H-21

Terahertz Spectroscopy for Pharmaceutical PAT Applications: Quantification Methods for Studying Theophylline Tablet Compositions

H. Wu1 , E. J. Heilweil2 , A. S. Hussain3 , M. A. Khan1 , 1Division of Product Quality Research, OTR/OPS/CDER, FDA, Silver Spring, MD, 2Optical Technology Division, Physics Laboratory, NIST, Gaithersburg, MD, 3Global Biopharmaceutical Development, Sandoz, Princeton, NJ


Background: Quantification study using THz spectroscopy for determining pharmaceutical tablet compositions is novel for THz pharmaceutical PAT applications.

Methods:Crushed theophylline tablets were mixed with polyethylene (PE) powder (~10 wt%), then compressed into THz transparent disks of thickness 2 mm in a 13 mm die at approximately 200 psi pressure. Prior to acquiring THz spectra, each disk was dried in the air-purged FTIR instrument (Nicolet Magna 550 with THz Si-coated beamsplitter and DTGS detector) for 1hr to remove moisture from the instrument. THz spectra were recorded at 4 cm-1 spectral resolution and 64 spectral averages. THz absorption spectra were corrected for PE matrix by ratioing against a standard pure PE pellet spectrum. Characteristic peak method, linear superposition method, and multivariate data analysis method were employed to correlate the THz spectra to the compositions of tablet constituents.

Results: Characteristic peak method works better for crystalline components than for amorphous ones. The best R2 values obtained for theophylline, lactose, MCC, MgS, and starch are 0.55, 0.73, 0.86, 0.72, and 0.45, respectively. Linear superposition method provides indication that there are possible interactions between tablet components. Both PCR and PLS1 generate excellent correlations: the R2 values are 0.96, 0.96, 0.97, and 0.23 for theophylline, lactose, MCC, and starch, respectively. MgS concentration can not be extracted via THz spectra directly due to low amounts present.

Conclusions: Theophylline tablet THz spectra were analyzed using characteristic peak method, linear superposition method, and multivariate data analysis methods. The best quantification was achieved using PCR and PLS1 except starch.
2006 FDA Science Forum | FDA Chapter, Sigma Xi | CFSAN | FDA
Last updated on 2006-MAR-27 by frf