Introducing SpectraFit™
Starting with Version 2.5 of KinTek Explorer™, we include an add-on, SpectraFit™, for fitting time-resolved spectra by singular value decomposition (SVD). If you collect time-resolved spectra as part of your kinetic analyses, you will find SpectraFit™ an indispensable and easy-to-use tool! Unlike other implementations of SVD for fitting time-resolved spectra which allow fitting to only a small number of canned models, SpectraFit allows you to fit your data to any model you input. Moreover, because the fitting is based upon numerical integration of rate equations, any concentrations of reactants can be used without the limitation imposed by pseudo-first order kinetics assumptions. Buy the SpectraFit™ add-on now.
Time-resolved spectra are de-convoluted by singular value decomposition (SVD) and the resulting amplitude vectors are then fit to the model programmed by the user to derive the spectra and time-dependence of individual species. Unlike other implementations of SVD, the user is not restricted to preprogrammed models with simplifying assumptions; rather, users fit data directly to the model based upon numerical integration of the rate equations with no simplifying assumptions. Moreover, multiple experiments with various signals and protocols can be fit simultaneously, taking advantage of the full power of KinTek Explorer™. In addition, KinTek Explorer™ provides comprehensive error analysis in fitting multiple experiments to a single model.
When data are imported into KinTek Explorer™, SVD is performed automatically. The user then enters a model, specifies the starting conditions for the experiment and defines the observable species. Fitting the data to the model resolves the individual spectra and time-dependence of each species. Finally, the model is used to reconstitute the original time-resolved spectra.
Below is a brief description of how time-dependent spectra are imported into KinTek Explorer™, and fit to the kinetic model of the user's choice. For a description with much more detail, look in the Users Manual, Chapter 10.
1. Time-resolved spectra are collected. The user prepares a data file in a tab-delimited text format. Data can be edited in MS Excel and then saved as tab-delimited text.
2. SVD analysis produces Eigenvectors for the spectra and amplitudes, representing a mathematical de-convolution without constraints of physical reality. For example note the negative absorbance and negative concentrations!
In this example, there are 5 Eigenvectors, displayed in the figures (five different colors). The user can select which Eigenvectors they think represent distinct species and de-select those they believe to represent noise. In this example, the light blue one is deemed to be noise, and was removed by the user (not shown).
3. Fitting the remaining SVD amplitudes to a model allows resolution of real spectra and the time dependence of species.
KinTek Explorer™ searches for a solution to the time dependence of the observables while attempting to eliminate negative absorbance values and negative concentrations. Using the example data set, the SVD Model produces the figures shown on the right.
4. The original time-resolved spectra are regenerated from the model.
The final solution depends upon fitting the SVD amplitudes, with arbitrary scaling factors, to derive an appropriate model and set of rate constants. In other words, the numerical integration of the rate equation produces the time dependence of individual species in absolute concentration, while the SVD amplitude vectors consist of a sum of these species with arbitrary scaling factors. The complete fitting of time-resolved spectra is simple and relies on an intuitive, interactive user-interface, yet the fitting is done to the most rigorous standard with no simplifying assumptions. Multiple experiments of any time, such as concentration series at a single wavelength, rapid quench-flow-data, or additional time-resolved spectra derived at different starting concentrations, can all be fit simultaneously to a single unifying model. There is no better way to fit kinetic data!
Buy the SpectraFit™ add-on now.