The radiation chemistry of acyclic hydrofluoro and perhalogenated ether and hydrocarbon compounds

Article abstract of DOI:10.1016/S0022-1139(03)00008-3

The radiolytic stability of some hydrofluoroethers and hydrofluorocarbons was investigated and compared with those of perfluoropolyethers (PFPEs) and the CCl₂FCClF₂ (CFC 113). The experimental results indicate that stability depends mainly on the relative abundance of hydrogen atoms in the molecule; however, a significant role is played also by the chemical structure (i.e. the relative positions of the hydrogen atoms in the molecule). As a result, molecules containing hydrogen atoms as -OCF₂H chain ends show a higher stability compared with the other hydrofluoro compounds. Based on the analysis of the end products and on the nature of radicals detected by EPR, radiolysis mechanisms are proposed and discussed. Due to their high dipole moments the hydrofluoro compounds and CCl₂FCClF₂ degrade mainly through an ionic mechanism.

Full text of DOI:10.1016/S0022-1139(03)00008-3

Journal of Fluorine Chemistry 121 (2003) 153–162
The radiation chemistry of acyclic hydrofluoro and
perhalogenated ether and hydrocarbon compounds
G. Marchionnia,*, P.A. Guardaa, A. Buttafavab, A. Faucitanob,1
aSolvay Solexis Research and Development Centre, Viale Lombardia 20, 20021 Bollate, MI, Italy
b
`
Dipartimento di Chimica Generale, Universita di Pavia, Viale Taramelli 12, 27100 Pavia, Italy
Received 23 September 2002; received in revised form 18 December 2002; accepted 23 December 2002
Abstract
The radiolytic stability of some hydrofluoroethers and hydrofluorocarbons was investigated and compared with those of perfluoropo-
lyethers (PFPEs) and the CCl2FCClF2 (CFC 113). The experimental results indicate that stability depends mainly on the relative abundance of
hydrogen atoms in the molecule; however, a significant role is played also by the chemical structure (i.e. the relative positions of the hydrogen
atoms in the molecule). As a result, molecules containing hydrogen atoms as –OCF2H chain ends show a higher stability compared with the
other hydrofluoro compounds. Based on the analysis of the end products and on the nature of radicals detected by EPR, radiolysis mechanisms
are proposed and discussed. Due to their high dipole moments the hydrofluoro compounds and CCl2FCClF2 degrade mainly through an ionic
mechanism.
# 2003 Elsevier Science B.V. All rights reserved.
Keywords: Hydrofluoropolyethers; Hydrofluoroethers; Hydrofluorocarbons; Radiation chemistry; EPR spectroscopy; Radicals
1. Introduction
on both ionic and excitation reaction paths, the latter being
of prominent importance. Further objectives of this research
were the characterisation of novel classes of intermediate
fluorinated radicals and their nitroxyl adducts and the
exploitation of thermal stability and decomposition modes
of the elusive fluoroether cation-radicals and anion radicals
near 77 K [10,11]. Other major investigations in this field
pertain to the e-beam chemistry of perfluoropolyethers
based mainly on IR spectroscopy analysis [18–21]; the
electron curing of perfluoropolyethers using low energy
electrons in the 10–90 eV was also investigated with deter-
mination of the negative ions, relative abundances and the
crosslinking reactions [22,23].
Recently new classes of low environmental impact hydro-
fluoroether (HFE) and hydrofluoropolyether (HFPE) fluids
have been developed. In order to evaluate the radiolytic
behaviour of such materials in comparison with perfluorinated
or chlorofluorinated (CFC) fluids, a number of compounds
belonging to the chlorofluorocarbon, hydrofluoropolyether,
perfluoropolyether, hydrofluoroether and hydrofluorocarbon
classes have been subjected to gamma irradiation. In parti-
cular in the present study, HCF2OCF2CF2OCF2H (HFPE),
C4F9OCH3 (HFE), and CF3CF2CFHCFHCF3 (HFC) have
been compared with a polydispersed perfluoropolyether
(PFPE) having general formula CF3O(CF2CF(CF3)O)nCF3
and with CF2ClCFCl2 (CFC 113).
The radiation chemistry of perfluopolyethers (PFPEs) has
been the subject of several investigations [1–11] owing to
their radiation stability which makes them interesting for
applications as lubricant fluids in space and nuclear indus-
tries as well as in several other fields of radiation technology.
A major effort aimed at the elucidation of perfluoropolyether
radiation degradation mechanisms is that based on the
application of a concerted methodology including 19F
NMR and gas chromatograph–mass spectrometry (GC–
MS) for end product and chain structure modifications
analysis, and methods for the identification and character-
isation of the transient intermediates, such as EPR and UV/
Vis spectroscopy, radiothermoluminescence (RTL), pulse
radiolysis and charge scavenger experiments and theoretical
simulations [12–17]. The key aspects of the radiolysis
mechanisms were thus determined and found to be based
* Corresponding author. Tel.: þ39-02-3835-6289;
fax: þ39-02-3835-2152.
E-mail addresses: giuseppe.marchionni@solvay.com (G. Marchionni),
antonio.faucitano@unipv.it (A. Faucitano).
1 Co-corresponding author. Tel.: þ39-03-8250-7344;
fax: þ39-03-8252-8544.
0022-1139/03/$ – see front matter # 2003 Elsevier Science B.V. All rights reserved.
doi:10.1016/S0022-1139(03)00008-3

154
G. Marchionni et al. / Journal of Fluorine Chemistry 121 (2003) 153–162
2. Results and discussion
2.1. Radiolysis products
wt.%) are reported for each compound at three radiation
doses. In the same tables, the mean values and standard
deviation of the radiolytic yields (expressed as m mol JÀ1
)
are presented, both for the radiolysis products and for the
degraded starting material. Satisfactory linear fits of the
weight percent degradation of the starting material as a
In Tables 1–5, the products generated after radiolysis and
the amount of the degraded starting material (expressed as
Table 1
Radiolysis of compound CF2ClCFCl2 (CFC 113) as a function of the absorbed dose and average radiolytic yields
Radiolysis products
Absorbed dose (kJ kgÀ1
)
Radiolytic yield (m mol JÀ1
)
45.1
81.6
129.8
CF3Cl
2.0
15.1
64.2
13.0
85.0
17.0
17.0
22.0
85.0
5.2
27.4
138
24.0
175
7.3
31.4
223
(5.22 Æ 0.93) Â 10À3
(3.52 Æ 0.63) Â 10À2
(1.33 Æ 0.14) Â 10À1
(1.60 Æ 0.18) Â 10À2
(1.49 Æ 0.1) Â 10À1
(1.50 Æ 0.08) Â 10À2
(2.03 Æ 0.16) Â 10À2
(1.82 Æ 0.09) Â 10À2
(5.77 Æ 0.37) Â 10À2
CF2HCl
CF2Cl2
CF2ClCF2Cl
CFCl3
31.0
274
C3F5Cl3
C2F2Cl4
C3F4Cl4
C4F6Cl4
29.0
35.0
37.0
44.0
57.0
58.0
218
139
CFC (degraded)
320
610
944
(3.89 Æ 0.1) Â 10À1
Values in columns 2–4 are conversions of the starting product expressed as weight percent (Â103).
Table 2
Radiolysis of compound HCF2OCF2CF2OCF2H (HFPE) as a function of the absorbed dose and average radiolytic yields
Radiolysis products
Absorbed dose (kJ kgÀ1
)
Radiolytic yield (m mol JÀ1
)
45.8
82.7
93.4
131.6
138
COF2
43.8
11.6
13.0
4.0
(2.37 Æ 0.20) Â 10À1
(4.47 Æ 0.76) Â 10À2
(2.55 Æ 0.24) Â 10À2
(7.90 Æ 1.40) Â 10À3
(1.83 Æ 0.08) Â 10À1
(4.18 Æ 0.04) Â 10À2
–
CHF3
27.6
24.5
9.2
46.5
CF3CF2H
44.7
16.3
CF3–O–CF2H
CF2H–O–CF2H
CF2H–O–CF2CF2H
High MW products
94.2
29.0
114
181
58.2
249
295
101
401
HFPE (degraded)
310
643
1043
(3.20 Æ 0.27) Â 10À1
Values in columns 2–4 are conversions of the starting product expressed as weight percent (Â103).
Table 3
Radiolysis of compound CF3O(CF2CF(CF3)O)nCF3 (PFPE) as a function of the absorbed dose and average radiolytic yields
Radiolysis products
Absorbed dose (kJ kgÀ1
)
Radiolytic yield (m mol JÀ1
)
45.3
81.9
38.4
130.2
42.6
CF4
20.0
24.3
4.0
3.1
4.0
0
(4.68 Æ 0.86) Â 10À2
(1.28 Æ 0.11) Â 10À1
(6.82 Æ 0.73) Â 10À3
(1.09 Æ 0.16) Â 10À2
(5.21 Æ 0.46) Â 10À3
(2.92 Æ 2.5 ) Â 10À3
(7.88 Æ 0.29) Â 10À3
–
COF2
50.9
8.7
70.1
11.5
9.2
C2F6
CHF3
CF3–O–CF3
C3F8
7.3
6.4
9.7
7.0
10.3
21.6
CF3–O–CF2CF3
High MW products
7.0
177
13.2
490
903
PFPE (degraded)
239
622
1078
(2.07 Æ 0.46) Â 10À1
Values in columns 2–4 are conversions of the starting product expressed as weight percent (Â103).

G. Marchionni et al. / Journal of Fluorine Chemistry 121 (2003) 153–162
155
Table 4
Radiolysis of compound C4F9OCH3 (HFE) as a function of the absorbed dose and average radiolytic yields
Radiolysis products
Absorbed dose (kJ kgÀ1
45.8
)
Radiolytic yield (m mol JÀ1
)
82.9
131.8
1.9
CH4
0
0
0
3.01 Â 10À3
CF4
COF2
0
9.5
27.9
4.9
2.73 Â 10À3
8.5
1.0
13.2
7.0
53.7
133
417
10.0
2.0
(3.92 Æ 1.1) Â 10À2
(3.96 Æ 1.2) Â 10À3
(7.65 Æ 1.4) Â 10À2
(8.78 Æ 1.4) Â 10À3
(5.53 Æ 0.6) Â 10À2
(1.31 Æ 0.1) Â 10À1
–
CHF3
CH3F
C3F6
17.0
11.0
80.0
37.8
14.5
C4F10, C3F7H, C4F8
C4F8H–O–CH3
High MW products
156
250
420
815
1360
HFE (degraded)
633
1185
2037
(5.80 Æ 0.34) Â 10À1
Values in columns 2–4 are conversions of the starting product expressed as weight percent (Â103).
Table 5
Radiolysis of compound CF2CF2CFHCFHCF3 (HFC) as a function of the absorbed dose and average radiolytic yields
Radiolysis products
Absorbed dose (kJ kgÀ1
45.5
)
Radiolytic yield (m mol JÀ1
)
82.2
10.6
130.8
15.0
CF4
3.0
2.1
(1.17 Æ 0.38) Â 10À2
(6.55 Æ 0.81) Â 10À3
(6.30 Æ 1.2) Â 10À3
(8.70 Æ 1.2) Â 10À2
–
CHF3
4.2
8.5
5.2
C2F6
C5F9Ha
High MW products
3.2
11.5
94.0
444
187
840
224
1320
HFC (degraded)
546
1050
1578
(4.87 Æ 0.17) Â 10À1
Values in columns 2–4 are conversions of the starting product expressed as weight percent (Â103).
a The quantification also includes C3F6H2, C4F6H2, C4F7H and C5F10 present in small amount.
function of the radiation dose are obtained for all the five
compounds (see Fig. 1); slopes (a) and correlation coeffi-
cients (R2) of the best fits:
The radiolytic yields are found to be independent of the
radiation dose (see Fig. 2 and Tables 1–5); this implies that
secondary reactions from radiolysis of primary products are
negligible, as expected from the low conversion (1–2%) of
the starting material. Both the weight percent degradation
and the overall radiolytic yield lead to the conclusion that
weight percent degradation ¼ a  absorbed dose ðkJ kgÀ1
are reported in Table 6.
Þ
Fig. 1. Products weight percent degradation as a function of the absorbed radiation dose: (&) CFC 113, (&) HFPE, (~) PFPE, (*) HFE, (^) HFC.

156
G. Marchionni et al. / Journal of Fluorine Chemistry 121 (2003) 153–162
Table 6
Slopes (a) and correlation coefficients (R2) for the linear fit of the weight
products (Table 4). The radiolysis of the compound HFC
(Table 5) is characterised by a prevalence of CF4 and C5F9H,
the latter originating from dehydrofluorination of the start-
ing compound, together with condensation products.
percent degradation as a function of the absorbed dose
Compound
a
R2
CFC
0.0073
0.0078
0.0079
0.0150
0.0122
0.998
0.991
0.952
0.990
0.995
2.2. EPR detection of intermediate radicals
HFPE
PFPE
HFE
In the case of CCl2FCClF2 both the spin-trapping and
direct EPR detection have afforded evidence for the forma-
tion of –CF2ꢀ and –CFꢀ– radicals which are expected to arise
from C–Cl and C–C scissions. The species ClCF2CFꢀCl,
formed by C–Cl breaking at the CFCl2 site, is the most
abundant as expected from the favourable statistical factor.
The spin-trapping with 2-methyl-2-nitrosopropane
(MNP) occurs according to the reaction:
HFC
CFC, HFPE and PFPE have higher radiolytic stabilities
compared with HFE and HFC compounds. Moreover, the
overall radiolytic yield, intrinsically independent of the
molecular weight of the starting material, allows us to
differentiate the behaviour of the more stable compounds,
giving the following order of stability:
ðCH3Þ3CÀN ¼ O þ Rf ! ðCH3Þ3CÀNðRfÞÀOꢀ
(1)
ꢀ
where the Rfꢀ radicals are radiolysis products which are
trapped at the melting of the compounds and thus detected
from the corresponding EPR spectra of the liquid phase
nitroxyl adducts. ꢀThe hyperfine (hf) coupling constants,
a(X), of the –CF2 nitroxyl adduct were:
PFPE > HFPE > CFC
Tables 1–5 show that the formation of products having
molecular weight higher than that of the starting material is a
common outcome of all the compounds examined; more-
over, such products are the main radiolysis products (in most
cases).
Cl2CFCF2NðOꢀÞR : aðNÞ ¼ 1:16 mT;
2aðFÞb ¼ 2:01 mT; aðFÞg ¼ 0:25 mT
Because of the lack of resolution, the Cl and F splittings in
the spectrum of the ClCF2CFClN(Oꢀ)R adduct could not be
fully determined. However, a partial simulation, adequate
for the radical identification, was achieved with the set of
coupling constants reported in the literature [24]:
In the case of CCl2FCClF2 (Table 1), a mixture of
chlorofluorinated compounds of variable chlorine content
and molecular size (C1–C4) was obtained, indicating that
both C–Cl and C–C bond ruptures take place. In the case of
HFE, PFPE and HFPE one of the main degradation products
is COF2, generated with radiolytic yields (G), respectively,
of 0.04, 0.1 and 0.2 m mol JÀ1. With HFPE (Table 2), the
major product (besides the high molecular weight products
and COF2) is HCF2OCF2H (G ¼ 0:183 m mol JÀ1). The
radiolysis of PFPE gives COF2, CF4, CHF3 and condensa-
tion products with the highest radiolytic yields (Table 3).
The main products in the radiolytic degradation of HFE are
COF2, CH3F and C4F8HOCH3, together with condensation
aðNÞ ¼ 1:15 mT; 2aðFÞCF ¼ 0:11 mT;
2
aðFÞCF ¼ 0:40 mT; aðClÞ ¼ 0:4 mT
The radicals stemming from the irradiation of HFE as
detected from the corresponding spin adducts were:
Fig. 2. Overall radiolytic yields of the investigated compounds.

G. Marchionni et al. / Journal of Fluorine Chemistry 121 (2003) 153–162
157
ꢀ
CF3C(CF3)CF2OCH2 (ꢀdominant species in the 77 K spec-
The fraction of radical cations and electrons which escapes
from the geminate recombination gives rise to ionic degra-
dation paths (electron capture with formation of radical
anions, degradation of radical anions and cations, etc.).
The nature of the compound can influence the reaction
mechanism, since the probability of the geminate recombi-
nation in condensed media depends on the chemical–phy-
sical characteristics of the medium, such as polarity, electron
affinity, etc. If the geminate recombination is highly
favoured, the mechanism is mainly through the excitation
path; otherwise ionic pathways involving the chemistry of
radical cations and anions can be prevalent. Often, both the
excitation and the ionic paths coexist. The relative role
played by the two basic components in a radiolysis mechan-
ism can be quantitatively evaluated. The free ions yield in a
compound undergoing radiation induced decomposition is
expressed by the equation [25]:
trum), CF3CF(CF3)CF2 , (CF3)2CFꢀ and CF3Cꢀ(CF3)CF2-
OCH3. The first three species were also detected from their
anisotropic a-F and a-H components in the 77 K spectra
which are characterised by the parallel principal values
A(X) :
k
AðHÞkCH ¼ 2:7 mT;
AðFÞkCF ¼ 23:6 mT and
2
2
AðFÞkCF ¼ 22:2 mT
The hf couplings of the nitroxyl adducts used in the com-
puter simulations are shown below:
ðCF3Þ CFCF2NðOꢀÞR : aðNÞ ¼ 1:17 mT;
2aðFÞb2 ¼ 0:241 mT
ðCF3Þ CFCF2OCH2NðOꢀÞR : aðNÞ ¼ 1:13 mT;
2aðHÞ2¼ 1:86 mT; 2aðFÞd ¼ 0:065 mT
Z
ðCF3ÞCFÀNðOꢀÞR : aðNÞ ¼ 1:22 mT;
6aðFÞg ¼ 0:11 mT; aðFÞb ¼ 0:22 mT
GðfiÞ ¼ GðtÞ fðrÞ Á PðrÞ dr
(2)
where G(fi) is the free ion yield, G(t) the total ion yield, f(r)
the distribution function of the electron-ion distances and
P(r) the probability of electron escape, expressed by the
equation:
ðCF3Þ CðCF2OCH3ÞNðOꢀÞR : aðNÞ ¼ 1:1 mT;
6aðFÞ2¼ 0:11 mT; 2aðFÞ ¼ 0:46 mT
The irradiation of HFPE affords species of type –CF2 as
ꢀ
ꢀ
ꢁ
the major products revealed by AðFÞkCF ¼ 23:5 mT. Also
2
r
PðrÞ ¼ exp À
(3)
present are minor amounts of radicals arising from C–F
rupture showing major features at ca. Æ10 mT.
rc
The nitroxyl adducts detected are consistent with two
ꢀ
In the above equation, rc is the Onsager radius, which is
˚
equal to 250 A for hydrocarbons with dielectric constant
types of –CF2 , namely:
˚
e ¼ 2 and 7 A for H2O where e ¼ 80. The distance distribu-
ÀOCF2NðOꢀÞR : aðNÞ ¼ 1:24 mT; 2aðFÞ ¼ 2:38 mT
tion function is not known in advance and frequently a
Gaussian distribution is assumed [25]. With molecular
dipole moments of 1.0–2.0 D (the H2O molecule has
m ¼ 1:8 D) the ionic component is of major importance.
In contrast, dipole moments near zero, as in hydrocarbons,
lead to dominance of the excitation pathway. The dipole
moment of the compounds investigated in this work as
obtained from the Debye equation [26] are:
ÀCF2CF2NðOꢀÞR : aðNÞ ¼ 1:25 mT;
2aðFÞ ¼ 1:97 mT; 2aðFÞg ¼ 0:04 mT
The EPR spectrum obtained from the HFC compound
after gamma irradiation at 77 K has the features expected
from –CFꢀ– radicals arising from C–H bond scissions. The
spin trapping with MNP supports this interpretation and also
ꢀ
reveals the presence of minor amounts of –CF2 species
resulting from C–C bond breaking.
a
ClCF2CFCl2
m ¼ 1:56 D
m ¼ 1:98 D
m ¼ 1:77 D
m ¼ 1:90 D
CF3CFðCF3ÞCF2OCH3
HCF2OCF2CF2OCF2H
CF3CF2CFHCFHCF3
2.3. Radiolysis mechanism
a Calculated using the MNDO-PM3 MO method.
A characteristic common to all the materials here exam-
ined is the formation after irradiation of considerable
amounts of products having molecular weights higher than
that of the starting compound. Such products can be
explained partly as a result of the coupling reactions of
intermediate radicals, the presence of which has been con-
firmed by the EPR experiments.
It is well known that g rays produce ionisation of materi-
als, giving radical cations and electrons. These primary
intermediates can quickly recombine (geminate recombina-
tion) producing excited molecules which can loose energy
thermally or by giving mono and bimolecular reactions.
The polarity of such compounds points to the ionic
radiolysis mechanisms being dominant.
The excitation pathways are expected to play a major
role in perfluoro ethers and perfluoropolyethers such as
those with repeating units –CF2CF(CF3)O– (KrytoxTM),
–CF2CF2CF2O– (DemnumTM), –CF2CF(CF3)O– and
–CF2O– (Fomblin YTM), –CF2CF2O– and –CF2O– (Fomblin
ZTM), since dipole moments similar to those of the hydro-
carbons are found in these cases. This conclusion was
confirmed by electron scavenging experiments with CF3Br

158
G. Marchionni et al. / Journal of Fluorine Chemistry 121 (2003) 153–162
from a –CF2– group which has a high activation energy. Also
a two step mechanism involving the dissociative electron
capture of a C–F bond to give FÀ and a –CFꢀ– radical is
considered unlikely because of its endothermicity.
ꢀ
The formation of the –OCF2 species can proceed via
proton transfer from the primary cation radicals to the parent
molecule or to an anion:
ÀOCF2Heꢀþ þ HCF2OCF2CF2OCF2H
! ÀOCF2ꢀ þ HCF2OCF2CF2OðHþÞCF2H‘
(5)
According to the results of MNDO-PM3 calculations (see
Fig. 3), additional reaction paths for the molecular cation
radicals are the decomposition by C–C and C–O bonds
breaking, giving neutral free radicals and carbonium ions.
This mechanism can account for some of the major radicals
and end radiolysis products, such as COF2 and HCF2OCF2H
Scheme 1. Reaction mechanism for the radiolysis of compound
ClCF2CFCl2 (CFC 113). (a) Dissociative electron capture pathway; (b)
excitation pathway.
which were consistent with a 70–80% of the excitation
contribution [10,11].
Due to the high electron affinity combined with polarity, a
dominant role by the dissociative electron capture of the C–
Cl bond in the CCl2FCClF2 is expected (see Scheme 1). The
coupling reactions of the radical intermediates thus formed
give rise to a distribution of products with molecular weight
higher than that of the starting material. The identified stable
end products also afford evidence of C–C scission events,
which may result from the cation-electron neutralisation
(see Scheme 1); the C1 radicals thus produced can abstract
chlorine from a CFC molecule to yield chlorofluorinated C1
species. On the other hand, the self-coupling reactions of C1
radicals and their cross-coupling reactions with C2 radicals
can account, respectively, for the observed C2 and C3 end
products with variable chlorine content (see Table 1).
As far as the radiolysis of HFPE is concerned, a major
difference with respect to perfluoropolyethers is represented
by the enhancement of the ionic contribution and by the
possibility of radical H abstraction reactions from the –
OCF2H end groups as indicated by the presence of the
radiolysis product HCF2OCF2CF2H (see Table 2). The
formation of radiolysis products with a molecular weight
higher than the parent compounds (the ‘‘higher molecular
weight products’’) is rationalised via radical coupling reac-
tions. Ionic reactions can be envisaged to explain the for-
mation of the majority of the radiolysis products. The
detection of secondary >CFꢀ radicals, coupled with the
formation of compounds characterised by perfluoromethyl
chain ends, can be rationalised by a concerted cation–anion
neutralisation reaction involving FÀ ion transfer:
Fig. 3. MNDO-PM3 fully optimised geometries of the HFPE cation and
anion radicals. The arrow shows the lengthening of the bond undergoing
˚
scission (bond length in A).
À
ÀCF2þ þ HCF2OCF2CF2OCF2Heꢀ
! ÀCF3 þ HCF2OCFꢀCF2OCF2H
(4)
This mechanism is expected to be favoured enerꢀgetically
Scheme 2. Reaction mechanism (excitation path) for the radiolysis of
compound HFPE (in bold identified end-products and radicals).
with respect to the F atom abstraction by a –CF2 radical

G. Marchionni et al. / Journal of Fluorine Chemistry 121 (2003) 153–162
159
Scheme 3. Reaction mechanism for the radiolysis of compound HFE (in bold identified end-products and radicals).
(see Scheme 2 and Table 2). As far as the anion radical is
concerned, the MO calculations predict C–C bond splitting
to be favoured (see Fig. 3 and Scheme 2).
Experimental evidence for such mechanisms was
obtained from solid state EPR electron capture experiments
performed at 77 K on HFPE using low energy photoelec-
trons generated from tetramethyl-paraphenylene-diamine
(TMPD) [10,11].
The neutralization reactions between charged species,
diamagnetic cations and anions, stemming from the primary
cation and anion radicals can justify the formation of other
major radiolytic products, such as HCF2CF2OCF2H (see
Scheme 2 and Table 2).
result from H abstraction reactions at the –OCH3 group
or deprotonation of the precursors cation radicals.
The proton transfer mechanism, because of its lower
activation energy, is more suited to account for the observed
ꢀ
C4F9OCH2 radical formation during the irradiation at 77 K.
The formation of the major radiolysis products C4HF8OCH3
and CH3F is justified via the neutralization reaction of methyl
cations with the parent anion radicals leading to CH3F and to
the tertiary radical (CF3)2CꢀCF2OCH3 (detected by EPR)
which is the precursor of the C4HF8OCH3 compound through
an H abstraction reaction. The neutralization reaction is
predicted to be exothermic by the MNDO-PM3 method with
a DHꢁ ¼ À193:9 kcal. However, other reactions leading to
C4HF8OCH3 must be considered since the yield of this
product is about twice that of CH3F. The MO method
also predicts (see Fig. 4) for the parent molecular anion
and cation of HFE, monomolecular decomposition reactions
with breaking of the C–C and C–O bonds, respectively.
These mechanisms can easily be rationalised with some of
the radicals detected and stable end products (see Table 4),
namely, CF3H (proton abstraction by CF3À from the proto-
nated ether) and (CF3)2CFCÀF3 (neutralization with FÀ
The higher radiolytic yield from HFE is rationalised by
the presence of the alkoxymethyl end group and the related
probability of O–CH3 and C–H bond rupture which are
favoured events in the radiolysis of ethers [27]. A ionic
radiolysis mechanism due to the high polarity of this com-
pound has been considered in the reaction Scheme 3. Pro-
ducts such as COF2, C3F7H, CF3H are consistent with
radical fragmentation reactions followed by hydrogen trans-
fer at the –OCH3 group. On the other hand, radical couplings
and ion neutralisations can account for the formation of
products with a molecular weight greater than the parent
eꢀ
transfer between C4F9OCH3
and (CF3)2CFCF2eþ).
HFC shows a reduced radiolytic stability as compared to
HFPE (Tables 2 and 5); this enhanced reactivity can perhaps
ꢀ
molecule. The major radical intermediate RfOCH2 can
Fig. 4. Semiempirical MO MNDO-PM3 fully optimised geometries of the HFE cation and anion radicals showing a lengthening of the bonds undergoing
˚
scission (bond lengths in A).

160
G. Marchionni et al. / Journal of Fluorine Chemistry 121 (2003) 153–162
Scheme 4. Reaction mechanism for the radiolysis of compound HFC (in bold identified end-products and radicals).
be rationalised by the presence of activated secondary C–H
bonds which may favour dehydrofluorination and C–H
bond ruptures reactions (see Scheme 4). Accordingly EPR
spectroscopy confirms the presence of (Rf)2CFꢀ species
and the GC analysis shows significant amounts of the
dehydrofluorination compound C5F9H (see Table 5).
C–C scissions events within the radiolysis mechanism of
HFC are witnessed by the formation of CHF3 and C2F6
among the major products. Such low molecular weight
radiation sensitivity of the compound. However, the position
of the hydrogen atoms and the chemical structure of the
molecule can also play an important role, as evidenced by
the higher stability of HFPE in comparison with HFC.
Judging from the dielectric constants and calculated
dipole moments, which are enhanced by the H substitutions,
the radiolysis mechanisms of HFE, HFPE, HFC are
expected to be mainly ionic; accordingly the reaction
Schemes 2–4 are predominately based on ionic pathways,
which are consistent with products and intermediate radi-
cals analysis.
A different mechanism applies to the PFPE for which
previous experimental evidence [10,11] and the magnitude
of the dielectric constant, point to a prevalence of the
excitation component. Such mechanistic difference may
play a key role, together with the lack of reactive C–H
bonds, in determining the greater radiation stability of the
perfluoropolyethers.
ꢀ
compounds can be related to CF3 radical intermediates
undergoing H abstraction and coupling reactions. The self
coupling between parent molecular radicals like (Rf)2CFꢀ
and cross coupling of these species with other radicals
are believed to be responsible for the formation the high
molecular weight compounds which are the major radiolysis
products (see Scheme 4).
3. Conclusions
The radiation chemistry of hydrofluoro- and chlorofluoro-
compounds has been investigated on a comparative base
with full analysis of the stable end products and intermediate
radicals. By considering the experimental overall decom-
position yields the following radiation stability scale, which
includes also the perfluoropolyethers previously investi-
gated, was established:
4. Experimental
4.1. Materials
HCF2OCF2CF2OCF2H (HFPE) was obtained by frac-
tional distillation of an oligomeric mixture of a,o-hydro-
perfluoropolyethers, with a 99% purity (GC).
C4F9OCH3 (HFE) and CF3CF2CFHCFHCF3 (HFC) are
commercial products from 3 M and DuPont, respectively,
and were used as received.
C4F9OCH3 is a mixture of linear and branched C4F9-
chain isomers, with a ratio n–C4F9–:(CF3)2CFCF2– of about
70:30.
PFPE > HFPE > CFC 113 > HFC > HFE
The key structural factor behind this stability trend seems
to be the presence of C–H bonds which enter reaction
pathways (H abstraction, proton transfer, C–H bond scis-
sions) which are forbidden to C–F bonds mainly because
of the greater bond breaking energy. Within this framework,
the greater the number of C–H bonds molÀ1 the higher the
Cl2CFCF2Cl (CFC 113) was purchased from Aldrich with
a 99.9% purity and used as received.

G. Marchionni et al. / Journal of Fluorine Chemistry 121 (2003) 153–162
161
CF3O(CF2CF(CF3)O)nCF3 (PFPE) is a PFPE polydis-
persed fraction, obtained by distillation, having a bp of
of the atoms constituting the molecule and ðZ=AÞFrieke
0:5534.
¼
55 8C and an average molecular weight of 350.
As far as the unknown molecular mass products are
concerned, it was not possible to calculate the radiolytic
yield but only the weight percent value of their formation.
The overall radiolytic yield was calculated from the weight
percent degradation of the starting material.
4.2. Sample irradiation and analysis of
the end products
The gamma irradiations were performed using a 60Co
source at a dose rate of 0.13 Mrad hÀ1. Aweighed amount of
each material (about 4 g) was put in a quartz tube
(volume ¼ 5 ml), evacuated from air after freezing in liquid
nitrogen and sealed with a break-seal. Three tubes were
prepared for each compound and irradiated at room tem-
perature with doses of 5.2, 9.4 and 15.0 Mrad, respectively.
After irradiation, the tubes were connected to a vacuum line
having a small and known volume, then the break seal was
broken and nitrogen was added until the system reached
room pressure. After equilibration, both the liquid and the
vapour phase were analysed by GC–MS using a HP 5890 gas
chromatograph equipped with Porabond Q column, 25 m,
0.32 mm i.d. and coupled with an HP 5988 mass spectro-
meter.
4.3. EPR measurements
The EPR measurements were carried out on a Brucker
EMX/300 spectrometer equipped with a variable tempera-
ture control. Two different techniques were employed in
order to detect the intermediate radicals:
(a) vacuum irradiation of the neat compound at 77 K
followed by thermal annealing up to the melting of the
matrix with contemporaneous recording of the spectra;
(b) spin-trapping with 2-methyl-2-nitrosopropane (MNP).
The samples containing 0.1% of MNP were irradiated
under vacuum at 77 and 195 K and subsequently allowed to
warm up to room temperature with continuous recording of
the spectra. The low temperature irradiation allows the
radiolysis process to be quenched at the initial stages so
that the primary species can be detected.
The quantitative analysis was performed by GC, using an
HP 5890 instrument with a Poraplot Q capillary column and
a TC detector.
In most cases it was not possible to determine the
molecular mass of a part of the products having elution
times longer than that of the starting material in the GC–
MS analysis: the elution order however indicates that
these products have a higher molecular mass than the
starting material, and have been labelled as ‘‘high mole-
cular weight products’’. GC in any case allowed the quanti-
tative evaluation of these products as weight percent
from the peak areas (this is a reasonable assumption when
a TC detector is used).
In the presence of the spin-trapping agent, the radicals are
trapped at the melting of the compounds (see reaction 1) and
thus detected from the corresponding EPR spectra of the
liquid phase nitroxyl adducts.
Radical identifications were based on computer simula-
tion of the experimental spectra.
Acknowledgements
As the volume of the vapour phase and the weight of the
liquid phase were known, it was possible to calculate the
amount of each product formed after radiolysis and conse-
quently the amount of the degraded starting material. The
amount of products in the vapour phase was generally low
compared with the amount in the liquid. The vapour phase
was rich in low boiling products, such as COF2, CF4, etc. In
some cases SiF4 was found in the vapour phase, due to the
reaction of COF2 with the glass surface:
The GC–MS analyses by Dr. F. Morandi and Dr. P.
Dardani are gratefully acknowledged.
References
[1] A. Faucitano, A. Buttafava, F. Martinotti, G. Caporiccio, C. Corti, S.
Maini, G.T. Viola, J. Fluorine Chem. 16 (1980) 649–650.
[2] A. Faucitano, A. Buttafava, F. Martinotti, G. Caporiccio, C. Corti, J.
Chem. Soc. Perkin 2 (1981) 425–431.
[3] A. Faucitano, A. Buttafava, F. Martinotti, G. Marchionni, Tetra-
hedron Lett. 29 (36) (1988) 4611–4614.
2COF2 þ SiO2 ! 2CO2 þ SiF4
(6)
[4] S. Mori, Y. Shitara, T. Seguchi, J. Nishimura, Jpn. J. Tribol. 36 (10)
(1991) 1177–1186.
SiF4 was not considered in the calculations because CO2
is stoichiometrically formed upon the reaction of COF2 and
is evaluated together with COF2 by gas cromatography.
Based on the amount of products and on the radiation
dose absorbed by the material, the radiolytic yields have
been calculated for each product at the three radiation
doses. The determination of the radiation dose absorbed
by the material was based on Frieke dosimetry and on the
[5] A. Faucitano, A. Buttafava, V. Comincioli, G. Marchionni, R.J. De
Pasquale, J. Phys. Org. Chem. 4 (1991) 293–300.
[6] A. Faucitano, A. Buttafava, F. Martinotti, G. Marchionni, R.J. De
Pasquale, Radiat. Phys. Chem. 37 (3) (1991) 493–498.
[7] A. Faucitano, A. Buttafava, P.A. Guarda, G. Marchionni, J. Fluorine
Chem. 64 (1993) 189–199.
[8] A. Faucitano, A. Buttafava, V. Patruno, P.A. Guarda, G. Marchionni,
Radiat. Phys. Chem. 45 (1) (1995) 23–30.
evaluated electron density ratios ðZ=AÞcompound=ðZ=AÞFrieke,
were Z refers to the atomic number and A to the atomic mass
[9] A. Faucitano, A. Buttafava, P.A. Guarda, G. Marchionni, S.
Karolczak, Radiat. Phys. Chem. 48 (1) (1996) 143–145.

162
G. Marchionni et al. / Journal of Fluorine Chemistry 121 (2003) 153–162
[10] A. Faucitano, A. Buttafava, S. Karolczak, P.A. Guarda, G.
Marchionni, Radiat. Phys. Chem. 54 (1999) 179–182.
[11] A. Faucitano, A. Buttafava, S. Karolczak, P.A. Guarda, G.
Marchionni, Radiat. Phys. Chem. 54 (1999) 183–186.
[12] S. Matsunuma, T. Miura, H. Kataoka, Tribol. Trans. 39 (2) (1996)
380–385.
[20] J. Packansky, R.J. Waltman, M. Maier, J. Phys. Chem. 91 (1987)
1225–1236.
[21] J. Packansky, R.J. Waltman, D. Jebens, O. Heery, J. Fluorine Chem.
82 (1997) 85–90.
[22] G.H. Vurens, C.S. Gudeman, L.J. Lib, J.S. Foster, Polym. Preprints
Am. Chem. Soc. 31 (2) (1990) 700.
[13] S. Matsunuma, Y. Hosoe, Tribol. Int. J. 30 (2) (1997) 121–128.
[14] P.H. Kasai, Macromolecules 25 (1992) 6791–6799.
[15] P.H. Kasai, Chem. Mater. 6 (1994) 1581.
[16] K.J.L. Paciorek, S.R. Masuda, W.H. Lin, J.H. Nakahara, J. Fluorine
Chem. 76 (1996) 21–27.
[23] G.H. Vurens, C.S. Gudeman, L.J. Lin, J.S. Foster, Langmuir 8 (1992)
1165–1169.
[24] L. Gille, R. Stroesser, J. Fluorine Chem. 69 (1994) 191–194.
[25] J.M. Warman, The dynamics of electrons and ions in non polar
liquids, in: J.H. Baxendale, F. Busi (Eds.), The Study of Fast
Processes and Transient Species by Electron Pulse Radiolysis,
Reidel, Dordrecth, 1982, 433 pp.
[17] A. Faucitano, A. Buttafava, G. Caporiccio, C.T. Viola, J. Am. Chem.
Soc. 106 (1984) 4172–4174.
[18] J. Packansky, R.J. Waltman, J. Phys. Chem 95 (1991) 1512–
1518.
[26] G. Marchionni, M. Bassi, G. Fontana, P. Maccone, G. Ajroldi, J.
Fluorine Chem. 98 (1999) 41–54.
[19] J. Packansky, R.J. Waltman, C. Wang, J. Fluorine Chem. 32 (1986)
283–297.
[27] R.J. Woods, A.P. Pikaev, Applied Radiation Chemistry: Radiation
Processing, Wiley, New York, 1994, 183 pp.