Probing individual steps of dynamic exchange with 31P EXSY NMR spectroscopy: Synthesis and characterization of the [E7PtH(PPh3)]2- Zintl ion complexes [E = P, As]

Article abstract of DOI:10.1021/ja000988x

Ethylenediamine (en) solutions of E₇^3- (E = P, As) react with Pt(PPh₃)₂(C₂H₄) in the presence of 4,7,13,16,21,24-hexaoxa-1,10-diazobicyclo[8.8.8]hexacosane (2,2,2-crypt) to give the transition-metal Zintl ion complexes [P₇PtH(PPh₃)]^2- (1) and [As₇PtH(PPh₃)]^2- (2) as isomorphic [K(2,2,2-crypt)]⁺ salts, respectively. The hydride ligand originates from the solvent, and deuterium labeling studies show that the order of hydrogen sources follows the sequence en > CH₃CN > DMSO > DMF in the synthesis of 1. The [E₇PtH(PPh₃)]^2- anions have C₁ point symmetry with unperturbed nortricyclane-like E₇^3- ligands bound η² to square planar Pt(II)H(PPh₃)⁺ centers. Complex 1 is dynamic in solution whereby the PtH(PPh₃)⁺ fragment shifts from the top to bottom of the P₇ cluster (an η² → η⁴ → η² shift) and spins in the P₄ face in a concerted process to give apparent C(2v) symmetry. We have shown that ³¹P EXSY NMR experiments at -78 °C permit the separation of the two processes and show that shifting is ~20 times faster than spinning (k(shift) = 160 s^-1, k(spin) = 8 s^-1).

Full text of DOI:10.1021/ja000988x

J. Am. Chem. Soc. 2000, 122, 11101-11107
11101
31
Probing Individual Steps of Dynamic Exchange with P EXSY NMR
Spectroscopy: Synthesis and Characterization of the [E7PtH(PPh3)]2-
Zintl Ion Complexes [E ) P, As]
Banu Kesanli,† Scott Charles,† Yiu-Fai Lam,† Simon G. Bott,‡ James Fettinger,† and
Bryan Eichhorn*,†
Contribution from the Department of Chemistry and Biochemistry, UniVersity of Maryland,
College Park, Maryland 20742, and Department of Chemistry, UniVersity of Houston, UniVersity Park,
Houston, Texas 77204
ReceiVed March 20, 2000
3-
Abstract: Ethylenediamine (en) solutions of E7 (E ) P, As) react with Pt(PPh3)2(C2H4) in the presence of
4,7,13,16,21,24-hexaoxa-1,10-diazobicyclo[8.8.8]hexacosane (2,2,2-crypt) to give the transition-metal Zintl
ion complexes [P7PtH(PPh3)]2- (1) and [As7PtH(PPh3)]2- (2) as isomorphic [K(2,2,2-crypt)]+ salts, respectively.
The hydride ligand originates from the solvent, and deuterium labeling studies show that the order of hydrogen
sources follows the sequence en > CH3CN > DMSO > DMF in the synthesis of 1. The [E7PtH(PPh3)]2-
anions have C1 point symmetry with unperturbed nortricyclane-like E7 ligands bound η2 to square planar
3-
PtIIH(PPh3)+ centers. Complex 1 is dynamic in solution whereby the PtH(PPh3)+ fragment shifts from the top
to bottom of the P7 cluster (an η2 f η4 f η2 shift) and spins in the P4 face in a concerted process to give
apparent C2V symmetry. We have shown that 31P EXSY NMR experiments at -78 °C permit the separation
of the two processes and show that shifting is ∼20 times faster than spinning (kshift ) 160 s-1, kspin ) 8 s-1).
Introduction
of nuclei such as 13C,6 31P,3,7,8 119Sn,9 and 131Xe10 to name a
few.3
For the past few decades, variable-temperature one-dimen-
sional NMR spectroscopy has been the technique most fre-
quently used to study dynamic processes in solution.1,2 In the
simplest case, the free energy of activation (∆G‡) can be
extracted from the coalescence temperature in a two-site
exchange. Recent developments in band-shape analysis allow
for quantitative analysis of rate constants at temperatures below
coalescence, provided the rate of exchange is commensurate
with the chemical shift time scale of the experiment.3,4 Exchange
processes that are slow on the chemical shift time scale or
systems that are difficult to model using band-shape methods
(i.e., second-order spin systems) require more sophisticated
saturation-transfer techniques for analysis of dynamic behav-
ior.3,5 However, these 1D methods are essentially limited to two-
site exchange systems.
In the past 10 years, two-dimensional exchange spectroscopy
(2D EXSY) has become the method of choice to study multisite
exchange and dynamic processes with high energy barriers.
EXSY NMR techniques allow one to study processes that are
slow on the chemical shift time scale and distinguish exchange
partners in multistep processes.3 While EXSY spectra have been
used in a qualitative sense for several years, recent advances in
software analysis have made EXSY a more quantitative tool.
In addition, 2D EXSY spectra have been reported for a variety
Baudler has studied the dynamic behavior of many poly-
phosphides and their alkylated derivatives11-14 and showed that
31P EXSY spectra could be used to qualitatively assess the
3-
fluxional processes in the P7 Zintl ion.11 The fluxional
3-
behavior in P7 and other polyphosphides involves multisite
exchange and second-order spin systems which make them
difficult to study by one-dimensional NMR techniques.13 Recent
3-
studies of amorphous (Li‚glyme)3P7 have shown that the P7
cluster undergoes intramolecular exchange faster in the solid
state than in solution.15 The transition-metal Zintl ion complexes
can also show multisite fluxional behavior that can also occur
by intramolecular multistep exchange mechanisms. The origin
of the enhanced fluxionality of the transition-metal Zintl ion
complexes often resides in the Zintl ions themselves. The
mechanisms of the fluxional processes include facile bond
breaking and bond making as well as inversion of configuration
at a single phosphorus site.12,13,16-18 In high-symmetry com-
pounds, the mechanism of exchange is often discernible from
(6) Farrugia, L. J. J. Chem. Soc., Dalton Trans. 1997, 1783.
(7) Bircher, H.; Bender, B. R.; von Philipsborn, W. Magn. Reson. Chem.
1993, 31, 293.
(8) Heise, J. D.; Raftery, D.; Breedlove, B. K.; Washington, J.; Kubiak,
C. P. Organometallics 1998, 17, 4461.
(9) Willem J. Organomet. Chem. 1995, 485, 263.
(10) Brotin, T.; Lesage, A.; Emsley, L.; Collet, A. J. Am. Chem. Soc.
2000, 122, 1171.
(11) Baudler, M.; Ternberger, H.; Faber, W.; Hahn, J. Z. Naturforsch.
1979, 34B, 1690.
† University of Maryland.
‡ University of Houston.
(1) Sandstro¨m, J. Dynamic NMR Spectroscopy; Academic Press: New
York, 1982; p 78.
(2) Encyclopedia of Nuclear Magnetic Resonance; Grant, D. M., Harris,
R. K., Eds.; Wiley: Chichester, UK, 1996; Vols. 1-8.
(3) Orrell, K. G. Annu. Rep. NMR Spectrosc. 1999, 37, 1.
(4) Bain, A. D.; Duns, G. J. Can. J. Chem. 1996, 74, 819.
(5) Bain, A. D.; Cramer, J. A. J. Magn. Reson., A 1993, 103, 217.
(12) Baudler, M.; Deriesemeyer, L. Z. Naturforsch. 1996, B51, 101.
(13) Baudler, M.; Glinka, K. Chem. ReV. 1993, 93, 1623.
(14) Baudler, M.; Glinka, K. Chem. ReV. 1994, 94, 3.
(15) Sen, T.; Poupko, R.; Fleischer, U.; Zimmermann, H.; Luz, Z. J.
Am. Chem. Soc. 2000, 122, 889.
(16) Charles, S.; Bott, S. G.; Rheingold, A. L.; Eichhorn, B. W. J. Am.
Chem. Soc. 1994, 116, 8077.
10.1021/ja000988x CCC: $19.00 © 2000 American Chemical Society
Published on Web 10/27/2000

11102 J. Am. Chem. Soc., Vol. 122, No. 45, 2000
Kesanli et al.
1D variable-temperature 31P NMR studies,13 but extracting rate
data and activation parameters is hampered by second-order spin
systems and multisite exchange.
distilled from K4Sn9, and stored under nitrogen. DMF-d7, DMSO-d6,
and Tol-d8 were purchased from Cambridge Isotopes and used as
received.
Synthesis. Preparation of [K(2,2,2-crypt)]2[P7PtH(PPh3)]. A modi-
fication of the published procedure18 was used for the synthesis of this
complex. In a drybox, K3P7 (30 mg, 0.089 mmol), Pt(PPh3)3 (88 mg,
0.089 mmol), and 2,2,2-crypt (100 mg, 0.26 mmol) were dissolved in
ca. 3 mL of en in a vial. The reaction mixture was stirred for 12 h,
producing a dark red solution with a brown precipitate. The reaction
mixture was heated gently (∼45 °C) to dissolve the precipitate. The
resulting dark red, clear solution was immediately filtered through
tightly packed glass wool in a pipet. Dark red crystals formed in the
reaction vessel after 2 days (88 mg, 59%). 31P{1H} NMR (DMF-d7,
Herein we show that 2D 31P EXSY spectroscopy can be used
as a quantitative tool to unravel the complicated multistep
exchange process in metalated Zintl ion [P7PtH(PPh3)]2- (1).
The details of the synthesis and characterization of 1 and its
arsenic analogue, [As7PtH(PPh3)]2- (2), are also described. A
preliminary communication of the [P7PtH(PPh3)]2- structure has
been reported.18
Experimental Section
1
2
25 °C): δ (ppm) 24 (PPh3, pentet, JPt-P ) 2720 Hz, JP-P ) 8 Hz),
-39 (1P, second-order multiplet), -51 (2P, second-order multiplet),
General Data. All reactions were performed under nitrogen
atmosphere in a drybox (Vacuum Atmospheres Co.). General operating
procedures in our laboratories have been described elsewhere.16 1H and
31P{1H} NMR, 31P 2D EXSY NMR, and 31P VT NMR spectra were
recorded on a Bruker DRX500 AVANCE spectrometer. The spectrom-
eter was run locked for 1H and 31P NMR data collections using DMF-
d7/tol-d8 mixtures as an internal reference for low-temperature studies.
31P{sel-31P} (1H broad band decoupled) homonuclear decoupling NMR
experiments were carried out by combining two 31P RF pulses with a
Mini-Circuits Inc. directional coupler.
-120 (4P, br). 1H NMR (25 °C): δ (ppm) -10.1 (multiplet, 1JPt-H
)
2
1080 Hz, JP-H ≈ 14 Hz). For labeling studies, [P7PtD(PPh3)]2- was
prepared as above but using either DMSO-d6 or DMF-d7 as the solvent.
IR (KBr pellet): ν(Pt-H) (cm-1) 1945, 1881. Anal. Calcd for
C56H96N6O12K2P8Pt: C, 42.94; H, 6.18; N, 5.37; P, 15.82. Found: C,
42.11; H, 5.90; N, 4.81; P, 16.13. The calculated formula contains one
en solvate molecule.
Identical procedures were used to make the title compound from
Pt(PPh3)4 and Pt(PPh3)2(C2H4) precursors.
31P{1H} EXSY spectra were recorded at -78 °C with a pulse
sequence from the Bruker pulse library NOESYTP, 90°-t1-90°-tm-
Preparation of [K(2,2,2-crypt)]2[As7PtH(PPh3)]. In vial 1, 57 mg
(0.089 mmol) of K3As7 and 100 mg (0.26 mmol) of 2,2,2-crypt were
dissolved in 3 mL of en. In vial 2, 66 mg (0.089 mmol) of
Pt(PPh3)2C2H4 was dissolved in 1 mL of toluene, generating a slurry
of partially dissolved complex. The contents of vial 2 were added to
vial 1, resulting in a dark red solution. The reaction mixture was stirred
for 1 h. The solution was concentrated in vacuo to 2 mL and filtered
through tightly packed glass wool in a pipet. Small red crystals formed
in the reaction vessel after 2 days (25 mg, 16%). 31P NMR (DMSO-d6,
25 °C): δ (ppm) 26.7 (PPh3, singlet with 195Pt satellites, 1JPt-P ) 2770
1
90°-acqu, modified to have H broad band decoupled in the phase-
sensitive TPPI (time proportion phase increment) mode. A total of 128
scans per FID of 2K data points were used per time increment. A total
of 425 time increments were collected. Recycle times were 1 s, which
was 5-6 times the measured T1 values. Typical measurement times
were 16 h. The resolutions of the FIDs were 25 and 118 Hz/pt in T2
and T1 domains of the raw data sets, respectively. The EXSY
measurements were repeated with a series of mixing times, tm ) 5 µs,
0.5 ms, 2 ms, 10 ms, 50 ms, 75 ms, and 100 ms.
1
2
Hz). H NMR (25 °C): δ (ppm) -10.8 (1JPt-H ) 1080 Hz, JP-H
)
17.6 Hz). IR (KBr pellet): ν(Pt-H) (cm-1) 1951, 1887. Anal. Calcd
for C63H104N6O12K2As7PPt: C, 38.45; H, 5.30; N, 4.24. Found: C,
38.50; H, 5.22; N, 3.69. The calculated formula contains one en and
one toluene solvate molecule per formula unit.
The forward linear prediction mode from Bruker XWINNMR
software version 2.1 was used to process the data, yielding the final
2D matrix of 2K by 1K. The same software was used to determine the
diagonal and cross-peak volumes. Cross-peaks were normalized ac-
cording to the procedure reported by Ramachandran et al.19 The
equilibrium magnetization, Mj0, necessary for the calculation of rate
constants was determined by acquiring the 2D EXSY spectrum with
mixing time equal to zero, since Iij(0) ) Mj0.
The slopes of the buildup curves at tm ) 0 were determined by linear
regression analysis by using the software KaleidaGraph, version 3.0.4.
The matrix calculations were performed numerically by using the
software package Mathcad, version 7.0.
Electrospray mass spectra were recorded from DMF solutions on a
Finnigan-type mass spectrometer through direct injection. The samples
were ionized by using an ESI probe and detected in the negative ion
mode. Elemental analyses were performed under inert atmospheres by
Desert Analytics, Tucson, AZ, and Atlantic Microlab, Inc., Norcross,
GA.
Chemicals. Melts of nominal composition K3E7 (E ) P, As) were
prepared by high-temperature fusion (∼1000 °C) of stoichiometric ratios
of the elements. The chemicals were sealed in evacuated, silica tubes
and carefully heated with a natural gas/oxygen flame. Caution: Alkali-
metal polyphosphorus compounds are known to spontaneously detonate
eVen under rigorously anaerobic conditions. Therefore, these com-
pounds should only be prepared in small quantities (<0.5 g) and should
be handled with caution. 4,7,13,16,21,24-Hexaoxa-1,10-diazobicyclo-
[8.8.8]hexacosane (2,2,2-crypt) was purchased from Aldrich. Pt(PPh3)3,
Pt(PPh3)4, and Pt(PPh3)2(C2H4) were purchased from Strem and used
without further purification. Anhydrous ethylenediamine (en) and DMF
were purchased from Fisher, distilled over calcium hydride, vacuum
Crystallography. [K(2,2,2-crypt)]2[P7PtH(PPh3)]. A dark red block
with crystal dimensions 0.11 × 0.14 × 0.17 mm was mounted on a
glass fiber in a random orientation. Data collection was performed at
25 °C with Mo KR radiation (λ ) 0.710 73 Å) on an Enraf-Nonius
CAD-4F diffractometer. Data were collected by using an ω-2θ scan
mode with a variable scan rate (0.67-8 deg/min). Periodic monitoring
of three check reflections throughout data collection showed less than
1% decay. An empirical absorption correction (DIFABS) and Lorentz
and polarization corrections were applied.
The data were indexed on an orthorhombic cell, and the systematic
absences indicated space group P212121. The initial structure was
determined from a Patterson map, and the remaining atoms were located
by successive least-squares refinements and difference Fourier synthe-
ses. The hydrogen atoms were placed in idealized positions. The
structure was successfully refined (MOLEN, Enraf-Nonius) using full-
matrix least-squares with the phosphorus, platinum, and potassium
atoms anisotropic in the final cycles. All carbon atoms were refined
isotropically, and one of the cryptands showed multisite disorder. The
highest peak in the final difference map was 1.33 e/Å3, which was
located near Pt.
[K(2,2,2-crypt)]2[As7PtH(PPh3)]. A dark red crystal with ap-
proximate dimensions 0.20 × 0.20 × 0.20 mm was placed on the Enraf-
Nonius CAD-4 diffractometer. Data were collected [Mo KR] with
ω-2θ scans over the θ range 2.02-19.94°. A total of 3660 unique
reflections were collected. Minor variations in intensity were observed
(1-3%), and data were not corrected. Four ψ scan reflections were
collected over the range 7.8-10.3°, and the absorption correction was
applied with transmission factors ranging from 0.6347 to 0.9893, the
average being 0.8364. Data were corrected for Lorentz and polarization
(17) Charles, S.; Danis, J. A.; Eichhorn, B. W.; Fettinger, J. C. Inorg.
Chem. 1997, 36, 3772.
(18) Charles, S.; Eichhorn, B. W.; Bott, S. G.; Fettinger, J. C. J. Am.
Chem. Soc. 1996, 118, 4713.
(19) Ramachandran, R.; Knight, C. T. G.; Kirkpatrick, R. J.; Oldfield,
E. J. Magn. Reson. 1985, 65, 136.
2
2
factors and reduced to Fo and σ(Fo ).
The systematic absences clearly indicated the orthorhombic space
group P212121 (no. 19), which was confirmed by successful solution

Dynamic Exchange with 31P EXSY NMR Spectroscopy
J. Am. Chem. Soc., Vol. 122, No. 45, 2000 11103
and refinement of the structure. The structure was determined by direct
methods using the program SHELXS-86 (Pt and seven As atoms) and
refined with SHELXL.20 Hydrogen atoms were placed in calculated
positions. All of the non-hydrogen atoms were refined anisotropically,
and the structure was refined to convergence. A final difference Fourier
map possessed three peaks (∼0.953 e/Å3) within 1.2 Å of the Pt and
As atoms.
Table 1. Summary of Deuterium Labeling Studies from Different
Solvent Mixtures
solvent mixturea
productb
solvent mixturea
productb
en
DMF-d7
DMSO-d6
CH3CN-d3
en/DMF-d7
en/DMSO-d6
DMF/DMSO-d6
Pt-H
Pt-D
Pt-D
Pt-D
Pt-H
Pt-H
Pt-D
en/CH3CN-d3
DMF/CH3CN-d3
en/THF-d8
DMF/THF-d8
THF/CH3CN-d3
THF
Pt-H
Pt-D
Pt-H
Pt-H
Pt-D
NRc
Results and Discussion
NEt3
NR
Zerovalent triphenylphosphine platinum complexes Pt(PPh3)4,
3-
a pKa values in DMSO solutions (pKDMSO) are (en) ∼36, (DMSO)
35, (THF) >40, and (CH3CN) 31.3. Data from ref 43. pKa values for
DMF are not available to our knowledge but probably exceed 35. b Pt-
Pt(PPh3)3, and Pt(PPh3)2(C2H4) react with the E7 Zintl ions
(E ) P, As) in polar aprotic solvents (en, DMF, DMSO, CH3-
CN) in the presence of 2,2,2-crypt to give the [E7PtH(PPh3)]2-
ions where E ) P (1) or As (2) according to eq 1.
H ) [P7PtH(PPh3)]2-; Pt-D ) [P7PtD(PPh3)]2-
.
Table 2. Summary of Crystallographic Data for
[K(2,2,2-crypt)]2[P7PtH(PPh3)] and [K(2,2,2-crypt)]2[As7PtH(PPh3)]a
K3E7 + Pt(PPh3)2Ln + solv(Hm) + 2(2,2,2-crypt) f
[K(2,2,2-crypt)]2[E7PtH(PPh3)] + PPh3 + Ln + K+ +
A
B
-
empirical formula
formula weight
temperature (K)
wavelength (Å)
crystal system
space group
C54H87N4O12 P8PtK2 C54H87As7K2 N4O12PPt
solv(Hm-1
)
(1)
1505.40
293(2)
1812.98
293(2)
E ) P, As; Ln ) PPh3, 2PPh3, C2H4 + PPh3; solv(Hm) )
0.709 30
orthorhombic
P212121
0.709 30
orthorhombic
P212121
en, DMF, DMSO, CH3CN
unit cell dimensions (Å) a ) 17.088(1)
a ) 17.303(2)
b ) 20.1199(14)
c ) 20.161(2)
7018.7(12)
4
1.716
5.471
The hydride ligands originate from the solvent (see below). The
complexes precipitate from en solutions as dark red crystals of
the [K(2,2,2-crypt)]+ salts in moderate yields (40-60%). 31P
NMR spectroscopic analysis of the crude reaction mixtures
indicates that the reactions are virtually quantitative. The salts
and their solutions are air and moisture sensitive. The complexes
have been characterized by 1H and 31P NMR spectroscopy, IR
spectroscopy, negative ion electrospray mass spectrometry (NI-
EMS), microanalysis, and single-crystal X-ray diffraction.
The negative ion mass spectrum of 1 shows the protonated
molecular ion [HP7PtH(PPh3)]- in which the proton presumably
originates from the solvent matrix and is attached to the P7 cage
(cf. [HP7W(CO)3]2-).21 The deuteride derivative of 1 also
appears as the protonated species [HP7PtD(PPh3)]-. Slight
discrepancies in the intensities and broadness of the peaks in
the deuteride spectrum presumably result from small amounts
of the hydride complex in the sample. Attempts to prepare the
protonated monoanions in solution were unsuccessful.
b ) 19.900(1)
c ) 20.030(2)
volume (Å3)
Z
6811(1)
4
density (g/cm3)
1.468
2.443
abs coeff (mm-1
)
no. of data/restraints/
params
3144/0/380
3660/0/375
final R indices
R1 ) 0.0817
R1 ) 0.0913
[I > 2σ(I)]b
wR2 ) 0.0988
wR2 ) 0.1975
1.604
largest diff peak (e/Å3) 1.33
a A ) [K(2,2,2-crypt)]2[P7PtH(PPh3)]; B ) [K(2,2,2-crypt)]2-
[As7PtH(PPh3)]. b R1 ) Σ|Fo Fc|/ΣFo; wR2 ) (Σw|(Fo)2 (Fc)2|2/
-
-
ΣwFo )1/2
.
2
using both 1H NMR and NI-EMS to determine whether hydride
(Pt-H) or deuteride (Pt-D) products were formed. The
competition experiments give rise to the following order of
hydride sources: en > CH3CN > DMSO > DMF.
In contrast to similar reactions with zerovalent 12-electron
organometallic precursors, the putative [P7Pt(PPh3)]3- is ap-
parently unstable and activates solvent in a formal oxidative
addition process to form 1. The formation of 1 has been
monitored at low temperatures (-78 °C, DMF-d7/tol-d8 mixture)
by using 31P NMR spectroscopy. At short reaction times, only
Structural Studies. The [K(2,2,2-crypt)]+ salts of 1 and 2
are isomorphic, space group P212121, and have been character-
ized by single-crystal X-ray diffraction. A summary of the
crystallographic data is given in Table 2, and a listing of selected
bond distances and angles for both ions is given in Table 3. An
ORTEP drawing of 2 is shown in Figure 1. The structure of 1
is virtually identical to that of 2 (see below) and has been
previously communicated.18 For comparison, a common num-
bering scheme has been used for both compounds.
3-
the starting materials, P7 and Pt(PPh3)2Ln, and free PPh3 are
observed in solution. At longer reaction times (∼1 h, -78 °C),
the resonances for 1 grow in but no intermediates are observed.
The absence of a [P7Pt(PPh3)]3- intermediate contrasts the
chemistry of the Ni(CO)2(PPh3)2 that gives the [P7Ni(CO)]3-
as a stable complex (eq 2).
The ions have C1 point symmetry with unperturbed E7 clusters
bound η2 to [PtH(PPh3)] metal centers. The hydride and PPh3
ligands are cis to one another in the square planar array. The
E(6) and E(7) atoms of the E7 cages are trans to the hydride
and PPh3 ligands, respectively, and complete the square planar
coordination sphere. For electron counting purposes, the com-
P73- + Ni(CO)2(PPh3)2 f
[η4-P7Ni(CO)]3- + CO + 2 PPh3 (2)
plexes can be viewed as η2-E7 ligands bound to 12-electron
3-
Although the formation of 1 appears to be a simple protonation
of a [P7Pt(PPh3)]3- intermediate, the mechanism is clearly more
complicated. Various combinations of solvents (deuterated and
nondeuterated) were used in eq 1 chemistry, and a summary of
the results is given in Table 1. The products were assayed by
PtH(PPh3)+ fragments. In this charge-separated model, each of
the metal-bound pnictides in the E7 cages donate 2 electrons to
give 16-electron square planar Pt(II) complexes.
The Pt-PPh3 bond distances in 1 and 2 are equivalent at
2.22(2) Å (av). The hydride ligands were not located crystal-
lographically but were evident from the H NMR spectra (for
1
(20) Sheldrick, G. Acta Crystallogr. 1990, A46, 467.
(21) Charles, S.; Danis, J. A.; Mattamana, S. P.; Eichhorn, B. W.;
Fettinger, J. C. Z. Anorg. Allg. Chem. 1998, 624, 823.
1
195Pt-1H
1, δ(Pt-H) ) -10.1 ppm, multiplet, J
) 1080 Hz,
2J
≈ 14 Hz; for 2, δ(Pt-H) ) -10.8 ppm, doublet,
31P-1H

11104 J. Am. Chem. Soc., Vol. 122, No. 45, 2000
Kesanli et al.
Table 3. Selected Bond Lengths (Å) and Angles (deg) for [K(2,2,2-crypt)]2[P7PtH(PPh3)] and [K(2,2,2-crypt)]2[As7PtH(PPh3)]
1 (E ) P)
2 (E ) As)
1 (E ) P)
2 (E ) As)
Pt(1)-P(8)
Pt(1)-E(7)
Pt(1)-E(6)
E(1)-E(2)
E(1)-E(3)
E(2)-E(4)
E(2)-E(5)
E(3)-E(6)
2.23(1)
2.40(1)
2.42(1)
2.16(2)
2.15(2)
2.22(2)
2.26(2)
2.18(2)
2.211(12)
2.499(4)
2.509(5)
2.347(7)
2.347(7)
2.437(7)
2.480(7)
2.403(7)
E(3)-E(7)
E(4)-E(6)
E(4)-E(5)
E(5)-E(7)
P(1)-C(11)
P(1)-C(31)
P(1)-C(21)
2.15(2)
2.14(2)
2.26(2)
2.16(2)
1.80(4)
1.88(4)
1.81(4)
2.414(6)
2.403(7)
2.459(7)
2.403(6)
1.81(4)
1.83(4)
1.91(4)
P(8)-Pt(1)-E(7)
P(8)-Pt(1)-E(6)
E(7)-Pt(1)-E(6)
E(2)-E(1)-E(3)
E(1)-E(2)-E(4)
E(1)-E(2)-E(5)
E(4)-E(2)-E(5)
E(1)-E(3)-E(6)
E(1)-E(3)-E(7)
E(6)-E(3)-E(7)
E(6)-E(4)-E(2)
174.5(4)
98.6(4)
78.2(4)
97.3(6)
107.7(8)
108.5(7)
60.5(5)
104.4(6)
108.2(7)
89.2(6)
104.6(7)
174.9(3)
94.6(3)
83.3(2)
99.0(2)
108.5(3)
107.7(2)
60.0(2)
106.0(3)
106.9(3)
87.4(2)
104.5(2)
E(6)-E(4)-E(5)
E(2)-E(4)-E(5)
E(7)-E(5)-E(4)
E(7)-E(5)-E(2)
E(4)-E(5)-E(2)
E(3)-E(6)-E(4)
E(3)-E(6)-Pt(1)
E(4)-E(6)-Pt(1)
E(5)-E(7)-E(2)
E(5)-E(7)-Pt(1)
E(3)-E(7)-Pt(1)
101.6(6)
60.6(6)
99.3(6)
104.1(7)
58.9(5)
101.0(7)
87.9(5)
93.0(5)
101.9(6)
94.1(5)
81.1(5)
100.7(2)
60.9(2)
100.2(2)
104.9(2)
59.1(2)
100.7(2)
88.2(2)
91.3(2)
100.6(2)
91.9(2)
88.2(2)
Figure 1. ORTEP drawing of the [As7PtH(PPh3)]2- ion.
1
2
31
195
1
1
J
Pt- H ) 1080 Hz, J P- H ) 18 Hz). The average Pt-P bond
to the P7 cage in 1 is 2.41(2) Å and is significantly longer than
Pt-P bonds involving phosphine ligands (2.2-2.3 Å, av). The
Pt-As bonds in 2 (2.50(1) Å, av) are approximately 0.1 Å
longer than the Pt-P7 contacts in 1, which is consistent with
the 0.12 Å increase in the covalent radius of As relative to P.22
There is an apparent lengthening of the Pt-E(6) bonds in both
complexes (the E atom trans to the hydrides) relative to the
Pt-E(7) bonds; however, the differences are not statistically
relevant. The E(7)-Pt-E(6) angle is more acute in 1 relative
to 2 due to the smaller bite angle of the P7 cage relative to As7.
The E-E bonds within the E7 cages are quite similar to those
Figure 2. Variable-temperature 31P{1H} NMR spectra for [P7PtH-
(PPh3)]2- recorded in DMF-d7 at 202.4 MHz. The numeric labels on
the low-temperature limiting spectrum are peak assignments corre-
sponding to the atom labeling scheme used for the complex.
) 3.38 Å),31 and relatively compressed with respect to the
neutral R3As7 compounds (h ) 3.47 Å, av).32
Spectroscopic Studies. The variable-temperature 31P{1H}
NMR spectra for 1 are shown in Figure 2. The low-temperature
limiting spectrum is consistent with the solid-state structure (C1
symmetry) showing seven independent 31P multiplets arising
from the P7 cluster and a downfield broad doublet with 195Pt
satellites due to the PPh3 ligand. All resonances can be assigned
through combinations of 1D selective decoupling and 31P-31P
COSY experiments (see below), and the peak labels in Figure
3-
of the unligated E7 parent ions. The As-As bonds in 2 are,
on average, 0.19-0.25 Å longer than the corresponding P-P
bonds in 1.22 The heights of the E7 fragments of the compounds
(i.e., the distance from the E(2)-E(4)-E(5) basal planes to the
apical E(3) atoms, h) are 3.04 and 3.37 Å for 1 and 2,
respectively. These heights are indicative of unperturbed E7
cages and are consistent with the trends in E-E bonds in related
compounds.23,24 For comparison, the distances from the apical
phosphorus atom to the center of the basal triangular plane are
elongated by 0.15 Å in the neutral R3P7 molecules (h ) 3.15
Å)25-28 relative to the parent P73- ion (h ) 3.00 Å).23,29,30 The
height of 2 is equivalent to that of the As73- ion in Ba3As14 (h
(25) Kova´cs, I.; Baum, G.; Fritz, G.; Fenske, D.; Wiberg, N.; Schuster,
H.; Karaghiosoff, K. Z. Anorg. Allg. Chem. 1993, 619, 453.
(26) Fritz, G.; Schneider, H.-W. Z. Anorg. Allg. Chem. 1990, 584, 12.
(27) Fritz, G.; Hoppe, K. D.; Ho¨nle, W.; Weber, D.; Mujica, C.;
Manriquez, V.; von Schnering, H. G. J. Organomet. Chem. 1983, 249, 63.
(28) Mujica, C.; Weber, D.; von Schnering, H. G. Z. Naturforsch. 1986,
B41, 991.
(29) Dahlmann, W.; von Schnering, H. G. Naturwissenschaften 1972,
59, 420.
(22) Huheey, J. E.; Keiter, E. A.; Keiter, R. L. Inorganic Chemistry.
Principles of Structure and ReactiVity, 4th ed.; Harper & Row: New York,
1993; p 300.
(23) Korber, N.; von Schnering, H.-G. J. Chem. Soc., Chem. Commun.
1995, 1713.
(30) von Schnering, H.-G.; Ho¨nle, W.; Marinquez, V.; Meyer, T.;
Mensing, C.; Giering, W. 2nd Int. Conf. Solid State Chem., Collect. Abstr.
FV54; Eindhoven, 1982.
(31) Belin, C. H. E.; Mercier, H.; Bonnet, B.; Bernard, M. C. R. Acad.
Sci. 1988, 307, 549.
(24) Mattamana, S. P.; Phomphrai, K.; Fettinger, J. C.; Eichhorn, B. W.
Inorg. Chem. 1998, 37, 6222.
(32) Ho¨nle, W.; Wolf, J.; von Schnering, H.-G. Z. Naturforsch., B 1988,
43, 219.

Dynamic Exchange with 31P EXSY NMR Spectroscopy
J. Am. Chem. Soc., Vol. 122, No. 45, 2000 11105
Scheme 1. Mechanism of Exchange for [P7PtH(PPh3)]2-
°C becomes a pentet flanked by Pt satellites (1J
P ) 2720
195Pt-31
Hz) at 25 °C (Figure 3a). Similarly, the hydride resonance is
transformed from a doublet of multiplets at -78 °C to a single
multiplet at 25 °C, which is also flanked by Pt satellites (1J
195Pt-1H
) 1080 Hz). These data are consistent with the intramolecular
two-step fluxional process outlined in Scheme 1. In this process,
the P7 shifts from an η2 f η4 coordination mode to generate an
18-electron [(η4-P7)PtH(PPh3)]2- intermediate, I1. A second η4
f η2 shift leaves P(5) and P(4) bound to Pt and forms a bond
between P(6) and P(7). The shifting process effectively ex-
changes P(2) with P(3), P(6) with P(4), and P(7) with P(5). This
type of η2 f η4 f η2 shift was observed in the reversible
reaction of CO with [η4-HP7W(CO)3]2- to give the [η2-HP7W-
(CO)4]2- ion as outlined in eq 3.
Figure 3. Variable-temperature NMR spectra for [P7PtH(PPh3)]2-
showing (a, top) the 31P{1H} NMR resonance of the PPh3 ligand and
(b, bottom) the 1H NMR resonance of the hydride ligand. The
experimental temperatures are (from bottom to top) -78, -50, -20,
and 25 °C for both sets of spectra.
2 correspond to the atom labeling scheme used in Figure 1 and
subsequent drawings.
[η4-HP7W(CO)3]2- + CO h [η2-HP7W(CO)4]2- (3)
The low-temperature resonances of the hydride and the PPh3
ligands are convenient starting points for assigning the peaks
of the P7 cage. The variable-temperature 31P NMR spectra for
However, due to the asymmetry in the PtH(PPh3) group, shifting
alone does not exchange all four of the Pt-bound phosphorus
atoms. An additional spinning step (I1 f I2, Scheme 1) is
required in the [(η4-P7)PtH(PPh3)]2- intermediate to exchange
the four Pt-bound phosphorus atoms. For the isoelectronic [η4-
P7W(CO)3]3- ion, spinning of the C3V W(CO)3 fragment attached
to the C2V P7 cluster is rapid on the NMR time scale at -60 °C
(80 MHz, DMF-d7).16 Because the two processes in Scheme 1
occur simultaneously, determination of the relative rate constants
for the two processes (kshift and kspin) is not possible by line
shape analysis of the variable-temperature 1D spectra. The
broad, second-order nature of the limiting spectra further
complicates a potential line shape simulation, and even an
average rate constant, kobs, would be difficult to extract by this
method.
2D EXSY NMR is a technique which is ideal for studying
multisite exchange pathways under equilibrium conditions. The
EXSY experiment is similar to a NOESY experiment except
cross-peaks are observed between nuclei that are in dynamic
chemical exchange.33 The rate constants associated with the
exchange process can be calculated from the cross-peak intensi-
ties of the 2D EXSY spectrum without a line shape analysis.
Several approaches for calculating rate constants from eq 4
have been developed.34 The equation relates the cross-peak
intensities (Iij)
1
the PPh3 region and H NMR spectra for the hydride region
are shown in parts a and b, respectively, of Figure 3. Both
resonances are doublets of multiplets in their limits as is
expected due to large couplings from the trans phosphorus atoms
in the square planar coordination sphere. Selective decoupling
experiments at -78 °C allow for the assignments of the
phosphorus atoms trans to these ligands. First, selective 31P-
decoupled 1H NMR experiments, 1H{sel-31P} NMR, show that
individually decoupling the 31P resonance at -40.0 ppm results
in the collapse of the hydride doublet of multiplets into a single
multiplet while individual decoupling of the other six resonances
has no appreciable effect. As such the resonance at -40.0 ppm
is assigned to P(6) of the P7 cage. Similarly, 31P{sel-31P} NMR
experiments show that irradiation of the -67.4 ppm resonance
collapses the PPh3 resonance to a single multiplet whereas
irradiation of the other resonances has no effect. Therefore, we
can assign the -67.4 ppm resonance to P(7). With these two
assignments, the remaining five resonances can be assigned from
the -78 °C 31P-31P COSY NMR spectrum and are labeled in
Figure 2.
Upon warming, the resonances for P(6), P(7), P(4), and P(5)
coalesce into the baseline and reemerge above -20 °C as a
single resonance of intensity 4 (Figure 2). The resonances for
P(2) and P(3) coalesce at -60 °C to give a single peak of
intensity 2 at -53.5 ppm. The P(1) resonance is not affected
by the exchange process and remains sharp at all temperatures.
In addition, the PPh3 resonance which is a broad doublet at -78
(33) Jeener, J.; Meier, B. H.; Bachmann, P.; Ernst, R. R. J. Chem. Phys.
1979, 71, 4546.
(34) Perrin, C. L.; Dwyer, T. J. Chem. ReV. 1990, 90, 935.

11106 J. Am. Chem. Soc., Vol. 122, No. 45, 2000
Kesanli et al.
Iij(tm) ) Mj0(e-Rt
)
(4)
m
ij
at mixing time tm to the rate matrix R35,36 and the initial
equilibrium magnetization vector Mj0 of the nuclei in site j. The
matrix R contains the rate constant kji, which is the first-order
rate constant for chemical exchange from site j to site i, and
the spin lattice relaxation rates Fi and Fj (i.e., the T1 values).
Fj
-kij
R )
[
]
-kji Fi
Two methods were used in this study for calculating the rate
constants. Method A, the initial rate approximation,37 is based
on eq 4 above where, at short mixing time, the exponential
simplifies to give eq 5.
Iij(tm) ≈ -tmRijMj0 ≈ kjitmMj0
i * j
(5)
In this form, the 2D spectrum can be directly related to the
exchange matrix R. The rate constant, kji, can be determined
from the slope of a plot of Iij/Mj0 versus tm. This method allows
for the separation of direct cross-peaks (i.e., those in nonsat-
urated direct chemical exchange), which have a nonzero slope,
from the indirect cross-peaks, which have a zero slope. It is
necessary to use sufficiently short mixing times to have reliable
rate constants because the slopes of the buildup curves change
drastically at longer tm values. The disadvantage of method A
is the need for multiple tm values, which requires considerable
spectrometer run time.
The matrix method (method B) for calculating rate constants
was first described by Perrin and Gipe36 and subsequently by
Bremer et al.38 This method consists of diagonalizing the rate
matrix R from all peak volume data at a single tm. The values
of the forward and reverse rate constants can be read from the
matrix directly.
The matrix equation is given by eq 6, where Aij ) Iij(tm)/Mj0
and X is the square matrix of eigenvectors of A,
X(ln Λ)X-1
ln A
R ) -
) -
(6)
tm
tm
such that X-1AX ) Λ ) diag(λi) and ln Λ ) diag(ln λi), where
λi is the ith eigenvalue of A. Although eq 6 looks formidable,
it is quite easy to extract the rate constants using commercial
software. Experimental intensities are the only input necessary
to solve for the eigenvalues and eigenvectors of A and perform
matrix multiplications. Application of this technique to 31P
exchange has been described by von Philipsborn,7 Kubiak,8 and
others.39,40
31P EXSY spectra for 1 were recorded at -78 °C in DMF-
d7/tol-d8 mixtures with mixing times in the range of 5 µs to
100 ms. Spectra with tm ) 10, 50, and 100 ms are shown in
Figure 4. To confirm that the observed cross-peaks were due
to exchange and not a result of either NOE or J leakage (i.e.,
unfiltered COSY peaks), the following experiments were
performed. First, with tm ) 5 µs (the instrumental limit),
chemical exchange is too slow to be detected in the EXSY
Figure 4. 31P{1H} EXSY NMR spectra for [P7PtH(PPh3)]2- recorded
at -78 °C, 202.4 MHz, and various mixing times, tm. The numeric
labels on the horizontal spectra correspond to the phosphorus peak
assignments.
(35) Macura, S.; Ernst, R. R. Mol. Phys. 1980, 41, 95.
(36) Perrin, C. L.; Gipe, R. K. J. Am. Chem. Soc. 1984, 106, 4036.
(37) Kumar, A.; Wagner, G.; Ernst, R. R.; Wuthrich, K. J. Am. Chem.
Soc. 1981, 103, 3654.
(38) Bremer, J.; Mendz, G. L.; Moore, W. J. J. Am. Chem. Soc. 1984,
106, 4691.
(39) Li, D. W.; Bose, R. N. J. Chem. Soc., Dalton Trans. 1994, 3717.
(40) Malaisselagae, F.; Willem, R.; Penders, M.; Malaisse, W. J. Mol.
Cell. Biochem. 1992, 115, 137.
experiment and only residual COSY peaks should be observed.
The EXSY spectrum for 1 with tm ) 5 µs showed no cross-
peaks, indicating that J leakage is not occurring in these
experiments. Second, a ROESY experiment at -78 °C, which
separates NOE and exchange effects, showed no NOE cross-

Dynamic Exchange with 31P EXSY NMR Spectroscopy
J. Am. Chem. Soc., Vol. 122, No. 45, 2000 11107
Table 4. Rate Constantsa for the Dynamic Exchange Processb in
peaks between P(6)-P(7) and P(4)-P(5) nuclei; however, the
large frequency separation between the P(6)-P(4) and P(7)-
P(5) pairs of resonances precluded analysis of NOE effects due
to dynamic range limitations of the instrument. However, EXSY
spectra recorded at -63 °C yielded rate constants ca. 6 times
larger than those at -78 °C (see below), which is consistent
with cross-peaks due to chemical exchange (a temperature-
dependent process) and not NOE (virtually temperature-
independent). As such, the EXSY data can be analyzed on the
basis of chemical exchange, and other extraneous contributions
can be neglected.
[P7PtH(PPh3)]2-
k (s-1
)
method A method B k (s-1
Fast Process (Shifting)b
)
method A method B
k64
k46
150
159
153
176
k75
k57
137
156
156
161
Slow Process (Spinning)b
k67
k76
6.8
9.7
7.0
8.5
k45
k54
7.4
6.2
7.9
7.4
a Rates determined from volume integration of EXSY off-diagonal
peaks. Method A ) initial rate approximation method. Method B )
matrix analysis method. See the Experimental Section for details. b See
Scheme 1 for definitions of rate constants and the steps in the exchange
mechanism. Data were collected at -78 °C.
Visual inspection of the EXSY spectra in Figure 4 reveals
three different exchange situations. At long mixing times (tm )
100 ms, Figure 4a (top)), spinning and shifting (Scheme 1) are
fast on the NMR time scale and equal-intensity cross-peaks are
observed for all of the P(4), P(5), P(6), and P(7) pairs. This
situation indicates direct and indirect exchange is occurring and
the cross-peak intensity is saturated. At shorter mixing times
(tm ) 50 ms, Figure 4b (middle)), large exchange cross-peaks
are observed between the P(4)-P(6) and P(5)-P(7) pairs of
resonances, with the remaining peaks being of diminished
intensity. In this regime, the EXSY experiment can separate
the different steps, one fast and one slow, in the exchange
Conclusions
von Philipsborn and co-workers first showed that inequivalent
31P-31P EXSY NMR peak intensities could be used to identify
inter- and intramolecular exchange processes in [Rh(S,S-
chiraphos)(mac)]+ complexes (mac ) R-N-acetylcinnamate).7
The use of 31P-31P EXSY NMR spectroscopy in the study of
the dynamic behavior of 1 is, to our knowledge, the first example
of an intramolecular multistep fluxional process in which the
individual steps could be quantitatively separated. In future
studies, we shall show that precise kinetic data can be extracted
in related dynamic systems where traditional line shape analysis
hampered by second-order spin systems limited access to
limiting spectra.
process outlined in Scheme 1. At shorter mixing times (tm
)
10 ms), only the fast step of the exchange mechanism is
detectable, which gives rise to cross-peaks between the P(4)-
P(6) and P(5)-P(7) pairs (Figure 4c (bottom)). Qualitatively,
these data indicate that the P(4)-P(6) and P(5)-P(7) exchange
is fast (the η2 f η4 f η2 shift, Scheme 1) relative to exchange
between the P(6)-P(7) and P(4)-P(5) pairs (spinning, Scheme
1). Chemical exchange between the P(6)-P(5) and P(7)-P(4)
pairs must occur by indirect means involving a shift + spin
process. The rate of this indirect process is governed by the
rate-determining step of the shift + spin exchange process,
which is the spinning step. Therefore, exchange between P(6)-
P(7) and P(4)-P(5) pairs, which occurs through a direct
(spinning) process, has the same rate constant as the indirect
(spinning + shifting) exchange process involving the P(6)-
P(5) and P(7)-P(4) pairs. While we cannot exclude other
exchange mechanisms, the data seem most consistent with the
process outlined in Scheme 1. An alternative process involving
an η2 f η2 f η2 “walk” of the PtH(PPh3) unit down an edge
of the P7 cage is unlikely. In such a mechanism, the PtH(PPh3)
unit would bridge an edge in the exchange process (e.g., the
P(7)-P(5) edge) rather than form the η4 intermediate (e.g., I1,
Scheme 1). From a P(7)-P(5) edge-bridged η2 transition state,
the probability of the PtH(PPh3) unit moving to the P(1)-P(7)
edge is the same as that for moving back to the P(7)-P(6) edge.
Repetition of this process would exchange all phosphorus nuclei
in the P7 cage to generate a single, time-averaged resonance in
the fast exchange limit. Since the P(1) resonance does not
become involved in the exchange process and the other nuclei
exchange in a pairwise fashion, this mechanism is inconsistent
with the observed data.
The formation of compounds 1 and 2 from zerovalent Pt
precursors is an unusual example of metal-based chemistry with
transition-metal Zintl ion compounds. Related compounds such
as [η4-P7Ni(CO)]3-, [η2-P7W(CO)4]3-, and [η4-E7M(CO)3]3-
,
where E ) P, As, and Sb and M ) Cr and Mo, all show ligand-
based chemistry with protons and other electrophiles (i.e.,
reactions at the E7 cluster).16-18,21,41,42 Pt complexes are of course
well-known for C-H activation chemistry, but the trends in
C-H activation in eq 1 are unusual. While we do not understand
the detailed mechanism of hydride formation at present, the
reaction appears to require a coordinating ligand for hydride
formation and is not a simple protonation. Spectroscopic studies
3-
on the [E7M(CO)3]3- ions have shown the E7 ligands to be
potent electron donors. As such, a zerovalent 16-electron [η2-
P7Pt(PPh3)]3- intermediate would be very nucleophilic and
prone to oxidative addition chemistry. It is interesting to note
that addition of a proton to the P7 cage leaves the metal in the
zerovalent state (e.g., [η4-HP7Ni(CO)]2-) whereas the present
compounds formally contain an oxidized Pt(II) center.
Acknowledgment. We thank the donors of the Petroleum
Research Fund and the NSF (Grant CHE 9500686 for support
of this work. We are indebted to Dr. Kangyan Du for the NI-
EMS data and Dr. Sundeep Mattamana for experimental
assistance. We thank Prof. Wolfgang von Philipsborn for very
helpful discussions.
Rate constants were extracted from the EXSY data by using
the method of initial rates (method A) and matrix analysis
method (method B) described above. A summary of the rate
data is given in Table 4. Importantly, the two methods of
analysis give the same rate constants within experimental error.
In addition, the rate constants show that the η2 f η4 f η2 shift
(kshift ) 160 s-1, av) is approximately 20 times faster than
spinning (kspin ) 8 s-1, av). The rate constants at -63 °C are
significantly faster (kshift ≈ 103 s-1, av; kspin ≈ 60 s-1, av), which
is again consistent with an all-exchange-type process.
Supporting Information Available: Complete listing of
crystallographic data for [K(2,2,2-crypt)]2[As7PtH(PPh3)] (PDF).
This material is available free of charge via the Internet at
http://pubs.acs.org.
JA000988X
(41) Charles, S.; Fettinger, J. C.; Eichhorn, B. W. J. Am. Chem. Soc.
1995, 117, 5303.
(42) Charles, S.; Eichhorn, B. W.; Fettinger, J. C.; Bott, S. G. Inorg.
Chem. 1996, 35, 1540.
(43) Bordwell, F. G. Acc. Chem. Res. 1988, 21, 456.