Particulate Collection Efficiency

Collection efficiency is the fraction of what enters a control device that fails to leave it, $\eta = (C_{in} - C_{out})/C_{in}$. A baghouse taking 12 g/m³ down to 0.06 g/m³ is running at $11.94/12 = 0.995$, or 99.5 %. The complementary quantity, penetration, is what actually reaches the atmosphere: 0.5 % here, and it is the far more useful number once efficiencies get high. Moving a device from 99 % to 99.9 % sounds like a tenth of a percentage point and is in fact a tenfold reduction in emissions. Anyone arguing about a control upgrade in efficiency units rather than penetration units is making a large improvement sound trivial.

Penetration is also what makes devices in series easy to reason about, because penetrations multiply while efficiencies do not. A cyclone at 90 % followed by a scrubber at 90 % passes $0.1 \times 0.1 = 0.01$, which is 99 % overall, not 180 % of anything. This is exactly how most real trains are built: a cyclone as a pre-cleaner to take the coarse load and protect the expensive device behind it, then a fabric filter or a precipitator for the fine fraction. Rough figures for a first pass are 70 to 90 % for a single cyclone, 90 to 98 % for a wet scrubber with venturi performance rising steeply with pressure drop, 99 to 99.9 % for an electrostatic precipitator, and 99.9 % or better for a well-maintained baghouse.

The number this equation produces is a MASS efficiency, and mass efficiency flatters every device on the list, because mass is dominated by the largest particles and every one of these mechanisms — inertia, impaction, interception, electrostatic migration — works best on large particles. The fractional efficiency curve tells the honest story, and for most collectors it dips to a minimum somewhere around 0.1 to 0.3 µm, where the particle is too small to be caught by inertia and too large for Brownian diffusion to help. That minimum is the most-penetrating particle size, and it falls squarely in the range that matters for PM2.5 and for health. A device advertised at 99.9 % on total mass can be well under 90 % on the fine fraction, and both statements are true at once.

Three practical cautions when the number comes out of a real test. Inlet and outlet must be sampled on the same basis, isokinetically and with the same moisture and temperature corrections, or the difference between two large numbers carries the error of both. Cleaning cycles matter: a baghouse re-entrains a puff of dust each time a row is pulsed and a precipitator does the same each time the plates are rapped, so an efficiency measured between cleaning events is not the one the stack sees over a shift. And an efficiency measured at design load says nothing about part load — a cyclone's efficiency falls off with the square root of gas velocity, so the same device at half flow is a noticeably worse collector, while a fabric filter usually gets slightly better. Efficiency is a property of the device AND the operating point, never of the device alone.