The desire to understand how light interacts with atoms and molecules has been one of the persistent themes of science in the 20th century. This quest led to the development of quantum theory, and the resulting dramatic progress in our understanding of matter on the atomic scale. In the 21st century, this high level of understanding inevitably leads to the question that is now a major theme in atomic and molecular science: given that we understand the basics of how atoms and molecules work, can we manipulate and control them in usefulways? Since atoms and molecules interact with their environment entirely by means of electro-magnetic fields, the preceding question can be formulated more specifically as follows: is to possible to "design" a pulse of light that, when incident on an atom or molecule, can precisely control its evolution in any desired way?

In experimental and theoretical work at JILA, we have taken this concept to the extreme. By manipulating a light pulse on the fastest possible time-scale relevant to atomic and molecular processes - that corresponding to less than a single undulation of the electromagnetic field of the light - we have manipulated atomic and molecular quantum wave functions in demonstrably useful ways. Experimentally, we found that very subtle changes in the shape of a light pulse measuring less than 20 fs (20x10^(-15) secs) in duration can dramatically change the spectrum of high-harmonic emission, making the emission process more efficient and channeling the emission preferentially into a single wavelength. We showed that this selectivity results from manipulating the wave function of the radiating electron during the interval between when it is ionized by the laser and when it "recollides" with the atom and emits the x-ray. This interval is < 1 femtosecond in duration. In general, if the laser pulse is not carefully-shaped, the x-ray bursts from adjacent cycles of the laser pulse will not necessarily add completely in phase, leading to a partial destructive interference that limits the intensity of the x-ray emission. In contrast, for an optimally shaped laser pulse, the x-ray bursts emitted from adjacent cycles add constructively, leading to a stronger x-ray emission. This work is one of the first results in the new field of "attosecond science", and demonstrates that we can control the shape of a radiating electron wave function on sub-angstrom length scales and sub-femtosecond time scales.

Fig. 2. An optimized laser pulse results in a series of x-ray bursts from an atom that are all in phase, thus enhancing the x-ray output.
PRL 86, 5458 (2001); Chemical Physics 267, 277 (2001).