Sunday, May 18, 2014

Modeling “large” evaporating sources (backpressure)

Recent blogs have covered the subject of surprises.   During some of my early work on modeling I got a surprise relative to the calculated airborne saturation concentration of an evaporating source.  Like almost all scientific surprises, this one was very useful and lead to the development of a modified model for estimating airborne concentrations above and around large evaporating sources which I termed the backpressure model.

If you put an evaporating source into a closed vessel and have enough of the material within the vessel, ultimately the saturation concentration (Csat) will be obtained in the headspace air volume of the vessel.   Csat is the highest concentration of vapors possible from the evaporating source.   Csat is directly related to the vapor pressure as follows:

Csat = (vapor pressure/atmospheric pressure)(1,000,000) 
The units of Csat are ppm volume/volume or ppmv.

Now take a relatively small sample of the material of interest and put it on a watch-glass on an open scale and allow it to evaporate so that you can measure its evaporation rate in mg/hour.   This should equal the generation rate of an evaporating source (G).

We know from previous work that the concentration in any volume with a constant generation rate (G) and ventilation rate (Q) is:

Ceq = G/Q

Remember our watch glass experiment above.   We now have a value for G in mg/hr.    Say the evaporating surface area was about 5 cm2.    So now normalize this rate per unit area so we can express this G rate as mg/((cm2)(hr)).  

So we used this G to estimate the Ceq  (=G/Q) of a room with a large (10 m2) spill of toluene in a reasonably well ventilated room.   We calculate G by multiplying our experimental G rate per cm2 by the area of the spill which is 100,000 square centimeters. 

When we calculate Ceq we get a value that is MUCH larger than Csat which is physically impossible!  Surprise surprise!

What is going on here?   The answer is backpressure.   The entire driving force for evaporation is the diffusion of the molecules from high concentration in and immediately above an evaporating liquid and the relatively low concentration of the same molecules in the receiving air volume.   When we release relatively few molecules from the source compared to the receiving volume, this driving force is maintained and G is constant.   However, when the receiving volume starts to contain a built up concentration of the evaporating molecules the driving force is diminished and the evaporation rate (G) decreases. 

Think of it this way.   In the desert with very low humidity in the air, wet clothes dry very quickly.   In more humid environs when the relatively humidity is 50% clothes dry only half as fast and at 100% humidity they do dry at all. 

In a large spill the initial generation rate (G) is maximized because there are no molecules of the evaporating liquid in the receiving air; however, relatively soon thereafter the molecules build up in the air and begin to retard the generation rate (G).   Thus, the generation rate is not a constant but is variable with time:

              G = Gmax (1-C/Csat)

Gmax is the initial generation rate at the beginning (t = 0)
C is the concentration in the receiving volume (which is a variable function of time until it reaches Ceq)
Csat is defined above

If C in any volumne ever gets to the point where it equals Csat  (Ceq = Csat), as it does in a closed vessel, then G becomes equal to zero.

Backpressure is always present over evaporating or vaporizing sources even in ventilated volumes; however, we usually do not have to account for it because its effect is relatively small when we have evaporating sources with small area-to-receiving volume ratios.   When it becomes important is when we have LARGE evaporating sources such as a large area spill indoors or large vaporizing sources like off-gassing wall paint or carpets.

There is a module in IH MOD for considering backpressure (The Well-Mixed Room Model with Backpressure).   Because backpressure shows itself mostly in indoor sources with large areas, this is the correct model in which to evaluate its effects.

The original paper on backpressure has a lot more detail and I would be happy to send it to whomever is interested and writes to me at

1 comment:

  1. ERATA: I mention a "watch glass" in the above blog. Perry Logan reminds me that because of the geometry of a watch glass (a kind of concave dish) that the surface area changes during evaporation and this is an artifact of the watch-glass and not the area-normalized evaporation we are attempting to study. Thus, it would be better to use a flat bottom plate like a petri dish; however, a petri dish has relatively high sides which can disturb the airflow. I still think a thin layer of evaporating VOC in a petri dish might give reasonable results. 3M actually machined a block of aluminum with a shallow wall for their experiments and this is perhaps the best approach if you are going to do a lot of these experiments.