PDF Archive

Easily share your PDF documents with your contacts, on the Web and Social Networks.

Share a file Manage my documents Convert Recover PDF Search Help Contact



J Environ Sci Heal A 44, 2009, 317 326 .pdf


Original filename: J Environ Sci Heal A 44, 2009, 317-326.pdf

This PDF 1.5 document has been generated by / iText 2.1.3 (by lowagie.com), and has been sent on pdf-archive.com on 03/11/2015 at 04:25, from IP address 71.17.x.x. The current document download page has been viewed 395 times.
File size: 240 KB (11 pages).
Privacy: public file




Download original PDF file









Document preview


This article was downloaded by: [Rayne, Sierra][Canadian Research Knowledge Network]
On: 30 January 2009
Access details: Access Details: [subscription number 783016891]
Publisher Taylor & Francis
Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House,
37-41 Mortimer Street, London W1T 3JH, UK

Journal of Environmental Science and Health, Part A
Publication details, including instructions for authors and subscription information:
http://www.informaworld.com/smpp/title~content=t713597268

Computational approaches may underestimate pK<i>a</i> values of longer-chain
perfluorinated carboxylic acids: Implications for assessing environmental and
biological effects
Sierra Rayne a; Kaya Forest b; Ken J. Friesen a
a
Department of Chemistry, The University of Winnipeg, Winnipeg, MB, Canada b Department of Chemistry,
Okanagan College, Penticton, BC, Canada
Online Publication Date: 01 March 2009

To cite this Article Rayne, Sierra, Forest, Kaya and Friesen, Ken J.(2009)'Computational approaches may underestimate pK<i>a</i>

values of longer-chain perfluorinated carboxylic acids: Implications for assessing environmental and biological effects',Journal of
Environmental Science and Health, Part A,44:4,317 — 326
To link to this Article: DOI: 10.1080/10934520802659620
URL: http://dx.doi.org/10.1080/10934520802659620

PLEASE SCROLL DOWN FOR ARTICLE
Full terms and conditions of use: http://www.informaworld.com/terms-and-conditions-of-access.pdf
This article may be used for research, teaching and private study purposes. Any substantial or
systematic reproduction, re-distribution, re-selling, loan or sub-licensing, systematic supply or
distribution in any form to anyone is expressly forbidden.
The publisher does not give any warranty express or implied or make any representation that the contents
will be complete or accurate or up to date. The accuracy of any instructions, formulae and drug doses
should be independently verified with primary sources. The publisher shall not be liable for any loss,
actions, claims, proceedings, demand or costs or damages whatsoever or howsoever caused arising directly
or indirectly in connection with or arising out of the use of this material.

Journal of Environmental Science and Health Part A (2009) 44, 317–326
C Taylor & Francis Group, LLC
Copyright
ISSN: 1093-4529 (Print); 1532-4117 (Online)
DOI: 10.1080/10934520802659620

Computational approaches may underestimate pKa values
of longer-chain perfluorinated carboxylic acids: Implications
for assessing environmental and biological effects
SIERRA RAYNE1 , KAYA FOREST2 and KEN J. FRIESEN1
1
Downloaded By: [Rayne, Sierra][Canadian Research Knowledge Network] At: 18:24 30 January 2009

2

Department of Chemistry, The University of Winnipeg, Winnipeg, MB, Canada
Department of Chemistry, Okanagan College, Penticton, BC, Canada

Acidity constants were calculated using the semiempirical PM6 pKa estimation method for all C2 through C9 perfluoroalkyl carboxylate (PFCA) congeners and the straight-chain C10 through C13 isomers. According to the PM6 estimates, the linear congeners
within each PFCA homologue group have the highest pKa values by up to 6 units depending on the degree of branching in the
perfluoroalkyl chain. In general, the higher the degree of branching in the perfluoroalkyl chain within a homologue group, the lower
the estimated pKa value. When the branching is closest to the terminal carboxylate group, the effect on the calculated pKa is greatest.
Although the PM6 calculated pKa values agree well with previously reported estimates for selected linear PFCA congeners using the
SPARC and COSMOtherm approaches, all computational approaches only show good agreement with reported experimental values
for short chain PFCAs (C2 through C5 ). Increasing divergences are observed between calculated and experimental results by up to
several pKa units as the perfluoroalkyl chain length increases beyond C5 . The findings demonstrate a need for additional experimental
pKa measurements for an expanded set of both linear and branched PFCA congeners to confirm previous experimental reports that
are potentially in error, and upon which to calibrate existing computational methods and environmental, toxicological, and waste
treatment method models.
Keywords: Perfluorinated compounds, perfluorinated carboxylic acids, physicochemical properties, environmental partitioning, acidity
constants, pKa values, computational methods.

Introduction
Perfluorinated compounds (PFCs) are used in a wide variety of industrial and consumer products.[1] A subclass of
PFCs, the perfluorinated carboxylic acids (PFCAs) have
been produced since the 1940s (Fig. 1). Recent applications of PFCAs have exploited their surface active properties and stability in materials such as water repellants,
fire suppressants, cleaning solutions, lubricants, and fuels,
among others.[2] PFCAs can also be formed via the degradation of other PFCs, such as the perfluorinated sulfonic
acids and telomer alcohols and olefins.[2,3] Consequently,
PFCAs are found widely in the environment with concentrations in surface, ground, and drinking waters generally
in the ng/L range but up to the µg/L or mg/L range in contaminated regions and wastewater effluents.[4,5] In addition,
PFCAs are ubiquitous in global wildlife and humans, and
Address correspondence to Sierra Rayne, Department of Chemistry, The University of Winnipeg, Winnipeg, MB, Canada;
E-mail: s.rayne@uwinnipeg.ca
Received July 26, 2008.

longer chain members of this contaminant class are known
to bioaccumulate and bioconcentrate.[6−8]
As evidence for the widespread occurrence of PFCAs in
environmental media increases, the research focus includes
new methods for degrading these compounds and their
fate in existing treatment processes.[4,9−13] Key to predicting
both the efficacy of treatment processes on PFCAs, as well
as the construction of multimedia environmental models to
better understand their environmental chemistry,[2,5,14−16]
is the development of reliable databases containing their
physico chemical properties. Estimating the partitioning behavior of different PFC classes, including the PFCAs, is an
active area of research with a number of studies using computational approaches such as SPARC, COSMOtherm,
EPI SUITE, and ClogP.[17−20] However, because of the amphiphilic structure of PFCAs,[2] they behave in a surfactantlike manner in aquatic systems. Unfortunately, our current
computational algorithms are generally not well suited for
modeling surfactants, as evidenced by the often reported
differences between experimental and estimated environmental constants of several orders of magnitude, and between varying computational approaches, for PFCs.[18]

318

Downloaded By: [Rayne, Sierra][Canadian Research Knowledge Network] At: 18:24 30 January 2009

Fig. 1. General structure of perfluoroalkyl carboxylic acids.

A primary physico chemical property of PFCAs that has
attracted much recent attention—and debate—in the literature is their acidity constants, or pKa values, which form
a core input in current modeling efforts. Here we examine a new semiempirical computational chemistry method
(PM6) to estimate the pKa values for not only the wellknown linear PFCA congener, but also the much larger
number of branched congeners for which there is negligible
information on either their presence in the environment or
their partitioning constants. In addition, we provide a comparative analysis of the PM6 results with prior SPARC and
COSMOtherm estimates and existing experimental data,
and discuss the implications of any deviations between estimated and experimental pKa values on our understandings
of PFCA environmental fate, toxicology, and waste treatment methods.

Materials and methods
Congener specific PFC identifications in the manuscript
refer to the general numbering approach published
elsewhere.[21] Gas phase molecular structures were initially
optimized with the MM2 molecular mechanics energyminimization method[22] using a minimum RMS gradient
of 0.100. The gas phase structures were then further optimized using the PM3[23] semiempirical method in MOPAC
2000[24] with a minimum RMS gradient of 0.100 and a
closed shell (restricted) wave function. The gas phase PM3
optimized geometries from MOPAC2000 were used as the
input geometries for aqueous phase PM6 geometry optimization in MOPAC2007.[25] Geometry optimizations and
pKa calculations in MOPAC2007 were conducted with the
following keywords in the input file header: PM6; PKA;
BONDS; CHARGE = 0; SINGLET; LET; DDMIN = 0.0;
GNORM = 0.01.

Results and discussion
Estimated pKa values range from +0.1 to −5.9 among
all congeners in the C2 through C9 PFCA homologue
groups and the straight chain C10 through C13 members
using the PM6 semiempirical basis set calculated using the
MOPAC 2007 platform (Table 1). Within this computationally derived dataset, several pKa trends depend on the perfluoroalkyl chain structure. Linear congeners within each

Rayne et al.
homologue group are predicted to have the highest pKa values by up to 6 units depending on the degree of branching in
the perfluoroalkyl chain. In general, the higher the degree
of branching in the perfluoroalkyl chain, the lower the estimated pKa value. When the branching is closest to the terminal carboxylate group, the branching effect on the calculated pKa is greatest. For example, the 1,1 ,2-trimethylbutyl
C8 PFCA-12 is estimated to have a pKa 4.2 units lower
than the corresponding linear congener (i.e., C8 PFCA-39),
whereas moving this trimethyl branching pattern farther
along the perfluoroalkyl backbone away from the carboxylate group leads to a much closer agreement with the linear
congener (e.g., 2,3,3 -trimethylbutyl C8 PFCA-18 is estimated to have a pKa only 1.0 units lower than the linear
isomer).
There is limited experimental and computational work
with which to compare our findings. Early work by Brace
found a pKa value of 2.8 for the high-profile linear perfluorooctanoic acid (linear-PFOA; or C8 PFCA-39).[26] Hennie
and Fox reported pKa values of 0.3 for trifluoroacetic acid
(i.e., C2 PFCA) and 0.2 for the linear C4 PFCA (i.e., perfluorobutanoic acid).[27] Generally overlooked in the PFCA
literature is the more recent work of Moroi et al.,[28] who
experimentally determined the acidity constants for the C2
through C12 straight chain PFCAs. Moroi et al.[28] found
that the pKa of straight chain PFCAs decreases slightly
between C2 and C4 , and then begins to increase rapidly
between C5 and C12 (Table 2). A recent report that used
titrimetric and 19 F-NMR based methods to determine the
pKa of linear PFOA corroborates the trend of increasing
pKa with increasing perfluoroalkyl chain length above C5
PFCAs, providing a pKa range for PFOA between 2.3 to
3.4[29] that brackets the previous report by Brace of 2.8[26]
and the estimated value (∼2.5) from a trendline between
the C6 to C10 compounds by Moroi et al.[28]
Similarly, few studies have attempted to estimate the pKa
of PFCAs. Steinle-Darling and Reinhard[10] used SPARC to
estimate the pKa values of the linear C5 through C12 values,
and reported values of −0.2 for all compounds except for
a value of −0.1 for perfluoropentanoic acid. These values
are in excellent agreement with our PM6 estimates of pKa
values between −0.1 and −0.2 for this homologue range. By
comparison, Goss has recently argued that the SPARC and
COSMOthermo computational approaches underestimate
acidities for the C2 through C12 PFCAs by between one and
two orders of magnitude, particularly for the linear-PFOS
congener.[17] Despite SPARC and COSMOtherm pKa value
estimates for linear PFOA of −0.1 and 0.7, respectively,
Goss proposed that the correct value is probably closer to
−0.5. This report thus calls into question all prior use of a
linear PFOA pKa near 2.8[26] in a variety of environmental
modeling investigations.[2,5,14,15] The estimated pKa value
of Goss[17] has also received recent criticism as part of a
study examining the water-air transport of PFOA,[16] whose
authors cited indirect experimental evidence supporting a
significantly higher pKa than Goss’ estimate of −0.5.

319

Long-chain PFCA congeners and pKa values

Table 1. Computationally estimated pKa values for all C2 through C9 and the straight chain C10 through C13 PFCAs using the pKa
estimation function in the PM6 semiempirical basis set.
HGa

Perfluoroalkyl chain substitution

C2
C3
C4

methyl
ethyl
n-propyl
isopropyl
1,1 -dimethylethyl
1-methylpropyl
2-methylpropyl
n-butyl
1-ethylpropyl
1,1 -dimethylpropyl
1,2-dimethylpropyl
2,2 -dimethylpropyl
1-methylbutyl
2-methylbutyl
3-methylbutyl
n-pentyl
1-ethyl-1 -methylpropyl
1-ethyl-2-methylpropyl
1,1 ,2-trimethylpropyl
1,2,2 -trimethylpropyl
1-ethylbutyl
2-ethylbutyl
1,1 -dimethylbutyl
1,2-dimethylbutyl
1,3-dimethylbutyl
2,2 -dimethylbutyl
2,3-dimethylbutyl
3,3 -dimethylbutyl
1-methylpentyl
2-methylpentyl
3-methylpentyl
4-methylpentyl
n-hexyl
1,1 -diethylpropyl
1-ethyl−1 ,2-dimethylpropyl
1-ethyl-2,2 -dimethylpropyl
1-isopropyl-2-methylpropyl
1,1 ,2,2 -tetramethylpropyl
1-ethyl−1 -methylbutyl
1-ethyl-2-methylbutyl
1-ethyl-3-methylbutyl
2-ethyl−1-methylbutyl
2-ethyl-2 -methylbutyl
2-ethyl-3-methylbutyl
1,1 ,2-trimethylbutyl
1,1 ,3-trimethylbutyl
1,2,2 -trimethylbutyl
1,2,3-trimethylbutyl
1,3,3 -trimethylbutyl
2,2 ,3-trimethylbutyl
2,3,3 -trimethylbutyl
1-isopropylbutyl
1-propylbutyl
1-ethylpentyl

Downloaded By: [Rayne, Sierra][Canadian Research Knowledge Network] At: 18:24 30 January 2009

C5

C6

C7

C8

Isomer
no.[21]

pKa

HGa

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

−0.2
−0.2
+0.1
−2.1
−3.4
−1.7
−1.1
−0.1
−1.7
−3.3
−2.1
−1.1
−1.8
−0.9
−0.3
−0.1
−3.8
−2.4
−4.3
−3.2
−2.2
−1.0
−3.3
−1.8
−2.1
−1.0
−1.2
−0.5
−1.7
−0.9
−0.3
−0.2
−0.1
−3.8
−4.2
−2.8
−2.6
−5.1
−4.0
−1.5
−2.5
−1.8
−1.4
−2.4
−4.4
−3.6
−2.7
−2.8
−3.2
−1.5
−1.2
−2.8
−2.5
−2.1

C9

Perfluoroalkyl chain substitution
2-isopropyl-3-methylbutyl
1-ethyl-1 ,2-dimethylbutyl
1-ethyl-1,3-dimethylbutyl
1-ethyl-2,2 -dimethylbutyl
1-ethyl-2,3-dimethylbutyl
1-ethyl-3,3 -dimethylbutyl
2-ethyl−1,1 -dimethylbutyl
2-ethyl−1,2 -dimethylbutyl
2-ethyl-1,3-dimethylbutyl
2-ethyl-2 ,3-dimethylbutyl
2-ethyl-3,3 -dimethylbutyl
1,1 ,2,2 -tetramethylbutyl
1,1 ,2,3-tetramethylbutyl
1,1 ,3,3 -tetramethylbutyl
1,2,2 ,3-tetramethylbutyl
1,2,3,3 -tetramethylbutyl
2,2 ,3,3 -tetramethylbutyl
1-methyl-1 -propyl butyl
2- methyl-1-propyl butyl
3-methyl-1-propylbutyl
1-ethyl-1 -methylpentyl
1-ethyl-2-methylpentyl
1-ethyl-3-methylpentyl
1-ethyl-4-methylpentyl
2-ethyl-1-methylpentyl
2-ethyl-2 -methylpentyl
2-ethyl-3-methylpentyl
2-ethyl-4-methylpentyl
3-ethyl-1-methylpentyl
3-ethyl-2-methylpentyl
3-ethyl-3 -methylpentyl
3-ethyl-4-methylpentyl
1-isopropylpentyl
2-isopropylpentyl
1,1 ,2-trimethylpentyl
1,1 ,3-trimethylpentyl
1,1 ,4-trimethylpentyl
1,2,2 -trimethylpentyl
1,2,3-trimethylpentyl
1,2,4-trimethylpentyl
1,3,3 -trimethylpentyl
1,3,4-trimethylpentyl
1,4,4 -trimethylpentyl
2,2 ,3-trimethylpentyl
2,2 ,4-trimethylpentyl
2,3,3 -trimethylpentyl
2,3,4-trimethylpentyl
2,4,4 -trimethylpentyl
3,3 ,4-trimethylpentyl
3,4,4 -trimethylpentyl
1-propylpentyl
2-propylpentyl
1-ethylhexyl
2-ethylhexyl

Isomer
no.[21]

pKa

12
−1.9
13
−5.2
14
−4.2
15
−3.7
16
−2.1
17
−3.1
18
−4.6
19
−3.1
20
−2.6
21
−1.7
22
−1.2
23
−5.2
24
−5.2
25
−4.0
26
−2.9
27
−2.2
28
−3.5
29
−5.1
30
−1.6
31
−2.9
32
−4.0
33
−2.0
34
−2.8
35
−2.2
36
−2.1
37
−1.4
38
−1.8
39
−1.9
40
−1.9
41
−1.6
42
−0.5
43
−0.5
44
−1.9
45
−0.7
46
−3.6
47
−3.4
48
−3.4
49
−2.6
50
−2.2
51
−2.6
52
−2.4
53
−2.5
54
−1.9
55
−2.8
56
−2.0
57
−1.0
58
−1.1
59
−1.6
60
−0.5
61
−0.4
62
−2.7
63
−2.3
64
−1.4
65
−1.7
(Continued on next page)

320

Rayne et al.

Table 1. (Continued.)

Downloaded By: [Rayne, Sierra][Canadian Research Knowledge Network] At: 18:24 30 January 2009

HGa

C9

a

Perfluoroalkyl chain substitution

Isomer
no.[21]

pKa

2-ethylpentyl
3-ethylpentyl
1,1 -dimethylpentyl
1,2-dimethylpentyl
1,3-dimethylpentyl
1,4-dimethylpentyl
2,2 -dimethylpentyl
2,3-dimethylpentyl
2,4-dimethylpentyl
3,3 -dimethylpentyl
3,4-dimethylpentyl
4,4 -dimethylpentyl
1-methylhexyl
2-methylhexyl
3-methylhexyl
4-methylhexyl
5-methylhexyl
n-heptyl
1-ethyl-1 ,2,2 -trimethylpropyl
1-ethyl-1 -isopropylpropyl
1-isopropyl-1 ,2-dimethylpropyl
1-isopropyl-2,2 -dimethylpropyl
1-tert-butylbutyl
1,1 -diethylbutyl
1,2-diethylbutyl
2,2 -diethylbutyl
1-isopropyl-1 -methylbutyl
1-isopropyl-2-methylbutyl
1-isopropyl-3-methylbutyl

22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
1
2
3
4
5
6
7
8
9
10
11

−1.8
−0.3
−3.3
−1.7
−3.0
−1.8
−1.1
−1.1
−1.9
−0.5
−0.3
−0.3
−1.7
−1.3
−0.2
−0.1
−0.1
−0.2
−5.9
−4.2
−4.6
−3.0
−3.0
−5.7
−2.3
−1.5
−3.4
−2.3
−3.0

HGa

Perfluoroalkyl chain substitution

C10
C11
C12
C13

3-ethylhexyl
4-ethylhexyl
1,1 -dimethylhexyl
1,2-dimethylhexyl
1,3-dimethylhexyl
1,4-dimethylhexyl
1,5-dimethylhexyl
2,2 -dimethylhexyl
2,3-dimethylhexyl
2,4-dimethylhexyl
2,5-dimethylhexyl
3,3 -dimethylhexyl
3,4-dimethylhexyl
3,5-dimethylhexyl
4,4 -dimethylhexyl
4,5-dimethylhexyl
5,5 -dimethylhexyl
1-methylheptyl
2-methylheptyl
3-methylheptyl
4-methylheptyl
5-methylheptyl
6-methylheptyl
n-octyl
n-nonyl
n-decyl
n-undecyl
n-dodecyl

Isomer
no.[21]

pKa
−0.4
−0.2
−3.3
−1.7
−3.1
−1.7
−1.8
−1.0
−1.9
−1.5
−0.9
−0.5
−0.3
−0.4
−0.2
−0.2
−0.2
−1.7
−0.8
−0.3
−0.1
−0.1
−0.1
−0.1
−0.1
−0.1
−0.1
−0.1

66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89

HG=PFCA homologue group.

It is important to note that Goss[17] cited, incorrectly, a
pKa value of −0.3 for trifluoroacetic acid (instead of the
actual value of 0.3) based on the historical work of Hennie
and Fox,[27] and that the more recent pKa measurement
of 0.6 from Moroi et al.[28] (and which is also more widely

cited in the literature) should also have been considered.
The pKa value cited, in comparison to the SPARC and
COSMOtherm estimates of 1.4 and 0.9, respectively,
was one key to Goss[17] argument that these methods
overestimate the pKa values for PFCAs, leading Goss[17]

Table 2. Comparison between available experimentally determined and corresponding computationally estimated pKa values for the
straight chain C2 through C12 PFCAs using the pKa estimation function in the PM6 semiempirical basis set (current work) and the
SPARC and COSMOtherm methods previously reported in the literature.
Compound
linear C2 -PFCA
linear C3 -PFCA
linear C4 -PFCA
linear C5 -PFCA
linear C6 -PFCA
linear C8 -PFCA
linear C10 -PFCA
linear C11 -PFCA
linear C12 -PFCA

Experimental pKa

PM6 estimate
(current work)

0.3 to 0.6[25,27]
0.5[27]
0.2 to 0.4[26,27]
0.5[27]
0.9[27]
2.3 to 3.4[25,28]
2.6[27]
2.7[27]
3.2[12]

−0.2
−0.2
0.1
−0.1
−0.1
−0.1
−0.1
−0.1
−0.1

SPARC
estimate[17]

SPARC
estimate[10]

1.4

0.9

0.4
−0.1

0.2

COSMOtherm
estimate[17]

0.7
−0.1
−0.2
−0.2
−0.2
−0.2
−0.2

0.7

0.8

Downloaded By: [Rayne, Sierra][Canadian Research Knowledge Network] At: 18:24 30 January 2009

Long-chain PFCA congeners and pKa values
to rationalize an approximate order of magnitude lowering
of the estimated pKa value for PFOA down to −0.5 from
the computational results of between −0.1 and 0.7. In
light of the results presented here, and the apparent error
in citing the literature pKa , we argue that Goss’ approach
likely needs to be reconsidered.
The known difficulty in determining pKa values for surfactants such as the PFCAs raises reasonable concerns as to
the relative reliability of the experimental versus the computational data. Moroi et al. noted explicitly that micellar
aggregates were likely present during their pKa determinations for the C6 through C12 PFCAs, leading to the measurement of oligomeric acidity constants rather than the corresponding monomeric values obtained for the C2 through
C5 PFCAs.[28] However, in their recent pKa investigations
on linear PFOA, Webster[29] found that the aggregate pKa
for this longer chain PFCA was lower than the more environmentally relevant monomeric pKa by about one unit.
Thus, even if the C6 through C12 PFCA pKa data of Moroi et al.[28] represent aggregate pKa values, Webster’s[29]
subsequent findings that aggregate pKa values underestimate the monomeric pKa values further suggests that Moroi et al.’s[28] pKa values are actually underestimates, rather
than overestimates. Consequently, it appears that the current computational approaches may be underestimating the
pKa of longer chain PFCAs by at least several pKa units,
and that Goss’[17] estimate of a pKa of −0.5 for linear PFOA
may be too low by more than three units. In the absence of
other contradictory evidence, the current experimentally
determined average pKa of about 2.8 should likely be used
in modeling the environmental fate of linear PFOA until the
discrepancies between the computational and experimental
approaches can be resolved.
One potential explanation that warrants investigation
for the increasing pKa trend that Moroi et al.[28] observed
with increasing chain length for the C5 through C12 PFCAs involves potential stabilization of the undissociated
acid by intramolecular hydrogen bonding between the carboxylic acid and the terminal CF3 group. The C2 through
C4 PFCAs are not likely able to form stable cyclic structures
whereby the acid function interacts with the terminal CF3
group. However, for the C5 PFCAs and higher homologues,
cyclic structures are likely possible, and this may explain
the reported increasing pKa trend with increasing perfluoroalkyl chain length. Hennie and Fox[30] used the inverse
of this argument to rationalize why pKa values declined
in their series of terminal fluorine substituted carboxylic
acids. For hydrocarbon chain carboxylic acids, one would
expect that increasing chain length, and the corresponding
increased ability to form cyclic structures, would increase
the acidity (i.e., decrease pKa values), since the cyclization
would allow the dissociated acid to intramolecularly hydrogen bond to alkyl CH2 units.
This process would promote acid dissociation, due to the
stabilization of the dissociated form, more than would be
expected on a simple chain length argument. In contrast,

321
the fluorines in the PFCA chains carry net partial negative
charges, and any cyclic structures for the longer chain
PFCAs would promote a stabilization of the undissociated acid. This process possibly explains the published
experimental pKa datasets for these contaminants, and
why the work of Moroi et al.[28] reports a critical point at
the C5 PFCAs for the slope of pKa versus perfluoroalkyl
chain length. In other words, the existing published PFCA
datasets show that from C2 through C4 , the slope of pKa
versus chain length is negative which changes to a positive
slope for the C5 and higher homologues, possibly due
to the influence of this intramolecular cyclization and
hydrogen bonding argument.
Our current computational methods (e.g., PM6, SPARC,
COSMOtherm) are not able to explicitly account for the
potential contributions of cyclization in aqueous solution
and resulting intramolecular hydrogen bonding effects such
as these. This perhaps explains the apparent deviation between experimental and computational acidity data for the
C6 and higher PFCA homologues, as well as the known
difficulty in predicting the partitioning behavior of these
compounds.[18] We also note the inability to correct the computationally estimated pKa values using available experimental data shown in Table 2. There is no significant change
in the estimated pKa values with chain length, precluding
any reliable regression analyses against the experiment data.
The accurate determination and estimation of PFCA
pKa values greatly impacts our mechanistic understanding and ability to experimentally test and model their environmental and toxicological fate and propensity for bioaccumulation and bioconcentration. These compounds are
known to concentrate in biological tissues, preferentially in
regions such as the liver, kidney, and blood plasma.[31−36]
The toxicity of PFCAs is also thought to increase with increasing perfluoroalkyl chain length,[32,37,38] and the clearance rates decrease with longer chains.[35] For example,
PFCA distributions into rat urine or bile depend on carbon chain length, with C9 or longer perfluoroalkyl chains
preferentially partitioning into bile. Similarly, peroxisomal
enzyme activity induction requires chain lengths of C8 and
higher, and impacts on hepatic phospholipid metabolism is
only observed for C9 or longer chains.[37]
It is unclear whether these effects are due to the increasing hydrophobicity of the chain itself (and hence, bioaccumulation potential)[33,34,39] or due to changes in the acidity
of the carboxylate group, or both. Limited work to date
on branched versus linear PFOA mixtures suggests no significant difference in their toxicity.[40] Although this may
suggest that the carboxylate acidity plays a negligible role
in the toxicology, further studies are needed — particularly
on a congener-specific basis. At present, we have no reliable data on the relative bioaccumulation and bioconcentration tendencies of the branched versus linear congeners,
although our group has recently put forward initial efforts
in developing a framework for assessment based on the hydrophobicity of the perfluoroalkyl chain.[39]

Downloaded By: [Rayne, Sierra][Canadian Research Knowledge Network] At: 18:24 30 January 2009

322
Better understanding of PFCA ionization properties
also has potential practical applications including in the
design and optimization of environmental treatment processes by helping to reduce current uncertainties regarding
mechanisms of PFCA remediation processes, including
thermal approaches,[41] sonochemical methods,[42,43]
redox techniques,[44,45] direct and catalyzed photodegradation,[46−50] as well as fate during water and wastewater treatment[4,9,10−13,51,52] where pKa dependent physico
chemical processes (e.g., sorption, volatilization) may be important during various unit processes and operations (e.g.,
activated carbon and zeolite sorption, scavenging during
activated sludge settling and metal salt/polymer induced
coagulation/flocculation, dissolved air flotation, nanofiltration, etc.). In addition, a better understanding of PFCA
solution phase conformations, as inferred by 19 F-NMR
studies in DMSO-d6 and D2 O, could be obtained and possibly assist in explaining current reports of distinct changes
in geometry and perfluoroalkyl chain rigidity between 8
and 10 carbon lengths.[53] Furthermore, optimizing both
LC[54−60] and GC based analytical methods and possible
prior derivatization approaches[61] also depends on a more
complete knowledge base regarding PFCA ionization.
The relative contributions of branched and linear PFCAs
in environmental samples has been put forward as a means
of discriminating between potential sources and evaluating
the relative magnitudes of aqueous and atmospheric transport pathways.[2] For example, it has been hypothesized that
PFCA signatures with dominantly linear congeners in each
homologue are likely derived from abiotic and biotic oxidation of the predominantly linear telomer alcohol and olefin
perfluoroalkyl precursors,[3,62−71] whereas signatures with
substantial contributions (perhaps up to 30%) of branched
congeners are likely directly derived from electrochemical fluorination based PFCA commercial products[2] and
degradation products of perfluoroalkyl sulfonamides.[72,73]
We stress that recent environmental monitoring data reasonably calls into question the validity of assuming that
branched:linear congener ratios measured in biological
monitoring studies represent exposure ratios, with studies proposing that differential pharmacokinetic behavior
plays a role in the observed branched:linear ratios found
in biological samples.[6,74] The possible higher acidities of
the branched versus linear congeners, as our PM6 results
presented here suggest, may be an underlying cause of this
proposed pharmacokinetic effect if ionization behavior results in lower bioconcentration/bioaccumulation and/or
higher clearance rates for the branched isomers. Alternately,
as we have already proposed elsewhere,[39] branched congeners are expected to have less hydrophobic perfluoroalkyl
chains than their linear counterparts. If chain hydrophobicity dominates PFCA bioaccumulation and bioconcentration behavior, then there is no need to invoke a contributory carboxylate ionization mechanism. However, the
relative importance of these mechanisms is speculative in
the absence of reliable congener-specific pKa data.

Rayne et al.
As well, if branched and linear congeners have differing
acidities within a homologue group (as the computational
data suggests), and the pKa values change substantially
with the length of the perfluoroalkyl chain (as the experimental data suggests, but with which the computational
data disagrees), this may help explain dissolved versus particulate congener signatures in aquatic systems. Although
limited work has been conducted on sorption of PFCAs to
inorganic mineral surfaces in aquatic systems, with the only
study using iron oxide,[5] the hydrocarbon fatty acid analogs
are known to be significantly sorbed to cations (e.g., Ca2+ )
in minerals such as gypsum, dolomite, calcite, and clays in
real and model solutions, with levels of sorption increasing
with longer chain lengths.[75−80] In aqueous solutions, we
may expect the order of ion-pairing—also important for in
vivo processes—for freely dissolved dissociated PFCAs to
be Na+ >Li+ >K+ >NH+
4 based on analogous work with
other carboxylates.[81] Overall, there remains a high level of
uncertainty as to the importance of these possible carboxylate group acidity driven sorption and ion-pairing processes
for PFCAs.
In general, the acidity of the carboxylate function—
and particularly the uncertainty about the corresponding constants—has not been explicitly incorporated into
partitioning[5,20] and local- through global-scale multimedia transport and fate models.[14,82] The acidity of these
compounds will also govern their partitioning into surface water microlayers and[83] atmospheric aerosols,[16]
a potentially important component of their long-range
transport.[84,85] Given that the pH of atmospheric aerosols
can range from zero to >5 (i.e., spanning across the range of
experimental pKa values for all PFCAs),[86] accurate knowledge of PFCA pKa values will greatly assist in our modeling
of sorption to, and incorporation into, these phases. Of immediate relevance, reliable pKa data and estimation methods for PFCAs also allows us to better model atmospheric
partitioning processes and interpret reported samples[87−92]
and to assess the potential risks posed by extraction into
food and consumer products.[93−96]
For example, if the computational pKa estimates of −0.1
to +0.7 for PFOA are more accurate than the reported
experimental values of 2.3 to 3.4, then current estimates
of air-water partitioning coefficients that only used pH
0.6 solutions[97] would be incorrect. Accurate determinations of the Henry’s law constant and rates of atmospheric
scavenging for airborne ionizable contaminants are pKa
dependent,[98] such that for a generic acid HA, KH =
aHA /pHA , where KH (mol/kg/atm) is the Henry’s law constant, aHA (mol/kg) is the activity of the undissociated acid,
and pHA is the partial pressure of HA in the gas phase
(atm).[99] Previous work on trifluoroacetic acid has indicated that the majority of the ±25% uncertainty in its KH
value is due to corresponding uncertainty regarding its pKa
value. The sensitivity of KH estimates on pKa values is evident where a 0.05 pH unit variation in the estimated pKa
value of trifluoroacetic acid changes the KH estimate by

323

Downloaded By: [Rayne, Sierra][Canadian Research Knowledge Network] At: 18:24 30 January 2009

Long-chain PFCA congeners and pKa values
∼8%.[100] Thus, the absolute value of the PFCA acidity
constants will also determine whether wet or dry deposition modes are more important in atmospheric scavenging
processes.
As a specific example regarding the importance of reliable
computational methods for estimating PFCA pKa values
as the research community moves toward congener-specific
environmental modeling of this contaminant class, we have
calculated the percent differences in KH values for the C2
through C11 PFCAs where the computational pKa data in
Table 2 is used in place of the available experimental values
for each straight chain isomer over the pH range from zero
to +14 (Fig. 2). To perform these calculations, we have followed the theoretical framework set out by Barton et al.,[15]
whereby the pure component air-water partition coefficient
(e.g., KH ) must be corrected to properly reflect the proportions of volatile (i.e., undissociated acid) and non-volatile
(i.e., dissociated acid) components at specific environmental
pH values. The corrected parameter is the effective Henry’s

law constant (KH,eff ), which can be calculated at a specific
pH using the following modified form of the HendersonHasselbach equation: KH,eff =KH /(1+10pH−pKa ). Since
KH is a constant, the ratio of the effective KH values using
the computationally estimated (KH,eff(est) ) and experimental (KH,eff(exp) ) pKa values (pKa,est and pKa,exp , respectively)
can be calculated as follows:
KH,eff(est) /KH,eff(exp)
= (1 + 10pH - pKa,exp )/(1 + 10pH - pKa,est ).
As is evident in Fig. 2, all computational approaches
lead to significant errors in the estimated Keff value, with
ratios ranging from about 10−4 to 101 at environmentally
relevant pH values for the C12 linear PFCA. Thus, the
high sensitivity of atmospheric modeling constants to the
uncertainty in PFCA acidity attests to the need for using
caution in computationally derived data for this group of
contaminants.

Fig. 2. Ratio (KH,eff(est) /KH,eff(exp) ) of the effective Henry’s law constant values for the linear C2 through C11 PFCAs using the
computationally estimated pKa values (yielding KH,eff(est) ) from the PM6 (solid line), SPARC (dashed line), and COSMOtherm (dotdot-dashed line) methods relative to the reported experimental pKa values (yielding KH,eff(exp) ) in Table 2 over the pH range from
zero to +14. Where more than one experimental pKa value has been reported for a particular PFCA isomer, the following composite
values were used for the calculations: C2 -PFCA, 0.45; C4 -PFCA, 0.3; C8 -PFCA, 2.5.

324

Downloaded By: [Rayne, Sierra][Canadian Research Knowledge Network] At: 18:24 30 January 2009

Conclusion
The PM6 semiempirical basis set derived full congener C2
through C9 and straight chain C10 through C13 PFCA
pKa data presented here agree very well with previously
published estimates using the SPARC and COSMOtherm
computational approaches for a subset of the C2 through
C12 straight chain congeners. However, all computational
methods used to date appear to only agree with the established experimental data for the shorter straight chain C2
through C5 compounds. As the perfluoroalkyl chain length
increases beyond C6 , all approaches underestimate the experimental pKa . This deviation between experimental and
estimated values increases from about one pKa unit for C6
to about three pKa units for C12 . Thus, it appears that our
current computational tools may be incapable of reliably
estimating pKa values for the longer-chain PFCAs. This
uncertainty in PFCA acidity has large impacts on our understanding regarding the mechanisms and magnitudes of
environmental fate processes and toxicological effects, warranting additional fundamental studies and the development of improved computational tools for estimating the
congener-specific physico-chemical properties of this contaminant class.

Acknowledgments
S.R. thanks the Natural Sciences and Engineering Research
Council (NSERC) of Canada for financial support.

References
[1] Kissa, E. Fluorinated surfactants and repellants, 2nd ed.; Marcel
Dekker: New York, 2001.
[2] Prevedouros, K.; Cousins, I.T.; Buck, R.C.; Korzeniowski, S.H.
Sources, fate and transport of perfluorocarboxylates. Environ. Sci.
Technol. 2006, 40, 42–44.
[3] Schenker, U.; Scheringer, M.; Macleod, M.; Martin, J.W.; Cousins,
I.T.; Hungerbuhler, K. Contribution of volatile precursor substances to the flux of perfluorooctanoate to the arctic. Environ.
Sci. Technol. 2008, 42, 3710–3716.
[4] Plumlee, M.H.; Larabee, J.; Reinhard, M. Perfluorochemicals in
water reuse. Chemosphere 2008, 72, 1541–1547.
[5] Higgins, C.P.; Luthy, R.G. Sorption of perfluorinated surfactants
on sediments. Environ. Sci. Technol. 2006, 40, 7251–7256.
[6] Conder, J.M.; Hoke, R.A.; de Wolf, W.; Russell, M.H.; Buck, R.C.
Are PFCAs bioaccumulative? A critical review and comparison
with regulatory criteria and persistent lipophilic compounds. Environ. Sci. Technol. 2008, 42, 995–1003.
[7] Houde, M.; Martin, J.W.; Letcher, R.J.; Solomon, K.R.; Muir,
D.C.G. Biological monitoring of perfluoroalkyl substances: A review. Environ. Sci. Technol. 2006, 40, 3463–3476.
[8] Giesey, J.P.; Kannan, K. Perfluorochemical surfactants in the environment. Environ. Sci. Technol. 2002, 36, 146A–152A.
[9] Ochoa-Herrera, V.; Sierra-Alvarez, R. Removal of perfluorinated
surfactants by sorption onto granular activated carbon, zeolite and
sludge. Chemosphere 2008, 72, 1588–1593.

Rayne et al.
[10] Steinle-Darling, E.; Reinhard, M. Nanofiltration for trace organic
contaminant removal: Structure, solution, and membrane fouling
effects on the rejection of perfluorochemicals. Environ. Sci. Technol. 2008, 42, 5292–5297.
[11] Becker, A.M.; Gerstmann, S.; Frank, H. Perfluorooctane surfactants in waste waters, the major source of river pollution. Chemosphere 2008, 72, 115–121.
[12] Boulanger, B.; Vargo, J.D.; Schnoor, J.L.; Hornbuckle, K.C. Evaluation of perfluorooctane surfactants in a wastewater treatment
system and in a commercial surface protection product. Environ.
Sci. Technol. 2005, 39, 5524–5530.
[13] Sinclair, E.; Kannan, K. Mass loading and fate of perfluoroalkyl
surfactants in wastewater treatment plants. Environ. Sci. Technol.
2006, 40, 1408–1414.
[14] Armitage, J.; Cousins, I.T.; Buck, R.C.; Prevedouros, K.; Russell, M.H.; Macleod, M.; Korzeniowski, S.H. Modeling globalscale fate and transport of perfluorooctanoate emitted from direct
sources. Environ. Sci. Technol. 2006, 40, 6969–6975.
[15] Barton, C.A.; Kaiser, M.A.; Russell, M.H. Partitioning and removal of perfluorooctanoate during rain events: The importance
of physical-chemical properties. J. Environ. Monitor. 2007, 9, 839–
846.
[16] McMurdo, C.J.; Ellis, D.A.; Webster, E.; Butler, J.; Christensen,
R.D.; Reid, L.K. Aerosol enrichment of the surfactant PFO and
mediation of the water-air transport of gaseous PFOA. Environ.
Sci. Technol. 2008, 42, 3969–3974.
[17] Goss, K.-U. The pKa values of PFOA and other highly fluorinated
carboxylic acids. Environ. Sci. Technol. 2008, 42, 456–458.
[18] Arp, H.P.H.; Niederer, C.; Goss, K.-U. Predicting the partitioning
behavior of various highly fluorinated compounds. Environ. Sci.
Technol. 2006, 40, 7298–7304.
[19] Goss, K.-U.; Bronner, G.; Harner, T.; Hertel, M.; Schmidt, T. The
partition behavior of fluorotelomer alcohols and olefins. Environ.
Sci. Technol. 2006, 40, 3572–3577.
[20] Higgins, C.P.; Luthy, R.G. Modeling sorption of anionic surfactants onto sediment materials: An a priori approach for perfluoroalkyl surfactants and linear alkylbenzene sulfonates. Environ.
Sci. Technol. 2007, 41, 3254-3261.
[21] Rayne, S.; Forest, K.; Friesen, K.J. Congener-specific numbering
systems for the environmentally relevant C4 through C8 perfluorinated homologue groups of alkyl sulfonates, carboxylates, telomer
alcohols, olefins, and acids, and their derivatives. J. Environ. Sci.
Health A 2008, 43, 1391–1401.
[22] Allinger, N.L. Conformational analysis. 130. MM2. A hydrocarbon force field utilizing V1 and V2 torsional terms. J. Amer. Chem.
Soc. 1977, 99, 8127–8134.
[23] Stewart, J.J.P. Optimization of parameters for semi-empirical methods I—Method. J. Comp. Chem. 1989, 10, 209—220.
[24] Stewart, J.J.P. MOPAC: A general molecular orbital package.
Quant. Chem. Prog. Exch. 1990, 10, 86.
[25] MOPAC2007. James J.P. Stewart, Stewart Computational Chemistry, Colorado Springs, CO, USA. http://OpenMOPAC.net.
[26] Brace, N.O. Long chain alkanoic and alkenoic acids with perfluoroalkyl terminal segments. J. Org. Chem. 1962, 27, 4491–
4498.
[27] Henne, A.L.; Fox, C.J. Ionization constants of fluorinated acids.
J. Am. Chem. Soc. 1951, 73, 2323–2325.
[28] Moroi, Y.; Yano, H.; Shibata, O.; Yonemitsu, T. Determination
of acidity constants of perfluoroalkanoic acids. Bull. Chem. Soc.
Japan 2001, 74, 667–672.
[29] Webster E. The Canadian Environmental Modeling Network
newsletter. Autumn 2007. www.trentu.ca/cemn/NewsReports/
CEMNnews200710.pdf. Accessed July 24, 2008.
[30] Henne, A.L.; Fox, C.J. Ionization constant of fluorinated acids II.
J. Am. Chem. Soc. 1953, 75, 2750–2751.
[31] Vanden Heuvel, J.P.; Kuslikis, B.I.; Van Rafelghem, M.J.; Peterson, R.E. Tissue distribution, metabolism, and elimination of


Related documents


j environ sci heal a 44 2009 317 326
environ sci technol 43 2009 7155 7156
j environ sci health a 44 2009 1374 1387
j environ sci health a 44 2009 1388 1399
j environ sci health a 44 2009 598 604
j environ sci health a 44 2009 1145 1199


Related keywords