White PapersMore >>
- High Performance Computing in Finance: Best Practices Revealed
- Catch the Security Breach Before It’s Out of Reach
Camera in hand, I visit the National Institute of Standards and Technology (NIST) Ion Storage Group and photograph the Ion Trap Quantum Computing experiment described in this blog last week.
If you haven't read that previous article , you'll probably want to do so now in order to understand what we are looking at today: a very constrained state-of-the-art quantum computing experiment which proves that ion-trap technology can be employed with a reasonable degree of accuracy to perform two-qubit computations.
The cartoon at the left by John Jost of NIST represents ions used for quantum computations (in red) paired with their cooling ions (in green). One pair is being transitioned by laser light.
Atomic Clock Watcher
You really cannot wander the halls of NIST without getting quantum mechanics all over you!I was early for my appointment, so public information officer James Burrus killed time with me the same way hosts have killed time at NIST with guests since at least 1991 (the last time I visited): he took me to see the Atomic Clock, which physicist Thomas P. Heavner of the Atomic Standards Group explained to me.
Cesium is cooled to near absolute zero and then excited by laser to shoot up and fall back down a sort of chimney, thus passing twice through a microwave aperture. Microwave energy nominally at the cesium frequency of 9192631770.0 Hz is projected through the cesium, which fluoresces in proportion to the right-on-ness of the microwave frequency. The frequency is corrected by the dual observation, one pair of observations per second, and used as a oscillation signal to the clock.
The Nearly Mystical Apparatus
When the time came, I was received very cordially by Jonathan Home and David Hume who explained their experiment and the apparatus, as well as allowing me to take photographs (while keeping me from angles of attack that would expose my eyes to injury by the lasers). Their partners in the effort, John Jost and David Hanneke, were not present.
Pretty much everything you see in the picture above (including the physicist!) is busy when the experiment is running. The ion trap itself (described below) is a small part of the collation. Most of the equipment in the picture consists of lenses, mirrors and crystals. The lenses and mirrors control the angle of the electric field vector of the light. The crystals attune laser light to frequencies necessary to transition quantum states in the ions which are the subject of the experiment.
In a existential juncture between quantum mechanics and swords-and-sorcery, crystals are impinged upon by acoustic waves (sine waves generated by realtime control programs) which makes a diffraction grating to tune light to the desired frequencies.
Be Vewy, Vewy Qwiet, I'm Trapping Ions
Here the ion trap (a printed multilayer wafer with a channel in which the ions can move one-dimensionally) is mounted in the "round generator-looking thingy with red wires on it" in the center of the picture. Photo: Jack Woehr
The ion trap is a slot in two wafers. They each have a set of 7 gold electrodes. The ions sit in a vacuum between the electrodes, which are sequenced to move ions along the center of the trap.The flat wafer, much like a printed circuit board, is the white object seen edge-on in the picture. Laser light can be projected through the trap to perform quantum operations on the ions.
The gold leads control the electrodes used to move the ions up and down the slot. The order of the ions in the slot cannot be changed in this experiment; motion is one-dimensional only. In a practical quantum computer you would want to be able to move the ions two-dimensionally so that they could pass around one another and into another region, i.e., into another qubyte register. In the present experiment, they must remain in order. The trap holds up to 10 ions.
Free the Beryllium Two!
Two beryllium atoms in the Trap (Photo: J. Jost/NIST). Two magnesium atoms are also present but do not fluoresce so do not appear in the image.
Every 20 millisecond experiment results in a number of photons which are counted by a photomultiplier tube. Hundreds of sequences are run and the photon count averaged. The point is that ions in one orientation of the quantum state being used as 1|0 in the quantum computation will fluoresce under laser light, and in the opposite orientation will not do so. Thus, the correctness of the result can be evaluated by counting photons.
Homebrew Conventional Software Tools
The experiments are run by a C++ program running on a workstation. Bison is used to create a compiler. Text is written in the ".dc" files and then compiled into FPGA control files by this compiler. C++ is also used for organizing data storage, and the graphical interface to control the experiments. It also interfaces in real time with the FPGA to extract the data, though not on the timescales of a single sequence in the experiment. Typically every 200 sequences there is an interrupt from the computer, which allows the C++ generated program on the computer to download information from the FPGA and upload new control variables.
Red and Green Laser Light Seen As Festive
Here is the laser source for the experiment. The rigid tube glowing green on the right is full of iodine ions and the pink plastic tube at the bottom is full of dye which gives off light of the right color when excited by the laser light. Photo: James Burrus
The ions engaged in computation are beryllium. Each beryllium ion is paired with a magnesium ion. Laser light is used to perform quantum transitions on the beryllium ions, but also to cool the magnesium ions (whose entanglement with the beryllium ions does not include the state being used for computation) thus cooling the beryllium ions.
The laser light is derived from "dye lasers", in which jets of dye are shot up into green "pump" laser light which excites electrons in the dye. As they cascade down they get to a transition and decay emitting red light. The decay is enhanced by a four-mirror cavity which allows light on the desired frequency to build up. The length of the cavity determines the frequency. The green laser light is used to access a particular transition of the red dye; it is the red dye which is emitting the light.
The actual desired light is twice the energy of the red light produced by the dye. So there is a second set of cavities that build up a lot of red light and send it into a crystal, which converts the red light into ultraviolet light.
The nastiest job in the lab, Jonathan and David told me, is cleaning the dye bucket. Just one of those sacrifices we make for science, I suppose ...