Altered States and Logic Gates

NEWS for June 2008www.jqi.umd.edu

JQ I Joint Quantum Institute

Quantum-Entangled Images: Mix and Squeeze

Whether it’s quantum-mechanical or classical, information pro-cessing must consist of stupendous numbers of very simple ac-tions performed by “logic gates” that obey if-then, and-or rules. For example: If X has some value, then change Y to the same value. Or change Y to the opposite value. And so forth. In con-ventional computers, transistors perform those tasks and pass the results along to other transistors through electronic connec-tions, building up elaborate logic chains seriatim.

In a quantum computer, however, that kind of direct contact be-tween components could easily destroy fragile quantum states. So scientists are searching for ways to create uniquely quantum logic gates in which one object can switch the state of another object with minimal destructive interaction.

continued, page 6 Trapped atoms named for J. Rydberg (1854-1919) -- the logical gate choice?

Classical and Quantum Cir-cuits -- TogetherAt Last! Page 2

INSIDE

Using a convenient and flex-ible method for creating twin light beams, JQI researchers from the National Institute of Standards and Technology have produced “quantum im-ages,” pairs of information-rich visual patterns whose features are “entangled,” or inextricably linked by the laws of quantum physics.

In addition to promising bet-ter detection of faint objects and improved amplifica-tion and positioning of light beams, the researchers’ tech-nique for producing quantum

continued, page 4

Recent PhDs and PostdocPlacements, Page 9

1

How to Talk to Your Quantum ComputerNo matter how exotic the in-nards of tomorrow’s quantum computers may be, users will still have to communicate with them using classical electronic circuits.

That deceptively simple state-ment conceals a host of very complicated problems. Every physical system under consid-eration as a quantum data bit (“qubit”) -- whether it consists of neutral atoms, ions, elec-tron/nuclear spins, quantum dots or Josephson junctions -- must be able to exist, at least temporarily, in a delicate quantum-mechanical condi-tion called “superposition,” in which the bit has values of 0, 1 or both at once. Virtually any contact with the outside world destroys that condition.

But in order to operate the computer, research-ers must be able to set and control the qubits, manipulate logic gates, correct errors, remove “noise” in processing, and record the computed values. All those functions will have to be con-trolled using some digital electronic technology.

The problem is notably acute for qubits construct-ed from Josephson junctions -- superconducting devices which are under intense study by several JQI Fellows including Chris Lobb, Fred Wellstood, Bob Anderson and Alex Dragt. Josephson junc-tions are extremely attractive candidates for qubits because the technology is well character-ized and readily scalable, and junction response times are in the sub-nanosecond range (one bil-lionth of a second).

That’s important because the JQI junctions’ coher-ence time (duration of sustained superposition before the system collapses into “decoherence”) is currently 50 to 100 nanoseconds. That’s pretty brief. But a Josephson junction qubit can knock out dozens of operations in that period.

However, Josephson junction qubits operate in ultra-cold conditions -- around 30 milliKelvin (thousandths of a degree above absolute zero) for the configurations that JQI is investigating. Even the chilliest of conventional electronic compo-nents would drastically overheat the system.

Moreover, no existing classical control/read-out system can react fast enough to take full advan-tage of the junctions’ blazing speed. As a result, physicists have had to artificially slow their sys-tems down in order to conduct experiments.

Obviously, some radically new technology is needed. And that’s where Anna Herr comes in.

From the University of Maryland, the JQI Associ-ate Research Scientist coordinates an interna-tional collaboration that is designing and testing control and read-out circuits that are cold, fast and very sensitively coupled to Josephson junc-tions. Recently, the consortium produced the first working device in which classical and quantum circuits are integrated on the same chip.

continued on next page

Associate Research Scientist Anna Yurievna Herr shows a tunable filter chip she and collaborators will test in tandem with Josephson junction qubits.

2

3

Temperature was the main initial concern. Herr and colleagues overcame it by improving on an existing technology called Rapid Single Flux Quantum (RSFQ) logic, a cryogenic digital system that uses Josephson junctions instead of transistors to switch signals, and represents information as patterns of electrical pulses that travel on superconducting transmission lines. One major advantage of RSFQ is that its pulses are single quanta -- indivisible units of magnetic flux -- and thus cannot change form as they move.

The researchers designed a two-stage device that has a com-bined access and operation time of 0.5 nanosecond. (Allowing a “clock speed’” of about 2 GHz.)The first stage consists of an RSFQ circuit that operates down to 30 mK right next to the Jo-sephson qubit. (See illustration above.) It doesn’t touch the qubit. The circuit and the qubit com-municate by electromagnetic induction, acting as two coils in the same tiny transformer.

This arrangement is connected by transmission lines to a second stage that runs at a compara-tively balmy 4.2 K, the temperature of liquid helium. The second stage is the main process-

ing unit and controls the pulses, which are used both to send signals for error correction and to query the qubit and register its state. The “0” and “1” conditions of the qubit are manifested in the clockwise or counterclockwise directions in which a current travels through the Josephson junction. The RSFQ circuit detects that direction in its inductively coupled flux comparator (see diagram below) and sends the result back to the processor as a digital signal.

The RSFQ system can rapidly read out as many as 20 separate qubits thanks to a fortunate quirk of quantum physics: As a single pulse quantum ( called a “fluxon”) passes near a qubit, it changes speed depending on the qubit’s state. So if it passes a long row of qubits, the fluxon’s propagation time will be a sensitive measure of the collective

states of all the qubits it passed.

The next steps for Herr and the consortium involve scaling up the system, creating a work-ing prototype with multiple qubits, and finding ways to transfer the state of one qubit to an-other one -- a process essential for control and error correction. Whatever the final design, one thing is certain: It will be very cool.

How to Talk to Your Quantum Computer, continued

4.2 K

Pulse control

50 mK

SFQ pulses

Phase qubit

RSFQ

Diagram of one configuration devised by an international consortium to connect digital control and read-out circuits to a Josephson junction phase qubit. JTL is Josephson transmission line.

Quantum-Entangled Images, continued from page 1

4

images—unprecedented in their simplicity, ver-satility, and efficiency—may someday be useful for storing patterns of data in quantum comput-ers and transmitting large amounts of highly secure encrypted information. The research team, led by JQI’s Paul Lett, describes the work in the June 27 issue of Science.

“Images have always been a preferred method of communication because they carry so much information in their details,” says Vincent Boyer, lead author of the new paper. “Up to now, however, cameras and other optical detectors have largely ignored a lot of useful information in images. By taking advantage of the quantum-mechanical aspects of images, we can improve ap-plications ranging from taking pictures of hard-to-see objects to storing data in futuristic quantum computers.”

Conventional photographic films or digital camera sensors only record the color and intensity of a light wave striking their surfaces. A hologram additionally records a light wave’s “phase”—the precise locations of the crests and valleys in the wave. How-ever, much more happens in a light wave. Even the most stable laser beam brightens and dims randomly over time because, as quantum mechanics has shown, light has inherent “uncer-tainties” in its features, manifested as moment-to-moment fluctuations in its properties. Controlling these fluctu-ations—which represent a sort of “noise”—can improve detection of faint objects, produce bet-ter amplified images, and allow workers to more accurately position laser beams.

Quantum mechanics has revealed light’s un-avoidable noise, but it also provides subtle ways of reducing it to values lower than physicists once imagined possible. Researchers can’t com-pletely eliminate the noise, but they can rear-range it to improve desired features in images. A

quantum-mechanical technique called “squeez-ing” lets physicists reduce noise in one proper-ty—such as intensity—at the expense of increas-ing the noise in a complementary property, such as phase. Modern physics not only enables useful noise reduction, but also opens new applications

for images—such as transferring heaps of en-crypted data protected by the laws of quantum mechanics and performing parallel processing of information for quantum computers.

Perhaps most strikingly, the quantum images produced by these researchers are born in pairs. Transmitted by two light beams originating from the same point, the two images are like twins separated at birth. Look at one quantum

continued on next page

A laser beam (“probe”) first passes through a mask that imprints a visual pattern. Along with a second laser beam (marked “pump”), it enters a cell containing a gas of rubidium atoms. Interactions between the rubidium gas and the beams produce an amplified version of the imprinted image as well as a second version of the image, rotated 180 degrees around the pump. The bottom panel shows, from left to right, an incoming probe beam imprinted with the letters “N” and “T,” an outgoing probe beam with an amplified image, and an upside-down version of the letters. The middle im-age is “entangled” with the rightmost image; the images’ changes over time are highly related to one another.

Credit: V. Boyer et al., JQI

Quantum-Entangled Images, continued

5

image, and it displays random and unpredictable changes over time. Look at the other image, and it exhibits very similar random fluctuations at the same time, even if the two images are far apart and unable to transmit information to one another. They are “entangled”—their properties are linked in such a way that they exist as a unit rather than individually. Together, they are squeezed: Matching up both quantum images and subtracting their fluctuations, their noise is lower—and their information content potentially higher—than it is from any two classical images.

To create quantum images, the researchers use a simple yet powerful method known as “four-wave mixing,” in which incoming light waves enter a gas and interact to produce outgoing light waves. In the setup, a faint “probe” beam passes through a stencil-like “mask” with a visual pattern. Imprinted with an image, the probe beam joins an intense “pump” beam inside a cell of rubidium gas. The atoms of the gas interact with the light, absorbing energy and re-emitting an amplified version of the original image. In addition, a complementary second image is created by light emitted by the atoms. To satisfy nature’s requirement for the set of outgoing light beams to have the same energy and momentum as the set of incoming light beams, the second image comes out as an inverted, upside-down copy of the first image, rotated by 180 degrees with respect to the pump beam and at a slightly different color.

One breakthrough in the experiment is that each image is made of up to 100 distinct regions, akin to the pixels forming a digital image, each with its own independent optical and noise properties. A pixel on one image forms a partnership with a pixel on the other image.

Look at two unrelated pixels—for example, a pixel in the top row of the first image and a pixel in the top row of the second image—and they appear to be doing their own random thing. But for two en-tangled pixels—the upper left pixel in the first image and the lower right pixel in the second image—their random fluctuations over time are eerily similar—one could predict many of the properties in the second pixel just by looking at the first.

“Making entangled quantum images is really striking, but what is most impressive to us is that the technique for making them is so much easier than what was possible before,” says Lett.

Previous efforts at making quantum images have been limited to building them up with “photon counting”—collecting one photon at a time over a long period of time, or having very specialized “images” such as something that could only be constructed from a dot and a ring. In contrast, the new method produces an entire image at one time and can make a wide variety of images in any shape. Moreover, those earlier efforts have been difficult to implement—some setups required light to bounce back and forth between tightly controlled, precisely spaced mirrors. By contrast, the four-wave mixing approach requires easy-to-prepare laser beams and a small cell of rubidium vapor.

A next goal is to produce quantum images with slowed-down light; such slowed images could be used in information storage and processing as well as communications applications.

Lab set-up. P: pump, Pr: probe, C: conjugate, LO:local oscillator, BS: 50/50 beamsplitter, PBS: polar-izing beamsplitter, PZT: piezoelectric actuator, Rb: Rubidium vapor cell, SA: spectrum analyzer

Quantum Logic Gates, from page 1

6

That’s not easy. Transistors transfer cleanly de-fined, fixed-value (0 or 1, on or off, high or low voltage, etc.) information units between gate components. Quantum computer logic gates, however, will have to transfer superpositions of multiple simultaneous states (0 and 1 in some proportion) from one object to another. Very few objects or systems are suitable.

Current favorites include trapped ions, “cavity QED” (in which single atoms and photons are coupled in microscopic enclosures) and trapped electrically neutral atoms. The last are particu-larly promising because neutral atoms have relatively long coherence times (in which they can remain in the superposition of states) and scientists have developed proven procedures to control them through cooling and trapping.

But the “down-side of neutral atoms,” says JQI Fellow Steve Rol-ston, “is that they don’t interact much. And because they’re neu-tral, it’s usually a comparatively weak interaction.”

Moreover, getting them to influence one another usually requires some sort of collision, which runs the risk that mechani-cal motion from the impact will hasten collapse of the superposi-tion, a condition called “decoher-ence.” So about 10 years ago, Rolston and a few collaborators began looking at a novel notion for how to construct a two-atom logic gate.

They were interested in taking advantage of peculiar phenom-ena that occur in Rydberg atoms. An atom in the Rydberg state, named for Swedish physicist Johannes Rydberg (1854-1919), has been super-excited; its out-ermost electron has absorbed so much energy that it has nearly left the atom. It remains at-

tached, but orbits at an exaggerated distance from the nucleus, giving those atoms certain special properties.

One is electric polarizability, and hence a height-ened response to electrical and magnetic fields -- itself a potentially exploitable feature. But perhaps even more useful is a behavior called the “Rydberg blockade.” It works like this: Take two adjacent atoms of the same element, and put one in a Rydberg state by hitting it with just the right frequencies of laser light. Then give the other atom the same treatment. Even though it has been excited identically, the second atom will not enter the Rydberg state. The presence of an existing Rydberg atom prevents the creation of a second one within a certain radius.

continued on next page

Quantum Logic Gates with Rydbergs, continued

7

The blockade effect provides three considerable advantages as a basis for future logic gates: It happens very fast; the interaction is strong (or-ders of magnitude stronger than the attractive van der Waals forces that typically bind neutral atoms); and it does not require physical contact between atoms.

Rolston, Peter Zoller of the University of Innsbruck and collaborators there and else-where proposed a scheme for using Rydberg states as quantum gates in 2000. Now sever-al labs around the world are study-ing ways to make that happen.

Eventually, Rol-ston’s group at the University of Maryland (UMD) -- working with JQI colleagues at the National Institute of Stan-dards and Tech-nology (NIST) and the Condensed Matter Theory Center at UMD -- hopes to be able to confine Rydberg atoms in an optical lattice.

Optical lattices are grids formed when laser beams overlap and interfere with one another. The result is a geometrical pattern of energy gra-dients. When atoms are placed in the lattice, they naturally settle into the minimum-energy spots with neat, regular spacing. At that point, re-searchers could use very carefully directed laser

beams to manipulate the atoms into controlled interactions for use as logic gates.

“But first,” Rolston says, “there’s the matter of making it actually work experimentally.”

That will entail sev-eral different lines of research. One in-volves learning how to fine-tune control of Rydberg atoms, which are so highly excited that they can easily lose their outer electrons, either spontaneously or by overstimulation from the laser beams. That process, called ionization, destroys the atoms’ utility as logic-gate compo-nents.

In addition, inves-tigators have to develop the ability to target and hit only selected atom pairs in the lattice, and to create a lattice with the right spac-ing interval to make targeting possible.

At present, Rolston’s group is working

with rubidium at-oms. Like its chemical

cousins sodium, potassium and the other alkali metals, rubidium has a single electron in its out-ermost orbital, and the specific laser frequencies necessary to excite that lone electron are obtain-able in the laboratory.

The final experimental configuration has not been determined. But the general arrangement of the apparatus is depicted on the next page.

Graduate student Jenn Robinson examines the laboratory apparatus used in the Rydberg experiments.

Quantum Logic Gates with Rydbergs, continued

8

Isolating and testing Rydberg atoms begins in a high vacu-um chamber [1] where a gas of approximately 500 million rubidium atoms is caught in a magneto-optical trap [2]. On the “ceiling” of the trap space is a device (photo at left) that produces a magnetic field gradient which herds the atoms into the desired location. Many evaporate away.

The laser beams excite a fraction of the atoms into the Rydberg states. After the Rydberg at-oms are formed, they are drawn off into a sepa-rate chamber (photo at right) in which they are exposed to strong electrical fields from a grid inside the chamber. In that field, the atoms lose their outer electrons and become positively charged ions. That process makes it convenient to collect them and count their numbers.

Once in position and cooled to a few hundred-bil-lionths of a degree above absolute zero, about 40 percent of the atoms remain, and are exposed to laser beams that enter through the side ports [3].

1

2

3

Moving On and Moving Up

Summer Schedule . . .

Joint Quantum InstituteDepartment of Physics, Univ. of MarylandCollege Park, MD 20742E-mail: jqi_info@squid.umd.eduTelephone: (301) 405-6129

JQI is a joint venture of the University of Maryland and the National Institute of Stan-dards and Technology, with support from the Laboratory for Physical Sciences.

The next issue of this newsletter will be pub-lished on August 15, and will include reports on developments during July and early August.

Thereafter, the regular monthly schedule will re-turn with the September issue on September 30.

As always, JQI welcomes comments and sugges-tions for coverage and content from all sources. We hope to expand our news reports, Web site scope and outreach activities in coming months. As that effort begins, it will be extremely impor-tant to know what sorts of information and out-reach opportunities are most desirable and use-ful to the quantum-science community and the public alike. Please send ideas by e-mail to:jqi_info@squid.umd.edu.

9

JQI is committed to inspiring and training quantum scientists to meet the na-tion’s research needs – both now and in the future. A sample of recent PhDs and postdocs advised and supervised by JQI Fellows includes the following :

Alex Dragt -- Frederick Strauch (2004 PhD thesis on “Theory of Supercon-ducting Phase Qubits”) has been appointed an Assistant Professor in the Physics Department of Williams College. Prior to that he was a National Research Council Postdoctoral Research Associate at NIST.

Fred Wellstood -- Sudeep Dutta got his PdD in 2006, stayed on as a post-doc, and is going on to another postdoc position in biophysics at George-town University, working with Prof. Daniel Blair. Hanhee Paik got her degree in 2007, and has been working as a postdoc at the Laboratory for Physical Sciences with Kevin Osborn on superconducting QC.

Alan Migdall -- Matt Eisaman, currently a NRC/NIST postdoc, has accepted a research staff position as an applied physicist at Palo Alto Research Cen-ter (PARC) in Palo Alto, CA, where he will focus ondeveloping technologies related to renewable energy and sustainability as part of PARC’s Cleantech Innovation Program.

Luis Orozco -- Jietai Jing finished a postdoctoral appointment, and is now an Assistant Professor at East China Normal University in Shanghai.

Steve Rolston -- Robert (Scott) Fletcher (2008 PhD thesis on “Three-body recombination and Rydberg atoms in ultracold plasmas”) has taken a position with Northrop-Grumman in Virginia, working with the Defense Threat Reduction Agency.

DAS SARMA WINS KIRWAN AWARD

JQI Fellow Sankar Das Sarma, Director of the Condensed Matter Theory Center at the University of Maryland, has been awarded the university’s 2008 Kirwan Faculty Research Prize for ground-breaking work in quantum information science. The university-wide honor rec-ognizes a faculty member for a highly significant work of research, scholar-ship or artistic creativity .