Microsoft word - analyticalcume_091208_solution.doc

Questions on this cume are drawn from the attached article, entitled Quantitative Nondestructive Methods for the Determination of Ticlopidine in Tablets
Using Reflectance Near-Infrared and Fourier-Transform Raman Spectroscopy
Where specific passages are relevant, they have been marked in the paper with the corresponding question number. The questions are marked with point values, partly to provide an indication of the relative emphasis and expected level of detail. Keep your answers succinct and on-topic; points will be deducted for irrelevance. Many of these questions can be fully answered in 2-3 short sentences. Write clearly – what cannot be read cannot be graded. Question 1 (20 points): The principle of a FTIR spectrometer and the principle of a Raman
spectrometer are very straightforward. But how does a FT-Raman spectrometer operate?
Discuss the principle of operation including data acquisition and basic data treatment. A
schematic might be very helpful for your discussion.
Question 2 (4 points): The authors have used a laser emission wavelength of 1064nm. Give
an advantage and a disadvantage of using a NIR laser.
Advantage: less fluorescence; disadvantage: NIR laser produce weaker Raman signal compared to shorter wavelengths Question 3 (16 points): This paper describes the use of Raman spectroscopy of a powdery
mixture/tablet made from powders. The authors state that the laser spot on the sample usually
has a diameter of 1mm. What is the problem to begin with and how did the authors work around
it?
Due to the grainy nature of the sample, the Raman signal changes depending on how much of the laser spot covers analyte or filling material. The authors average over several sample location or rotate the sample while illuminating it. Thus, they average out fluctuation in the signal. Question 4 (12 points): For quantitation, the authors have applied the chemometric method
Partial Least-Squares (PLS). Describe in general terms what PLS’s (or similar chemometric
algorithms) advantages are over direct application of Beer’s Law.
Chemical matrix and only target analytes need not be known Using many measurement points as opposed to only one enhances precision Question 5 (8 points): P. 253 (right column) paragraph ‘Raman Calibration Model for TCL
Tablets’: Do you agree or disagree with the two reasons why peak intensities have been
chosen for quantification. Give reasons.
(1) Several Raman peaks show a concentration dependency; using only one measurement point reduces the measurement’s precision; ‘simplicity’ of a measurement is not a very strong argument here (2) Overlapping peaks can be handled by means of PLS; the statement that peak heights are immune to overlapping bands is way to general. Plus, the authors had to correct for a background due to the ‘placebo’ anyway (Fig. 2). In conclusion – these arguments are weak especially under the light that other parts of the paper use PLS – why not here? Question 6 (10 points): In table 1, a FT-Raman and NIR spectroscopic measurement series
are compared. Which statistical test would you use to test whether the mean values (bottom row
of table 1) are significantly different or not at a 95% confidence level (see hints)? Are they
significantly different or not?
For this question a pocket calculator may be used – if you have none available fill in the right numbers in the right equation and describe what you would test for. Full credit will be given for a sufficient explanation and description of the calculation even if calculations are not done. Thus, the means are not significantly different. Here it was assumed that the means had been obtained from five repeated measurement. The table footnote is ambiguous – it could be interpreted that at each of the three days five measurements were made. In that case n = n = 15 would be used. Both answers are accepted. Question 7 (15 points): Compare HPLC, NIR and Raman spectroscopy for this application –
give advantages and limitations of each technique.
HPLC advantage: very precise, used in the paper as gold standard HPLC disadvantage: measurement takes longer than spectroscopy, sample prep required NIR advantage: cheap instrumentation, faster (1 min) than Raman NIR disadvantage: less precise than HPLC, spectroscopic signatures broad and not overly specific -> selectivity can be a problem Raman advantage: faster than HPLC, spectroscopic signatures are very specific Raman disadvantage: expensive instrumentation, measurement takes longer than NIR Question 8 (15 points): Fig 4 and related paragraph on top of figure: What mistake have the
authors made while evaluating using PLS these transmission spectra?
The transmission spectra are non-linear in concentration according to Beer’s Law I (λ) = I (λ)⋅exp(− L ⋅ε (λ)⋅c). PLS and many other chemometric standard techniques require signals that are linear in concentration. The authors should have converted the transmission in absorbance spectra prior to applying PLS. s ⋅ n − + s ⋅ n − x , x mean values of data set #1 and #2, respectively; s , s standard deviation of data set #1 and #2, respectively; n , n number of repeated measurements in data set #1

Source: https://ion.chem.utk.edu/cumes/Analytical/9-12-08-analytical.pdf