Monday, September 30, 2013

Gifts from the Netherlands – More UNIFAC and more

Compared to the US, the Netherlands is a relatively small country but it is gigantic in the stature and generosity of its scientists doing human health risk assessment.  Of course, these folks  are not the only other good guys in the world making great contributions to our field from relatively small countries (Demark and the UK come to mind) but one Dutch colleague really highlighted this for me recently because of his kindness.   My friend and colleague, Theo Scheffers (formal title:  Ir. Theo Scheffers RAH) read the last blog and sent me an email.   He asked if I was aware of an Excel sheet (XLUnifac.xls) that has been highlighted as an advanced tool for REACh.    REACh stands for Registration, Evaluation, Authorisation and Restriction of Chemicals.  It is a very big deal in Europe and something I will cover in a future blog or two.

Turns out I had not heard of XLUnifac.xls so Theo sent me a copy along with a pdf manual for the spreadsheet.    The program and manual were written by Preben Randhol and Hilde K. Engelien for their students at the Norwegian University of Science and Technology (NTNU).   It will do 15 mixture components (UNIFACAL does 5) and really goes into the background of the UNIFAC method which the UNIFACAL spreadsheet discussed last week does not.  So if you really like looking under the hood this is the program and document for you.  You can find the program with a Google search but if that gives you trouble just email me ( and I will send it to you along with the manual.

It further turns out that Theo has sent me a link to a site with quite a few very interesting tools for the Industrial Hygiene community.   He apologizes that some of the material is in Dutch but there is enough there for any of us to find of interest.   The site is: 
I will let you folks explore its contents for yourself but wanted to put up two sites that really caught my eye: The first is  This site presents a very extensive database of Occupational Exposure Limits from all over the world.   It includes OSHA and NIOSH limits but not the ACGIH TLVs. 

Another site that I found very interesting and potentially quite valuable:   Below is a cut and paste from this site:

“ChemSpider is a free chemical structure database providing fast text and structure search access to over 29 million structures from hundreds of data sources.
our introduction video.”  
Pretty cool!

Down in the Dutch language part of the TSAC web site is a section entitled:   SKINPERM.    In this section there is a link to the home page of another outstanding contributor from the Netherlands to our science:  Dr. Wil tenBurg.    Wil’s site: has been up for many years and contains a lot of great stuff.    The page is not flashy and has the very modest introduction reproduced below:

   "Main areas of interest are:
  • Estimation of the permeation rate of substances through the skin .
  • Estimation of dose/response relationships for acute inhalation toxicity, controlled by exposure concentration, exposure period and other independent variables"
The “skin” link is the program SKINPERM which estimates dermal exposure to chemicals applied to the skin and in the air over skin.  It does this as well or better than anything else I have seen or I have been able to find after many year of searching.   The “relationships” link presents his ground breaking software to calculate acute dose-response modeling (usually lethality) as a function of both time and concentration.   This is a very big deal in setting Emergency Response Planning Guidelines (ERPGs) and is used all over the world.   Wil provides these remarkable tools for free.   I will be discussing the technical details of all this in future blogs but I just wanted show off some of these links and to tip my hat to our colleagues from the Netherlands.  

Monday, September 23, 2013

Getting Activity Coefficients for Mixtures

In last week’s  blog I talked about non-ideal solution mixtures that can go far afield of the predictions of vapor pressure over a mixture from Raoult's Law.    The fix for this situation was to modify Raoult’s Law to include an activity coefficient (AC):


VPM = vapor pressure of the compound of interest over the mixture
VPP = vapor pressure of pure compound
MF =  mole fraction of the compound

The thermodynamic AC can be very large (>1000) as in the case of benzene coming out of water or pretty close to one (1.0) as in the case of the mixture of compounds with similar structures like methanol in ethanol or a mixture of aromatics hydrocarbons.  

So how do we come up with a numeric AC is for a compound of interest in a mixture of chemicals?   Well there are models, of course.    In this case there are relatively complicated physical-chemical models that use the structural characteristics of the molecules in the mixture to estimate the AC values of each component.  Here again we needed someone to do the computing coding so we would not have to wade through all the math.  We need a dedicated user-friendly program.  I think that the most useful one that I have found over the years of looking has been:  UNIFACAL.exe.    This little program (1.7mb) can do wonders.   The screen shot below shows what I mean: 

Notice that there are places for 6 mixture components.   I have put the two component (binary) system that I talked about in the last blog:  1.7 grams (solubility limit) of benzene in 1 liter of water.   UNIFACAL comes with a modest database of benzene, chlorobenzene, ethylbenzene, toluene and water.   It is is a relatively simple matter to add compounds using the database.  For example, I added ethanol by adding up 1 x CH3 group, 1 x CH2 group, and 1 x OH group.   You will find it is even easier than it sounds once you know the structure of the chemical you want to add and move to actually put it into the UNIFACAL Database.    All the halogens are in the database, along with some silicon, sulfur and nitrogen containing moieties. 

Notice that the AC of benzene in this aqueous mixture is predicted to be over 2400 which is essentially the full expression of the vapor pressure (VPP) of pure benzene.   The reasons for this are discussed in the previous blog. 

The latest version of this program was written and has been shared as freeware since 1998 by Bruce Choy and Danny D. Reible from The University of Sydney Australia and Louisiana State University and we owe them a debt of gratitude for their generosity. 

You can download this remarkable program at:  

Another way to estimate vapor pressure of compounds in water is to use Henry's Law Constant (HLC).
Henry's Law is simply a variation on Raoutl's Law.   It says that the amount of a mixture's component vapor in the equilibrium head-space will be a constant proportion to the amount of that component dissolved in the liquid mixture.  This constant ratio is called Henry's Law Constant (HLC) and if one knows the concentration in water and the HLC then he or she can calculate the concentration (i.e., vapor pressure) in the headspace.   Remember that the saturation  head-space concentration is convertible to the partial pressure which is the vapor pressure over the mixture (VPM).   We went through this calculation in a previous blog.    Many compounds have published HLC but these, of course, are only for water.   I must say that water is used a lot as a solvent but if you are not dealing with water and have dissimilar organics (e.g., straight and branch chain hydrocarbon and lower alcohol) you need to use the AC approach above.

Note that the AC is a function of concentration (i.e., molefraction) so you need to understand what might happen to the concentration of the evaporating liquid over the time of exposure.   If the time of exposure is pretty short or only a small portion of the evaporating liquid is expected to disappear during the exposure then you do not need to worry about it much; however, if the composition is expected to change significantly during the exposure then you need to account for this.   I usually do this by taking worst case.   That is, what is the worst emission rate that might occur during the exposure and use that for the entire exposure period knowing that it is a purposeful overestimate that still might be useful.  If you cannot live with the overestimation there are other ways to approach the problem but they always require more work. 

In the next blog I am going to talk about practical approaches to estimating emission source rates by various means under different circumstances.   If you send me some of the specifics of what has been challenging you I may use it as an example.   Give me as much detail as possible and let me know what I might need to "blind" for reasons of confidentiality.


Monday, September 16, 2013

Evaporation Rates Over Mixtures

When you good folks comment on these blogs I get a much better idea of what you want to hear about in the world of exposure modeling.  I received a number of comments like the one below from anonymous. 

“I would like to see more on vapor generation rates. Calculating this from a chemical's vapor pressure, liquid:vapor equilibrium and evaporation rate can be complicated for chemical (solvent) mixtures. Is there an easier way?”

Indeed, pure substances are relatively rare and we are almost always dealing with mixtures.   Anonymous is correct, the situation is somewhat complicated but I hope to present it in a form that is reasonably easy to follow  over the span of the next few blogs. 

Anonymous is also correct that one of the keys to determining evaporation rate is a chemical’s vapor pressure over and out of the mixture.  So at this point we are going to concentrate on determining the vapor pressure of various compounds over a liquid mixture.  

As an example let’s consider a mixture of toluene, xylene and benzene.   In this particular mixture 50% of the molecules are xylene, 49% of the molecules are toluene and 1% of the molecules are benzene.   This means that the mole fraction of benzene in this mixture is 0.01.  Mole fraction is the number of moles of benzene over the total number of moles in the mixture. Just to review a mole is 6.023 x 1023 molecules.  The molecular weight (MW) of a compound is the weight of one mole of its molecules.   For benzene its MW is 78.11 g/mole.   So if we know the contributing weight and the MW of all the components we can calculate the mole fraction of any component by adding up all the moles of the compounds and dividing this number into the individual moles of each.  I will present an example later in this blog.

So here we have a mixture in which 1% of the molecules (or moles) are benzene.  Presummably, the benzene is evaporating from the mixture but it is NOT evaporating at nearly the rate that pure benzene would evaporate.  In 1882 some smart guy by the name of Fran├žois-Marie Raoult came up with the idea that the vapor pressure of benzene over a mixture will be directly proportional to its mole fraction within the mixture.   Raoult’s Law states:


VPM = vapor pressure of the compound of interest over the mixture
VPP = vapor pressure of pure compound
MF =  mole fraction of the compound

If Raoult was correct then the vapor pressure of benzene over our mixture will be 1% of the vapor pressure of pure benzene.   The room temperature vapor pressure of pure benzene is 95 mmHg.  Thus, Raoult’s Law would predict 0.95 mmHg vapor of benzene over this mixture.   As it turns out Raoult was essentially correct for this mixture and for any mixture in which the compounds within the mixture were reasonably similar in structure. These are called ideal solutions.  In this case all these compounds are aromatics or benzene derivatives and so it works as pretty close to ideal. The vapor pressures of the xylene and toluene over this mixture are about half of their vapor pressures as pure materials.

Wow, life should be so simple but it’s not.   As soon as some of the major components of the mixture become very different from one another relative to a chemical structure standpoint then Raoult’s Law starts to break down.  In some instances it breaks down dramatically. 

Consider benzene in water.   Benzene has a water solubility of about 1.7g/Liter or about 1.7/78.11 = 0.022 moles in a liter of water which is 55.5 moles of H2O.   Thus the mole fraction of benzene in water at its solubility limit is 0.022/(55.5 + 0.022) = 0.0004.    Raoult’s Law would predict:   VPM  =  (95 mmHg) (0.0004) = 0.038 mmHg.   In reality the VPM of benzene as it approaches the solubility limit in water is essentially as high as the vapor pressure of pure benzene!     Raoult missed by a factor of about 95/0.038 or >2000!

How could Raoult get it so wrong?   The answer lies in the thermodynamic “activity” of the benzene molecule within this aqueous mixture.   In a mixture with its similar “bothers” (e.g., toluene and xylene) the benzene has more or less neutral activity and tends to obey Raoult’s Law.   In water, however, it is so different than the solvent that it is somewhat freaked out and highly active.   It does not want to be in the water and really wants out; thus, it expresses its vapor pressure at a relatively high level. 

High activity relative to the ideal Raoult prediction is not just for organics out of water which is perhaps one of the more extreme examples.   However it can also happen, for example, if you mix dissimilar organics like straight chained or brained hydrocarbons and lower alcohols.   Also intermediate deviations (AC = 2 to 20) from Raoult’s Law will also occur in some mixtures.  As such, we definitely need (and have) a modification and correcting generalization of Raoult’s Law:

VPM = (VPP) (MF) (AC)

VPM = vapor pressure of the compound of interest over the mixture
VPP = vapor pressure of pure compound
MF  = mole fraction of the compound in the mixture
AC = activity coefficient

The activity coefficient (AC) is about 1.0 when the molecule is neutral as the are in ideal mixtures and thus obeys Raoult’s Law.  As we have seen above, it can be very large (e.g. 2000 or more) when the molecule is very different and really wants out of the mixture.  

Getting the AC can be problematic but I will discuss how one might get it in the next blog.  I will also discuss how one can estimate vapor pressure of organics out of water in another way.  

Monday, September 9, 2013

The Most Versatile and Well-Tested Inhalation Model

The standard box model is simple and relatively easy to understand but it has some problems.   To understand the standard box model all you have to do is consider a box of air and keep track of what goes into it and what comes out.   Typically the “box” is a room and what goes into it’s air volume is the airborne contaminant; that is, the emission or generation rate.   The other stuff that goes into it (and out of it) is air from outside the box; that is, the ventilation rate.

Like I said it is pretty simple, just keep track of what goes in and what comes out and you can estimate the amount that is in the box.     C = (amount in – amount out)/box volume.   Indeed, we have equations that will estimate C for any point-in-time or as time-weighted averages.    What could be easier? However, the careful reader will notice that the airborne concentration (C) in this equation is an AVERAGE within the box.   Averages are good except when there is a lot of variation.   Indeed, if you are boiling hot on one side of your body while freezing on the other it could render an average of 75F but that average does not mean much to your comfort level!

We have all seen high breathing zone concentrations near point or concentrated emission sources within a room that are not representative of the average concentration within the entire room.   Consider a solvent spill in a large garage.   Folks cleaning it up can get a lot of exposure while those in the far corner of the room may be exposed to essentially nothing in the time it takes to clean it.   

The model works great when the source is relatively large within the room or there are a lot of sources spread more or less evenly throughout the box.   Also if there is very good mixing of the source(s) within the box as might occur with fans, then the average concentration determined by the model is reasonably accurate and useful.   As such we now called it the Well Mixed Box Model (WMB) and it is available in IH MOD.

This basic problem of the WMB model inability to handle near field sources plagued modelers for many years and then Mark Nicas and other workers developed what has come to be known as the 2 zone (or Near Field(NF)/Far Field (FF) model.    In this model there is a far field (FF) or outer zone which is typically the room volume and a near field (NF) or inner zone which contains the source and the breathing zone of the exposed person.   The exact origin of this model goes back a while.  As best as we can determine, the original concept for the 2 zone model NF/FF model was first put forth by W.C.L. Hemeon in the 1950s.  In a paper on general ventilation and the limitations of the WMB model, Hemeon outlined the basic concepts of the 2 zone model.  In 1996, both Mark Nicas and Ed Furtaw published papers in separate journals further developing the concept.  It was Mark Nicas who derived the dynamic concentration equations for the model.   Also he has continuously demonstrated and promoted its utility in industrial hygiene during the last 17 years.   

So what do you need to run the model?  In addition to the typical emission rate, you need a measure of the random air speed within a room along with the room’s volume and the general ventilation rate within the room.   The equations are relatively complicated but easily handled using the IH MOD freeware spreadsheet (as described in the previous blog).  

A critical parameter for the 2 zone model is the size of the near field (NF). In general, it should be large enough to include both the breathing zone of the exposed person and the emitting source. In the case of hair spray (or similar cosmetic spray applications) it has been estimated to be less than 1 m3 in volume.  In recent work designed to estimate exposure from an evaporating spill in a laboratory or plant, it has been set at a volume of 25 m3 – the volume of a 2 m diameter, 2 m high hemisphere centered over the spill.

This model allows one to estimate the exposure to people close to the source (NF) but also allows for an estimated exposure for persons and in the same room but not close to the source (FF).

So how good is the model at predicting breathing zone concentrations near sources?   As it turns out, it is quite good and its performance is well documented in a recent paper which lays out the capabilities of the model relative to a legal standard as set out by the “Daubert” decision.   Ref: Jayjock MA,  TW Armstrong and M Taylor:  The Daubert Standard as Applied to Exposure Assessment Modeling Using the Two-Zone (NF/FF) Model Estimation of Indoor Air Breathing Zone Concentration as an Example,  Journal of Occupational and Environmental Hygiene, 8: D114–D122, ISSN: 1545-9624 print / 1545-9632 online, November 2011.

One of the supported conclusions of this work was that the NF/FF model predictions were usually within the range of 0.5- to 2-fold of the measured concentration.  I will send you a copy of this paper if you write to me at

For all of these reasons I consider this model to be the most versatile and well vetted tool of its type available.

Monday, September 2, 2013

Modeling Math made Easy or at least Easier

Exposure models involve mathematical calculations and that is often a problem for those of us prone to make math errors.  When I first started working with exposure models calculators were plentiful but personal computers were just getting started.    I would run through a series of equations on a calculator and then to check myself I would run it again.   Often I got a different answer!    Clearly I was making simple math errors and just as clearly it was very frustrating.    Enter the IBM Compatible PC and an early programming language, BASICA.   Now I could invest the time in a simple program that could automatically run a long series of calculations and render consistent results.   Once I invested the time to program and check the answer(s), I had it made.

Enter, Lotus 1-2-3 and later Microsoft Excel, now we could do the same thing in a spreadsheet program which was considerably easier.   It also provided some really nice graphing capabilities.    Now we really had it made!   Everyone pretty much knew how to operate a spreadsheet and we could much more easily spread the word about modeling in our classes.

Flash forward a number of years and enter some remarkably talented and dedicated colleagues like Daniel Drolet and Tom Armstrong.    Tom is a very capable modeler and Daniel is a technology artist who is truly gifted in his ability to program model into Visual Basic.    Together with some of our other colleagues they developed IH MOD which is a compilation of inhalation concentration models and modeling tools.    Just about every important model of this type is included in this super spreadsheet.  Input fields are clearly noted and graphics and output columns are built-in.     Most important, it is an evergreen tool which has undergone numerous revisions and will undergo many more in the years to come as it gets better and better.   It is available in multiple languages and it is offered free of charge.   In my opinion it is the single most important tool ever developed in the history of occupational exposure modeling.

There are currently 12 models (modules) in IH MOD.   I am not going to list them all here but will recount some of the most important or most used models, some of which should look familiar if you have been reading this blog:

  • Well Mixed Box Model with Constant Emission Rate
  • Well Mixed Box Model with Exponentially Decreasing Emission Rate
  • Well Mixed Box Model with Backpressure
  • Well Mixed  Box with Purging (decay) from ventilation
  • The Two Zone Model: Near field Far field Constant Emission Rate
  • The Two Zone Model: Near field Far field Decreasing Mass Emission Rate
  • Estimating Contaminant Evaporation Emission Rate from small spills
The IH MOD super spreadsheet (along with other tools) is available online for download from a link at:

The documentation for all of the models presented in IH MOD is available in the publication:  Mathematical Models for Estimating Occupational Exposure to Chemicals, 2nd Edition
Edited by Charles B. Keil, Catherine E. Simmons, and T. Renee Anthony.

All of this is for the estimation of airborne concentration for inhalation modeling but there is also another super spreadsheet dedicated to dermal exposure.   Its name is IHSkinPerm which will be the subject of a future blog.

Also covered in future blogs will be the evaporation estimating module mentioned above and the rest of the modules in IHMOD.