If you see the interference pattern for exciting radiation, the distribution of emission intensity will probably be affected as well, that is, the answer to your first question is yes, you'll have to deal with distortions of the fluorescence spectra. I am not aware of any existing methods to correct for this (which does not mean they don't exist, I haven't really searched), but as an idea I would try to measure the spectra in the very same arrangement you used for Brewster-angle absorbance. To take polarized excitation into account, you probably have to rotate the sample to prove that the shape of the spectrum does not change. Maybe you also could think about diffuse excitation, that is to put a UV-translucent strongly scattering filter between the light source and your sample, so that the interference pattern would be smeared out because light would simultaneously come from all directions.

Florescence intensity will track with the light coupled into the film during absorbance so it's true that it should vary. However non-stimulated fluorescence in itself should not demonstrate an interference effect on its own because when it emits I think it will emit without any mutual coherence.

Light is always coherent with itself! Remember clasical Young's experiments. Also remember fresnel's bi-prism and e.t..c. If fluorecence spectrum of the thin film is broad then you always will have influence of entererence effects on the emission spectrum. If you would tilt your sample then you cold see a spectral shift of these interference features. It could be an experimental prove of the influence of the interference on the emission spectrum.

To remove these interference features from the emission spectrum I would advice to model such an interference with apropriate refractive index (possibly spectral dependant) and than to try to substract it from the experimental spectrum.

Point taken. Thanks

If you fluorescence is bright enough and film thickness is hundreds of nm you can try to excite normally to surface and collect fluorescence from the film edge using cylindrical lens and fiber

Not entirely sure from your description, but you may have constructed the conditions for a dye-doped DFB laser. See Wright, et al., J. Opt. Sci Am. B, Vol. 21 (2004), pp. 944-950 as an example of this. I agree with Igor regarding the right way to observe the fluorescence to minimize the effects of the grating.

Suppression of Interference Effects in Spectroscopy Using an Integrating Sphere

http://www.hpl.hp.com/techreports/96/HPL-96-69.pdf

There is a robust way to measure the absorbtion in a single pass within the film or interference-less. You need to measure the magnitude Transmittance/(1-Reflectance). But your substrate must be transparent and of course you have yo measure R and T independently (see paper). Alternatevely you can obtain the interference modulated absorption A by using A=1-R-T. Finally you perhaps may correlate this absorption with fluorescent emission by capturing this emission with an integrating sphere. For all the measurements you may use an integrating sphere conected with an optical fiber to a spectrophotometer...

Here the link to a paper connecting to the use of magnitude T/(1-R)

Thanks to all of you for your comments and suggestions and links. I am still hoping to find a journal article somewhere that will tell me a formula to make corrections for the fluorescence spectra taken on a conventional fluorimeter. Otherwise I may have to find a fluorimeter with an integrating sphere.

Even for the side emission you will find interference effects. A hard problem to simulate-calculate due to the many angles and layers.

Integrating sphere seems to be mandatory to have an averaged "smoothed" efficiency