Performance Analysis of an Integrated Process Raman Analyzer Using a Multiplexed Transmission Holographic Grating, CCD Detection, and Confocal Fiber-Optic Sampling

Abstract

The performance of an integrated turn-key on-line Raman spectrometer based upon a multiplexed transmission holographic grating, holographic laser rejection filters, and a charge-coupled-device (CCD) detector is described and compared with a laboratory-based research-grade FT-Raman spectrometer. Data for the dispersive system were acquired with a noncontacting, confocal dual-fiber probe of up to 100-m length, with Raman shifts down to 50 cm−1 and no apparent fiber background features. Despite the fact that the FT-Raman system was directly coupled to the sample (i.e., not through fibers), the sensitivity of the dispersive system was almost three orders of magnitude greater for equivalent incident laser powers and accumulation times. The potential for performing multivariate calibrations with the dispersive equipment was also investigated. With samples of known density, it was possible to produce a partial least-squares calibration for poly(thylene terephthalate) (PET) density with a precision of 0.002 g cm−3, with the use of a two-factor model. This precision compared favorably with previous calibrations using FT-Raman data. The effect of changing the f-number of the collection optics of the fiber probe head (to allow variation in the working distance) was also examined. It was found that, for transparent samples, the f-number could be changed by a factor of three without greatly affecting the Raman signal intensity, whereas for an opaque solid increasing the f-number greatly reduced the detected intensity. The reasons for this difference are discussed