In Brief: February 19

Feb 19 2019 | By Holly Evarts | Capponi Photo Credit: Timothy Lee Photographers | Micrograph Image Credit: Allan Brooks | Solar Flare Photo Credit: NASA | Plasmon Image Credit: B.Y. Jiang/UCSD

Recent publications from faculty in Industrial Engineering and Operations Research, Chemical Engineering, and Applied Physics and Applied Mathematics

Who Likes Large Uninformed Orders? Anyone?

A new paper by Agostino Capponi, assistant professor of industrial engineering and operations research (IEOR), and Hongzhong Zhang, IEOR associate research scientist, with Albert Menkveld, professor of finance at VU University Amsterdam, was published on SSRN, the Social Science Research Network. The study focuses on market quality in the presence of “whale,” or uninformed, orders and provides a tractable model for who benefits—market makers, large sellers or investors—when large institutional sellers trade to gain liquidity.

Increased intermediation has made some investors "too large" for their markets. If such investors need to sell quickly, then they cannot reach buyers who arrive later. Market makers then supply liquidity by taking on inventory to sell to future buyers, splitting their order optimally through time, often constrained by a time limit.

The researchers set out to answer the question of how does algorithmic execution of an uninformed order, i.e. an order driven by liquidity needs and not by informational advantage, affects other market participants, such as market makers and other investors.

They set up a model that features strategic liquidity supply and demand. In a game setting, a large institutional seller has to trade before a particular time limit. He will do so optimally, anticipating the response of high-frequency market makers. The market makers supply liquidity optimally by setting bid-and-ask prices to earn off the spread, while minimizing their cost of holding inventory. Other investors arrive randomly and trade at the quotes they observe when entering the market.

The researchers show that in nearly all market scenarios, all the parties benefit. Moreover, their analysis reveals that sharing knowledge of how long the large seller's order will last benefits all market participants. Stealth trading by the seller turns out to be costly, privately and socially, because market makers experience additional cost not knowing when execution ends. If they do know, then the price pressure might subside before execution ends, rationalizing such patterns observed in the data.

The model shows how large liquidity demands affect institutional investors and intermediaries. It allows policy makers to protect retail investors operating in markets with an increasing number of “whale” orders. Because the model is tractable, it yields precise analytic results on such sunshine trading as compared to stealth trading. The ability to relate market outcomes to model parameters allows institutional investors to understand why the market responds the way it does to their orders, and enables those who offer optimal-execution services such as Goldman Sachs, BlackRock, or ITG to adjust the sell rate of large orders to changing market conditions.

Capponi notes that analysis of how the execution affects the market at large will allow policy makers to regulate markets that increasingly feature whale orders. Given regulators' mandate to protect retail investors, they will be able to use the model to learn how large liquidity demand affects retail investors and to determine if they benefit or suffer from large price pressures due to sudden strong liquidity demand.
 

Agostino Capponi

Major Advance in Microrobotics

Chemical Engineering Associate Professor Kyle Bishop had a paper recently published in Nature Communications on catalytic micromotors that use chemical fuels to “swim”’ through fluid surroundings. Being able to control these micromotors, essentially mobile robots capable of transporting micrometer-sized components, would be a major step forward in the field of microrobotics. This is especially true in biology where micromotors could be used for drug delivery, disease detection, and microsurgeries.

In order to program swimmers that have functional dynamics, such as corkscrew motions that enhance navigation through porous media, researchers must understand the propulsion mechanism and its dependence on particle shape.

Bishop’s study—“Shape-directed rotation of homogeneous micromotors via catalytic self-electrophoresis”—makes two complementary advances. First, it shows—for the first time—that self-propelled motions of catalytic micromotors can be designed by controlling the geometric shape of a catalytic particle. Second, in explaining the origins of shape-directed motions, it solves a longstanding mystery regarding the propulsion of platinum-catalyzed motors, one of the most well-studied systems.

The paper demonstrates that platinum micromotors move due to self-electrophoresis, the process by which chemical reactions at the particle surface induce local electric fields that then propel fluid motions. Bishop’s findings show that shape controls the rate at which chemical reactants are delivered to the particle surface and thereby determines the location of anodic and cathodic surface regions that induce local electric fields. The team developed a mathematical model of this mechanism to explain their experimental observations on the direction and speed of particle motions as a function of particle shape.
 

Scanning electron micrograph of platinum microparticles.

Debunking the Solar-Cycle/North Atlantic Winter Weather Connection

A new Nature Geoscience study from Applied Physics and Applied Mathematics (APAM) and Lamont Doherty researchers has upended a commonly accepted theory that variations in the energy emitted by the sun affect weather patterns in the North Atlantic and the likelihood of storms and floods over Europe. The North Atlantic Oscillation (NAO) is considered a key driver of winter weather patterns over the northern hemisphere. A positive NAO is linked with more windstorms, and mild and wet winters in Europe. A negative NAO indicates snowy and cold winters in Europe.

In recent years, published research has claimed the existence of a correlation between the NAO and the 11-year solar cycle, a periodic change in the sun’s activity. That claim has held that the connection between the NAO and the solar cycle is strong enough to inform predictions of the NAO as much as a decade in advance, which would in turn, enable scientists to predict winter weather patterns as many as ten years in advance.

APAM research scientist Gabriel Chiodo and PhD student Jessica Oehrlein, together with Professor Lorenzo Polvani (also a professor of Earth and Environmental Sciences)  showed in the new study that there is no robust connection between the solar cycle and the NAO. Their paper essentially debunks what was considered a “demonstrated link” between the 11-year sun cycle and winter weather over the northern hemisphere and found that it is, for the most part, a coincidental alignment. Using sophisticated computer modeling and extended observations, they demonstrated that before 1960 evidence of any correlation simply vanishes.

The implications of these new findings are substantial both for Europe and for science. The older correlation claim, if substantiated, would have meant great advantages to societies in the northern hemisphere, giving enough warning of periods of intense storms and flooding to inform community planning efforts. But this new finding—“Insignificant influence of the 11-year solar cycle on the North Atlantic Oscillation”—will be important for climate research into the future, as it implies that the causes for decadal weather changes over Europe lies elsewhere, not in the solar variations.
 

This image shows the bright light of a solar flare on the left side of the sun and an eruption of solar material shooting through the sun’s atmosphere, called a prominence eruption. Shortly thereafter, this same region of the sun sent a coronal mass ejection out into space.

First “Quantum Photonic Crystal” in Graphene

Sai Swaroop Sunku, a PhD student in the department of applied physics and applied mathematics, is the lead author of a study, “Photonic crystal for nano-light in moiré graphene superlattices,” conducted in the lab of Physics Professor Dmitri Basov. Published in Science, the paper examines surface plasmons, or nano-light, which are hybrids of light (photons) and electrons in graphene.

Since the 1980s, scientists have been trying to manipulate plasmons to build smaller and faster interfaces between the transistors in computers and the fiber optic cables that carry data. Plasmons could also be used to transfer data on a chip at low power with little energy loss.

Studies done over the past five years, including in the labs of Basov, Mechanical Engineering Professor Jim Hone, and others at Columbia, have shown that plasmons in graphene show great promise for such applications because they are much smaller than actual light waves and can travel long distances. But even though the small size of the graphene plasmons is what scientists want for applications such as data transfer, it has been extremely challenging to make structures small enough to manipulate the nanolight.

Sunku and his team are the first to make a “quantum photonic crystal” in graphene, a structure that can control and direct the nano-light using quantum effects. More importantly, their approach avoids the complications of making tiny structures.

The researchers discovered that they could enhance the plasmonic properties of graphene by introducing and then rotating a second graphene layer so that there is a slight angle between the atomic registry of the layers. Twisting these layers forms a moiré pattern with confined conducting channels that reflect the plasmons in a specific, tunable way. The size of this moiré pattern can be made as small as necessary simply by changing the twist angle between the layers. Thus, by controlling the structure of the pattern, the team was able to create a pathway for a nanophotonic platform that can be used for data transfer on a chip and perhaps even computation.

The team is now running more experiments at low temperatures, trying to observe novel effects such as one-dimensional plasmons that are also predicted to occur in this system.

Schematic illustration of the plasmon travelling in graphene photonic crystal. The plasmon is excited by the laser (red) and studied by the tip (grey).

Stay up-to-date with the Columbia Engineering newsletter

* indicates required