Atom level control

Welcome to the atomic ion trap effort in quanta lab! Our goal is to learn to control atomic ions in traditional linear Paul traps and in new experimental planar traps. By fabricating, characterizing, and using these traps for quantum control experiments, we will make a unique contribution to the worldwide effort to build a scalable quantum computer based on trapped ions.

Ions and Lasers

Fig. 1: Relevant level structure for the strontium 88 ion.

current ion of choice is strontium 88; its relevant level structure is shown in Fig. 1. One nice thing about this species is that all the lasers can be constructed from extended cavity diode lasers (ECDL's). There are three transitions shown in the figure. The blue (422nm) laser pumps the principal S-P transition. The rapid spontaneous emission rate (~20MHz) makes this the radiation with which to observe ions and Doppler cool them. So far, temperatures below 100mK have been reached. The infrared (1092nm) laser repumps ions which, in its absence, become "shelved" to a metastable D state. The red (674nm) laser addresses the electric quadrupole allowed transition from S to (another) D. This laser, due to the extremely narrow natural linewidth of the transition (1Hz!), is valuable for precise measurements of the ion temperatures, for performing single qubit gates using the optical {|S>,|D>} qubit, and for cooling on the motional sidebands to the zero point of motion. We have designed and built a 422nm diode laser which uses a filter cavity and optical feedback, under vacuum and vibration isolation, to produce Gaussian, low noise, frequency-stable output. Our red laser system uses the novel technique of locking a cavity to an atomic line and then locking a second red laser to this stable cavity; due to the narrowness of the S -> D transition, this laser must be stable to within 10kHz.

Paul traps

Fig. 2: Schematic of a generic linear Paul trap setup.

A linear Paul trap consists of four rods arranged in a quadrupole configuration (Fig. 2). Charged electrodes on the ends (endcaps) prevent escape along the trap axis. A radiofrequency potential (~5MHz, ~800V) is applied to two rods on a diagonal while the other two are held at DC. This creates a rotating saddle potential that forms a stable, roughly harmonic well. Loading is conventionally done by evaporating hot neutral atoms from a resistive oven and ionizing them inside the trapping region with a photoionization laser. Ion motion in the trap is well described by slow oscillations at the frequency defined by the pseudopotential superposed with small oscillations at the drive frequency, dubbed micromotion. Micromotion leads to heating of the ions, so its influence is minimized by using DC potentials to push the ions to the node of the rf electric field, where micromotion does not occur. The secular motion of cool ions can then be used as a bus for performing two ion quantum operations, an essential ingredient of any quantum computation.

Surface-electrode ion traps

Traditional ion trap traps are three dimensional systems operated at room temperature. We have experimentally implemented a new kind of ion trap, based on semiconductor lithography, fabricated at MIT, and operated at liquid helium temperature. This ion trap chip, pictured below, has enabled us to trap single ions of strontium for extremely long times, held ~100 micrometers above an electrode surface, and laser-cooled to its quantum ground state with better than 90% fidelity.

Fig. 3: MIT Cryogenic ion trap.

Heating rates and sideband cooling

A significant problem facing ion traps is the excess noise which grows in power as 1/d⁴, for trap size d, due probably to fluctuations of surface charges. This noise significantly degrades the performance of two-qubit quantum logic gates such as the controlled-NOT gate. Our silver-on-sapphire ion trap chip reduces the heating rate to less than one quantum of increase in harmonic motion per second, which is two orders of magnitude better than previous results in comparably sized traps, and seven orders of magnitude less than the heating rate observed in a trap of the same design at room temperature. This remarkable result suggests that surface charge noise strongly depends on temperature, and indicates the importance of studying surface electrode materials and fabrication methods for trapped ion chips.

These measurements were made possible by pulsed sideband cooling of a strontium ion to its motional quantum ground state, as evidenced by observations of Rabi oscillations, and asymmetry in the red and blue motional sidebands around the 674nm S to D transition of the ion. This transition has a natural linewidth less than 1 Hz wide, and the experiment employed a laser stabilized by a ULE cavity operated in vacuum, with finesse of 300,000.

Fig. 4: Sideband cooling of Sr+ at 4K.

Current & future work