Abstract
Journal of Near Infrared Spectroscopy
Volume 13 Issue 1, Pages 1–8 (2005)
doi: 10.1255/jnirs.451
The rapid, at-line determination of starch content in sucrosestarch blends using near-infrared reflectance spectroscopy; a process analytical technology initiative
Brad Swarbrick,a Bronwyn Groutaand James
Nossb
a Process Analytical Technology Group, Pfizer Australia, 38-42 Wharf Road, West Ryde, Sydney 2114,
Australia
bTechnical Services Department, Pfizer Australia, 38-42 Wharf Road, West Ryde, Sydney 2114, Australia
A near infrared (NIR) reflectance spectroscopic method was developed for the rapid determination of starch content in the excipient material sucrose special powder, 4.6% starch. It was previously determined that the variability of starch content in individual bags within each delivery of this material was high. This required appropriate sample selection to provide a controlled content of starch before use in the pharmaceutical industry. Therefore, an NIR method was developed to reduce long laboratory testing time and to facilitate testing of the starch content of the material at-line in the receiving warehouse, [standard error of prediction (SEP)NIR = 0.16% compared with 0.15% for the laboratory method]. Another important feature of the NIR starch model was its use as an at-line tool for in-house blending of sucrosestarch blends to determine blend homogeneity. In one particular application, a sucrosestarch blend, known to contain low starch levels, was spiked with commercially available starch and blended. The NIR method was used to make rapid decisions on content uniformity and the results obtained were in close agreement with laboratory assay data (mean starch content using the NIR method 4.5% compared with 4.6% for the laboratory method)
Keywords: near infrared reflectance spectroscopy, starch, process analytical technology, partial least squares regression, pharmaceutical industry
Full-text article (496 kB) (subscribers only)
Buy article on-line for £20 (get immediate access)
Permalink: http://dx.doi.org/10.1255/jnirs.451
QR Code (what is this?):



