Notes on a possible household air filter

I was thinking about filtering the air in my house, which is on a major thoroughfare with lots of diesel buses, burning that dirty high-sulfur diesel fuel which produces a lot of soot.

The standard approach, of course, is HEPA filters, which are made of disposable paper. I was thinking that a possible improvement would be to bubble the air through liquids, which could catch not only particulates but also a number of gases. Also, if you color the liquids, illuminate them with LEDs, and keep them in transparent tanks, it would look a lot cooler than a HEPA filter. You could use a series of such tanks to eliminate a lot of different particulates.

After the bubbles finish the tanks, they would need to pass through a fine screen to filter out any remaining droplets.

My apartment is 101 m³; filtering all of the air in it every six hours would thus require 4.7 ℓ/s of airflow, which is 10 cfm; as described in House scrubber, you can get that out of a 75mm-diameter 4-watt 3000 RPM fan designed for a CPU cooler. However, that fan won’t generate the head needed to bubble the air through a liquid; if we have a series of four tanks, each with liquid 100mm deep and of more or less the density of water, then we need 18 watts just to counteract the hydrostatic pressure, let alone producing the necessary flow rate through the bubble orifices. This suggests that perhaps using a few axial fans in series, each designed for a substantially higher flow rate, would likely do the trick.

XXX how do you calculate the head to produce a given flow through an orifice?

If the tanks are 10% bubble by volume, which seems like achievable and difficult to exceed by much, and the bubbles take 500ms to rise to the surface, then you need 23½ liters in each tank. If the tanks are 2 meters long, then they need to be 117½ mm wide in order to contain this much liquid.

You would need some system for replacing dirty filter liquids at some point, either manually or automatically. This is ideally done in batches to minimize the mixing of old dirty liquid in the new clean filter liquid.

Possibly useful liquids include:

• Plain water, which increases humidity, absorbs droplets of water-soluble liquids airborne from previous tanks, and absorbs particulate pollutants that aren’t hydrophobic. It will also dissolve small amounts of gaseous contaminants, but probably not enough to be useful.
• Vegetable oil, which absorbs hydrophobic particulate pollutants. Any oil would work for this, but vegetable oil has the advantages that it’s nontoxic, has very low vapor pressure at room temperature, and has somewhat higher surface tension than many alternative oils, reducing droplet formation. Used frying oil would likely work.
• Propylene glycol, which is nontoxic, has very low vapor pressure, won’t rot, and will absorb both hydrophilic and hydrophobic particulates, as well as vapors of many volatile organic compounds. It has relatively high surface tension (36 mN/m), reducing droplet formation. A lot of VOCs that won’t dissolve significantly in vegetable oil are completely miscible with propylene glycol. It’s also quite hygroscopic, so it serves to reduce humidity, but this could be a problem if it results in substantial dilution in normal use. At room temperature, propylene glycol reaches equilibrium with 60%-humidity air only once it’s absorbed about 20% of its own weight in water!
• An aqueous solution of CaCl₂ or a similar salt will reduce humidity, potentially down to a very low level. Aside from directly controlling the humidity of the output air, this could be used to reduce the dilution by humidity of a later propylene-glycol or similar stage. Calcium chloride is also nontoxic. (An anhydrous calcium-chloride desiccator is actually what Dow’s A Guide to Glycols recommends for drying air that will be in contact with propylene glycol if a nitrogen pad is not feasible.)
• An aqueous solution of a weak acid or weak base, such as sodium bicarbonate, could remove reactive gases such as SO₂ from the air. If you use calcium hydroxide (not weak!) or calcium carbonate for this, you produce gypsum as a bonus.
• Calcium hydroxide would also remove CO₂ from the air. If this is desirable, ethanolamine, diethanolamine or triethanolamine (in aqueous solution) may be a better choice, because it’s substantially easier to “regenerate” by heating (to 120° in the case of ethanolamine) to drive out the carbon dioxide (somewhere outside). These unfortunately require at least 5 but ideally 200 atmospheres to absorb the carbon dioxide.

Regeneration

I mentioned above that the ethanolamines absorb a lot of carbon dioxide, which can be driven back out by heating, “regenerating” them. The topic of regeneration is interesting in general: rather than discarding the dirty filter liquid, you go through some kind of process to clean it, thus extending its life.

Particulates can be removed by filtration, but in some sense this is not a solution — you could have just filtered the air. More interesting is if you can remove them by centrifugation or flocculation.

Hygroscopic solutions such as propylene glycol and aqueous calcium chloride can also be regenerated by heating to drive off some of the water.

Pebble-bed alternatives

As mentioned earlier, Dow’s recommendation for drying air that will be used to pad propylene glycol is to use an anhydrous calcium chloride desiccator. As I understand it, this is a pebble-bed kind of affair, with solid crystals of calcium chloride with air space between them.

Other kinds of pebble-bed-like things include the following:

• Platinum or palladium catalytic converters to remove organic compounds from the atmosphere, including even methane, as well as nitrogen oxides and ozone. Since a substantial part of the pollution from both Otto and diesel engines consists of unburnt hydrocarbons, nitrogen oxides, and ozone, this could be very helpful in the city. These need to be hot to work, typically requiring a refractory substrate such as alumina, and they produce carbon dioxide, which may need to be managed.
• Activated carbon to adsorb many kinds of contaminant gases.
• Oxide or hydroxide of calcium, lithium, sodium, magnesium, or even potassium, to combine with carbon dioxide or (I assume) nitrogen oxides.
• Sodium bicarbonate is famous for adsorbing unpleasant odors, and would also eat acid gases like SO₂ or nitrogen oxides, though not carbon dioxide.

In general, pebble beds have the advantage over bubble tanks that they have no minimum pressure to operate. Also, as desiccators, they can reach lower humidities than aqueous solutions of salts can. They have the distinct disadvantage in this case of looking substantially less bitchin.

Also, pebble beds are less suited to continuous-flow processes. You can regenerate pebble beds in place by taking them out of service and passing a regenerant over them — typically hot air, but activated carbon needs a hot non-oxidizing gas instead, such as hot carbon dioxide. Steam is not suitable, as it degrades the carbon to produce highly toxic water gas. (It would be nice to have a non-oxidizing gas that isn’t flammable or absurdly reactive, and is liquid or solid at room temperature, so that your activated-carbon regeneration gas doesn’t pose a suffocation hazard. But nothing occurs to me at the moment.)

Plasma alternatives

If you want an air purifier that looks really cool, nothing can beat plasma, especially a reduced-pressure plasma with different alkali metals evaporating into it (from oxide or carbonate feedstocks applied to your electrodes, presumably). This will look especially cool if it uses high-frequency AC and it’s inside a thin glass envelope so you can guide the plasma arcs with your fingers! But probably corona discharge in approximately atmospheric pressure is more practical.

Like a catalytic converter, this will also burn unburnt volatile hydrocarbons, and maybe also particulates, but the resulting gas is very far from breathable — it contains a substantial fraction of brown nitrogen oxide (NO₂), plus ozone and nitrous oxide (NO₂). The N₂O is reactive enough that you can combine it with just about anything (maybe bubbling it through sodium bicarbonate would be the easiest choice) but I’m not sure what to do about the ozone and nitrous other than using a catalytic converter from a car.

Oxidizing sodium, lithium, potassium, or even calcium or magnesium into the plasma, in addition to producing super awesome colors, might help to cut down on the nitrogen oxide production, too. But then you need to make sure you filter the generated metal oxides out of the air before you breathe it. Maybe some kind of hot acid refractory would work. Silica, for example, famously combines with sodium hydroxide to produce sodium silicate.

(Sodium and potassium nitrates are “saltpeter”, a stable mineral that acts as the oxidant in gunpowder. Calcium nitrate, “norwegian saltpeter”, also works for this. Magnesium nitrate is also stable. These are mostly used as fertilizer these days.

??? What are sodium, lithium, potassium, calcium, and magnesium nitrates like?

??? What are the other acid refractories?

Maybe the ozone could be made safe by passing the resulting gas over a “pebble bed” of something like used yerba mate or coffee grounds, thus converting it into relatively harmless carbon dioxide, and maybe a bit of water.