MIT News

When (and where) work disappears

newsoffice@mit.edu (Peter Dizikes, MIT News Office) — Fri, 24 Feb 2012 05:00:01 +0000

As the United States seeks to reinvigorate its job market and move past economic recession, MIT News examines manufacturing’s role in the country’s economic future through this series on work at the Institute around manufacturing.

The loss of U.S. manufacturing jobs is a topic that can provoke heated arguments about globalization. But what do the cold, hard numbers reveal? How has the rise in foreign manufacturing competition actually affected the U.S. economy and its workers?

A new study co-authored by MIT economist David Autor shows that the rapid rise in low-wage manufacturing industries overseas has indeed had a significant impact on the United States. The disappearance of U.S. manufacturing jobs frequently leaves former manufacturing workers unemployed for years, if not permanently, while creating a drag on local economies and raising the amount of taxpayer-borne social insurance necessary to keep workers and their families afloat.

Geographically, the research shows, foreign competition has hurt many U.S. metropolitan areas — not necessarily the ones built around heavy manufacturing in the industrial Midwest, but many areas in the South, the West and the Northeast, which once had abundant manual-labor manufacturing jobs, often involving the production of clothing, footwear, luggage, furniture and other household consumer items. Many of these jobs were held by workers without college degrees, who have since found it hard to gain new employment.

“The effects are very concentrated and very visible locally,” says Autor, professor and associate head of MIT’s Department of Economics. “People drop out of the labor force, and the data strongly suggest that it takes some people a long time to get back on their feet, if they do at all.” Moreover, Autor notes, when a large manufacturer closes its doors, “it does not simply affect an industry, but affects a whole locality.”

In the study, published as a working paper by the National Bureau of Economic Research, Autor, along with economists David Dorn and Gordon Hanson, examined the effect of overseas manufacturing competition on 722 locales across the United States over the last two decades. This is also a research focus of MIT’s ongoing study group about manufacturing, Production in the Innovation Economy (PIE); Autor is one of 20 faculty members on the PIE commission.

The findings highlight the complex effects of globalization on the United States. “Trade tends to create diffuse beneficiaries and a concentration of losers,” Autor says. “All of us get slightly cheaper goods, and we’re each a couple hundred dollars a year richer for that.” But those losing jobs, he notes, are “a lot worse off.” For this reason, Autor adds, policymakers need new responses to the loss of manufacturing jobs: “I’m not anti-trade, but it is important to realize that there are reasons why people worry about this issue.”

Double trouble: businesses, consumers both spend less when industry leaves

In the paper, Autor, Dorn (of the Center for Monetary and Fiscal Studies in Madrid, Spain) and Hanson (of the University of California at San Diego) specifically study the effects of rising manufacturing competition from China, looking at the years 1990 to 2007. At the start of that period, low-income countries accounted for only about 3 percent of U.S. manufacturing imports; by 2007, that figure had increased to about 12 percent, with China representing 91 percent of the increase.

The types of manufacturing for export that grew most rapidly in China during that time included the production of textiles, clothes, shoes, leather goods, rubber products — and one notable high-tech area, computer assembly. Most of these production activities involve soft materials and hands-on finishing work. “These are labor-intensive, low-value-added [forms of] production,” Autor says. “Certainly the Chinese are moving up the value chain, but basically China has been most active in low-end goods.”

In conducting the study, the researchers found more pronounced economic problems in cities most vulnerable to the rise of low-wage Chinese manufacturing; these include San Jose, Calif.; Providence, R.I.; Manchester, N.H.; and a raft of urban areas below the Mason-Dixon line — the leading example being Raleigh, N.C. “The areas that are most exposed to China trade are not the Rust Belt industries,” Autor says. “They are places like the South, where manufacturing was rising, not falling, through the 1980s.”

All told, as American imports from China grew more than tenfold between 1991 and 2007, roughly a million U.S. workers lost jobs due to increased low-wage competition from China — about a quarter of all U.S. job losses in manufacturing during the time period.

And as the study shows, when businesses shut down, it hurts the local economy because of two related but distinct “spillover effects,” as economists say: The shuttered businesses no longer need goods and services from local non-manufacturing firms, and their former workers have less money to spend locally as well.

A city at the 75th percentile of exposure to Chinese manufacturing, compared to one at the 25th percentile, will have roughly a 5 percent decrease in the number of manufacturing jobs and an increase of about $65 per capita in the amount of social insurance needed, such as unemployment insurance, health care insurance and disability payments.

“People like to think that workers flow freely across sectors, but in reality, they don’t,” Autor says. At a conservative estimate, that $65 per capita wipes out one-third of the per-capita gains realized by trade with China, in the form of cheaper goods. “Those numbers are really startling,” Autor adds.

The study draws on United Nations data on international trade by goods category among developing and developed countries, combined with U.S. economic data from the Census Bureau, the Bureau of Economic Analysis and the Social Security Administration. The study received funding from the National Science Foundation, Spanish Ministry of Science and Innovation, and the Community of Madrid.

The paper has already generated discussion among economists focused on this issue. The large effect found in the paper of overseas competition on U.S. unemployment is “quite plausible, from my experience of surveying and talking to manufacturers,” says Nicholas Bloom, an economist at Stanford University. “That was an important figure, it’s been very well-estimated, they’ve done as good a job as you can.” Bloom also thinks the paper’s analysis of the rise in social insurance payments is “ingenious,” and of value to policymakers.

New policies for a new era?

In Autor’s view, the findings mean the United States needs to improve its policy response to the problem of disappearing jobs. “We do not have a good set of policies at present for helping workers adjust to trade or, for that matter, to any kind of technological change,” he says.

For one thing, Autor says, “We could have much better adjustment assistance — programs that are less fragmented, and less stingy.” The federal government’s Trade Adjustment Assistance (TAA) program provides temporary benefits to Americans who have lost jobs as a result of foreign trade. But as Autor, Dorn and Hanson estimate in the paper, in areas affected by new Chinese manufacturing, the increase in disability payments is a whopping 30 times as great as the increase in TAA benefits.

Therefore, Autor thinks, well-designed job-training programs would help the government’s assistance efforts become “directed toward helping people reintegrate into the labor market and acquire skills, rather than helping them exit the labor market.”

Still, it will likely take more research to get a better idea of what the post-employment experience is like for most people. To this end, Autor, Dorn and Hanson are conducting a new study that follows laid-off manufacturing workers over time, nationally, to get a fine-grained sense of their needs and potential to be re-employed.

“Trade may raise GDP,” Autor says, “but it does make some people worse off. Almost all of us share in the gains. We could readily assist the minority of citizens who bear a disproportionate share of the costs and still be better off in the aggregate.”

Rolling in the chip

newsoffice@mit.edu (Jennifer Chu, MIT News Office) — Fri, 24 Feb 2012 05:00:03 +0000

Cell rolling is a common mechanism cells use to navigate through the body. During inflammation, for example, the endothelial cells that line blood vessels present certain molecules that attract white blood cells just enough to divert them from the rest of the vessel’s cellular traffic. White blood cells then roll along the vessel wall, slowing down to help in the healing of inflamed areas.

Researchers at MIT and Brigham and Women’s Hospital have now designed a cell-sorting microchip that takes advantage of this natural cell-rolling mechanism. The device takes in mixtures of cells, which flow through tiny channels coated with sticky molecules. Cells with specific receptors bind weakly to these molecules, rolling away from the rest of the flow, and out into a separate receptacle.

The cell sorters, about the size of postage stamps, may be fabricated and stacked one on top of another to sift out many cells at once — an advantage for scientists who want to isolate large quantities of cells quickly. The device doesn’t necessarily require an external pump to push cells through the chip, which makes it a portable, affordable option for use in laboratories or clinics, where cell samples may be taken and sorted without specialized equipment.

“We’re working on a disposable device where you wouldn’t even need a syringe pump to drive the separation,” says Rohit Karnik, the d’Arbeloff Assistant Professor of Mechanical Engineering at MIT. “You could potentially buy a $5 or $10 kit and get the cells sorted without needing any kind of [additional] instrument.”

Karnik collaborated with postdoc Sung Young Choi of MIT and Jeffrey Karp, co-director of the Center for Regenerative Therapeutics at Brigham and Women’s. The team reported their findings in a paper posted online in the journal Lab on a Chip.

While current cell-sorting technologies separate large batches of cells quickly and efficiently, they have several limitations. Fluorescence-activated cell sorting, a widely used technique, requires lasers and voltage to sort cells based on their electric charge — a complex system requiring multiple parts. Researchers have also used fluorescent markers and magnetic beads that bind to desired cells, making them easy to spot and sift out. However, once collected, the cells need to be separated from the beads and markers — an added step that risks modifying the samples.

Going with the flow

Karnik’s team designed a compact cell sorter that requires no additional parts or steps. The team built upon their 2007 work with MIT’s Robert Langer and others, in which they first came up with the sorting-by-rolling principle. Since then, the group has been turning principle into practice, designing a working device to sort cells. The initial proof-of-principle design was relatively simple: Cells were injected into a single inlet, which gave way to a large chamber coated on one side with sticky, roll-inducing molecules. The incoming cells flowed through the chamber; the cells that bound to the molecules rolled to one side, then out to a collection chamber.

However, the researchers found that in order to allow target cells to first settle on the chamber’s surface, long channels were required, which would make the device too large. Instead, Choi came up with a surface pattern that causes cells to circulate within the chamber. The pattern comprises 10 parallel channels with 50 ridges and trenches, each ridge about 40 microns high. The researchers coated the ridges with P-selectin, a well-known molecule that promotes cell rolling. They then injected two kinds of leukemia cells: one with receptors for P-selectin, the other without.

They found that once injected, the cells entered the chamber and bounced across the top of the ridges, exiting the chip through an outlet. The cells with P-selectin receptors were “caught” by the sticky molecule and flipped into trenches that led to a separate receptacle. Through their experiments, the team successfully recovered the cells they intended to sift out with 96 percent purity.

Karnik says the device may be replicated and stacked to sort large batches of cells at relatively low cost. He and his colleagues are hoping to apply the device to sort other blood cells, as well as certain types of cancer cells for diagnostic applications and stem cells for therapeutic applications. To do that, the team is investigating molecules similar to P-selectin that bind weakly to such cells. In the future, Karnik envisions tailor-made cell rolling, designing molecules and surfaces that weakly adhere to any desired type of cell.

“It’s really the ability to design molecules to separate cells of interest that will be powerful,” Karnik says. “There’s no reason to believe it cannot be done, because nature has already done it.”

The device is a “smart design,” says Milica Radisic, an associate professor of biomedical engineering at the University of Toronto, who was not involved in this research. Radisic says because the device relies on hydrodynamics within the chamber, it doesn't require external equipment.

“The design is probably good as it is for separation of leukemia cell lines,” Radisic says. “The question is if it can be adopted for other receptor/ligand pairs.”

Making droplets drop faster

newsoffice@mit.edu (David L. Chandler, MIT News Office) — Thu, 23 Feb 2012 05:00:00 +0000

The condensation of water is crucial to the operation of most of the powerplants that provide our electricity — whether they are fueled by coal, natural gas or nuclear fuel. It is also the key to producing potable water from salty or brackish water. But there are still large gaps in the scientific understanding of exactly how water condenses on the surfaces used to turn steam back into water in a powerplant, or to condense water in an evaporation-based desalination plant.

New research by a team at MIT offers important new insights into how these droplets form, and ways to pattern the collecting surfaces at the nanoscale to encourage droplets to form more rapidly. These insights could enable a new generation of significantly more efficient powerplants and desalination plants, the researchers say.

The new results were published online this month in the journal ACS Nano, a publication of the American Chemical Society, in a paper by MIT mechanical engineering graduate student Nenad Miljkovic, postdoc Ryan Enright and associate professor Evelyn Wang.

Although analysis of condensation mechanisms is an old field, Miljkovic says, it has re-emerged in recent years with the rise of micro- and nanopatterning technologies that shape condensing surfaces to an unprecedented degree. The key property of surfaces that influences droplet-forming behavior is known as “wettability,” which determines whether droplets stand high on a surface like water drops on a hot griddle, or spread out quickly to form a thin film.

It’s a question that’s key to the operation of powerplants, where water is boiled using fossil fuel or the heat of nuclear fission; the resulting steam drives a turbine attached to a dynamo, producing electricity. After exiting the turbine, the steam needs to cool and condense back into liquid water, so it can return to the boiler and begin the process again. (That’s what goes on inside the giant cooling towers seen at powerplants.)

Typically, on a condensing surface, droplets gradually grow larger while adhering to the material through surface tension. Once they get so big that gravity overcomes the surface tension holding them in place, they rain down into a container below. But it turns out there are ways to get them to fall from the surface — and even to “jump” from the surface — at much smaller sizes, long before gravity takes over. That reduces the size of the removed droplets and makes the resulting transfer of heat much more efficient, Miljkovic says.

One mechanism is a surface pattern that encourages adjacent droplets to merge together. As they do so, energy is released, which “causes a recoil from the surface, and droplets will actually jump off,” Miljkovic says. That mechanism has been observed before, he notes, but the new work “adds a new chapter to the story. Few researchers have looked at the growth of the droplets prior to the jumping in detail.”

That’s important because even if the jumping effect allows droplets to leave the surface faster than they would otherwise, if their growth lags, you might actually reduce efficiency. In other words, it’s not just the size of the droplet when it gets released that matters, but also how fast it grows to that size.

“This has not been identified before,” Miljkovic says. And in many cases, the team found, “you think you’re getting enhanced heat transfer, but you’re actually getting worse heat transfer.”

In previous research, “heat transfer has not been explicitly measured,” he says, because it’s difficult to measure and the field of condensation with surface patterning is still fairly young. By incorporating measurements of droplet growth rates and heat transfer into their computer models, the MIT team was able to compare a variety of approaches to the surface patterning and find those that actually provided the most efficient transfer of heat.

One approach has been to create a forest of tiny pillars on the surface: Droplets tend to sit on top of the pillars while only locally wetting the surface rather than wetting the whole surface, minimizing the area of contact and facilitating easier release. But the exact sizes, spacing, width-to-height ratios and nanoscale roughness of the pillars can make a big difference in how well they work, the team found.

“We showed that our surfaces improved heat transfer up to 71 percent [compared to flat, non-wetting surfaces currently used only in high-efficiency condenser systems] if you tailor them properly,” Miljkovic says. With more work to explore variations in surface patterns, it should be possible to improve even further, he says.

The enhanced efficiency could also improve the rate of water production in plants that produce drinking water from seawater, or even in proposed new solar-power systems that rely on maximizing evaporator (solar collector) surface area and minimizing condenser (heat exchanger) surface area to increase the overall efficiency of solar-energy collection. A similar system could improve heat removal in computer chips, which is often based on internal evaporation and recondensation of a heat-transfer liquid through a device called a heat pipe.

Chuan-Hua Chen, an assistant professor of mechanical engineering and materials science at Duke University who was not involved in this work, says, “It is intriguing to see the coexistence of both sphere- and balloon-shaped condensate drops on the same structure. … Very little is known at the scales resolved by the environmental electron microscope used in this paper. Such findings will likely influence future research on anti-dew materials and … condensers.”

The next step in the research, underway now, is to extend the findings from the droplet experiments and computer modeling — and to find even more efficient configurations and ways of manufacturing them rapidly and inexpensively on an industrial scale, Miljkovic says.

This work was supported as part of the MIT S3TEC Center, an Energy Frontier Research Center funded by the U.S. Department of Energy.

Unique languages, universal patterns

newsoffice@mit.edu (Peter Dizikes, MIT News Office) — Thu, 23 Feb 2012 05:00:01 +0000

You don’t have to be a language maven to find the direct object in a basic English-language sentence. Just look next to the verb. Take a simple sentence: “I gave a book to Mary.” In this case the verb, “gave,” is quickly followed by “book,” the direct object. The sentence’s indirect object, “Mary,” lies farther away from the verb.

Things look quite different in Japanese, however, where direct objects pop up all over the place, and are signified by the presence of a language particle, -o. For example: The Japanese sentence, “Taroo-wa hon-o kinoo katta,” means “Taro bought a book yesterday.” But as written in Japanese, the word order is “Taro a book yesterday bought.” The word “hon-o,” or book, is the direct object with the particle, but it is not adjacent to “katta,” which is the verb “bought.”

To the chagrin of anyone who knows one of these languages but not the other, then, English and Japanese appear to be frustratingly different tongues governed by drastically different rules. And yet, under the surface, English and Japanese have deep similarities, as MIT linguist Shigeru Miyagawa argues in his new book, Case, Argument Structure, and Word Order, published this month in Routledge’s “Leading Linguists” series.

In turn, the similarities between English and Japanese underscore a larger point about human language, in Miyagawa’s view: All its varieties exist within a relatively structured framework. Languages are different, but not radically different. Dating to the 1950s, in fact, much of MIT’s linguistics program has aimed to identify the similar pathways that apparently unrelated languages take.

“There is this very interesting tension in language between diversity and uniformity,” says Miyagawa, the Kochi Prefecture-John Manjiro Professor of Japanese Language and Culture at MIT. “Human languages are diverse in stunning ways. Each one has some unique property that distinguishes it from 6,500 or maybe 7,000 other languages. But when you look as a linguist, you begin to notice that there are uniform properties shared by languages.”

English and Japanese may be different, but, as Miyagawa shows in his book, when it comes to denoting a direct object, they have performed a kind of grand historical flip-flop: Each has adopted rules that the other language has abandoned. In Old Japanese, in the eighth and ninth centuries, direct objects existed without the particle –o attached to them. In the sentence “Ware-wa imo omou,” or, “I think of my wife,” the word “imo,” or “wife,” lacks a particle. Instead, particles were used to mark points of emphasis: In Old Japanese, “kono tosi goro-o” means “during this year.”

By contrast, Old English, dating to the same time, used case markings (the equivalent of the –o particle) to specify that all direct objects take the accusative case, a rule derived from the structure of Latin. And unlike today, Old English word order was more flexible: Direct objects could appear in many sentence locations.

In this grammatical regard, at least, “Old Japanese is modern English,” Miyagawa says. “And Old English and Latin are modern Japanese. It is really quite remarkable.”

Compound interest

Indeed, English and Japanese effectively swapped rules during a time when they could not have influenced each other directly. But the nature of language is such that those changes “cannot just be anything,” as Miyagawa says. And the nature of linguistics is such that these parallels are not always obvious; many patterns emerge only after years of scholarly analysis.

Video: Lucy Lindsey and Melanie Gonick

Miyagawa’s book summarizes work he has done over three decades of research. He analyzes recent findings by other scholars in the area, engages with recent critiques of his work — “You have to be ready for that,” he says — and assesses the current state of knowledge in his own area of the field.

Recent work by linguist Yuko Yanagida at Tsukuba University in Japan seems to have strengthened Miyagawa’s suggestion that there are parallels between Old Japanese and modern English, and Old English and modern Japanese. Yanagida has shown that Old Japanese had an alternative way of denoting direct objects, which also surfaces in modern English. This is “compounding,” the joining of verbs and direct objects into new words.

Thus the Old Japanese sentence, “Sirokane-no su-wo hitobito tuki-sirohu,” has a compound verb at the end: “tuki-sirohu” literally means “poke each other.” (The sentence as a whole, literally “silver cover people poke each other,” is probably best rendered as, “People laugh amongst themselves at the silver cover.”)

This type of word formation occurs occasionally in English today. We join a verb and a direct object in words such as “bird-watching.” And linguists find the same habit elsewhere. In the Chukchee language of Russia, the sentence transliterated as “ytlygyn qaa-tym-ge” means “father deer-killed,” or “father killed a deer.”

Scholars who work on the evolution of language have welcomed the arrival of Miyagawa’s book. “There aren’t many languages in the world where we have historical records, only a handful where one can work on change [in language], and most of those languages are Indo-European,” says David Lightfoot, a linguist at Georgetown University, who has read the manuscript. “So it’s enormously valuable to have a very well-analyzed treatment of change in Japanese.”

Curiosity, but no gloriosity

Case, Argument Structure, and Word Order analyzes several other ways in which English and Japanese, despite superficial appearances to the contrary, actually converge. Japanese is regarded as having extensive word formation rules whereas English, at a glance, does not. On closer inspection, however, English does have a system governing word formation. We can turn “curious” into “curiosity,” for instance, but we don’t change “glorious” into “gloriosity.” Why not? Because English already has a relevant noun, “glory,” in place. That exact same rule — a “blocking effect,” as linguists say — holds in Japanese, too, as Miyagawa first asserted.

“In Japanese, we see this blocking effect in a very extensive manner,” he says. “But no one had ever really perceived this comparison before.”

John Whitman, a linguist at Cornell University who has read the book, thinks its impact “will really be lasting,” and increasingly so in Japan. “Linguists within one national tradition tend to think their language has always existed within the same basic ground plan. But Shigeru Miyagawa’s work shows that, no, Japanese 1,000 years ago was a very different thing.” As Whitman sees it, the “next step” for researchers “is to look in more detail at specific periods. He has a broad sweep over hundreds of years, and we would also like to look at 50-year slices.”

Beyond making the case for the similarities of Japanese and English, Miyagawa says, he hopes his work will reveal the excitement of discovery in linguistics — and the larger fascination in pondering language’s apparent universalism.

“It’s so exciting to see languages and know there is this diversity we should celebrate,” Miyagawa says. “And when we look closely we see they all work with the same mechanisms. That’s one thing that is so interesting about human language.”

A new twist on nanowires

newsoffice@mit.edu (David L. Chandler, MIT News Office) — Wed, 22 Feb 2012 05:00:01 +0000

Nanowires — microscopic fibers that can be “grown” in the lab — are a hot research topic today, with a variety of potential applications including light-emitting diodes (LEDs) and sensors. Now, a team of MIT researchers has found a way of precisely controlling the width and composition of these tiny strands as they grow, making it possible to grow complex structures that are optimally designed for particular applications.

The results are described in a new paper authored by MIT assistant professor of materials science and engineering Silvija Gradečak and her team, published in the journal Nano Letters.

Nanowires have been of great interest because structures with such tiny dimensions — typically just a few tens of nanometers, or billionths of a meter, in diameter — can have very different properties than the same materials have in their larger form. That’s in part because at such minuscule scales, quantum confinement effects — based on the behavior of electrons and phonons within the material — begin to play a significant role in the material’s behavior, which can affect how it conducts electricity and heat or interacts with light.

In addition, because nanowires have an especially large amount of surface area in relation to their volume, they are particularly well-suited for use as sensors, Gradečak says.

Her team was able to control and vary both the size and composition of individual wires as they grew. Nanowires are grown by using “seed” particles, metal nanoparticles that determine the size and composition of the nanowire. By adjusting the amount of gases used in growing the nanowires, Gradečak and her team were able to control the size and composition of the seed particles and, therefore, the nanowires as they grew. “We’re able to control both of these properties simultaneously,” she says. While the researchers carried out their nanowire-growth experiments with indium nitride and indium gallium nitride, they say the same technique could be applied to a variety of different materials.

These nanowires are far too small to see with the naked eye, but the team was able to observe them using electron microscopy, making adjustments to the growth process based on what they learned about the growth patterns. Using a process called electron tomography, they were able to reconstruct the three-dimensional shape of individual nanoscale wires. In a related study recently published in the journal Nanoscale, the team also used a unique electron-microscopy technique called cathodoluminescence to observe what wavelengths of light are emitted from different regions of individual nanowires.

Precisely structured nanowires could facilitate a new generation of semiconductor devices, Gradečak says. Such control of nanowire geometry and composition could enable devices with better functionality than conventional thin-film devices made of the same materials, she says.

One likely application of the materials developed by Gradečak and her team is in LED light bulbs, which have far greater durability and are more energy-efficient than other lighting alternatives. The most important colors of light to produce from LEDs are in the blue and ultraviolet range; zinc oxide and gallium nitride nanowires produced by the MIT group can potentially produce these colors very efficiently and at low cost, she says.

While LED light bulbs are available today, they are relatively expensive. “For everyday applications, the high cost is a barrier,” Gradečak says. One big advantage of this new approach is that it could enable the use of much less expensive substrate materials — a major part of the cost of such devices, which today typically use sapphire or silicon carbide substrates. The nanowire devices have the potential to be more efficient as well, she says.

Such nanowires could also find applications in solar-energy collectors for lower-cost solar panels. Being able to control the shape and composition of the wires as they grow could make it possible to produce very efficient collectors: The individual wires form defect-free single crystals, reducing the energy lost due to flaws in the structure of conventional solar cells. And by controlling the exact dimensions of the nanowires, it’s possible to control which wavelengths of light they are “tuned” to, either for producing light in an LED or for collecting light in a solar panel.

Complex structures made of nanowires with varying diameters could also be useful in new thermoelectric devices to capture waste heat and turn it into useful electric power. By varying the composition and diameter of the wires along their length, it’s possible to produce wires that conduct electricity well but heat poorly — a combination that is hard to achieve in most materials, but is key to efficient thermoelectric generating systems.

The nanowires can be produced using tools already in use by the semiconductor industry, so the devices should be relatively easy to gear up for mass production, the team says.

Zhong Lin Wang, the Regents’ Professor and Hightower Chair in Materials Science and Engineering at the Georgia Institute of Technology, says that being able to control the structure and composition of nanowires is “vitally important for controlling their nanoscale properties. The fine-tuning in the growth behavior” of these materials “opens the possibility for fabricating new optoelectronic devices that are likely to have superior performance.”

In addition to Gradečak, the Nano Letters paper was co-authored by MIT graduate student Sam Crawford, Sung Keun Lim PhD ’11 and researcher Georg Haberfehlner of the research and technology organization CEA-Leti in Grenoble, France. The Nanoscale paper was co-authored by MIT graduate student Xiang Zhou, Megan Brewster PhD ’11 and postdoc Ming-Yen Lu. The work was supported by the MIT Center for Excitonics, the U.S. Department of Energy, the MIT-France MISTI program and the National Science Foundation.

3 Questions: Adam Berinsky on the unpredictable GOP campaign

newsoffice@mit.edu (Peter Dizikes, MIT News Office) — Wed, 22 Feb 2012 15:42:00 +0000

The 2012 Republican primary season has featured many sharp swings in the polls, some of which have caught seasoned political professionals by surprise. MIT News spoke with Associate Professor of Political Science Adam Berinsky, an expert in public opinion and director of MIT’s Political Experiments Research Lab (PERL), about the unpredictable nature of the ongoing campaign.

Q. By some counts, at least 10 potential candidates have topped various national polls covering the Republican presidential nomination. When was the last time we had a presidential primary season like this one, with so much uncertainty and upheaval in the polls?

A. In 2008, we saw something a little similar when John McCain was written off early, then came from way back to be the eventual Republican nominee, but much of that happened before the formal primary season took place. And this time, before the primaries, we saw Herman Cain surging and then trailing, Rick Perry surging and then trailing, and so on. What’s unusual about this primary season is that we’re watching it unfold in real time, during the election campaign itself.

To see this kind of unsettled race emerge in the heat of the campaign, I think you have to go back to 1976, when Jimmy Carter got the Democratic nomination. It used to be this kind of dynamic wasn’t strange, but over the last 30 or 40 years, there’s been a compression of the primary calendar, in the sense that the race has often been over by Super Tuesday in March, except for the Obama-Clinton race in 2008. That’s one thing that makes this seem so unusual.

The other element is that it hasn’t been a two-person race. We’re used to seeing two politicians compete, but here, it’s clearly Mitt Romney and the not-Romney candidate, whether it be Newt Gingrich, Rick Santorum or a candidate who could emerge at a brokered convention, as there has been talk of that again. Santorum got written off, then he won in Iowa, it seems, but then Gingrich took that role, and now Gingrich has faded.

Q. What accounts for this dynamic?

A. I think the big reason is that there wasn’t a clear front-runner coming in. There was a belief that Romney would eventually be the nominee, and I think he still will be, but what determines if a candidate is atop the polls is partially the enthusiasm of the base, and also if party elites — who help manage and determine where the money goes from big donors — are coalescing around a candidate. Also, I don’t think you can ascribe an organizational role to the Tea Party, but there’s no clear candidate who supports that point of view, which is another factor we continue to see in the polls.

Q. It seems that part of the argument for Romney is that he would be the most electable in November. But how much does the electability argument really help candidates, or can it be a kind of house of cards that eventually collapses?

A. There are two important decision points, one being the collective decisions that ordinary voters make. … We know from political science research that what matters a lot in primaries is momentum. Gingrich had it for a while, Santorum has it now, and momentum matters because it’s a lot easier for the candidate who is perceived as surging to raise money. I just read that in January, Romney spent more money than he took in.

Second, it’s partially about other politicians jumping on board with endorsements. There is a general sense that Romney is endorsed by more party insiders than other candidates, but he hasn’t lined them up at quite the rate he would have liked. So it’s not just the decisions of voters, but also party insiders, and electability is a large concern for them. I’m surprised that we’re seeing Santorum again. I thought that once Romney blitzed Gingrich with money in Florida, the race was almost done, but clearly there is a sense of unease with Romney as a candidate that’s still coming through.

A faster way to catch cells

newsoffice@mit.edu (Anne Trafton, MIT News Office) — Wed, 22 Feb 2012 05:00:00 +0000

Separating complex mixtures of cells, such as those found in a blood sample, can offer valuable information for diagnosing and treating disease. However, it may be necessary to search through billions of other cells to collect rare cells such as tumor cells, stem cells or fetal cells. “You’re basically looking for a needle in a haystack,” says Sukant Mittal, a graduate student in the Harvard-MIT Division of Health Sciences and Technology (HST).

Mittal and his colleagues at MIT and Massachusetts General Hospital (MGH) have now demonstrated a new microfluidic device that can isolate target cells much faster than existing devices. Such technology could be used in applications such as point-of-care diagnostics and personalized medicine.

The researchers describe their results in the Feb. 21 issue of Biophysical Journal. Other authors of the paper are Ian Wong, a postdoc at MGH and Harvard Medical School (HMS), MIT chemical engineering professor William Deen, and Mehmet Toner, a professor of biomedical engineering at MGH, HMS and HST.

Researchers have used a number of techniques to sort cells based on differences in size, density or electrical properties. However, since the physical characteristics of cells can vary significantly, these techniques risk separating cells incorrectly, leading to an erroneous diagnosis. A more specific way to isolate cells is to use antibodies that latch on to distinctive molecules displayed on the surfaces of the target cells.

However, this selective approach only works if the target cells come into contact with the antibodies designed to capture them. This is unlikely to happen when the cells are moving at relatively high speeds.

“Imagine you’re standing on a bridge over a river, and you throw a message in a bottle out in the middle,” Wong says. “If the river is moving really slowly, you could imagine that eventually the bottle will drift over to the riverbank and somebody can grab it. But if the river is flowing too quickly, then the bottle is swept downstream without ever approaching the sides.”

That’s the problem the team needed to solve, Wong says: “Can we steer the bottle toward the riverbank so that it can get caught?” To achieve that, the MIT and MGH researchers designed their device to guide the fluid toward the bottom of the channel as it flows, bringing more of the cells in contact with the antibodies. Key to their new design is the use of a soft membrane with nanoscale pores, which separates two adjacent microchannels.

Cells enter one channel only, and as they flow through the channel, the fluid is rapidly drawn to the porous divider, bringing the cells with it. Fluid can pass into the other channel, but the cells cannot. Once they reach the surface, they start rolling — slowly enough that target cells have time to attach to the antibodies and get captured, but fast enough to keep the other cells moving. Such rolling behavior is similar to how white blood cells or stem cells selectively “home in” to sites of infection and injury in the body.

Shashi Murthy, an associate professor of chemical engineering at Northeastern University, says the device is simple but very well-designed. “The field of microfluidics is very largely done by experimental trial and error,” says Murthy, who was not involved in this research. “One seldom sees as in-depth an analysis, and one so well-grounded in theory.”

One potential application for these devices is to isolate cancer cells from patient blood samples. Toner’s group has previously shown that the number of circulating tumor cells in the bloodstream correlates with the clinical response to treatment in a given patient, suggesting the potential for personalized medicine for cancer patients.

“Considerable validation and testing will be necessary before this early-stage device can be deployed in the clinic,” Toner says. “Nevertheless, this novel approach may enable exciting diagnostic and therapeutic opportunities that are not feasible using existing technologies.”

Toying with biological systems

newsoffice@mit.edu (Anne Trafton, MIT News Office) — Tue, 21 Feb 2012 05:00:00 +0000

Bacteria don’t normally take photographs. Nor do they attack tumor cells or produce chemicals. But with some help from biological engineer Chris Voigt, they can do all that and more.

Voigt, who joined MIT’s faculty in July as an associate professor of biological engineering, likes to tinker with bacteria and other microbes to get them to perform myriad useful tasks that nature never intended — an approach known as synthetic biology.

For example, to develop their “bacterial camera,” Voigt and his students inserted a light-detecting sensor from an alga into the bacterium E. coli, coupled with a gene that causes the bacterium to make a black pigment. A sheet of these bacteria acts as the “film,” and when a stencil is laid over the film and light shone upon it, an image of the stencil forms on the sheet of bacteria.

Likewise, his tumor-targeting E. coli incorporate genes from other bacteria that detect low oxygen levels and high cell density, both conditions often found in tumors. Voigt, who had been on the faculty of the University of California at San Francisco before coming to MIT, then linked those genes with a cell circuit that triggers production of a protein called invasin that enables E. coli to invade mammalian cells.

Despite all Voigt has accomplished in synthetic biology, he got into the field almost by accident. As an undergraduate at the University of Michigan, he majored in chemical engineering, focusing mainly on theoretical studies of reaction mechanics and catalysis. But one day, he was in the chemistry building picking up an exam, and a professor who saw him standing near his door invited him in, thinking he was a student who had applied for a summer job. (That student never showed up.)

“I happened to be standing there, so we started talking,” Voigt recalls. “He was doing protein folding and because of the work that I had been doing on catalysis, I was able to converse with him on some of the theoretical underpinnings of protein folding, and that’s how I got the job. That’s how I got into biology, but I had never really taken any classes in it.”

In graduate school at the California Institute of Technology, he started to work on directed evolution of proteins — specifically, developing a computer program that would identify locations in a protein where mutations would produce a better protein. During a postdoctoral stint at the University of California at Berkeley, he became interested in synthetic biology, which was then just emerging as a new field, based on the idea that novel biological circuits could be assembled from a set of standardized parts — in this case, genes.

At Berkeley, Voigt worked on extracting genetic circuits from a bacterium called B. subtilis and reconstituting them in E. coli, so they could be studied in isolation. Upon joining the faculty at UCSF, he started working on building simple circuits such as a sensor that would respond to a specific stimulus, which led to his bacterial camera.

Complex circuits

Voigt, who is co-directing MIT’s new Synthetic Biology Center, is now working on building larger, highly interconnected systems that include sensors and circuits that can respond to sensors’ input.

“Right now we’re integrating components into an individual cell. One of the problems with that is getting all of the pieces to interact with each other,” he says. Another challenge is preventing pieces that are not supposed to interact from doing so.

“If you want to create a system with 50 circuits that are all working together as part of a computation that the cell is running, then you need each one of those individual circuits to not interfere with all the others. So it becomes an exponentially challenging problem to build each additional new circuit and show that it doesn’t interact with all the others,” Voigt says.

Such complex circuits could form the basis of microbes that can regulate their own fermentation processes — for example, the yeast that ferment biomass into ethanol, Voigt says. Ethanol fermentation produces acetate, which is toxic to yeast, as a byproduct. Therefore, fermentation vats must be equipped with sensors that detect dangerous acetate levels and take corrective action, such as slowing down delivery of the microbes’ food supply (glucose).

Using synthetic biology, it’s possible that this monitoring process could be transferred into the cells themselves. Yeast cells would sense the elevated acetate levels and shut off their own glucose transporters until acetate levels go down again.

Voigt says he came to MIT in part because of its focus on biological engineering as a way to impact a variety of fields — not just medicine but also agriculture, energy, industrial chemistry, environmental cleanup and materials. To that end, the new Synthetic Biology Center has recruited researchers from a wide range of backgrounds.

“Our hope is that this place really brings together different people with the same objectives who can think innovatively about the types of systems we can design,” Voigt says.

An element that's rare on Earth is found far, far away

newsoffice@mit.edu (Jennifer Chu, MIT News Office) — Fri, 17 Feb 2012 05:00:00 +0000

Nearly 13.7 billion years ago, the universe was made of only hydrogen, helium and traces of lithium — byproducts of the Big Bang. Some 300 million years later, the very first stars emerged, creating additional chemical elements throughout the universe. Since then, giant stellar explosions, or supernovas, have given rise to carbon, oxygen, iron and the rest of the 94 naturally occurring elements of the periodic table.

Today, stars and planetary bodies bear traces of these elements, having formed from the gas enriched by these supernovas over time. For the past 50 years, scientists have been analyzing stars of various ages, looking to chart the evolution of chemical elements in the universe and to identify the astrophysical phenomena that created them.

Now a team of researchers from institutions including MIT has detected the element tellurium for the first time in three ancient stars. The researchers found traces of this brittle, semiconducting element — which is very rare on Earth — in stars that are nearly 12 billion years old. The finding supports the theory that tellurium, along with even heavier elements in the periodic table, likely originated from a very rare type of supernova during a rapid process of nuclear fusion. The researchers published their findings online in Astrophysical Journal Letters.

"We want to understand the evolution of tellurium — and by extension any other element — from the Big Bang to today," says Anna Frebel, an assistant professor of astrophysics at MIT and a co-author on the paper. "Here on Earth, everything's made from carbon and various other elements, and we want to understand how tellurium on Earth came about."

'In the halo of the Milky Way,' a rare element found

The team analyzed the chemical composition of three bright stars located a few thousand light-years away, "in the halo of the Milky Way," Frebel says. The researchers looked at data obtained from the Hubble Space Telescope's spectrograph, an instrument that splits light from a star into a spectrum of wavelengths. If an element is present in a star, the atoms of that element absorb starlight at specific wavelengths; scientists can observe this absorption as dips in the spectrograph's data.

Frebel and her colleagues detected dips in the ultraviolet region of the spectrum — at a wavelength that matched tellurium's natural light absorption — providing evidence that the rare Earth element does indeed exist in space, and was likely created more than 12 billion years ago, at the time when all three stars formed.

The researchers also compared the abundance of tellurium to that of other heavy elements such as barium and strontium, finding that the ratio of elements was the same in all three stars. Frebel says the matching ratios support a theory of chemical-element synthesis: namely, that a rare type of supernova may have created the heavier elements in the bottom half of the periodic table, including tellurium.

No ordinary supernova

According to theoretical predictions, elements heavier than iron may have formed as part of the collapsing core of a supernova, when atomic nuclei collided with huge amounts of neutrons in a nuclear fusion process. For 50 years, astronomers and nuclear physicists have modeled this rapid process, named the r-process, in order to unravel the cosmic history of the elements.

Frebel's team found that the ratios of heavy elements observed in the three stars matched the ratios predicted by these theoretical models. The findings, she says, confirm the theory that heavier elements likely formed from a rare, extremely rapid supernova.

"You can make iron and nickel in any ordinary supernova, anywhere in the universe," Frebel says. "But these heavy elements seem to only be made in specialized supernovas. Adding more elements to the observed elemental patterns will help us understand the astrophysical and environmental conditions needed for this process to operate."

Jennifer Johnson, an associate professor of astronomy at Ohio State University, says tellurium has been a "tough" element to detect, since it absorbs light in the ultraviolet spectrum, which is impossible for ground-based telescopes to spot. The team's findings, she says, are a first step in identifying some of the most elusive elements in the universe.

"If you look at the periodic table, tellurium is right in the middle of these elements that are hard for us to measure," Johnson says. "If we need to understand how [the r-process] works in the universe, we really have to measure this part of the periodic table. It's really cool that they got this element in this sea of unknown-ness."

Frebel is continuing the search for heavy elements in space. For example, selenium — which is similar to tellurium — has yet to be detected in the universe. Tin, Frebel says, is also a difficult element to spot, as are many elements along the same row as tellurium in the periodic table.

"There are still quite a few holes," Frebel says. "Every now and then, we can add an element, and it adds another data point that makes our work easier."

Successful human tests for first wirelessly controlled drug-delivery chip

newsoffice@mit.edu (Anne Trafton, MIT News Office) — Thu, 16 Feb 2012 19:00:00 +0000

About 15 years ago, MIT professors Robert Langer and Michael Cima had the idea to develop a programmable, wirelessly controlled microchip that would deliver drugs after implantation in a patient’s body. This week, the MIT researchers and scientists from MicroCHIPS Inc. reported that they have successfully used such a chip to administer daily doses of an osteoporosis drug normally given by injection.

The results, published in the Feb. 16 online edition of Science Translational Medicine, represent the first successful test of such a device and could help usher in a new era of telemedicine — delivering health care over a distance, Langer says.

“You could literally have a pharmacy on a chip,” says Langer, the David H. Koch Institute Professor at MIT. “You can do remote control delivery, you can do pulsatile drug delivery, and you can deliver multiple drugs.”

In the new study, funded and overseen by MicroCHIPS, scientists used the programmable implants to deliver an osteoporosis drug called teriparatide to seven women aged 65 to 70. The study found that the device delivered dosages comparable to injections, and there were no adverse side effects.

These programmable chips could dramatically change treatment not only for osteoporosis, but also for many other diseases, including cancer and multiple sclerosis. “Patients with chronic diseases, regular pain-management needs or other conditions that require frequent or daily injections could benefit from this technology,” says Robert Farra, president and chief operating officer at MicroCHIPS and lead author of the paper.

“Compliance is very important in a lot of drug regimens, and it can be very difficult to get patients to accept a drug regimen where they have to give themselves injections,” says Cima, the David H. Koch Professor of Engineering at MIT. “This avoids the compliance issue completely, and points to a future where you have fully automated drug regimens.”

Achieving precision

The MIT research team started working on the implantable chip in the mid-1990s. John Santini, then a University of Michigan undergraduate visiting MIT, took it on as a summer project under the direction of Cima and Langer. Santini, who later returned to MIT as a graduate student to continue the project, is also an author of the new paper.

In 1999, the MIT team published its initial findings in Nature, and MicroCHIPS was founded and licensed the microchip technology from MIT. The company refined the chips, including adding a hermetic seal and a release system that works reliably in living tissue. Teriparatide is a polypeptide and therefore much less chemically stable than small-molecule drugs, so sealing it hermetically to preserve it was an important achievement, Langer says.

The human clinical trial began in Denmark in January 2011. Chips were implanted during a 30-minute procedure at a doctor’s office using local anesthetic, and remained in the patients for four months. The implants proved safe, and patients reported they often forgot they even had the implant, Cima says.

Professors Robert Langer, right, and Michael Cima speak in Cima's lab at the Koch Institute.
Photo: M. Scott Brauer

Chips used in the study stored 20 doses of teriparatide, individually sealed in tiny reservoirs about the size of a pinprick. The reservoirs are capped with a thin layer of platinum and titanium that melts when a small electrical current is applied, releasing the drug inside. MicroCHIPS is now working on developing implants that can carry hundreds of drug doses per chip.

Because the chips are programmable, dosages can be scheduled in advance or triggered remotely by radio communication over a special frequency called Medical Implant Communication Service (MICS). Current versions work over a distance of a few inches, but researchers plan to extend that range.

Consistent results

In the Science Translational Medicine study, the researchers measured bone formation in osteoporosis patients with the implants, and found that it was similar to that seen in patients receiving daily injections of teriparatide. Another notable result is that the dosages given by implant had less variation than those given by injection.

Henry Brem, professor of neurosurgery, ophthalmology, oncology and biological engineering at Johns Hopkins University School of Medicine, called the results “stunning.”

“It’s very rare to find a paper that is really a breakthrough in technology,” says Brem, who was not part of the research team. “It fulfills the promise of polymer drug delivery and the incredible sophistication of microchip capabilities.”

Once a version of the implant that can carry a larger number of doses is ready, MicroCHIPS plans to seek approval for further clinical trials, Farra says. The company has also developed a sensor that can monitor glucose levels. Eventually such sensors could be combined with chips that contain drug reservoirs, creating a chip that can adapt drug treatments in response to the patient’s condition.