вход по аккаунту


Evidence for [PdSiO] from IR Spectroscopy.

код для вставкиСкачать
ry ion source constantly releases xenon as a background gas
into the vacuum up to a pressure of ca. lo-’ mbar, which
leads to a great number of collisions of the ions excited to a
larger cyclotron orbit and thus to their dissociation
x 11 eV [32]). Finally, the signals of the residual mother
ion and the new fragment ion Fef are shown in Figure 2d.
The whole reaction cycle is repeated once again in an
analogous way in Figure 3. (Although the observed intensity
inevitably decreases with every step, in this experiment even
at the end it is still three to four times more intense than the
background noise.)
320 m/z
320 r n h 200
320 m/z
Fig. 3. a) Signal of the isolated Fef cluster ions (2nd generation; the signal
intensities were increased by a factor of 2.5). b) F T mass spectrum of the adduct
ions [Fe,(C,H,),]+ (m = 1-3) following the reaction of Fef with C,H, as in
Figure 2 b (2nd generation; the signal intensities were increased by a factor of
2.0). c) Signal of the isolated adduct cluster ions [Fe,(C,H,),]+ (2nd generation). d) FT mass spectrum of the mother- and fragment ions following CID of
[Fe,(C,H,),]+, 2nd generation with xenon as collision gas ( j%lo-’
E,, z 11 eV).
Thus, the originally produced Fe: ions have run through
the same reaction cycle twice. This demonstrates that a
naked metal cluster in the gas phase can indeed act as a
catalytic center and repeatedly build up a larger molecule
from smaller units.
as in earlier experiments mainly the first possibility was used [8,10] the
ethene in this case was added through the second pulsed valve.
For more details see a) M. P. Irion, A. Selinger, Z . Phys. Chem. 1989, 161,
233; b) M. P. Irion, A. Selinger, R. Wendel, Int. J. Mass Spectrom. Ion
Processes 1990,96,27; c) A. Selinger, P. Schnabel, W. Wiese, M. P. Irion,
Ber. Bunsenges. Phys. Chem. 1990,94, 1278.
See for example a) F. M. Devienne, J.-C. Roustan, Org. Mass Spectrom.
1982, f7,173; b) I. Katakuse, T. Ichihara, Y Fujita, T. Matsuo, T. SakUVdi,
H. Matsuda, Int. J. Mass. Spectrom. Ion Processes 1985, 67, 229.
P. Schnabel, M. P. Irion, K. G. Weil, Ber. Bunsenges. Phys. Chem. 1991,
95, 197.
a) M. P. Irion, A. Selinger, P. Schnabel, Z . Phys. D 1991,19,393; b) M. P.
Irion, P. Schnabel, A. Selinger, Ber. Bunsenges. Phys. Chem. 1990, 94,
1291; c) M. P. Irion, P. Schnabel, 1 Phys. Chem. 1991,95, 10596.
G . Ertl in Catalysis Science and Technology, Band 4 (Eds.: 1. R. Anderson,
M. Boudart), Springer, Berlin, 1983, p. 257-282.
a) P. Schnabel, M. P. Irion, K. G. Weil, J. Phys. Chem. 1991,95,9688; b)
Chem. Phys. Lett. 1992, 190, 255.
An ion type of a certain mass/charge ratio is isolated in the ICR cell from
ions of other masses, by intentionally exciting the ions of unwanted types
by a radio frequency (RF) “ejection pulse” thus eliminating them from the
cell. A particular variant of this technique is “FERETS’ (R. A. Forbes, F.
A. Laukien, J. Wronka, Int. 1 Mass Spectrom. Ion Processes 1988,83,23).
The center-of-mass energy, E,,, defines the maximum energy, which is
transmitted by the collision of a molecule M with an ion I. It is obtained
from the masses m, from the following formula: E,, = E,,.m,/(m, + mM).
M. M. Kappes, R. H. Staley, 1 Am. Chem. Sac. 1981, 103.1286.
D. Schroder, H. Schwarz, Angew. Chem. 1990, 102, 1466; Angew. Chem.
Int. Ed. Engl. 1990, 29, 1431.
D. Schroder, D. Sulzle, J. Hrusak, D. K. Bohme, H. Schwarz, In/. J. Mass
Spectrorn. Ion Processes 1991, 110, 145.
A complex experiment with several mass spectrometric steps (MS”) requires in a conventional instrument n sectors separated in space. However,
in an instrument with ion storage (FT-ICR, quadrupole ion trap) the
whole experiment is conducted in the time domain and nothing more than
a software extension to include all the pulses is needed. The number n of
possible steps is dependent upon the number of mother ions stored originally and the overall efficiency of the MS” process (fraction of the number
of end product ions and the number of original ions) (171. In FT-ICR the
finite ICR cell plate distance limits the cyclotron radius of the ions and thus
the maximum number of steps. Supposedly, the practical limit lies at n = 5
or n = 6 at most [18]. However, already in 1984 MS5 was demonstrated on
organic ions (FT-ICR at 1.4 T magnetic field strength) [19]. With the
application of high magnetic fields the experimentalist will run out of ideas
rather than of steps to be actually achieved for the MS” experiment; for
example a magnetic field strength of 7 T should allow even MSZ0[20].
[17] S. A. McLuckey, G. L. Glish, G. J. van Berkel, In/. J. Mass Spectrorn. Ion
Processes 1991, 106, 213.
(181 D. Schroder, H. Schwarz, Angew. Chem. 1990,102,925;Angew. Chem. Int.
Ed. Engl. 1990, 29,910.
[19] J. C. Kleingeld, Dissertation, Amsterdam, 1984.
[20] B. S. Freiser in Techniquesfor the Study of Ion-Molecule Reactions (Eds.:
J. M. Farrar, W. H. Saunders, Jr.), Wiley, New York, 1988, p. 76-78.
Received: December 11, 1991 [Z5067IE]
German version: Angew. Chem. 1992,104, 633
CAS Registry numbers:
Fez, 73145-65-0; Fe’, 7439-89-6; ethene, 74-85-1; ethyne, 74-86-2; benzene,
[I] 1 Chem. Soc. Faraday Trans. 1990,862343-2551 (Proceedings 25th Faraday-Symposium: “Large Gas Phase Clusters”), in which notably: A.
Kaldor, D. M. Cox, pp. 2459-2463.
[2] Z . Phys. D. 1991, 19, 20 (Proceedings ISSPIC-5).
[3] A. Kaldor, D. M. Cox, M. R. Zakin, Adv. Chem. P/7ys. 1988, 70, 220.
[4] The FT-ICR mass spectrometer used was home-built with a cubic ICR cell
of side length 80 mm, placed in the homogeneous field of an Oxford Instruments 7.05T superconducting magnet and a Spectrospin data system. Via
an Aspect 3000 computer it provided the cell with pulses, registered and
Fourier transformed the transient signal.[5] The metal cluster ions are
produced by the bombardment of the corresponding metal foil in an external chamber with 20 keV Xe+ ions from a duoplasmatron ion source [6].
They are injected into the differentially pumped ICR cell by a system of ion
optics. A grid in front of the cell decelerates the external ions electrostatically and controls their entrance into the cell. By using a home-made
piezoelectric valve at a distance of 8 cm, neon is pulsed in up to ca.
mbar, which by means of collisions helps in trapping the incoming
ions and cools them to approximately room temperature.[7] To initiate
chemical reactions with the thus thermalized ions, gases can be admitted
via continually adjustable leak valves or via a second pulsed valve. Where-
0 VCH Verlagsgesellschaft mbH, W-6940 Weinheim, 1992
Evidence for [PdSiO] from IR Spectroscopy**
By Thomas Mehner, Ralf Koppe, and Hansgeorg SchnockeP
Dedicated to Professor Wolfgang Beck
on the occasion of his 60th birthday
For a long time simple, coordinatively unsaturated transition-metal carbonyls have been known to exist under matrix
conditions. Lately theoretical methods have also been used
to study these small molecules such as [PdCO].[1-31Attempts to generate transition-metal complexes with SiO as
ligand in solid argon led to the preparation of [AgSi0].[41
Several theoretical[5.6 ] and experimental investigationst7]
are concerned with the clarification of the bonding relationships in this species. Apparently it is an ion pair AgfSi0-,[91
[*] Prof. Dr. H. Schnockel, Dr. T. Mehner, Dipl.-Chem. R. Koppe
Institut fur Anorganische Chemie der Universitat
Meiserstrasse I , D-W-8000 Munich 2 (FRG)
[**I This work was supported by the Deutsche Forschungsgemeinschaft and
the Fonds der Chemischen Industrie.
Angew. Chem. Int. Ed. Engl. 31 (1992) No. 5
as is shown by the formation of SiO- with Na' or Kf.["]
Here we report the first transition-metal complex containing
SiO as ligand in solid argon. With respect to bonding it is
similar to carbonyl complexes.
When SiO and palladium atoms were condensed together
with argon on a helium-cooled copper surface, we observed
in the IR spectrum (Fig. 1) besides the absorption of molecular SiO at 1226.0 cm-' a new band at 1246.3 cm-', which
we assigned to the complex [PdSiO] (1).[81Moreover, the
complex [PdCO] (2) is generated (f = 2050.0 cm-'), since
CO (2138 cm-') is always present in these experiments as
is calculated to be 8.92 mdyn A-' (SiO in argon: 9.02) and
that of Pd-Si, 2.69 mdyn A- '.
The linear, silicon-bridged structure for 1 (deduced from
the spectroscopic data) and the force constants (based on
experiment) were confirmed convincingly by ab initio calculations. The relativistic model potential of Sakai et al.,"'] as
implemented in the program package CADPAC 4.0L'91that
we applied here, was used as basis set for palladium. The
quality of this basis[201was tested by calculations of the
excitation energy 4d"('S) -+ 4d95s'('D) at MP-2 level
(exp.: 91.7, calcd: 80.0 kJmol-') and by optimization of the
geometry of PdH (%+) (exp.: v(PdH) =153.5, calcd:
r(PdH) =I51 pm). The basis sets for Si, C, and 0 were of
TZ2P quality.["] We performed all calculations at MP2 level
taking into account basis-set superposition error.
The results for the structure and for the atomic charges
from Mulliken population analyses are summarized in
Table 1. The geometry and the force constants correspond to
Table 1. Calculated (MP2) geometries, force constants, and atomic charges (Mulliken)
for the linear molecules [PdCO] (2) and [PdSiO] (l),and the isolated molecules CO and
[PdCO] 1.86
[PdSiO] 2.15
CO[b] SiO [c]
12LO 1220 1200 1180
Fig. 1. Section of the IR matrix spectrum after the condensation of Pd atoms
with Si160/Si'80.
Annealing the matrix doubles the intensity of the band due
to 1 and leads to an additional absorption band at
1237.1 cm- '. Simultaneously [Pd(CO),] (2044 cm-') is
formed from 2 and CO. Since under these conditions no
palladium carbonyl complex with more than one Pd atom is
generated, the absorption at 1237 cm-' may be attributed to
the complex [(OC)Pd(SiO)] formed from CO and 1. The
formation of [Pd(SiO),] can be ruled out because the dimerization of SiO is energetically more favorable than the coordination of a second SiO to 1.[", 12]
At high CO concentrations (Ar/CO mixtures of 20: I),
[(OC)Pd(SiO)] (1237 cm- ') is observed alongside the complexes [Pd(CO),] ( n = 2-4) even without annealing. Since no
CO absorption is detected," 31 the assignment for the absorption at 1237 cm- is based initially on theoretical consideration of the bonding in possible products that can form
under these conditions but also on the specific experimental
conditions under which this absorption is observed.['51
Qualitative confirmation of the assignment of the SiO
band of 1[16] is provided by the position of the SiO band of
the isotopomer [PdSi"O] at 1205.8 cm-'. Absorption by the
isotopomers with 29Si and 'OSi are not observed. They apparently lie beneath the band at 1237.3 cm- ([(CO)Pd(SiO)])
and 1226.0cm-' (uncoordinated SiO). The band at
1237.1 cm-' is shifted to 1197.1 cm-' by l80substitution,
which similarly indicates a Si=O vibrational band.
Conclusive confirmation for the assignment of the band at
1246.3 cm- ' to [PdSiO] (1) is found in normal-coordinate
analysis on assumption of a simple valence force field
(f(PdSi/SiO) = 0). The measured 160/'80
shift of the SiO
stretching frequency is satisfactorily reproduced only for a
linear silicon-bridged Pd-Si-0 molecule[' 1' on assumption of
a Pd-Si frequency of 375 cm-'. The SiO force constant in 1
Angew. Chem. Int. Ed. Engl. 31 (1992) No. 5
f(PdSi/SiO) = 0.007 mdyn k ' .[b] E$"
[a] f(PdC/CO) = 0.66 mdyn k ' ,
-113.143575 Hartree. [c] EtFZ = - 364.293021 Hartree.
those deduced from the spectra. The Pd-Si bond in 1 is
surprisingly strong, since the enthalpy of reaction for Pd
SiO .+ [PdSiO]is - 182 kJmol-' (basis-set superposition error is taken into account). It is thus of the same order as that
for the formation of [PdCO] (- 162 kJmol-'). Furthermore, the frequencies resulting from the calculations
(Table 2) confirm the experimental blue shift of the SiO band
observed when SiO is coordinated in [PdSiO] (AV,,~,=
-20.3, AvCalcd= -25.5 cm-').[221
Table 2. Measured and calculated (MP2) frequencies for CO, SiO, [PdCO],
and [PdSiO] (frequencies in cm- ', intensities in parentheses in kmrnol-').
359.0 (6.4)
462.0 (27.2)
1203.4 (9.9)
1228.9 (65.6)
2133.0 (30.2)
1969.2 (632.2)
164.3 (8.9)
343.7 (30.4)
The similar bonding situation in 1 and 2 suggested by
these results is underscored by population analyses. The
charges on palladium show similar distributions for these Pd
complexes (2: Pd
P ' l oPol;
p0 .1 2
The theoretical and experimental
results presented here show that the bonding relationships in
[PdSiO] can be compared with those in [PdCO]; that is, the
(J donor bond is strengthened by R acceptor components.
The proportion of cr and R bonding, however, differs in 2 and
1, resulting in a red shift in the frequency of the uncoordinated ligand on coordination in 2, but a blue shift in 1.
Verlagsgesellschaft mbH, W-6940 Weinheim, 1992
Received: October 31, 1991 [Z5000IE]
German version: Angew. Chem. 1992, 104, 653
CAS Registry numbers:
SiO, 10097-28-6; Pd, 7440-05-3; [PdSiO], 139495-54-8; fPdCO], 41772-86-5;
[(OC)Pd(SiO)], 139523-85-6; CO, 630-08-0.
[I] G. W. Smith, E. A. Carter, J. Phys. Chem. 1991, 98, 2327, and references
cited therein.
[2] C. M. Rohlfing, P. J. Hay, J. Chem. Phys. 1985,83, 4641.
(31 P. Schwerdtfeger, J. S. McFeaters, J. J. Moore, D. M. McPherson, R. P.
Cooney, G. A. Bowmaker, M. Dolg, D. Andrae, Langmuir, 1991, 7,
[4] T. Mehner, H. Schnockel, M. J. Almond, A. J. Downs, J. Chem. SOC.Chem.
Commun. 1988, 117.
[5] G. E. Quelch, R. S. Grev, H. F. Schaefer 111, J. Chem. Sor. Chem. Commun. 1989, 1498.
[6] J. S. Tse, J. Chem. SOC.Chem. Commun. 1990, 1119.
[7] B. Chenier, H. A. Joly, J. A. Howard, B. Mile, P. L. Timms, J. Chem. SOC.
Chem. Commun. 1990, 581.
[8] SiO was synthesized by passing 0, over hot silicon (1200°C); Pd atoms
were generated by vaporization of the metal at 1450 "C. The cryostat and
vaporization conditions have been described previously: R. Ahlrichs, R.
Becherer, M. Binnewies, H. Borrmann, M. Lakenbrink, S. Schunck, H.
Schnockel, 1 Am. Chem. SOC.1986,108,7905. The IR spectra were recorded on an IFS-113 v and an IFS-66v FT-IR spectrometer in absorption with
the help of a reflection unit.
[Y] R. Koppe, H. Schnockel, C. Jouany, F. X. Gddea, J. C. Barthelat, Heterout. Chem., in press.
[lo] R. Koppe, H. Schnockel, Heteroat. Chem., in press.
[I I] H. Schnockel, T. Mehner, H. S. Plitt, S. Schunck, J. Am. Chem. Soc. 1989,
1121 From an ab initio calculation for linear [(OC)Pd(SiO)] ( E z p =
- 573.873915 Hartree; r ( C 0 ) = 1.135, v(PdC) = 2.02, r(PdSi) = 2.25,
r(Si0) = 1.525 A) erhalt man folgende Reaktionsenergien.
[PdCO] + SiO 4 [(OC)Pd(SiO)] BE = -107.4 kJmol-'
[PdSiO] + CO 4 [(OC)Pd(SiO)] A E = - 87.4 kJmol-'
Pd CO + SiO + [(OC)Pd(SiO)] AE = - 296.4 kJmol-'
[I 31 The corresponding bands in the carbonyl region are possibly hidden by the
CO bands of the various carbonylpalladium complexes.
[I41 E. P. Kundig, M. Moskovits, G. A. Ozin, Can. J. Chem. 1972, 50, 3587.
1151 The lowering of the frequency on changing from [PdSiO] to [Pd(SiO)(CO)]
is reasonable, because the Pd-Si bond is weakened as a result of a smaller
backbonding component. The Si-0 bond and its stretching frequency
should thus be more similar to the their counterparts in uncoordinated SiO
(1226 cni-I). This is also supported by ah initio calculations [12]: both the
Pd-C distance in 2 and the Pd-Si distance in 1 are shorter than in
[Pd(SiO)(CO)]. The bonding relationships in 2 and [Pd(CO),] yield additional evidence for the correctness of these arguments: The CO force constantflC0) is a bonding parameter free from coupling influences. Its value
in 2 of 16.97 m d y n k ' is smaller than that in [Pd(CO),] (17.16 m d y n k '
[14]).ThustheCObondoffreeCOU[CO) =18.36mdyn.k1inanargon
matrix) more closely resembles that of [Pd(CO),] than that of 2. Consequently both the CO bond in 2 and the SiO bond in 1 become more like the
bond in uncoordinated CO or SiO on addition of a CO ligand at Pd.
1161 <(Si'60) =1226.0; <((si'"O) =1182.2 cm-' in an argon matrix.
[17] In an oxygen-bridged complex, the silicon isotope shifts (Pd-O-"Si
would be much smaller and would therefore be able to be
observed as they would not overlap with strong absorptions of othercompounds such as [(SiO)Pd(CO)] or SiO. Calculation of the frequencies for
[Pd29Si0] and [Pd"SiO] confirm that these bands are hidden by overlap.
[I81 Y. Sakai, E. Miyoshi, M. Klobukowski, S. Huzinaga, J. Compul. Chem.
1987, 8, 226.
1191 R. D. Amos, J. E. Rice, CADPAC: The Cambridge Analytic Derivatives
Package, issue 4.0, Cambridge, 1987.
[20] The 5 s and 4d as well as the 4 p electrons are treated explicitly in triple-zeta
contraction to include any possible contribution of these electrons to the
bond formation. As suggested by Sakai and Huzinaga et al. 1181, we added
two p-like polarization functions (Pd (MP) (Sp 8 s 5d 2p)/[3 p 3 s 3 d 2 p].
[21] S. Huzinaga, Approximate Atomic Functions. Technical Report 1971; University of Alberta, Canada, 1971.
[22] The structural data and bond energies we calculated for [PdCO] agree well
with earlier MP2 calculations [2] (where r(Pd-C) =1.882, r(C0) =1.185A, t ( P d C 0 ) =180", Pd-C bond energy = -156kJmol-').
In a new detailed theoretical work 131, many computational methods and
better hasis sets than in our calculations were used to investigate the bonding relationships in 2. Among the computational methods incorporating
electron correlation (MP2, MP3, CISD. CISC), the MP2 calculation
yielded the largest bond energy (- 109 kJ mol- I ) ; that is, also in our MP2
calculations the strength of the Pd-Si bond is a little overestimated: the
Pd-Si distance is too small and the bond energy (Pd + SiO + [PdSiO]) too
large. This is in accord with the difference in the frequency shift calculated
by us (v,,(CO~v,,(PdCO)) of 164 or 122 cm-' [3] and the experimental
value of 88 cm-I.
Q VCH Verlag.~gesellscho~
mbH, W-6940 Weinheim. 1992
The Structure of C,o: Orientational Disorder in
the Low-Temperature Modification of C,, **
By Hans-Beat Biirgi, Eric Blanc, Dieter Schwarzenbach,*
Shengzhong Liu, Ying-jie Lu, Manfred M . Kappes,
and James A . Ibers
The availability of gram quantities of fullerenes['l and the
discovery of superconductivity in C,, doped with an alkali
metalL2'have led to a broad spectrum of investigations dealing with diverse chemical and physical properties of solid
C,, . Models of the structure of C,, are the starting point for
much of this work. In particular for investigations of the
solid-state properties, reliable information on the crystal and
molecular structure of C,, is indispensable.
On the assumption of icosahedral symmetry, the molecules of C,, are approximately spherical. They crystallize in
a cubic close-packed arrangement.13]Above 249 K, the molecules are orientationally disordered in the crystal (the probable space group is F m h ) . X-ray data of C,, measured at
110 K on a merohedrally twinned crystal were interpreted in
terms of an ordered model in the space group Pa3 with
molecular site symmetry 3.[31However, the lengths of bonds
which are chemically equivalent and related by the icosahedral symmetry mT5 of the molecule differ considerably in this
model; the distances also show deviations of up to 0.1 A
from those determined in an electron-diffraction study on
gaseous C6,.L4] It was notedL3]that the agreement between
the observed and calculated structure amplitudes is not nearly as good as would be expected from the internal agreement
among symmetry-equivalent reflections, and it was suggested that there may be deficiencies in the model concerning
thermal motion or twinning. Analysis['] of the atomic displacement tensors U from ref.[31and subsequent inspection
of the residual displacements with the graphics program
PEANUT[,] showed them to be incompatible with the expected motion of rigid molecules and also suggested the presence of disorder in the structure. In an independent study of
C,, based on neutron-diffraction measurements on a powder at 5 KL7]and a Rietveld refinement with bond angles
restrained to idealized values, the same packing as in ref.[3]
was found with bond lengths in good agreement with those
obtained from electron-diffraction studies of gaseous C,,
and with no indication of disorder. The agreement of observed and calculated diffraction profiles was satisfactory.
Displacement parameters were not reported in ref.[71
We have now carried out additional analyses of the singlecrystal X-ray diffraction data of ref.[31collected at 110 K
(Cu,,) and also of data collected in a similar manner from a
second twinned crystal at 153 K (Cu,,) and from a third
twinned crystal at 200 K and 100 K (Mo,,). These analyses
show that at these temperatures the molecules are statistically distributed in two orientations: one with a large population (major orientation) as has been found in the earlier
[*] Prof. D. Schwarzenbach, E. Blanc
Institute of Crystallography, University of Lausanne
BSP, CH-1015 Lausanne (Switzerland)
Prof. H.-B. Biirgi
Laboratory of Crystallography, University of Bern
S. Liu, Y-J. Lu, Prof. M. M. Kappes, Prof. J. A. Ibers.
Department of Chemistry, Northwestern University
[**I This research was supported by the Swiss National Science Foundation,
grants 31-194.91 (H.-B.B.) and 20-28930.90 (D.S.), the U.S. National Science Foundation, grant CHE-8922754 (J.A.I.), the Northwestern University Materials Research Center, NSF grant DMR-8821571 (M.M.K.), and
the Science and Technology Center for Superconductivity, NSF grant
DMR-8809854 (J.A.I.). Note added in proof (April 28, 1992): After this
paper was submitted, the results of a high-resolution neutron-diffraction
experiment on C,, powder were published 1131.
0570-0833f92f0508-0640$3.50+ ,2510
Angew. Chem. In[. Ed. Engl. 31 (1992) No. 8
Без категории
Размер файла
417 Кб
spectroscopy, pdsio, evidence
Пожаловаться на содержимое документа