A convenient and practical method for the selective preparation of deuterofluorocarbons

Article abstract of DOI:10.1016/j.jfluchem.2015.10.002

A detailed study of the development of efficient and practical conditions for the selective synthesis of 1-deuterononafluorobutane from 1-iodononafluorobutane is reported. The optimal conditions involve treatment of the iodo-precursor in D₂O at ～170 °C in the presence of metallic zinc in a sealed Schlenk tube to give a 59% yield of 1-deutero-1,1,2,2,3,3,4,4,4-nonafluorobutane. The same method was applied successfully to two higher homologues to produce 1-deutero-1,1,2,2,3,3,4,4,5,5,5-undecafluoropentane and 1-deutero-1,1,2,2,3,3,4,4,5,5,6,6,6-tridecafluorohexane in yields of 64% and 56%, respectively. Surprisingly, even the non-perfluorinated product 6-deutero-1,1,1,2,2,3,3,4,4-nonafluorohexane could be synthesized in 69% yield with this method.

Full text of DOI:10.1016/j.jfluchem.2015.10.002

Journal of Fluorine Chemistry 180 (2015) 208–215
Contents lists available at ScienceDirect
Journal of Fluorine Chemistry
journal homepage: www.elsevier.com/locate/fluor
A convenient and practical method for the selective preparation
of deuterofluorocarbons
*
David P. Richardson , Gordon S. Bauer, Andrew A. Bravo, Michael Drzyzga, Tina Motazedi,
Emma M. Pelegri-O’Day, Shirish Poudyal, Daniel L.M. Suess, John W. Thoman Jr
Williams College, Department of Chemistry, 47 Lab Campus Drive, Williamstown, MA 01267, United States
A R T I C L E I N F O
A B S T R A C T
Article history:
A detailed study of the development of efficient and practical conditions for the selective synthesis of 1-
deuterononafluorobutane from 1-iodononafluorobutane is reported. The optimal conditions involve
treatment of the iodo-precursor in D2O at ꢀ170 8C in the presence of metallic zinc in a sealed Schlenk
tube to give a 59% yield of 1-deutero-1,1,2,2,3,3,4,4,4-nonafluorobutane. The same method was applied
successfully to two higher homologues to produce 1-deutero-1,1,2,2,3,3,4,4,5,5,5-undecafluoropentane
and 1-deutero-1,1,2,2,3,3,4,4,5,5,6,6,6-tridecafluorohexane in yields of 64% and 56%, respectively.
Surprisingly, even the non-perfluorinated product 6-deutero-1,1,1,2,2,3,3,4,4-nonafluorohexane could
be synthesized in 69% yield with this method.
Received 25 August 2015
Received in revised form 28 September 2015
Accepted 4 October 2015
Available online 8 October 2015
Keywords:
Deuteration
Perfluoroalkanes
Deuteroperfluroroalkanes
Iodoperfluoroalkanes
Zinc
ß 2015 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND
license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
Zinc chloride
19F NMR spectroscopy
1. Introduction
with CH3OD, CD3OH, or CD3OD) that demonstrated this conversion
proceeds through either an anion-mediated mechanism (high
As part of a research program directed at the study of
vibrational overtone dynamics in polyfluorinated hydrocarbons
[1], we required a source of highly isotopically enriched mono-
deuterofluorocarbons (DFCs). These compounds are generally not
commercially available and literature reports of their synthesis,
and of their hydroisotopomers, are limited. For example, Hudlicky
et al. reported in 1992, in a study of potentially anesthetic mono-
and di-hydrofluorocarbons (HFCs), that 1,1,1,2,2,3,3,4,4-nona-
fluorobutane 1 (‘‘4H-nFB’’) could be prepared by addition of
4-iodononafluorobutane 2 (‘‘4I-nFB’’) to a warm mixture of zinc
metal, catalytic zinc chloride, and methanol, followed by reflux
(Scheme 1) [2]. This approach was based on seminal studies of the
preparation and reactivity of perflouoroorganozinc iodides by
Miller et al. [3] and by Haszeldine [4,5].
temperature, methanol solution, no radical initiators present;
Scheme 2) or a radical-mediated mechanism (low temperature,
methanol solution, elemental iodine present as radical initiator;
Scheme 3).
We reasoned that simple extension of Hudlicky et al.’s method
for zinc-mediated transformation of iodofluorocarbons to include a
suitable deuterium source would provide access to our desired
mono-deuterofluorocarbons, and might also provide a general
route to selectively deuterated synthetic targets. Selective
deuterium incorporation plays a critical role in a broad range of
experimental applications, from mechanistic studies employing
kinetic isotope effects to spectroscopic work in which deuterium
substitution provides easier spectral analysis and peak assign-
ment. Moreover, recent pharmaceutical research has focused on
the preparation and study of selectively deuterated drugs in order
to increase overall drug efficacy by decreasing toxicity and slowing
excretion rates, and to accomplish patent lifetime stretching [7].
Examples include deuterium-enriched N-pyrrolidinyl arylamides
for treatment of urinary stress incontinence [8] and deuterium-
enriched montelukast for treatment of asthma and allergies [9].
Deuterated organic compounds are also useful in the construction
of light-emitting devices and photodiodes [10]. The research
described herein details our attempts to maximize the efficiency of
In related subsequent work, Howell et al. reported that primary
iodoperfluoroalkanes also can be converted efficiently to the
corresponding primary hydroperfluoroalkanes by treatment with
sodium methoxide in methanol [6]. Howell, et al. conducted
mechanistic studies (including deuterium transfer experiments
*
Corresponding author. Tel.: +1 413 597 3201; fax: +1 413 597 4150.
E-mail address: drichard@williams.edu (D.P. Richardson).
http://dx.doi.org/10.1016/j.jfluchem.2015.10.002
0022-1139/ß 2015 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/
4.0/).

D.P. Richardson et al. / Journal of Fluorine Chemistry 180 (2015) 208–215
209
F
F
F
could be extended to produce 3 using a suitably deuterated form
of methanol as solvent, and to develop a method that could
conveniently provide highly pure product on a relatively modest
scale (1–20 g).
F
F
F
F
F
Zn, ZnCl2
CH3OH
I
H
F
F
F
F F
F
F F F F
1
2
We found that conducting the reaction in a 50 mL Schlenk tube
was the easiest way to decrease its scale. Subsequent transfer to
a second Schlenk tube, using standard vacuum line transfer
techniques, then accomplished product isolation. In a preliminary
trial, we found that simply shaking 2 in a sealed Schlenk tube
containing zinc dust (1.4 eq.), a catalytic amount of anhydrous
ZnCl2 and CH3OD ([2] ꢀ0.8 M) gave an exothermic reaction and
produced a mixture of 3 (1D-nFB) contaminated with unreacted 2,
the hydroisotopomer 1 (4H-nFB), perfluorooctane 6 (‘‘dimer’’), and
traces of methanol, following product isolation and 19F NMR
analysis (Scheme 4, R = CH3) [11]. Because reaction was so
spontaneous and vigorous under these conditions, we took pains
to temporarily isolate the ZnCl2 catalyst by placing it in a small test
tube that was carefully positioned in the bottom of the Schlenk
tube. After being sealed, the Schlenk tube was carefully tipped,
allowing the reaction solvent to enter the test tube. Gentle shaking
then allowed the catalyst to mix with the other reactants leading to
vigorous bubbling and heat release.
Scheme 1. Zinc-mediated synthesis of 4H-nonafluorobutane 1.
RF
I
RF
CH3O
CH3O
I
CH3O
F
F
F
F
4
RF
H
RF
CH3O H
F
F
F
F
Scheme 2. Proposed [6] anion-mediated mechanism for transformation of primary
iodoperfluoroalkanes.
Na
I
CH3O Na
CH3O I
I2
We observed low levels of the dimeric side product
6
throughout our studies of zinc-mediated transformation of 2.
19F NMR and GC–MS data analyses of product mixtures were
consistent with a molecular formula of C8F18 for 6, clearly
suggesting that two nonafluorobutane fragments combined during
its formation. Howell et al.’s work [6] indicated that either anionic
(4, Scheme 2) or radical (5, Scheme 3) intermediates can be
produced during transformation of primary iodoperfluoroalkanes.
Hence, formation of 6 in our experiments could result from SN2
reaction between an anionic intermediate (2i) and a molecule of 2,
or from coupling of a pair of radical intermediates (2r), as shown in
Scheme 5.
CH3O
RF
I
CH3O
I
I
RF
I2
I
F
F
F F
5
RF
H
RF
CH2OH
CH3O
H
F
F
F
F
Our initial study sought to maximize conversion to and yield of
1D-nFB 3 while minimizing contamination with unreacted 2,
isotopomer 1, ‘‘dimer’’ 6, and solvent. Parameters varied included
reaction time, reaction temperature, concentration, and deuterium
source (i.e. CH3OD and CD3OD). Results for these experiments are
summarized in Table 1. In this Table, and all that follow, we include
data that we measured in an effort to determine both the
conversion of starting iodide to the desired deuterated product and
the isolated yield of the product. Due to the extremely volatile
nature of the deuterated products, and the necessarily small
reaction scales involved, however, it was only possible in our
experiments to estimate conversion and yield based on NMR
analysis of products that could be only incompletely purified by a
single stage of low-temperature/low-pressure bulb-to-bulb distil-
lation. Thus, Trials 1–3 explored the use of CH3OD as deuterium
source (Scheme 4, R = CH3). Trial 1 examined the effect of
conducting the reaction at an elevated temperature: the reactants
were mixed briefly at room temperature and then the Schlenk tube
was immediately placed in a heating bath (60 8C). Under these
conditions, crude yield of 3 appeared moderate (43%) and the very
low level (<1%) of unreacted 2 contaminating the isolated product
indicated conversion was high. However, contamination with the
undesired hydroisotopomer 1 (7%) and reaction solvent (ꢀ45%)
were far from ideal. Contamination with solvent was essentially
Scheme 3. Proposed [6] radical-mediated mechanism for transformation of primary
iodoperfluoroalkanes.
the synthesis of singly deuterated fluorocarbons by the simplest,
mildest, and cheapest method possible through the use of D2O as
deuterium source.
2. Results and discussion
2.1. General aims and preliminary results
Our primary goal was to develop a convenient, efficient and
economical method to synthesize 1-deutero-1,1,2,2,3,3,4,4,4-
nonafluorobutane
3
(1-deuterononafluorobutane, ‘‘1D-nFB’’).
Hudlicky et al.’s preparation of the hydroisotopomer of 3 (1,
Scheme 1) from 2 demonstrated that a readily protonated reactive
intermediate was accessible from primary iodoperfluoroalkanes
under very mild conditions (Zn8/catalytic ZnCl2/refluxing metha-
nol). On the downside, isolating the highly volatile 1 (b.p. = 14 8C)
in reasonable yield (72%) using this approach required conducting
the reaction on large scale (100 g), the use of a dry ice trap and
product purification by spinning-band distillation. Hence, our
initial aims were to demonstrate that the Hudlicky et al. approach
F
F
F
F
F
F
F
F
F
F
F
F
F F F F F
F
F
F
Zn, ZnCl2
F
F
I
D
H
+
+
F
F
F
F
ROD
F
F F
F
F
F F
F
F
F F
F
F F F F F F
F
(R=CH3, CD3, D)
3
1
6
2
Scheme 4. Zinc-mediated synthesis of 1D-nonafluorobutane 3.

210
D.P. Richardson et al. / Journal of Fluorine Chemistry 180 (2015) 208–215
F
F
F
anionic
pathway
radical
pathway
F
F
F F
F
F
F
F
F
I
F
F
F
F
F F
F
F F F F
F
F F
F
2i
2
2r
F
F
F
F
F
F
F
F
I
F
F
F
F
F F F F F
F
F
F F
F
F
F F F
F
F
F
F
F F F F F
F
6
Scheme 5. Possible competing dimerization pathways in zinc-mediated synthesis of 1D-nFB 3.
Table 1
Zinc-mediated preparation of 1D-nFB (3) from 4I-nFB (2) in CH3OD or CD3OD.
Trial
Solvent
Scale (g 2)
Reactant ratios
(mmol reactant/
mmol 2)a
T (8C)b
Rxn time (h)
Relative product composition (%)c
Crude yield (%)d
Solvent
Zn8
3
2
1
6
1
2
3
4
CH3OD
CH3OD
CH3OD
CD3OD
2.7
2.7
31.8
31.8
8.0
1.4
1.4
1.4
1.4
60
0.5
45
89
84
99
<1
<1
<1
<1
7
9
2
43
39
30
46
0, 22, 60
0.5 at each T
<1
<1
<1
2.1
22
22
0.5
0.5
16
15.8
6.1
<1
a
All reactions included a catalytic amount of ZnCl2: Trials 1–3: ꢀ10 mg; Trial 4: ꢀ60 mg.
b
c
Temperature of external heating/cooling bath.
Relative composition was determined by integration of appropriate 19F NMR signals in the product mixture following bulb-to-bulb transfer from the reaction flask.
Percentages listed as <1% indicate that a trace of the specific product could be detected but not accurately integrated.
d
Crude (e.g. uncorrected) yield based on total mass of product mixture collected relative to theoretical yield of 3 after bulb-to-bulb transfer from the reaction flask. Thus,
these numbers are only roughly indicative of the actual yield of 3.
eliminated in subsequent trials (Trials 2–4) by passing the
volatilized product through a tube filled with freshly activated
in the contribution of 3 to the overall product composition was due
to removal of solvent during product isolation. Trial 3 was
performed at room temperature but at a significantly higher
reactant concentration ([2] in CH3OD ꢀ3.1 M) than Trials 1 and 2;
the apparent increase in production of 1 under these conditions
was traced to incomplete exclusion of moisture. Ultimately, we
minimized formation of 1 with rigorously dry technique (including
the use of a glove box for loading the Schlenk tube) and by using
deuterated solvents with very high isotopic purity (99 at% D in
trials with CD3OD).
˚
molecular sieves (4 A) during the bulb-to-bulb transfer process
(Fig. 1).
In Trial 2, we investigated the effect of a stepwise increase in
temperature, beginning at 0 8C and ending at 60 8C. This had little
effect on yield and contamination with 1; the significant increase
Trial 4 explored the use of completely deuterated methanol
(Scheme 4, R = CD3, 99.8 at% D), at an even higher concentration
([2] in CD3OD ꢀ4.1 M), and provided the purest sample overall of 3,
although the yield was still modest (46%). While this set of
experiments did not exhaustively optimize zinc-mediated trans-
formation of 2 to 3, the outcome of Trial 4 provided a sample of 3
that was both large and pure enough to make our vibrational
overtone dynamics studies possible.
2.2. Preparation of 1D-nFB in D2O, with and without ZnCl2 catalyst
Of course, the solvents/deuterium sources (CH3OD and CD3OD)
used in our studies above are prepared commercially from D2O,
which is a significantly less expensive potential deuterium donor
($/mole: D2O = 1.00, CH3OD = 1.77, CD3OD = 16.71). These facts led
us to wonder if transformation of 2 to 3 might be accomplished
directly in D2O (Scheme 4, R = D) as effectively as in CH3OD and
CD3OD. Results from our experiments exploring this possibility are
summarized in Table 2. In each of these experiments, held constant
were the amounts of 2 (2.7 g) and Zn8 (1.5 equiv.) while reaction
temperature, reaction time, external energy source, concentration,
zinc metal/ion source, and amount of ZnCl2 catalyst were varied.
Fig. 1. Vacuum apparatus for isolation of reaction product: (A) Schlenk reaction tube
(50 mL); (B) test tube for isolation of ZnCl2 catalyst; (C) solvent trap filled with
˚
˚
molecular sieves (4 A for methanol; 3 A for water); (D) Schlenk collection tube (25
or 50 mL); (E) vacuum valve.

D.P. Richardson et al. / Journal of Fluorine Chemistry 180 (2015) 208–215
211
Table 2
Zinc-mediated preparation of 1D-nFB (3) from 4I-nFB (2) in D2O.
Trial
Reaction conditionsa
Relative product composition (%)h
Crude
yield (%)i
Calc’d
yield (%)j
T (8C)b
22
Rxn time (h)
Other modifications
3
2
1
6
5
1
1
–
58
74
85
95
88
90
94
94
87
81
90
55
70
50
87
40
2
3
4
2
3
5
3
3
4
6
2
<1
<1
2
124
28
82
90
58
18
66
85
74
69
57
75
48
55
45
45
73
59
59
6a
6b
6c
6d
6e
7a
7b
7c
7d
8a
8b
9a
9b
9c
115
115
115
135
135
35
–
23
8
5
–
5
2ꢂ D2O vol.
2ꢂ D2O vol.
2ꢂ D2O vol.; nanoZnc
Sonicationd
Sonication
Sonication, ZnOe
Sonication
Microwave reactorf
<1
9
2
9
1
91
80
24
1
3
2
3
1
64
81
58
73
53
92
121
149
72
35
5
1
1
35
5
7
1
40
13
1
12
7
1
60
1
g
170
110–140
165
170
2
Microwave reactorf
21
27
46
7
24
2
<1
1
15
15
6
19 mol% ZnCl2
48 mol% ZnCl2
No ZnCl2
3
1
4
2
a
Trials 5, 6a–6e, 7a–7d, 8a–8b: reaction conducted with 2.7 g (7.7 mmol) of 2 and 0.755 g (11.46 mmol; 1.5 equiv.) of Zn8 dust in D2O (4.84 mL; 268 mmol; 34.8 equiv.,
except Trials 6c, 6d, and 6e which used 9.68 mL of D2O (536 mmol; 69.6 equiv.)) and included a catalytic amount (ꢀ10 mg, ꢀ0.07 mmol, ꢀ0.1mol%) of ZnCl2. Trials 9a,b:
reaction conducted with 7.7 mmol of 2 and 1.5 equiv. of Zn8 dust in 9.68mL (535.9 mmol, 69.6 equiv.) of D2O in a 50 mL Schlenk tube. Trial 9c: reaction conducted with
4.0 mmol of 2 and 1.5 equiv. of Zn8 dust in 6.26 mL (344 mmol, 86 equiv.) of D2O in a 50 mL Schlenk tube.
b
Temperature of external heating/cooling bath.
c
Nanoparticle Zn (Sigma-Aldrich 578002-5G, particle size < 50 nm).
d
Branson Model 1510 Sonicator; sonicator bath filled with water.
e
Reaction included a catalytic amount (ꢀ10 mg) of ZnO instead of ZnCl2.
f
Biotage ‘‘Initiator’’ Microwave Reactor; reaction conducted in 20 mL microwave vial.
g
Microwave apparatus recorded a pressure of 15 bar during this run.
Relative composition was determined by integration of appropriate 19F NMR signals in the product mixture following bulb-to-bulb transfer from the reaction flask.
h
Percentages listed as <1% indicate that a trace of the specific product could be detected but not accurately integrated.
i
(Crude reaction mass/theoretical mass of 3) ꢂ 100.
j
See Section 4.4 for details.
We discovered in preliminary trials that mixtures of 2, Zn8, and
ZnCl2 in D2O did not react spontaneously at room temperature.
Therefore, in the trials summarized in Table 2, reactants were
simply added to a Schlenk tube in the glove box at room
temperature, after which the sealed tube was subjected to the
indicated conditions. As before, these trials culminated with a
bulb-to-bulb vacuum transfer of the product mixture through a
least 115 8C. In addition, running the reaction at lower concentra-
tion (e.g. [2] in D2O ꢀ 0.8 M, Trials 6c, 6d) appeared to somewhat
improve overall product purity. Under these conditions, crude
yields of 3 settled out at about 90% (or, after correcting the crude
yield for the presence of unreacted 2 and side products 1 and 6, at a
calculated yield of ꢀ74–85%). These data also showed that overall
conversion of 2 to 3 (including Trials from 6b through 7c and 8a)
fluctuated a bit, but is generally quite high (as suggested by the 3:2
ratio observed in the isolated products: from ꢀ10:1 in Trial 6d to a
significantly higher value in Trial 6c). Running the reaction under
these improved conditions but with 1.5 equiv. of Zn8 nanoparticles
in place of Zn8 dust did not appreciably affect the overall outcome
(Trial 6e). We had hoped these conditions might increase the
overall Zn8 particle surface area available to 2 thereby increasing
conversion and yield.
Trials 7a–7d investigated accelerating the reaction with
sonication rather than heat. Ultimately, although sonication at
35 8C for 5 h (Trial 7b) appeared to be somewhat superior to
heating the for the same time period (T = 115 8C, Trial 6b), we
found that sonication gave results that were difficult to reproduce
and understand (e.g. Trial 7d: sonication for a long period resulted
in a lower conversion of 2 to 3). Trial 7c showed that replacing
ZnCl2 with a different source of Zn2+ catalyst (ZnO) also did not
improve the reaction outcome.
˚
tube containing freshly activated molecular sieves (3 A, Fig. 1) into
a second, tared Schlenk tube. The mass of the product isolated at
this stage provided the ‘‘Crude Yield’’ values reported in Table 2.
Subsequent preparation of an NMR sample of product mixtures
and analysis by 19F NMR spectroscopy then provided the relative
product composition data presented which, in turn, provided
the corresponding ‘‘Calculated Yield’’ values. Since the techniques
developed earlier minimized undesired proton sources, we were
not surprised to observe that crude product mixtures contained
very little 1; integration of appropriate 19F NMR signals showed
that 1 typically was present at a level of ꢀ 3% ꢁ 2%. (The
uncertainties inherent in our NMR signal integration protocol suggest
that these estimates actually overstate actual contamination by 1).
Trial 5, in which the reactants were simply shaken periodically
at room temperature for 1 h, provided baseline results for this
entire study. Thus, we were gratified to learn that transformation
of 2 to 3 can be accomplished in D2O ([2] in D2O ꢀ 1.6 M), although
NMR analysis showed that more than a third of the isolated
product was unreacted starting material, indicating that the
desired reaction is fairly slow under these conditions. However, we
were encouraged that the crude product was otherwise relatively
pure, containing only a trace of 1 and an even smaller amount of
the dimer 6. Subsequent experiments aimed to improve these
results.
Trials 8a and 8b illustrate several attempts to activate the
desired transformation using microwave irradiation. Relatively
brief (1 h) treatment in a sealed microwave reaction vessel at
moderate temperature (60 8C, Trial 8a) gave overall results similar
to those achieved by heating for 5 hr at 115 8C (Trial 6b); doubling
irradiation time and increasing temperature (170 8C; Trial 8b) gave
inferior results.
Trials 6a–6d examined the effect of running the reaction at
higher temperatures, for longer periods of time, and at lower
concentration. These results show that the reaction appears to
operate best when conducted for at least 5 h at a temperature of at
Trials 9a–9d summarize experiments that led to an optimized
method for preparation of 3. These experiments involve the effect
of varying the amount of ZnCl2 catalyst present, reaction time and
reaction temperature. Trials 9a and 9b demonstrate that increasing

212
D.P. Richardson et al. / Journal of Fluorine Chemistry 180 (2015) 208–215
the amount of ZnCl2 catalyst surprisingly caused a marked
decrease in the conversion of 2 to 3, even when reaction
temperature is raised significantly. We don’t have an explanation
for this outcome but speculate that high levels of ZnCl2 may coat
the particles of metallic zinc, decreasing metal surface availability
and even significantly decreasing the volume of the reaction
Schlenk tube from 50 mL to 3 mL. We were led to decreasing the
reaction tube volume by reasoning that since 11 is significantly
more volatile than 2, it will be present to a larger extent in the head
space of the reaction tube, decreasing its contact with the zinc-
based reagent and, hence, its conversion to the deuterated product.
In the smaller reaction vessel, containing the usual amounts of Zn8
and D2O and 20 mol% of ZnCl2, heated to 223 8C for 45 h, the best
outcome we could achieve was a product mixture showing 80%
conversion of 11 to 12, but, significantly, with none of the dimeric
side product perfluorohexane observed. Although this result
represented a significant improvement overall, we deemed it
inferior, in overall practicality, to our result for synthesis of 12,
reported in Table 3, using the original method with CD3OD as
deuterium source: 12 produced in 39% isolated yield (contami-
nated with 7% of the dimeric side product). We have not yet
studied deuteration of 1-iodoperfluorethane using D2O, but our
results indicate that this process is difficult with substrates smaller
than 4-iodononafluorobutane.
for reaction with substrate. Trial 9c reports
a completely
serendipitous result: conversion of 2 to 3 proceeded (6 h at
170 8C with 1.5 equiv. of Zn8) even when ZnCl2 was mistakenly
omitted from the reaction mixture! Moreover, this reaction
proceeded in 59% yield and with only 2% contamination with
dimer 6. Ultimately, we found it was most convenient to conduct
reactions under these simplified conditions by simply heating
‘‘overnight’’ (12–18 h) at 170 8C with only Zn8 and D2O present. We
believe the overall modest yield for this process mainly reflects the
challenges inherent in removing the product from the reaction
system and then separating it from solvent.
2.3. Method generality: Preparation of deuterofluorocarbons closely
related to 4D-nFB
The final entry in Table 3 reports a completely unanticipated
result: a nonperfluorinated primary iodosubstrate also smoothly
underwent zinc-mediated deuteration. Thus, substrate 13
(1,1,1,2,2,3,3,4,4-nonafluoro-6-iodohexane), when treated with
Zn8 alone in D2O at 170 8C for 9 h produced the corresponding
deuterated product 14 in 69% yield, with no unreacted starting
material or possible dimeric products detected by 1H or 19F NMR
analysis. This surprising outcome allows us to draw two
conclusions: (a) that fluorine substitution at the carbon atom
undergoing zinc-mediated deuteration is unnecessary to allow C–I
bond scission (surprising if a carbanion intermediate, as suggested
Successful preparation of 3 with D2O led us to attempt our
simplified deuteration method with other suitable substrates;
results are reported in Table 3. For example, deuteration of
primary iodoperfluroroalkanes closely related to 2 proceeded
smoothly in good yield and high levels of purity. Thus, 5-
iodoundecafluoropentane (7) and 6-iodotridecafluorohexane (8),
homologues of 2, gave the corresponding deuterated products 9
and 10 [12] when heated overnight (12 h for 9; 18 h for 10) at
170 8C in D2O with 1.5 equiv. of Zn8. In both of these examples, if
the corresponding dimeric and protonated side products were
produced, levels were so low they could not be detected by NMR
spectroscopy.
above in Scheme 2, is actually involved), and (b) that possible
b-
elimination of H–I from 13 (which would produce alkene 15,
Scheme 6) does not compete with the desired deuteration
process. The lack of elimination side products is particularly
surprising since the deuteration process must produce Zn(OD)2 as
it proceeds. In broader terms, efficient production of 14 suggests
that substrates other than just iodoperfluorocarbons are suscepti-
ble to zinc-mediated deuteration in D2O. We are currently
conducting a more extensive exploration of this general method-
ology. We can make a preliminary report here, however, in support
of our conclusion (a), above, that under our general reaction
Unfortunately, our deuteration method with D2O only and no
ZnCl2 is less successful with 1-iodoheptafluoropropane, 11, the
homologous substrate one CF2 unit smaller than 2: our standard
reaction conditions (170 8C, 18 h in
a 50 mL Schlenk tube)
produced a mixture of the desired product 12 [13] contaminated
with unreacted starting material and the ‘‘dimeric’’ side product,
tetradecafluorohexane, in a 1.0:2.6:0.53 ratio. Extensive experi-
mentation aimed at improving this outcome explored increasing
the reaction temperature, returning to the use of catalytic ZnCl2,
Table 3
Zinc-mediated preparation of deuterofluorocarbons closely related to 4D-nFB.
ꢄ
Zn ;ꢁZnCl
2
RF ꢃ I
ꢃ!
RF ꢃ D
XOD X¼CD ;D
ð
Þ
3
Substrate, RF=
Deuteron
source
Reaction
time (h)
Deuteration product
(calculated yield, %)b
Comments
7, CF3CF2CF2CF2CF2
8, CF3CF2CF2CF2CF2CF2
11, CF3CF2CF2
D2Oa
12
18
0.5
9
9 (64)
RF–H, RF–RF contaminants below NMR detection limits
RF–H, RF–RF contaminants below NMR detection limits
Product contained 7% of RF–RF dimer
D2Oa
10 (56)
12 (39)
14 (69)
CD3ODc
D2Oa
13, CF3CF2CF2CF2CH2CH2
RF–H, RF–RF contaminants below NMR detection limits
a
Standard reaction conditions: used 2.5–6 mmol (1 equiv.) of RF–I and Zn8 dust (1.5 equiv.) in D2O (69 equiv.; [RF–I] ꢀ 0.8 M) overnight (9–18 h) at 170 8C in a 50 mL
Schlenk tube.
b
Yield was calculated (see Table 2) by integration of appropriate NMR signals (19F, and 1H for 14) in the isolated product mixture following bulb-to-bulb transfer from the
reaction flask.
c
Reaction conducted using 11 (13.8 g, 46.7 mmol), Zn8 dust (4.58 g, 70.1 mmol, 1.5 equiv.) and catalytic ZnCl2 (ꢀ60 mg) in CD3OD (11.26 mL, 277mmol, 5.93 equiv.) at room
temperature (0.5 h) in a 50 mL Schlenk tube.
H
F
F
F
F F
F
F H H
F
F H H
F
F
Znο
F
H
D
I
F
F
F
D2O
F
F F F
F
F F F H H
F
F F F H H
H
15
14
13
Scheme 6. Zinc-mediated synthesis of 6D-1,1,1,2,2,3,3,4,4-nonafluorohexane 14.

D.P. Richardson et al. / Journal of Fluorine Chemistry 180 (2015) 208–215
213
conditions 1-iodobutane gives 1-deuterobutane in greater than
80% yield. We will report full details in a future publication.
solvents/deuterium sources, was a hand-fashioned drying tube
(forged by joining two 2 ‘‘long, 9 mm diameter Pyrex tubes to a 5’’
long, 14 mm Pyrex tube; see Fig. 1) filled with lightly pulverized,
˚
3. Conclusions
activated molecular sieves (3 A sieves were used to remove D2O;
˚
4 A sieves were used to remove CH3OD and CD3OD). Cooling of
Our studies provide a practical route for the preparation of
monodeuterofluorocarbons from the corresponding monoiodo-
fluorocarbons using zinc dust and a suitable solvent as deuterium
source in sealed Schlenk tubes. Reaction with CD3OD and Zn8 (with
or without ZnCl2 catalyst) with primary iodo-perfluorocarbons
provides the corresponding 1-deuteroperfluorocarbons in moder-
ate yield and good purity. With 4-iodononafluorobutane (2), 5-
iodoundecafluoropentane (7) and 6-iodotridecafluorohexane (8),
the reaction can be simplified by elimination of the ZnCl2 catalyst
and replacing CD3OD with (significantly less expensive) D2O. These
conditions require longer reaction times and higher temperatures
but produce the corresponding 1-deuterononafluorobutane (3),
1-deuteroundecafluoropentane (9) and 1-deuterotridecafluorohex-
ane (10) products in excellent yield and purity. Surprisingly, these
conditions work smoothly and efficiently with 1,1,1,2,2,3,3,4,4-
nonafluoro-6-iodohexane (13) to produce 14, indicating that the
deuteration method can tolerate iodofluorocarbon substrates
Schlenk tubes during the bulb-to-bulb transfer operations, and
freeze/pump/thaw degassing steps, was provided by immersion in
liquid-nitrogen-filled Dewar flasks.
4.2. Exclusion of proton sources
In order to minimize the inclusion of proton (1H) sources during
the synthesis of deuterofluorocarbons, all necessary steps were
taken to exclude water. All glassware, syringe needles, and tools
were washed with water, rinsed with methanol, and then oven-
dried at 160 8C (Fisher Isotemp 230G oven) and allowed to cool to
room temperature in a desiccator before use. Used reaction vessels
containing zinc were washed with cycles of concentrated HCl (aq),
distilled water, 1 M NaOH, distilled water, and methanol, followed
by oven drying. Schlenk tube valves and syringe needles were
cleaned by rinsing with methanol and wiping dry with Kimwipes
followed by storage overnight in a desiccator before reuse.
lacking fluorine atoms at the
a- and b-positions.
4.3. NMR analysis
4. Experimental
1H and 19F NMR spectra were measured using a Bru¨ker Avance
DRX 500 MHz NMR spectrometer operating at frequencies of
500.150 (1H) and 470.611 (19F) MHz using the proton coil (which
was separately retuned to observe 19F frequencies) in a standard
5 mm broadband multinuclear (PABBO) probehead (908 pulse
4.1. General information
All reactions were performed in 50 mL Schlenk tubes (Chem-
glass Life Sciences, AF-0096) that were charged with reagents and a
small magnetic stir bar at room temperature under an atmosphere
of dry nitrogen in a glove box (Vacuum Atmospherics Co., HE-493)
before the Schlenk tube valve was closed and they were removed
from the glove box. Liquids were added to the Schlenk tube using
new disposable plastic syringes (NormJect) of the appropriate size
and 6-inch stainless steel needles (Sigma-Aldrich, Z219363).
Reactions were conducted in a fume hood behind an explosion
shield. Reactions conducted above room temperature were heated
at the desired temperature by inserting the Schlenk tube in a sand-
filled heating mantle; gentle agitation was provided by spinning
the magnetic stir bar.
D2O was obtained from Cambridge Isotope Laboratories
(99.9 atD %, DLM-4-100). 4-iodononafluorobutane (98%, 1100-J-
06), 5-iodoundecafluoropentane (97%, 1100-J-51), 6-iodotrideca-
fluorohexane (98%, 1100-J-12), and 1-iodoheptafluoropropane
(95%, 1100-J-07) were obtained from SynQuest Labs, Inc.
1,1,1,2,2,3,3,4,4-nonafluoro-6-iodohexane (95%, 07387-5G-F) and
CH3OD (99 atD %, 15,193-9) were obtained from Sigma-Aldrich.
CD3OD was obtained from Cambridge Isotope Laboratories
(99.8 atD %, DLM-24-S-10). These reagents were stored in the glove
box; iodine-containing precursors were wrapped in foil to minimize
widths: 1H, 11.5
m m
s; 19F, 10.0 s). 1H NMR spectra were measured
at 298 K while 19F NMR spectra were measured at 273 K using the
NMR spectrometer’s BVT-2000 variable temperature unit. Mea-
suring 19F spectra at the lower temperature reduced 19F signal
complexity, which we tentatively ascribed to a decrease in
conformational isomer distribution. 19F NMR spectra were not
proton decoupled. 1H chemical shifts (ppm) were measured
relative to internal TMS (
relative to internal CFCl3
d
(
= 0 ppm); 19F shifts were measured
d
= 0 ppm). Spectral windows for 19F
acquisitions were set at 200 ppm; the recycle delay (D1) was set at
5 s. Sample spectra for compounds 3, 9, 10, 12, and 14 are
presented in the Supplementary Data section.
In order to assess reaction product purity, and to provide a
calculated reaction yield, a small sample of each reaction product
was transferred, using the vacuum transfer line, to a 5 mm NMR
tube equipped with a J. Young valve (Wilmad 507-JY-7). Samples
were prepared by pre-filling an NMR tube with ꢀ700
premade mixture of CDCl3 and CFCl3 (ꢀ0.05% (v/v)) using a
1000 L gas tight syringe. The J. Young valve was then closed and
mL of a
m
the NMR tube was attached to the vacuum manifold and immersed
in the liquid nitrogen bath and then evacuated to 17–20 mTorr. A
Schlenk collection tube containing a reaction product was then
attached to the vacuum line and immersed in liquid nitrogen. Once
the sample was cold, the Schlenk collection tube valve was opened
and the liquid nitrogen Dewar flask under the tube was removed.
Transfer of the reaction product into the NMR tube was then
allowed to proceed until it was evident that the column of liquid in
the NMR tube had increased in depth by a few millimeters. The J.
Young valve and the Schlenk collection tube valves were then
closed and both tubes were allowed to warm to room temperature.
exposure to light. Zinc dust (98%, <10
mm, 209988-100G), anhy-
drouszincchloride(99.99%,429430-5G)andnanopowderzinc(99%,
578002-5G) were obtained from Sigma-Aldrich. CDCl3 (99.8%D with
0.05% TMS, DLM-7-t-b) was obtained from Cambridge Isotope
Laboratoriesandwasstoredina desiccator;CFCl3 (99%, 254991)was
obtained from Sigma-Aldrich and was stored in a refrigerator.
˚
Molecular sieves were obtained from Sigma-Aldrich (3 A, 208574;
˚
4 A, 208590) and were activated by heating in a vacuum drying oven
(National Appliance Co., Model 5831) overnight at 160 8C followed
by cooling to room temperature in a desiccator.
Reaction products (and any un-reacted starting materials) were
separated from zinc, zinc chloride and reaction solvents/deuterium
sources by bulb-to-bulb distillation into a dry, tared 25 mL Schlenk
tube employing a vacuum transfer line. Interposed between the
reaction and collection Schlenk tubes, and used to remove reaction
4.4. Assessment of product purity and reaction yield calculation by
NMR analysis
Relative product composition (Tables 1–3) and calculated yield
(Table 2) for all deuteration reactions were determined by

214
D.P. Richardson et al. / Journal of Fluorine Chemistry 180 (2015) 208–215
measurement of relative integral areas of 19F NMR signals for the
deutero-product, hydro-product, unreacted starting material, and
‘‘dimer’’ following assignment of unique and well-resolved
characteristic signals for each material. Relative integral areas
(which are equivalent to the mole fraction of each material in the
isolated product mixture) led directly to the relative product
composition data reported. Calculated yield was determined by
comparingtheactualmoles ofdeutero-productproduced,measured
by 19F NMR signalintegration, withthetheoreticalmoles of deutero-
product calculated from the mass of starting iodofluorocarbon used
in each reaction. The total moles of all materials in the isolated
product could be calculated by dividing the total isolated product
mass by the sum of the mole fraction for each material times their
respective molecular weights. Actual moles of deutero-product
could be calculated by dividing the numerical difference of the total
isolated product mass and the measured mass of the deutero-
product by the molecular weight of the deutero-product. Finally,
the measured mass of the deutero-product was equal to the total
moles of all materials in the isolated product times the sum of the
mole fractions for the hydro-product, unreacted starting material
and dimer times their respective molecular weights.
While the reaction mixture was rapidly agitated with a magnetic
stirbar, thereaction tubevalve wasopened slowlyuntilthereaction
mixture began visibly bubbling. During this process, the denser and
more volatile deuterofluorocarbon product could be seen rising
through the D2O solvent. Care was taken to avoid ‘‘bumping’’ of the
solvent:if D2O rosenearthelevelof theneckoftheflask, theproduct
transfer rate was slowed by slightly closing the reaction Schlenk
tube valve. Product transfer progress was evidenced both by a
decrease in the pressure observed in the transfer line and by the
physical descent of the light gray, oxidized zinc solid in the reaction
tube. (Variation in the transfer efficiency of the variety of
fluorinated materials involved in this study made it difficult to
standardize the process for bulb-to-bulb transfer. In some cases, the
reaction tube was connected to a second, clean collection flask and
theprocess was repeated if the yield fromthe first transfer appeared
abnormally low.) To complete the transfer, the collection tube valve
was closed. At this point, if the vacuum line pressure had not
dropped below 1 Torr, or if there had been bumping during the
transfer, the reaction flask was placed in a liquid-nitrogen-filled
Dewar flask in order to condense and retain any volatile compounds
remaining in the line. The transfer line was only opened to
the vacuum pump once both Schlenk flasks had been closed. The
collection flask was then allowed to warm to room temperature.
The product was a homogenous colorless liquid when the
collection tube was immersed in liquid nitrogen, but was invisible
gas at room temperature. The mass of the collection tube was then
measured; the isolated product weighed 0.631 g (59% calculated
yield; Table 2). By 19F NMR analysis, this sample was composed of a
4.5. 1-Deutero-1,1,2,2,3,3,4,4,4-nonafluorobutane, 3
The preparation of 1-D-nonafluorobutane 3 (1D-nFB) serves as
a
general guide for the synthesis of deuterofluorocarbons.
Preparation of compounds reported in Sections 4.6–4.8 below
employed these methods.
Inside a glove box, reactants were added to a 50 mL Schlenk
tube (containing a 10 ꢂ 5 mm Teflon-coated magnetic stir bar)
under an atmosphere of nitrogen. Zinc dust (0.497 g, 7.6 mmol,
1.5 equiv.) was added to the Schlenk tube using a spatula, taking
care to keep the neck of the Schlenk tube free of zinc dust, which
could impair sealing the tube. Deuterium oxide (D2O, 6.26 mL,
mixture of 3 (87%), 2 (7%), 1 (4%), and 6 (2%). 3: 19F NMR
d
ꢃ81.5 (t,
3F, F-4); ꢃ128.3 (m, 2F, F-3); ꢃ130.9 (m, 2F, F-2); 138.4 (m, 2F, F-1).
4.6. 1-Deutero-1,1,2,2,3,3,4,4,5,5,5-undecafluoropentane, 9
A
50-mL Schlenk flask was charged with 2.42 g of 5-
345 mmol, 86 equiv.) and 4-iodononafluorobutane
2
(4I-nFB,
iodoundecafluoropentane 7 (1.15 mL, 6.12 mmol), 7.5 mL of D2O
(421 mmol, 69 equiv.), 0.596 g of Zn dust (9.16 mmol, 1.5 equiv.),
and a magnetic spin bar under an atmosphere of nitrogen inside a
glove box. The Schlenk tube was sealed, removed from the glove
box and then heated in a sand bath at 170 8C for 12 h. The tube
was then removed from the sand bath, allowed to cool to room
temperature, and the crude product mixture was transferred, using
a vacuum transfer line, into a tared 25-mL Schenk tube. The
0.65 mL, 4.0 mmol, 1.0 equiv.) were added by gas-tight syringe.
The Schlenk tube was sealed, removed from the glove box and then
shaken vigorously for a few minutes to mix the contents.
The Schlenk tube was then heated at 170 8C in a sand bath for
6 h. Vigorous bubbling was observed to commence in the tube
shortly after the heating period began. The tube was removed from
the sand bath, allowed to cool to room temperature, and then was
attached to a vacuum transfer line. To isolate the desired product
from the reaction vessel, a bulb-to-bulb transfer was performed,
beginning with the removal of nitrogen gas (from the glove box
operations) through a series of freeze-pump-thaw cycles using
liquid nitrogen. While removing nitrogen gas, this step retained
D2O and all reactants and reaction products as frozen solids. This
degassing was repeated (usually 3–7 cycles) until the vacuum
manifold pressure remained below 20 mTorr when the Schlenk
tube was opened to the frozen solids.
product (1.038 g, 64% yield) was a colorless oil. 19F NMR
d
ꢃ81.1
(t, 3F, F-5); ꢃ124.6 (m, 2F, F-4); ꢃ126.6 (m, 2F, F-3); ꢃ129.8 (m, 2F,
F-2); ꢃ137.9 (m, 2F, F-1).
4.7. 1-Deutero-1,1,2,2,3,3,4,4,5,5,6,6,6-tridecafluorohexane, 10
A
50-mL Schlenk flask was charged with 2.23 g of 6-
iodotridecafluorohexane 8 (1.08 mL, 5.0 mmol), 6.26 mL of D2O
(345 mmol, 69 equiv.), 0.497 g of Zn dust (7.60 mmol, 1.5 equiv.),
and a magnetic spin bar under an atmosphere of nitrogen inside a
glove box. The Schlenk tube was sealed, removed from the glove
box and then heated in a sand bath at 170 8C for 18 h. The tube was
then removed from the sand bath, allowed to cool to room
temperature, and the crude product mixture was transferred, using
a vacuum transfer line, into a tared 25-mL Schenk tube and
weighed. The product (0.900 g, 56% yield) was a colorless oil. 19F
After this initial de-gassing process was complete, a collection
apparatus, including a clean, empty Schlenk tube and a drying tube
˚
(containing oven-dried 3 A molecular sieves, see above, and Fig. 1)
was attached to the vacuum system in order to isolate the product
from the crude reaction mixture. The product isolation process
began by flushing the collection tube and drying tube with
nitrogen to displace water, and then evacuating. The collection
tube valve was closed and the collection tube was removed from
the vacuum line and tared. The collection tube was then returned
to the vacuum line, its valve was opened, and then reaction tube
was cooled in a water ice bath to prevent excessive vaporization of
D2O during the bulb-to-bulb transfer. The evacuated collection
tube was then placed in a liquid-nitrogen-filled Dewar flask and
main vacuum valve E (Fig. 1) was closed in order to isolate the
transfer line system from the vacuum pump.
NMR
d
ꢃ81.2 (t, 3F, F-6); ꢃ123.6 (m, 2F, F-5); ꢃ124.2 (m, 2F, F-4);
ꢃ126.9 (m, 2F, F-3); ꢃ130.1 (m, 2F, F-2); ꢃ138. 3 (m, 2F, F-1).
4.8. 6-Deutero-1,1,1,2,2,3,3,4,4-nonafluorohexane, 14
A
50-mL Schlenk flask was charged with 0.93 g of
1,1,1,2,2,3,3,4,4-nonafluoro-6-iodohexane 13 (0.48 mL, 2.5 mmol),
3.13 mL of D2O (173 mmol, 69 equiv.), 0.244 g of Zn dust

D.P. Richardson et al. / Journal of Fluorine Chemistry 180 (2015) 208–215
215
(3.75 mmol, 1.5 equiv.), and
a magnetic spin bar under an
References
atmosphere of nitrogen inside a glove box. The Schlenk tube
was sealed, removed from the glove box and then heated in a sand
bath at 170 8C for 9 h. The tube was then removed from the sand
bath, allowed to cool to room temperature, and the crude product
mixture was transferred, using a vacuum transfer line, into a tared
25-mL Schenk tube and weighed. The product (0.432 g, 69% yield)
[1] B.G. Saar, G.P. O’Donoghue, A.H. Steeves, J.W. Thoman Jr., Chem. Phys. Lett.
417 (2006) 159–163.
[2] T. Hudlicky, R. Fan, J.W. Reed, D.R. Carver, M. Hudlicky, E.I. Eger II, J. Fluorine
Chem. 59 (1992) 9–14.
[3] W.T. Miller Jr., E. Bergman, A.H. Fainberg, J. Am. Chem. Soc. 79 (1957)
4159–4164.
[4] R.N. Haszeldine, Nature 168 (1951) 1028–1031.
[5] R.N. Haszeldine, E.G. Walaschewski, J. Chem. Soc. (1953) 3607–3610.
[6] J.L. Howell, B.J. Muzzi, N.L. Rider, E.M. Aly, M.K. Abouelmagd, J. Fluorine Chem.
72 (1995) 61–68.
was a colorless oil. 1H NMR
19F NMR
d 1.13 (br. t, 2H, H-6); 2.10 (m, 2H, H-5).
d
ꢃ81.5 (t, 3F, F-1); ꢃ117.4 (m, 2F, F-2); ꢃ125.2 (m, 2F, F-
3); ꢃ126.8 (m, 2F, F-4).
[7] A.T. Yarnell, Sci. Technol. 87 (2009) 36–39.
[8] A.W. Czarnik Protia, LLC, USA (2009) U.S. Pat, Appl. Publ. (2009) US
2009062373 A1 20090305.
[9] A.W. Czarnik Protia, LLC, USA (2008) U.S. Pat, Appl. Publ. (2008) US
2008319007 A1 20081225.
Acknowledgments
This work was supported by the President and Trustees of
Williams College. We thank Deborah Morandi for assistance with
preparation and submission of the manuscript.
[10] K.C. Hwang, C.C. Tong, T.H. Yeh U.S. Pat, Appl. Publ. (2009) US 2009295274 A1
20091203.
[11] (a) For references detailing 1H and 19F NMR data for 1 and 3 H. Blancou, P.
Moreau, A. Commeyras, Tetrahedron, 33 (1977) 2061–2067;
(b) A. Foris, Magn. Reson. Chem. 42 (2004) 534–555;
(c) J. Gil-Rubio, J. Vicente, J. Guerrero-Leal, M. Blaya, D. Bautista, P.G. Jones,
Organometallics 31 (2012) 1287–1299.
Appendix A. Supplementary data
ˇ
´
´ˇ
´
ˇ
´ˇ
[12] J. Paterova´, M. Skalicky´, M. Ryba´ckova, M. Kvıcalova, J. Cvacka, J. Kvıcala,
Supplementary material related to this article can be found, in the
online version, at http://dx.doi.org/10.1016/j.jfluchem.2015.10.002.
J. Fluorine Chem. 131 (2010) 1338–1343.
[13] G. Giacometti, E.W.R. Steacie, Can. J. Chem. 36 (1958) 1483–1500.