DFT modeling of sorbate-sorbent interactions for

DFT modeling of sorbate-sorbent interactions for PCBs with natural organic matter and black carbon Preliminary Report Upon Completion of Time at the British University in Egypt Nathan L. Howell1†, Tamer Shoeib2,3 1

PhD Candidate, University of Houston, TX, USA Assistant Professor, British University of Egypt, El Sherouk City, Egypt 3 Research Fellow, Loughborough University, Leicestershire, United Kingdom † Report preparer 2

Abstract Density functional theory (DFT) modeling is used to examine the mechanisms and Gibbs free energies ( of three specific PCB congeners (11,126,209) with model sorbates of natural organic matter (NOM) and black carbon. The DFT specifically employed is through Gaussian software and uses B3LYP with the split-valence basis set 6-311G**. The model sorbates for NOM are three humic acid moieties that focus on the following chemical structures and groups—aromatic rings, aliphatic chain, alcohols, phenols, and carboxylic acids. Black carbon sheets are approximated as coronene PAH. This purely theoretical study examines the ground state conformation geometries and associated thermochemical data of individual PCB congeners and model sorbents and the interactions between contaminant and sorbent. Not all results have been obtained and examined, but what has been examined reveals the following preliminary conclusions. Multiple conformations exist for each individual structure either because of actual geometric changes or because of protonation-deprotonation. DFT is successful at providing reasonable terms for the conformations that are consistent with chemical intuition. Sorbent-sorbate interactions are quantitatively examined according to free energy to assess the likelihood of their formation in vacuo in comparison to continued separation of sorbent and sorbate. Out of four available sorbent-sorbate results, only one seems to be energetically favorable. Current hypotheses for the favorability of this aromatic humic structure with PCB 11 are acid deprotonation, hydrogen bonding, and π-π electron donar-acceptor interactions. Recommended future activities are to perform due diligence on all modeling that has begun and to attempt to find ways to augment the modeling to it more realistic for the natural environment. These augmentations include the presence of water molecules and specific conductivity as the presence of ionic species. It is also recommended to model biphenyl contaminants of bromine and fluorine as well as PCB so that the work has broader application.

Density functional theory (DFT) modeling of sorbate-sorbent interactions for polychlorinated biphenyls, Howell, N.L., Shoeib, T. Revision 1, June 15, 2010

Introduction The importance of pollutant adsorption to soils, sediment, and air particulates cannot be overstated. The sorbent particles themselves may be classified in many ways according to, for example, method of generation (geologic, biogenic, anthropogenic) or, in aquatic environments, by their origin in relation to the hydrosphere (autochthonous or allochthonous) . The chemical nature of the pollutant is important in terms of its relative hydrophobicity/hydrophilicity. Hydrophilic pollutants such as ionized metals tend to exist in dissolved state in the water and thus are often deemed more “mobile”. Contrastingly, hydrophobic pollutants often exist in hydrophobic “pockets” such as lipid tissue, at various phase interfaces, and within and on the surface of particles. From these hydrophobic microenvironments, the pollutant is sequestered to varying degrees depending on the particular chemical-physical interaction that maintains the sequestration, and over time the pollutant can gradually leach into the dissolved water phase as high as the solubility limit. Environmental risk assessment would sometimes prefer to keep contaminants sequestered and immobilized so that the material holding the contaminant can be removed. This strategy is often the case in the soil subsurface. In such a situation, conditions that may force mobilization such as pH or conductivity changes are mitigated. Some contaminated sites would be better served by the mobilization of the contaminant. Such is often the case in contaminated sediments in a hydrodynamic flow environment (e.g. rivers, estuaries, some lakes) because the rate of advection is high enough to flush out the pollution in a shorter period of time. The key difference between the subsurface example and the riverine example is the velocity. If velocity is high, it is often preferable to move the contaminant to a dissolved state where it can be diluted, flushed, and volatilized removing the risk. If the velocity is low, as in groundwater (on the order of 0.1 ft/day), sequestration either natural or through human intervention may be the better decision. The type of environment in view in this work is the high flow environment of a river or estuary in the presence of cohesive sediment. Cohesive and non-cohesive sediments are usually treated separately from a physical perspective because the sediment motion of cohesive sediment is different. Cohesive sediments do not maintain a constant grain size due to flocculation, and they often require greater shear stress at a channel bottom to initiate motion than what would be expected for their size and mass because of the cohesive strength between the grains. This distinction would be irrelevant from an environmental contaminant perspective if not for two important consequences both related to adsorption onto the grains. If a contaminant is adsorbed onto any type of sediment and can remain there for even a few days, then the advection of the contaminant throughout the water is primarily determined by advection of the sediment. As cohesive sediment motion can be significantly different from non-cohesive sediment, understanding the distinction between sediment types is critical for contaminant transport. Secondly, cohesive sediments are cohesive primarily due to differences in mineralogy from that of non-cohesives. Those mineralogic differences ultimately create differences in physical behavior due to electrostatic and organic material interactions between the cohesive sediment grains. These interactions make the sediment cohesive, but they also can influence the strength and quantity of contaminant sorption, especially organic material. S.W. Karickhoff is often credited with some of the first experimental work to illustrate the importance of organic carbon (OC) in controlling the transport of pollutant in soils and sediment (Karickhoff, 1979, Page 2 of 68


1981). Karickhoff and other researchers eventually reduced much of contaminant fate and transport understanding between solid, air, and water interfaces down to a single factor, the octanol-water partition coefficient (Kow). Any environmental contaminant of any consequence has had its Kow fairly well determined through experiment, and that value, usually given as log Kow, can be related to what is ultimately sought after in understanding and modeling environmental systems with regard to sorption, the linear water-organic carbon partition coefficient (Koc). With an accurate value for this coefficient, it should be possible to predict what the equilibrium distribution of a contaminant is between a soil, sediment, or some other environmental solid and water using only the OC content of the solid. That has been done most often in the literature through bench-scale experiments that provide an adsorption isotherm similar to idealized isotherms given in Figure 1.

Figure 1. Idealized adsorption isotherms. Taken from Goldberg et al. (2007) What the figure shows are three idealized isotherms: Distribution Coefficient or Simple Linear, Freundlich, and Langmuir. If one applies a partition (distribution) coefficient indiscriminately for a single contaminant, then the implicit assumption is the use of the Distribution Coefficient model. As with any model, this model has its limitations. Bench studies of sorption with hydrophobic contaminants for a large range of dissolved concentrations show what would be a varying Koc (Di Toro and Horzempa, 1982; Horzempa and Di Toro, 1983). If one took a tangent of the Freundlich or Langmuir isotherms, a different Koc would be obtained at every location. In environmental media, the water concentration is usually extremely low (ppb or ppt), and so it is assumed that one is in the lower segment of a system more resembling a non-linear isotherm. In that lower concentration segment, the range of concentration is also low. Thus, a linear assumption is more justifiable. Another complication with the linear approach is the existence of different strengths of sorption for fractions of the contaminant. It is often found that when an adsorption curve is generated that the path of adsorption is not always the same as desorption, a hysteresis (Di Toro and Horzempa, 1982). To explain the phenomenon, the adsorption is conceptualized as sorbed contaminant fractions. One such conceptualization is to divide the solid concentration into that which is easily desorbed, a labile fraction, that which is resistantly sorbed, and that which is irreversibly sorbed. Another problem in characterizing environmental contaminant sorption is the heterogeneity of soils and sediments. Karickhoff and others have provided evidence that OC is a controlling agent for sorption, but more recent research has indicated that the amount of OC is not enough to quantify adsorption in every Page 3 of 68


environment. Using only the quantity of OC, a one parameter linear free energy relationship (LFER) model that requires only knowledge of the Kow has been successfully fit to many datasets. One notable example for polychlorinated biphenyls (PCBs) is that of Seth et al. (1999), (1) where Koc and Kow have been defined previously. These authors note that their equation fits a large of amount of field environmental data examined, but they qualified the fit by nothing that the log-log slope value could vary by a factor of 2.5 either direction. Even with that “rule of thumb” variation, only 80% of their data fell within the sweep of the linear isotherm. Seth et al. attribute significant differences in the one parameter LFER between them and other authors (some even using the same datasets) to “variability in the organic matter of the soils used”. Studies since that time have tried to provide multiple parameter LFER models to describe the adsorption of hydrophobic pollutants to soils and sediments. Other parameters in those models besides total organic carbon (TOC) include amorphous organic carbon (AOC) and black carbon (BC). Arp et al. (2009) reviewed much of this data up to the present time and found that TOC and AOC-BC models could be used to describe environmental data for hydrophobic organic pollutants (HOPs), but in the end a large set of environmental data was still required. It would be preferable to be able to measure sediment, soil, and aquatic particle OC in some fashion without the need for a large number of data points sufficient to generate a portion of an adsorption isotherm. Field-based environmental sorption in surface water is still not well developed enough in theory and in practice to provide a simpler and easier characterization of sorption. Figure 2 presents the study area in view for this work as the Houston Ship Channel (HSC), a microtidal estuarine river that, as a saline influenced body (> 0.5 ppthou salinity), runs approximately 50 km from downtown Houston to the mouth of the river that empties into Upper Galveston Bay (UGB). The river is generally 2-6 meters deep in most places except in the navigationally dredged channel that is anywhere from 13-18 meters deep. It is industrially impacted by many historical and current contaminants with much pollution from PCB, the contaminant of concern in this study, suspected to be sourced from two industrial sites shown in the map.

Page 4 of 68


Figure 2. The Houston Ship Channel (HSC) in Houston, TX, USA. Red markers signify distance in kilometers downstream from Buffalo Bayou at Shepherd drive in central Houston. A plot of some water column based dissolved and suspended PCB concentrations (collected in the summer of 2009 from the HSC) for PCB 11 (3,3’-DiCB), PCB 126 (3,3',4,4',5-PentaCB), and PCB 209 (DecaCB) is given in Figure 3. The data shown does not actually present concentrations of PCB as “truly dissolved” but rather what is “operationally dissolved”. The method of sample collection makes an artificial demarcation for suspended to dissolved based on a 1-um filter. Nonetheless, there are two major significant trends in the data. One is that a line of best fit (albeit with a low coefficient of determination) would yield mass-based partition coefficients (Kd) in units of L/Kg that are very different for different congeners. One would expect that as chlorination and hydrophobicity increases that Kd, the apparent dry weight (L/Kg) partition coefficient, would also increase, and this expectation is supported in the data. The second trend seen is that any line of best fit, if plotted with 95% confidence intervals or some other form of uncertainty, would have large range in the slope and thus in the K d. This data does not yield a Koc (because particulate organic carbon (POC) data were not able to be obtained), but even if it were the general quantity of variation in that data can be supposed (based on variation in bed sediment TOC) to be insufficient to collapse the scatter onto a single line or thin linear region. Of course the trends here are only acceptable to some degree due to the presence of < 1 um colloidal particles included in the “dissolved” concentration, which further obfuscates the estimation of Kd or Koc (Butcher et al., 1998). One would need to be more rigorous in sample collection and chemical analysis to ascertain the truly dissolved and suspended sediment concentrations. This data also does not include any reference to the PCB partitioning that occurs in bed sediments as no pore water data to match with bed sediment concentration has been collected.

Page 5 of 68

40

PCB 11

35

PCB 126

30

PCB 209

25 20 15 10 5

0 0.00

0.05

0.10

0.15

0.20

PCB Suspended (> 1 um) Concentration (ng PCB/g dry total suspended solid)

PCB Suspended (> 1 um) Concentration (ng PCB/g dry total suspended solid)


0.25

PCB Dissolved (< 1 um) Concentration (ng PCB/L)

0.50 PCB 11 PCB 126 PCB 209

0.45 0.40 0.35 0.30 0.25

0.20 0.15 0.10 0.05

0.00 0.000

0.002

0.004

0.006

PCB Dissolved (< 1 um) Concentration (ng PCB/L)

Figure 3. Suspended ( > 1 um) PCB concentrations for PCB 11, 126, and 209 from summer 2009 sampling in the HSC. The number of detectable concentration datapoints (out of 48 samples collected) in both media concurrently are 45, 7, and 40 for PCB 11, 126, and 209, respectively. One extremely high PCB 209 sample (1800 ng PCB/g dry TSS) was removed. The three PCB congeners give in the figure are chosen because they are particularly relevant to the HSC, and are given in simple structural diagram in Figure 4. PCB 11 is important globally because it is not part of the traditional Aroclor PCB mixtures (Frame et al., 1996) but rather is now linked to a novel source in diarylide yellow paint pigments (Rodenburg et al., 2009). It is present in the HSC, at different locations though the locations are not consistently high at all sampling events. PCB 126 is a dioxin-like congener with a high Toxic Equivalency Factor (TEF) from WHO2005, the highest of all PCB (0. 1). It is seen in samples of aquatic biota in the HSC but not in near as many abiotic examples, which defies sourcereceptor explanations to some degree (unpublished results from HSC). Lastly, PCB 209 is the highest chlorinated congener and is unusually high in some places in the HSC. Because it sorbs so strongly (highest Kow), there is some question about how long it will remain in the HSC if its sorptive strength provides only slow removal from the HSC.

Figure 4. Three PCB congeners to use in density functional theory (DFT) modeling.

Page 6 of 68


As mentioned previously, there are two main forms of OC that have been noted recently in the literature—AOC and BC. AOC is the OC portion of what is called natural organic matter (NOM), one major component of which is humic substances, the three most important of which are humic acids (HAs), fulvic acids, and humins shown elementally in Figure 5. A few example structures are given in Figure 6. While much research has focused on BC especially in explaining HOP and PCB adsorption, not all particles in surface water will contain a large percentage of BC. Nor is AOC likely to be inconsequential because BC is present.

Figure 5. Elemental composition of various humic substances from different source of organic matter. Image taken from Abdoul-Kassim and Simoneit (2001).

Page 7 of 68


Figure 6. Examples of proposed humic acid structures.

Page 8 of 68


It has already been shown to some degree that sorption has been examined extensively as BC, AOC-BC, or TOC by experiment at a lab or field scale. Efforts have also been made to explain the sorption interactions of HOPs through theoretical studies employing computational chemistry1. The first relevant study is one that examines the role of fulvic acids to bind inorganic colloidal particles together (Seijo et al., 2009). Brownian Dynamics Simulation was used for this. There was also a study of pesticide sorption to kaolinite, HAs, and lignin oligomers through the use of molecular dynamics (Yan and Bailey, 2000). Other studies attempted to understand the conformations of large scale NOM structures through molecular mechanics. Schulten and Schnitzer (1997) attempted to fully characterize the 3D structure of HAs on soil particles and with water interactions. Their results hinted at applications for contaminant sorption while Kubicki and Apitz (1999) went even further and modeled the interactions of some PAH with HA, fulvic acid, lignin, and soot particles. They were able to make a first order approximation of sorption energies between PAH and the sorbents that could indicate how much sorption is possible depending on the sorbent type and local environmental conditions. Other studies that involve NOM and 3D molecular mechanics modeling have also been conducted (Maurice and Namjesnik-Dejanovic, 1999; Schulten, 1999; Govers et al., 2003; Sutton et al., 2005; Alvarez-Puebla et al., 2006). It is noted that there are few studies that use ab initio/density functional theory (DFT) methods to understand hydrophobic contaminant sorption and none for PCB specifically. This theoretical approach of DFT is what will be used to examine potential interactions of PCB with AOC and BC. Specifically, the objectives of the efforts are 1. To identify some of the mechanisms, strength, and quantity of sorption for particular representative PCB congeners on particular representative AOC and BC sorbents. 2. To apply what is learned from DFT modeling to interpret current HSC PCB data related to particle sorption and to have it inform future sample collections.

1

Detailed structural and mechanistic chemical analysis has also been attempted to explain contaminant sorption though it is not as pertinent in relation to this work since it is theoretical. One good example is a specific examination of metals with humic acid using Fourier transform infrared spectroscopy (Ibrahim and El-Aal, 2008).

Page 9 of 68


Methods and Modeling The structures modeled from the contaminant sorbate class are PCB 11, 126, and 209 (Figure 4), and the sorbents are modeled only as pieces of a larger HA structure. The three HA moieties are given in Figure 7 and described as an aliphatic humic (ALHUMIC), an aromatic humic (ARHUMIC), and a simple humic (SIMPLEHUMIC) acid structures. Each represents a different kind of environment on a humic acid that a PCB congener could sorb to or desorb from, and nearly any functional group that exists in a humic acid is represented. The distinction was made between aliphatic and aromatic moieties because research by others (Keiluweit and Kleber, 2009) on PAHs, all of which also have multiple phenyl rings, has shown that the attraction of the two π electron systems of the sorbate and sorbent can be major way that sorption occurs. Also, the figure shows that ARHUMIC has a carboxylic acid functional group one of the phenyl rings. This provides one source of variation in modeling because it can be protonated or deprotonated. The pKa is not known exactly for this structure, but a similar structure, benzoic acid, has a pKa in water of 4.2. Given a pH of 7.7 typical of marine surface water, the percentage of deprotonation by such a carboxylic acid structure is estimated at 0.00082%. To reach a significant level of deprotonation of 5% would require a pH of 0.001. The lowest conceivable range of pH in natural surface water is around 6.5, which gives a deprotonation rate of 0.001%. Nonetheless, deprotonated ARHUMIC is examined. Coronene (Figure 8) was chosen to represent BC because it is relatively small but exists as a planar sheet of PAH, which is similar to what is known of BC (Smith and Chughtai, 1995). Coronene has also been used previously to examine BC-PAH sorption (Kubicki, 2006).

Figure 7. Chemical moieties taken from humic acid models. (A) A small aromatic structure (ARHUMIC). (B) An aliphatic structure (ALHUMIC). (C) A simple aromatic structure (SIMPLEHUMIC). HAR stands for humic acid remainder and indicates where the rest of the humic acid would connect to the moiety. In a real humic acid all three of the structures, or something similar to them, would exist in multiple instances.

Page 10 of 68


Figure 8. Coronene structure that will be used as a model for the multiple PAH sheets found in black carbon. It is important to note the source of the data that will be used in the modeling. Data on PCB concentration and sediment characteristics exists for three different large scale sampling events in the HSC. However, none of that data is being directly employed here. It was used to help decide on the representative PCB congeners to use in the study, but the modeling itself is being conducted as close to a “first principles” approach in the sense that ab initio normally implies. The actual DFT method used is Becke, three-parameter, Lee-Yang-Parr (B3LYP ) exchange-correlation functional, and the basis set used is the split-valence 6-311G**. Gaussian09 and Gaussian03 were used for all DFT calculations. This basis set is recommended by Gaussian for systems involving C, O, Cl, and Br. The general method employed is as follows. For any structure or pair of structures considered, an optimization was conducted to find energy minima ground states for as many conformations as could be conceived. Starting geometries were manipulated to discern multiple stable configurations. Ground states were confirmed and distinguished from transition states using the predicted frequency values for rotation, translational, and vibration within the structure. Single point energy, zero point energy, thermal correction (to 298.15 K) to internal energy, thermal correction to enthalphy (to 298.15 K), and enthalpy were determined. These values were used to derive a Gibbs free energy at 298.15 K ( ) for all structures based on the same reference state. All individual structures to be used were optimized for ground states and thermochemical data first. Then combinations of the individual structures were modeled being certain to find many different mechanisms of interaction between the various functional structures. Known ground states configurations of the individual molecules were used to try and find stable multiple structure states. Data analysis methods are still being developed. What will be and has been done to some extent is to compare the final structures according to their resultant . Thermochemical values are obtained from the Gaussian simulations for each ground state structure. These values can be combined to arrive at ΔG˚298.15 K according the equation,

(2)

where

is the single point energy of the electronic interactions only (protons and

electrons) at absolute zero (Kcal/mol),

is the zero point energy (Kcal/mol), Page 11 of 68

is the


thermal correction to take the electronic energy to 298.15 K (Kcal/mol), is the thermal correction required to bring to an enthalpy value at 298.15 K (Kcal/mol) , and is the entropy (Kcal/mol-K). can be used to assess the likelihood of particular sorption interactions occurring, to ascertain what interactions between two structures are most favorable, to make inferences about the relative strength of an interaction between a congener and the sorbent at the bulk scale, and to determine how various chemical variations (e.g. position of the halogen, identity of the halogen, number of halogens, bond angles, position of π electrons) influence adsorption.

Page 12 of 68


Work Accomplished The work accomplished consists almost entirely of Gaussian model runs and the collection of final geometries, bond lengths, and thermochemical information all of which is given in full detail in the Appendix. Far more model runs were planned and have begun than were able to be completed for this report. Table 1 gives a summary of the Gaussian model runs and their current status. There is no real distinction between the names in the table under than a test number, but the difference is in the starting geometry used for each run. Many initial positions were tried, and it is not certain if all runs will generate a convergent state be it either a transition or ground state. Images and configuration files of the original run geometry exist if needed, but as they will be obsolete as soon as the model run is complete. Thus, they are not included.

Page 13 of 68


Table 1. Gaussian model run summary. Some runs were complete in Gaussian09 and some in Gaussian03. The results should not differ. Molecules Involved

Run Type

Descriptor

Run Status

Frequency Analysis Performed

Aliphatic Humic

Individual Conformer

Linear Chain

Complete

Yes

Aromatic Humic (prot) Individual Conformer

Protonated

Complete

Yes


Flat Sheet

Complete

Yes

PBB 11


"Cis" Structure

Complete

Yes

1

PBB 126


Slight Ring Rotation

Complete

Yes

1

1

PBB 209


Antiplanar

Complete

Yes

PCB11test01

1

1

PCB 11


"Cis" Structure

Complete

Yes

8

PCB11test02

2

1

PCB 11


"Trans" Structure

Complete

Yes

9

PCB126

1

1

PCB 126


Slight Ring Rotation

Complete

Yes

10

PCB209

1

1

PCB 209


Antiplanar

Complete

Yes

11

SIMPLEHUMIC

1

1

-

Awaiting Submission

No

ARHUMICtest02

2

1

Simple Humic Aromatic Humic (deprot)


12


Deprotonated

Complete

Yes

13

ALHUM-PCB11test01

1

2

Aliphatic Humic, PCB 11

Sorbate-Sorbent

Middle OH turned towards one phenly ring

Complete

Yes

14

ALHUM-PCB11test02

2

2


Sorbate-Sorbent

-

Permanently Killed (Bad Initial Geometry)

No

15

ALHUM-PCB11test03

3

2


Sorbate-Sorbent

-


No

16

ALHUM-PCB126test01

1

2


Sorbate-Sorbent

Suspected H-Bond Carbonyl to ortho H's

Complete

Yes

17

ALHUM-PCB126test02

2

2


Sorbate-Sorbent

-


No

18

ALHUM-PCB126test03

3

2


Sorbate-Sorbent

-


No

19

ALHUM-PCB209test01

1

2

Sorbate-Sorbent

-


No

Killed

Killed

Complete

Yes

Complete Convergence

Energy Minimum

Awaiting Submission

No

Awaiting Submission Awaiting Submission

Awaiting Submission

No


Run Count

Run Name

1

ALHUMICtest01

1

1

2

ARHUMICtest01

1

1

3

CORONENE

1

1

Coronene

4

PBB11

1

1

5

PBB126

1

6

PBB209

7

20

ARHUM-PCB11test01

Test No Molecule Count

1

2

21

ARHUM-PCB11test02

2

2

22

ARHUM-PCB11test03

3

2

Aliphatic Humic, PCB 209 Aromatic Humic (Prot), PCB 11 Aromatic Humic (Prot), PCB 11 Aromatic Humic (Deprot), PCB 11

Sorbate-Sorbent Sorbate-Sorbent Sorbate-Sorbent

Phenol group to pi electrons All phenyl rings matched, PCB 11 in "cis" All phenyl rings matched, PCB 11 in "cis"

Page 14 of 68

Convergence Status

Resultant State

Complete Energy Minimum Convergence Non-Convergent in Energy Minimum Max Diplacement Complete Energy Minimum Convergence Complete Energy Minimum Convergence Complete Energy Minimum Convergence Complete Energy Minimum Convergence Complete Energy Minimum Convergence Complete Energy Minimum Convergence Complete Energy Minimum Convergence Complete Energy Minimum Convergence Awaiting Submission Awaiting Submission Complete Energy Minimum Convergence Complete Energy Minimum Convergence Permanently Killed Permanently Killed (Bad Initial (Bad Initial Geometry) Geometry) Permanently Killed Permanently Killed (Bad Initial (Bad Initial Geometry) Geometry) Complete Energy Minimum Convergence Permanently Killed Permanently Killed (Bad Initial (Bad Initial Geometry) Geometry) Permanently Killed Permanently Killed (Bad Initial (Bad Initial Geometry) Geometry)



Molecules Involved

Run Status


Awaiting Submission

No


Awaiting Submission

No


Complete

Yes

Complete Convergence

Energy Minimum

Currently Running

No

Currently Running

Currently Running

Currently Running

No

Currently Running

Currently Running

Currently Running

No

Currently Running

Currently Running

Currently Running

No

Currently Running

Currently Running

Currently Running

No

Currently Running

Currently Running

Awaiting Submission

No


Awaiting Submission

No


Awaiting Submission

No


Awaiting Submission

No


Awaiting Submission

No


Currently Running

No

Currently Running

Currently Running

Currently Running

No

Currently Running

Currently Running

PCB 11 "cis" and right angle to coronene with chlorines out

Currently Running

No

Currently Running

Currently Running

Sorbate-Sorbent

PCB 11 "cis" and right angle to coronene on the side of coro

Currently Running

No

Currently Running

Currently Running

Sorbate-Sorbent

PCB 11 antiplanar and above coronene rings with one chlorine in

Currently Running

No

Currently Running

Currently Running

Run Count

Run Name

23

ARHUM-PCB11test04

4

2

24

ARHUM-PCB11test05

5

2

25

ARHUM-PCB11test06

6

2

Aromatic Humic (Deprot), PCB 11

Sorbate-Sorbent

26

ARHUM-PCB11test07

7

2

Aromatic Humic (Prot), PCB 11

Sorbate-Sorbent

27

ARHUM-PCB11test08

8

2


Sorbate-Sorbent

28

ARHUM-PCB11test09

9

2


Sorbate-Sorbent

29

ARHUM-PCB11test10

10

2


Sorbate-Sorbent

30

ARHUM-PCB11test11

11

2


Sorbate-Sorbent

31

ARHUM-PCB126test01

1

2


Sorbate-Sorbent

32

ARHUM-PCB126test02

2

2

33

ARHUM-PCB126test03

3

2

34

CORO-PCB11test01

4

2

Coronene, PCB 11

Sorbate-Sorbent

35

CORO-PCB11test02

5

2

Coronene, PCB 11

Sorbate-Sorbent

36

CORO-PCB11test03

6

2

Coronene, PCB 11

Sorbate-Sorbent

37

CORO-PCB11test04

7

2

Coronene, PCB 11

Sorbate-Sorbent

38

CORO-PCB11test05

8

2

Coronene, PCB 11

Sorbate-Sorbent

39

CORO-PCB11test06

9

2

Coronene, PCB 11

40

CORO-PCB11test07

10

2

Coronene, PCB 11

Aromatic Humic (Prot), PCB 11 Aromatic Humic (Deprot), PCB 11

Aromatic Humic (Prot), PCB 11 Aromatic Humic (Deprot), PCB 126

Run Type

Descriptor

Sorbate-Sorbent Sorbate-Sorbent

Sorbate-Sorbent Sorbate-Sorbent

PCB 11 in "cis", chlorines turned towards phenyl rings PCB 11 in "cis", chlorines turned towards phenyl rings PCB 11 antiplanar, one ring matched with coplanar ARHUMIC PCB 11 antiplanar, one ring matched with coplanar ARHUMIC PCB 11 as "cis" but rings at slight twist, ARHUMIC twist to match PCB 11 as "cis" but rings at slight twist, ARHUMIC twist to match ARHUMIC coplanar, PCB 11 anti with Cl turned toward carboxyl ring ARHUMIC coplanar, PCB 11 anti with Cl turned toward carboxyl ring ARHUMIC coplanar, PCB 126 slight twist with one ring aimed at OH end of ARHUM Both coplanar with open rings matched together Both coplanar with open rings matched together PCB 11 "cis" at right angle to coronene at distance PCB 11 "cis" and rings alligned with coronene PCB 11 "cis" and right angle to coronene with chlorines in PCB 11 "trans" and right angle to coronene with one chlorine in

Page 15 of 68

Convergence Status

Resultant State


Run Count

Run Name


Molecules Involved

Run Type

41

CORO-PCB11test08

11

42

CORO-PCB11test09

43

CORO-PCB11test10

Descriptor

2

Coronene, PCB 11

Sorbate-Sorbent

12

2

Coronene, PCB 11

Sorbate-Sorbent

13

2

Coronene, PCB 11

Sorbate-Sorbent

PCB 11 antiplanar and above coronene rings with one chlorine out PCB 11 "cis" and in same plane as coronene on it side (Cl in towards coronene) PCB 11 "cis" and in same plane as coronene on it side (Cl away from coronene)

Page 16 of 68

Run Status


Convergence Status

Resultant State

Currently Running

No

Currently Running

Currently Running

Awaiting Submission

No


Awaiting Submission

No



Results and Data Analysis Complete raw results of all converged ground state structures are given in the Appendix. This section contains some early preliminary analysis. More comprehensive analysis will be conducted when more model runs are complete.

Individual Structure Results values for each individual molecule model run that was completed are given in Figure 9. Some trends are evident in this data although nothing can be too conclusive about much of this analysis for two reasons. The first is that not all conformations of individual structures are yet known, and the knowledge of what those conformations are can affect the ranking of one molecule against another as well as how each conformation is viewed in light of all other possible conformations. The second reason for caution is that there is far more model data that needs to be examined for each stable configuration besides the thermochemical data alone. Examination of the rest of the model data (e.g. bond lengths, bond frequencies, inter-atomic forces) will proceed at a later time. The first obvious trend is the amount of atoms in the structure as compared with the free energy. All PCBs and PBBs have the same number of atom centers (22), and yet there are wide variations in . It would appear then that the thermodynamic free energy of a structure is more influenced by the other structural elements such as atomic distances, atomic number, bond lengths, and bond angles. It is also seen that all of the sorbent structures have a higher energy state than the sorbate structures. The aromatic humic structure is very similar in many respects to PCBs and PBBs. It is reasonable then to suggest that the presence of halogens provides more stability to the biphenyl structure than alcohols, carboxylic acids, and ethers. Lastly, the figure indicates the generally lower energy geometries are formed with the heavier bromine substitution over chlorine. PCB 11 is the sister congener to PBB 11 in the brominated class of contaminants. The conformers of PCB11test1 and PBBtest1 are visually identical in terms of pertinent bond angles and ring rotations. The main difference is the identity of the halogen. It is likely that the presence of a more electronegative ring substitute, bromine over chlorine, increases overall stability. Such stability is most contextually relevant in this study to conformers and adsorption, but it cannot be missed that the increased stability would likely correspond to greater recalcitrance in the environment as well.

Page 17 of 68

ΔG˚298 (Mcal/mol)

PBB209 (22)

PBB126 (22)

PBB11 (22)

PCB209 (22)

PCB126 (22)

PCB11test2 (22)

PCB11test1 (22)

ARHUMICtest1 (35)

ARHUMICtest2 (34)

CORONENE (36)

ALHUMICtest1 (29)


0 -2,000 -4,000 -6,000 -8,000 -10,000 -12,000 -14,000 -16,000 -18,000

Figure 9. Gibbs free energy of formation from reference state at 298.15 K for individual molecule geometries gathered from Gaussian. Energy is ordered left to right from highest to lowest energy. Numbers in parentheses indicate the number of atoms in each structure. Figure 10 and Figure 11 compare pairs of stable conformers for PCB 11 and aromatic humic, respectively. The difference in conformation with PCB 11 is a result of the starting geometry given to Gaussian while the conformational difference in the aromatic humic is caused by the deprotonation of the carboxylic acid for ARHUMICtest2. The PCB 11 test1 and test 2 conformers can be analogized to “cis” and “trans” configurations, respectively, though in a technical sense those terms do not apply since the critical bond of separation between them is not a double bond. The trans conformer is lower in over the cis conformer by 0.44 Kcal/mol, which may be within the range of model uncertainty though it is still considered here to be a real difference in free energy. The trans conformer would logically be lower in energy because the repulsive forces between electronegative chlorines achieve the greater separation in PCB11test2 (9.0104 Angstroms) over PCB11test1 (7.6634 Angstroms). In the aromatic humic structures the difference in the structure is primarily in the phenyl rings. The protonated aromatic humic is more antiplanar while the deprotonated structure is nearly coplanar. The protonated anti-planar conformation is energetically favored over the deprotonated aromatic humic by 0.59 Kcal/mol.

Page 18 of 68


Figure 10. Two PCB 11 ground state conformers found through Gaussian modeling. Neither is completely coplanar. Test1 can bed loosely termed a "cis" conformer while Test2 would be a "trans" conformer.

Figure 11. Protonated and deprotonated forms of the aromatic humic acid moiety. Both are ground state geometries but are likely not the only conformer that exists for the molecule.

Page 19 of 68


Sorbent-Sorbate Structure Results Many more sorbent-sorbate modeling trials must be run to even begin to describe the mechanisms, strength, and amount of sites for sorption for PCB congeners. The few trials that have generated complete data are still informative even at this preliminary stage. Figure 12 presents the relative Gibbs free energy

, which is the sum of the individual

sorbent and sorbate modeled subtracted from the interaction geometries . If the value is negative it means that the free energy of the system in vacuo is lowered by the sorption interaction predicted by Gaussian. If it is higher upon interaction, then the thermodynamic direction for that one interaction is away from that interaction and towards either another interaction or separation of the sorbent and sorbate. Interestingly, the DFT modeling was able to so far find four energy minima for some sorbent-sorbate combinations, and yet only one of the minima was more favored compared with the individual separated structures. Highlighting the preliminary nature of this data, the figure does not indicate that the aliphatic humic-PCB11 and aliphatic humic-PCB126 interactions cannot be energetically favorable. It simply means that the two that have been obtained thus far do not, in this idealized model system, are not favored.

Figure 12. Relative Gibbs free energy ( difference (Modeled Structures Together - Sum of Individual Structures) for sorbent-sorbate interactions. Page 20 of 68


Rather than attempt to understand exactly why the energetically disfavored sorbent-sorbate interactions are so, it is the strategy to make inferences for why the two aromatic humic-PCB11 interactions differ in the sign of between the sum of sorbate and sorbent and the sorbentsorbate interaction. Figure 13 presents the two model trials in space. Two major differences seen are that the PCB 11 conformers are different. ARHUM-PCB11test1’s PCB 11 conformer is far more of the cis variety (PCB11test1) while that of ARHUM-PCB11test6 is closer to trans (PCB11test2). The second difference is in the aromatic humic acid moiety. ARHUM-PCB11test1 uses the protonated aromatic humic (ARHUMICtest1) while ARHUM-PCB11test6 use the deprotonated form (ARHUMICtest2). If one looks only at those two differences and assumes that the conformations of the sorbent and sorbate are not that different from the individual molecule conformations just referenced, then the from ARHUM-PCB11test6 to ARHUM-PCB11test6 is +0.44 Kcal/mol and -593.44 Kcal/mol for individual PCB 11 and aromatic humic conformers, respectively. The resultant sum (-593 Kcal/mol) is sufficient to explain the free energy differences in the two trials depicted in Figure 12. A direct comparison of test1 to test6 gives of -338.78 Kcal/mol in favor of test1, the interaction of positive in Figure 12. Thus, the sorbate-sorbent free energy is lower for test1 over test6, but the constituent sorbent in test6 (deprotonated aromatic humic) is even lower than test1 sorbate-sorbent. More definite conclusions about the value of protonation-deprotonation of the carboxylic acid cannot be made without more knowledge of other ground state configurations.

Page 21 of 68


Figure 13. Ground state configurations for ARHUM-PCB11test1 and test6. The geometries of PCB 11 are different between test 1 (cis) and test 6 (trans), and the aromatic humic structures are protonated (test1) and depronated (test6). A closer look into ARHUM-PCB11test6 shows one other point of interest, the role of π-π electron interactions between sorbate and sorbent. Several angles of view for ARHUM-PCB11test6 are shown in Figure 14 that allow the relative positions of all four rings to be examined. The role of aromatic π electrons in the sorption of PAHs has been a subject of interest amongst researchers of late (Keiluweit and Kleber, 2009), but PCBs have not been so directly on the topic. Zhu et al. (2004) studied π-π electron acceptor-donor (EDA) relationships of PAHs with model humic acids and discovered that those interactions are characterized by conformations that are parallel-planar but slightly offset in the π donar and acceptor systems. The side views of Figure 14 are at first glance similar to this description. Far more analysis would need to be conducted to attribute the favorable interaction observed to π-π EDA, and such interactions have been shown to have energies only as high as 4.06 Kcal/mol for PAHs (Meyer et al., 2003). π-π EDA may explain part of the sorption potential in this interaction and others that may later be found, but other attractive forces such as hydrogen bonds need also be considered.

Page 22 of 68


Figure 14. Different views of the ARHUM-PCB11test6 interaction. The orientations of the pairs of phenyl rings with each other may suggest a π- π electron donor-acceptor mechanism to explain at least part of the thermodynamic favorability of the structure.

Page 23 of 68


Conclusions and Future Work Conclusions from this preliminary work in DFT modeling of PCB sorption to NOM and BC are as follows. The modeling was able to predict single molecule and multiple molecule ground state geometries that are reasonable given chemical intuition and general sorption behaviors observed in the environment. The final geometries generated thermochemical data that can be used to arrive at free energies of formation that provide a metric for comparison of various individual structure and sorbate-sorbent interactions. The thermochemistry quantities seem to be reasonable as trends seen in them are consistent with spatial configurations. The use of spatial reasoning with comparative free energies allows inferences to be made concerning mechanism and cause for the conformations. These inferences are primarily speculative and will need to be investigated individually for confirmation. This work could be continued in many different directions, but some future work is fairly obvious. The modeling itself is not complete nor is have all the current results been sufficiently examined. What has been planned for modeling needs to be completed and analyzed. PCB 11 was the only structure that was examined for multiple conformations. When current modeling efforts are complete, it will need to be made certain that all reasonable conformers have been found. If the lowest or highest energy conformer is somehow missed, it could affect later conclusions. These future activities can be classified as simply “due diligence” to the direction that the work has already taken. Other activities that are valuable to attempt are first to remove some of the model assumptions that have been used thus far. Rather than simulate in vacuo, water molecules can be added to see how certain structures may be stabilized or to more clearly define how “hydrophobic interactions” are or are not significant to adsorption. Natural surface water also have varying pH, temperature, and specific conductivity. An attempt should be made to examine these influences, especially specific conductivity since it is composed of ionic species. One last future step is to examine the effects of halogen substitution including bromination, fluorination, and mixed halogen biphenyl contaminants. When all or much of this activitivity has been completed, it is hoped that the result will help to better explain bulk sorption behavior previously seen and that will be investigated in the HSC. It may even be possible to quantitate a bulk property such as number of sorption sites, sorptive strength, or solid-water partition coefficient using the DFT results.

Page 24 of 68


References Abdoul-Kassim, T., Simoneit, B., 2001. Pollutant-Solid Phase Interactions. SpringerVerlag, Berlin. Alvarez-Puebla, R.A., Valenzuela-Calahorro, C., Garrido, J.J., 2006. Theoretical study on fulvic acid structure, conformation and aggregation - A molecular modelling approach. Sci. Total Environ. 358, 243-254. Arp, H.P.H., Breedveld, G.D., Cornelissen, G., 2009. Estimating the in situ SedimentPorewater Distribution of PAHs and Chlorinated Aromatic Hydrocarbons in Anthropogenic Impacted Sediments. Environ. Sci. Technol. 43, 5576-5585. Butcher, J.B., Garvey, E.A., Bierman, V.J., Jr., 1998. Equilibrium partitioning of PCB congeners in the water column: field measurements from the Hudson River. Chemosphere 36, 3149-3166. Di Toro, D.M., Horzempa, L.M., 1982. Reversibe and resistant components of PCB adsorption desorption - isotherms. Environ. Sci. Technol. 16, 594-602. Frame, G.M., Cochran, J.W., Bowadt, S.S., 1996. Complete PCB congener distributions for 17 aroclor mixtures determined by 3 HRGC systems optimized for comprehensive, quantitative, congener-specific analysis. Journal of High Resolution Chromatography 19, 657-668. Goldberg, S., Criscenti, L.J., Turner, D.R., Davis, J.A., Cantrell, K.J., 2007. Adsorption Desorption Processes in Subsurface Reactive Transport Modeling. Vadose Zone J. 6, 407-435. Govers, H.A.J., Van Roon, A., Parsons, J.R., 2003. Calculation of the interaction between dissolved organic carbon and organic micropollutants by three dimensional force field methods. Environ. Toxicol. Chem. 22, 753-759. Horzempa, L.M., Di Toro, D.M., 1983. The Exent of Reversibility of Polychorinated Biphenyl Adsorption. Water Research 17, 851-859. Ibrahim, M., El-Aal, M.A., 2008. Spectroscopic study of the interaction of heavy metals with organic acids. Int. J. Environ. Pollut. 35, 99-110. Karickhoff, S.W., 1979. Sorption of hydrophobic pollutants on natural sediments. Water Research 25, 241-248. Page 25 of 68


Karickhoff, S.W., 1981. Semiempirical Estimation of Sorption of Hydrophobic Pollutants on Natural Sediments and Soils. Chemosphere 10, 833-846. Keiluweit, M., Kleber, M., 2009. Molecular-Level Interactions in Soils and Sediments: The Role of Aromatic pi-Systems. Environmental Science & Technology 43, 34213429. Kubicki, J.D., 2006. Molecular simulations of benzene and PAH interactions with soot. Environmental Science & Technology 40, 2298-2303. Kubicki, J.D., Apitz, S.E., 1999. Models of natural organic matter and interactions with organic contaminants. Organic Geochemistry 30, 911-927. Maurice, P.A., Namjesnik-Dejanovic, K., 1999. Aggregate structures of sorbed humic substances observed in aqueous solution. Environmental Science & Technology 33, 1538-1541. Meyer, E.A., Castellano, R.K., Diederich, F., 2003. Interactions with aromatic rings in chemical and biological recognition. Angew. Chem.-Int. Edit. 42, 1210-1250. Rodenburg, L.A., Guo, J., Du, S., Cavallo, G.J., 2009. Evidence for Unique and Ubiquitous Environmental Sources of 3,3â€²-Dichlorobiphenyl (PCB 11). Environmental Science & Technology. Schulten, H.R., 1999. Analytical pyrolysis and computational chemistry of aquatic humic substances and dissolved organic matter. Journal of Analytical and Applied Pyrolysis 49, 385-415. Schulten, H.R., Schnitzer, M., 1997. Chemical model structures for soil organic matter and soils. Soil Science 162, 115-130. Seijo, M., Ulrich, S., Filella, M., Buffle, J., Stoll, S., 2009. Modeling the Adsorption and Coagulation of Fulvic Acids on Colloids by Brownian Dynamics Simulations. Environmental Science & Technology 43, 7265-7269. Seth, R., Mackay, D., Muncke, J., 1999. Estimating the organic carbon partition coefficient and its variability for hydrophobic chemicals. Environ. Sci. Technol. 33, 2390-2394. Smith, D.M., Chughtai, A.R., 1995. The Surface-Structure and Reactivity of Black Carbon. Colloid Surf. A-Physicochem. Eng. Asp. 105, 47-77. Page 26 of 68


Sutton, R., Sposito, G., Diallo, M.S., Schulten, H.R., 2005. Molecular simulation of a model of dissolved organic matter. Environ. Toxicol. Chem. 24, 1902-1911. Yan, L., Bailey, G.W., 2000. Molecular dynamics modeling of sorption of pesticides onto the surfaces of kaolinite. American Chemical Society, pp. COMP-128. Zhu, D., Hyun, S., Pignatello, J.J., Lee, L.S., 2004. Evidence for pi-pi Electron DonorAcceptor Interactions between pi-Donor Aromatic Compounds and pi-Acceptor Sites in Soil Organic Matter through pH Effects on Sorption. Environmental Science & Technology 38, 4361-4368.

Page 27 of 68


Appendix There are two forms of model data in this appendix. The first is the set of thermochemical parameters predicted for each ground state atom system. The second set of data are tables of bond lengths for all covalent bonds known to exist in the system. Extensive labeling of atom centers is provided to help identify exactly where each bond is in the structure. Inter-atomic distance that are not noted specifically as “bonds” would need to be obtained from the Cartesian coordinates files, which are not included in this appendix.

Page 28 of 68


Thermochemical Data Sorbents

Page 29 of 68


Page 30 of 68


Contaminants

Page 31 of 68


Page 32 of 68


Page 33 of 68


Page 34 of 68


Sorbate-Sorbent Structures Aliphatic Humic

Other interactive model runs with the aliphatic humic are postponed until other runs with aromatic humic and coronene can be completed.

Page 35 of 68


Aromatic Humic

Several additional model runs are still active for PCB11 and ARHUMIC. There are 11 runs begun so far. Coronene Model runs still active, and most are with PCB 11. There are 10 total coronene-PCB 11 interactions that still need results.

Page 36 of 68


Bond Lengths Sorbents

ALHUMICtest1

Bond 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17

Start Atom 1 1 1 1 2 2 2 4 4 4 6 7 7 7 9 10 11

End Atom 2 3 19 26 4 5 29 7 8 27 26 12 22 28 27 29 28

Bond Order 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 Page 37 of 68

Rotatable True False True False True False False True False False False False True False False False False

Length (Angstroms) 1.5378 1.1000 1.5231 1.4134 1.5481 1.0958 1.4306 1.5323 1.0941 1.4274 0.9654 1.0910 1.5412 1.4335 0.9680 0.9661 0.9697


18 19 20 21 22 23 24 25 26 27 28

13 13 13 13 14 14 14 19 19 22 22

14 15 16 24 17 18 25 20 21 23 24

1 1 1 1 1 1 1 2 1 2 1

Page 38 of 68

False False False True False False False False False False True

1.5174 1.0888 1.0912 1.4509 1.0937 1.0917 1.0906 1.2029 1.1096 1.2089 1.3401


ARHUMICtest1

Bond 1 2 3 4 5 6 7 8 9 10

Start Atom 1 1 1 2 2 3 3 4 4 5

End Atom 2 6 28 3 7 4 15 5 8 6

Bond Order 2 1 1 1 1 2 1 1 1 2

Rotatable Bond Length (Angstroms) False False False False False False True False False False

Page 39 of 68

1.3874 1.4032 1.3616 1.4034 1.0831 1.3995 1.4851 1.3941 1.0828 1.3872


Bond 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 20 27 28 29 30 31 32 33 34 35 36

Start Atom 5 6 10 10 10 11 11 12 12 13 13 14 14 18 19 19 19 20 20 20 26 28 30 32 32 34

End Atom 9 30 11 15 16 12 26 13 18 14 32 15 17 19 20 21 22 23 24 25 27 29 31 33 34 35

Bond Order 1 1 2 1 1 1 1 2 1 1 1 2 1 1 1 1 1 1 1 1 1 1 1 2 1 1

Rotatable Bond Length (Angstroms) False False False False False False False False True False True False False True False False False False False False False False False False False False

Page 40 of 68

1.0861 1.3762 1.3884 1.4009 1.0830 1.4052 1.3580 1.4047 1.3736 1.4028 1.4869 1.3970 1.0801 1.4537 1.5135 1.0893 1.0966 1.0930 1.0932 1.0912 0.9693 0.9658 0.9618 1.2093 1.3589 0.9680


ARHUMICtest2

Bond

Start Atom

End Atom

Bond Order

Rotatable

Bond Length (Angstroms)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22

33 21 27 26 8 16 11 11 9 10 32 32 22 4 4 12 12 5 18 15 15 13

32 19 26 11 4 10 10 12 5 15 13 34 19 5 3 18 13 6 19 3 14 14

2 1 1 1 1 1 2 1 1 1 1 1 1 2 1 1 2 1 1 1 2 1

False False False False False False False False False False True False False False False True False False True True False False

1.2562 1.0905 0.9693 1.3295 1.0795 1.08 1.3632 1.4741 1.0844 1.4273 1.5199 1.2289 1.087 1.3736 1.4328 1.2901 1.4406 1.4026 1.4979 1.4375 1.4287 1.3717

Page 41 of 68


Bond

Start Atom

End Atom

Bond Order

Rotatable

Bond Length (Angstroms)

23 24 25 26 27 28 29 30 31 32 33 34 35

3 14 31 19 6 6 2 2 1 28 20 20 20

2 i7 30 20 30 1 1 7 28 29 24 25 23

2 1 1 1 1 2 1 1 1 1 1 1 1

False False False False False False False False False False False False False

1.4182 1.0802 0.9673 1.5146 1.3324 1.4289 1.3757 1.0804 1.3388 0.9676 1.0932 1.0897 1.0916

Page 42 of 68


CORONENEtest1

Bond 1 2 3 4 5 6 7 8 9 10 11 12 13

Start Atom 1 1 1 2 2 3 3 4 4 5 5 6 7

End Atom 2 6 19 3 23 4 7 5 8 6 16 20 10

Bond Order 1 2 1 2 1 1 1 2 1 1 1 1 1

Rotatable Bond Length (Angstroms) False 1.4224 False 1.3695 False 1.0849 False 1.4191 False 1.4225 False 1.4257 False 1.4257 False 1.4192 False 1.4256 False 1.4224 False 1.4225 False 1.0850 False 1.4256

Page 43 of 68


Bond 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42

Start Atom 7 8 8 9 9 10 11 11 12 12 13 13 14 15 15 16 23 23 24 24 25 28 28 29 29 30 33 33 35

End Atom 25 9 14 10 11 30 12 33 13 21 14 22 15 16 18 17 24 26 25 27 28 29 31 30 32 35 34 35 36

Bond Order 2 1 2 1 2 2 1 1 2 1 1 1 1 2 1 1 2 1 1 1 1 2 1 1 1 1 1 2 1

Rotatable Bond Length (Angstroms) False 1.4191 False 1.4257 False 1.4191 False 1.4257 False 1.4191 False 1.4192 False 1.4225 False 1.4224 False 1.3694 False 1.0849 False 1.4224 False 1.0849 False 1.4224 False 1.3693 False 1.0850 False 1.0849 False 1.3694 False 1.0849 False 1.4224 False 1.0849 False 1.4224 False 1.3693 False 1.0850 False 1.4225 False 1.0849 False 1.4224 False 1.0849 False 1.3695 False 1.0850

Page 44 of 68


Contaminants

PCB11test1

Bond 1 2 3 4 5 3 7 8 9 10 11 12 13 14 15 16 17 18 19

Start Atom 1 1 1 2 2 4 3 4 4 5 5 6 11 11 11 12 12 13 14

End Atom 2 6 7 3 8 2 9 5 21 9 10 13 12 16 17 13 18 14 15

Bond Order 2 1 1 1 1 2 1 1 1 2 1 1 2 1 1 1 1 2 1

Rotatable False False False False False False False False False False False True False False False False False False False

Page 45 of 68

Bond Length (Angstroms) 1.3908 1.4019 1.0833 1.3927 1.0840 1.3902 1.0820 1.3888 1.7608 1.4015 1.0822 1.4857 1.3908 1.3927 1.0840 1.4019 1.0833 1.4015 1.3888


Bond 20 21 22 23

Start Atom 14 15 15 16

End Atom 19 16 22 20

Bond Order 1 2 1 1

Rotatable False False False False

Page 46 of 68

Bond Length (Angstroms) 1.0822 1.3902 1.7608 1.0820


PCB11test2

Bond 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19

Start Atom 6 1 1 3 2 4 9 21 5 10 5 13 16 12 11 13 12 14 19

End Atom 1 2 7 2 8 3 3 4 4 5 6 6 11 11 17 12 18 13 14

Bond Order 2 1 1 2 1 1 1 1 2 1 1 1 2 1 1 2 1 1 1


Page 47 of 68



Bond 20 21 22 23


End Atom 15 15 16 16

Bond Order 2 1 1 1


Page 48 of 68



PCB126test1

Bond 1 2 3 4 5 6 7 8 9 10 11 12 13 14

Start Atom 1 1 1 2 2 3 3 4 4 5 5 6 10 10

End Atom 2 6 7 3 8 4 21 5 18 6 9 12 11 15

Bond Order 2 1 1 1 1 2 1 1 1 2 1 1 2 1

Rotatable False False False False False False False False False False False True False False

Page 49 of 68

Bond Length (Angstroms) 1.3874 1.4012 1.0831 1.3933 1.0823 1.3977 1.7447 1.3915 1.7469 1.3984 1.0821 1.4838 1.3890 1.4014


Bond 15 16 17 18 19 20 21 22 23

Start Atom 10 11 11 12 13 13 14 14 15

End Atom 22 12 16 13 14 17 15 19 20

Bond Order 1 1 1 2 1 1 2 1 1

Rotatable False False False False False False False False False

Page 50 of 68

Bond Length (Angstroms) 1.7469 1.3979 1.0815 1.3979 1.3890 1.0816 1.4015 1.7466 1.7348


PCB209test1

Bond 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21

Start Atom 1 1 1 2 2 3 3 4 4 5 5 6 7 7 7 8 8 9 10 10 11

End Atom 2 6 19 3 20 4 16 5 13 6 21 9 8 12 17 9 18 10 11 22 12

Bond Order 2 1 1 1 1 2 1 1 1 2 1 1 2 1 1 1 1 2 1 1 2

Rotatable False False False False False False False False False False False True False False False False False False False False False

Page 51 of 68

Bond Length (Angstroms) 1.3990 1.3982 1.7393 1.4018 1.7343 1.4018 1.7338 1.3990 1.7343 1.3982 1.7393 1.4924 1.3990 1.4018 1.7343 1.3982 1.7394 1.3982 1.3990 1.7394 1.4018


22

11

14

1

False

23

12

15

1

False

Start Atom 1 1 1 2 2 3 3 4 4 5 5 6 7 7 7 8 8 9

End Atom 2 6 19 3 16 4 22 5 14 6 17 9 8 12 21 9 15 10

Bond Order 2 1 1 1 1 2 1 1 1 2 1 1 2 1 1 1 1 2

Bond 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18

Rotatable False False False False False False False False False False False False False False False False False False

Page 52 of 68

1.7343 1.7338

Bond Length (Angstroms) 1.3908 1.4015 1.0834 1.3931 1.0841 1.3907 1.0819 1.3891 1.9193 1.4023 1.0821 1.4860 1.3908 1.3931 1.0841 1.4015 1.0834 1.4023

PBB11test1


Bond 19 20 21 22 23

Start Atom 10 10 11 11 12

End Atom 11 18 12 13 20

Bond Order 1 1 2 1 1

Rotatable False False False False False

Page 53 of 68

Bond Length (Angstroms) 1.3891 1.0821 1.3907 1.9193 1.0819


PCB126test1 1

Bond 1 2 3 4 5 6 7 8 9 10 11 12 13

Start Atom 1 1 1 2 2 3 3 4 4 5 5 6 7

End Atom 2 6 22 3 19 4 14 5 17 6 20 9 8

Bond Order 2 1 1 1 1 2 1 1 1 2 1 1 2

Order False False False False False False False False False False False True False

Page 54 of 68

Bond Length (Angstroms) 1.3875 1.4008 1.0834 1.3951 1.0821 1.3974 1.9058 1.3932 1.9081 1.3986 1.0821 1.4840 1.3902


Bond 14 15 16 17 18 19 20 21 22 23

Start Atom 7 7 8 8 9 10 10 11 11 12

End Atom 12 13 9 18 10 11 21 12 15 16

Bond Order 1 1 1 1 2 1 1 2 1 1

Order False False False False False False False False False False

Page 55 of 68

Bond Length (Angstroms) 1.4028 1.9096 1.3977 1.0814 1.3977 1.3903 1.0815 1.4030 1.9094 1.8982


PBB209test1

Bond 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19

Start Atom 1 1 1 2 2 3 3 4 4 5 5 6 7 7 7 8 8 9 10

End Atom 2 6 15 3 16 4 17 5 22 6 18 9 8 12 13 9 14 10 11

Bond Order 2 1 1 1 1 2 1 1 1 2 1 1 2 1 1 1 1 2 1


Page 56 of 68



Bond 20 21 22 23


End Atom 19 12 20 21

Bond Order 1 2 1 1


Page 57 of 68



Sorbate-Sorbent Interactions

ALHUM-PCB11test1

Bond 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

Start Atom 1 1 1 1 2 2 2 4 4 4 6 7 7 7 9

End Atom 2 3 19 26 4 5 29 7 8 27 26 12 22 28 27

Bond Order 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1

Rotatable True False True False True False False True False False False False True False False

Page 58 of 68

Bond Length (Angstroms) 1.5388 1.1000 1.5222 1.4132 1.5474 1.0958 1.4297 1.5330 1.0958 1.4258 0.9655 1.0911 1.5402 1.4314 0.9689


Bond 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 33 39 40 41 42 43 44 45 46 47 43 49 50 51

Start Atom 10 11 13 13 13 13 14 14 14 19 19 22 22 30 30 30 31 31 32 32 33 33 34 34 35 40 40 40 41 41 42 42 43 43 44 44

End Atom 29 28 14 15 16 24 17 18 25 20 21 23 24 31 35 36 32 37 33 38 34 45 35 39 51 41 45 46 42 47 43 48 44 50 45 49

Bond Order 1 1 1 1 1 1 1 1 1 2 1 2 1 2 1 1 1 1 2 1 1 1 2 1 1 2 1 1 1 1 2 1 1 1 2 1

Rotatable False False False False False True False False False False False False True False False False False False False False False True False False False False False False False False False False False False False False

Page 59 of 68

Bond Length (Angstroms) 0.9666 0.9692 1.5166 1.0880 1.0914 1.4527 1.0935 1.0917 1.0908 1.2029 1.1100 1.2106 1.3386 1.3926 1.3902 1.0820 1.3916 1.0839 1.4025 1.0837 1.4018 1.4859 1.3887 1.0822 1.7609 1.3924 1.4038 1.0822 1.3940 1.0841 1.3902 1.0820 1.3894 1.7599 1.4020 1.0821


ALHUM-PCB126test1

Bond 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17

Start Atom 1 1 1 1 2 2 2 4 4 4 6 7 7 7 9 10 11

End Atom 2 3 19 26 4 5 29 7 8 27 26 12 22 28 27 29 28

Bond Order 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1

Rotatable Bond Length (Angstroms) True 1.5394 False 1.0991 True 1.5190 False 1.4133 True 1.5456 False 1.0957 False 1.4290 True 1.5321 False 1.0946 False 1.4260 False 0.9656 False 1.0915 True 1.5414 False 1.4318 False 0.9681 False 0.9657 False 0.9692

Page 60 of 68


Bond 18 19 20 21 22 23 24 25 26 27 20 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51

Start Atom 13 13 13 13 14 14 14 19 19 22 22 30 30 30 31 31 32 32 33 33 34 34 35 38 38 38 39 39 40 40 41 41 42 42

End Atom 14 15 16 24 17 18 25 20 21 23 24 31 35 49 32 36 33 43 34 37 35 47 48 39 43 44 40 45 41 50 42 51 43 46

Bond Order 1 1 1 1 1 1 1 2 1 2 1 2 1 1 1 1 2 1 1 1 2 1 1 2 1 1 1 1 2 1 1 1 2 1

Rotatable Bond Length (Angstroms) False 1.5170 False 1.0888 False 1.0912 True 1.4513 False 1.0936 False 1.0917 False 1.0905 False 1.2062 False 1.1073 False 1.2087 True 1.3397 False 1.3884 False 1.4006 False 1.7487 False 1.3986 False 1.0813 False 1.3990 True 1.4847 False 1.3887 False 1.0812 False 1.4015 False 1.7487 False 1.7364 False 1.3872 False 1.4018 False 1.0830 False 1.3927 False 1.0822 False 1.3971 False 1.7465 False 1.3912 False 1.7498 False 1.3998 False 1.0828

Page 61 of 68


ARHUM-PCB11test1

Bond 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

Start Atom 1 1 1 2 2 3 3 4 4 5 5 6 10 10 10 11

End Atom 2 6 28 3 7 4 15 5 8 6 9 30 11 15 16 12

Bond Order 2 1 1 1 1 2 1 1 1 2 1 1 1 1 1 1

Rotatable Bond Length (Angstroms) False False False False False False True False False False False False False False False False

Page 62 of 68

1.3866 1.4032 1.3651 1.4036 1.0833 1.3994 1.4848 1.3940 1.0827 1.3876 1.0860 1.3727 1.3882 1.4010 1.0830 1.4052


Bond 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55

Start Atom 11 12 12 13 13 14 14 18 19 19 19 20 20 20 26 28 30 32 32 34 36 36 36 37 37 38 38 39 39 40 40 41 46 46 47 47 48 48 49

End Atom 26 13 18 14 32 15 17 19 20 21 22 23 24 25 27 29 31 33 34 35 37 41 42 38 46 39 43 40 44 41 45 57 47 51 48 52 49 56 50

Bond Order 1 2 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 1 1 2 1 1 1 1 2 1 1 1 2 1 1 2 1 1 1 2 1 1

Rotatable Bond Length (Angstroms) False False True False True False False True False False False False False False False False False False False False False False False False True False False False False False False False False False False False False False False

Page 63 of 68

1.3580 1.4051 1.3733 1.4029 1.4871 1.3970 1.0801 1.4534 1.5135 1.0893 1.0966 1.0929 1.0932 1.0912 0.9693 0.9681 0.9620 1.2094 1.3584 0.9680 1.4007 1.3889 1.0821 1.4027 1.4858 1.3921 1.0833 1.3930 1.0840 1.3898 1.0820 1.7628 1.4012 1.4017 1.3889 1.0824 1.3903 1.7614 1.3927


Bond 56 57 58 59


End Atom 53 51 54 55

Bond Order 1 2 1 1

Rotatable Bond Length (Angstroms) False False False False

Page 64 of 68

1.0821 1.3913 1.0838 1.0835


ARHUM-PCB11test6

Bond 1 2 3 4 5 6 7 8 9 10 11 12 13

Start Atom 1 1 1 2 2 3 3 4 4 5 5 6 10

End Atom 2 6 28 3 7 4 15 5 8 6 9 30 11

Bond Order 2 1 1 1 1 2 1 1 1 2 1 1 2

Rotatable False False False False False False True False False False False False False

Page 65 of 68

Bond Length (Angstroms) 1.3891 1.4002 1.3665 1.4041 1.0828 1.4042 1.4847 1.3936 1.0830 1.3870 1.0861 1.3899 1.3827


Bond 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 30 37 39 39 40 41 42 43 44 45 46 47 48 49 50 51 52

Start Atom 10 10 11 11 12 12 13 13 14 14 18 19 26 19 20 20 20 26 28 30 32 32 35 35 35 36 36 37 37 38 38 39 39 40 45 45 46 46 47

End Atom 15 16 12 26 13 18 14 32 15 17 19 20 19 22 23 24 25 27 29 31 33 34 36 40 41 37 45 38 42 39 43 40 44 56 46 50 47 51 48

Bond Order 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 2 2 1 1 1 1 2 1 1 1 2 1 1 2 1 1 1 2

Rotatable False False False False False True False False False False True False False False False False False False False False False False False False False False True False False False False False False False False False False False False

Page 66 of 68

Bond Length (Angstroms) 1.4044 1.0829 1.4079 1.3759 1.4005 1.3864 1.4026 1.5720 1.3966 1.0832 1.4486 1.5144 1.0881 1.1001 1.0951 1.0950 1.0906 0.9724 0.9661 0.9627 1.2529 1.2440 1.4032 1.3887 1.0820 1.4027 1.4866 1.3905 1.0833 1.3923 1.0842 1.3898 1.0821 1.7667 1.4025 1.4016 1.3876 1.0822 1.3894


Bond 53 54 55 56 57 58

Start Atom 47 48 48 49 49 50

End Atom 55 49 52 50 53 54

Bond Order 1 1 1 2 1 1

Rotatable False False False False False False

Page 67 of 68

Bond Length (Angstroms) 1.7680 1.3935 1.0822 1.3926 1.0848 1.0834


EXTRA PAGE (Needed for bibliography) (Ibrahim and El-Aal, 2008)

Page 68 of 68