can be tightly focused to a very small spot, but with little power, Huang explains. Overlapping many similar beams at differing wavelengths, however, produces a beam that focuses on a small spot, making it very intense. And the number of overlapping beams with differing wavelengths can be very high.
In the early 2000s, Huang, Chann, and Lincoln Laboratory colleagues built a few prototype lasers based on WBC technology. One, which reached a power level of 50 watts, “was a world’s record for diode laser brightness at that time,” Huang says.
In 2009, Huang and Chann — along with Fred Leonberger, a former Lincoln Laboratory staffer who now serves on TeraDiode’s board of directors, and former CEO David Sossen — launched TeraDiode, now in Wilmington, Mass., to bring the technology to market. (The company’s current CEO is photonics entrepreneur Parviz Tayebati.)
Today, the WBC-based TeraBlade is a laser module that contains diode laser bars (long arrays of diode lasers), a transform lens, a diffraction grating, and an output lens. The light from the diode lasers passes through a transform lens, onto the carefully positioned diffraction grating, a plate of glass scratched with parallel lines. However, instead of dispersing light at different angles — which it’s designed to do — the grating forces the beams into the same direction, superimposing them on one another.
There are a few other multikilowatt direct-diode lasers, but they run on another popular and similar power-scaling technique, called side-by-side or spatial beam combining, that joins together the beams of similar wavelengths. As the number of diode lasers increases, the beam quality degrades, resulting in a large focused spot, limiting the beam’s intensity.
This means the TeraBlade outputs a beam roughly 100 times brighter than these scaled-up direct-diode laser models, Huang says.
Each TeraBlade module outputs about 1,000 watts and can be scaled to increase power. The company has also developed a commercial TeraBlade system: a 3-foot cube that comes with multiple laser engines, a control computer, power supplies, and an output head for welding and cutting, among other components.
Sky’s the limit
Increasingly, the TeraDiode technology is finding customers in countries such as Japan and Germany, where energy costs are high, Huang says.
In April, the company began shipping its system to Panasonic Welding Systems in Europe and Japan. Other customers include top global builders of industrial laser-based machines and system integrators.
“More broadly, our vision is to revolutionize the laser industry,” Huang says, by introducing powerful direct-diode lasers to various applications across the globe.
In the future, he adds, the company is also looking toward defense applications. One idea is to build a laser that acts as a heat-seeking missile deterrent: It fires infrared laser light at the missile, which would confuse the missile’s programming, and cause it to lose its target. The laser’s compact design would allow it to be mounted on a fighter jet.
With the TeraBlade technology, Huang says, “The sky is the limit — literally.” | | 12:00a |
Building up bamboo Bamboo construction has traditionally been rather straightforward: Entire stalks are used to create latticed edifices, or woven in strips to form wall-sized screens. The effect can be stunning, and also practical in parts of the world where bamboo thrives.
But there are limitations to building with bamboo. The hardy grass is vulnerable to insects, and building with stalks — essentially hollow cylinders — limits the shape of individual building components, as well as the durability of the building itself.
MIT scientists, along with architects and wood processors from England and Canada, are looking for ways to turn bamboo into a construction material more akin to wood composites, like plywood. The idea is that a stalk, or culm, can be sliced into smaller pieces, which can then be bonded together to form sturdy blocks — much like conventional wood composites. A structural product of this sort could be used to construct more resilient buildings — particularly in places like China, India, and Brazil, where bamboo is abundant.
Such bamboo products are currently being developed by several companies. The MIT project intends to gain a better understanding of these materials, so that bamboo can be more effectively used structurally. To that end, MIT researchers have now analyzed the microstructure of bamboo and found that the plant is stronger and denser than North American softwoods like pine, fir, and spruce, making the grass a promising resource for composite materials.
“Bamboo grows extensively in regions where there are rapidly developing economies, so it’s an alternative building material to concrete and steel,” says Lorna Gibson, the Matoula S. Salapatas Professor of Materials Science and Engineering at MIT. “You probably wouldn’t make a skyscraper out of bamboo, but certainly smaller structures like houses and low-rise buildings.”
Gibson and her colleagues analyzed sections of bamboo from the inside out, measuring the stiffness of each section at the microscale. As it turns out, bamboo is densest near its outer walls. The researchers used their data to develop a model that predicts the strength of a given section of bamboo.
The model may help wood processors determine how to assemble a particular bamboo product. As Gibson explains it, one section of bamboo may be more suitable for a given product than another: “If you wanted a bamboo beam that bends, maybe you’d want to put the denser material at the top and bottom and the less dense bits toward the middle, as the stresses in the beam are larger at the top and bottom and smaller in the middle. We’re looking at how we might optimize the selection of bamboo materials in the structure that you make.”
Gibson and her colleagues have published their results in the Journal of the Royal Society: Interface.
A look at bamboo, from the inside out
For their experiments, the researchers analyzed specimens of moso, the main species of bamboo used in China. Like most types of bamboo, moso grows as hollow, cylindrical stalks, or culms, segmented by nodes along the length of a stalk. Bamboo can reach heights of 20 meters — as tall as a six-story building — in just a few months. The stalks then take another few years to mature — but still much faster than a pine tree’s statelier, decades-long growth.
“One of the impressive things is how fast bamboo grows,” Gibson notes. “If you planted a pine forest versus a bamboo forest, you would find you can grow far more bamboo, and faster.”
Researchers used electron microscopy to obtain images of the bamboo microstructure and create complete, microscale cross-sections of the entire culm wall at different heights along the stalk.
The resulting images showed density gradients of vascular bundles — hollow vessels — that carry fluid up and down the stalk, surrounded by solid fibrous cells. The density of these bundles increases radially outward — a gradient that seems to grow more pronounced at higher positions along a stalk.
The researchers cut sections of bamboo from the inside out, noting each sample’s radial and longitudinal position along a culm, then gauged the stiffness and strength of the samples by performing bending and compression tests. In particular, they performed nanoindentation, which uses a tiny mechanical tip to push down on a sample, to gain an understanding of bamboo’s material properties at a finer scale. From the results of these mechanical tests, Gibson and her colleagues found that in general, bamboo is stiffer and stronger than most North American softwoods commonly used in construction, and also denser.
The researchers then used the stiffness and density data to create a model that accurately predicts the mechanical properties of bamboo as a function of position in the stalk. Gibson says wood processors that she works with in Canada may use the model as a guide to assemble durable bamboo blocks of various shapes and sizes.
Going forward, the processors, in turn, will send the MIT team composite samples of bamboo to characterize. For example, a product may be processed to contain bamboo along with other materials to reduce the density of the product and make it resistant to insects. Such composite materials, Gibson says, will have to be understood at the microscale.
“We want to look at the original mechanical properties of the bamboo culm, as well as how processing affects the product,” Gibson says. “Maybe there’s a way to minimize any effects, and use bamboo in a more versatile way.”
Oliver Frith, acting director of programme for the International Network for Bamboo and Rattan, headquartered in Beijing, says that very few species of bamboo have been classified, and the lack of knowledge of the material’s microstructure has impaired efforts to design efficient, optimal structural products.
“MIT’s work is very timely and has great potential to support development of the sector,” says Frith, who was not involved in the research. “While bamboo has similarities to wood, as this study shows, the material also has very distinct properties. Although current approaches to developing structural engineered bamboo have tended to focus on mimicking engineered wood products, the future will probably lie in innovating new approaches that can better enhance the natural advantages of this unique material.” | | 9:45a |
Essays in English yield information about other languages Computer scientists at MIT and Israel’s Technion have discovered an unexpected source of information about the world’s languages: the habits of native speakers of those languages when writing in English.
The work could enable computers chewing through relatively accessible documents to approximate data that might take trained linguists months in the field to collect. But that data could in turn lead to better computational tools.
“These [linguistic] features that our system is learning are of course, on one hand, of nice theoretical interest for linguists,” says Boris Katz, a principal research scientist at MIT’s Computer Science and Artificial Intelligence Laboratory and one of the leaders of the new work. “But on the other, they’re beginning to be used more and more often in applications. Everybody’s very interested in building computational tools for world languages, but in order to build them, you need these features. So we may be able to do much more than just learn linguistic features. … These features could be extremely valuable for creating better parsers, better speech-recognizers, better natural-language translators, and so forth.”
In fact, Katz explains, the researchers’ theoretical discovery resulted from their work on a practical application: About a year ago, Katz proposed to one of his students, Yevgeni Berzak, that he try to write an algorithm that could automatically determine the native language of someone writing in English. The hope was to develop grammar-correcting software that could be tailored to a user’s specific linguistic background.
Family resemblance
With help from Katz and from Roi Reichart, an engineering professor at the Technion who was a postdoc at MIT, Berzak built a system that combed through more than 1,000 English-language essays written by native speakers of 14 different languages. First, it analyzed the parts of speech of the words in every sentence of every essay and the relationships between them. Then it looked for patterns in those relationships that correlated with the writers’ native languages.
Like most machine-learning classification algorithms, Berzak’s assigned probabilities to its inferences. It might conclude, for instance, that a particular essay had a 51 percent chance of having been written by a native Russian speaker, a 33 percent chance of having been written by a native Polish speaker, and only a 16 percent chance of having been written by a native Japanese speaker.
In analyzing the results of their experiments, Berzak, Katz, and Reichart noticed a remarkable thing: The algorithm’s probability estimates provided a quantitative measure of how closely related any two languages were; Russian speakers’ syntactic patterns, for instance, were more similar to those of Polish speakers than to those of Japanese speakers.
When they used that measure to create a family tree of the 14 languages in their data set, it was almost identical to a family tree generated from data amassed by linguists. The nine languages that are in the Indo-European family, for instance, were clearly distinct from the five that aren’t, and the Romance languages and the Slavic languages were more similar to each other than they were to the other Indo-European languages.
What’s your type?
“The striking thing about this tree is that our system inferred it without having seen a single word in any of these languages,” Berzak says. “We essentially get the similarity structure for free. Now we can take it one step further and use this tree to predict typological features of a language for which we have no linguistic knowledge.”
By “typological features,” Berzak means the types of syntactic patterns that linguists use to characterize languages — things like the typical order of subject, object, and verb; how negations are formed; or whether nouns take articles. A widely used online linguistic database called the World Atlas of Language Structures (WALS) identifies nearly 200 such features and includes data on more than 2,000 languages.
But, Berzak says, for some of those languages, WALS includes only a handful of typological features; the others just haven’t been determined yet. Even widely studied European languages may have dozens of missing entries in the WALS database. At the time of his study, Berzak points out, only 14 percent of the entries in WALS had been filled in.
The new system could help fill in the gaps. In work presented last month at the Conference on Computational Natural Language Learning, Berzak, Katz, and Reichart ran a series of experiments that examined each of the 14 languages of the essays they’d analyzed, trying to predict its typological features from those of the other 13 languages, based solely on the similarity scores produced by the system. On average, those predictions were about 72 percent accurate.
Branching out
The 14 languages of the researchers’ initial experiments were the ones for which an adequate number of essays — an average of 88 each — were publicly available. But Katz is confident that given enough training data, the system would perform just as well on other languages. Berzak points out that the African language Tswana, which has only five entries in WALS, nonetheless has 6 million speakers worldwide. It shouldn’t be too difficult, Berzak argues, to track down more English-language essays by native Tswana speakers.
“There are folks who have been debating to what degree, when you learn a second language, you’re just starting over and learning the structure of the language,” says Robert Frank, chair of Yale University’s linguistics department. “Another hypothesis is that you think of the new language as a modified version of your language. Some researchers think that such modifications take place at a fairly superficial level. But others think they operate over bits of abstract grammar."
The MIT researchers’ technique could help sharpen that debate, Frank says. The ability to predict features of speakers’ native languages from the syntax of their written English, he says, shows that “there is clearly a reflection of the grammar of the original language. So it’s not that they just start over from scratch.”
In French, for instance, objects follow verbs — unless they’re pronouns, in which case they precede the verbs. In Yiddish, both pronouns and definite objects precede the verb, but other objects do not. So are French and Yiddish verb-object languages or object-verb languages?
Frank would like to see how well the MIT researchers’ technique predicts classifications under other, more fine-grained and abstract typological systems. “What are the underlying features that would determine, ‘Oh, it’s just the pronoun, or it’s only objects that have particular kinds of properties?’” Frank says. “Cautiously, I’m optimistic. I’m excited about the possibility that even more abstract properties are going to be reflected in the English production.” |
|