1. Chapter 6: Air-Organic Solvent and Air-Water Partitioning in other words Henry’s Law equilibrium partitioning between air and water

2. Air Water Octanol A gas is a gas is a gas T, P Fresh, salt, ground, pore T, salinity, cosolvents NOM, biological lipids, other solvents T, chemical composition Pure Phase (l) or (s) Ideal behavior PoLPoL C sat w C sat o K H = P o L /C sat w K oa KHKH K ow = C sat o /C sat w K ow K oa = C sat o /P o L

3. Ranges of K H

4. Partitioning between air and any solvent (recall that in an ideal solution,  = 1) If  is constant, even close to solubility, then: units of pressure over mole fraction (no one uses) units of pressure over molar conc Pa-m 3 /mol or Pa-L/mol “dimensionless” units or L water /L air

5. VP/solubility if activity coefficients do not change, even as the chemical approaches saturation, then Henry’s law may be estimated as the compounds vapor pressure divided by its aqueous solubility this is, I think, a useful concept that has been lost in the new edition of the text. If a compound has both a low VP and a low solubility, it can be difficult to judge what its HLC will be.

6. VP ranges over 10 12 solubility ranges over 10 12 HLC ranges over 10 7

7. Factors influencing HLC Temperature Salinity Cosolvents

8. Temperature dependance of HLC  H “Henry” =  H vaporization minus the excess enthalpy of solubilization When solvent is similar to solute,  H E may be negligible water Pure liquid air HEHE  vap H  H “Henry” Note: you can use any units for K aw in this equation except dim’less units for K aw in this equation must be pressure-L/mol (and must match R) If you want to use dim’less units, use this form of the equation

9.  If you can’t find H E, then just use  H vap

10. Effect of salinity and cosolvents on HLC Salinity will increase HLC by decreasing the solubility (increasing the activity coefficient) of the solute in water. Account for salinity effects via Setschenow constant: Cosolvents will decrease HLC by increasing the solubility (decreasing the activity coefficient) of the solute in water. Account for cosolvent effects via:  i c is the cosolvent term, which depends on the identity of both the cosolvent and solute f v is the volume fraction of cosolvent

11. LFERs relating partition constant in different air-solvent systems Once again, partitioning depends on size, polarity/polarizability, and H-bonding IF these interactions are similar in both solvents, then a simple LFER is sufficient:

12. A familiar estimation technique Note that this is a generic equation for estimating the partition of a compound between air and any solvent. It is similar to the equation we used to estimate vapor pressure and solubility, but is slightly less complicated molar volume describes vdW forces refractive index describes polarity additional polarizability term H-bonding

13. Table 6.2

14. For water: That darn cavity term is back!

15. Measurement of Henry’s Law Relatively few measured values available. Hard to measure when solubility is low. Two approaches: static and dynamic

16. Static determination Static equilibration between air and water in a vessel such as a gas-tight syringe See problem 6.5

17. Dynamic determination batch air or gas stripping first must generate an aqueous solution containing a relatively high concentration of analyte first order process: where G = volume of gas V w = volume of water

18. Henry’s Law Constants of Polychlorinated Biphenyl Congeners and Their Variation with Temperature Holly A. Bamford, † Dianne L. Poster, ‡ and Joel E. Baker*,† paper

19. HLC of all 209 PCB congeners HLC (and VP, too) are lowest for congeners with few ortho chlorines

20. Does HLC go up or down with MW? Bamford shows HLC going up with MW. Solubility and VP both go down with MW. Which goes down faster? VP pretty well known, soly much less certain Bond contribution methods such as Hine and Mookerjee, Meylan and Howard suggest Kh goes down as MW goes up (based on limited data) Brunner (the only other good experimental data set) shows Kh goes down as MW goes up other VP/soly estimates show that Kh goes down as MW goes up for congeners with same number of ortho chlorines.

22. Old vs. new measurements

23.  S vs  H for air-water eqbm

24. Enthalpy questions Enthalpy of HLC is bigger than  H vap for many congeners. Environmental data shows that the slope of logP vs. 1/T plots increases as MW of PCBs increases. This is true at sites over land AND near water (such as Sandy Hook). If HLC relative to other temperatures is wrong, how can HLC at any given temp be correct?

25. Entropy questions  S of melting and vaporization are pretty much constant at 56.5 and 88 J/molK, respectively.  S of solubilization rises slowly with MW (from ~47 to ~62 J/molK) for PAHs Holly’s  S range from 49 J/molK (reasonable) to 500 J/molK (not reasonable!) –How is an entropy change of 500 J/molK possible? –  S fus for water = 22 J/molK Holly’s  S are a strong function of  H (R 2 = 0.999)

26. Estimation technique Vapor Pressure/Solubility how good is either?

27. Estimation Technique: Bond contribution methods In the absence of any other info, QSAR methods give good approximation. Hine and Mookerjee 1975 –bond contribution method –292 compounds Nirmalakhadan and Speece, 1988 –connectivity indexes –same data set as H&M but excludes amines, ethers, aldehydes & ketones –good to within a factor of 1.8 for most compounds Meylan and Howard 1991 –bigger data set (345 compounds) –also good to within 1.8 Pitfalls –How good are the calibration data? Measured or estimated from VP/soly? –Human error? –How big is the data set?

28. K H from fragment constants: structure-property relationships structure-property relationships used to predict many things  specific structural units increase or decrease and compound's K H by about the same amount. K H estimation method: where f are factors for structural units, and F are correction factors for affects such as polyhalogenation, etc. Note: factors for fragments attached to aliphatic carbons (C-H) are not the same as those attached to aromatic carbons (C ar -H) Example: C-Cl = -0.30 C ar -Cl = +0.14

29. table 6.4 Benzene biphenyl aliphatic alcohols

30. Examples: hexane: log K iaw (n-hexane) = 14(C-H) + 5(C-C) + 0.75 0.75 is the correction factor for a linear or branched alkane log K iaw (n-hexane) = 14*0.1197 + 5*-0.1163 + 0.75 = 1.84 experimental value is 1.81 benzene: logK iaw (benzene) = 6(C ar -H) + 6(C ar -C ar ) logK iaw (benzene) = 6(0.1543) + 6(-0.2638) = -0.66 experimental value is –0.68

31. Example: PCBs by M&H method Calibration set includes 12 halogenated benzenes: mean error = 21% and 3 PCBs error = 47% (is this good enough?) Validation set includes some PCBs and chlorobenzenes, they are predicted OK. Best to start with a known compound: –4-CBP logKh = -0.63 2-CBP log Kh = -0.09 –subtract C ar -H = -0.1543 –add one C ar -Cl = +0.0241 –result = -0.76 (err = 7%)-0.22 (err = 78%) –measured: 4,4’ CBP = -0.79; 2,5 CBP = -0.47 Cl in the 2 position has a large effect on Kh. These estimation methods cannot account for that.

32. Other properties can be used to predict HLC works best when compounds are closely structurally related.

33. PAHs

34. PCBs- chlorine number

35. Problem 6.3 1,1,1-TCA C air = 0.9 mg/m 3 C water = 2.5 mg/m 3 Is this compound volatilizing from, or absorbing into, the arctic ocean at 0C and at 10C? Salinity = 0.35% o

36. homework Problems 6.5 and 6.1