Gas phase UV and IR absorption spectra of CxF2x+1CHO (x = 1-4)

Article abstract of DOI:10.1016/j.jfluchem.2004.07.006

The UV and IR spectra of C_xF_{2x +1}CHO (x = 1-4) were investigated using computational and experimental techniques. C_xF_2x+1CHO (x = 1-4) have broad UV absorption features centered at 300-310 nm. The maximum absorption cross-section increases significantly and shifts slightly to the red with increased length of the C_xF_2x+1 group: CF₃CHO, 3.10 × 10^-20 (300 nm); C ₂F₅CHO, 6.25 × 10^-20 (308 nm); C ₃F₇CHO, 8.96 × 10^-20 (309 nm); and C ₄F₉CHO, 10.9 × 10^-20 (309 nm). IR spectra for C_xF_2x+1CHO were recorded, calculated, and assigned. Results are discussed with respect to the literature data and to the atmospheric fate of CxF2x+1CHO.

Full text of DOI:10.1016/j.jfluchem.2004.07.006

Journal of Fluorine Chemistry 125 (2004) 1925–1932
www.elsevier.com/locate/fluor
Gas phase UV and IR absorption spectra of C F CHO (x = 1–4)
x 2x+1
a
Y. Hashikawa , M. Kawasaki , R.L. Waterland , M.D. Hurley , J.C. Ball ,
a
b,
d
d
*
d
T.J. Wallington , M.P. Sulbaek Andersen , O.J. Nielsen
c
c
a
Department of Molecular Engineering and Graduate School of Global Environmental Studies, Kyoto University,
Kyoto 615-8510, Japan
b
DuPont Engineering Research and Technology, E. I. du Pont de Nemours and Co. Inc., P. O. Box 80249,
Wilmington, DE 19880-0249, USA
c
University of Copenhagen, Universitetsparken 5, DK-2100 Copenhagen, Denmark
d
Ford Motor Company, P.O. Box 2053, Dearborn, MI 48121-2053, USA
Received 24 June 2004; received in revised form 21 July 2004; accepted 21 July 2004
Available online 11 September 2004
Abstract
The UVand IR spectra of CxF2x+1CHO (x = 1–4) were investigated using computational and experimental techniques. CxF2x+1CHO (x = 1–
) have broad UV absorption features centered at 300–310 nm. The maximum absorption cross-section increases significantly and shifts
4
slightly to the red with increased length of the C
ꢁ20
ꢁ20
x
F
2x+1 group: CF
3
CHO, 3.10 ꢀ 10
(309 nm). IR spectra for C F
x 2x+1
(300 nm); C
2
F
5
CHO, 6.25 ꢀ 10
(308 nm);
CHO were recorded, calculated, and
ꢁ20
ꢁ20
C F CHO, 8.96 ꢀ 10
(309 nm); and C F CHO, 10.9 ꢀ 10
3
7
4
9
assigned. Results are discussed with respect to the literature data and to the atmospheric fate of CxF2x+1CHO.
2004 Elsevier B.V. All rights reserved.
#
Keywords: Absorption spectra; PFOA; PFCA
1
. Introduction
blends treated with such polymers. Their study, which was
conducted at conditions typical for municipal incinerators,
concluded that no detectable levels of PFOA result from
incineration of these fluorinated materials.
There is growing interest in the environmental fate and
effects of fluorinated chemicals. Perfluorooctane sulfonate
ꢁ
(
PFOS, C F SO ) has been shown to be widely distributed
17
Fluorinated surfactants are commonly used as processing
aids in the production of fluoropolymers and fluoroelasto-
mers [7]. In most commercial processes, fluoropolymers are
heated in excess of 350 8C to allow melt processing or
sintering of the resin particles and it has been proposed that
thermal degradation of residual surfactant during such
processing could lead to PFCAs or their derivatives.
However, a recent kinetic study of the thermal degradation
8
3
in trace quantities in the Northern Hemisphere [1], and very
recently, derivatives of perfluorooctanoic acid (PFOA,
C F COOH) and of other perfluoroalkyl carboxylic acids
7
15
(
PFCAs, CxF2x+1COOH, where x = 6–12) have been
observed in trace quantities in fish [2,3] and mammals [4]
in remote locations.
The source of PFCA derivatives is unknown at present but
several possibilities have been examined. Mabury and
coworkers [5] have studied the thermolysis of fluoropoly-
mers as a potential source of PFCAs in the environment.
Recently, Yamada and Taylor [6] have studied the
incineration of fluorinated acrylic polymers and fabric
ꢁ
of ammonium perfluorooctanoate (APFO, C F COO
15
8
+
NH ), the carboxylate surfactant most commonly used as
4
a processing aid, showed no evidence for production of
PFOA [8].
PFCAs observed in remote locations may arise from a
variety of competing pathways including, but not limited
to, atmospheric direct and indirect transport and poorly
understood food chain processes. Much attention has
*
Corresponding author. Tel.: +1302 6951511; fax: +1 302 6958805.
E-mail addresses: robert.l.waterland@usa.dupont.com (R.L. Waterland),
twalling@ford.com (T.J. Wallington).
0022-1139/$ – see front matter # 2004 Elsevier B.V. All rights reserved.
doi:10.1016/j.jfluchem.2004.07.006

1926
Y. Hashikawa et al. / Journal of Fluorine Chemistry 125 (2004) 1925–1932
focussed on PFCA production from atmospheric degrada-
tion of precursor chemicals. Unfortunately, the identity of
any atmospheric precursors and the mechanism by which
they are converted into PFCAs are unclear at the present
time.
photodissociation pathway as an atmospheric loss process
for CxF2x+1CHO.
CxF2xþ1CHO þ hn ! CxF2xþ1 þ HCO
(13)
Photolysis via reaction will compete with reaction and
suppress the formation of PFCAs by the mechanism
described above. Two pieces of information are required
to incorporate reaction (13) into models of the atmospheric
chemistry of FTOHs and assess the importance of this
process. First, we need the UV absorption spectra of
CxF2x+1CHO. Second, we need photodissociation quantum
yields for CxF2x+1CHO as a function of wavelength.
Fluorotelomer alcohols, (FTOHs, CxF2x+1CH CH OH)
2
2
are chemical intermediates commonly used in the manu-
facture of fluorotelomer-based products. Although, the
environmental fate of fluorotelomer-based substances is
not fully determined, it has been suggested that atmos-
pheric oxidation of FTOHs may be a source of PFCAs
in the environment [9]. Sulbaek Andersen et al. [10]
have proposed a three step mechanism. First, FTOHs
react with OH radicals leading to formation of perfluorinated
aldehydes. In the second step, OH radicals react with
the newly-formed perfluorinated aldehydes to give per-
While the UV absorption spectrum of CF CHO has been
3
measured in several studies [12–16] and is well defined,
there is little information available concerning the UV
absorption spectra of longer chain, and potentially
environmentally more significant, perfluoroaldehydes. At
the start of the present work there was no information
available concerning the photodissociation quantum yields
of CxF2x+1CHO. During the course of the present work,
Sellev a˚ g et al. [16] reported an upper limit of 0.02 for the
fluoroacetyl peroxy radicals, CxF2x+1C(O)O . Finally,
2
reaction of CxF2x+1C(O)O2 with HO2 radicals produces
PFCAs [10].
ꢂ
CxF2 CH CH OH þ OH ! CxF
CH C HOH þ H O
xþ1
2xþ1
2xþ1
2xþ1
2xþ1
2
2
2xþ1
2
2
(
1)
2)
3)
effective photodissociation quantum yield of CF CHO in
3
1
atm of air with 290–400 nm radiation.
To facilitate the inclusion of reaction in atmospheric
CxF
CxF
CxF
CxF
CH CHOH þ O ! C F
CH CHO þ HO2
2
2
x
2xþ1
2
(
(
chemistry models and the experimental study of the
atmospheric chemistry of CxF2x+1CHO, a computational
and experimental study of the UV and IR spectra of
CxF2x+1CHO (x = 1–4) was performed in our laboratories.
Results are reported herein. Straight chain isomers n-
C F CHO and n-C F CHO were studied in the present work.
ꢂ
CH CHO þ OH ! C F
CH C ðOÞ þ H O
2
x
2xþ1
2
2
ꢂ
ꢂ
2
CH C ðOÞþ O þ M ! C F
CH CðOÞO þM
2
2
x
2xþ1
2
3
7
4 9
(4)
For simplicity we will refer to these species as C F CHO and
3
7
CH CðOÞO2ꢂ þ RO
2
2
C F CHO in the rest of this article.
4 9
ꢂ
!
CxF
CH CðOÞO þ RO þ O
(5)
(6)
(7)
2
xþ1
2
2
ꢂ
ꢂ
CxF2 CH CðOÞO ! CxF
CxF2 CH2 þ O þ M ! CxF
CxF2 CH O þ RO ! CxF
CH2 þ CO2
xþ1
2
2xþ1
2. Experimental and computational details
ꢂ
ꢂ
CH O2 þ M
CH O þ RO þ O
xþ1
2
2xþ1
2
Samples of CxF2x+1CHO (x = 1–4), were synthesized
ꢂ
ꢂ
xþ1
2
2
2
2xþ1
2
2
from the corresponding hydrates CxF2x+1CH(OH) (x = 1–4)
2
(
8)
(9)
(10)
using the procedure previously described for C F CHO [17].
2
5
ꢂ
The samples were purified by vacuum distillation.
CxF2 CH O þ O ! CxF
CHO þ HO2
xþ1
2
2
2xþ1
ꢂ
CxF2xþ1
CxF
CHO þ OH ! C F
x
2xþ1
CðOÞ þ H O
2
2.1. UV measurements
CðOÞ þ O þ M ! C F2xþ1CðOÞO2ꢂ þ M (11)
ꢂ
2xþ1
2
x
UV absorption spectra of CxF2x+1CHO (x = 1–4) were
ꢂ
CxF2xþ1CðOÞO2 þ HO ! CxF2xþ1CðOÞOH þ O3
(12)
2
measured in the 200–500 nm wavelength range using a
This mechanism will not be effective in air masses with high
ꢂ
commercial dual beam UV spectrometer (Lambda 18,
Perkin-Elmer) operated at a spectral resolution of 1.0 nm.
Gaseous samples of CxF2x+1CHO were introduced using a
Pyrex gas handling system with greaseless valves into a
NO (e.g., urban areas) where loss of C F2x+1C(O)O2 will
x
x
be dominated by reaction with NO . For the above mechan-
x
ism to be a significant source of PFCAs, reaction with OH
radicals (reaction 10) must be an important loss mechanism
for CxF2x+1CHO. Photolysis is a likely alternate loss
mechanism that will compete directly with reaction (10).
Non-halogenated aldehydes absorb in the UV at 250–
5.8 cm long Pyrex cell equipped with BaF windows. Gas
2
pressures were in the range 15–150 Torr as measured using
with capacitance manometers (MKS). No diluent gas was
added. All measurements were performed at 296 K. The
purity of compounds was checked using FTIR spectroscopy
before and after the UV measurements; no impurities were
evident in the IR spectra.
3
50 nm. It is well established that UV-induced C–C bond
scission is an important atmospheric loss mechanism for
aldehydes [11]. It seems reasonable to propose an analogous

Y. Hashikawa et al. / Journal of Fluorine Chemistry 125 (2004) 1925–1932
1927
2
.2. IR measurements
benzene and methane gave the correlation
Eexpt ¼ 1:144 E
with an R of 0.961.
ꢁ 0:553 eV
calc
The experimental setup consisted of a Mattson Instru-
2
ments, Sirius 100 Fourier transform infrared spectrometer,
interfaced to a 140 l, 2 m long evacuable Pyrex chamber
described elsewhere [18]. The spectrometer was operated at
For the estimation of IR spectra, we re-optimized all
molecular geometries using B3LYP with the 6-31G* basis
set [43] and computed harmonic vibrational frequencies and
intensities at the same level of theory on the resultant
optimized geometries. Vibrational frequencies were scaled
by 0.961 as recommended by Scott and Radom [44].
ꢁ
1
a spectral resolution of 0.50 cm . Infrared spectra were
derived from 32 co-added interferograms. Spectra were
recorded at 296 K in the presence of 700 Torr of air diluent.
We estimate the absorption cross-section measurement to be
accurate to within Æ5% [18].
3. Results
2
.3. Computational details
3.1. Measured UV spectra
Time-dependent density functional theory (TD-DFT) is
rapidlybecoming the standard method for rapid and accurate
prediction of the properties of electronically excited states
and for the estimation of UV and visible spectra. TD-DFT
was first proposed by Ando [19] in the 1970s. Despite
subsequent development and application to atoms [20–22]
and solids [23,24], TD-DFT has only recently been applied
to molecules. The first substantial molecular calculations
were published in the mid to late 1990s [25,26]. Recently,
a number of groups have used TD-DFT to predict UV
and visible spectra for a broad array of molecular species
As an initial control experiment, the UV spectrum of
CH CHO was measured using the same techniques and
3
apparatus used to study CxF2x+1CHO. The spectrum of
CH CHO measured in the present work (solid trace) is
3
compared to the recommended literature spectrum [45]
(circles) in the bottom panel of Fig. 1. The excellent
agreement between the spectrum of CH CHO measured in
3
[
27–33] and in each of these studies, comparison of cal-
culated and experimental transition energies showed that
TD-DFT significantly outperforms older Hartree–Fock-
based methods such as configuration interaction singles
(
CIS) [34]. Of particular importance for the present work,
studies by Matsuzawa and co-workers [35] for about 40
molecules spanning a broad variety of fluorinated organics
including ethers, aldehydes, acids and esters and Waterland
et al. [36] for fluoroalkanes demonstrated that TD-DFT
calculations of photoabsorption spectra, which incorporate
empirical correction of the transition energy are very useful
aids for molecular design.
In the present work, all calculations were performed
using the Gaussian 03 suite of programs [37]. We used the
gradient-corrected level of density-functional theory (DFT)
utilizing Becke’s three-parameter exchange functional [38]
and the Lee–Yang–Parr correlation functional [39]
(
B3LYP).
Molecular geometries were optimized using the DFT-
derived DZVP basis set [40] and the same basis set was used
for frequency calculations to ensure that computed
geometries corresponded to bound molecular states. Vertical
excitation energies and oscillator strengths were calculated
with time-dependent density-functional theory (TD-DFT)
Fig. 1. UV spectra for CH
and C CHO, C CHO, and C
present work (solid traces). The circles in the lower panel are the spectrum
of CH
CHO recommended by N o¨ lle et al. [45]. The filled triangles, open
triangles, open circles, and blue circles in the middle panel are the data for
CF CHO from Borkowski and Ausloos [12], Francisco and Williams [14],
Meller et al. [13], and Sellev a˚ g et al. [16], respectively. Triangles and
diamond in the top panel are data for C CHO from Borkowski and
Ausloos [12] and Pritchard et al. [46], respectively. The inverted open
triangle is a datum for C CHO from Pritchard et al. [46].
3
CHO (bottom panel), CF
3
CHO (middle panel)
2
F
5
3
F
7
4 9
F
CHO (upper panel) recorded in the
[
25,41]. The TD-DFT calculations employed 50 states and
3
used the DZVP basis set augmented with Dunning and
Hay’s Rydberg functions [42] on all heavy atom centers.
We incorporated energy re-scaling in the same manner as
Matsuzawa et al. [35] and Waterland et al. [36]. Our empi-
rical fitting of the experimental spectra of formaldehyde,
3
2 5
F
3 7
F

1928
Y. Hashikawa et al. / Journal of Fluorine Chemistry 125 (2004) 1925–1932
the present work and the literature data provides confidence
in our experimental methodology.
As seen from the top panel of Fig. 1, the spectra of
C F CHO and C F CHO measured in the present work are
2
5
3 7
UV spectra of CxF2x+1CHO (x = 1–4) were measured over
the wavelength range 200–500 nm. The UV spectra of at
least three samples of each gas were measured as a check of
both experimental reproducibility and linearity of absorption
with reagent concentration. The insert in the top panel of
Fig. 1 shows the observed absorbance at 309 nm as a
function of the concentration of C F CHO and serves to
in good agreement with the absorption cross-sections
reported by Borkowski and Ausloos [12], and Pritchard et
al. [46] for these species. There are no literature data for
C F CHO with which we can compare our results.
4
9
3.2. Measured IR spectra
4
9
illustrate the typical linearity and reproducibility observed
during the present work. Linear least squares analysis of the
data in the insert gives a value of s(309 nm, C F CHO) =
IR spectra of CxF2x+1CHO (x = 1–4) measured in the
present work are shown in Fig. 2. As seen from the bottom
panel the measurements in the present work are in excellent
agreement with the previous work by Sellev a˚ g et al. [16].
Integrated absorption intensities over the wavenumber range
4
9
ꢁ19
2
ꢁ1
1
.09 ꢀ 10
cm mol . As indicated by the data in the
insert in Fig. 1, the measurements were reproducible to
within Æ3%. We estimate that systematic uncertainties
associated with the pressure and path length measurements
add an additional 5% uncertainty to give a final value of
ꢁ
1
670–3000 cm measured in the present work were S = 1.30,
ꢁ1
ꢁ
16
1.67, 1.97, and 2.29 ꢀ 10
cm mol
C F CHO, C F CHO, and C F CHO, respectively. Fran-
for CF CHO,
3
2
5
3
7
4 9
ꢁ
19
2
ꢁ1
s(309 nm, C F CHO) = (1.09 Æ 0.06) ꢀ 10
cm mol .
cisco and Williams [14] have reported integrated absorption
4
9
ꢁ
intensities over the range 801–2890 cm for CF CHO of
3
1
This value was used as a scaling point to derive the UV
spectrum of C F CHO shown in the top panel of Fig. 1. The
spectra of CF CHO, C F CHO, and C F CHO shown in Fig.
ꢁ
17
ꢁ16 ꢁ1
4
9
2.7 ꢀ 10
for pure vapor and 1.07 ꢀ 10
cm mol in
the presence of 150–750 Torr of argon diluent. Sellev a˚ g et al.
3
2
5
3 7
1 were derived in a similar fashion. As seen from Fig 1, the
maximum absorption cross-section increases significantly
and shifts slightly to the red with increased length of the
ꢁ
20
CxF2x+1 group: CF CHO, (3.10 Æ 0.18) ꢀ 10
(300 nm);
(308 nm); C F CHO,
3
ꢁ
20
C F CHO, (6.25 Æ 0.36) ꢀ 10
2
5
3 7
ꢁ
20
(
8.96 Æ 0.52) ꢀ 10
(309 nm); and C F CHO, (10.9 Æ
x 2x+1
4
9
ꢁ
20
0
.6) ꢀ 10
–4) at 10 nm intervals are given in Table 1.
As shown in Fig. 1, our spectrum for CF CHO agrees
(309 nm). Spectral data for C F
CHO (x =
1
3
well with the previous measurements by Borkowski and
Ausloos [12], Francisco and Williams [14], Meller et al.
[
15], and Sellev a˚ g et al. [16]. The spectrum for CF CHO
3
reported by Lucazeau and Sandorfy [13] is similar in shape
to that measured herein but is approximately 30% less
intense. In light of the excellent agreement between the
spectral data reported herein and by Borkowski and Ausloos
[12], Francisco and Williams [14], Meller et al. [13], and
Sellev a˚ g et al. [16] it would appear that Lucazeau and
Sandorfy [13] underestimated the absorption of CF CHO by
3
approximately 30%.
Table 1
ꢁ
20
2
cm mol ) measured in the present work
ꢁ1
Absorption cross-sections (10
Wavelength (nm) CF CHO
3
C
2
F
5
C
3
F
7
4 9
CHO CHO C F CHO
2
2
2
2
2
3
3
3
3
3
3
50
60
70
80
90
00
10
20
30
40
50
0.38
0.84
1.46
2.20
2.86
3.01
2.83
2.15
1.33
0.57
0.22
0.31
0.86
1.78
3.13
4.64
5.75
6.14
5.66
3.84
1.94
1.01
0.42
1.08
2.39
4.16
6.24
7.91
8.76
8.32
5.62
2.94
1.56
0.42
1.35
2.89
5.01
7.63
9.75
10.74
10.22
6.92
3.56
1.82
Fig. 2. IR spectra measured in the present work for C
x
F
2x+1CHO. The insert
in panel D shows a comparison of the spectrum for CF CHO reported by
Sellev a˚ g et al. [16] (circles) and that measured in the present work (solid
3
trace).

Y. Hashikawa et al. / Journal of Fluorine Chemistry 125 (2004) 1925–1932
1929
3
Fig. 3. Lowest energy conformation of CF CHO.
[
16] report an integrated absorption intensity over the
ꢁ1
wavenumber range 775–3000 cm
of S = 1.274 ꢀ
x
Fig. 4. Lowest energy conformer of C F2x+1CHO (x = 2,4).
ꢁ
16
ꢁ1
cm mol for pure CF CHO vapor. As seen from
3
1
0
Fig. 2, there is a small absorption feature centered at
ꢁ
1
7
06 cm
which was not recorded by either Francisco
conformers of the higher aldehydes, CxF2x+1CHO (x =
2,4), have the carbonyl oxygen eclipsing the C–C single
bond of the first tetrafluoroethylene segment. These
conformations are sketched in Fig. 4. C F CHO and
and Williams [14] or Sellev a˚ g et al. [16]. Exclusion of
absorption from this feature gives a value of S = 1.25 ꢀ
ꢁ
16
ꢁ1 ꢁ1
cm mol at 800–3000 cm in the present work. This
1
0
2
5
value is in excellent agreement with the previous work by
Sellev a˚ g et al. [16] and in fair agreement (15% discrepancy)
with that reported by Francisco and Williams [14] in 150–
C F CHO have C symmetry with ground electronic state
3 7 s
1
0
A but, for C F CHO the perfluorinated develops a helical
9
4
structure and the symmetry is lowered to C . The ground
1
1
electronic state of C F CHO is A.
7
50 Torr of argon diluent. The simplest explanation for the
4
9
anomalously low absorption observed by Francisco and
Williams [14] in the absence of diluent gas is saturation of
the CF CHO absorption features in their low pressure
In Table 2 we show the wavelength and oscillator strength
for the first 12 electronic transitions (singlet or triplet) in
each of the perfluoroalkyl aldehydes studied. These
transitions include all those with wavelengths greater than
3
experiments.
150 nm. Since, the solar spectrum below 290 nm is
3
.3. Calculated UV spectra
completely blocked by stratospheric trace gases, a necessary
condition for photolysis in the lower atmosphere is that the
compound absorb UV at wavelengths above 290 nm. We
have included lower wavelengths for completeness.
Before we measured the UV spectra of CxF2x+1CHO (x =
–4), we first calculated them to guide the subsequent
1
experimental measurements. Our first task was to determine
the lowest energy molecular conformers for each of the four
perfluoroalkyl aldehydes. As sketched in Fig. 3, the lowest
energy conformation of CF CHO has C symmetry with the
Singlet–triplet transitions are spin-forbidden and are
marked with an ‘f’ in Table 2. We would not expect any
intensity from these transitions in the experimental spectra.
More surprisingly, in the critical wavelengths above 290 nm,
all of the singlet–singlet transitions have oscillator strength
less than or equal to 0.0001, i.e. they are predicted to be
3
s
carbonyl oxygen eclipsing one of the C–F bonds. The
1
ground electronic state is A . Energetically preferred
0
Table 2
x
Computed singlet and triplet transitions for C F2x+1CHO (x = 1–4)
1
/ A )
0
1
/ A )
0
1
/ A )
0
1
CF CHO (C
3
s
C
2
F
5
CHO (C
s
C
3
F
7
CHO (C
s
4 9 1
C F CHO (C / A)
Upper
state
l
(nm)
calc
Oscillatory
strength
Upper
state
l
calc (nm)
Oscillatory
strength
Upper
state
l
calc (nm)
Oscillatory
strength
Upper
state
l
calc
(nm)
Oscillatory
strength
3
1
3
3
3
1
3
3
3
1
1
1
00
00
0
00
0
0
0
0
0
3
1
3
3
1
3
1
3
3
1
1
1
00
00
0
00
00
0
0
00
0
3
1
3
3
1
3
1
3
3
1
1
1
00
00
0
00
00
0
0
00
0
3
A
A
A
A
A
A
A
A
A
A
A
A
374.7
300.4
219.6
156.3
156.2
153.8
143.0
140.5
140.2
140.2
140.1
139.7
f
A
A
A
A
A
A
A
A
A
A
A
A
380.1
305.2
220.0
167.2
159.8
155.2
152.2
150.4
146.6
143.7
142.5
140.6
f
A
A
A
A
A
A
A
A
A
A
A
A
380.4
306.1
219.5
172.4
166.9
160.2
155.3
152.9
150.1
147.4
147.0
145.4
f
A
380.6
306.5
219.7
177.7
172.2
164.8
158.8
154.0
153.0
151.9
151.5
150.1
f
1
<0.0001
f
<0.0001
f
<0.0001
f
A
0.0001
3
A
f
3
f
f
f
A
f
1
f
0.0005
f
0.0001
f
A
0.0005
f
V
0.0155
f
A
1
0.0238
f
0.0408
f
A
0.0608
f
3
0.0307
f
A
3
f
f
A
f
00
0
0
0
0
00
0
00
0
1
0.0017
0.0118
0.0030
0.0041
0.0371
0.0004
0.0038
0.0002
0.0135
A
0.0052
0.0643
0.0001
1
A
1
A

1930
Y. Hashikawa et al. / Journal of Fluorine Chemistry 125 (2004) 1925–1932
Table 3
x
Observed and calculated vibrational bands (cm ), calculated infrared intensities (km mole ), and approximate band assignments for C F2x+1CHO
ꢁ1
ꢁ1
CF
3
CHO
C
2
F
5
CHO
C
3
F
7
CHO
C
4
F
9
CHO
Approximate assignment
n
obs
n
calc
IR intensity
n
obs
n
calc
IR intensity
n
obs
n
calc
IR intensity
n
obs
n
calc
IR intensity
7
8
06
41
673
34.4
38.0
718
794
664
741
18.4
2.7
745
996
725
965
7.2
745
888
717
934
19.7
5.99
815
70.1
1
1
1
074 1084 116.5
112 1136 155.6
164 1163 168.9
1104
a
1135
1116 174.6
1144 88.4
1174 112.3
1117 172.5
a
b
1144 1149
42.5
Backbone bend
1189
1178 186.5
1220 407.6
a
a
b
1
1
1
1
2
192 1168
459.7
1224 1209 378.2
1243 1216
1311
618.7
68.1
16.4
98.4
63.7
1247
Backbone bend
303
383
787
863
a
1275 136.6
1303
1345 1385
1776 1793
2864 2857
83.6
10.8
99.5
61.2
1310
1374
90.9
6.6
C–C stretch
C–C–H bend
C=O stretch
C–H stretch
1367
1798
2870
35.0
84.7
60.3
1357
1778
2868
1373
1795
2852
1357
1778
2868
1794 102.8
2851 65.6
Superposition of two or more bands: ncalcis set to the mean frequency and intensity is set to sum of individual band intensities.
Out of plane vibrational mode.
b
weak. Strong absorption is predicted to occur at wavelengths
below 290 nm. For example, CF CHO has a strong singlet–
singlet transition at 153.8 nm.
Table 3 shows the experimentally observed vibrational
bands and the corresponding computed vibrational counter-
parts. Theory and experiment agree well for the location of
the bands. Taking all four molecules together, for those
transitions for which both experimental and theoretical data
3
Similar behavior has been reported for formaldehyde.
Experimentally, the UV spectrum of formaldehyde has a
single broad absorption peak centered at 326 nm [47]. For
ꢁ1
exist, the root mean square error is 24 cm
and the
1
*
ꢁ1
this transition, the upper state is designated A (n,p ) [48]
2
maximum absolute error is 54 cm which are typical for
this level of theoretical treatment. Examining the molecules
individually, the root mean square and maximum absolute
and, under electric dipole rules, it is actually forbidden. That
this transition is, in fact, experimentally observed is due to
vibronic interaction as discussed by Dieke and Kistiakowski
ꢁ
1
errors are respectively, 18 and 33 cm for CF CHO, 32 and
3
ꢁ ꢁ1
54 cm for C F CHO, 17 and 27 cm for C F CHO, and
2 5 3 7
ꢁ
27 and 46 cm for C F CHO. Theoretical predictions show
3 7
1
[
49]. Calculation of the vibronic structure of molecules as
1
large as those that are the focus of this study is not feasible at
present but as a check, we calculated an UV spectrum for
formaldehyde using the same method described above. The
calculated spectrum has one singlet–singlet transition at
a modest systematic error for the highest intensity peaks.
However, examination of the three most intense peaks for
ꢁ
1
each molecule shows that the largest error is 28 cm . The
relative computed IR intensities also agree quite well with
experiment. Fig. 5 shows the computed spectrum of
3
19 nm in good agreement with experiment but, in
concordance with the fluorinated analogue, this transition
has oscillator strength less than 0.0001. By analogy,
although we cannot be certain, we think it likely that the
low intensity singlet–singlet transitions shown in Table 2
correspond, in fact, to experimentally observable transitions.
In summary, the ab initio calculations identify the
predicted spectral peak locations and how these peaks shift
as the perfluorinated chain lengthens but do not give the
relative peak intensity. For the fluoraldehydes, Table 2 shows
that only one peak is predicted above 290 nm. The peak
location shifts slightly to the red as the perfluorinated chain
ꢁ1
CF CHO in the region 1100–1500 cm . The theoretical
3
spectrum has been artificially broadened to allow compar-
ison with the corresponding experimental spectrum in Fig.
2D. The tendency to underpredict peak location is seen here
lengthens. The largest change is from CF CHO (300 nm) to
3
C F CHO (305 nm) but slight additional red shifts occur for
2
5
C F CHO (306 nm) and C F CHO (307 nm).
3
7
4 9
3
.4. Calculated IR spectra
The calculation of vibrational normal modes and the
corresponding harmonic vibrational frequencies is a routine
operation for modern ab initio methods. In this work, we
have followed the recommendation of Scott and Radom
ꢁ
1
[
44]. This method gave a root mean square error of 34 cm
Fig. 5. Computed IR spectrum of CF
3
CHO. The lines have been artificially
for a set of 122 molecules and 1066 experimental vibrational
frequencies.
broadened with a Lorentzian lineshape (full width at half maximum of
ꢁ
1
20 cm ) for comparison with Fig. 2D.

Y. Hashikawa et al. / Journal of Fluorine Chemistry 125 (2004) 1925–1932
1931
but the relative intensity of the three strong peaks is well
reproduced.
oalkyl aldehydes, CxF2x+1CHO, and it is not possible to
calculate their photolysis rate at this time.
By visual inspection of each computed normal mode,
we have attempted to describe the character of the modes
in terms of stretching and bending deformations. These
descriptions are approximate and the true normal
modes usually involve a complicated mixture of atomic
motions.
As noted above, it is clear that absorption of solar UV in
the lower atmosphere will be substantially greater for the
perfluoroaldehydes than their non-halogenated analogues.
However, in the absence of quantum yield data it is unclear
whether the rate of photolysis of long chain perfluoroalde-
hydes is greater than, comparable to, or less than that of the
corresponding normal aldehydes. Studies of the photolysis
quantum yields for CxF2x+1CHO (x > 1) under atmospheric
conditions are needed to quantify the degree to which
reaction (13) competes with reaction (10) and hence
suppresses, and perhaps eliminates, formation of perfluor-
ocarboxylic acids via the CxF2x+1C(O)O + HO reaction.
4
. Discussion
Our theoretical and experimental investigations show
impressive agreement and lead us to several important
conclusions. We have established that the perfluoroalkyl
aldehydes CxF2x+1CHO (x = 1–4), absorb strongly in the
critical UV region above 290 nm. The absorption is
concentrated in a broad single band and, for the higher
aldehydes (x = 3,4), the peak absorption is at 309 nm. We
anticipate that the higher homologues CxF2x+1CHO (x > 4),
will absorb at a very similar wavelength.
2
2
Acknowledgements
Ole John Nielsen thanks the Danish Natural Science
Research Council for financial support. In addition, we
thank Robert C. Buck, Paul J. Krusic and Mary A. Kaiser for
their critical review of the manuscript.
The maximum in the absorption cross-section increases
monotonically and rapidly with increasing length of the
perfluorinated tail but we cannot say if this trend will hold
for n > 4. The peak absorption of C F CHO is about 3.5
4
9
References
times greater than that of CF CHO.
3
The higher perfluoroalkyl aldehydes absorb more
strongly than their corresponding non-halogenated counter-
parts [50]. Propionaldehyde has a peak absorption of 5.88 ꢀ
[
[
1] J.P. Giesy, K. Kannan, Environ. Sci. Tech. 35 (2001) 39.
2] C.A. Moody, J.W. Martin, W.C. Kwan, D.C.G. Muir, S.A. Mabury,
Environ. Sci. Tech. 36 (2002) 545.
ꢁ
20
ꢁ20
for C F CHO while
1
0
compared to 6.25 ꢀ 10
2
5
[3] C.A. Moody, W.C. Kwan, J.W. Martin, D.C.G. Muir, S.A. Mabury,
Anal. Chem. 73 (2001) 2200.
the absorption cross-section for n-butyraldehyde peaks at
ꢁ20
compared to 8.96 ꢀ 10ꢁ20 cm mol for
2
ꢁ1
[
[
[
[
4] J.W. Martin, M.M. Smithwick, B.M. Braune, P.F. Hoekstra, D.C.G.
Muir, S.A. Mabury, Environ. Sci. Tech. 38 (2004) 373.
6
C F CHO.
.17 ꢀ 10
3
7
5] D.A. Ellis, S.A. Mabury, J.W. Martin, D.C.G. Muir, Nature 412 (2001)
In addition, the non-halogenated aldehydes have their
peak absorption much closer to the critical 290 nm
threshold. The acetaldehyde cross-section peaks at
6
844.
6] P. Taylor, T. Yamada, R. Giraud, M. Kaiser, R. Buck, Environ. Sci.
Tech., in press.
7] E. Kissa, Fluorinated surfactants and repellents, in: Surfactant Science
Series, Marcel Dekker, New York, 2001.
2
90 nm, propionaldehyde at 293 nm and n-butyraldehyde
at 294 nm. This is important because the UV flux at the
earth’s surface increases rapidly as a function of wavelength
for l > 290 nm. For a solar zenith angle of 08, the UV flux at
[
[
8] P.J. Krusic, D.C. Roe, Anal. Chem. 76 (2004) 3800.
9] D.A. Ellis, J.W. Martin, A.O. De Silva, S.A. Mabury, M.D. Hurley,
M.P. Sulbaek Andersen, T.J. Wallington, Environ. Sci. Tech. 38 (2004)
3316.
3
09 nm is about 1400 times that at 294 nm [51]. Thus, we
[
[
10] M.P. Sulbaek Andersen, M.D. Hurley, T.J. Wallington, J.C. Ball, J.W.
Martin, D.A. Ellis, S.A. Mabury, Chem. Phys. Lett. 381 (2003) 14.
11] J.G. Calvert, R. Atkinson, J.A. Kerr, S. Madronich, G.K. Moortgat,
T.J. Wallington, G. Yarwood, Mechanisms of the Atmospheric Oxida-
tion of the Alkenes, Oxford, University Press, 2000, ISBN: 0-19-
513177-0.
may safely conclude that absorption of solar UV in the lower
atmosphere will be substantially greater for the perfluor-
oalkyl aldehydes as compared to their non-halogenated
counterparts.
It is well established that the normal aldehydes,
C H2x+1CHO, undergo UV-induced C–C bond scission with
quantum yields at 300 nm in the range 0.4–1.0 for x = 1–3
[
[
[
[
12] R.P. Borkowski, P. Ausloss, J. Am. Chem. Soc. 84 (1962).
13] G. Lucazeau, C. Sandorfy, J. Mol. Spectrosc. 35 (1970) 14.
14] J.S. Francisco, I.H. Williams, Mol. Phys. 6 (1992) 1433.
15] R. Meller, D. Boglu, G. K. Moortgat, STEP-HALOCSIDE/AFEAS
Workshop, 23–25 March, Dublin, Ireland, 1993, p. 130–138.
x
o
11]. For an altitude of 0.5 km and solar zenith angle of 0 it
[
has been calculated that HCHO, CH CHO, and C H CHO
3
2 5
ꢁ4
ꢁ6
photolyze at rates of 1 ꢀ 10 , 7 ꢀ 10 , and 2.5 ꢀ
[16] S.R. Sellev a˚ g, T. Kelly, H. Sidebottom, C.J. Nielsen, Phys. Chem.
ꢁ
5
ꢁ1
Chem. Phys. 6 (2004) 1243.
1
0
s , respectively [11]. Photolysis is an important
loss atmospheric mechanism for normal aldehydes,
[
[
17] M.P. Sulbaek Andersen, M.D. Hurley, T.J. Wallington, J.C. Ball, J.W.
Martin, D.A. Ellis, S.A. Mabury, O.J. Nielsen, Chem. Phys. Lett. 379
C H2x+1CHO. While, Sellev a˚ g et al. [16] have reported an
x
(
2003) 28.
upper limit of 0.02 for the photolysis quantum yield of
CF CHO, no data are available for longer chain perfluor-
18] S. Pinnock, K.P. Shine, T.J. Smyth, M.D. Hurley, T.J. Wallington, J.
Geophys. Res. 100 (1995) 23227.
3

1932
Y. Hashikawa et al. / Journal of Fluorine Chemistry 125 (2004) 1925–1932
[
[
[
[
[
[
[
[
19] T. Ando, Z. Phys. B 26 (1977) 263.
Stratmann, O. Yazyev, A.J. Austin, R. Cammi, C. Pomelli, J.W.
Ochterski, P.Y. Ayala, K. Morokuma, G.A. Voth, P. Salvador, J.J.
Dannenberg, V.G. Zakrzewski, S. Dapprich, A.D. Daniels, M.C.
Strain, O. Farkas, D.K. Malick, A.D. Rabuck, K. Raghavachari,
J.B. Foresman, J.V. Ortiz, Q. Cui, A.G. Baboul, S. Clifford, J.
Cioslowski, B.B. Stefanov, G. Liu, A. Liashenko, P. Piskorz, I.
Komaromi, R.L. Martin, D.J. Fox, T. Keith, M.A. Al-Laham, C.Y.
Peng, A. Nanayakkara, M. Challacombe, P.M.W. Gill, B. Johnson, W.
Chen, M.W. Wong, C. Gonzalez, J.A. Pople, Gaussian 03, Revision
B.05, Gaussian, Inc., Pittsburgh, PA, 2003.
20] A. Zangwill, P. Soven, Phys. Rev. A 21 (1980) 156.
21] K. Nuroh, M.J. Stott, E. Zaremba, Phys. Rev. Lett. 49 (1982) 862.
22] G.D. Mahan, J. Chem. Phys. 76 (1982) 493.
23] S. Baroni, P. Giannozzi, A. Testa, Phys. Rev. Lett. 58 (1987) 1861.
24] Z.H. Levine, D.C. Allan, Phys. Rev. Lett. 63 (1989) 1719.
25] R. Bauernschmitt, R. Ahlrichs, Chem. Phys. Lett. 256 (1996) 454.
26] M.E. Casida, C. Jamorski, K.C. Casida, D.R. Salahub, J. Chem. Phys.
1
05 (1998) 4953.
27] M.E. Casida, C. Jamorski, K.C. Casida, D.R. Salahub, J. Chem. Phys.
08 (1998) 4439.
28] C. Jamorski, M.E. Casida, D.R. Salahub, J. Chem. Phys. 104 (1999)
134.
[
[
1
[38] A.D. Becke, J. Chem. Phys. 98 (1993) 5648.
[39] C. Lee, W. Yang, R.G. Parr, Phys. Rev. B 37 (1988) 785.
[40] N. Godbout, D.R. Salahub, J. Andzelm, E. Can. Wimmer, J. Chem. 70
(1992) 560.
5
[
[
[
29] M.E. Casida, D.R. Salahub, J. Chem. Phys. 113 (2000) 8918.
30] C. Adamo, G.E. Scuseria, V. Barone, J. Chem. Phys. 111 (1999) 2889.
31] K.B. Wiberg, R.E. Stratmann, M.J. Frisch, Chem. Phys. Lett. 297
[41] M.E. Casida, in: D.P. Chong (Ed.), Recent Advances Density Func-
tional Theory Methods Part I, World Scientific, Singapore, 1995.
[42] T.H. Dunning, P.J. Hay, in: H.F. Schaefer, III (Ed.), Methods of
Electronic Structure Theory, 3, Plenum Press, New York, 1977.
[43] W.J.L. Hehre, L. Radom, P.V.R. Schleyer, J.A. Pople, Ab Initio
Molecular Orbital Theory, John Wiley& Sons, New York, 1986.
[44] A.P. Scott, L. Radom, J. Phys. Chem. 100 (1996) 16502.
[45] A. N o¨ lle, F. P a¨ R. Meller, G.K. Moortgat, E.R. Roeth, R. Ruhnke, H.
Keller-Rudek, UV–vis Spectra of Atmospheric Constituents, DLR-
DFD (Deutsches Fernerkundungsdatenzentrum), 1998, ISBN 3-
89100-030-8.
(1998) 60.
[
[
32] Z.-L. Cai, J.R. Reimers, J. Chem. Phys. 112 (2000) 527.
33] S.J.A. van Gisbergen, J.A. Groeneveld, A. Rosa, J.G. Snijders, E.J.
Baerends, J. Phys. Chem. A 103 (1999) 6835.
[
[
34] J.B. Foresman, M. Head-Gordon, J.A. Pople, M.J. Frisch, J. Phys.
Chem. 96 (1992) 135.
35] N.N. Matsuzawa, A. Ishitani, D.A. Dixon, T. Uda, J. Phys. Chem. A.
105 (2001) 4953;
N.N. Matsuzawa, S. Mori, E. Yano, S. Okazaki, A. Ishitani, DA.
Dixon, Proc. SPIE 3999 (2000) 375;
[46] G.O. Pritchard, G.H. Miller, J.K. Foote, Can. J. Chem. 40 (1962) 1830.
[47] W.B. DeMore, S.P. Sander, C.J. Howard, A.R. Ravishankara, D.M.
Golden, C.E. Kolb, R.F. Hampson, M.J. Kurylo, M.J. Molina, in:
Chemical Kinetics and Photochemical Data for Use in Stratospheric
Modeling, NASA Panel for Date Evaluation, Evaluation no. 10, Jet
Propulsion Laboratory Publication 92-20, August 15, 1992, p. 155.
[48] D.C. Moule, A.D. Walsh, Chem. Rev. 75 (1) (1975) 67.
[49] G.H. Dieke, G.B. Kistiakowski, Phys. Rev. 45 (1934) 4.
[50] R.D. Martinez, A.A. Buitrago, N.W. Howell, C.H. Hearn, J.A. Joens,
Atmos. Environ. A. 26 (1992) 785.
N.N. Matsuzawa, A. Ishitani, D.A. Dixon, T. Uda, Proc. SPIE 4345
(2001) 396.
[
[
36] R.L. Waterland, K.D. Dobbs, A.M. Rinehart, A.E. Feiring, R.C.
Wheland, B.E. Smart, J. Fluor. Chem. 122 (2003) 37.
37] M.J. Frisch, G.W. Trucks, H.B. Schlegel, G.E. Scuseria, M.A. Robb,
J.R. Cheeseman, J.A. Montgomery, Jr., T. Vreven, K.N. Kudin, J.C.
Burant, J.M. Millam, S.S. Iyengar, J. Tomasi, V. Barone, B. Mennucci,
M. Cossi, G. Scalmani, N. Rega, G.A. Petersson, H. Nakatsuji, M.
Hada, M. Ehara, K. Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T.
Nakajima, Y. Honda, O. Kitao, H. Nakai, M. Klene, X. Li, J.E. Knox,
H.P. Hratchian, J.B. Cross, C. Adamo, J. Jaramillo, R. Gomperts, R.E.
[51] B.J. Finlayson-Pitts, J.N. Pitts, Jr., in: Atmospheric Chemistry, John
Wiley & Sons, Inc., New York, 1986, p. 110 (See Table 3.5).