Number 624, February 13, 2003
by Phil Schewe, James Riordon, and Ben Stein
A Pinpoint Precision MAP
A pinpoint precision map of the cosmic microwave background, reported
this week at a press conference by scientists associated with the orbiting
Wilkinson Microwave Anisotropy Probe (WMAP), brings the early universe
into sharper focus.
The credibility of WMAP's pronouncement rests on three things: its
angular resolution is some 40 times better than that of its microwave
predecessor, the Cosmic Background Explorer (COBE); it comprehensively
surveyed the entire sky for a whole year (3 more years of data is yet
to come); and it measures the polarization of the microwave radiation;
the orientation of the radiation arises partly from the last scattering
of light at the time of "recombination," when stable atoms
formed for the first time, and partly from the time when ultraviolet
radiation strewn by the first generation of stars ionized once again
a lot of atoms in space.
Here are a few of the salient numbers coming out of the WMAP analysis:
the time of recombination was 380,000 years after the big bang
the era of the first stars was about 200 million years after the
big bang (surprisingly early)
the age of the universe is 13.7 billion years
the accounting of matter in the universe is as follows: atomic matter
makes up about 4%, dark matter about 23%, and dark energy 73%.
Solid, liquid, melting, and freezing are concepts that refer to bulk
matter, and not to individual atoms. But what about a cluster of a dozen
atoms?
Louis Bloomfield (University of Virginia) has assembled a nano-sized
grain of salt, a seven-atom blob of consisting of 4 cesium atoms and
3 iodide atoms. Compare this to an ordinary salt grain, with a size
of 0.2 mm and about 1.5 million atoms along each side of its cubical
structure.
By spraying this cluster with picosecond pulses of light, Bloomfield
has been able to make a "movie" of sorts showing how the cluster
rearranges its geometry: sometimes a 2 x 2 x 2 cube, sometimes a flat
2 x 4 ladder, sometimes an octagonal ring, all by virtue of the cluster's
own internal thermal energy; they don't image the cluster directly,
but their locations can be inferred from a mixture of measurement and
theory (for figures and cool movie, see http://rabi.phys.virginia.edu/research/
). Separate laser pulses are used to heat or to view the clusters.
One outcome of the experiment: "melting" of the tiny crystal
begins at a "temperature" of 225 C rather than 626 C, the
melting temperature of the bulk material. Studies like this are pertinent
to the production of nm-sized circuitry since one should know whether
a wire or some other structure will retain its basic shape or shift
into something else over time. (Dally
and Bloomfield, Physical Review Letters, 14 February 2003,
bloomfield@virginia.edu, 434-924-4576; see also How
Things Work: The Physics of Everyday Life, chapter 15)
Ultraviolet Lithography
Ultraviolet lithography can produce lines for integrated circuits as
small as 39 nm in one recent test. To help sustain Moore's law and cram
more and more gates and memory units into a given space, manufacturers
of microchips must make the lines in their circuitry ever smaller. This
usually means working with a shorter-wavelength light beam for creating
the patterns used for inscribing fine features on silicon or metal surfaces.
The form of lithography currently in mass production now can produce
a half-pitch size (equal lines and spaces in between) of 90 nm and isolated
line widths of 65 nm. To produce a later generation after that you would
need even shorter wavelengths.
At the Advanced Light Source at the Lawrence Berkeley National Lab
(LBNL) a government-industry consortium of scientists is trying out
this future lithography. Using a beam of synchrotron radiation in the
extreme ultraviolet range they have produced 70-nm line/space intervals
and isolated lines 39 nm wide (see figure).
By the time this type of lithography comes into play, by about 2007,
these numbers should be 45 and 25 nm, respectively. The consortium consists
of a government side, the "Virtual National Lab" (LBNL, Livermore,
and Sandia), and an industrial component comprising Intel, AMD, IBM,
Infineon, Micron, and Motorola. (Naulleau
et al., Journal of Vacuum Science Technology B, Nov/Dec
2002; contact Patrick Naulleau, pnaulleau@lbl.gov)