compatible microcomputer. Thus, this procedure is applicable to real time, on-line kinetic analysis and model testing.Precision in the rate constants of the two more slowly decomposing anion radicals was excellent. However, a visible trend seen in Tables 111 and IV is the deterioration of precision in k at higher rates. While work is being done to fully explain the cause of such behavior, we feel that it is mainly a limitation of electrode size and data acquisition rate of 1 point/ms. Analysis of data for faster reactions is more heavily influenced by instrumental errors in the measured charge. Also, deviation from the model due to the time dependence of charging of the electrode double layer (Qdl) is most pronounced in the first few milliseconds after the potential step. This also contributes to degraded precision for larger rate constants where most of the kinetic information is contained at short times.* This limitation may possibly be removed by using ultramicroelectrodes, decreasing the time window into the submillisecond range and increasing the rate at which data are collected. Such an approach may also require incorporation of the time dependence of Q d ] in the model.Orthogonalization necessitates deriving an orthogonal form of the model. After the values of regression parameters are obtained by use of such a model, deconvolution to recoup the original set of parameters is required. Fortunately, the nature of GramSchmidt orthogonalization should eventually enable fully automated transitions between the original basis set and the orthogonal one. It is not necessary or practical to use an orthogonalized model for every problem to be solved by nonlinear regression analysis. However, our results demonstrate the usefullness of orthogonalization when serious correlation between parameters creates problems in convergence of nonlinear regression analyses. Since the transformation is made on the model, the method is compatible with currently used general programs for nonlinear regression.The thermal decomposition of pyrrole was studied behind reflected shocks in a pressurized driver single-pulse shock tube over the temperature range 1050-1450 K and overall densities of -3 X mol/cm3. Under these conditions the nitrogen-containing products found in the postshock mixtures were cis-CH,CH=CHCN, HCN, CH2=CH-CH2CN, trans-CH,CH=CHCN, CH,CN, CH2=CHCN, C2H5CN, CHcC-CN, and small quantities of C6HSCN, C6HSCH2CN, CH2=C=CHCN and CH,C=C-CN which began to appear at the high end of the temperature range. Products without nitrogen were CH,C=CH, CH=CH, CH2=C=CH2, CH4, C2H4, and small quantities of C4H6, C4H4, C4H2, C6H6, C6H5C=CH, and C6H5CH3 which appeared only at high temperatures. The main reaction of pyrrole under these conditions is a simultaneous unimolecular bond cleavage in the 1-5 (1-2)-position and a hydrogen atom transfer, followed by electronic rearrangement and (1) isomerization to cis-crotonitrile, (2) dissociation to HCN + C3H4 (mainly propyne) and (3) isomerization to allyl cyanide, with a branching ratio of approx...

No abstract

No abstract

The thermal decomposition of pyrrolidine was studied behind reflected shocks in a single-pulse shock tube over the temperature range 900-1400 K and overall densities of ~3 X 10~1 2345 mol/cm3. Under these conditions the following reaction products were found in the postshock mixtures: H2, CH4, C2H4, C2H6, C2H2, C3H6, CH2==C=CH2, CH3C=CH, CH2=CHCH2CH3, HCN, CH3CN, CH2=CHCN, C2H5CN, and small quantities of pyrrole and butadiene. Studies with a mixture of pyrrolidine and pyrrolidine-i/g show that C2H4 and C3H6 are produced by direct ring cleavage with the following rate parameters: pyrrolidine -C2H4 + (CH2)2-NH, k = 3.42 X 1016 exp(-75.2 X 103/ ) s"1 and pyrrolidine -C3H6 + CH2=NH, k = 1.35 X 1016 exp(-80.4 X 103/RT) s"1, where R is expressed in units of cal/(K mol). Hydrogen cyanide has the highest concentration among the nitrogen-containing products, followed by acetonitrile. No reaction products resulting from pyrrolidine isomerization were observed.

No abstract