Energetics of halogen bonding of group 10 metal fluoride complexes

Article abstract of DOI:10.1021/ja203320y

A study is presented of the thermodynamics of the halogen-bonding interaction of C₆F₅I with a series of structurally similar group 10 metal fluoride complexes trans-[Ni(F)(2-C₅NF ₄)(PCy₃)₂] (2), trans-[Pd(F)(4-C ₅NF₄)(PCy₃)₂] (3), trans-[Pt(F){2-C₅NF₂H(CF₃)}(PR ₃)₂] (4a, R = Cy; 4bR = iPr) and trans-[Ni(F){2-C ₅NF₂H(CF₃)}(PCy₃)₂] (5a) in toluene solution. ¹⁹F NMR titration experiments are used to determine binding constants, enthalpies and entropies of these interactions (2.4 ≤ K₃₀₀ ≤ 5.2; -25 ≤ ΔH^o ≤ -16 kJ mol^-1; -73 ≤ ΔS^o ≤ -49 J K^-1 mol^-1). The data for -ΔH^o for the halogen bonding follow a trend Ni < Pd < Pt. The fluoropyridyl ligand is shown to have a negligible influence on the thermodynamic data, but the influence of the phosphine ligand is significant. We also show that the value of the spin-spin coupling constant J_PtF increases substantially with adduct formation. X-ray crystallographic data for Ni complexes 5a and 5c are compared to previously published data for a platinum analogue. We show by experiment and computation that the difference between Pt-X and Ni-X (X = F, C, P) bond lengths is greatest for X = F, consistent with F(2pπ)-Pt(5dπ) repulsive interactions. DFT calculations on the metal fluoride complexes show the very negative electrostatic potential around the fluoride. Calculations of the enthalpy of adduct formation show energies of -18.8 and -22.8 kJ mol ^-1 for Ni and Pt complexes of types 5 and 4, respectively, in excellent agreement with experiment.

Full text of DOI:10.1021/ja203320y

ARTICLE
pubs.acs.org/JACS
Energetics of Halogen Bonding of Group 10 Metal Fluoride Complexes
Torsten Beweries,†,^ Lee Brammer,‡ Naseralla A. Jasim,† John E. McGrady,§ Robin N. Perutz,*,† and
Adrian C. Whitwood†
†Department of Chemistry, University of York, Heslington, York, United Kingdom YO10 5DD
‡Department of Chemistry, University of Sheffield, Sheffield, United Kingdom S3 7HF
§Department of Chemistry, Inorganic Chemistry Laboratory, University of Oxford, South Parks Road, Oxford,
United Kingdom OX1 3QR
S Supporting Information
b
ABSTRACT: A study is presented of the thermodynamics of the
halogen-bonding interaction of C6F5I with a series of structurally
similar group 10 metal fluoride complexes trans-[Ni(F)(2-C5NF4)-
(PCy3)2] (2), trans-[Pd(F)(4-C5NF4)(PCy3)2] (3), trans-[Pt(F)-
{2-C5NF2H(CF3)}(PR3)2] (4a, R = Cy; 4b R = iPr) and
trans-[Ni(F){2-C5NF2H(CF3)}(PCy3)2] (5a) in toluene solution.
19F NMR titration experiments are used to determine binding
constants, enthalpies and entropies of these interactions (2.4 e K300 e
5.2; ꢀ25 e ΔHo e ꢀ16 kJ molꢀ1; ꢀ73 e ΔSo e ꢀ49 J Kꢀ1 molꢀ1).
The data for ꢀΔHo for the halogen bonding follow a trend Ni < Pd < Pt. The fluoropyridyl ligand is shown to have a negligible
influence on the thermodynamic data, but the influence of the phosphine ligand is significant. We also show that the value of the
spinꢀspin coupling constant JPtF increases substantially with adduct formation. X-ray crystallographic data for Ni complexes 5a and
5c are compared to previously published data for a platinum analogue. We show by experiment and computation that the difference
between PtꢀX and NiꢀX (X = F, C, P) bond lengths is greatest for X = F, consistent with F(2pπ)ꢀPt(5dπ) repulsive interactions.
DFT calculations on the metal fluoride complexes show the very negative electrostatic potential around the fluoride. Calculations of
the enthalpy of adduct formation show energies of ꢀ18.8 and ꢀ22.8 kJ molꢀ1 for Ni and Pt complexes of types 5 and 4, respectively,
in excellent agreement with experiment.
’ INTRODUCTION
electron-deficient region formed when polarizable halogen
atoms are bound to electronegative groups as in C6F5I.10 Reports
of the energetics of halogen bonding in solution are very limited
indeed, although there was indirect evidence from crystal struc-
tures that halogen bonds compete successfully with hydrogen
bonds.5,11 Laurence et al. assumed that relationships of shifts in IR
bands observed on hydrogen bonding could be transferred to
halogen bonds in order to determine trends in enthalpies of complex-
ation.12 Corradi et al. obtained the first calorimetric estimate of the
enthalpy of adduct formation by adding 1-iodoperfluorohexane
Intermolecular interactions such as hydrogen and halogen
bonding have found applications in various fields of chemistry
and interdisciplinary sciences.1,2 The interactions of halogens are
of great interest since they can exhibit both Lewis acidic and basic
character. Structural characterization of these interactions is
established in a wide range of situations encompassing organic
compounds, main group elements, and transition metals.1a,3 In
recent years, halogen bonding has emerged as a powerful,
directional, and very useful noncovalent force relevant to mo-
lecular recognition with wide-ranging applications.2 For example,
the benefits of halogen bonding were described in a recent
communication which showed that such an interaction can help
to control the luminescence properties of bimetallic silverꢀgold
clusters.4 Moreover, examples of the application of halogen
bonding in self-assembly processes,5 in designing liquid crystals,6
in conducting materials,7 and in structural biology8 have been
reported.
to excess tetramethylpiperidine, obtaining a value of 31 kJ molꢀ1
,
assuming that solvent effects were negligible.5a Very recently, the
free energies of halogen bonding of a variety of substituted
iodoperfluoroarenes to organic bases were reported, and subtle
differences from hydrogen bonding were detected.13aꢀc The free
energies, enthalpies, and entropies of complexation of toluene-d8
with 1-C3F7I (ΔHo = ꢀ2.9 kJ molꢀ1) and 2-C3F7I (ΔHo =
ꢀ2.7 kJ molꢀ1) were reported by 19F NMR spectroscopy while
this article was under review. DFT calculations indicated that
The phenomenon of halogen bonding has been studied for
small gaseous molecules in detail by rotational spectroscopy9 and
by computational methods, leading to a better understanding
of this noncovalent force. Theoretical results suggest that halo-
gen bonding results from interaction with the “σ-hole”, an
these adducts involve CꢀI π interaction.13d
3 3 3
Received: April 11, 2011
Published: August 18, 2011
r
2011 American Chemical Society
14338
dx.doi.org/10.1021/ja203320y J. Am. Chem. Soc. 2011, 133, 14338–14348
|

Journal of the American Chemical Society
ARTICLE
Chart 1. Hydrogen and Halogen Bonding of trans-[Ni(F)-
(C5NF4)(PEt3)2] (D = Hydrogen-Bond Donor or Halogen-
Bond Donor)
Chart 2. Group 10 Metal Fluorides for Halogen Bonding and
Structural Studies
Our recent determination of the enthalpy, entropy, and
temperature-dependent equilibrium constants of formation of
halogen-bonded adducts between a nickel fluoride complex and
iodopentafluorobenzene remains the most detailed source for
the thermodynamics of halogen bonding in solution.14 Transi-
tion metal fluorides have been studied extensively to elucidate
their unusual reactivity and their role in CꢀF activation15 and
CꢀF bond-formation processes.16 We showed that trans-[Ni-
(F)(C5NF4)(PEt3)2] (1) acts as a Lewis base in the formation of
hydrogen and halogen bonds with hydrogen bond donors (e.g.
indole) and halogen bond donors (e.g., iodopentafluorobenzene,
Chart 1).14 The formation of the adducts was monitored by
multinuclear NMR spectroscopy, and the thermodynamics of the
interactions were quantified by using the 19F NMR shift of the
metal fluoride as a spectroscopic handle. We found that the
enthalpy of halogen bond formation was ꢀ16(1) kJ molꢀ1 in
toluene but ꢀ26(1) kJ molꢀ1 in heptane. Entropies of adduct
formation in the two solvents were ꢀ42(4) and ꢀ63(4) JKꢀ1 molꢀ1,
respectively. The 19F NMR chemical shift changed by more than
20 ppm at low temperature on adduct formation, although other
NMR parameters were unaffected. We were also able to compare
the energetics of halogen bond formation to those for hydrogen
bonding to the same complex. The hydrogen bonding results
may be compared to previous measurements, including some on
other metal fluoride complexes.17
been reported only for binary complex ions, [MF6]nꢀ and for
complexes of early transition metals bearing two fluoride
ligands.18 However, a series of such complexes bearing only
one fluoride ligand is desirable for comparative studies of halogen
bonding in order to avoid multiple interactions. Moreover, they
should be soluble in nonpolar solvents in order to avoid
competing interactions with the solvent. The synthesis of such
a series has not been achieved previously.
We have shown that oxidative addition of aromatic CꢀF
bonds at zero-valent group 10 metals can be used to generate a
range of metal fluoride complexes. Oxidative addition of penta-
fluoropyridine proceeds smoothly with nickel(0) and palladium(0)
in the presence of monodentate phosphines, and leads to CꢀF
activation at the 2-position for nickel,19 but the 4-position for
palladium (Chart 2).20a Fluoride products can only be generated
for platinum using 2,3,5-trifluoro-4-trifluoromethylpyridine;
otherwise alkyl migration from the phosphine to the metal
occurs with formation of a fluorophosphine ligand. Such reac-
tions occur with Pt(PCy3)2 and pentafluoropyridine and involve
a phosphine-assisted pathway with fluorine transfer from the
pyridine to the phosphine.20b,21
During our studies, we focused on group 10 metal fluorides
and tried to synthesize a series of complexes by CꢀF activation
with a minimum of structural differences. With these considera-
tions in mind, we concentrated on square-planar complexes with
two trans-tricyclohexylphosphine ligands (PCy3) but with var-
ious fluoropyridyl groups trans to the fluoride ligand trans-[M-
(F)(pyF)(PCy3)2]. We synthesized two new nickel complexes by
reaction with pentafluoropyridine and 2,3,5-trifluoro-4-trifluor-
omethylpyridine, yielding 2 and 5a, respectively (Chart 2). The
metal is bound at the 2-position of the fluoropyridyl in both these
products. For the palladium derivative, we synthesized the pre-
viously described 4-tetrafluoropyridyl derivative, 3, while for
platinum we used the known compound 4a derived from reaction
with 2,3,5-trifluoro-4-trifluoromethylpyridine. This methodology
provided us with a pair of structurally identical nickel and
platinum complexes and a separate pair of very similar nickel
and palladium complexes. Unfortunately, we were unable to
In this article, we present a comparative study of halogen
bonding between iodopentafluorobenzene and members of a
series of structurally similar group 10 metal fluoride complexes in
toluene solution. We show that the halogen-bond enthalpy increases
down the series, Ni, Pd, Pt. We also determined experimental and
calculated metalꢀfluorine bond lengths for closely related nickel
and platinum complexes in order to elucidate the effect of the
change of metal on metalꢀfluorine bonding.
’ RESULTS AND DISCUSSION
Our strategy in this paper is to synthesize a series of closely
related metal fluoride complexes so that we can identify periodic
trends in the thermodynamics of halogen bond formation. The
synthesis of isostructural metal fluoride complexes has previously
14339
dx.doi.org/10.1021/ja203320y |J. Am. Chem. Soc. 2011, 133, 14338–14348

Journal of the American Chemical Society
ARTICLE
Scheme 1. Synthesis of Nickel Fluoride Complexes 2, 5a,
and 5c
Figure 1. Molecular structure of complex 2. Hydrogen atoms are
omitted for clarity. The displacement ellipsoids correspond to 50%
probability. Selected bond lengths (Å) and angles (deg.): Ni(1)ꢀF(1)
1.8518(15), Ni(1)ꢀP(1) 2.2300(4), Ni(1)ꢀC(1) 1.865(2); F(1)ꢀ
Ni1ꢀP(1) 87.358(13), P(1)ꢀNi(1)ꢀC(1) 92.892(14).
downfield shift of the signals of the fluorine atoms bound in
o- and p-positions to the metal atom relative to the corresponding
signals of 2. As for compound 2, the molecular ion was observed
using LIFDI+ mass spectrometry (m/z 819.4). Complex 5c was
synthesized similarly, and its spectra are illustrated in the
Supporting Information (Figures S9 and S10) (SI).
isolate the palladium analogues of 4a and 5a. The variations in
the ligand set of complexes 2, 3, 4a, and 5a involve the
fluoropyridyl group only. We have extended the series in two
ways to allow us to explore the influence of the phosphine ligand in
halogen bonding. Complex 2 could be compared to its triethyl-
phosphine analogue 1 that was the subject of a previous
communication,14 while complex 4a can be compared to its tri-
isopropylphosphine analogue 4b.
The structural studies also required comparison of Pt com-
plexes with the closest possible Ni analogues. We had previously
determined the structure of the platinum bis(tricyclopentyl-
phosphine) complex 4c,20b but crystals of 4a (for which an exact
Ni analogue exists) were of poor quality. To afford a direct com-
parison, we have therefore also synthesized the nickel analogue
of 4c, trans-[Ni(F){2-C5NF2H(CF3)}(PCyp3)2] (5c) (Cyp =
cyclopentyl).
Preparation of Nickel Fluoride Complexes and Structural
Studies. The nickel fluorides 2, 5a, and 5c were obtained directly
from Ni(COD)2 by reaction with two equivalents of the phos-
phine and the appropriate fluorinated pyridine in n-hexane, a
protocol reported previously for other nickel monofluoride com-
plexes (Scheme 1).19 Crystallization from n-hexane at ꢀ25 ꢀC
afforded yellow crystals of complex 2 suitable for X-ray analysis.
Crystalline samples of compounds 5a and 5c were obtained from
saturated toluene solutions at room temperature.
Complex 2 was assigned as a metal fluoride on the basis of the
19F resonance at δ ꢀ378.0. The four resonances due to the
tetrafluoropyridyl fragment appear as multiplets of equal integration
and resemble those of the known complex 1.19 A doublet was
observed in the 31P NMR spectrum at δ 19.3 (JPF = 39 Hz) due
to the two phosphines coupled to the fluoride ligand. The
presence of the nickel fluoride was demonstrated by mass
spectrometry (LIFDI+), showing the molecular ion at m/z
787.6.22 The spectroscopic data of complex 5a resemble those
of 2 (19F: NiF δ ꢀ375.1 and 31P: PCy3 δ 19.0, JPF = 37 Hz).
The 1H NMR signal of the trifluoromethyl-substituted pyridine
ligand was observed downfield at δ 8.50, which is close to the
value found before in the isostructural Pt complex (δ 8.54).20b
The electron-withdrawing nature of the CF3 group causes a
The molecular structures of complexes 2, 5a, and 5c are shown
in Figures 1 and 2. The nickel centers are coordinated in a slightly
distorted square-planar geometry. There is disorder between
N(1) and C(4)ꢀF(5) in complex 2 which corresponds to
rotation about Ni(1)ꢀC(1), as found in analogues studied
previously. The CF3 group in each of 5a and 5c is also disordered.
In addition, there is a ring-flip disorder in the cyclopentyl groups
of 5c. Bond distances and angles are in the expected range.
Notably, NiꢀF distances (Ni(1)ꢀF(1) 1.8518(15), 1.8551(11),
and 1.8591(10) Å for 2, 5a, and 5c, respectively) are similar to
those found for other square-planar nickel monofluoride com-
plexes (e.g., 1.856(2) Å in trans-[Ni(F)(C5NHF3)(PEt3)2].19
The nickelꢀfluorine bonds are significantly shorter than the
MꢀF bonds in the similar palladium and platinum complexes 3
and 4c (2: 1.8518(15), 3: 2.041(3) Å,20a 4c: 2.0402(19) Å20b).
The angle between the coordination plane C(1)P(1)P(2)F(1)
and the C5N plane of the fluoropyridyl ring is 90.00ꢀ for 2,
88.30(8)ꢀ for 5a, and 83.43(6)ꢀ for 5c.
A comparison of the structural data of complexes 5a and 5c
with those of the known platinum tricyclopentylphosphine com-
plex trans-[Pt(F){2-C5NF2H(CF3)}(PCyp3)2] (4c)20b (analogue
of 4a) nicely reveals the influence of the metal on the MꢀX
distances. The lengthening of the MꢀF distance from nickel to
platinum is more pronounced than the lengthening of the other
bonds in the first coordination sphere (Table 1).23 The formation
and structures of bis(phosphine) substituted group 10 metal
fluorides were investigated earlier by theoretical methods,24 where
the exceptional increase in the MꢀF bond length on replacing
nickel by platinum was first predicted.24f Table 1 includes a com-
parison of calculated structures of trans-[M(F){2-C5NF2H(CF3)}-
(PMe3)2] (M = Ni, Pt) with the same pyridyl ligand as in the
experimental systems 4 and 5 but a simplified phosphine. The
differences Δdcalc between the Pt and Ni bond lengths are remark-
ably close to the experimental data. These calculations are consistent
14340
dx.doi.org/10.1021/ja203320y |J. Am. Chem. Soc. 2011, 133, 14338–14348

Journal of the American Chemical Society
ARTICLE
with our earlier studies that indicated that these PtꢀF bonds are
weaker than NiꢀF bonds, in contrast to the usual trends where
PtꢀX (X = H, C, P) bonds are usually stronger than the
corresponding NiꢀX bonds.24f
the ring and the resulting high electrophilicity of the iodine
atom.25 Upon addition of the halogen bond donor, the 19F NMR
fluoride signal of the corresponding metal fluoride of all the
complexes exhibited a substantial downfield shift reaching ca.
10 ppm at room temperature and ca. 20 ppm at low temperature
with excess C6F5I (Figure S11, SI). In contrast, the 19F reso-
nances for the fluorinated pyridine ligand and the 31P NMR
signals for the phosphine ligands remain almost unchanged
(maximum change at room temperature 0.24 and 0.39 ppm,
respectively). The 19F resonances of C6F5I are also shifted
negligibly since C6F5I is present in excess. Thus, the formation
of the halogen-bonded adduct can be monitored by 19F NMR
spectroscopy using the fluoride resonance. The ratio of the metal
fluoride to C6F5I concentrations was varied, giving titration
curves that can be fitted using the simple model of eq 1 to yield
the equilibrium constant for formation of the adduct K and the
change in the 19F chemical shift between the free metal fluoride
and the pure adduct (Δδfit):
Halogen Bonding of Metal Fluoride Complexes and C6F5I.
Titration experiments were carried out in protio-toluene using
the series of structurally related metal fluoride complexes shown
in Chart 2 (other than 5c) and C6F5I. The latter has been found
to be a very good halogen bond donor due to the fluorination of
Thermodynamic data (ΔHo and ΔSo) can be obtained by
recording titration curves at various temperatures and plotting
ln K vs Tꢀ1 (Van’t Hoff Plot). In all cases, the data could be fitted
satisfactorily assuming exclusive formation of a 1:1 adduct. The
1:1 stoichiometry was confirmed by Job plots for complexes 2
and 4a. For the Job plots, we measured the change in chemical
shift of the 19F resonance (Δδ) as a function of the mol fraction
of metal complex (X) while keeping the sum of the concentra-
tions [metal complex] + [C6F5I] = 0.01 mol dmꢀ3. The graph of
XΔδ vs X showed a maximum at X = 0.5 as anticipated for a 1:1
complex (Figures S1 and S6 in SI).26 In our previous commu-
nication, the formation of a 1:2 adduct was observed when
changing the solvent from toluene to n-heptane14 to prevent
interaction of the perfluorinated arene with the π-electron
system of the solvent molecules.27 However, this was not
possible for the examples described here due to the limited
solubility of the metal fluorides in nonpolar solvents.
As an example, Figure 3 shows the titration data of the
experiment using the system 2/C6F5I. The thermodynamic
values for all experiments can be found in Table 2. All titration
curves show a significant temperature dependence of the chemi-
cal shift caused by the interaction with the halogen bond donor.
This effect is distinct for higher concentrations of C6F5I, whereas
for lower concentrations the difference in chemical shifts is small.
Analysis of the titration data gave excellent fits in all cases,
resulting in Van’t Hoff plots with correlation coefficients (R2) of
Figure 2. Molecular structures of complexes 5a (above) and 5c
(below). Hydrogen atoms are omitted for clarity. The displacement
ellipsoids correspond to 50% probability. Disorder of the CF3 groups is
shown for 5a and 5c. Three of the six cyclopentyl groups in 5c are also
disordered. Selected bond lengths (Å) and angles (deg.): 5a Ni(1)ꢀ
F(1) 1.8551(11), Ni(1)ꢀP(1) 2.2249(5), Ni(1)ꢀP(2) 2.2242(5), Ni(1)ꢀ
C(1) 1.8732(19), C(1)ꢀNi(1)ꢀP(1) 94.36(6), F(1)ꢀNi(1)ꢀP(1)
87.66(4), C(1)ꢀNi(1)ꢀP(2) 92.52(6), F(1)ꢀNi(1)ꢀP(2) 85.62(4).
5c Ni(1)ꢀF(1) 1.8591(10), Ni(1)ꢀP(1) 2.2262(5), Ni(1)ꢀP(2)
2.2299(5),Ni(1)ꢀC(1) 1.8613(17), C(1)ꢀNi(1)ꢀP(1) 96.09(5), F(1)ꢀ
Ni(1)ꢀP(1) 85.56(3), C(1)ꢀNi(1)ꢀP(2) 95.32(5), F(1)ꢀNi(1)ꢀ
P(2) 82.94(3).
Table 1. Comparison of MꢀX Bond Distances (Å) in Nickel and Platinum Fluoride Complexes trans-[M(F){2-C5NF2H(CF3)}-
(PR3)2] (M = Ni, R = Cy or Cyp; M = Pt, R = Cyp) by X-ray Crystallography and DFT-Optimized Structures for M = Ni, R = Me;
M = Pt, R = Me
experimental distances/Å
calculated distances/Å
M = Pt
MꢀX
4c20b
5a
5c
Δdexp (4cꢀ5a)
Δdexp (4cꢀ5c)
M = Ni
Δdcalc
MꢀC
1.988(3)
1.8732(19)
1.8551(11)
2.2249(5)
2.2242(5)
1.8613(17)
1.8591(10)
2.2262(5)
2.2299(5)
0.115(4)
0.185(2)
0.0975(9)
0.0983(9)
0.127(4)
0.181(2)
0.0992(9)
0.0926(9)
1.91
1.83
2.30
2.31
2.00
2.05
2.38
2.37
0.09
0.22
0.08
0.06
MꢀF
2.0402(19)
2.3224(8)
2.3225(7)
MꢀP(1)
MꢀP(2)
14341
dx.doi.org/10.1021/ja203320y |J. Am. Chem. Soc. 2011, 133, 14338–14348

Journal of the American Chemical Society
ARTICLE
greater than 0.99 (Figure 4, see Figures S2ꢀS5 and S7ꢀS10 in SI
for titration curves and Van’t Hoff plots for other complexes).
In the titration experiment with complex 4a and C6F5I, the
platinum fluoride signal was absent for higher concentrations of
C6F5I ([C6F5I]/[4a] = 5.3). Evidently, the platinum fluoride
reacted with C6F5I to form a new complex, which has not been
characterized yet. However, since the chemical shifts for low
ratios of donor/metal fluoride determine the quality of the fit,
thermodynamic parameters should still be reliable.
Complexes 4a and 5a differ only in the metal, but the results
show a markedly higher binding enthalpy for platinum than for
nickel. The palladium complex 3 has an enthalpy of formation
that lies between the values of the nickel and platinum complex,
but a direct comparison is not possible because its fluoropyridyl
group differs from those in 4a and 5a. However, the nickel
complexes 2 and 5a offer an opportunity to determine indepen-
dently the influence of the substitution pattern at the fluoropyr-
idyl ligand on the thermodynamic parameters, since they have
identical phosphines but different fluoropyridyl ligands. The
difference in values for ΔHo and ΔSo for compounds 2 and 5a
is insignificant, suggesting that the influence of the different
fluoropyridyl ligand on halogen bonding is negligible. Thus, a
comparison of the tetrafluoropyridyl-substituted complexes 2
and 3 with complexes 4a and 5a, bearing a trifluoromethyl-
substituted pyridine ligand is valid. These arguments enable us to
deduce that the enthalpy of halogen bonding, ꢀΔHo, increases in
the order Ni < Pd < Pt. There is evidence that ꢀΔHo correlates
with ꢀΔSo, leading to compensation in the values of ΔGo and K.
However, the correlation is not close, and K300 varies as Ni ≈ Pt < Pd,
in contrast to the trend in ꢀΔHo. There is little change between
the metals in the chemical shift difference between free metal
fluoride and that of the adduct, Δδfit (Table 2).
The electronic influence of the phosphine substituents of the
metal fluoride on halogen bonding was investigated by compar-
ing nickel complexes 1 and 2 with PEt3 and PCy3 ligands,
respectively, and by comparing platinum complexes 4a and 4b
with PCy3 and PiPr3 ligands, respectively. In each pair the
fluoropyridyl group is identical. The results show that the
phosphine has a significant influence on ΔHo with the strongest
binding for PCy3 with both nickel and platinum (Table 2). The
experimental titration data and the calculated thermodynamic
data show that the difference in the chemical shifts Δδ is not
correlated with the enthalpy and entropy.
The platinum complexes 4a and 4b offer the opportunity of
monitoring the effect of halogen bonding on the value of the
platinumꢀfluorine coupling constant for the metal fluoride
resonance, JPtF. The data for 4b show that the measured JPtF
grows with increasing [C6F5I]/[MF] and decreasing tempera-
ture (Figure 5). Although there is some scatter in the data due to
broadening of the platinum satellites especially at low tempera-
ture and low [C6F5I]/[MF], the upward trend is unmistakable.
Because of the scatter in the data, we could not determine the
equilibrium constant K and the limiting values of the coupling
constant for the 1:1 adducts, JPtFadd, independently. The data for
257, 268, and 279 K were fitted satisfactorily using the values of K
at those temperatures derived from the chemical shift data,
allowing us to determine JPtFadd. This approach yielded values
of JPtFadd of 338 ( 27, 363 ( 54, and 392 ( 59 Hz, respectively,
compared with 240ꢀ250 Hz for the free molecule in toluene
solution, JPtF0. Considering the large uncertainties in the values
of JPtFadd, it is not appropriate to draw any conclusions concern-
ing the temperature dependence of JPtFadd, but we can conclude
Figure 3. Fit of the titration curves at different temperatures, showing
observed values for the 19F chemical shift of the metal fluoride δF vs ratio
of molar concentrations of C6F5I and complex 2.
Figure 4. Van’t Hoff plot for halogen bonding of C6F5I and complex 2.
Table 2. Thermodynamic Data for Halogen Bonding of Group 10 Metal Fluorides with C6F5I in Toluenea
ΔHo kJ molꢀ1
ΔSo J Kꢀ1 molꢀ1
K300
K200
Δδfit
303K
b
c
Metal fluoride (MF)
[Ni(F)(2-C5NF4)(PEt3)2] (1)14
ꢀ16(1)
ꢀ42(2)
ꢀ54(2)
ꢀ59(1)
ꢀ73(4)
ꢀ54(1)
ꢀ49(2)
3.41(9)a
2.49(16)
5.2(3)
96.6
109
33.7
[Ni(F)(2-C5NF4)(PCy3)2] (2)
ꢀ18.6(4)
ꢀ21.7(4)
ꢀ24.5(11)
ꢀ18.9(4)
ꢀ16.9(6)
36.0
35.5
37.9
35.6
27.7
[Pd(F)(4-C5NF4)(PCy3)2] (3)
385
385
130
[Pt(F){2-C5NF2H(CF3)}(PCy3)2] (4a)
[Ni(F){2-C5NF2H(CF3)}(PCy3)2] (5a)
[Pt(F){2-C5NF2H(CF3)}(PiPr3)2] (4b)
2.9(4)
3.06(15)
2.4(2)
71.5
a Errors are 95% confidence limits and are based on statistics of fits. b Extrapolated to 200 K. c Δδfit = Chemical shift difference between free metal fluoride
and metal fluoride-C6F5I adduct, calculated by the fitting routine.
14342
dx.doi.org/10.1021/ja203320y |J. Am. Chem. Soc. 2011, 133, 14338–14348

Journal of the American Chemical Society
ARTICLE
Figure 5. Plot of the spinꢀspin coupling constant JPtF vs the concen-
tration ratio [C6F5I/[PtF] for 4b measured in toluene.
that formation of the C6F5I
FꢀPt adduct results in a very
3 3 3
add
large increase in JPtF (JPtF . JPtF0). The data for 4a show
similar trends, but the broadening is too severe for quantitative
analysis.
Calculations on isotropic spinꢀspin coupling constants of
binary metal fluorides have shown that both the Fermi Contact term
and the paramagnetic spinꢀorbit term make major contributions
to JMF and that they may have opposite signs.28 Thus, a simple
inverse correlation between adduct strength and JPtF is not to be
expected.29
Computational Studies of Halogen Bonding. The opti-
mized structures (BHandHLYP) of the model metal fluo-
ride complexes trans-[Pt(F){2-C5NF2H(CF3)}(PMe3)2] and
trans-[Ni(F){2-C5NF2H(CF3)}(PMe3)2], along with their
C6F5I adducts, are summarized in Figure 6. The adduct geome-
Figure 6. Optimized structures of trans-[Ni(F){C5NF2H(CF3)}(PMe3)2],
trans-[Pt(F){C5NF2H(CF3)}(PMe3)2], and their C6F5I adducts. Dis-
tances shown in Å.
the electrostatic potentials for the nickel and platinum complexes.
On descending group 10 from Ni to Pt, we naturally anticipate an
increase in metalꢀligand bond lengths, but we have shown (both
experimentally and computationally, see above) that MꢀF
bonds are affected to a greater extent than comparable MꢀC
bonds. This lengthening indicates that the PtꢀF bond is some-
what weaker than its NiꢀF counterpart, an effect that we ascribe
to large 5dπꢀpπ repulsions.24f,32 The electrostatic potentials
(Figure 7) suggest that the region of negative potential around
the fluoride ligand is slightly more radially extended in the
platinum case, and the potential minimum along the extension of
tries describe a CꢀI FꢀM halogen bond in which the halogen
3 3 3
bond distance I F is substantially shorter than the correspond-
3 3 3
ing sum of van der Waals radii (3.45 Å)30 (I F(Pt) 2.73 Å, RIF
3 3 3
0.79; I F(Ni) 2.78 Å, RIF 0.81).31 In both adducts the Cꢀ
3 3 3
I
I
F angles are close to linear (Pt 179ꢀ, Ni 180ꢀ). The
FꢀM moiety is slightly bent for Ni (167ꢀ), but distinctly
3 3 3
3 3 3
bent for Pt (145ꢀ). Such angles are consistent with observations
in a survey of H FꢀM angles for hydrogen bonding involving
3 3 3
the MꢀF vector is deeper for platinum fluoride (ꢀ302 kJ molꢀ1
)
metal fluorides.1a We note, also, that the I FꢀM bending
3 3 3
than nickel fluoride (ꢀ280 kJ molꢀ1). Collectively, the differ-
ences allow for a more effective electrostatic interaction with the
σ-hole of the iodine center in the case of the platinum fluoride.
A complementary perspective on halogen bonding can be
obtained from an analysis of the topology of the electron density
modes have very low frequencies. The formation of the adducts
results in a lengthening of both MꢀF and CꢀI bonds, although
the effects are marginal. Otherwise, the structures of both
subunits remain unperturbed by the intermolecular interaction.
The counterpoise-corrected interaction enthalpies (defined as
ΔH(adduct) ꢀ ΔH(C6F5I) ꢀ ΔH(complex)) are ꢀ18.8 and
ꢀ22.8 kJ molꢀ1 for Ni and Pt, respectively, in excellent agreement
with the experimentally determined enthalpies of ꢀ18.9 and ꢀ24.5
kJ molꢀ1 for 5a and 4a, respectively. Thus, the observed trend
toward stronger bonding in the platinum complex is reproduced
by the calculations with the binding energy calculated to be 4.0 kJ
molꢀ1 greater than that in the nickel analogue. Although this
energy difference is small, the very similar structures of the two
adducts under consideration means that sources of error (finite
basis set, solvation) should be the same in both cases, providing
some confidence that this number is meaningful.
along the I F axis.33 In both nickel and platinum cases a bond
3 3 3
critical point (BCP) is present between F and I, with a rather low
electron density (∼0.02 e Åꢀ3) and a positive value of the
2
Laplacian, r F, both of which are typical of closed-shell inter-
actions. Significantly, however, the electron density at the critical
point is marginally greater in the platinum case than in nickel
(0.024 vs 0.020 e Åꢀ3). Sosa et al.34 and Zou et al.35 have
independently established approximately linear relationships
between electron density at the critical point and interaction energy
in a series of halogen-bonded adducts, with gradients in the range
1400ꢀ1800 kJ Å3 molꢀ1
e
ꢀ1. Assuming a gradient of similar
magnitude in this case, the difference in F of 0.004 e Åꢀ3 between
nickel and platinum complexes equates to a difference in inter-
action energy of 5ꢀ7 kJ molꢀ1, in excellent agreement with experi-
ment and computed interaction energies. Thus, the subtle
Halogen bonding has usually been discussed in the context of
an electrostatic model, whether referring to halogen bonds with
MꢀX, X = Cl, Br, I,3h,i or halogen bonds involving organic
bases.12,13 We investigated an electrostatic model by calculating
14343
dx.doi.org/10.1021/ja203320y |J. Am. Chem. Soc. 2011, 133, 14338–14348

Journal of the American Chemical Society
ARTICLE
the metal fluoride resonance is very sensitive to adduct forma-
tion, moving ∼35 ppm downfield in the limit for the 1:1 adduct.
There is no evidence for a 1:2 adduct. The value of JPtF increases
from about 245 in the free molecule to about 360 Hz in the
adduct of trans-[Pt(F){2-C5NF2H(CF3)}(PiPr3)2] 4b. We have
shown that a comparison of the thermodynamics of binding for
these structurally similar complexes is acceptable due to the
negligible influence of the pyridyl substituents on the thermo-
dynamic data. The identity of the phosphine has a much more
marked influence on the thermodynamics. The value of ꢀΔHo
increases markedly from Ni to Pd to Pt (ꢀΔHo Ni 18.9, Pd 21.7,
Pt 24.5 kJ molꢀ1). The values of ΔSo range from ꢀ54 to
ꢀ73 J Kꢀ1 molꢀ1 for the corresponding complexes, but the
correlation of ΔSo with ꢀΔHo is less than perfect, rendering it
hard to discern systematic trends. The trends in the equilibrium
constants are significantly different from those in ꢀΔHo, making
the value of temperature-dependent data clear. At 300 K, the
equilibrium constants range from 2.4 to 5.2, while at 200 K they
range from 72 to 385. Although the biggest equilibrium constant
at 300 K is found for palladium complex 3, the biggest equilib-
rium constant below 200 K belongs to platinum complex 4a. The
trends in halogen bonding down a transition metal group may be
compared with trends in dihydrogen bonding. The best examples
are probably those of M(P(CH2CH2PPh2)3(H)2 (M = Fe, Ru,
Os) interacting with alcohols which also demonstrate a trend of
increasing ꢀΔHo down the group.17d
The DFT studies of halogen bonding provide estimates of the
interaction enthalpy of halogen bonding of ꢀ18.8 kJ molꢀ1 and
ꢀ22.8 kJ molꢀ1 for models of 5a (Ni) and 4a (Pt), respectively,
using PMe3 in place of PCy3. These compare with the experi-
mental values of ΔHo of ꢀ18.9(4) and ꢀ24.5(11) kJ molꢀ1 for
5a (Ni) and 4a (Pt). The calculated I F(M) distance is slightly
3 3 3
shorter for M = Pt, in keeping with the stronger halogen bond.
The calculations show a bond critical point between F and I, and
a very marked region of negative potential around fluorine.
In summary, we have found that these group 10 metal
fluorides are excellent halogen bond acceptors, giving enthalpies
for these interactions that are even greater than the value
obtained in previous studies for trans-[Ni(F)(2-C5NF4)-
(PEt3)2]. Moreover, these data provide quantitative data that
may be compared with models and imply that halogen bonding is
a force comparable in strength to hydrogen bonding.
Figure 7. Electrostatic potentials in the metal coordination plane,
shown with zero and negative contours at intervals of 5 kcal molꢀ1
(20.9 kJ molꢀ1) for (upper panel) trans-[Ni(F){C5NF2H(CF3)}-
(PMe3)2] and (lower panel) trans-[Pt(F){C5NF2H(CF3)}(PMe3)2.
Red indicates most negative potential, blue is most positive. Views in
the plane of the pyridine ring are shown in Figure S14 (SI).
differences in interaction energy are consistent with the accumu-
lation of greater charge in the intermolecular region in the
platinum case.
’ EXPERIMENTAL SECTION
’ CONCLUSION
General Procedures. All operations were performed under an
argon atmosphere, either on a high-vacuum line (10ꢀ4 mbar) using
modified Schlenk techniques, on standard Schlenk lines (10ꢀ2 mbar) or
in a glovebox. Solvents for general use (n-hexane, toluene) were of AR
grade, dried by distillation over standard reagents, or with a solvent
purification system, and stored under Ar in ampules fitted with Young’s
PTFE stopcocks.
Deuterated solvents were dried by stirring over potassium, and were
distilled under high vacuum into small ampules with potassium mirrors.
All NMR spectra were recorded on Bruker AMX500 spectrometers in
the same type of tube. All 1H NMR spectra were recorded at 500.2 MHz;
chemical shifts are reported in ppm (δ) relative to tetramethylsilane and are
referenced using the chemical shifts of residual protio solvent resonances
(benzene, δ7.16). The 31P{1H} NMR spectra were recorded at 202.5 MHz
and are referenced to external H3PO4. 19F NMR spectra were recorded
at 470.5 MHz and referenced to external CFCl3 at δ 0 or internal C6F6 at
δ ꢀ162.9. The temperature of the probe was calibrated according to
We have established a series of structurally similar group 10
metal fluorides bearing two mutually trans PCy3 ligands by
synthesis of three new nickel fluoride complexes by CꢀF
oxidative addition. The Ni and Pt complexes are isostructural,
but the palladium complex has a different pyridyl substituent that
results from a change in selectivity of the pathway of oxidative
addition of the fluorinated pyridine.20 We have examined quan-
titatively the extension of bond lengths for platinum compounds
relative to related nickel compounds by experiment and compu-
tation. The biggest extension of the PtꢀX bonds (X = C, P, F)
relative to the NiꢀX bonds is found for X = F, probably as a result
of pπꢀdπ repulsive interactions.24f
We have measured the 19F NMR chemical shifts of these
complexes dissolved in toluene in the presence of excess C6F5I
from 250 to 300 K and used the data to determine the thermo-
dynamics of halogen bond formation. The 19F chemical shift of
14344
dx.doi.org/10.1021/ja203320y |J. Am. Chem. Soc. 2011, 133, 14338–14348

Journal of the American Chemical Society
ARTICLE
published procedures.36 Mass spectra were recorded by the University of
York analytical services on a Waters GCT instrument fitted with a
Linden LIFDI probe and are quoted for 58Ni. Infrared spectra were
recorded as KBr disks (prepared in the glovebox) on a Unicam-Mattson
Research Series FTIR spectrometer fitted with a KBr beam splitter. The
sample chamber was purged with dry, CO2-free air. Samples for
elemental analysis were prepared in a glovebox, sealed under vacuum,
and measured by Elemental Microanalysis Ltd., Okehampton. Chemicals
were obtained from the following sources: bis(1,5-cyclooctadiene)-
nickel(0), iodopentafluorobenzene, and potassium tetrachloroplatinate
from Aldrich (or recovered from platinum residues by a standard procedure);
allylpalladium chloride dimer, PiPr3, and PCy3 from Strem, pentafluoro-
pyridine and 2,3,5-trifluoro-4-trifluoromethyl-pyridine from Fluorochem.
trans-[Pd(F)(4-C5NF4)(PCy3)2] (3),20a trans-[Pt(F){2-C5NF2H(CF3)}-
(PCy3)2] (4a),20b and trans-[Pt(F){2-C5NF2H(CF3)}(PiPr3)2] (4b)20b
were synthesized according to known literature procedures.
ꢀ145.85 (q, 1F, JFF = 22.2 Hz, p-F), ꢀ102 (q, 1F, JFF = 22.5 Hz o-F),
41 Hz). MS m/z (LIFDI+): 735.3 (M+) 100%.
ꢀ57.2 (t, JFF = 20 Hz, 3F, CF3), 31P (300 K, benzene-d6): δ 17.7 (JPF
=
X-ray Crystallography. Diffraction data were collected on a
Bruker SMART APEX I diffractometer with Mo KR radiation (λ =
0.71073 Å). Diffractometer control, data collection, and initial unit cell
determination were performed using “SMART” (v5.625, Bruker AXS).
Frame integration and unit cell refinement were carried out with
“SAINT+” (v6.22, Bruker AXS). Absorption corrections were applied
empirically (SADABS, v.2.10) based upon symmetry-equivalent reflec-
tions combined with measurements at different azimuthal angles.37 The
structures were solved by direct methods using SHELXS-97 and refined
by full-matrix least-squares methods using SHELXL-97.38 Hydrogen
atoms were placed at calculated positions and were included in the
refinement using a riding model. For complex 2, the crystals exhibited
disorder of both the perfluoropyridyl ligand and the toluene of crystal-
lization. Each was modeled over two sites with relative occupancy
63.9:36.1(6). For the toluene, the aromatic CꢀC bond lengths re-
strained to be equal. For complex 5a, the structure was modeled with the
CF3 group disordered over two positions in a 81.7:18.3(4) ratio. The
CꢀF bonds of this group were restrained to 1.34 Å and the distances
between F atoms of the same residue restrained to be equal. The atomic
displacement factor of F(5a) was restrained to be approximately
isotropic. Large ADP max/min ratios are explained as a consequence
of rotation about the CF3 axis. For complex 5c, the CF3 group was
disordered and modeled over two positions the relative occupancy of
which refined to 63.7:36.3(4). Three of the cyclopentyl groups exhibited
disorder and for each group one carbon atom was modeled in two
positions (C(19/19a), C(25/25a), and C(35/35a)) the relative occu-
pancy of which refined to 0.545:0.455(7), 0.736:0.264(8), and
0.558:0.442(8), respectively. See Table 3 for data for 2, 5a, and 5c.
NMR Titrations and Analysis of Data. The equilibrium con-
stants were determined through NMR titration at a series of tempera-
tures, by following the chemical shift of the fluoride ligand coordinated
to the transition metal. The volumes of the solutions were assumed to be
the sum of the volumes of the components, thereby enabling the
densities of the solutions to be calculated. Test solutions of iodopenta-
fluorobenzene in protio-toluene, in concentrations similar to those used
in the titrations, showed no significant deviation from this assump-
tion. The activities of the species were assumed equal to their molar
concentration. The calculations for the equilibrium constants were
carried out with Microsoft Excel, using a macro programmed by Prof.
Christopher A. Hunter of the University of Sheffield. There are two
parameters to be fitted: the equilibrium constant K and the downfield
shift from the signal of free metal fluoride for the coordinated fluoride in
the adduct, ΔδFit. The two parameters can be fitted for the whole range
of temperatures without any restraints. ΔHo and ΔSo were calculated
from the Van’t Hoff plots of the equilibrium constants.
trans-[Ni(F)(2-C5NF4)(PCy3)2] (2). Ni(COD)2 (0.500 g, 1.82 mmol)
and PCy3 (1.274 g, 4.54 mmol) were suspended in 15 mL of n-hexane. The
orange-red mixture was stirred for 10 min at room temperature, and
pentafluoropyridine (0.240 mL, 2.19 mmol) was added at this temperature.
The color of the reaction mixture changed to orange. After 2 h of stirring at
room temperature, the suspension was filtered to give an orange solution.
Storage at ꢀ25 ꢀC overnight yielded yellow crystals, which were isolated,
washed with cold n-hexane, and dried under vacuum; yield: 0.94 g (1.19
mmol, 66%). Anal. Calcd for C41H66F5NNiP2 (rmm 788.6): C, 62.44; H,
8.44; N, 1.78. Found: C, 62.92; H, 7.80; N, 1.86. IR (KBr disk, cmꢀ1): 1620
(m), 1595 (m), 1581 (s), 1481 (s), 1449 (s), 1420 (s), 1402 (s), 1381 (s),
1342 (m), 1266 (m), 1199 (m), 1174 (m), 1125 (m), 1085 (s), 1048 (w),
1003 (m), 994 (vs), 915 (m), 848 (s), 806 (vs), 736 (s), 711 (w), 676 (w),
605 (w), 530 (s), 513 (s), 491 (m), 450 (w), 407 (w). NMR: 1H (300 K,
benzene-d6): δ0.8ꢀ2.2 (m, C6H11). 19F (300 K, benzene-d6): δꢀ378.0 (t,
J
PF = 37 Hz, 1F, NiF), ꢀ175.1 (m, 1F, F5), ꢀ152.7 (m, 1F, F4), ꢀ126.6
(t, JFF = 28 Hz, 1F, F3), ꢀ85.8 (m, 1F, F6). 31P (300 K, benzene-d6): δ 19.3
(d, JPF = 37 Hz). MS, m/z (LIFDI+): 787.6 (M)+ 100%, 769.6 (MH ꢀ
F)+ 16%.
trans-[Ni(F){2-C5NF2H(CF3)}(PCy3)2] (5a). Ni(COD)2 (0.248 g,
0.90 mmol) and PCy3 (0.505 g, 1.80 mmol) were suspended in 10 mL
of n-hexane. The orange-red mixture was stirred for 10 min at room
temperature, and 2,3,5-trifluoro-4-trifluoromethyl-pyridine (0.200 g, 0.99
mmol) was added at this temperature. The color of the reaction mixture
changed to red, and after one hour a yellow precipitate formed. The solid
was separated by filtration, washed with n-hexane, and dried in vacuum.
Crystals suitable for X-ray analysis were obtained from saturated toluene
solutions at room temperature; yield: 0.589 g (0.72 mmol, 79%). Anal.
Calcd for C42H67F6NNiP2 (rmm 820.62): C, 61.47; H, 8.23; N, 1.71.
Found: C, 61.01; H, 8.24; N, 1.60. IR (KBr disk, cmꢀ1): 1595 (m), 1542
(m), 1466 (m), 1446 (s), 1352 (s), 1340 (s), 1296 (s), 1266 (m), 1227
(w), 1194 (m), 1171 (s), 1143 (s), 1108 (m), 1079 (s), 1048 (w), 1025
(s), 1003 (m), 914 (w), 887 (w), 848 (m), 804 (w), 766 (w), 736 (s), 708
(s), 664 (s), 649 (m), 575 (s), 563 (m), 511 (s), 489 (m), 450 (w). NMR:
1H (300 K, benzene-d6): δ 0.8ꢀ2.3 (m, 66H, C6H11), 8.50 (s, 1H,
H-pyridyl). 19F (300 K, benzene-d6): δ ꢀ375.1 (t, JPF = 39 Hz, 1F, NiF),
ꢀ146.3 (q, JFF = 21 Hz, 1F, p-F), ꢀ101.0 (q, 1F, JFF = 20 Hz, o-F), ꢀ56.8
(t, 1F, JFF = 21 Hz, CF3). 31P (300 K, benzene-d6): δ 19.0 (d, JPF = 39 Hz).
MS, m/z (LIFDI+): 819.4 (M+) 100%.
trans-[Ni(F){2-C5NF2H(CF3)}(PCyp3)2] (5c). This complex was
prepared in a fashion identical to that for the synthesis of (5a). Yield 0.97
g (1.22 mmol, 72%). Anal. Calcd for C36H55F6NNiP2 (rmm 736.46): C,
58.71; H, 7.53; N, 1.90. Found C, 58.52; H, 7.72; N, 1.86. IR (KBr
disk, cmꢀ1): 2952 (vs), 2868 (s), 1996 (w), 1540 (w), 1449 (m sharp),
1355 (s sharp), 1297 (vs sharp), 1166 (m), 1140 (vs), 1080 (w), 1025
(m), 921 (w), 804 (w), 708 (w) 665 (w), 548 (w), 500 (m). NMR 1H
(300 K, benzene-d6) δ 1.5ꢀ2.3 (m, 54H, C5H9), 8.52 (s, 1H, H-Pyridyl).
19F (300 K, benzene-d6): δ ꢀ377.62 (t, 1F, JPF = 40.8 Hz, PtꢀF),
General Procedure for the Preparation of the NMR Sam-
ples. The solutions for titration were prepared as follows. A solution of
known concentration of the metal fluoride was prepared by dissolving
the complex in toluene, and a solution of C6F5I (the titrant solution, also
of known concentration, see Table S1 in SI) by dissolving C6F5I in
toluene. Approximately equal amounts of the solution of the complex
were transferred into each of ten labeled NMR tubes in the glovebox; the
mass of the solution contained in each tube was recorded. In one sample,
no C6F5I solution was added, providing a reference; in the other tubes,
variable amounts of the solution of C6F5I were added by volume, in
quantities ranging from 5 μL up to 160 μL. The 19F NMR spectra of all
samples were recorded at various temperatures. The samples were kept
in a bath close to the temperature of the probe and left to equilibrate
inside the probe for two minutes before the spectrum was recorded. The
19F NMR spectra were collected unlocked. However, for each tempera-
ture the spectrometer was shimmed with a solution of the corresponding
14345
dx.doi.org/10.1021/ja203320y |J. Am. Chem. Soc. 2011, 133, 14338–14348

Journal of the American Chemical Society
ARTICLE
Table 3. Crystallographic Data for Complexes 2, 5a, and 5c
2 toluene
5a
5c
3
cryst syst
space group
formula
a [Å]
orthorhombic
monoclinic
P21/c
monoclinic
Pmn21
P21/n
C41H66F5NNiP2 C7H8
C42H67F6NNiP2
13.6755(13)
18.4389(17)
17.6985(16)
111.439(2)
4154.1(7)
4
C36H55F6NNiP2
14.5066(3)
12.3458(2)
19.7390(3)
91.3446(17)
3534.18(12)
4
3
13.9158(12)
16.9189(14)
9.7820(8)
90.00
b [Å]
c [Å]
β [deg]
V [Å3]
2303.1(3)
2
Z
density [g cmꢀ3
]
1.270
1.312
1.384
μ(Mo KR) [mmꢀ1
]
0.544
0.601
0.698
T [K]
110(2)
25854
110(2)
110(2)
no. of rflns (measd)
no. of rflns (indep)
42705
19381
6904
10329
10333
no. of rflns (obsd and used in refinement, n)
5924
8147
8174
no. of params, p
332
497
455
GOF on F2 (all data)
R1(F) (I > 2σ(I))
wR2(F2) (all data)
1.016
1.036
1.038
0.0348
0.0816
831711
0.0434
0.0398
0.1139
0.0831
CCDC number
831710
831709
metal fluoride in toluene-d8 and maintained with the same settings
throughout.
’ AUTHOR INFORMATION
Corresponding Author
robin.perutz@york.ac.uk
For the Job plots, a stock solution of C6F5I in toluene of concentra-
tion 0.01 mol dmꢀ3 was prepared in the glovebox. A stock solution of the
metal complex was prepared with the same concentration. NMR tubes
were made up with measured aliquots of each solution such that the total
volume in the NMR tube was 0.400 mL.
Present Addresses
^Leibniz-Institut f€ur Katalyse e.V. an der Universit€at Rostock,
Albert-Einstein-Strasse 29a, 18059 Rostock, Germany.
Computational Methods. All calculations were performed using
density functional theory as implemented in the Gaussian03 package.39
The BHandHLYP functional was used in conjunction with the SDD
effective core potential and associated basis set on P, Ni, and Pt and a
6-31G* basis on all main group atoms. This functional has been shown to
provide accurate energetics for noncovalently bonded systems.40 Diffuse
functions were also added to all fluorine atoms (6-31+G*). All phos-
phine ligands were modeled as PMe3. Equilibrium structures were fully
optimized, including counterpoise corrections, starting from a geometry
where the CꢀI and FꢀM groups were colinear. An ultrafine grid was
used together with tight optimization criteria. Frequency calculations
confirmed all reported stationary points to be minima, although in the
case of the adducts, three low-lying modes were present that related to
the bending of the C6F5I unit relative to the metal complex. Interaction
enthalpies were corrected for basis set superposition error (BSSE) using
the counterpoise correction of Boys, and also for the effect of zero-point
energies.
’ ACKNOWLEDGMENT
This work was supported by the Deutsche Forschungsge-
meinschaft (research fellowship for T.B., Grant No. BE 4370/1-1)
and the EPSRC. We thank Dr. Marius V. C^ampian (University
of York), Stefano Libri (University of Sheffield), and Robert J.
Thatcher (University of York) for discussions and suggestions,
and Jason Loader (University of Sheffield) for preparation of
electrostatic potential diagrams. Prof. Christopher A. Hunter
(University of Sheffield) provided the macro used in calculation
of the equilibrium constants and help with the application of the
program.
’ REFERENCES
(1) (a) Brammer, L.; Bruton, E. A.; Sherwood, P. Cryst. Growth Des.
2001, 1, 277. (b) Steiner, T. Angew. Chem., Int. Ed. 2002, 41, 48. (c)
Desiraju, G. R.; Steiner, T. The Weak Hydrogen Bond; Oxford University
Press: Oxford, 1999.
’ ASSOCIATED CONTENT
(2) (a) Metrangolo, P.; Neukirch, H.; Pilati, T.; Resnati, G. Acc.
Chem. Res. 2005, 38, 386. (b) Metrangolo, P.; Resnati, G., Eds. Halogen
Bonding: Fundamentals and Applications, Structure & Bonding; Springer:
Berlin, 2007. (c) Metrangolo, P.; Meyer, F.; Pilati, T.; Resnati, G.;
Terraneo, G. Angew. Chem., Int. Ed. 2008, 47, 6114. (d) Brammer, L.;
Minguez Espallargas, G.; Libri, S. CrystEngComm 2008, 10, 1712. (e)
Metrangolo, P.; Carcenac, Y.; Lahtinen, M.; Pilati, T.; Rissanen, K.; Vij,
A.; Resnati, G. Science 2009, 323, 1461. (f) Fourmiguꢀe, M. Curr. Opin.
Solid State Mater. Sci. 2009, 13, 36. (g) Ritter, S. K. Chem. Eng. News
2009, 87, 39.
S
Supporting Information. Crystallographic data for 2, 5a,
b
and 5c; titration data and Van’t Hoff plots for 3, 4a, 4b and 5b;
Job plots for 2 and 4a; temperature dependence of derived
equilibrium constants for C6F5I adducts; 31P{1H} and 19F NMR
spectra of 5c; 19F NMR spectra of 4b in presence and absence of
C6F5I; author list for Gaussian 03 (ref 39), packing diagram for
5a showing short NiꢀF HꢀC hydrogen bonds. Cartesian coordi-
3 3 3
nates and total energies for all computed structures. This material
is available free of charge via the Internet at http://pubs.acs.org.
14346
dx.doi.org/10.1021/ja203320y |J. Am. Chem. Soc. 2011, 133, 14338–14348

Journal of the American Chemical Society
ARTICLE
(3) (a) Brammer, L.; Swearingen, J. K.; Bruton, E. A.; Sherwood, P.
Proc. Nat. Acad. Sci., U.S.A. 2002, 99, 4956. (b) Mareque Rivas, J. C.;
Brammer, L. Inorg. Chem. 1998, 37, 4746. (c) Lewis, G. R.; Orpen, A. G.
Chem. Commun. 1998, 1873. (d) Banerjee, R.; Desiraju, G. R.; Mondal,
R.; Howard, J. A. K. Chem.—Eur. J. 2004, 10, 3373. (e) Walsh, R. B.;
Padgett, C. W.; Metrangolo, P.; Resnati, G.; Hanks, T. W.; Pennington,
W. T. Cryst. Growth Des. 2001, 1, 165. (f) Bosch, E.; Barnes, C. L. Cryst.
Growth Des. 2002, 2, 299. (g) Brammer, L.; Mínguez Espallargas, G.;
Adams, H. CrystEngComm 2003, 5, 343. (h) Zordan, F.; Brammer, L.;
Sherwood, P. J. Am. Chem. Soc. 2005, 127, 5979. (i) Mínguez Espallargas,
G.; Brammer, L.; Sherwood, P. Angew. Chem., Int. Ed. 2006, 45, 435.
(j) Rosokha, S. V.; Neretin, I. S.; Rosokha, T. Y.; Hecht, J.; Kochi, J. K.
Heteratom. Chem. 2006, 17, 449. (k) Derossi, S.; Brammer, L.; Hunter,
C. A.; Ward, M. D. Inorg. Chem. 2009, 48, 1666.
(4) Laguna, A.; Lasanta, T.; Lꢀopez-de-Luzuriaga, J. M.; Monge, M.;
Naumov, P.; Olmos, M. E. J. Am. Chem. Soc. 2010, 134, 456.
(5) (a) Corradi, E.; Meille, S. V.; Messina, M. T.; Metrangolo, P.;
Resnati, G. Angew. Chem., Int. Ed. 2000, 39, 1782. (b) Sun, A. W.; Lauher,
J. W.; Goroff, N. S. Science 2006, 312, 5776. (c) Shirman, T.; Arad, T.;
van der Boom, M. E. Angew. Chem., Int. Ed. 2010, 49, 926. (d) Serpell,
C. J.; Kilah, N. L.; Costa, P. J.; Felix, V.; Beer, P. D. Angew. Chem., Int. Ed.
2010, 49, 5322. (e) Kilah, N. L.; Wise, M. D.; Serpell, C. J.; Thompson,
A. L.; White, N. G.; Christensen, K. E.; Beer, P. D. J. Am. Chem. Soc.
2010, 132, 11893.
(6) (a) Nguyen, H. L.; Horton, P. N.; Hursthouse, M. B.; Legon,
A. C.; Bruce, D. W. J. Am. Chem. Soc. 2004, 126, 16. (b) Pr€asang, C.;
Whitwood, A. C.; Bruce, D. W. Chem. Commun. 2008, 2137.
(7) (a) Fourmiguꢀe, M.; Batail, P. Chem. Rev. 2004, 104, 5379. (b)
Imakubo, T.; Sawa, H.; Kato, R. J. Chem. Soc., Chem. Commun.
1995, 1667. (c) Yamamoto, H. M.; Kosaka, Y.; Maeda, R.; Yamaura,
J.; Nakao, A.; Nakamura, T.; Kato, R. ACS Nano 2008, 2, 143.
(8) (a) Auffinger, P.; Hays, F. A.; Westhof, E.; Ho, P. S. Proc. Nat.
Acad. Sci. U.S.A. 2004, 101, 16789. (b) Howard, E.; Sanishvili, R.;
Cachau, R. E.; Mitschler, B.; Bart, P.; Lamour, V.; Van Zandt, M.; Sibley,
E.; Bon, C.; Moras, D.; Schneider, T. R.; Joachimiak, A.; Podjarny, A.
Proteins Struct., Funct., Genet. 2004, 55, 792. (c) Parisini, E.; Metrangolo,
P.; Pilati, T.; Resnati, G.; Terraneo, G. Chem. Soc. Rev. 2011, 40, 2267.
(9) (a) Legon, A. C. Angew. Chem., Int. Ed. 1999, 38, 2686. (b)
Legon, A. C. Phys. Chem. Chem. Phys. 2010, 12, 7736.
J.; Kinzel, T.; Buchwald, S. L. Science 2009, 325, 1661. (c) Engle, K. M.;
Mei, T.-S.; Wang, X.; Yu, J.-Q. Angew. Chem., Int. Ed. 2011, 50, 1478.
(17) (a) Richmond, T. G. Coord. Chem. Rev. 1990, 105, 221. (b)
Yandulov, D. V.; Caulton, K. G.; Belkova, N. V.; Shubina, E. S.; Epstein,
L. M.; Khoroshum, D. V.; Musaev, D. G.; Morokuma, K. J. Am. Chem. Soc.
1998, 120, 12553. (c) Lee, D.-H.; Kwon, H. J.; Patel, B. P.; Liable-Sands,
L. M.; Rheingold, A. L.; Crabtree, R. H. Organometallics 1999, 18, 1615.
(d) Gutsul, E. I.; Belkova, N. V.; Sverdlov, M. S.; Epstein, L. M.; Shubina,
E. S.; Bakhmutov, V. I.; Gribanova, T. N.; Minyaev, R. M.; Bianchini, C.;
Peruzzini, M.; Zanobini, F. Chem.—Eur. J. 2003, 9, 2219. (e) Belkova, N. V.;
Epstein, L. M.; Shubina, E.S . Acc. Chem. Res. 2005, 38, 624. (f) Brammer, L.
Dalton Trans. 2003, 3145.
(18) Examples of group 4 metal fluorides of the type Cp2MF2 and
Cp*2MF2: (a) Druce, P. M.; Kingston, B. M.; Lappert, M. F.; Spalding,
T. R.; Srivastava, R. C. J. Chem. Soc. (A) 1969, 2106. (b) Herzog, A.; Liu,
F.-Q.; Roesky, H. W.; Demsar, A.; Keller, K.; Noltemeyer, M.; Pauer, F.
Organometallics 1994, 13, 1251. (c) Murphy, E. F.; Yu, P.; Dietrich, S.;
Roesky, H. W.; Parisini, E.; Noltemeyer, M. J. Chem. Soc., Dalton Trans.
1996, 1983.
(19) Cronin, L.; Higgitt, C. L.; Karch, R.; Perutz, R. N. Organometallics
1997, 16, 4920.
(20) (a) Jasim, N. A.; Perutz, R. N.; Whitwood, A. C.; Braun, T.;
Izundu, J.; Neumann, B.; Rothfeld, S.; Stammler, H.-G. Organometallics
2004, 23, 6140. (b) Nova, A.; Erhardt, S.; Jasim, N. A.; Perutz, R. N.;
Macgregor, S. A.; McGrady, J. E.; Whitwood, A. C. J. Am. Chem. Soc.
2008, 130, 15499.
(21) Nova, A.; Reinhold, M.; Perutz, R. N.; Macgregor, S. A.;
McGrady, J. E. Organometallics 2010, 29, 1824.
(22) Dransfield, T. A.; Nazir, R.; Perutz, R. N.; Whitwood, A. C.
J. Fluorine Chem. 2010, 131, 1213.
(23) (a) A survey of crystal structures (CSD Conquest 1.12; Nov
2009 + 1 update)23b of square-planar bis(phosphines)nickel and bis-
(phosphines)platinum complexes containing a monodentate O-bonded
ligand show the same trend for MꢀO distances, viz. mean NiꢀO 1.899
Å (40 observations) and mean PtꢀO 2.107 Å (114 observations); Δd =
0.208 Å. (b) Allen, F. H. Acta Crystallogr. 2002, B58, 380.
(24) (a) Su, M.-D.; Chu, S.-Y. J. Am. Chem. Soc. 1997, 119, 10178.
(b) Bosque, R.; Clot, E.; Fantacci, S.; Maseras, F.; Eisenstein, O.; Perutz,
R. N.; Renkema, K. B.; Caulton, K. G. J. Am. Chem. Soc. 1998,
120, 12634. (c) Su, M.-D.; Chu, S.-Y. J. Am. Chem. Soc. 1999,
121, 1045. (d) Gꢀerard, H.; Davidson, E. R.; Eisenstein, O. Mol. Phys.
2002, 100, 533. (e) Gꢀerard, H.; Eisenstein, O. Dalton Trans. 2003, 839.
(f) Reinhold, M.; McGrady, J. E.; Perutz, R. N. J. Am. Chem. Soc. 2004,
126, 5268.
(25) Valerio, G.; Raos, G.; Meille, S. V.; Metrangolo, P.; Resnati, G.
J. Phys. Chem. A 2000, 104, 1617.
(26) For examples of similar methodology for Job plots, see (a)
McCall, A. S.; Wang, H.; Desper, J. M.; Kraft, S. J. Am. Chem. Soc. 2011,
1330, 1832. (b) Yoo, H.; Mirkin, C. A.; DiPasquale, A. G.; Rheingold,
A. L.; Stern, C. L. Inorg. Chem. 2008, 47, 9727.
(10) (a) Clark, T.; Hennemann, M.; Murray, J. S.; Politzer, P. J. Mol.
Model 2007, 13, 291. (b) Politzer, P.; Murray, J. S.; Clark, T. Phys. Chem.
Chem. Phys. 2010, 12, 7748.
(11) (a) Mínguez Espallargas, G.; Zordan, F.; Arroyo Marín, L.; Adams,
H.; Shankland, K.; van de Streek, J.; Brammer, L. Chem.—Eur. J. 2009,
15, 7554. (b) Reddy, L. S.; Chandran, S. K.; George, S.; Babu, N. J.; Nangia,
A. Cryst. Growth Des. 2007, 7, 2675. (c) Aaker€oy, C. B.; Fasulo, M.;
Schulteiss, N.; Desper, J.; Moore, C. J. Am. Chem. Soc. 2007, 129, 13772.
(12) (a) Laurence, C.; Queignec-Cabanatos, M.; Dziembowska, T.;
Queignec, R.; Wojtkowiak, B. J. Am. Chem. Soc. 1981, 103, 2567. (b)
Laurence, C.; Queignec-Cabanatos, M.; Wojtkowiak, B. J. Chem. Soc.,
Perkin Trans. 2 1982, 1605. (c) Laurence, C.; Quegnec-Cabanetos, M.;
Woitowiak, B. Can. J. Chem. 1983, 61, 135.
(13) (a) Cabot, R.; Hunter, C. A. Chem. Commun. 2009, 2005. (b)
Sarwar, M. G.; Dragisic, B.; Salsberg, L. J.; Gouliaras, C.; Taylor, M. S.
J. Am. Chem. Soc. 2010, 134, 1646. (c) Dimitijeviꢀc, E.; Kvak, O.; Taylor,
M. S. Chem. Commun. 2010, 46, 9025. (d) Hauchecorne, D.; van der
Veken, B. J.; Herrebout, W. A.; Hansen, P. E. Chem. Phys. 2011, 38, 5.
(14) Libri, S.; Jasim, N. A.; Perutz, R. N.; Brammer, L. J. Am. Chem.
Soc. 2008, 130, 7842.
(27) (a) Williams, J. H. Acc. Chem. Res. 1993, 26, 593. (b) Reich-
enbaecher, K.; S€uss, H. I.; Hulliger, J. Chem. Soc. Rev. 2005, 34, 22. (c)
Coates, G. W.; Dunn, A. R.; Henling, L. M.; Ziller, J. W.; Lobkovsky,
E. B.; Grubbs, R. H. J. Am. Chem. Soc. 1998, 120, 3641. (d) Vanspey-
brouck, W.; Herrebout, W. A.; van der Veken, B. J.; Lundell, J.; Perutz,
R. N. J. Phys. Chem. B 2003, 107, 13855.
(28) (a) Khandogin, J.; Ziegler, T. Spectrochim. Acta, Part A 1999,
55, 607. (b) Feindel, K. W.; Wasylishen, R. E. Magn. Reson. Chem. 2004,
42, S158.
(15) (a) Murphy, E. F.; Murugavel, R.; Roesky, H. W. Chem. Rev.
1997, 97, 3425. (b) Braun, T.; Perutz, R. N. Chem. Commun. 2002, 2749.
(c) Braun, T.; Perutz, R. N. Transition-Metal Mediated CꢀF Bond
Activation. In Comprehensive Organometallic Chemistry III; Crabtree,
R. H., Mingos, D. M. P., Eds.; Elsevier: Amsterdam, 2006; Vol. 1,
Chapter 26. (d) Clot, E.; Eisenstein, O.; Jasim, N.; Macgregor, S. A.;
McGrady, J. E.; Perutz, R. N. Acc. Chem. Res. 2011, 44, 333.
(29) We are aware that the Fermi Contact term and the paramag-
netic spinꢀorbit term derive principally from covalent bonding rather
than from electrostatic perturbation. We have considered several
possible scenarios: for instance, the paramagnetic spinꢀorbit term
should be strongly affected by the energy separation between dxy and
dx2ꢀy2 orbitals. This energy gap may be influenced by an electric field
gradient within an electrostatic model, or by a small covalent
component.
(16) (a) Ball, N. D.; Sanford, M. S. J. Am. Chem. Soc. 2009, 128, 7134.
(b) Watson, D. A.; Su, M.; Teverovskiy, G.; Zhang, Y.; García-Fortanet,
(30) Bondi, A. J. Phys. Chem. 1964, 68, 441.
14347
dx.doi.org/10.1021/ja203320y |J. Am. Chem. Soc. 2011, 133, 14338–14348

Journal of the American Chemical Society
ARTICLE
(31) RIF = d(I F)/rI + rF, where rI and rF are, respectively, the van
3 3 3
der Waals radii of I and F, following the definition in Lommerse, J. P. M.;
Stone, A. J.; Taylor, R.; Allen, F. H. J. Am. Chem. Soc. 1996, 118, 3108.
(32) (a) Pt is more electronegative than Ni.32b (b) Huheey, J. E.;
Keiter, E. A.; Keiter, R. L. Inorganic Chemistry, Principles of structure and
reactivity, 4th Ed; Harper Collins: New York, 1993.
(33) (a) Bader, R. F. W. Atoms in Molecules: A Quantum Theory,
Oxford University Press: Oxford, 1990. (b) Popelier, P. Atoms in
Molecules, An Introduction; Pearson Education, Harlow: New York,
2000. (c) Bader, R. F. W. Chem. Rev. 1991, 91, 893.
(34) Martinez Amezaga, N. J.; Pamies, S. C.; Peruchena, N. M.; Sosa,
G. L. J. Phys. Chem. A 2010, 114, 552.
(35) Lu, Y.-X.; Zou, J.-W.; Wang, Y. H.; Jiang, Y.-J.; Yu, Q.-S. J. Phys.
Chem. A 2007, 111, 10781.
(36) Ammann, C.; Meier, P.; Merbach, A. E. J. Magn. Reson. 1982,
46, 319.
(37) (a) Sheldrick, G. M. SADABS: Empirical absorption correction
program; University of G€ottingen: G€ottingen, 1995, based upon the
method of Blessing.37b (b) Blessing, R. H. Acta Crystallogr. 1995,
A51, 33.
(38) Sheldrick, G. M. Acta Crystallogr. 2008, A64, 112.
(39) Frisch, M. J. et al. . Gaussian 03, Revision C.02, Gaussian Inc:
Wallingford CT, 2004.
(40) (a) Csontos, J.; Palermo, N. Y.; Murphy, R. F.; Lovas, S.
J. Comput. Chem. 2008, 29, 1344. (b) Ruiz, E.; Salahub, D. R.; Vela, A.
J. Phys. Chem. 1996, 100, 12265. (c) Piacenza, M.; Grimme, S. J. Comput.
Chem. 2004, 25, 83. (d) Romero, C.; Fomina, L.; Fomine, S. Int. J.
Quantum Chem. 2005, 102, 200. (e) Waller, M. P.; Robertazzi, A.; Platts,
J. A.; Hibbs, D. E.; Williams, P. A. J. Comput. Chem. 2006, 27, 491.
14348
dx.doi.org/10.1021/ja203320y |J. Am. Chem. Soc. 2011, 133, 14338–14348