Selective and facile C-F bond activation of trifiuoromethyl groups on Cu(111) [7]

Full text of DOI:10.1021/ja991596x

8116
J. Am. Chem. Soc. 1999, 121, 8116-8117
Selective and Facile C-F Bond Activation of
Trifluoromethyl Groups on Cu(111)
Ya-Ching Hou and Chao-Ming Chiang*
Department of Chemistry
National Sun Yat-Sen UniVersity
Kaohsiung, Taiwan 80424
ReceiVed May 13, 1999
Activation of the strongest bond that carbon can form, the C-F
bond, remains a topic of recent interest.1 It has long been known
that interaction of fluorocarbons with a metal center may
ultimately lead to the cleavage of such robust bonds. Catalytic
activation of C-F bonds by metal complexes in solution is the
subject of several reviews.2,3 Trifluoromethyl transition-metal
complexes draw our particular attention because M-CF3 species
have been recognized as precursors to MdCF2 (difluorocarbene)
complexes.4 We are interested in such problems regarding
controlled C-F activation of small fluorocarbons bound to metal
surfaces. The chemistry of trifluoromethyl groups adsorbed on
metal single-crystal surfaces was explored by the reactions of
trifluoromethyl iodide (CF3I) with Ni,5-7 Ru,8,9 Pt10,11, and Ag.12
Except for silver (CF3 desorbs as a radical), experimental results
revealed that the C-F bonds could be ruptured, however, in either
uncontrolled fashion or at high temperatures. Removal of a single
fluorine atom from the CF3 fragment could not be achieved with
high selectivity, as displayed for metal complexes.
We speculate that the choice of metal surfaces might be pivotal.
Here we report a novel C-F activation in chemisorbed CF3
moieties using a Cu(111) surface. The key findings of this work
are 2-fold: First, only one C-F bond in the trifluoromethyl is
broken, suggesting selectiVe actiVation of carbon-fluorine bonds
rather than complete defluorination, and the formation of a
difluorocarbene intermediate on the surface. Without other
competing reaction channels, all the CF3 groups undergo this mild
C-F decomposition pathway at and below the monolayer
coverage. Second, C-F bonds are cleaved by the copper surface
below 160 K. This low reaction temperature can be translated, in
kinetics terms, into small activation energy; therefore, remarkably
facile bond activation is implied.
Figure 1. Multiple-ion TPR/D spectra of m/e 31 (CF+), 50 (CF2+), 69
(CF3+), 81 (C2F3+), 100 (C2F4+), 119 (C2F5+), and 196 (CF3I+) after the
adsorption of 1.25 L of CF3I on Cu(111) at 110 K. Ions were monitored
with the mass spectrometer ionizer energy set at 70 eV. The heating rate
was 2 deg K/s. It should be noted that the current levels on the Cu(111)
were less than 1 nA with our experimental configurations; therefore, the
desorption features were surely not stimulated by electrons emitted from
the ion source of the mass spectrometer. The inset shows the integrated
peak areas, calculated from m/e 81 (tetrafluoroethylene formation) and
196 (molecular desorption) TPR/D features, as a function of CF3I
exposure.
of 2 mm diameter to conduct temperature-programmed reaction/
desorption (TPR/D) studies. CF3I has been widely used as a source
of adsorbed CF3 moieties on a variety of metal surfaces5-12
because the C-I bond is weak, and is known to break readily
upon adsorption, yielding the desired CF3 species. As illustrated
in Figure 1, multiple-ion TPR/D survey shows a featureless trace
of m/e 196 following the adsorption of 1.25 L (1 L ) 10-6 Torr
s) of CF3I at 110 K. This observation indicates no molecular
desorption occurs from the surface at this exposure, suggesting
that all the adsorbed CF3I undergo C-I bond scission. Because
The experiments were performed in an ultrahigh vacuum
chamber described in detail elsewhere.13 In brief, the system is
equipped with an ion sputtering gun and a retarding-field analyzer
for both Auger electron spectroscopy and low-energy electron
diffraction. A quadrupole mass spectrometer is shielded in a
differentially pumped cylindrical cage assembly with an aperture
+
+
the ion signals of CF3 (m/e 69) and C2F5 (m/e 119) are also
absent, the possibility of forming the CF3 free radical, perfluo-
romethane, and perfluoroethane can be ruled out. The only
desorption state is featured by m/e 31 (CF+), 50 (CF2+), 81
+
+
(C2F3 ), and 100 (C2F4 ) with peak maxima at 250 K. The relative
ion abundance of these desorption profiles nearly duplicates the
fragmentation pattern of tetrafluoroethylene (CF2dCF2).14 Intu-
itively, coupling of two difluorocarbene (CF2) units would afford
tetrafluoroethylene. The intermediacy of a surface-bound difluo-
rocarbene, CF2(ad), can be invoked to account for the observed
* To whom correspondence should be addressed. Phone: 886-7-525-3939.
Fax: 886-7-525-3908. E-mail: cmc@mail.nsysu.edu.tw.
(1) For leading recent references, see: Hughs, R. P.; Linder, D. C.;
Rheingold, A. L.; Liable-Sands L. M. J. Am. Chem. Soc. 1996, 118, 1805.
Burdenic, J.; Jedlicka, B.; Crabtree, R. H. Chem. Ber. (Recl.) 1997, 130, 145.
(2) Kulawiec, R. J.; Crabtree, R. H. Coord. Chem. ReV. 1990, 99, 89.
(3) Kiplinger, J. L.; Richmond, T. G.; Osterberg, C. E. Chem. ReV. 1994,
94, 373.
(4) Brothers, P. J.; Roper, W. R. Chem. ReV. 1988, 88, 1293.
(5) Jones, R. G.; Singh, N. K. Vacuum 1988, 38, 213.
(6) Jensen, M. B.; Thiel, P. A. J. Am. Chem. Soc. 1995, 117, 438.
(7) Myli, K. B.; Grassian, V. H. J. Phys. Chem. 1995, 99, 1498; 1995, 99,
5581.
(8) Dyer, J. S.; Thiel, P. A. Surf. Sci. 1990, 238, 169.
(9) Jensen, M. B.; Myler, U.; Jenks, C. J.; Thiel, P. A.; Pylant, E. D.; White,
J. M. J. Phys. Chem. 1995, 99, 8736.
(10) Liu, Z.-M.; Zhou, X.-L.; Kiss, J.; White, J. M. Surf. Sci. 1993, 286,
233.
(11) Armentrout, D. D.; Grassian, V. H. Langmuir 1994, 10, 2071.
(12) Castro, M. E.; Pressley, L. A.; Kiss, J.; Pylant, E. D.; Jo, S. K.; Zhou,
X.-L.; White, J. M. J. Phys. Chem. 1993, 97, 8476.
(13) Wu, H.-J.; Hsu, H.-K.; Chiang, C.-M. J. Am. Chem. Soc. 1999, 121,
4433.
•
product. The lack of CF3, CF4 ,and C2F6 suggests complete
conversion from CF3(ad) to CF2(ad). CF2(ad) restricts itself from being
further decomposed into CF(ad) and carbidic carbon because the
Auger data reveal the disappearance of carbon signal above 250
K, which coincides with the desorption of CF2dCF2. The fate of
fluorine atoms on the surface, resulting from the selective C-F
bond activation CF3(ad) f CF2(ad) + F(ad), eventually leads to
etching of the substrate by evolving CuF in the gas phase above
700 K (TPR/D data are not shown here). Surface-bound iodine
due to the initial C-I bond dissociation ultimately desorbs as
iodine atoms in a broad temperature range from 600 to 900 K.
(14) McLafferty F. W.; Stauffer, D. B. The Wiley/NBS Registry of Mass
Spectral Data; Wiley: New York, 1989.
10.1021/ja991596x CCC: $18.00 © 1999 American Chemical Society
Published on Web 08/19/1999

Communications to the Editor
J. Am. Chem. Soc., Vol. 121, No. 35, 1999 8117
ensures the presence of surface difluorocarbene, which impera-
tively demands that C-F bond cleavage proceed in a selective
fashion. Furthermore, the 163 K liberation of CF2dCD2 reflects
that the C-F bond in CF3 has already been activated prior to or
at this temperature. That CF2dCF2 formed at 250 K is thus rate
limited by the CF2 coupling step instead of C-F activation. It is
apparent that the addition of scavengers helps to place the kinetic
upper limit for the C-F activation more accurately by 90 K (163
vs 250 K). In summary, we propose the following mechanism to
delineate the aforementioned events and relevant kinetic informa-
tion:
Figure 2. TPR/D spectra of three different types of ethylene due to
competing carbene coupling reactions after exposing the Cu(111) surface,
which was preadsorbed with 0.15 L of CF3I, to an X amount of CD2I2 (X
) 0, 0.5, 1, 2, 3 L exposures). The representative ions are m/e 32 (CD2d
CD2+, left), 66 (CF2dCD2+, middle), and 100 (CF2dCF2+, right),
respectively. It should be noted that the tiny feature at 158 K in all the
CF2dCF2 (m/e 100) spectra resulted from traces of tetrafluoroethylene
impurity in the CF3I gas manifold.
The role of Cu(111), the substrate used in this study, merits
comment. The catalytic activity of atomically clean copper
surfaces under ultrahigh vacuum (UHV) conditions can be
classified into two categories. The first encompasses dehydroge-
nation reactions, such as R- and â-elimination in alkyl groups.16
Methyl (CH3) is stable on the copper surface until approximately
450 K, where a number of competing reactions occur, including
R-hydrogen elimination to yield a methylene (CH2) intermediate.
Therefore, a similar reactivity toward CF3 is somewhat expected;
however, the extremely low temperature for the C-F activation
and the total domination of this reaction route are unusual. The
second category is associated with the facile C-C coupling
reactions to evolve alkanes17,18 and alkenes.17,19 This reactivity
becomes a mechanistic probe that allows us to approach the
kinetic limit for the C-F bond activation as exact as possible.
Activation of saturated C-F bonds has been investigated by
trifluoromethyl groups chemisorbed on Cu(111). To our knowl-
edge, this is the first example capable of readily and selectively
cleaving a single C-F bond to transform CF3 into CF2 moieties
on a well-defined metallic surface under UHV ambiences. In
perspective, this discovery provides a convenient method for
generating difluorocarbene intermediates. A variety of surface
reactions involving carbenes, such as abstraction, addition, and
insertion, can be studied and compared with its hydrocarbon
counterpart, methylene, to give insight into the effects of
fluorination on reactivity.
The reaction scheme described above holds for all the CF3I
exposures below 1.25 L. Due to active-site availability, intact
molecular desorption starts to compete with tetrafluroethylene
formation in the TPR/D spectra at higher doses. As shown in the
+
inset of Figure 1, according to the uptake curves of C2F3 (m/e
81) and CF3I+ (m/e 196), it is reasonable to assign 1.25 L to the
saturation coverage of one monolayer (ML) of CF3I.
The intermediacy of CF2(ad) deserves further scrutiny. We can
verify this intermediate state based upon purely chemical evidence.
Our approach, borrowed from the concept of “carbene scaven-
gers”, utilizes a labeled substance that is added to the surface to
counteract the self-coupling of CF2(ad) (if any) and meanwhile to
remove difluorocarbene as an easily identifiable species. Here
this substance is chosen to be deuterium-labeled diiodomethane
(CD2I2), a well-known precursor for generating surface methylene.
The chemistry of CD2I2 on Cu surfaces studied by Chiang et al.15
reveals facile dissociation of C-I bonds to form CD2(ad) followed
by methylene recombination, yielding CD2dCD2 as the sole gas-
phase product. Figure 2 juxtaposes the TPR/D spectra of three
different types of ethylene, CF2dCF2 (m/e 100, right panel), CF2d
CD2 (m/e 66, middle panel), and CD2dCD2 (m/e 32, left panel),
in a series of experiments by adding various amounts of CD2I2
(scavengers) on Cu(111) which was preadsorbed by 0.15 L of
CF3I (∼1/10 ML). These data demonstrate that the formation of
CF2dCF2 (250 K), originating from CF3(ad), is being suppressed
Acknowledgment. This research was supported by the National
Science Council of the Republic of China under Contract No.
88-2113-M-110-013.
and finally blocked entirely as a function of increasing CD2(ad)
.
JA991596X
Concomitantly, yields of CD2dCD2 from recombination of CD2(ad)
at 160 K are found to increase. The most informative results
consist of the emerging CF2dCD2, which is responsible for the
descending yields of CF2dCF2. This cross-coupling product
(16) Jenks, C. J.; Chiang, C.-M.; Bent, B. E. J. Am. Chem. Soc. 1991, 113,
3, 6308.
(17) Chiang, C.-M.; Wentzlaff, T. H.; Bent, B. E. J. Phys. Chem. 1992,
96, 1836.
(18) Paul, A.; Bent, B. E. J. Catal. 1994, 147, 264.
(19) Yang, M. X.; Eng, J., Jr.; Kash, P. W.; Flynn, G. W.; Bent, B. E.;
Holbrook, M. T.; Bare, S. R.; Gland, J. L.; Fischer, D. A. J. Phys. Chem.
1996, 100, 12431.
(15) Chiang, C.-M.; Wentzlaff, T. H.; Jenks, C. J.; Bent, B. E. J. Vac. Sci.
Technol. 1992, A10, 2185.