Mercury exposure from compact fluorescent lights

A commenter on a previous post asked about potential mercury exposure from broken compact fluorescent lights (CFLs). There are also various stories floating around the internet (such as this one) about the dangers of mercury in CFLs. So I thought I’d take the opportunity to look at how much mercury someone could be exposed to from a broken CFL, and whether there were any risks from that level of exposure.

A CFL typically contains about 4 mg of mercury (according to US EPA); a lot of newer CFLs contain 1 mg or less. I’ll look at the worst-case exposure, so let’s go with 4 mg of mercury in a bulb. In reality that mercury isn’t going to all be in the air right away – the evaporation rate of mercury is about 56 micrograms per hour per square centimetre – but figuring out the rate at which it enters the air requires assumptions about the area covered by the spilled mercury, temperature, pressure, etc. To keep things simple and to make sure I’m considering the absolute worst case, I’ll assume that all of that mercury instantly volatilizes.

After the mercury is released into the air, two things are going to happen. First, it will spread out, occupying a larger volume of air. Second, air movement in and out of the room will remove mercury over time. So let’s look at a couple of specific scenarios:

1. Let’s assume that right after the bulb breaks a person bends down to pick up the pieces. To keep the math simple I’ll say the mercury has spread into a volume of 1 m3 at this stage – large enough that the person’s head is within the affected volume and representing pretty much a worst-case exposure. In reality the mercury concentration within this 1 m3 volume would be highest right next to the floor (presuming the broken bulb is on the floor) and decreasing as you go up, but again to keep it simple and look at an absolute worst-case I’ll simplify things and pretend the mercury is evenly mixed over that entire 1 m3. The resulting concentration would be 4 mg mercury per m3 air. This concentration would not last very long at all and represents a worst-case instantaneous exposure.

2. If the CFL bulb broke in a fairly standard-sized bedroom (around 3 m x 3 m x 2.4 m, or 21.6 m3 – I’ll round down to 20 m3), then over time the mercury would mix throughout the volume of the room, with an average concentration of 4/20 = 0.25 mg/m3. However, air is also going to move in and out of the room (and the house). Even at the low end in winter with windows closed, an air exchange rate of about 1 air change per 3 hours is fairly conservative, so about 1/3 of the mercury would be removed after every hour. Therefore after 1 hour the concentration would be 0.17 mg/m3, after 2 hours it would be 0.11 mg/m3, after 3 hours it would be 0.074 mg/m3, and so on.

These concentrations are probably unrealistically high, since I’ve assumed all of the mercury instantly volatilized. A more realistic concentration would consider the evaporation rate (56 μg/h/cm2). The volume of 4 mg mercury would be about 0.0003 cm3; if it formed a hemispherical drop (probably a reasonable approximation) the exposed surface area would be about 0.017 cm2. Assuming room temperature and sea-level atmospheric pressure, the concentration in the bedroom would peak at about 0.0003 mg/m3 (again based on mixing throughout the bedroom; concentrations would be higher if you put your head right next to the mercury on the floor).

So now we have a very conservative prediction of how much mercury could be in the air under worst-case conditions. What does this mean in terms of potential for effects? The infamous National Post story I mentioned above talked about a Maine Department of Environmental Protection employee comparing concentrations measured in a bedroom to a state exposure limit of 0.0003 mg/m3. This exposure limit is actually the US Environmental Protection Agency “reference concentration” (RfC). However, if a Maine DEP employee really was using this value here, that employee is utterly incompetent or at least unqualified for that sort of assessment – it is a completely inappropriate value to use. The RfC is a conservative estimate of the average concentration you could be breathing in over your entire life without expecting any adverse effects. I’ve seen some other evaluations use occupational exposure limits – again this isn’t ideal, since they’re based on a worker being exposed every day over their working career.

The Agency for Toxic Substances and Disease Registry compiled data on short-term exposures to mercury vapours. The data on short-term human exposures are fairly limited, but measurable effects were associated with very high mercury exposures (of the cases where concentrations were known, they exceeded 40 mg/m3). There are also some animal studies using short-term exposures; generally effects were only observed at concentrations above about 27 mg/m3. The one exception is developmental toxicity; there were a couple of animal studies showing that when pregnant rats were exposed to concentrations as low as about 0.05 mg/m3 over 6 or 7 days at a key stage of foetal brain development, their offspring were hyperactive and showed effects on spatial learning.

Overall, a broken CFL bulb won’t result in mercury concentrations in air anywhere near as high as the concentrations at which effects on humans or even animals have been observed. However, given that mercury is a known developmental toxin, and since concentrations could be higher very close to the spill or after disturbance (e.g. during cleanup), it’s probably not a bad idea for pregnant women to avoid cleaning up broken CFL bulbs if possible; there probably wouldn’t be an effect but it doesn’t hurt to err on the side of caution if possible.

The commenter I mentioned at the beginning of this post mentioned a different scenario where her husband drove in a car for 3 months in winter with a case of broken bulbs. A study of air exchange rates of stationary vehicles suggests between 1 and 3 exchanges per hour with windows closed and no ventilation, or between 1.8 and 3.7 exchanges per hour with windows closed and the fan on recirculation; a compact car has an interior volume of about 3 m3. If the case had 12 CFLs, that would be a total of 48 mg of mercury potentially released; if it was released all at once this would result in an initial concentration of about 16 mg/m3 which would rapidly decrease. Again this mercury probably would not be volatilized to the air all at once though. Rather mercury would be released gradually over time. It’s honestly almost impossible to calculate the likely concentrations, particularly without knowing things like the temperature, how the car was used, etc. If we assume room temperature (e.g. the heater was on) and that each bulb resulted in a hemispherical mercury drop as described above, then a concentration as high as about 0.004 mg/m3 could be calculated, but there are so many assumptions and unknown variables that I wouldn’t want to rely on that number. The concentration probably isn’t in the range where short-term effects are likely, but there are so many unknowns I wouldn’t feel comfortable making predictions about long-term effects, particularly if the driver was spending long periods of time in the car.