J. Segatz et al., DETAILED NUMERICAL SIMULATIONS IN FLOW REACTORS - A NEW APPROACH IN MEASURING ABSOLUTE RATE CONSTANTS, Journal of physical chemistry, 100(22), 1996, pp. 9323-9333
A detailed investigation is presented to simulate the reactive flow fi
eld in a low-pressure flow reactor for kinetic studies. This has been
done to improve existing methods for evaluating data from isothermal f
low kinetic measurements. The full Navier-Stokes equations for compres
sible flows including transport phenomena and chemical reactions have
been solved numerically for the low Mach number case. By using splitti
ng techniques for variables and spatial dimensions, the calculation ti
me could be reduced by more than 2 orders of magnitude. This reduction
allows the repeated application of the solver to adjust parameters in
the kinetic model organized as an optimization problem and give best
agreement between experiment and calculation. The model results for th
e nonreactive flow field have been verified by comparison to imaging m
easurements via two-dimensional laser-induced fluorescence of acetone
tracer gases for the visualization of diffusive mixing. Numerical resu
lts of full reactive flow simulation have been compared with the measu
rement of elementary relaxation processes and vibrational energy trans
fer in collisions of vibrationally excited hydrogen and deuterium mole
cules. Spatially resolved axial and radial concentration profiles of b
oth species were obtained at room temperature using coherent anti-Stok
es Raman spectroscopy (CARS). From the detailed numerical simulation e
valuated wall deactivation probabilities at 300 K for H-2(v=1) --> (wa
ll) H-2(v=0) (Ia) of gamma(w) = (1.5 +/- 0.3) x 10(-3) s and thermal r
ate constants for vibrational energy transfer H-2(v=1) + D-2(v=0) -->
H-2(v=0) + D-2(v=1) (IIa) of k(vv) = (6 +/- 0.5) x 10(9) cm(3) mol(-1)
s(-1) were derived using an optimization procedure specially adapted
to the present kinetic problem. They are larger, respectively, by a fa
ctor of 2 (wall deactivation) and 1.4 (vibrational energy transfer) co
mpared with values obtained from a plug-flow evaluation. While the k(v
v) data from different experiments are now in excellent agreement, the
oretical results using recent ab initio potentials still differ by a f
actor of 2.