Professional Documents
Culture Documents
Since 1994, when plasma crystals were observed for To measure the 3D structure (z corresponds to the ver-
the first time in experiments [13], there have been a num- tical direction and xy is the horizontal plane) a vertical
ber of studies of crystal structure and phase transitions; see, laser sheet of 160 mm thickness, parallel to the xz plane,
e.g., [4]. Most of the experiments were performed in radio- moves in the y direction over a distance of 5.5 mm with
frequency (rf) discharges, where the micron sized particles a constant velocity y 0.30 mms. The duration of the
are levitated in the electric field of the lower electrode whole scan is about 18 s. The particles are imaged from
sheath in ordered structures with one or a few horizontal the side (in the y direction) by an external charge-coupled
particle layers. Three-dimensional (3D) crystals were also device (CCD) camera. The camera moves together with the
observed in some experiments, where the particles order laser sheet in order to keep the particles in focus [5]. Im-
themselves in fcc, bcc, and hcp lattices, as well as string- ages were recorded 15 min after particle injection, to allow
like structures [57]. the lattice to anneal. The crystallization started from the
In this Letter, we report measurements of a plasma bottom.
crystal with 19 horizontal lattice planes (layers). We im- In order to determine the particle positions in 3D, a
aged a volume with about 2 3 104 microspheres, which set of consecutive video frames (full frame rate 1t
is more than in previous reports (see, e.g., [5]). From 25 framess, resolution 768 3 576 pixelsframe, 8 bit
the images, we identify the xyz coordinates of all the gray scale steps) is digitized. Each particle is represented
particles in the sample volume using a technique that on a video frame by a light spot with an intensity dis-
we devised. Analysis of the individual particle positions tribution Ii x, z. An intensity threshold Ith is chosen to
shows that the crystal contains regions of either fcc or identify a particle from background fluctuations. We set
hcp lattices, distributed in different domains and sepa- Ii 0 when Ii , Ith . Then the coordinates R of the ith
rated by a transition region typically 23 lattice planes particle
R on the frame are X i , Zi x, zIi x, z dx dz
wide. The particle coordinates are then used not only Ii x, z dx dz. To measure the third coordinate Yi , we
to identify the local crystalline lattice type, but also to calculate the mean intensity of the ith particle for each
measure the three-dimensional pair correlation function. consecutiveRframe. For the nth R frame, the mean intensity
We determine the volume ratio of the two lattice struc- is Ii n P Ii x, z dx P
dz dx dz. Then Yi Ni yt,
tures coexisting in the crystal sample, which allows us to where Ni nIi n Ii n. The particle coordinates
make a comparison with results of molecular dynamics in a frame, Xi , Zi , have an error of one pixel or better
simulations [8]. (&9 mm), the error of Yi is about yt2 6 mm. The
The experiments were performed in a symmetrically thermal motion of the particles around their equilibrium
driven directly coupled rf chamber. The same setup was position is small, with a mean amplitude of less than one
used earlier in parabola and rocket flights in order to in- pixel.
vestigate the behavior of complex plasmas under micro- Using the measured 3D coordinates we found that the
gravity conditions [9]. The term complex plasma refers to horizontal interparticle distance Dxy decreases 40% from
composition (ions, electrons, charged microspheres) in top to bottom as shown in Fig. 1. We attribute this to the
analogy to the term complex fluids in colloids [10]. compression by the gravitational force. The particle num-
Polymer microparticles of melamine formaldehyde with a ber per unit length in the vertical z direction as well as
diameter of 3.375 6 0.102 mm were injected through a the total number of particles in each horizontal layer is
dispenser into an argon plasma at a pressure of 0.47 mbar. shown in Fig. 2. The distribution of particles in each layer
The peak-to-peak rf voltage on the lower electrode was is rather narrow, so that we can easily distinguish neigh-
61.8 V. boring planes. Figure 3 shows that the thickness of layers
4065
VOLUME 85, NUMBER 19 PHYSICAL REVIEW LETTERS 6 NOVEMBER 2000
4066
VOLUME 85, NUMBER 19 PHYSICAL REVIEW LETTERS 6 NOVEMBER 2000
4067