Reactions of laser-ablated Ni, Pd, and Pt atoms with carbon monoxide: Matrix infrared spectra and density functional calculations on M(CO)n (n = 1-4), M(CO)n- (n = 1-3), and M(CO)n+ (n = 1-2), (M = Ni, Pd, Pt)

Article abstract of DOI:10.1021/jp993646q

There has always been extra focus on the bonding characteristics of monocarbonyls since they are deemed as the models of the CO binding to the metal surface. Laser-ablated Ni, Pd, and Pt atoms were reacted with CO molecules during condensation in a neon matrix at 4 K. Annealing, photolysis, and isotopic substitution experiments identified metal carbonyl anions [M(CO)_n^- (n = 1-3)] and cations [Ni(CO)_n⁺ (n = 1-4); Pd(CO)_n⁺ (n = 1,2); Pt(CO)_n⁺ (n = 1-3)], and neutrals [M(CO)_n (n = 1-4)]. Doping with the CCl₄ electron trap increased cation and decreased anion absorptions and supported the identification of the ionic species. The density functional theory (DFT) calculations showed that experimental results agreed excellently with frequencies and isotopic frequency ratios, confirming the vibrational assignments and the identification of these metal carbonyl complexes. All the monocarbonyls were linear, except PdCO^- and PtCO^-, which were computed to be bent by both DFT/B3LYP and BP86 functionals. Natural bonding orbital analysis on the monocarbonyls, conducted to describe the bonding of CO to transition metals, showed that: C-O bond orders were cations > neutrals > anions, indicating that C-O stretching frequencies have the same order as seen for other transition metals; and the various configurations of metal atoms in anions, cations, or neutrals could be employed to explain the different geometries.

Full text of DOI:10.1021/jp993646q

J. Phys. Chem. A 2000, 104, 3905-3914
3905
Reactions of Laser-Ablated Ni, Pd, and Pt Atoms with Carbon Monoxide: Matrix Infrared
-
Spectra and Density Functional Calculations on M(CO)n (n ) 1-4), M(CO)n (n ) 1-3),
+
and M(CO)n (n ) 1-2), (M ) Ni, Pd, Pt)
Binyong Liang, Mingfei Zhou, and Lester Andrews*
Department of Chemistry, UniVersity of Virginia, CharlottesVille, Virginia 22901
ReceiVed: October 12, 1999
Laser-ablated Ni, Pd, and Pt atoms have been reacted with CO molecules followed by condensation in excess
-
neon at 4 K. Besides the neutral metal carbonyl species, metal carbonyl anions [M(CO)
n
(n ) 1-3)] and
+
+
+
cations [Ni(CO)
Doping with the CCl
n
(n ) 1-4), Pd(CO)
n
(n ) 1,2), and Pt(CO)
n
(n ) 1-3)] have been produced and identified.
electron trap increases cation and decreases anion absorptions and confirms the
4
identification of the ionic species. The excellent agreement between observed bands and frequencies and
isotopic frequency ratios from the DFT calculations supports the vibrational assignments and the identification
of these metal carbonyl complexes.
I. Introduction
laser-ablation method to systematically study the reactions of
Ni group atoms, cations, and electrons with CO in excess neon,
Binary transition-metal carbonyls are among the earliest
and then use infrared spectroscopy coupled with CCl4 doping
1
known and most commonly encountered organometallic com-
25
to trap electrons to identify the metal carbonyl neutrals, anions,
2
pounds. Metal carbonyl fragments are ubiquitous components
and cations. We will also perform density functional theory
of other organometallic compounds and larger metal clusters,
and they are frequently invoked as active species in homo-
geneous catalytic cycles. Moreover, the monocarbonyls are
(DFT) calculations to model these carbonyl species, including
their vibrational analyses, ground state geometries, and bonding
properties.
3
often considered as models for CO binding to the metal
surface.4 Nickel tetracarbonyl was the first metal carbonyl to
,5
II. Experimental and Computational Methods
1
be prepared, and nickel group carbonyls are among the
The experimental method for laser ablation and matrix
most studied metal carbonyls both experimentally by matrix-
2
6-29
isolation has been described in detail previously.
Briefly,
6
7
isolation spectroscopy and theoretically. The corresponding
neutral Pd(CO)x and Pt(CO)x (x ) 1, 2, 3, 4) carbonyls have
been prepared using the matrix isolation technique, and their
infrared spectra in argon and krypton matrices have been
the Nd:YAG laser fundamental (1064 nm, 10 Hz repetition rate
with 10 ns pulse width) is focused to irradiate the rotating metal
targets. Typically, low laser power (3-5 mJ/pulse) is used,
which favors the stabilization of ionic species and minimizes
metal cluster formation. Laser-ablated metal atoms are co-
deposited with carbon monoxide (0.1%-0.4%) in excess neon
onto a 4 K CsI cryogenic window at 2-4 mmol/h for 0.5-1 h.
8
-10
11-15
16
reported.
It is generally accepted that NiCO,
PdCO,
1
7
1 +
and PtCO possess Σ electronic states with linear geometries.
1
6
The dicarbonyls have also been examined theoretically.
Finally, the vibrational spectra18 and the force fields of the
19
1
3
16
12 18
Carbon monoxide (Matheson) and isotopic C O and C O
Cambridge Isotopic Laboratories) and selected mixtures are
tetracarbonyls have been reported.
On the other hand, comparatively less work has been done
on the metal carbonyl cations and anions, especially their
vibrational spectra. The photoelectron spectra (PES) of Ni(CO)n
have given electron affinities and vibrational fine structure of
neutral carbonyls. Recently, the anion PES of PdCO has been
observed, and the electron affinity and CO vibrational frequency
of PdCO have been deduced.21 Collision-induced dissociation
coupled with mass spectrometry have been employed to study
(
used in different experiments. Fourier-transformed infrared
-1
(
7
FTIR) spectra are recorded at 0.5 cm resolution on a Nicolet
-
-1
50 spectrometer with 0.1 cm accuracy using a mercury
-1
cadmium telluride (MCTB) detector down to 400 cm . Matrix
samples are annealed at different temperatures, and selected
samples are subjected to photolysis using a medium-pressure
mercury lamp (240-580 nm) with the globe removed, or a
tungsten lamp and applying optical filters when needed.
DFT calculations are performed on all of the proposed
reaction products using the GAUSSIAN 94 program and the
hybrid B3LYP functional. The 6-311+G* basis set is used
for carbon, oxygen, and nickel, and the LanL2DZ effective
20
-
-
+
sequential bond energies in Ni(CO)n (n ) 1-3) and Ni(CO)n
2,23
(
n ) 1-4).2 Bond energies of the anions combined with the
30
electron affinity data have been used to determine bond energies
of the neutral species.22 Finally, ionic Ni group carbonyls have
received little theoretical attention; to our knowledge, only the
mono- and dicarbonyl cations of Ni and Pd have been examined
by theory.24
31
32
33
core potential is employed for palladium and platinum. Natural
bonding orbital (NBO) analyses are done for all of the nine
34
monocarbonyls.
Recently, we reported argon matrix infrared spectra of
-
25
III. Results and Discussion
A. Infrared Spectra and Assignments. Infrared spectra of
laser-ablated Ni, Pd, and Pt atom, cation and electron reaction
Ni(CO)n (n ) 1-3). It will be interesting to determine the
neon matrix interaction for the negative carbonyls and to find
how Pd and Pt compare to the Ni counterparts. We will use the
1
0.1021/jp993646q CCC: $19.00 © 2000 American Chemical Society
Published on Web 04/04/2000

3
906 J. Phys. Chem. A, Vol. 104, No. 17, 2000
Liang et al.
assignment
TABLE 1: Infrared Absorptions (cm-1) from Reaction of Laser-Ablated Ni Atoms with CO in Excess Neon at 4 K
1
2
16
13 16
12 18
13 16
13 16
C O
C O
C O
C O + C O
2347.6, 2282.0
R(12/13)
R(16/18)
2
2
2
2
2
2
2
2
2
2
2
2
2
2
1
1
1
1
1
1
1
1
1
1
1
347.6
209.2
206.5
205.3
192.4
186.2
176.2
141.1
091.4
089.7
056.3
026.9
025.0
006.6
980.5
978.9
873.2
870.5
867.7
864.7
860.6
841.4
817.1
815.4
813.8
2282.0
2159.9
2156.9
2156.1
2312.5
2157.6
2156.0
2153.8
1.0288
1.0228
1.0230
1.0228
1.0152
1.0239
1.0234
1.0239
CO
Ni(CO)
NiCO
Ni(CO)
Ni(CO)
Ni(CO)
Ni(CO)
CO
2
+
2
site
+
2206.5, 2156.7
2205.3, 2165.9, 2156.1, 2223.8 (sym)
+
+
+
+
2
3
3
4
2135.2
2124.9
2089.7
2046.9
2045.3
2009.2
1980.1
1978.3
1965.4
1934.5
1932.9
1.0239
1.0241
1.0246
1.0217
1.0217
1.0234
1.0236
1.0236
1.0210
1.0238
1.0238
2129.5
2093.6
2042.1
2040.3
2009.8
1981.3
1979.6
1958.1
1936.4
1934.7
1832.5
1830.9
1823.9
1821.0
1815.6
1798.5
1774.8
1773.2
1771.6
1.0219
1.0227
1.0241
1.0242
1.0231
1.0230
1.0229
1.0248
1.0228
1.0228
1.0221
1.0216
1.0240
1.0240
1.0248
1.0239
1.0238
1.0238
1.0238
2141.1, 2093.6
2091.4, 2071.6, 2042.1
2089.7, 2069.7, 2040.3
2056.3, 2037.4, 2026.4, 2017.2, 2009.8
2026.8, 2004.7, 1992.3, 1981.3
2025.1, 2003.6, 1990.6, 1979.7
2006.5, 1958.2
1980.6, 1953.7, 1936.4
1978.9, 1951.9, 1934.7
1873.0, 1832.5
Ni(CO)
Ni(CO)
Ni(CO)
Ni(CO)
Ni(CO)
NiCO
2
2
4
3
3
site
site
site
Ni(CO)
Ni(CO)
2
2
Ni
Ni
Ni(CO)
Ni(CO)
x
CO site
1830.6
1824.9
1823.4
1821.0
1801.5
1777.6
1775.9
1774.3
1870.5, 1831.0
1.0218
1.0235
1.0226
1.0217
1.0221
1.0222
1.0222
1.0223
x
CO
-
1867.9, 1824.0, 1848.8, 1834.9
1864.8, 1845.9, 1832.0, 1821.0
1860.6, 1815.5
3
3
site
-
-
NiCO
1841.1, 1798.6
x
Ni CO
-
-
-
1817.2, 1790.0, 1774.9
1815.4, 1788.2, 1773.7
1813.8, 1786.8, 1771.7
Ni(CO)
Ni(CO)
Ni(CO)
2
2
2
site
site
TABLE 2: Infrared Absorptions (cm-1) from Reaction of Laser-Ablated Pd Atoms with CO in Excess Neon at 4 K
1
2
16
13 16
12 18
12 16
13 16
C O
C O
C O
C O + C O
R(12/13)
R(16/18)
assignment
+
2210.5
2206.4
2074.6
2068.2
2065.8
2064.7
2062.7
2057.3
2056.4
1909.0
1902.7
1878.8
1872.2
1870.1
1834.3
1832.3
2161.1
2156.8
2028.4
2022.9
2020.5
2018.5
2016.7
2009.6
2008.7
1864.6
1859.8
1835.6
1829.2
1827.1
1791.4
1789.5
2158.7
2155.4
2026.0
2019.0
2016.5
2210.5, 2169.8, 2161.1, 2225.0 (sym)
2206.4, 2156.8
2074.6, 2052.3, 2042.2, 2034.8, 2028.4
2068.2, 2038.7, 2023.0, 2128.4 (sym)
2065.8, 2036.3, 2020.5, 2125.2 (sym)
1.0228
1.0230
1.0228
1.0224
1.0224
1.0229
1.0228
1.0238
1.0238
1.0238
1.0231
1.0235
1.0235
1.0235
1.0239
1.0239
1.0240
1.0237
1.0240
1.0244
1.0244
Pd(CO)
Pd(CO)
Pd(CO)
Pd(CO)
Pd(CO)
Pd(CO)
Pd(CO)
2
+
4
2
2
3
3
site
site
2014.4
2012.4
2011.4
1868.2
1859.3
1836.4
1831.9
1828.8
1794.6
1792.9
2062.7, 2037.4, 2026.0, 2016.8
2057.3, 2009.6
2056.4, 2008.7
1909.0, 1864.5
1902.4, 1959.8
1.0240
1.0224
1.0224
1.0218
1.0233
1.0231
1.0226
1.0226
1.0221
1.0220
PdCO site
PdCO
-
Pd(CO)
Pd
Pd
x
CO
CO
1878.8, 1835.6
x
-
-
-
-
1872.2, 1853.1, 1839.2, 1829.2
1870.1, 1851.5, 1837.7, 1827.1
1834.3, 1806.8, 1791.3
1832.3, 1805.0, 1789.5
Pd(CO)
Pd(CO)
Pd(CO)
Pd(CO)
3
3
2
2
site
site
TABLE 3: Infrared Absorptions (cm-1) from Reaction of Laser-Ablated Pt Atoms with CO in Excess Neon at 4 K
1
2
16
13 16
12 18
12 16
13 16
C O
C O
C O
C O + C O
2210.3, 2173.5
R(12/13)
R(16/18)
assignment
+
2
2
2
2
2
2
2
2
2
2
1
1
1
1
210.3
207.9
204.7
065.5
064.7
058.1
056.2
053.2
046.0
044.0
896.3
879.5
875.8
844.0
2159.0
2156.7
2156.7
2024.4
2023.4
2010.9
2007.5
2004.7
1999.7
1997.1
1858.7
1839.3
1835.7
1804.6
1.0238
1.0237
1.0223
1.0203
1.0204
1.0235
1.0242
1.0242
1.0232
1.0235
1.0202
1.0219
1.0218
1.0196
Pt(CO)
Pt(CO)
PtCO
2
3
+
2158.4
1.0229
+
2014.9
2013.9
2011.6
2010.8
2007.9
1999.8
1997.8
1850.0
1835.3
1831.7
1801.0
2065.5, 2014.9
2064.7, 2013.9
1.0251
1.0252
1.0231
1.0226
1.0226
1.0231
1.0231
1.0250
1.0241
1.0241
1.0238
PtCO
PtCO site
2058.1, 2037.3, 2026.6, 2018.5, 2011.6
2056.2, 2027.9, 2010.8
2053.2, 2025.1, 2007.9
2046.1, 2021.3, 2010.9, 1999.8
2044.0, 2020.0, 2007.7, 1997.8
1896.3, 1850.0
Pt(CO)
Pt(CO)
Pt(CO)
Pt(CO)
Pt(CO)
4
2
2
3
3
site
site
-
PtCO
-
Pt(CO)
Pt(CO)
Pt(CO)
3
3
2
site
-
-
1875.8, 1856.2, 1842.2, 1831.8
1844.0, 1817.3, 1801.0
1
3
-1
products with CO molecules in excess neon are displayed in
Figures 1-9, and the absorption bands in different isotopic
experiments are listed in Tables 1-3.
CO isotopic experiments, the 2006.6 cm absorption shows
-1
a doublet feature, while the 1978.9 cm band has a 1:2:1 triplet
pattern. These two bands are assigned as the C-O stretching
mode of NiCO and the antisymmetric C-O vibration of Ni-
(CO)2, respectively. The 1978.9 cm band exhibits a matrix
site splitting at 1980.5 cm . Compared to the NiCO and Ni-
Ni(CO)n (n ) 1-4). A strong absorption observed at 2006.6
-
1
-1
cm on deposition increases slightly on 6 K annealing but
-
1
-1
decreases during higher annealings. A band at 1978.9 cm is
also observed on deposition, and its intensity increases more
-
1
(CO)2 absorptions, the feature at 2025.0 cm is weak on
deposition, but increases steadily during annealings. This band
-
1
12
than the 2006.6 cm band during annealings. In mixed CO/

Reactions of Laser-Ablated Ni, Pd, and Pt Atoms with CO
J. Phys. Chem. A, Vol. 104, No. 17, 2000 3907
Figure 1. Infrared spectra in the 2220-2160 and 2100-1960 cm-1
regions for laser-ablated Ni co-deposited with 0.2% CO in neon at 4
K: (a) sample deposited for 30 min, (b) after 6 K annealing, (c) after
Figure 2. Infrared spectra in the 1890-1810 cm- region for laser-
ablated Ni co-deposited with 0.2% CO in neon at 4 K: (a) sample
deposited for 30 min, (b) after 8 K annealing, (c) after λ > 630 nm
photolysis, (d) after λ > 470 nm photolysis, (e) after 240-580 nm
photolysis, (f) after 10 K annealing.
1
8
K annealing, (d) after 10 K annealing, (e) after 12K annealing, (f)
after 240-580 nm photolysis.
has a 5:2:2:5 splitting pattern in the mixed 12CO/ CO isotopic
experiment, which is characteristic of the doubly-degenerate
mode of a trigonal species,35 and accordingly, the 2025.0 cm
band is assigned to the doubly degenerate C-O stretching mode
13
-1
-
1
of Ni(CO)3. The band at 2056.3 cm is almost invisible on
deposition and increases markedly in the annealings, and its
splitting pattern in 12CO/ CO experiment is 8:2:3:2:8. It is
13
35
assigned to the triply degenerate C-O stretching mode of Ni-
CO)4. The isotopic counterparts and the mixed isotopic
(
splittings of those bands are listed in Table 1. These results are
in good agreement with the previous and recent argon matrix
-
1
work; the Ni(CO)n bands show 12.2, 13.4, 7.8, and 5.1 cm
blue shifts in the neon matrix with n ) 1, 2, 3, and 4,
2
5,36
respectively.
is bent in an argon matrix.25,36 It is worth noting that the Ni-
CO)2 molecule is also bent in a neon matrix because a weak
Recent infrared spectra indicate that Ni(CO)2
Figure 3. Infrared spectra in the 2215-2200 and 2080-2050 cm-1
(
regions for laser-ablated Pd co-deposited with 0.2% CO in neon at 4
K: (a) sample deposited for 30 min, (b) after 6 K annealing, (c) after
8 K annealing, (d) after 10 K annealing, (e) after 240-580 nm
photolysis, (f) after 12 K annealing.
-
1
13 16
symmetric stretch is observed at 2089.7 cm . In the C O
and C O experiments, this band shifts to 2040.3 and 2045.3
cm , respectively. In the mixed CO/ CO experiment, a 1:2:1
relative intensity triplet displays an intermediate band at 2069.7
cm which is 4.7 cm higher than the average of two pure
isotopic bands. Correspondingly, the intermediate band of the
12 18
-1
12
13
-1
8,9
and 2044 cm , respectively, and with very recent argon
-
1
-1
36
matrix studies of Manceron et al. who observed 2050.3 and
044.2 cm bands for PdCO. The PdCO fundamental is blue-
-1
2
-
1
antisymmetric stretch of Ni(CO)2 is 4.9 cm lower than the
average of two pure isotopic bands, owing to the interaction
-1
shifted 6.1 and 12.2 cm by the neon matrix. However, the
-1
2
140 ( 60 cm fundamental for PdCO deduced from the
13
between the two stretching modes in the lower symmetry O C-
21
-
photoelectron spectrum of PdCO is too high and probably
suffers from accuracy of the weak fine structure band measure-
1
2
Ni- CO molecule.
The symmetric modes at 2006.6 and 2089.7 cm- for NiCO
1
-1
ment. Our neon matrix observation at 2056.4 cm is probably
and Ni(CO)2 may be compared to the PES fine structure
-1
within 10 cm of the true PdCO gas phase fundamental, which
2
0
-1
intervals 1940 and 2100 ( 80 cm ; the weaker band in the
-1
we predict at 2065 ( 5 cm .
The interesting question here is again the structure of
Pd(CO) . Both previous works observed Pd(CO) at 2044 cm
2 2
(6 cm below PdCO) and claimed that Pd(CO) is a linear
-
NiCO PES is more difficult to measure accurately.
-1
Pd(CO)n (n ) 1-4). Similar to Ni(CO)n, the absorptions of
Pd(CO)n are also observed and identified; the major difference
is that the Pd(CO)n absorptions are more congested as also found
-
1
2
8,9
molecule because of the missing symmetric mode. The
geometry of this molecule might be linear because the pure
isotopic symmetric stretching mode is not observed here.
2
8,9
in the thermal Pd atom work. The new band observed at
-
1
2
056.4 cm on deposition is assigned to the C-O stretching
mode of PdCO based on the doublet pattern in our mixed
CO/ CO experiment. Another prominent band on deposition
at 2065.8 cm is assigned as the antisymmetric stretching mode
of Pd(CO)2, which shows a 1:2:1 splitting in our mixed CO/
However, the bent structure for Ni(CO) and the DFT calculation
1
2
13
12
13
suggest a bent form. In our mixed CO/ CO experiment, a
new band at 2125.2 cm is observed in addition to the 1:2:1
-
1
-1
1
2
relative intensity feature of the antisymmetric mode of Pd(CO)2.
This new band tracks with the 1:2:1 triplet bands and is assigned
13
CO experiment. The weaker Pd(CO)3 and Pd(CO)4 absorptions
-
1
12
13
observed at 2062.7 and 2074.6 cm become prominent bands
during annealings (Figure 3). The annealing behaviors and the
mixed isotopic splittings are very similar to those of Ni(CO)n
as the symmetric mode of Pd( CO)( CO). The intermediate
-
1
-1
band of the triplet at 2036.3 cm is 6.9 cm lower than the
average of two pure isotopic bands, which indicates ν
sym
37
>
(n ) 1-4) bands.
νantisym. A similar spectrum was observed for Rh(CO) . By
analogy to the Ni(CO)2 case, the weak 2125.2 cm band should
be about 7 cm higher than the average of the two pure isotopic
2
-1
This result is in agreement with the previous argon and
krypton matrix work on PdCO, which reported bands at 2050
-1

3908 J. Phys. Chem. A, Vol. 104, No. 17, 2000
Liang et al.
Figure 4. Infrared spectra in the 1920-1820 cm-1 region for laser-
ablated Pd co-deposited with 0.2% CO in neon at 4 K: (a) sample
deposited for 35 min, (b) after 8 K annealing, (c) after 10 K annealing,
-1
Figure 5. Infrared spectra in the 1920-1780 cm region for laser-
12
13
ablated Pd co-deposited with 0.15% CO + 0.15% CO in neon at 4
K: (a) sample deposited for 60 min, (b) after 6 K annealing, (c) after
(d) after λ > 630 nm photolysis, (e) after λ > 510 nm photolysis, (f)
8
K annealing, (d) after 10 K annealing, (e) after λ > 470 nm photolysis.
after 12 K annealing.
bands. On the basis of this information and the isotopic ratio
from the DFT calculation, we estimate the symmetric modes
1
2
13
-1
of Pd( CO)2 and Pd( CO)2 to be near 2143 and 2093 cm .
Unfortunately, strong absorptions of 12CO and CO mask these
regions. Hence, the geometry of this molecule is still an open
13
-
1
question. Finally, the observation of Pd(CO)2 at 9 cm aboVe
PdCO in solid neon in contrast to 6 cm below in solid argon
and krypton suggests that Pd(CO)2 may be more bent in the
more polar matrices. Such has been found for Fe(CO)2 in solid
-
1
8
,9
38
neon and argon.
The 2062.7 and 2074.6 cm- bands are assigned to Pd(CO)3
1
and Pd(CO)4 in solid neon are in good agreement with the argon
-
1
8,9
matrix 2070 and 2058 cm counterparts, and the apparent 5
Figure 6. Infrared spectra in the 2215-2200 and 2070-2040 cm-
regions for laser-ablated Pt co-deposited with 0.2% CO in neon at 4
K: (a) sample deposited for 40 min, (b) after 6 K annealing, (c) after
1
-
1
37,38
cm blue shift in neon is reasonable.
Pt(CO)n (n ) 1-4). The absorptions of Pt(CO)n (n ) 1-4)
are similar to those for Ni(CO)n and Pd(CO)n. The 2065.5 and
1
0 K annealing, (d) after 11 K annealing, (e) after 12 K annealing, (f)
-
1
2
064.7 cm bands are assigned to the CO stretching mode of
after 240-580 nm photolysis.
PtCO in two matrix sites. The absorptions at 2053.2, 2044.0,
-
1
and 2058.1 cm are assigned as antisymmetric modes of
12
13
Pt(CO)2, Pt(CO)3, and Pt(CO)4, respectively. The CO/ CO
isotopic ratios for the Pt(CO)n (n ) 1, 2, 3, 4) carbonyls are
1
.0251, 1.0226, 1.0231, and 1.0231, respectively.
The early thermal Pt argon matrix work was complicated by
congestion of all Pt(CO)n bands into a 10 cm- region;
1
nevertheless, all four species were assigned with varying weight
1
0
12
13
of evidence. With 5% CO and 5% CO, a strong doublet at
2
-
1
053.3 and 2006.9 cm
with three weaker intermediate
components was assigned to tetrahedral Pt(CO)4, and with more
-
1
dilute 0.4% CO the strongest feature at 2052.0 cm was
attributed to PtCO. Unfortunately, the 12CO/ CO ratios are not
accurate because of band overlap, which fortunately does not
happen in solid neon as shown in Figure 8. Our PtCO band at
13
Figure 7. Infrared spectra in the 1900-1830 cm- region for laser-
ablated Pt co-deposited with 0.2% CO in neon at 4 K: (a) sample
deposited for 35 min, (b) after 8 K annealing, (c) after λ > 470 nm
photolysis, (d) after λ > 380 nm photolysis, (e) after 240-580 nm
photolysis, (f) after 12 K annealing.
1
-
1
2
065.5, 2064.7 cm clearly forms a mixed isotopic doublet
12
13
for a single CO group, and the high CO/ CO ratio (1.0251)
is appropriate for this motion of C between Pt and O as observed
for RhCO and IrCO. (Manceron finds two sites for PtCO at
37
-
1
36
2
056.8, 2057.9 cm in solid argon.) Our Pt(CO)2 absorption
-
1
matrix 2064.7 cm-1 and argon matrix36 2056.8 cm-1 assign-
ments to PtCO are compatible, but our neon matrix data cast
doubt on the argon matrix identifications of Pt(CO)2,3.
forms a 1:2:1 triplet pattern at 2053.2, 2025.1, 2007.9 cm ,
which is only slightly asymmetric. The neon matrix band at
046.0, 2044.0 cm- is due to Pt(CO)3; it shows one weaker
1
10
2
intermediate component when two are expected for the anti-
symmetric mode of a trigonal species,35 but band congestion
may obscure another weak feature for mixed isotopic Pt(CO)3.
The symmetric stretching mode of Pt(CO)2 is not observed
in our experiments, which might simply mean this is a linear
molecule. However, the DFT calculation again supports a bent
structure for this molecule. The missing symmetric mode could
-
1
Our neon matrix band at 2058.1 cm shows three possible
weaker intermediate mixed isotopic features, as is expected for
the triply-degenerate mode,35 and is compatible with the argon
be due to the low intensity or masking by other dominant peaks.
8-10,36
Compared to thermal metal atom reactions,
laser
-
1
matrix Pt(CO)4 band at 2053.5 cm . We believe that the neon
ablation also introduces cations and electrons into the reaction

Reactions of Laser-Ablated Ni, Pd, and Pt Atoms with CO
J. Phys. Chem. A, Vol. 104, No. 17, 2000 3909
-
1
Figure 8. Infrared spectra in the 2070-1980 cm region for laser-
-1
Figure 9. Infrared spectra in the 1900-1790 cm region for laser-
1
2
13
ablated Pt co-deposited with 0.15% CO + 0.15% CO in neon at 4
K: (a) sample deposited for 45 min, (b) after 6 K annealing, (c) after
K annealing, (d) after 10 K annealing, (e) after 12 K annealing.
12
13
ablated Pt co-deposited with 0.15% CO + 0.15% CO in neon at 4
K: (a) sample deposited for 45 min, (b) after 8 K annealing, (c) after
8
1
0 K annealing, (d) after λ > 470 nm photolysis, (e) after 240-580
nm photolysis.
system. Additional weak bands in the low C-O stretching
-
1
12
13
region (<1900 cm ) are observed in all three (Ni, Pd, Pt +
CO) systems. These bands exhibit similar photosensitive
behavior and are almost eliminated in CCl4 doped experiments,
CO/ CO experiment, this band shows a characteristic 1:2:1
triplet feature. Therefore, this photosensitive band is assigned
-
to the antisymmetric C-O stretching mode of Pd(CO)2 . The
25,37,38
-1
so anion species must be considered.
broad feature around 1872 cm increases during annealing and
-
-
-
Ni(CO)n (n ) 1, 2, 3). Three sets of weak photosensitive
bands are observed on deposition at 1867.7 and 1864.7 cm ,
at 1860.6 cm , and at 1817.1, 1815.4, and 1813.8 cm . The
860.6 cm band does not change during 6 and 8 K annealings
increases on λ > 630 nm photolysis when PdCO and Pd(CO)2
disappear and decrease. However, shorter wavelength photolysis
-
1
-
1
-1
12
13
(510-630 nm) destroys this band. In the mixed CO/ CO
-1
-1
1
experiment, the 1872.2, 1870.1 cm band shows a quartet
feature which is appropriate for a trigonal species.2
5,35
Hence,
but severely decreases during l0 and 12 K annealings and is
destroyed with λ > 470 nm photolysis (Figure 2). The band
this band is assigned to the antisymmetric C-O vibrational
has 13C O and C O counterparts at 1815.6 and 1821.0 cm
16
12 18
-1
,
mode of Pd(CO)3 .
-
-
with isotopic ratios of 1.0248 and 1.0217, and shows a doublet
Pt(CO)n (n ) 1, 2, 3). Similar to Ni and Pd experiments,
feature in the mixed 12CO/ CO experiment. This photosensitive
13
three photosensitive bands are also observed in the low carbonyl
-
-1
band is assigned as NiCO . The 1817.1, 1815.4, and 1813.8
vibrational region. The 1896.3 cm band shows doublet feature
-
1
12
13
cm band increases during annealings and is also destroyed
in the mixed CO/ CO experiment (Figure 9), is destroyed by
1
3
16
12 18
by λ > 470 nm photolysis. In C O and C O experiments,
photolysis(λ > 380 nm), and is assigned as the C-O stretching
-
1
-
-1
this band shifts to 1774.8, 1773.2, and 1771.6 cm , and to
777.6, 1775.9, and 1774.3 cm , respectively. The average
isotopic ratios are 1.0238 ( CO/ CO) and 1.0222 (C O/C O).
Each peak in this band shows a triplet feature in the mixed
CO/ CO experiment. This photosensitive band is assigned
mode of PtCO . The 1844.0 cm absorption shows a sharp
-
1
12
13
1
1:2:1 feature in the mixed CO/ CO experiment and is assigned
as the antisymmetric C-O stretching mode of Pt(CO)2 . The
broad band near 1875 cm is assigned to Pt(CO)3 , this
assignment is supported by the quartet feature in the mixed
CO/ CO experiment,
12
13
16
18
-
-
1
-
1
2
13
-
12
13
25,35
as the C-O antisymmetric stretching mode of Ni(CO)2 . The
864.7 and 1867.7 cm band increases even more during
and it increases on λ > 470 nm
-
1
-
-
1
photolysis at the expense of PtCO and Pt(CO)2 .
annealings. However, λ > 470 nm photolysis does not destroy
this band, but shorter wavelength photolysis (240-580 nm) does
destroy it. This band has two weak intermediate components
in the mixed 12CO/ CO experiment, which is appropriate for
the degenerate stretching mode of a trigonal species,2 and it
Besides the anion bands in low CO vibrational region, some
weak, photosensitive bands are also observed in the high C-O
stretching region (>2140 cm ). In contrast to the anion species,
the absorption intensities of these bands increase in the CCl4
-1
13
5,35
37,38
doping experiments. Hence, cation species are considered.
-
+
-1
is assigned to this C-O stretching mode of Ni(CO)3 .
Ni(CO)n (n ) 1-4). The band at 2206.5 cm is observed
on deposition, increases during lower temperature annealings,
and then decreases during higher temperature annealings. The
2205.3 cm band is also observed on deposition with an
intensity weaker than the 2206.5 cm band, but after higher
These results are in excellent agreement with the recent argon
matrix work.25 The neon matrix blue-shifts the Ni(CO)n bands
-
-
1
-1
1
3.6, 12.1, and 5.9 cm , n ) 1, 2, and 3, respectively.
-
-1
Pd(CO)n (n ) 1, 2, 3). The nickel carbonyl anions pave the
way to identify the palladium carbonyl anions. Three photo-
sensitive bands are formed in the low CO vibrational region on
temperature annealings, this band is stronger than the 2206.5
-
1
-1
cm band. The 2192.5 cm and another almost indiscernible
-
1
-1
-1
deposition: 1909.0 cm , broad feature at 1872 cm (splits to
2186.2 cm band are observed on deposition. They track
-1
1
872.2 and 1870.1 cm on late annealings), and 1834.3, 1832.3
together in all experiments, and their intensities increase steadily
-
1
-1
-1
cm . The 1909.0 cm band decreases slightly on annealing,
during annealings. The fourth band at 2176.2 cm is weak on
and λ > 630 nm photolysis destroys this band (Figure 4d). In
deposition and it increases on all annealings. All of these four
sets of bands are destroyed by the medium-pressure mercury
lamp photolysis (240-580 nm) (Figure 1f). These bands are
the 13C O and C O experiments, this band shifts to 1864.6
16
12 18
-
1
and 1868.2 cm , respectively, and it shows a doublet feature
1
2
13
-1
in the mixed CO/ CO experiment. This band is assigned as
assigned as nickel carbonyl cations: 2206.5 cm is the C-O
-
+
-1
the C-O stretching mode of the PdCO anion. The 1834.3,
stretching mode of NiCO , 2205.3 cm is the antisymmetric
-
1
+
-1
1
832.3 cm band increases during annealings, and photolysis
C-O stretching mode of Ni(CO)2 , and 2192.5 and 2186.2 cm
are assigned to the b2 and a1 C-O vibrational modes of the
(
λ > 630 nm) decreases the band by half. The 12/13 and 16/18
+
-1
isotopic ratios of this band are 1.0239 and 1.0220. In the mixed
Ni(CO)3 cation with a C2V symmetry. The 2176.2 cm band

3
910 J. Phys. Chem. A, Vol. 104, No. 17, 2000
Liang et al.
is assigned to the e mode of the Ni(CO)4+ cation with a D2d
symmetry. The observed isotopic bands are listed in the Table
Other Absorptions. In Ni and Pd + CO reactions, several
extra bands are observed in the low C-O stretching region or
1
. The mono- and dicarbonyl cation assignments are supported
the metal carbonyl anion region. In the Ni + CO experiment,
12
13
-1
by the mixed CO/ CO experiment, which shows doublet
feature at 2206.5 and 2156.0 cm for NiCO , and an asym-
metric 2:3:2 triplet feature at 2205.3, 2165.9, and 2156.1 cm
with a weaker associated 2223.8 cm band for the symmetric
stretching mode of the Ni( CO)( CO) cation. The isotopic
counterparts of bands of tri- and tetracarbonyl cations are
difficult to identify, because some of the counterparts hide in
the CO and CO aggregate bands which dominate around 2140
cm . In the CCl4-doped experiments, the absorptions of these
nickel carbonyl cation bands increase while the anion bands
are not observed. The CCl4 molecule is well-known for its high
electron capture cross section and resulting dissociative capture
absorptions at 1870.5 and 1841.4 cm
are observed on
-
1
+
deposition, increase during annealing, and are not eliminated
by photolysis. When higher laser power is employed, the
absorption intensities of both bands increase, but metal carbonyl
anions decrease with higher laser power. Both bands show
-
1
-
1
12
13
+
1
3
18
isotopic shifts in CO and C O experiments, and both become
1
2
13
doublets in the mixed CO/ CO isotopic experiments. There-
fore, Ni cluster monocarbonyls are reasonable candidates for
-
1
-1
those absorptions. The 1841.4 cm absorption is stronger than
-
1
the 1870.5 cm absorption on deposition, but grows slower
-1
than the 1870.5 cm absorption on annealings. We are inclined
to assign tentatively the 1841.4 cm absorption as Ni2CO, and
the 1870.5 cm absorption as Ni3CO. The absorption at 1873.2
-1
-
-1
process giving CCl3 and Cl . Hence, the assignments of the
ions are supported by the CCl4-doped experiments.
the observation of NiCO absorption at 2205.3 cm
consistent with the argon matrix work, which gave a weak
37,38
-1
-1
Finally,
cm is a matrix site of the 1870.5 cm absorption. Similar to
the Ni + CO experiment, there are also two bands observed in
the Pd + CO experiment. On the basis of the similar behaviors
as the Ni2CO and Ni3CO in the Ni + CO experiment, the 1878.8
+
-1
is
25
-
1
-1
+
2
176.3 cm absorption, 30.2 cm lower, for NiCO .
-
1
and 1902.7 cm bands are tentatively assigned as Pd2CO and
Pd3CO.
Our DFT calculations are in very good agreement with the
observations of these cations (scale factors 0.96-0.97). It is
interesting to point out that the equilibrium geometries of
B. Calculations: Theoretical Considerations. The results
of DFT calculations are listed in Tables 4-8. We have
calculated different geometries and different spin multiplicities
for these species. Only the most stable ones (the lowest energy,
and no negative vibrational frequencies) are listed.
+
+
Ni(CO)3 and Ni(CO)4 are C2V and D2d instead of D3h and Td.
+
The observation of two C-O stretching modes for Ni(CO)3
supports the calculated C2V structure.
Pd(CO)n+ (n ) 1, 2). Two photosensitive bands are observed
-
1
in the high C-O vibrational region. The 2206.4 cm band
As a first step, reproducing the correct ground state of the
metal atoms is a difficult task. From our knowledge in general
chemistry, the ground state configuration of Ni atom is [Ar]-
observed on deposition increases during annealings and de-
1
3
16
12 18
creases on photolysis. This band has C O and C O
-1
-1
8
2
10
14
9
1
counterparts at 2156.8 and 2155.4 cm . The weak 2210.5 cm
band also grows on annealing and decreases on photolysis, and
C O and C O isotopic counterparts are at 2161.1 and
158.7 cm . In the mixed CO/ CO isotopic experiment, the
206.4 cm band shows a doublet feature, and the 2210.5 cm
band splits to a triplet with a fourth band at 2225.0 cm . The
206.4 cm band is assigned as the C-O stretching mode of
Pd(CO) , and the 2210.5 cm band is the antisymmetric C-O
3d 4s , Pd is [Kr]4d, and Pt is [Xe]4f 5d 6s . In our DFT
calculation (B3LYP functional, 6-311+G* basis set on C, O,
and Ni or LANL2DZ on Pd and Pt), the ground states of Pd
and Pt are reproduced, but Ni is calculated to have the [Ar]-
1
3
16
12 18
-
1
12
13
2
2
-1
-1
9
1
3d 4s ground state configuration, as found by other DFT
-
1
11,14
calculations.
For CO, the calculated bond length is 1.128 Å
-
1
-1
2
and the harmonic vibrational frequency is 2211.8 cm . These
results are in excellent agreement with the experimental values,
+
-1
40
+
-1
stretching mode of Pd(CO)2 . The band at 2225.0 cm in the
mixed isotopic experiment is the symmetric stretching mode
of Pd( CO)( CO) . In the CCl4-doped experiment, both bands
CO bond length, 1.128 Å, and the harmonic vibrational
-
1
frequency (ωe), 2169.8 cm .
12
13
+
The binding energies of the carbonyls are listed in Tables
4-6. All of those values are positive, which means that adding
CO to unsaturated carbonyls are exothermic reactions. During
deposition, excess CO is condensed in the neon matrix so excess
CO can react with unsaturated metal carbonyls to produce higher
order carbonyls if enough energy is available to overcome any
barriers. In our annealing experiments, the intensities of the
higher order carbonyl absorptions increase while the lower order
carbonyls decrease. These observations are well supported by
the calculations. The previous experimental bond energy data
increase while the anion bands are absent from the 1820-1920
-
1
37,38
cm region.
Pt(CO)n+ (n ) 1, 2, 3). Three bands are observed at 2210.3,
-
1
2
207.9, and 2204.7 cm in the high C-O vibrational region
-
1
on deposition. With 0.1% CO/Ne, the 2204.7 cm band is the
-1
strongest, but with 0.2% CO/Ne, the 2210.3 cm band becomes
the strongest (Figure 6). In the 0.2% CO experiments, the 2204.7
-
1
cm band shows only a small increase after 6 K annealing,
then no change after 10 K annealing, and finally decreases after
-
1
of (CO) Ni-CO (n ) 0, 1, 2) are 40.5 ( 5.8, 47.1 ( 5.8, and
1
1 and 12 K annealings. The 2210.3 cm band increases by
n
2
2
-
2
8.3 ( 2.3 kcal/mol, (CO)nNi -CO (n ) 1,2) are 43.4 ( 5.8
half after 6 K annealing, increases again after 10 K annealing,
but the intensity of this band decreases after 11 and 12 K
2
2
+
and 38.5 ( 2.3 kcal/mol, (CO)nNi -CO (n ) 1,2,3) are
2
3
-
1
40.4 ( 2.5, 22.6 ( 1.4, and 18.4 ( 0.7 kcal/mol. Since the
annealings. The 2207.9 cm band is the weakest band on
deposition, but it keeps increasing on annealings; after 12 K
annealing, this band is the strongest among the three. Photolysis
after the annealings almost destroys the three bands. The three
zero-point energy (ZPE) and basis set superposition error
(BSSE) only account for several kcal/mol, we have not done
these corrections. Nevertheless, our “raw” calculation results
-
1
for binding energies of (CO) Ni-CO (n ) 0, 1, 2) 32.9, 40.8,
bands at 2210.3, 2207.9, and 2204.7 cm are assigned as
n
-
+
+
+
and 27.4 kcal/mol, (CO) Ni -CO (n ) 1,2) 43.0 and 39.3 kcal/
Pt(CO)2 , Pt(CO)3 , and PtCO , respectively. Some of isotopic
counterparts and intermediate bands in mixed isotopic experi-
ments are not observable due to the dominant CO and aggregate
absorptions in this region. As expected, in the 0.02% CCl4-
doped experiment, these three bands increase by at least half
again while the anion products are absent.
n
+
mol, and (CO)nNi -CO (n ) 1,2,3) 37.5, 21.0, and 16.8 kcal/
mol are close (within 10%-20%) to the experimental data.
Electron affinities (EA) of monocarbonyls are also listed in
Tables 4-6. Combined with the binding energies, we can get
the EAs of Ni(CO)n (n ) 1, 2, 3) as 13.8, 16.0, and 27.9 kcal/

Reactions of Laser-Ablated Ni, Pd, and Pt Atoms with CO
J. Phys. Chem. A, Vol. 104, No. 17, 2000 3911
TABLE 4: Ground Electronic States, Equilibrium Geometries, Binding Energies, and Frequencies Calculated (B3LYP/
6
-311+G*) for the Ni Carbonyls
electronic point binding energy
a
frequencies (intensity)g
species
state
group
(kcal/mol)
geometry (Å, deg)
1
1
1
+
CO
NiCO
Ni(CO)
Σ
Σ
C
C
C
∞V
∞V
2V
C-O: 1.128
2211.8(89)
+
32.9b
40.8
Ni-C: 1.685, C-O: 1.151
Ni-C: 1.781, C-O: 1.141,
345.0(9 × 2), 585.9(5), 2083.6(745)
2
A
1
426.2(4), 450.6(0), 496.0(153), 2073.3(2137), 2162.7(72)
∠
CNiC: 150.7, ∠ NiCO: 173.7
1
1
2
2
2
2
2
2
Ni(CO)
Ni(CO)
NiCO
3
A
A
1
′
1
+
3h
27.4
20.3
-13.8
43.0
39.3
175.1
37.5
21.0
Ni-C: 1.819, C-O: 1.139
Ni-C: 1.846, C-O: 1.138
Ni-C: 1.691, C-O: 1.172
Ni-C: 1.766, C-O: 1.175
Ni-C: 1.784, C-O: 1.166
Ni-C: 1.896, C-O: 1.118
Ni-C: 1.931, C-O: 1.118
Ni-C′: 1.961, C′-O′: 1.121,
Ni-C: 1.954, C-O: 1.119,
484.0(5 × 2), 2094.5(1378 × 2), 2173.1(0)
464.3(0), 2113.3(1032 × 3), 2186.2(0)
4
-
c
Σ
224.4(87 × 2), 573.2(0), 1937.6(1066)
-
-
Ni(CO)
Ni(CO)
NiCO
2
Π
u
∞h
402.2(17), 448.5(0), 458.5(1), 1867.3(2960), 1936.0(0)
514.0(9 × 2), 1913.4(2266 × 2), 1997.7(0)
294.5(0 × 2), 378.0(3), 2294.3(178)
3
2
A ′′
3h
+
+
d
Σ
+
+
+
Ni(CO)
Ni(CO)
2
Σ
g
379.4(28), 405.6(3 × 2), 2279.6(545), 2308.5(0)
387.7(4), 400.0(8), 2260.7(296), 2265.1(538), 2292.8(1)
3
A
1
∠
CNiC: 154.9, ∠ NiCO: 179.9e
+
2B
Ni(CO)
4
2
2d
16.8
Ni-C: 1.982, C-O: 1.121,
408.2(0), 2247.8(280), 2251.1(434 × 2), 2280.0(0)
f
∠
CNiC: 129.0, ∠ NiCO: 175.8
a
Binding energies refer to the reactions: Ni(CO)n-1 + CO f Ni(CO) q
q
, without ZPE and BSSE corrections. b The ground state of Ni atom is
n
9
1
c
d
calculated as [Ar]3d 4s configuration. Relative energy compared to NiCO (-EA of NiCO). Relative energy compared to NiCO (IE of NiCO).
C′ and O′ are on the C
e
f
g
2
axis. C and O refer to others. These two carbon atoms are in one dihedral plane. Highest frequencies listed, no negative
roots for any frequencies.
TABLE 5: Ground Electronic States, Equilibrium Geometries, Binding Energies, and Frequencies Calculated for the Pd
Carbonylsa
electronic point binding energy
b
frequencies (intensity)g
species
PdCO
state
group
(kcal/mol)
geometry (Å, deg)
1
+
Σ
C
C
∞V
2V
41.3
29.7
Pd-C: 1.873, C-O: 1.143
Pd-C: 1.965, C-O: 1.135,
242.4(7 × 2), 445.3(12), 2112.9(629)
1
Pd(CO)
2
A
1
321.0(3), 366.0(130), 383.0(0), 2121.1(1826), 2190.3(28)
∠
CPdC: 162.5, ∠ PdCO: 176.2
1
1
2
Pd(CO)
Pd(CO)
PdCO
3
A
A
A′
1
′
D
3h
11.6
4.7
-10.8
Pd-C: 2.020, C-O: 1.135
Pd-C: 2.070, C-O: 1.134
Pd-C: 1.873, C-O: 1.168,
413.6(2 × 2), 2120.2(1184 × 2), 2174.9(0)
369.2(2), 2129.9(901 × 3), 2183.0(0)
203.9(143), 478.4(38), 1914.2(1913)
4
1
T
d
-
c
C
s
∠
PdCO: 154.3
-
-
2
2
2
2
2
Pd(CO)
Pd(CO)
PdCO
2
Π
u
D
D
C
D
C
∞h
32.9
19.3
201.5
32.6
15.2
Pd-C: 1.935, C-O: 1.169
Pd-C: 1.976, C-O: 1.161
Pd-C: 1.996, C-O: 1.118
Pd-C: 2.061, C-O: 1.117
Pd-C: 2.088, C-O: 1.119,
Pd-C′: 2.138, C′-O′: 1.120
345.5(11), 347.4(17), 408.0(0), 1887.5(3436), 1943.4(0)
448.5(17 × 2), 1921.5(2523 × 2), 1990.2(0)
282.4(0 × 2), 327.3(4), 2287.4(174)
3
A
2
′′
3h
+
+
d
Σ
∞V
∞h
2V
+
+
+
Pd(CO)
Pd(CO)
2
Σ
g
301.7(0), 398.1(5 × 2), 2285.6(431), 2311.3(0)
337.5(3), 384.4(7), 2265.7(203), 2267.8(478), 2291.3(0)
3
A
1
∠
CPdC: 166.1, ∠ PdCO: 178.5e
+
2B
Pd(CO)
4
2
D
2d
10.5
Pd-C: 2.174, C-O: 1.121,
∠
324.3(0), 2253.9(53), 2254.4(427 × 2), 2276.9(0)
f
CPdC: 154.5, ∠ PdCO: 178.4
a
DFT calculation: B3LYP functional, 6-311+G* basis set on C and O, LANL2DZ basis set and ECP on Pd. The ground state of Pd atom is
10
0
b
q
q
n
calculated as [Kr]4d 4s configuration. Binding energies refer to the reactions: Pd(CO)n-1 + CO f Pd(CO) , without ZPE and BSSE corrections.
c
d
e
2
Relative energy compared to PdCO (-EA of PdCO). Relative energy compared to PdCO (IE of PdCO). C′ and O′ are on the C axis. C and
f
g
O refer to others. These two carbon atoms are in one dihedral plane. Highest frequencies listed, no negative frequencies.
mol. The experimental data are 18.5 ( 0.3, 14.8 ( 0.3, and
4.8 ( 0.3 kcal/mol.20 Similarly, the ionization energies (IE)
are higher than anions. The calculated frequencies as well as
experimental values have shown this trend. The detailed
comparison between calculated and experimental values is listed
in Table 7. The average of scale factors is 0.971 ( 0.007. The
B3LYP functional is well-known for higher frequencies, and
the scale factor for this method is generally around 0.95-1.00.41
On the other hand, the calculated harmonic frequencies should
be larger than the experimental fundamental modes which
include anharmonicity factors. Matrix shifts contribute to
the difference between calculated and experimental values. If
we take CO as our standard, which gives scale factor 0.968,
the average scale factor of the carbonyl complexes is quite
acceptable.
2
of the Ni(CO)n (n ) 1-4) are calculated as 175.1, 178.4, 184.8,
and 188.3 kcal/mol, and the experimental data are 168.3 ( 6.7,
2
2
1
79.6 ( 5.1, 177.3 ( 5.8, and 188.4 ( 5.3 kcal/mol. Both
EA and IE values are close to the experimental measurements.
Recently, the EA of PdCO has been deduced in anion photo-
electron spectroscopy,21 and our value (10.8 kcal/mol) is close
to the experimental value (13.9 ( 0.2 kcal/mol). Although there
is concern about the accuracy of DFT anion energies for the
39
calculation of electron affinity, the present B3LYP calculations
for EA’s and IE’s of Ni(CO)n are within 10% of experimental
measurements except for the EA of NiCO (25% too low).
The C-O stretching modes largely depend on the strengths
of the C-O bonds. The bond lengths of C-O units can be used
to reflect the strengths of the C-O bonds. In all these three
metal carbonyls, the cations have shorter C-O bonds than
neutrals, while the neutrals have shorter C-O bonds than anions.
As a result, the C-O vibrational frequencies (stretching modes)
of the cations are higher than neutrals, and those of neutrals
The isotopic frequency ratios of these C-O vibrational modes
are well-defined by the calculations (Table 7). They give strong
support in assigning these bands, especially when the C-O
stretching vibrations of Pd and Pt carbonyl neutrals are very
congested.
In the case of CO bonded to cation center, such “nonclassical”
carbonyls with frequencies higher than CO itself have been

3912 J. Phys. Chem. A, Vol. 104, No. 17, 2000
Liang et al.
TABLE 6: Ground Electronic States, Equilibrium Geometries, Binding Energies, and Frequencies Calculated for the Pt
Carbonylsa
electronic point binding energy
b
frequencies (intensity)g
species
PtCO
state
group
(kcal/mol)
geometry (Å, deg)
1
Σ+
1
C
C
∞V
2V
66.6
43.5
Pt-C: 1.781, C-O: 1.148
Pt-C: 1.909, C-O: 1.138,
404.7(2 × 2), 590.3(5), 2121.4(659)
1
Pt(CO)
2
A
452.4(0), 463.3(0), 520.5(0), 2108.9(1934), 2195.4(64)
∠
CPtC: 157.1, ∠ PtCO: 171.5
1
1
2
Pt(CO)
Pt(CO)
PtCO
3
A
A
A′
1
′
D
3h
12.1
3.7
-23.6
Pt-C: 1.956, C-O: 1.139
Pt-C: 1.996, C-O: 1.137
Pt-C: 1.820, C-O: 1.172,
507.9(14 × 2), 2104.2(1299 × 2), 2165.9(0)
472.6(18), 2114.5(1013 × 3), 2174.2(0)
191.4(137), 547.8(26), 1935.4(1132)
4
1
T
d
-
c
C
s
∠
PtCO: 165.9
-
-
2
2
2
2
2
Pt(CO)
Pt(CO)
PtCO
2
Π
u
D
D
C
D
C
∞h
41.5
20.7
216.5
47.9
24.2
Pt-C: 1.901, C-O: 1.172
Pt-C: 1.923, C-O: 1.165
Pt-C: 1.861, C-O: 1.124
Pt-C: 1.984, C-O: 1.119
Pt-C′: 2.004, C′-O′: 1.122,
Pt-C: 1.998, C-O: 1.121,
406.0(0), 489.4(1), 501.5(0), 1896.1(2898), 1975.5(0)
533.0(46 × 2), 1925.9(2135 × 2), 1994.5(0)
383.3(0 × 2), 492.1(18), 2262.2(321)
3
A
2
′′
3h
+
+
d
Σ
∞V
∞h
2V
+
+
+
Pt(CO)
Pt(CO)
2
Σ
g
421.1(0), 504.7(12 × 2), 2262.2(744), 2315.4(0)
436.7(8), 494.8(15), 2243.1(790), 2245.2(350), 2286.1(3)
3
A
1
∠
CPtC: 163.0, ∠ PtCO: 178.2 e
+
2B
Pt(CO)
4
2
D
2d
11.4
Pt-C: 2.030, C-O: 1.124,
433.4(14), 2212.5(870 × 2), 2217.0(99), 2252.9(0)
f
∠
CPtC: 153.3, ∠ PtCO: 173.6
a
DFT calculation: B3LYP functional, 6-311+G* basis set on C and O, LANL2DZ basis set and ECP on Pt. The ground state of Pt atom
1
4
9
1
b
q
q
n
is calculated as [Xe]4f 5d 6s configuration. Binding energies refer to the reactions: Pt(CO)n-1 + CO f Pt(CO) , without ZPE and BSSE
c
d
e
corrections. Relative energy compared to PtCO (-EA of PtCO). Relative energy compared to PtCO (IE of PtCO). C′ and O′ are on
f
g
2
the C axis. C and O refer to others. These two carbon atoms are in one dihedral plane. Highest frequencies listed, no negative roots for any
frequencies.
TABLE 7: Comparison of Computed and Experimental C-O Stretching Modes of the Ni Group Carbonylsa
computed
experimental
stretching
mode
scale factor
(exp/comp)
species
NiCO
freq (cm-1)
R(12/13)
R(16/18)
freq (cm-1
)
R(12/13)
R(16/18)
σ
2
2083.6
2073.3
2162.7
2094.5
2113.3
1937.6
1867.3
1913.4
2294.3
2279.6
2260.7
2265.1
2247.8
2112.9
2121.1
2120.2
2129.9
1914.2
1887.5
1921.5
2287.4
2285.6
2267.8
2254.4
2121.4
2108.9
2104.2
2114.5
1935.4
1896.1
1925.9
2262.2
2262.2
2243.1
2212.5
2211.8
1.0248
1.0232
1.0244
1.0233
1.0233
1.0250
1.0242
1.0244
1.0234
1.0231
1.0231
1.0230
1.0230
1.0239
1.0229
1.0230
1.0229
1.0239
1.0240
1.0238
1.0232
1.0230
1.0230
1.0229
1.0253
1.0232
1.0234
1.0233
1.0251
1.0243
1.0244
1.0241
1.0230
1.0231
1.0230
1.0228
1.0217
1.0241
1.0224
1.0239
1.0239
1.0212
1.0226
1.0227
1.0238
1.0243
1.0243
1.0244
1.0245
1.0230
1.0246
1.0244
1.0245
1.0231
1.0228
1.0233
1.0240
1.0244
1.0245
1.0246
1.0209
1.0240
1.0238
1.0240
1.0213
1.0225
1.0224
1.0227
1.0244
1.0243
1.0244
1.0248
2006.6
1978.9
2089.7
2025.0
2056.3
1860.6
1813.8
1864.7
2206.5
2205.3
2186.2
2192.4
2176.2
2056.4
2065.8
2062.7
2074.6
1909.0
1832.3
1872.2
2206.4
2210.5
1.0248
1.0228
1.0242
1.0229
1.0231
1.0248
1.0238
1.0240
1.0230
1.0228
1.0210
1.0238
1.0217
1.0236
1.0234
1.0217
1.0223
1.0226
1.0234
1.0239
1.0239
0.963
0.955
0.966
0.967
0.973
0.960
0.971
0.975
0.962
0.967
0.967
0.968
0.968
0.973
0.974
0.973
0.974
0.997
0.971
0.974
0.965
0.967
Ni(CO)
2
b
a
1
Ni(CO)
Ni(CO)
NiCO
3
e′
4
t
2
-
σ
-
-
Ni(CO)
Ni(CO)
NiCO
2
σ
u
3
e′
+
σ
σ
+
+
Ni(CO)
Ni(CO)
2
u
3
a
b
e
1
2
+
Ni(CO)
PdCO
4
σ
1.0238
1.0224
1.0228
1.0228
1.0238
1.0239
1.0235
1.0230
1.0228
1.0224
1.0244
1.0240
1.0240
1.0218
1.0220
1.0226
1.0237
1.0240
Pd(CO)
Pd(CO)
Pd(CO)
2
b
2
3
4
e′
t
2
-
PdCO
a′
-
-
Pd(CO)
Pd(CO)
PdCO
2
σ
e′
σ
u
3
+
+
+
+
Pd(CO)
Pd(CO)
Pd(CO)
PtCO
2
3
4
σ
2
u
b
e
σ
2065.5
2053.2
2044.0
2058.1
1896.3
1844.0
1875.8
2204.7
2210.3
2207.9
1.0251
1.0226
1.0231
1.0231
1.0250
1.0238
1.0241
1.0203
1.0242
1.0235
1.0235
1.0202
1.0196
1.0218
1.0223
1.0238
1.0237
0.974
0.974
0.971
0.973
0.980
0.973
0.974
0.975
0.977
0.984
Pt(CO)
Pt(CO)
Pt(CO)
2
b
2
3
e′
t
2
a′
4
-
PtCO
-
Pt(CO)
Pt(CO)
PtCO
2
σ
e′
σ
u
-
3
+
+
+
+
Pt(CO)
Pt(CO)
Pt(CO)
CO
2
3
4
σ
2
u
b
e
1.0229
1.0227
σ
2141.1
1.0246
0.968
a
There is no single negative mode in the vibrational frequencies calculated for all of these molecules.
explained by electrostatic effects. The agreement of NiCO+,
4
2
The bonding properties of monocarbonyls always receive
extra attention because they are considered as the models of
PdCO , and PtCO frequencies within 2 cm-1 shows that
electrostatic effects dominate here, even more so than predicted
by the B3LYP calculations, which suggest a 32 cm spread
for these carbonyl cation species.
+
+
1
3-17
the CO binding to the metal surface.
all the monocarbonyls are linear, except PdCO and PtCO
which are calculated to be bent by both DFT/B3LYP and BP86
In our calculations,
-
1
-
-

Reactions of Laser-Ablated Ni, Pd, and Pt Atoms with CO
J. Phys. Chem. A, Vol. 104, No. 17, 2000 3913
TABLE 8: Natural Bonding Orbital Analyses of Ni Group Monocarbonyls
species
conformn of Ni
BO of CO
species
conformn of Pd
BO of CO
species
PtCO
conformn of Pt
BO of CO
3d9 4s
.37
0.53
2.80
2.95
2.73
PdCO
PdCO
4d 5s
9.53 0.40
2.84
2.94
2.76
5d 6s
9.17 0.91
2.77
2.87
2.67
NiCO
NiCO
+
8.96 0.13
+
9.05 0.12
+
8.89 0.36
3d 4s
4d 5s
PtCO
5d 6s
-
9.25 1.21
-
9.40 1.06
-
9.12 1.35
NiCO
3d 4s
PdCO
4d 5s
PtCO
5d 6s
functionals. The description of the bonding of CO to transition
metals usually involves the mechanism of σ donation and π
back-donation. To address this bonding mechanism, we have
applied natural bonding orbital (NBO) analysis34 to the mono-
carbonyls, and results are listed in Table 8.
PES measurement.20 The uncorrected sequential binding ener-
gies of metal carbonyls agree well with the collision-induced
dissociation data. The NBO analyses have also been done on
all nine monocarbonyls; the only two nonlinear monocarbonyls
-
-
are PdCO and PtCO .
The PtCO complex observed here in solid neon at 2065 cm-
1
First, the cations have higher C-O bond orders than neutrals,
and neutrals have higher C-O bond orders than anions. Again,
this shows us that C-O stretching frequencies have the order
cations > neutrals > anions, as found for other transition
is just below value for low-coverage top site CO on Pt(111)
-
1
measured by ELS (258 meV ) 2080 cm ) and very low
-
1 43,44
coverage at 20 K by RAIRS at 2089 cm .
The RhCO and
3
7,38
37
metals.
IrCO frequencies in solid neon agreed with the low-coverage
-
1
Second, the different configurations of metal atoms in anions,
cations, or neutrals can be used to explain the different
geometries. Anions have the largest s electron component, while
neutrals and cations have smaller s character. The σ donation
in carbonyls is always explained as the 5σ electrons of CO
molecule donating to valence s orbital (or sdσ hybridized orbital)
of metal atom, while π back-donation is the valence dπ electrons
of metal atom donating to 2π* orbital of CO. The electrons in
valence s orbital (or sdσ hybridized orbital) of metal atom can
substantially decrease the M-CO bond strength because of the
repulsion between the metal s electrons and the CO 5σ electrons.
To relax this repulsion, it is energy wise for metal atom to
undergo s to d electron promotion. In the NBO analyses, this
promotion is confirmed by a comparatively smaller s component,
but a larger d component in the metal valence electron
configurations. In carbonyl anions, the metal atom has larger
valence s character than in the neutrals and cations, and hence,
the repulsive forces between metal s electrons and CO 5σ
electrons are larger. Because the 5σ orbital of CO does not point
directly toward the metal in bent geometry, the σ repulsion is
smaller. This repulsion probably acts as a driving force for
bending in the Pd and Pt carbonyl anions. The geometric
top-site measurements within 10 cm .
Catalyst systems with Ni(II) have provided carbonyl absorp-
-
1
tions in the 2204-2191 cm region attributed to Ni(II)-CO
4
5
on AlPO4, Al2O3, and SiO2 supports. The present observation
+
-1
of NiCO at 2206.5 cm provides strong evidence that the
local charge on these supported Ni(II) catalyst systems is +1.0
and not near the +2.0 implied by the oxidation state. The
average CO stretching frequencies of Pd(CO)2(SO3F)2 and
-
1 46,47
Pt(CO)2(SO3F)2 were reported as 2217 and 2201 cm .
Our
observations of neon matrix-isolated Pd(CO)2 and Pt(CO)2+
+
-1
at 2210.5 and 2210.3 cm strongly suggest that the local charge
on these solvated dicarbonyls is near +1 instead of +2 that
might be inferred from the II oxidation state in their molecular
formulas.
Acknowledgment. The authors gratefully acknowledge
National Science Foundation support from Grant CHE 97-00116
and a preprint from L. Manceron.
References and Notes
(1) Mond, L.; Langer, C.; Quincke, F. J. Chem. Soc. 1890, 749. Mond,
L.; Langer, C. J. Chem. Soc. 1891, 1090.
-
differences between linear NiCO and the other two carbonyl
(2) Cotton, F. A.; Wilkinson, G. AdVanced Inorganic Chemistry, 5th
ed.; Wiley: New York, 1988. Collman, J. P.; Hegedus, L. S.; Norton, J.
R.; Finke, R. G. Principles and Applications of Organotransition Metal
Chemistry; University Science Books: Mill Valley, CA, 1987.
anions are not easy to understand because in our NBO analysis,
Ni atom shows similar larger valence s component as Pd and
Pt atoms in their monocarbonyl anions. Additional theoretical
calculations are necessary to clarify this problem.
(
3) Tolman, C. A. Chem. Soc. ReV. 1972, 1, 337. Casey, C. P.; Cyr,
C. R. J. Am. Chem. Soc. 1973, 95, 2248. Mitchener, J. C.; Wrighton, M. S.
J. Am. Chem. Soc. 1981, 103, 975. Whetten, R. L.; Fu, K.; Grant, E. R. J.
Am. Chem. Soc. 1982, 104, 4270. Weitz, E. J. Phys. Chem. 1987, 91, 3945.
IV. Conclusions
(4) Walsh, S. P.; Goddard, W. A., III J. Am. Chem. Soc. 1976, 98,
7
908.
Laser-ablated Ni, Pd, and Pt atoms have been reacted with
CO molecules during condensation in a neon matrix at 4 K.
Through annealing, photolysis, and isotopic substitution experi-
(
(
5) Nekrylova, J. V.; Harrison, I. J. Chem. Phys. 1994, 101, 1730.
6) DeKock, R. L. Inorg. Chem. 1971, 10, 1205. Bach, S. B. H.; Taylor,
C. A.; Van Zee, R. J.; Vala, M. T.; Weltner, W., Jr. J. Am. Chem. Soc.
1986, 108, 7104. Joly, H. A.; Manceron, L. Chem. Phys. 1998, 226, 61.
(7) Howard, I. A.; Pratt, G. W.; Johnson, K. H.; Dresselhaus, G. J.
Chem. Phys. 1981, 74, 3415. Blomberg, M.; Brandemark, U.; Johanson,
J.; Siegbahn, P.; Wennerberg, J. J. Chem. Phys. 1988, 88, 4324. Blomberg,
M.; Siegbahn, P.; Lee, T. J.; Rendell, A. P.; Rice, J. E. J. Chem. Phys.
-
ments, metal carbonyl anions [M(CO)n (n ) 1-3)] and cations
+
+
+
[
(
Ni(CO)n (n ) 1-4); Pd(CO)n (n ) 1,2); Pt(CO)n
n ) 1-3)], as well as neutrals [M(CO)n (n ) 1-4)] have been
identified. Evidence for nonlinear geometries of dicarbonyls
1
991, 95, 5898. Sodupe, M.; Bauschlicher, C. W., Jr.; Lee, T. J. Chem.
Phys. Lett. 1992, 189, 266.
8) K u¨ ndig, E. P.; Moskovits, M.; Ozin, G. A. Can. J. Chem. 1972,
50, 3587.
(9) Darling, J. H.; Ogden, J. S. J. Chem. Soc., Dalton Trans. 1973,
079.
10) K u¨ ndig, E. P.; McIntosh, D.; Moskovits, M.; Ozin, G. A. J. Am.
[
M(CO)2] is provided. The CCl4-doped experiments are used
25,37,38
to confirm the identification of the ionic species.
Tentative
(
assignments for Ni and Pd cluster monocarbonyls [MxCO,
(x ) 2, 3; M ) Ni, Pd)] are also made.
1
The excellent agreement between experiment results and the
(
frequencies and isotopic frequency ratios from the DFT calcula-
tions (B3LYP functional, 6-311+G*/LANL2DZ basis sets)
supports the vibrational assignments and the identification of
these metal carbonyl complexes. The average scale factor for
the C-O vibrational modes is 0.971 ( 0.007. This scale factor
times the calculated symmetric mode for Ni(CO)3 predicts 2110
Chem. Soc. 1973, 95, 7234.
(11) Adamo, C.; Lelj, F. J. Chem. Phys. 1995, 103, 10605.
(12) Sosa, R. M.; Gardiol, P.; Beltrame, G. Int. J. Quantum Chem. 1998,
69, 371.
(
(
(
13) Fournier, R. J. Chem. Phys. 1993, 99, 1801.
14) Fournier, R. J. Chem. Phys. 1993, 98, 8041.
15) Xu, X.; L u¨ , X.; Wang, N.; Zhang, Q.; Ehara, M.; Nakatsuji, H.
-1
-1
cm , which is in excellent agreement with the 2100 ( 80 cm
Int. J. Quantum Chem. 1999, 72, 221.

3914 J. Phys. Chem. A, Vol. 104, No. 17, 2000
Liang et al.
(
16) P a´ pai, I.; Goursot, A.; St-Amant, A.; Salahub, D. R. Theor. Chim.
(31) Lee, C.; Yang, E.; Parr, R. G. Phys. ReV. B 1988, 37, 785.
Acta 1992, 84, 217.
(32) McLean, A. D.; Chandler, G. S. J. Chem. Phys. 1980, 72, 5639.
Wachters, A. J. H. J. Chem. Phys. 1970, 52, 1033. Hay, P. J. J. Chem.
Phys. 1977, 66, 4377. K. Raghavachari, K.; Trucks, G. W. J. Chem. Phys.
1989, 91, 1062.
(17) Basch, H.; Cohen, D. J. Am. Chem. Soc. 1983, 105, 3856.
(18) Jonas, V.; Thiel, W. J. Chem. Phys. 1995, 102, 8474.
(19) Jonas, V.; Thiel, W. J. Phys. Chem. A 1999, 103, 1381.
(20) Stevens, A. E.; Feigerle, C. S.; Lineberger, W. C. J. Am. Chem.
(33) Hay, P. J.; Wadt, W. R. J. Chem. Phys. 1985, 82, 270. Wadt, W.
R.; Hay, P. J. J. Chem. Phys. 1985, 82, 284. Hay, P. J.; Wadt, W. R. J.
Chem. Phys. 1985, 82, 299.
(34) Glendening, E. D.; Reed, A. E.; Carpenter, J. E.; Weinhold, F.
NBO Version 3.1.
(35) Darling, J. H.; Ogden, J. S. J. Chem. Soc., Dalton Trans. 1972, 2496.
(36) Manceron, L.; Alikhani, M. E. Chem. Phys. 1999, 244, 215.
Manceron, L.; Tremblay, B.; Alikhani, M. E. J. Phys. Chem. A 2000, in
press.
(37) Zhou, M.; Andrews, L. J. Phys. Chem. A 1999, 103, 7773
(Co, Rh, Ir + CO).
(38) Zhou, M.; Andrews, L. J. Chem. Phys. 1999, 110, 10370 (Fe + CO).
(39) Galbraith, J. M.; Schaefer, H. F. III J. Chem. Phys. 1996, 105,
862.
(40) Huber, K. P.; Herzberg, G. Constants of Diatomic Molecules; Van
Nostrand-Reinhold: New York, 1979.
(41) Bytheway, I.; Wong, M. W. Chem. Phys. Lett. 1998, 282, 219.
(42) Goldman, A. S.; Krogh-Jespersen, K. J. Am. Chem. Soc. 1996, 118,
12159.
Soc. 1982, 104, 5026.
21) Klopcic, S. A.; Moravec, V. D.; Jarrold, C. C. J. Chem. Phys. 1999,
10, 8986.
(
1
1
9
1
(22) Sunderlin, L. S.; Wang, D.; Squires, R. R. J. Am. Chem. Soc. 1992,
14, 2788.
(23) Khan, F. A.; Steele, D. L.; Armentrout, P. B. J. Phys. Chem. 1995,
9, 7819.
(24) Barnes, L. A.; Rosi, M.; Bauschlicher, C. W., Jr. J. Chem. Phys.
990, 93, 609.
(25) Zhou, M.; Andrews, L. J. Am. Chem. Soc. 1998, 120, 11499. Further
+
-1
work has identified NiCO at 2176.3 cm in solid argon. See references
for previous argon matrix work on Ni(CO)n.
(
26) Chertihin, G. V.; Saffel, W.; Yustein, J. T.; Andrews, L. Neurock,
M.; Ricca, A.; Bauschlicher, C. W., Jr. J. Phys. Chem. 1996, 100, 5261.
27) Andrews, L.; Chertihin, G. V.; Neurock, M. J. Phys. Chem. 1995,
00, 14609.
(
1
(28) Burkholder, T. R.; Andrews, L. J. Chem. Phys. 1991, 95, 8697.
(29) Hassanzadeh, P.; Andrews, L. J. Phys. Chem. 1992, 96, 9177.
(30) Gaussian 94, Revision B.1; Frisch, M. J.; Trucks, G. W.; Schlegel,
(43) Froitzheim, H.; Hopster, H.; Ibach, H.; Lehwald, S. Appl. Phys.
1977, 13, 147.
(44) Nekrylova, J. V.; French, C.; Artsynkhovich, A. N.; Ukrainkev,
V. A.; Harrison, I. Surf. Sci. Lett. 1993, 295, L987.
(45) Lindblad, T.; Rebenstorf, B. Acta Chem. Scand. 1991, 45, 342.
(46) Hwang, G.; Wang, C.; Bodenbinder, M.; Willner, H.; Aubke, F.
J. Fluorine Chem. 1994, 66, 150.
H. B.; Gill, P. M. W.; Johnson, B. G.; Robb, M. A.; Cheeseman, J. R.;
Keith, T. A.; Petersson, G. A.; Montgomery, J. A.; Raghavachari, K.; Al-
Laham, M. A.; Zakrzewski, V. G.; Ortiz, J. V.; Foresman, J. B.; Cioslowki,
J.; Stefanov, B. B.; Nanayakkara, A.; Challacombe, M.; Peng, C. Y.; Ayala,
P. Y.; Chen, W.; Wong, M. W.; Andres, J. L.; Replogle, E. S.; Gomperts,
R.; Martin, R, L.; Fox, D. J.; Binkley, J. S.; Defrees, D. J.; Baker, J.; Stewart,
J. P.; Head-Gordon, M.; Gonzalez, C.; Pople, J. A. Gaussian, Inc.:
Pittsburgh, PA, 1995.
(47) Wang, C.; Willner, H.; Bodenbinder, M.; Batchelor, R. J.; Einstein,
F. W. B.; Aubke, F. Inorg. Chem. 1994, 33, 3521.