Investigating the Performance and Discussing the Suitability of FTIR for TOC Emissions Monitoring

Fourier transform infrared (FTIR) spectroscopy is becoming increasingly used for emission monitoring applications for a range of inorganic gases. However, where TOC monitoring is also required generally a flame ionisation detector (FID) is used. The standard reference method (SRM) [2] is based on the use of flame ionisation detection. Whilst there are no technical issues in using FTIR and FID in tandem there are the practicalities and cost of the transportation and set-up of two analysers. Consequently, it would be advantageous to be able to measure all the required species by FTIR alone ? in addition, the FTIR also has the advantage of providing speciation information. Many UK VOC emissions are regulated and fall under either the waste incineration directive (WID) [3] or large combustion plant directive (LCPD) [4]. Consequently, regulatory monitoring must be carried out following the SRM or with an alternative reference method (ARM) where equivalence to the former has been demonstrated. Towards assessing the suitability of FTIR for TOC monitoring we have tested the performance of the Gasmet DX4000 (distributed in the UK by Quantitech Ltd) against that of a Sick Bernath FID for measuring VOC compositions generated in NPL?s Stack Simulator Facility. The FID is certified under MCERTS for TOC measurements whilst the DX4000 FTIR is certified for a range of inorganic gases (for example, NO, SO2, HCl) but, at the time of writing this article, no VOC species. Many performance parameters of the FTIR are well characterised due to the original MCERTS testing, for example, noise, drift and temperature sensitivity) so for the purposes of this investigation need not be repeated, consequently, we have focussed mainly on responses to different VOC mixtures and cross-sensitivities. The Stack Simulator Facility developed at NPL was used for the work so that testing was carried out under real stack conditions. The facility has been designed with a cross-stack pathlength of 1.5 m, four 5? BSP sample ports, 300 L capacity and is capable of velocities and temperatures of up to 10 m.s-1 and 200 ?C, respectively [5]. These specifications allow testing of instruments [6] and procedures [7], and the carrying out of proficiency testing schemes under real sampling conditions. The approach of the testing was to create test mixtures in the Stack Simulator based on the performance requirements for a low range (0 ? 20 mgC.m-3) TOC continuous emissions monitor (CEM) as detailed under BS EN 15267-38.