The spectrum will vary widely between aromatic and aliphatic hydrocarbons. A better approach might be to burn the sample and measure CO2 and water either using analytical chemistry or by IR spectroscopy.

I agree with Paul. UV/vis is not very specific, in particular when you analyse a complex fluid like petroleum. IR, NIR and Raman are better suited. the following text may be useful: http://dx.doi.org/10.3390/en8043165

Thank you for your recomendation. I know, that for TPH, PAH analyses mainly used Chromatography, IR Spectroscopy. For PAH also chemiluminesence. I think, that this compound may be have also absorbance peak in UV Vis range. In this case may be devoloped methods for determination this compound by UV vis spectroscopy

I agree with Johannes.  The problem here is not detection of a signal but how to weight all the possible constituent spectra- presumably you do not know the composition of the fluid ?

An alternative might be to use analytical chemistry to determine the total hydrocarbon content of a few "reference" samples and then use spectroscopy to interpolate on the "unknown" samples. Chemometrics (e.g. PCA) might help here although one would have to be very careful that the composition did not vary beyond the validity of the model.

Dear Paul,

Thank you! Yes, all quantity measurment we will use standard samples as reference. I know standard analyses method - EPA, ISO. But for me is interesting application UV Vis Spectroscopy for determination TPH and PAH. May be anybody was developed such kind of method

The UV approach can only be applied when composition of TPH does not change, because the method is more or less only sensitive to aromatic compounds. In other words the amount of 'inert' compounds will likely different between calibration and analytical determination, i.e. it will not be able to draw clear conclusions.

Good to al,l am presently working on some extracted compound for TPH and PAH analysi and I used UV Visible spectrophotometer for scanning and reasonable peaks are obtained at various level while 550 nm is constant in all which I assumed the peak is for the solvent used for the extraction i.e Dichloromethane while other peaks ranges between 200 and 900 nm, what can we say about this before I proceed for FTIR and GC-MS Analysis

In principal, an evaluated optical spectroscopy with a developed inverse algorithms of light transport in turbid media shall give you the total absorption coefficient (mu_a) and total scattering coefficient (mu_s). Now, assuming that petroleum hydrocarbon is an absorber, its concentration is related to mu_a as: mu_a = e*c*ln(2), where e is the extinction coefficient of petroleum hydrocarbon (unit of cm-1M-1), and c is molar concentration (mole/l). mu_a and e are function of wavelength. So as long as mu_a and e is known, plus you have a solid inverse model to simulate light propagation, you can use spectroscopy method to recalculate concentration of any absorber.