Laser light can lay lattices on substrates, opening the way to atom lithography that can write lines for semiconductors with resolutions in the tens of nanometers. In addition, the same technology could be used to move single atoms in ultra-dense memory applications. To explore possibilities for nanodeposition and fabrication, Yasushi Inouye and a team of researchers at Osaka University (Osaka, Japan) have built an atom-optical system that uses a narrowband diode laser to collimate and focus a thermal chromium (Cr) atom beam. "The narrow linewidth lets us concentrate the laser energy at the best frequency for cooling," says Inouye.
Inouye's team used a 0.5-W continuously tunable diode laser operating between 840 and 860 nm and fitted with a built-in amplifier. They directed the laser beam through a ring-cavity second harmonic generator and used lithium triborate (LBO) as the frequency-doubling crystal to convert the output to violet wavelengths. The laser is tuned so that the second harmonic matches the 7S3 → 7P4 transition in 52Cr, which is 425.55 nm in vacuum. The linewidth is also identical to that of Cr atomic resonance, which is Γ/2π = 5 MHz. They then split the laser beam with an acoustical-optical modulator to form beam pairs. They used the upper beam pair as the cooling source and the lower pair to construct a standing wave.
An AFM image shows the chromium line nanostructure diode-laser-focused atom deposition. (Osaka University)
"We effused chromium atoms from a cell with a 1 mm [aperture] by heating the cell to 1950 K, which is below the 2130 K melting point for that element," says Inouye. The group aimed the beam of Cr atoms through a 10-6 protactinium vacuum toward a substrate of <001>-oriented silicon. The team aligned the laser beam standing wave, which was nominally Gaussian in profile, so that it grazed across the sample with its peak intensity at the surface.
To get the proper flux and beam divergence, Inouye and his associates had to configure and optimize the atom emissions using three variablesthe heating temperature, aperture area, and aperture-to-substrate distance. Measurements and calculations estimated the probable atom velocity as 960 m/s and a deposition rate of about 0.1 nm/min.
The experimental setup precollimated the Cr atom beam with a 1-mm slit downstream of the aperture, then subjected it to 1-D Doppler cooling. The team used photoluminescent imaging to determine the angular distribution of the beam. The narrow 1-mm slit reduced divergence angles to less than 0.2 rad with the slit positioned 210 mm from the fusion cell. Moving the slit to 495 mm, a position 10 mm above the substrate, further reduced the divergence. "In fact," says Inouye, "after cooling, the beam divergence was beyond our detection limits. But, judging from the fabrication, the divergence was in the range of sub-mrad."
A dipole force of the standing wave generated by the laser focused the atom beam. Inouye's team found that the first-order electrical potential provides the spatial dependence that causes focusing. "In practice, each node of the standing wave acts as a lens, creating an array of lines with a spacing of half the interference laser wavelength," says Inouye.
Inouye's technique formed a structure of uniform lines. Their height was a function of the atom-beam intensity and deposition time; their width was a function of the atom-beam collimation, the standing-wave intensity and profile, and the wavelength and spectral width of the laser. The width was also influenced by the laser frequency and ambient oscillations. According to Inouye, the group formed a lattice with an average spacing between lines of 212.78 nm, which is one-half the interference laser wavelength. The full-width at half maximum measured 150 nm.
"The most significant step forward that Inouye has made is to utilize diode-laser technology to produce the laser light that does the collimating and focusing of the atoms," says Jabez McClelland, physicist at the National Institute of Standards and Technology (NIST; Gaithersburg, MD). "This will have technological impact because it makes the incorporation of the process into an actual fabrication implementation more realistic."