2π + 2π cycloaddition kinetics of some fluoro olefins and fluoro vinyl ethers

Article abstract of DOI:10.1016/j.jfluchem.2004.06.014

The second order 2π + 2π homo- and co-dimerization between various classes of fluorinated olefins has been investigated. The fluorinated olefins examined in this study were: (1) Rf-OCF=CF2 (perfluorinated vinyl ethers); (2) R_f-CF=CF₂ (perfluorinated terminal olefins); (3) R-CH₂-CF=CF₂; (4) Ph-OCF=CF₂ (aryl perfluorinated vinyl ethers). Homo-dimerizations between vinyl ethers have an E_a between 20 and 24 kcal 1 mol^-1 while homo-dimerizations between terminal olefins have an average E_a between 35 and 40 kcal mol^-1; vinyl groups have a second order cyclodimerization rate constant of formation between 1 × 10^-7 and 1 × 10^-4 M^-1 S^-1 while vinyl ethers have a second order cyclodimerization rate constant of formation = 1 × 10^-1 M^-1 S^-1. If there is a -CH₂- group α to the terminal olefin, the E_a of cyclodimerization is about 7 kcal mol^-1) lower with respect to those olefins with a -CF₂-α to the instauration. At 270 °C co-dimerizations have an average ΔS^≠ = -45 cal K^-1 mol^-1 and a second order rate constant of cyclodimerization ranging between 0.1 × 10^-4 and M^-1 S^-1 while homo-dimerizations have an average ΔS^≠ = -17 cal K^-1 mol^-1 and a second order rate constant which can span from 7 × 10^-1 M^-1 S^-1 to as much as 1 × 10^-1 M^-1 S^-1 depending on the electronic nature of the perfluorinated terminal olefin. A good correlation between the electronegativity x and the activation energy E_a demonstrates that polarizing groups, -O-, PhO, α to the olefin play an important role in the formation and stabilization of the cyclodimerization biradical intermediate.

Full text of DOI:10.1016/j.jfluchem.2004.06.014

Journal of Fluorine Chemistry 125 (2004) 1519–1528
www.elsevier.com/locate/fluor
2p + 2p cycloaddition kinetics of some fluoro olefins
and fluoro vinyl ethers
*
I. Wlassics
Solvay Solexis, Via S. Pietro, 50, 20021 Bollate, MI, Italy
Received 23 October 2003; received in revised form 24 March 2004; accepted 4 June 2004
Available online 6 August 2004
Abstract
The second order 2p + 2p homo- and co-dimerization between various classes of fluorinated olefins has been investigated. The fluorinated
olefins examined in this study were: (1) R –OCF=CF (perfluorinated vinyl ethers); (2) R –CF=CF (perfluorinated terminal olefins); (3) R–
(aryl perfluorinated vinyl ethers).
Homo-dimerizations between vinyl ethers have an E between 20 and 24 kcal mol while homo-dimerizations between terminal olefins
f
2
f
2
2 2 2
CH –CF=CF ; (4) Ph–OCF=CF
ꢀ1
a
ꢀ
1
have an average E between 35 and 40 kcal mol ; vinyl groups have a second order cyclodimerization rate constant of formation between
a
ꢀ
7
ꢀ4
ꢀ1 ꢀ1
ꢀ1
ꢀ1 ꢀ1
s . If
1
there is a –CH – group a to the terminal olefin, the E of cyclodimerization is about 7 kcal mol ) lower with respect to those olefins with a –
ꢁ 10 and 1 ꢁ 10
M
S
while vinyl ethers have a second order cyclodimerization rate constant of formation = 1 ꢁ 10
M
ꢀ
1
2
a
6
ꢀ1
ꢀ1
CF
cyclodimerization ranging between 0.1 ꢁ 10
2
– a to the instauration. At 270 8C co-dimerizations have an average DS = ꢀ45 cal K mol and a second order rate constant of
ꢀ4
ꢀ1 ꢀ1
ꢀ4
ꢀ1 ꢀ1
6¼
M
S
and 16 ꢁ 10
M
S
while homo-dimerizations have an average DS =
to as much as 1 ꢁ 10ꢀ1
ꢀ1
ꢀ1
ꢀ7
ꢀ1 ꢀ1
ꢀ1 ꢀ1
M S depending on
ꢀ
17 cal K mol and a second order rate constant which can span from 7 ꢁ 10
M
S
the electronic nature of the perfluorinated terminal olefin.
a
A good correlation between the electronegativity x and the activation energy E demonstrates that ‘‘polarizing’’ groups, –O–, PhO, a to the
olefin play an important role in the formation and stabilization of the cyclodimerization biradical intermediate.
2004 Elsevier B.V. All rights reserved.
#
Keywords: Cycloaddition; Kinetics; Fluoro olefins
1
. Introduction
substituted cyclobutane ring (almost exclusively a 1,2 dis-
ubstitution and not a 1,3 disubstitution) [12] and of its
stereochemistry [11]. Recent kinetic studies by Shtarov et
The thermal cycloaddition of terminal olefins to give
1
9
cyclobutane rings is a well-known reaction. It has been
utilized both as a methodology for small molecules [1–4]
and as a method to form high molecular weight materials
al. [13] employing gas phase F NMR confirm the results of
Lacher et al. Very recently, the biradical intermediate pro-
posed by Bartlett et al. has been observed spectroscopically
(EPR) for the reaction of an aryl trifluoro vinyl ether [14].
In spite of the literature regarding the synthetic aspects of
the [2 + 2] thermal cycloaddition of fluorinated olefins, little
has been done to study the effect (if any) of the substituents
on the n system of the olefins. In the present work we have
observed the effect of different substituents a to the trifluoro
vinyl group (substituents: perfluoroalkyl, oxa-perfluor-
oalkyl, fluoroalkyl of the type –CF –CH –, oxyaryl) on
[
5–10]. In sharp contrast with its hydrogenated counterpart,
Bartlett proposed that fluorinated olefins undergo ‘‘thermal
2 + 2] non-concerted cycloadditions’’ instead of ‘‘con-
[
certed’’ photochemical cycloadditions proceeding through
a ‘‘biradical’’ intermediate typical of hydrogenated olefins
[11]. Such a mechanistic explanation has also been proposed
by Lacher et al. on the basis of the regiochemistry of the
2
2
the energy of activation, E , the rate constant of formation
a
*
Tel.: +39-02-38356367; fax: +39-02-38356355.
E-mail address: ivan.wlassics@solvay.com.
of the cyclodimers, k , the second order half-life of
c
0022-1139/$ – see front matter # 2004 Elsevier B.V. All rights reserved.
doi:10.1016/j.jfluchem.2004.06.014

1520
I. Wlassics / Journal of Fluorine Chemistry 125 (2004) 1519–1528
cyclodimerization, t1/2, the frequency factor, A, the Eyring
6¼
and Lacher et al. [12], for cycloadditions with similar
olefins.
6
theory transition state activation parameters DH , DG , and
6¼
DS on the cycloaddition reaction. The effects of the chain
length on these kinetic parameters has also been investi-
gated.
From the data presented in Table 1, the following general
trends for homo-dimerizations are apparent:
1
. The presence of a vinyl ether (VE) lowers the E by about
a
ꢀ1
2
0 kcal mol
2. The rate of cyclodimerization between two alkyl vinyls
V/V) is slower by two–five orders of magnitude with
.
2
. Results and discussion
(
In Table 1 we show the energy of activation, E , the rate
a
respect to the rate of cyclodimerization between two
vinyl ethers (VE/VE).
3. The E of cyclodimerization of vinyls having a –CH – a
constants for the formation of the cyclodimers at the various
T
temperatures tested, k or k , and the second order half-life
c
a
2
c
RH RH
ꢀ1
of cyclodimerization, t1/2, for various classes of olefins. The
cycloaddition reactions between the olefins tested can be
grouped into three different classes shown in the columns
labeled ‘‘Reagents’’ and ‘‘System’’ (the superscript ‘‘S’’
to the insaturation (V /V ) is about 7 kcal mol
lower with respect to those olefins with a –CF – a to
the insaturation.
2
L
L
4. Long chainid olefins, V or VE (ꢂ5 carbon atoms) have
(
short) or ‘‘L’’ (long) in the ‘‘System’’ column indicates
a greater cyclodimerization rate constant, a greater fre-
quency factor and a more positive DS with respect to
6¼
whether the reaction is between olefins having <5 carbon
atoms, or olefins having ꢂ5 carbon atoms). The following
list summarizes the different types of reactions studied.
s
short chained, V or VE (<5 carbon atoms) olefins.
S
2.1. Effects of groups a to a perfluorovinyl group
ꢃ Reaction among vinyl ethers (‘‘VE’’);
ꢃ Reaction among alkyl vinyls (‘‘V’’);
RH
on E and k
a c
ꢃ Reaction between alkyl vinyls wherein a to the vinyl
A reasonable explanation of the experimental data
obtained can be given by observing the reaction mechanism
of 2 + 2 cycloaddition reactions between perfluorinated
olefins. The literature reports that the transition state of 2
+ 2 cycloaddition reactions between fluorinated olefins is a
di-radical [11,13,16,17]. Fig. 1 schematically depicts the
approach to one another of the two olefins right before the
group there is a –CH – group (‘‘V ’’).
2
The results reported for E and the frequency factor A
a
(
employed to calculate k and t1/2 and shown further on in
c
Table 2) are in good agreement with the data reported by
Shtarov et al. [13], Atkinson and Tsiamis [15], Babb [16],
Table 1
Kinetic parameters of 2 + 2 cycloaddition reactions
ꢀ
1 ꢀ1
ꢀ1
270 8C
c
ꢀ1 ꢀ1
270 8C
c
a
R
Reagents
System
T
r
(8K)
k
c
(M
s
)
C
0
(M)
E
a
(kcal mol
)
k
(M
s
)
k
(h)
L
S
(
(
(
(
(
(
(
III) + (II)
III) + (I)
I) + (II)
III)
1 VE /VE
443
ꢀ
4
15.8 ꢁ 10ꢀ4
11.4 ꢁ 10ꢀ4
4
83
43
6 ꢁ 10
14.5 ꢁ 10
ꢀ4
0.0075
12.5
16.3
23.4
2290
1652
ꢀ
4
5
L
S
2 VE /V
449
0.8 ꢁ 10
0.00157
155
404
ꢀ4
5
23
53
1 ꢁ 10
ꢀ
4
5
69 ꢁ 10
0.5 ꢁ 10
S
S
ꢀ5
1.4 ꢁ 10ꢀ5
1.3 ꢁ 10ꢀ1
4.04 ꢁ 10ꢀ5
6.9 ꢁ 10ꢀ7
4.7 ꢁ 10ꢀ4
3 V /VE
513
0.049
0.0089
0.457
1.35
18.6
24.1
34.7
41
20
ꢀ5
5
23
53
1 ꢁ 10
2 ꢁ 10
1 ꢁ 10
ꢀ5
5
L
L
ꢀ4
4 VE /VE
423
0.24
187500
59
ꢀ
2
1
4
83
43
0.55 ꢁ 10
0.55 ꢁ 10
ꢀ
5
RH RH
5 V /V
ꢀ7
V)
483
4 ꢁ 10
15
ꢀ
5
5
23
53
0.5 ꢁ 10
ꢀ5
5
4 ꢁ 10
1 ꢁ 10
4 ꢁ 10
1 ꢁ 10
S
S
ꢀ8
I)
6 V /V
493
299
1
ꢀ8
5
23
51
ꢀ6
5
L
L
ꢀ6
ꢀ
ꢀ
IV)
7 V /V
483
1.5 ꢁ 10
0.5 ꢁ 10
2.5 ꢁ 10
0.0235
44
25.2
682
4
5
23
43
4
5
k
c
is the second order rate constant of cycloaddition; C
0
the initial concentration of the starting monomer or of the limiting reagent in the case of co-
ꢄ
ꢄ
the energy of activation at k2
70
C
is the second order rate constant of cycloaddition at 2708C; t 21 =7 20 C is the second order half-life of
dimerizations; E
cycloaddition at 270 8C.
a
c
ꢄ
ꢄ
C
a
270
C
270
PFP
k
=k
:

I. Wlassics / Journal of Fluorine Chemistry 125 (2004) 1519–1528
1521
Table 2
Activation parameters for the 2 + 2 cycloaddition reaction
DH t2 s
(kcal mol
DG
ts
270
D G
ꢀ1
ts
DS
ts
Reagents
System
log A
E
a
ꢄ
ꢄ
ꢄ
270 C
70
C
C
270 8C
ꢀ1
ꢀ1
ꢀ1
(
kcal mol
)
)
(kcal mol
)
(cal K mol
)
L
S
S
(
(
(
(
(
(
(
III) + (II)
lll) + (l)
l) + (ll)
III)
V)
1)
1 VE /VE
2.25
3.64
2.66
8.85
9.63
10.33
14.4
12.5
16.3
18.6
24.1
34.72
41
11.4
15.2
17.5
23
39.3
39.6
44.3
34.5
43.19
47.6
40.6
ꢀ51.3
ꢀ44.9
ꢀ49.4
ꢀ21.14
ꢀ17.6
ꢀ14.4
4.31
L
2 VE A/V
S
S
3 V /VE
L
L
4 VE /VE
RH RH
5 V /V
33.65
39.8
42.9
s
s
6 V /V
7 V /V
L
L
IV)
44
2
transition state complex is achieved; as soon as the di-
radicals have aligned themselves (transition state complex)
the cyclodimerization is instantaneous [13].
fairly good (R = 0.941 for the least squares fit) and suggests
that as the electronegativity of the group a to the olefin
increases, the polarization of the olefin p bonds will be
easier. This will favor the formation of the di-radical inter-
mediate and consequently the 2 + 2 cycloaddition reaction.
0
In Fig. 1 X may be equal to X and is defined as one of the
following atoms or groups of atoms: –O–, –O–Ph, –CF –, –
2
CH –; R is either a long (ꢂ5 carbon atoms) or short (<5
2
0
carbon atoms) perfluoroalkyl chain. If X or X are a –CH –
2
2.2. Effects of groups a to a perfluorovinyl group on the
frequency factor and on the entropy of activation
group, then R = R (CH ) with n ꢂ1 where R = a per-
f
2 n
f
fluoroalkyl chain. In an independent work Babb [16], Lomas
0
and Tarrant [18] show that both X and X in Fig. 1 can
polarize the olefin, thereby favoring the di-radical inter-
mediate.
From the kinetic data obtained in Table 1, it is possible to
easily calculate the other activation parameters: the enthalpy
6
6¼
of activation, DH , the free energy of activation, DG , and
the entropy of activation, DS , at any desired temperature (in
0
6
The electronic nature of X and/or X in Fig. 1 must
therefore be fundamental for the polarization of the p bond
of the olefin thereby promoting the radical species and
consequently the 2 + 2 cycloaddition reaction. In order to
verify this hypothesis, we plotted the activation energy, Ea,
of four compounds examined in this study each with a
different group a to the olefin whose general formulae are
shown below (the different groups are shown in bold char-
acters) with respect to the effective electronegativities, x, of
the groups a to the olefin which are reported in the literature
by Powell and Lagowski [19], Huheey [20] and Morreland
our case, 270 8C) according to Eqs. (4)–(6) reported in the
literature [23] and discussed in detail in Section 4.
The activation parameters calculated for all of the 2 + 2
cycloaddition systems studied are summarized in Table 2
and they are in good agreement with values reported in the
literature [11,13,17] for similar olefins and reaction condi-
tions.
On average, 2 + 2 cyclodimerizations involving two
different types of olefins (co-dimerizations) have relatively
6
more negative values for DS and consequently lower
and Brey [21]. The E for the compound (4) shown below
was determined by the author [22]
frequency factor values with respect to cyclodimerizations
involving only one olefin (homo-dimerizations). This could
indicate a greater difficulty for each of the two olefins to
correctly orient itself in order to react and give the desired 2
a
R ꢀꢀ ðCF Þ ꢀꢀ CF ¼¼ CF
(1)
(2)
(3)
(4)
2
n
2
+
2 co-dimer. The data in Table 2 show that homo-dimer-
R ꢀꢀ ðCH Þ ꢀꢀ CF ¼¼ CF
2
n
2
izations require less energy in order to achieve the correct
orientation in order to react; in fact, homo-dimerizations
R ꢀ ðCF Þ ꢀꢀ CF ꢀꢀ O ꢀꢀ CF ¼¼ CF
6
2
n
2
2
have relatively more positive values of DS and conse-
quently a higher value for the frequency factor A. Further-
more, if the olefin in the homo-dimerization reaction has a
polarizing group (i.e. an oxygen atom) in the X position (see
Fig. 1) that system will react with relatively lower activation
R ꢀꢀ C H O ꢀꢀ CF ¼¼ CF
f
6
4
2
Fig. 2 shows the plot of E with respect to effective
a
electronegativity, x. The correlation between E and x is
a
energy E and with a relatively faster rate constant of
a
cyclodimer formation, k , at a given temperature T. If the
c
olefin involved in the homo-dimerization lacks a polarizing
group a to the terminal olefin it will react with a relatively
higher E and a relatively slower k at a given temperature T,
a
c
even if that system has relatively high frequency factor A and
6¼
therefore a relatively positive entropy of activation DS .
6¼
This observation indicates that relatively positive DS and
consequently high frequency factors are necessary, but are
Fig. 1. The di-radical intermediate of the 2 olefins in a 2 + 2 cycloaddition
reaction.

1522
I. Wlassics / Journal of Fluorine Chemistry 125 (2004) 1519–1528
a
Fig. 2. Correlation between electronegativity, x, and E for the 2 + 2 cycloaddition reaction.
not the only determining factors, for a 2 + 2 cyclodimeriza-
tion to occur.
chained system and is evidenced by a relatively more
6
¼
6¼
negative DS . A more positive DS indicates that in the
transition state the correct approach (for a cycloaddition
reaction to occur) of the two di-radical intermediates for
the long chained system is not as constraining as it is for
the short chained system. As a consequence of this ease in
achieving the proper conformation of the di-radical inter-
2
2
.3. Effect of chain length on kc 70 8C, E and ꢀS
ꢄ
270 C
a
The chain length of either vinyls or vinyl ethers seems to
6¼
influence the kinetic parameters of entropy of activation DS
demonstrating that a correct spatial orientation is necessary
for a cyclodimerization to occur. Comparing the 2 + 2
cyclodimerization at 270 8C (within the same system and
therefore with an equal polarizing effect on the terminal
mediate in the transition state, k is much greater as
c
compared to the short chained system. The literature does
report [12,13,15,24] that short, volatile and not particu-
larly polarized compounds such as CTFE and TFE in fact
have relatively slow second order cyclodimer rate con-
stants of formation which are of the same magnitude as the
kc values reported for the short chained non polarized
systems of this study.
s
s
olefin) of (I) (V /V ) and of CH CH (CF ) CF=CF (IV)
3
2
2 2,4
2
L
V /V ) shown in Tables 1 and 2, one notices that the system
L
(
of vinyls with a relatively longer chain length has a cyclo-
dimer formation rate constant, k , almost 3 orders of mag-
c
nitude greater with respect to the k of the shorter system.
c
The same trend just described is observed again for a
S
L
The V /V system has also a greater frequency factor A, a
L
S
different system of olefins: (I) + (II) (V /VE ) with respect
6¼
ꢀ1
s
L
more positive DS , and an E 3 kcal mol greater with
a
to (I) + HCF (CF ) OCF=CF (III) (V /VE ).
2 2 11 2
s
s
respect to the V /V system.
The greater E can be explained in terms of a greater
a
2.4. Co-dimerizations
steric hindrance of the long chained system with respect to
the short chained system.
Concerning the other two transition state activation
Co-dimerizations are a competition reaction and are
therefore discussed separately.
6
¼
parameters k and DS , one must consider the fact that
c
shorter molecules are more volatile and therefore more
‘‘energy rich’’ at equal temperatures with respect to
Table 3 shows the details of the three co-dimerizations
studied. For the two co-dimerizations involving the very
slow monomer (I) an initial mol ratio of 9:1 for (I) versus
either the fast monomer (II) or (III) had to be employed in
order to be able to accurately characterize the resulting co-
dimer (either I–II or I–III).
longer, less volatile compounds. To bring close together
two such ‘‘energy rich’’ compounds in the transition state
is probably a constraining situation with respect to the long

I. Wlassics / Journal of Fluorine Chemistry 125 (2004) 1519–1528
1523
Table 3
Co-dimer selectivity as a function of time and temperature
Reagents
t (h) T (8C) Experimental
mmol (I)
Selectivity
mmol cyclodimers (I–I)
mmol (II)
(II–II)
mmol
0
(I–II)
mmol
0
mmol
(%)
0
(%)
0
(%)
(
(
(
l) + (ll); reactor =10 ml
I) + (III); reactor = 45 ml
II) + (III); reactor = 65 ml
0
–
4401
489
0
0
0
4
8
0
240
4391
4381
4285
449
410
107
25
3.7
7.5
37
14.8
15
18.5
37
74
74
68
2.8
5.4
11
11
17
49.9
249.5
1
15
170
42.5
5
0
0
250
280
4354
4309
4030
430
368
9
53
21
40
39.6
37.6
27
51
5.3
12
10
11
12
1
4
106.5
425.5
54.5
51
160
37.6 215
50.5
50
5
0
0
4282
4149
3894
426
375
266
91
183
365
56
112
225
61.5
61
28
43
83
30.7
23.5
23
7
28
57
8
15
16
1
2
62
mmol (I)
mmol (III) mmol cyclodimers (I–I)
mmol
(III–III)
mmol
0
(I–III)
mmol
0
(%)
0
(%)
0
(%)
0
0
–
628
70
0
0
4
0
0
176
627.8
627.4
625.6
69
0.6
0.002
0.004
0.17
0.3
0.3
0.3
0.4
0.9
3.5
67
60
59
0.198 33
1
4
67.6
60.6
1.5
0.59
2.4
39
41
5.95
5
0
0
250
280
627.2
626.5
622.2
68
1.4
2.7
0.15
0.32
1.19
11
12
11
0.77 55
1.49 55
0.48
0.89
3.46
31
33
32
1
4
66.1
54.1
10.8
6.2
57
1
5
0
614
61
25.2
3
11.3
56
5.1
45
44
57
2.6
13.4
18.5
23
24
16
3.7
32
32
27
560.8
469
24.6
64.5
18
1
113
30
mmol (II) mmol (III) mmol cyclodimers (II–II)
mmol
(III–III)
mmol
0
(II–III)
mmol
0
(%)
0
(%)
0
(%)
0
0
–
490
490
0
0
1
4
0
170
483.2
459.8
428
481.8
454.2
406
7.5
2.2
9.8
29
30
27
2.9
12.6
31
38.7
38
2.4
10.6
22
32
32
30
.5
33
1
73
20
42
1
3
6
210
270
454.2
382
447.8
364
39
117
235
11.6
35
29
30
29
14.8
44
38
37
38
12.6
38
32
33
32
276
234
69
90
76
1
2
6
376
262
30
366
242
2
119
238
474
37
74
31
31
31
40
80
34
34
34
42
84
35
35
35
148
162
164
Observing Table 3, an interesting fact emerges: the
cycloaddition rate increases by two orders of magnitude.
This, coupled by the fact that (I) is nine times more con-
centrated than the fast co-dimer (II) or (III) explains the
trend observed in Table 3: as temperature increases more of
the monomer (I) has enough energy to react and therefore
the selectivity of the homo-dimer (I–I) increases and the
selectivity of homo-dimer (II–II) or (III–III) consequently
decreases.
selectivity of co-dimer formation is practically independent
of temperature. Instead, reactions involving monomers I and
II and I and III show a marked dependence of selectivity on
reaction temperature for the competing homo-dimmers I–I,
II–II and III–III. This can be explained by observing Table
1
which shows the cyclodimerization rates of the pure
monomers at various temperatures. At temperatures between
70 and 250 8C the rate of homo-dimerization of monomer
1
In Table 1 we notice that the co-dimerization rate con-
stant, k , is rather slow even between two vinyl ethers and at
ꢀ8
ꢀ1 ꢀ1
M s
(
I) is in the order of 10
while at 280 8C its
c

1524
I. Wlassics / Journal of Fluorine Chemistry 125 (2004) 1519–1528
temperatures as high as 270 8C ꢀ280 8C (monomers II and
III, the most reactive species according to the considerations
made in Sections 2.1 and 2.2) compared to the homo-
dimerization of vinyl ether (III) and are even not much
It has also been demonstrated that chain length does have
an effect on the kinetics of cyclodimerization, and is
reflected by a greater second order cyclodimerization rate
6
constant and a more positive (S for long chained molecules
different from the k involving, the slow vinyl monomer (I).
with respect to short chained (<5 carbon atoms) molecules.
Finally it has been shown that co-dimerizations require
greater energy to achieve the proper spatial orientation
c
6¼
70 C
co-dimerizations with respect to homo-dimerizations and
In Table 2 we also notice both the very negative DS2
ꢄ
for
6
6
the large gap in DS2
ꢄ
between co-dimerizations and
6¼
(relatively less positive (S ) with respect to homo-dimer-
izations.
70 C
homo-dimerizations. The very large negative DS2
values and the slow k values shown in Tables 1 and 2
ꢄ
70 C
can be explained in terms of the difficulty in obtaining the
correct spatial orientation and collision frequency for the di-
radical intermediate to be formed and therefore a cycloaddi-
tion to occur between two species rather than just one. This
is not just a result of the 9:1 mol ratio between (I) and either
4. Experimental
1
19
H and F NMR spectra were recorded on a Varian
spectrometer operating at 200 MHz using CFCl3 as an
internal standard. FT-IR spectra were recorded on a Nicolet
Avatar 360 FT-IR ESP interfaced with OMNIC software.
(II) or (III) (merely a difficulty of the monomer in excess (I)
in ‘‘finding’’ (II) or (III)); in fact, Table 2 shows that the co-
6¼
70 C
dimer (II–III) has even a more negative DS2
ꢄ
(and
consequently a lower value of log A) while the respective
concentrations of the starting monomers at t = 0 was 1:1 as
shown in Table 3.
4
.1. 2 + 2 homo-and co-dimerizations
The following compounds (I–V) were tested in this study;
in parenthesis we give their boiling point and the ‘‘system’’
Once the correct approach of the di-radical intermediate
in the transition state for the co-dimerization to occur has
been achieved, the actual 2 + 2 cycloaddition is not a very
energy demanding process as evidenced by the relatively
low E values reported in Table 1. Comparing the E values
(see Section 2 and Table 1) it belongs to:
S
ꢃ (I) CF CF=CF (bp = ꢀ28 8C;‘‘V ’’);
3
2
S
a
a
ꢃ (II) CF CF OCF OCF=CF (bp = 40 8C; ‘‘VE ’’);
3
2
2
2
for the co-dimerizations in Table 1 one notices that co-dimer
selectivity (see Table 3) is inversely proportional to E : as
ꢃ (III) HCF (CF ) OCF=CF (bp = 170 8C–180 8C;
2
2 11
2
L
‘VE ’’);
a
‘
co-dimer selectivity increases its E decreases.
a
ꢃ (IV) CH CH (CF ) , CF=CF (bp
=
75 8C–80 8C;
3
2
2 2 4
2
Finally, it must be stated that the 2 + 2 cyclodimerization
rates of the homo-dimers (I–I), (II–II) or (III–III), can not
be determined from the homo-dimer data reported in Table
L
‘V ’’);
‘
RH
ꢃ (V) CF =CF(CH ) CH=CH (bp = 110 8C; ‘‘V ’’).
2
2 4
2
3
. Separate experiments must be performed in order to
For a general perfluorinated terminal olefin R–CF=CF2,
all 2 + 2 cyclodimerization reactions, either homo-dimer-
izations or co-dimerizations, had products with the general
formulae (VIa, VIb and VIc) shown below and were
determined by comparing F NMR spectra reported in
the literature [13,18,25]. The general structure of formula
obtain the kinetic data for the homo-dimerizations reported
in Tables 1 and 2. This is because co-dimerization reactions
are a competition reaction and because, in the case of co-
dimers (I–II) and (I–III), the initial mol ratios were 9:1 for
19
(I) versus (II) or (III) as explained above. This causes a
marked ‘‘dampening’’ effect on the k of the rate limiting
c
(VI) (either a, b or c) is the one also proposed by Lomas and
monomer (as high as a 60–70% lowering of the homo-dimer
yield with respect the yield observed for the same homo-
dimer, but in a homo-dimerization experiment at equal
reaction temperature and times). The same ‘‘dampening’’
effect is observed for the monomer in excess, but to a much
lesser extent (just a 15–20% lowering of the homo-dimer
yield).
Tarrant [18], Bartlett et al. [25] and by Coffman et al. [26] for
cycloaddition reactions between similar perfluorinated term-
inal olefins.
3
. Conclusions
All 2 + 2 cyclodimerization reactions were carried out
either in a stainless steel reactor if the compound(s) to be
tested was a gas or had a low boiling point (compounds (I)
and (II)) or in a thick wall (3 mm) glass ampoule equipped
with a bleeding valve if the compound(s) to be tested had
high boiling points (compounds (III–V)). The desired moles
of fluoro olefins which were a liquid at room temperature
In this study we have demonstrated that the electronic
nature of groups a to the terminal perfluorinated olefin has a
considerable effect on the stabilization of the biradical
intermediate in 2 + 2 cyclodimerization reactions. This in
turn effects both the kinetics and the activation parameters of
such reactions.

I. Wlassics / Journal of Fluorine Chemistry 125 (2004) 1519–1528
1525
1
9
were transferred in the glass reactor simply by accurately
weighing the olefin on an analytical balance and transferring
it with a Pasteur Pipette. Instead, the correct amount of
moles of compounds (I) and (II) were transferred employing
the Ideal Gas Law, PV = ZnRT, and the following apparatus.
The cylinder containing (I) or (II) (the cylinder containing
F NMR for cyclodimer of system 3 (formula VIc):
ꢀ149 ppm (a), ꢀ118 ppm (b).
1
9
F NMR for cyclodimer of system 4 (formula VIc):
ꢀ148 ppm (a), ꢀ118 ppm (b).
1
9
F NMR for cyclodimer of system 5 (formula VIa):
ꢀ190 ppm (a), ꢀ117 ppm (b).
1
9
(
II) was kept on a heating mantle at a known temperature in
F NMR for cyclodimer of system 6 (formula VIa):
ꢀ194 ppm (a), ꢀ110 ppm (b).
order to have it completely in the gas phase) was directly
connected by a bleeding valve to a glass ampoule of known
volume which was in turn directly connected to a digital
pressure reader and to the stainless steel reactor by another
bleeding valve. Both the glass ampoule and the reactor were
previously evacuated. The moles needed for the reaction
were transferred from the cylinder into the glass ampoule by
reading the pressure on the digital reader. Once the correct
pressure was reached, the bleeding valve between the cylin-
der and the glass ampoule was closed. The reactor was
placed in liquid nitrogen and the desired amount of (I) or (II)
was condensed into the reactor by opening the bleeding
valve between the glass ampoule and the reactor.
1
9
F NMR for cyclodimer of system 7 (formula VIa):
ꢀ184 ppm (a), ꢀ118 ppm (b).
Mixed cyclodimerizations reactions have three different
cycloaddition products: the co-dimerization cycloaddition
product and two homo-dimers arising from each of the two
starting monomers. Table 3 reports the experimental details
of the three co-dimerizations studied in this work along with
the selectivities of the co-dimerization products reported
both as a function of time and temperature.
4.1.1. Sampling
In the case of reactions in the stainless steel reactor, the
sampling was performed directly during the course of the
reaction by placing a ‘‘T’’ between the neck of the reactor
and the transducer (apparatus, connected to a reader, that
measures the pressure inside the reactor). The ‘‘T’’ was
connected to a pressure reducer whose exit was connected to
a septum. This permitted to sample the reaction mixture
employing an appropriate gas syringe (which can be oper-
ated at pressures up to 2 atm) even if the pressure inside the
reactor was >2 atm at the time of the sampling. The same
was done for reactions which employed the thick-walled
glass reactor only that the pressure reducer was omitted
since the pressure inside the reactor was <2 atm. Fig. 3
All 2 + 2 cyclodimerizations were carried out at three
different temperatures ranging between 170 and 280 8C. For
every temperature, samplings were taken at four different
reaction times. The temperatures chosen for each cyclodi-
merization reaction (see Table 1 for details) and conse-
quently the pressure generated inside the reactor, were
such so as to guarantee, based upon the liquid–vapour
equilibrium of each compound (not shown) that the fluoro
olefins involved in the cycloaddition were always in the gas
phase and therefore homogeneously mixed. In order to
achieve this very important physical condition for a correct
kinetic analysis without excessive experimental or mathe-
matical complications, it was not always possible to have the
same initial concentration, C , for every experiment. This is
0
4.2. Kinetic parameters
clearly shown in detail in Table 1. In the case of co-
dimerizations, C is of the limiting reagent.
At the end of the reaction the crude reaction mixture was
It has already been reported in the literature [15,16,24]
that 2 + 2 cyclodimerization kinetics follow second order
kinetics according to Eq. (la) at a fixed temperature [23].
0
1
analyzed by F NMR, using CFCI as an internal standard,
9
3
in order to determine the conversion and the structure of the
desired cyclodimer. The chemical shifts already reported in
the literature [13,18] were used as an aid to correctly identify
the various cyclodimers obtained.
ꢀ
d½XŠ=dt
2
¼ k ½XŠ
c
(1a)
2
Solving la for X one obtains Eq. (1b)
A kinetic sample was considered valid when the reaction
conversion was between about 5 and 60%. Those reactions
that converted >60% or < 5% were repeated. This was done
in order to avoid deviations from linearity of the reaction
kinetics due to an excessive dilution of the reagents or of the
products (due respectively to a high or low conversion).
1
1
¼ 2kct þ
(1b)
½
XŠt ½XŠ0
where [X] is the concentration (in mol l ) of the starting
ꢀ
1
t
1
9
olefin X at reaction time t. It is determined by F NMR by
calculating the moles of starting olefin to give the desired
cyclodimer. The volume is that of the reactor employed. For
co-dimerizations, in order to account correctly for the
kinetics of just the co-dimer, only the moles of the co-dimer
1
9
The F NMR shifts relative to the cyclobutane ring are
reported below for each ‘‘system’’ (see Tables 1 and 2)
studied. Spectroscopic details of each of the R groups
synthesized are reported in Section 4.3.
1
9
were taken into consideration ( F NMR) for the determina-
tion of [X] . k is the second order rate constant of formation
1
9
F NMR for cyclodimer of system 1 (formula
VIb):ꢀ140 ppm (a), ꢀ118 ppm (b), ꢀ188 ppm (c).
t
c
ꢀ
1
ꢀ1
of cyclodimer (in M
(where T is temperature). If the reaction follows second
s ) which is known [23] to be a f(T)
1
9
F NMR for cyclodimer of system 2 (formula VIb):
151 ppm (a), ꢀ117 ppm (b), ꢀ192 ppm (c).
ꢀ
order kinetics, a plot of l/[X] with respect to time t for each
t

1526
I. Wlassics / Journal of Fluorine Chemistry 125 (2004) 1519–1528
Fig. 3. Plot of 1/[X] vs. time at different temperatures for the determination of the second order rate constant of cycloaddition, k
c
, as f(T) for the 2 + 2
cycloaddition of CH CH (CF )2,4CF=CF
3
2
2
2
.
ꢄ
3. k270 8C/k
270 C
PFP
= a parameter which gives the relative
temperature studied will give a straight line with a slope =
k and y-intercept = 1/[X] as shown in Fig. 4. t is the
reaction time (in seconds). [X] is the initial concentration
2
reactivities for the different cycloaddition reactions with
respect to the cyclodimerization of (I), the slowest of the
cycloadditions studied and permits one to easily evaluate
the different effects of the polarizing or non polarizing
groups a to the terminal vinyl group. The rate constant of
formation of the cyclodimers at any temperature T
(270 8C in our case) were calculated according to Eq.
(2b) [23].
c
0
0
ꢀ
1
(
in mol l ) of the olefin.
In this study the following kinetic parameters were
determined:
ꢀ
1
. E : the activation energy (in kcal mol ). The linear
1
a
regression of the plot hi k with respect to 1/T gives a
c
straight line. The Arrhenius equation shown in Eq. (2a)
[23] can therefore be used to determine both E and the
frequency factor, A, as shown in Fig. 4.
ꢀ
ꢁ
Ea
a
T
c
k ¼ A ꢁ e ꢀ
(2b)
RT
Ea
Figs. 3 and 4 respectively show, as a typical example, the
plots of 1/[X] with respect to t at different reaction tem-
In k ¼ ꢀ
þ ln A
(2a)
c
RT
where k is the previously defined second order rate constant
c
peratures, used to determine k for each temperature tested,
c
of formation of cyclodimer determined using Eq. (lb). ꢀE /R
and the plot of ln k with respect to 1/T used to determine E
a
a
c
is the slope of the straight line resulting from Eq. (2a). R =
and the frequency factor A for the 2 + 2 cyclodimerization of
CH CH (CF ) CF=CF2.
ꢀ1
ꢀ1
l.98 cal K mol and A is the Arrhenius frequency factor.
T
=2
3
2
2 2,4
2
. t1 is the half-life of the starting olefin to give the desired
cycloaddition product at temperature T. For a second
From the kinetic data obtained thus far, we calculate the
6¼
enthalpy, DH , the free energy, DG , and the entropy,
T
6¼
T
6
order reaction the half-life is determined as shown in Eq.
(3).
DS of activation, at any temperature T, according to Eqs.
(4)–(6) [23].
T
1
T
6
t
¼
(3)
DH ¼ E ꢀ RT
(4)
1
=2
T
a
kcX0
!
6
where k and X are the same parameters defined earlier in
c
0
k T
B
ꢀDG
k
T
kT ¼
ꢁ e
(5a)
Eq. (lb) or Eq. (2b).
h
RT

I. Wlassics / Journal of Fluorine Chemistry 125 (2004) 1519–1528
1527
c a
Fig. 4. Plot of In k vs. 1/T for the determination of the energy of activation, E , and the frequency factor, A, for the 2 + 2 cycloaddition of
CH
3
CH
2
(CF
2
)
2
2,4CF=CF .
ꢀ
ꢁ
T
k ꢁ h
kBTk
6
¼
After 48 h conversion was complete and the crude mixture
was filtered to eliminate the salts and residual Zn powder.
DG ¼ ꢀln
RT
(5b)
(6)
T
+
After repeated washings with H O , the crude product was
3
DG ¼ DH ꢀ TDST¼
6¼
6¼
6
taken up in C
F14 and dried over MgSO . After stripping of
6 4
T
T
the solvent, 10 g of pure (III) were obtained. Yield = 48%.
1
9
where E is the activation energy (in cal/mol) calculated
ꢀ
F NMR: ꢀ85 ppm (m, 2F, –OCF –), ꢀ112 ppm and
a
2
1
according to Eq. (2a) and Fig. 4, R =1.98 cal K mol ,
ꢀ1
ꢀ120 ppm (dd, m, 2F, CF =CF–), ꢀ122, ꢀ124, ꢀ126,
2
T = 543 K, k is the second order rate constant of cyclodi-
c
ꢀ130 ppm (s, 20F internal –CF –), ꢀ135 ppm (dd, IF,
2
merization determined at temperature T (270 8C in our
study) according to Eq. (2b). h is Planck’s constant: 6.62
CF =CF–O–), ꢀ138 (d, 2F, –CF CF H).
2
1
2
2
H NMR: 6.2 ppm (m, 1H, –CF CF H).
2
2
ꢀ
34
ꢀ1
ꢀ23
ꢀ1
ꢀ1
ꢀ1
ꢁ
10
k is the transmission coefficient = 1 l mol
J s , k = Boltzman’s constant: l.38 ꢁ l0
J K
,
) as defined
FT-IR: 3510 cm (–CF –H st), 1839 cm (–OCF=CF2
2
B
ꢀ
1
ꢀ2
ꢀ1
s
st), 1341, 1287, 1163 cm (CF st).
in the literature [13].
4.3. Synthesis of the cyclodimer precursors (I)–(V) and
compound 4
4.3.2. Synthesis of CH CH (CF ) , CF=CF (IV)
2
3
2
2 2 4
The precursor of (IV), C1CF CFC1(CF ) , CH CH I
2 2 2 4 2 2
(
IVa), was synthesized in several steps starting from CTFE
CF CF=CF (I) was purchased from Fluorochem Ltd.
and IC1 according to known literature procedures [31–34].
(IVa), 50 g, 0.117 mol, are dripped in a stirring solution of
100 ml of anhydrous DMF at 70 8C containing 0.17 mol of
3
2
and was used without further purification.
CF CF2OCF OCF=CF2 (II) was synthesized according
to literature procedure [27].
3
2
powdered Zn. After 4 h the crude mixture is filtered, taken
+
up in C6F14 and washed several times with H O . The
3
4
.3.1. Synthesis of HCF (CF ) CF OCF=CF (III)
organic layer is dried over MgSO , filtered and the solvent
4
2
2 10
2
2
The precursor of (III), CF (CF ) –OCFClCF Cl (IIIa),
is distilled. Twenty grams (yield = 74%) of (IV) are obtained
which appear as a clear, colourless liquid.
2
2 10
2
was synthesized in several steps starting from C F and SO
3
2
4
1
9
according to known literature procedures [28–31]. (IIIa),
.03 mol, were then placed in 60 ml iPrOH and were
vigorously stirred with 0.14 mol of Zn powder at 90 8C.
F NMR: ꢀ94 and ꢀ110 ppm (dd, m, 2F, CF =CF–),
2
0
ꢀ119, ꢀ120, ꢀ122, ꢀ128 ppm (s, 5F internal –CF –),
2
ꢀ190 ppm (dd, IF, CF =CF–CF –).
2
2

1528
I. Wlassics / Journal of Fluorine Chemistry 125 (2004) 1519–1528
1
[
10] D.W. Smith Jr., D.A. Babb, H.V. Shah, A. Hoeglund, R. Traiphol, D.
Perahia, H.W. Boone, C. Langhoff, M. Radler, J. Fluorine Chem. 104
2000) 109–117.
11] P.D. Bartlett, Science 159 (1968) 833–838.
H NMR: 1.3 ppm (t, 3H, CH CH –), 2.2 ppm (q, 2H,
3
2
CH CH –).
3
2
(
ꢀ
1
ꢀ1
(CH CH – st), 1790 cm
FT-IR: 2910, 2850 cm
CF =CF– st), 1250, 1150 cm (internal –CF– st).
3
2
[
ꢀ
1
(
2
[12] J.R. Lacher, G.W. Tompkin, J.D. Park, J. Am. Chem. Soc. 74 (1952)
693–1696.
1
[
13] B.A. Shtarov, P.J. Krusic, B.E. Smart, W.R. Dolbier, J. Am. Chem.
Soc. 123 (2001) 9956–9962.
4
.3.3. Synthesis of CF =CF(CH ) CH=CH (V)
2 2 4 2
The precursor of (V) BrCH CH (CH ) CF=CF (Va),
2
2
2 4
2
[
14] S.C. Ligon, D.W. Smith, S. Chen, S. Suresch, J. Jin, C.M. Topping, V.
Mellon, J. Ballato, S. Foulger, 16th Winter Conference on Fluorine
Chemistry; Ft. Lauderdale, FL, USA, 13–18 January, 2003.
was synthesized in several steps starting from Br and CTFE
2
according to known literature procedures [33,35], followed
by the same dechlorination procedure described earlier for
the conversion of (IIIa) to (III). Then, a potassium ethoxide
[15] B. Atkinson, C. Tsiamis, Intl. J. Chem. Kinetics (1979) 585–593.
[
[
[
16] D. Babb, Fluoropolymers 1: Synthesis, Hougham et al. (Ed.), Plenum
Press, New York, 1999, pp. 25–50.
(
0.123 mol)–ethanol solution is slowly dripped in a solution
17] N.N. Buravtsev, Y.A. Kolbanovsky, J. Fluorine Chem. 96 (1999)
of (Va), 0.112 mol, in 20 ml of anhydrous ethanol. The
solution is stirred at 80 8C until all of the KBr has pre-
cipitated. The crude mixture is filtered and washed with
3
5–42.
18] D. Lomas, P.J. Tarrant, Org. Chem. 34 (1968) 323–326.
[19] H.B. Powell, J.J. Lagowski, J. Chem. Soc. (1965) 1392–1396.
+
[20] J.E. Huheey, J. Phys. Chem. 69 (1965) 3284–3291.
[21] C.G. Morreland, W.S. Brey, J. Chem. Phys. 45 (1966) 803–808.
[22] I. Wlassics, V. Tortelli, A. Zompatori, M. Sala, Solvay-Solexis Internal
Report 07/200, pp. 26–27.
H O . The organic layer was distilled, bp 90 8C at
640 mmHg, obtaining 10.3 g of a clear, colourless liquid.
Yield = 65%.
3
1
9
F NMR: ꢀ108 and ꢀ126 ppm (dd, m, 2F, CF CF–CH –)
2
2
[23] D. Eisenberg, D. Crothers, Physical Chemistry, The Benjamin/Cum-
ming Publishing Co., 1979.
ꢀ
176 ppm (dd, IF, CF =CF–CH –).
2 2
1
[
[
24] B. Atkinson, M.J. Stedman, Chem. Soc. (1962) 512–519.
H NMR: 5.95 ppm (m, 2H –CH=CH ), 5.8 ppm (m, 1H
2
25] P.D. Bartlett, L.K. Montgomery, B. Siedel, J. Am. Chem. Soc. 86
–
CH=CH ), 3.7 ppm (m, 2H, –CH Br), 2,2 (m, 6H, internal
2 2
(
1964) 616–622.
–
CH –).
2
[
26] D.D. Coffinan, P.L. Barrick, R.D. Cramer, M.S. Raasch, J. Am. Chem.
Soc. 71 (1949) 490–496.
ꢀ
1
FT-IR: 2984 cm (–CH – st), 1798 cm (CF =CF st),
ꢀ1
2
2
ꢀ
1
1
644 cm (–CH=CH2 st).
Ph–OCF=CF (4) was synthesized according to literature
procedure [7,36,37].
[27] W. Navarrini, S. Corti, 14th European Symposium on Fluorine Chem-
istry, Oral Communication C-O-05, 11–16th July, Poznan, Poland, p.
2
3
16.
28] W. Navarrini, V. Tortelli, A. Russo, S. Corti, J. Fluorine Chem. 95
1999) 27–39.
29] F.J. Pavlick, US Patent 3,420,877 CA 70:67609 (1969).
[
[
(
References
[30] W. Grot, C. Molnar, P. Resnik, Belgian Patent 86122 (1978).
[
[
[
[
31] G. Baudin, B. Boutevin, R. Bertocchio, A. Lantz, C. Verge, J. Fluorine
Chem. 90 (1998) 107–115.
[
[
[
[
[
[
1] J.D. Roberts, C.M. Sharts, Org. React. 12 (1975) 1–56.
2] D.R.A. Perry, Fluorine Chem. Rev. 1 (1967) 253–313.
3] W.H. Sharkey, Fluorine Chem. Rev. 2 (1968) 1–36.
32] B. Ameduri, B. Boutevin, G. Kostov, P. Petrova, J. Fluorine Chem. 74
(
1995) 261–267, and references therein.
33] B. Ameduri, B. Boutevin, M. Nouiri, M. Taibi, J. Fluorine Chem. 74
1995) 191–197.
34] I. Wlassics, E. Barchiesi, M. Sala, J. Fluorine Chem. 123 (2003) 119–
26.
4] R.D. Chambers, R.H. Mobbs, Adv. Fluorine Chem. 4 (1965) 50–106.
5] R. Beckerbauer, US Patent 3,397,191 CA 67:100588 (1968).
6] D.A. Babb, B.R. Ezzell, K.S. Clement, W.F. Richey, A.P. Kennedy, J.
Polym. Sci., Part A: Polym. Chem. 31 (1993) 3465–3477.
7] D.W. Smith, D.A. Babb, Macromolecules 29 (1996) 852–860.
8] J. Rizzo, F.W. Harris, Poly. Pepr. 40 (2) (1999) 874–875.
9] D.W. Smith Jr., D.A. Babb, S. Kumar, J. Ballato, H. Shah, C. Topping,
S. Foulger, Adv. Mater. Sci. 14 (21) (2002) 1585–1589.
(
1
[
35] P. Tarrant, M.R. Lilyquist, J. Am. Chem. Soc. 77 (1955) 3640–3642.
36] X. Li, H. Pan, X. Jiang, Tetrahedron Lett. 25 (1984) 4937–
[
[
[
[
4
940.
[
37] H. Hoshiumi, H. Aoyama, EP 1247791 CA 135:108913 (2000).