In recent years, liquefied natural gas (LNG) has become a popular alternative to heavy fuel oils for use in marine engines due to its low cost and increasingly stringent emissions regulations from the International Maritime Organization (IMO). While natural gas (NG) combustion produces substantially less soot than other fossil fuels, its soot emissions are not negligible and have not been morphologically characterized. In this study, transmission electron microscopy (TEM) was used to interpret scanning mobility particle sizer (SMPS) distributions of catalytically-denuded particulate matter (soot) from an NG-powered commercial vessel. The vessel studied was powered by a dual-fuel compression ignition engine which uses a diesel pilot only to ignite the NG. Less than 2% of the engine's power is provided by the diesel pilot for engine loads above 50%. For comparison, tests were also run in diesel-only mode. During NG operation, the SMPS size distributions did not follow a lognormal distribution as expected. Instead, the distribution remained roughly constant from 10 nm to 100 nm. High resolution (HR)-TEM revealed particles as small as 30 nm in diameter were soot. From the TEM images, soot primary particle size and projected-area-equivalent diameter were estimated for individual soot particles. These data were then used to infer the effective density of the particles collected. In general, the primary particle size and correspondingly the estimated effective density of soot from this engine were larger than those from the “universal fit” proposed for soot in previous literature. For both operating conditions the source of the soot was likely the diesel pilot however, there were statistically significant differences between particles produced by the two fuelling modes. With LNG fueling, there were far fewer particles, and mean soot aggregate size decreased, but the primary particle size (for a 100 nm aggregate) increased, relative to diesel. Engine load did not have a statistically significant effect on soot morphology.
Read full abstract