Measurements of emissions of Particulate Matter (PM) and Semi-Volatile and Intermediate-Volatility Organic Compounds (S/IVOC) from refinery sources – Phase 1: A refinery fuel gas-fired heater

Report no. 1/25: This report provides the results from two measurement campaigns carried out on a gas-fired heater at a refinery in north-west Europe. The main purposes of the campaigns were i) to measure the emissions of primary particulate matter (PM) and also the volatile organic compounds (VOCs) which act as precursors to the formation of secondary PM and ii) to derive emission factors (EFs) for comparison with those published by the US EPA. In this report the EFs are expressed as g/GJ, with energy based on the net calorific value (NCV) of the fuel.

First campaign. In this campaign measurements were made of both filterable particulate matter (FPM) and condensable particulate matter (CPM).

Condensable particulate matter (CPM). Two different test methods to measure CPM were used simultaneously. Those methods were US EPA M202 (2017 version) which uses chilled impingers and ISO 25597:2013 which uses dilution to cool the sample.

In these tests, M202 provided CPM concentration values about six times higher than the ISO method. However, the measurement uncertainty of the ISO method was significantly greater than that of M202 due to the effect of sample dilution. The CPM emission factors derived using the two test methods were:
− Dilution (ISO 25597): 0.81 g/GJ
− Chilled impinger (EPA M202): 4.79 g/GJ

For comparison the factor published by the EPA, derived from tests using M202, is 2.67 g/GJ.

CPM speciation analysis was only undertaken on the samples from the M202 tests. Insufficient sample was collected from the ISO 25597 tests. The average composition comprised 14% organics and 86% inorganics by weight. The main inorganic constituents were 30% sulphate and 19% nickel. Further testing established that the nickel and majority of the sulphate were due to contamination. The main organic constituent was diethylene glycol dibenzoate. This was unexpected and contamination is also suspected, but it has not been possible to identify the source.

Filterable particulate matter (FPM). In all tests the FPM concentration was less than the limit of detection value (LDV) of 0.2 mg/Nm3. The FPM emissions, related to energy consumption, were therefore <0.075 g/GJ. For comparison the FPM EF factor published by the EPA is 0.89 g/GJ.

Second campaign. This campaign measured the emissions of VOCs classified as semi- and intermediate-volatility organic compounds (S/IVOCs). No tests to measure total S/IVOCs have been previously reported for gas-fired sources in European refineries. This initial test was used to provide a preliminary estimate of emissions rather than an accurate quantification.
Samples were collected on sorbent tubes. Analyses were carried out using automated thermal desorption (ATD) systems, one linked to a gas chromatograph (GC) with a flame ionisation detector (FID) for quantification, with the other linked to a GC fitted with a mass spectrometer (MS) for compound identification. The protocol used for VOC classification was that a compound with the same retention time as an n-alkane had the same classification as the latter. The indicative emission factors derived were 0.95 g/GJ (LHV) for IVOCs and 1.2 g/GJ (LHV) for SVOCs.