Showing papers by "David S. Cannell published in 1986"

Restructuring of colloidal silica aggregates.

Abstract: We present experimental results showing that colloidal silica spheres can be induced to aggregate either slowly or very rapidly. The slow process always yields clusters with a fractal dimensionality ${\mathrm{d}}_{\mathrm{f}}$=2.08\ifmmode\pm\else\textpm\fi{}0.05, but the rapid process can produce clusters with either ${\mathrm{d}}_{\mathrm{f}}$=1.75\ifmmode\pm\else\textpm\fi{}0.05 or ${\mathrm{d}}_{\mathrm{f}}$=2.08\ifmmode\pm\else\textpm\fi{}0.05. However, clusters with ${\mathrm{d}}_{\mathrm{f}}$=1.75 are always observed to restructure so as to yield ${\mathrm{d}}_{\mathrm{f}}$=2.08\ifmmode\pm\else\textpm\fi{}0.05.

Gelation of colloidal silica

Abstract: Classical light-scattering studies of colloidal silica gels show that they have a fractal structure for length scales shorter than a concentration-dependent crossover length $\ensuremath{\xi}$, and behave like a collection of nearly randomly distributed scattering centers for length scales greater than $\ensuremath{\xi}$. Our data show that $\ensuremath{\xi}$ has a power-law concentration dependence, $\ensuremath{\xi}\ensuremath{\propto}{c}^{\ensuremath{-}1.17\ifmmode\pm\else\textpm\fi{}0.1}$.

Wave-vector dependence of the initial decay rate of fluctuations in polymer solutions

Abstract: We present experimental results for the initial decay rate ${\mathrm{\ensuremath{\Gamma}}}_{1}$ of concentration fluctuations in solutions of linear polymers, for various molecular weights, concentrations, and wave vector q, in two rather different solvents. We find that under all conditions explored the dimensionless decay rate 6\ensuremath{\pi}${\mathrm{\ensuremath{\eta}}}_{0}$${\mathrm{\ensuremath{\Gamma}}}_{1}$/(${\mathrm{k}}_{\mathrm{B}{\mathrm{Tq}}^{3}}$) is a function only of the scaled wave vector q${\ensuremath{\xi}}_{\mathrm{H}}$. Here ${\mathrm{\ensuremath{\eta}}}_{0}$ is the solvent viscosity, and ${\ensuremath{\xi}}_{\mathrm{H}}$ is the length scale defined by ${\ensuremath{\xi}}_{\mathrm{H}}^{\mathrm{\ensuremath{-}}1}$\ensuremath{\equiv}${\mathrm{lim}}_{\mathrm{q}\ensuremath{\rightarrow}0}$6\ensuremath{\pi}${\mathrm{\ensuremath{\eta}}}_{0}$\ensuremath{\Gamma} 1/(${\mathrm{k}}_{\mathrm{B}}$${\mathrm{Tq}}^{2}$).

Effects of finite geometry on the wave number of Taylor-vortex flow.

Abstract: Richard Heinrichs, Guenter Ahlers, and David S. Cannell Department of Physics, University of California, Santa Barbara, Santa Barbara, California 93106 (Received 16 January 1986) We report measurements of the axial variation of the wave number q of Taylor-vortex flow in a system with aspect ratio 17 <. L <. 25 containing ten vortex pairs between rigid nonrotating ends. Near the critical Reynolds number Rc, qis very nonuniform when its average value ^differs significantly from its critical value qc. For sufficiently small \\q — qc\\, the finite geometry eliminates the Eckhaus instability. Our results agree quantitatively with solutions of the Ginzburg-Landau equation.

Aggregation of Colloidal Silica

Abstract: Spherical colloidal silica particles of ~ 7nm diameter may easily be induced to aggregate into “large” (≳5µm) structures even in dilute solution, by the addition of sufficient salt. The resulting aggregates appear to be fractal, scale invariant objects, as determined by classical light scattering. Studies carried out in 0.5M NaCl show that the fractal exponent characterizing the clusters is 2.08±0.04 over a wide range of pH, and that the aggregation kinetics are slow and exponential in time. Studies in 1M NaCl reveal a transition to very rapid aggregation at high pH or very low particle concentration. This rapid aggregation regime results in a fractal dimension of either 2.08 or 1.77 depending upon conditions.

Ahlers and Cannell respond.

Abstract: Ahlers and Cannell Respond: Folse and Mead make the interesting observation that the patterns observed by Ahlers, Cannell, and Steinberg^ rotate at a uniform rate for t ̂ 60. Here ris time measured in units of the horizontal thermal diffusion time L^/^, with t^^SJK and L=^ Rl d^ where d\\s the vertical dimension of the container of radius i?, and K is the thermal diffusivity. The original authors were aware of the fact that their patterns rotated (see Ref. 10 of Ref. 1), but they had not established the uniformity of that rotation. With the orientation B of the pattern in radians and the time I'm units of t^, we obtain 0-= d9/dl^2.0x lO\"\"\"* from the analysis of Folse and Mead. Rotation of a convective flow pattern in a circular container had been studied previously by Steinberg, Ahlers, and Cannell.^ Those authors found that rotation occurred at the uniform rate ^ = LOxlO\"\"^ for their case, but rotation ceased when a small change occurred in the pattern. This change yielded a pattern with a plane of mirror symmetry. Although more extensive experimental and theoretical studies of this problem are clearly desirable, it appears that the rotation is ''driven\" by departures of the pattern from mirror symmetry. On the basis of the observations of Steinberg, Ahlers, and Cannell, we assume that a pattern with mirror symmetry consisting of two identical grain boundaries and otherwise straight rolls in a cylindrical cell does not rotate.^ Thus we also presume that the rotation of the patterns of Ref. 1 was not strictly uniform. Whenever two nearly identical grain boundaries were established, rotation probably ceased. However, this situation prevailed only for a small fraction of the time of the experiment reported in Ref. 1. We also presume that the instantaneous rotation rate depended upon the extent of departures from mirror symmetry. However these short-time variations in 0 are not resolved by the data of Ref. 1 or the analysis of Folse and Mead. Rather, the analysis shows that the coarse-grained value of 9 which can be inferred from the experiment is constant. We therefore presume that the extent of deviations of the patterns from mirror symmetry, on the same coarse-grained scale, is time independent. We believe that this lends further support to the idea that the process involved in the time dependence reported in Ref. 1 is stationary and not relaxational.