Organometallic complexes for second-order non-linear optics: Synthesis and molecular quadratic hyperpolarizabilities of η5-monocyclopentadienyliron(II) nitrile derivatives with different phosphines. X-ray crystal structure of [FeCp(DPPE)(p-NCC<

Article abstract of DOI:10.1016/S0022-328X(00)00679-3

New complexes of the type [FeCp(P_P)(p-NCR)][PF₆] (P_P=(S)-PROLOPHOS, DPPE and (R)-PROPHOS ((S)-PROLOPHOS=(S)-N-(diphenylphosphino)-2- diphenylphosphinooxymethylpyrrolidine; DPPE=1,2-bis(diphenylphosphino)ethane; (R)-PROPHOS=(R)-(+)-1,2-bis(dip

Full text of DOI:10.1016/S0022-328X(00)00679-3

www.elsevier.nl/locate/jorganchem
Journal of Organometallic Chemistry 619 (2001) 252–264
Organometallic complexes for second-order non-linear optics:
synthesis and molecular quadratic hyperpolarizabilities of
h5-monocyclopentadienyliron(II) nitrile derivatives with different
phosphines. X-ray crystal structure of
[FeCp(DPPE)(p-NCC6H4NO2)][PF6]·CH2Cl2
M. Helena Garcia a,b,*, M. Paula Robalo a,c, A.R. Dias a, M. Fa´tima M. Piedade a,b,
Adelino Galva˜o a, Wim Wenseleers d, Etienne Goovaerts d
a Centro de Qu´ımica Estrutural, Instituto Superior Te´cnico, A6. Ro6isco Pais, 1049-001 Lisbon, Portugal
b Departamento de Qu´ımica e Bioqu´ımica, Faculdade de Cieˆncias da Uni6ersidade de Lisboa, Campo Grande, 1749-016 Lisbon, Portugal
Departamento de Qu´ımica, Uni6ersidade de E6ora, Cole´gio Lu´ıs Anto´nio Verney, Rua Roma˜o Ramalho n° 59, 7000-671 E6ora, Portugal
d Physics Department, Uni6ersity of Antwerp (UIA), Uni6ersiteitsplein 1, 2610 Wilrijk, Antwerp, Belgium
c
´
´
Received 13 July 2000; accepted 12 September 2000
Abstract
New complexes of the type [FeCp(P P)(p-NCR)][PF6] (P P=(S)-PROLOPHOS, DPPE and (R)-PROPHOS ((S)-PRO-
–
–
LOPHOS=(S)-N-(diphenylphosphino)-2-diphenylphosphinooxymethylpyrrolidine; DPPE=1,2-bis(diphenylphosphino)ethane;
(R)-PROPHOS=(R)-(+)-1,2-bis(diphenylphosphino)propane); R=C6H4N(CH3)2, C6H4NO2, (E)-C(H)ꢀC(H)C6H4NO2,
C6H4C6H4NO2) have been synthesized and characterized. Spectroscopic data were evaluated in order to correlate the tuning of the
electronic density at the metal center with the first molecular hyperpolarizabilities (i) determined by hyper-Rayleigh scattering at
1064 nm. Experimental i values increase with chain length and with the enhancement of the donor ability of the bidentate
phosphine. [FeCp(DPPE)((E)-p-NCꢁC(H)ꢀC(H)C6H4NO2)][PF6] is found to have the highest i value (570×10−30 esu) in the
series. Molecular orbital calculations were carried out to understand the geometric and electronic factors involved and also to
compute the static hyperpolarizabilities, i0. These results are consistent with the trend observed in the i0 values derived from
experiment on the basis of the two-level model. The complex [FeCp(DPPE)(p-NCC6H4NO2)][PF6] was crystallographically
characterized. The compound crystallizes in the centrosymmetric space group Pbcn. Bond distances within the nitrile ligand were
discussed in order to evaluate the extension of p-backdonation suggested by spectroscopic data. © 2001 Elsevier Science B.V. All
rights reserved.
Keywords: NLO materials; Iron; Benzonitrile; Non-linear optics; First hyperpolarizability; Crystal structures
ities of organometallic complexes has started to be
actively studied more recently [5].
1. Introduction
The general criteria for obtaining organic molecules
The development of novel materials for non-linear
with large first hyperpolarizability (i) are being applied
optical (NLO) applications has been a very active area
in the search for molecular organometallic NLO mate-
of research for a number of years [1–4]. Although a
great deal of work has been carried out on the investi-
gation of the non-linear optical properties in inorganic
materials and organic molecules, the optical non-linear-
rials. Accordingly, the basic strategy of using electron-
donor and electron-acceptor substituents to polarize the
p-electron system of organic molecules, thereby creating
the possibility of a second-order non-linear optical re-
sponse, has been extended to organometallic com-
pounds consisting of conjugated donor–acceptor
systems [1–5].
* Corresponding author.
E-mail address: i017@alfa.ist.utl.pt (M.H. Garcia).
0022-328X/01/$ - see front matter © 2001 Elsevier Science B.V. All rights reserved.
PII: S0022-328X(00)00679-3

M.H. Garcia et al. / Journal of Organometallic Chemistry 619 (2001) 252–264
253
Our earlier studies in this field [6] reported for
[FeCp((+)-DIOP)(p-NCC6H4NO2)][PF6] ((+)-DIOP
=(+) - 2,3 - O - isopropylidene - 2,3 - dihidroxy - 1,4 - bis-
(diphenylphosphino)butane) a SHG efficiency of 38
times the urea standard, measured by Kurtz powder
technique. Spectroscopic evidence was found to support
the idea that the [FeCp((+)-DIOP)]+ fragment was
acting as a p-donor via d-p*(NC) orbitals towards the
NO2 acceptor group, enhancing the molecular first hy-
perpolarizability (i) of the coordinated nitrile, which
was related to the increased non-linear optical proper-
ties. Further studies on ruthenium(II) analogs also
yielded significant SHG results [7,8].
X-ray diffraction structures of some members of the
families [9] showed the metal to be in the same plane as
the p-system of the chromophore, which is responsible
for the availability of d orbitals for p-delocalization
towards the organic conjugated ligand. These results
are in good agreement with results published before
[10], suggesting the use of this structural feature as a
design criterion to maximize the organometallic non-
linear response.
PROPHOS phosphine, was considered as a good alter-
native to (+)-DIOP, since it is less bulky and therefore
more appropriate for future studies of the NLO effect
in the solid state.
Spectroscopic data were evaluated in order to under-
stand the contribution of these coligands on the
donor–acceptor ability of the iron(II) fragment and the
degree of p-delocalization on the molecule as well.
Measurements of the molecular quadratic hyperpolariz-
abilities for some of these compounds were carried out
by HRS at the fundamental wavelength of 1064 nm in
liquid solution. Extended Hu¨ckel MO calculations were
carried out in two model compounds in order to under-
stand the geometric and electronic factors that may be
conditioning the design of new compounds showing
large hyperpolarizabilities and ab initio calculations
were performed to compute values of i0 for these two
model compounds.
Experimental i values were correlated with spectro-
scopic data in order to understand the molecular struc-
ture–NLO property relationship in view of a future
optimization of the macroscopic NLO properties of this
family of compounds.
Our further studies on a systematic series of metal
fragments (h5-C5H5)M(P P) (M=Co(III), Ni(II),
–
Ru(II) and Fe(II); P P=phosphine ligand) with para-
–
substituted benzonitrile ligands, by hyper-Rayleigh
scattering (HRS) at the fundamental wavelength of
1064 nm, showed also that the iron(II) fragment is the
most promising organometallic electron donor group
[11].
2. Results and discussion
2.1. Preparation of the complexes
[FeCp(P P)(p-NCR)][PF6]
–
These encouraging NLO results prompted us to ex-
tend our work on the h5-monocyclopentadienyliron(II)
system and to study the possibility of tuning the elec-
tron density at the metal center, through phosphine
Nitrile complexes with different phosphines were pre-
pared by iodide abstraction from the starting materials
[FeCp(P P)(I)]
(P P=(S)-PROLOPHOS,
(R)-
–
–
PROPHOS, DPPE (Fig. 1)) by TlPF6 in the presence of
a slight excess of nitrile, in dichloromethane at room
temperature, following the procedure published before
[6].
coligands. We report herein the synthesis of [FeCp(P
–
P)(p-NCR)]+ complexes with various bidentate phos-
phines, P P=(S)-PROLOPHOS, DPPE and (R)-
–
PROPHOS
((S)-PROLOPHOS=(S)-N-(diphenyl-
After workup, orange or red crystals of [FeCp(P
–
phosphino)-2-diphenylphosphinooxymethylpyrrolidine;
DPPE=1,2-bis(diphenylphosphino)ethane; (R)-PRO-
PHOS=(R)-(+)-1,2-bis(diphenylphosphino)propane),
with different donor abilities. Also a single crystal
X-ray study of [FeCp(DPPE)(p-NCC6H4NO2)][PF6] is
presented and discussed. The use of the chiral (R)-
P)(p-NCR)][PF6]
C6H4N(CH3)2 (1), R=C6H4NO2 (2); P P=DPPE,
(P P=(S)-PROLOPHOS,
R=
–
–
R=C6H4N(CH3)2 (3), R=C6H4NO2 (4), R=(E)-
C(H)ꢀC(H)C6H4NO2 (5), R=C6H4C6H4NO2 (6); P
P=(R)-PROPHOS, R=C6H4N(CH3)2 (7), R=C6H4-
NO2 (8)) were obtained in ca. 52–96% yield.
–
Fig. 1. Structural formulae of phosphines DPPE, (R)-PROPHOS and (S)-PROLOPHOS.

254
M.H. Garcia et al. / Journal of Organometallic Chemistry 619 (2001) 252–264
[Fe]+ꢁNꢂCꢁC6H4ꢁRl[Fe]2+ꢀNꢀCꢀC6H4ꢀR−
The compounds are fairly stable towards oxidation in
air and to moisture both in the solid state and in
solution. The formulation is supported by analytical
data, IR and 1H, 13C and 31P-NMR spectra (see Section
4). The molar conductivities of ca. 10−3 M solutions of
the complexes in nitromethane, in the range 70–86 V−1
cm2 mol−1, are consistent with values reported for 1:1
electrolytes [12].
Comparison of these spectral data with our results on
(+)-DIOP analogs [6] indicates that different phos-
phine coligands can tune the donor character of the
organometallic fragment, in the following trend:
(S)-PROLOPHOSB(+)-DIOPB(R)-PROPHOS
:DPPE
Typical IR bands confirm the presence of the cy-
clopentadienyl ligand (ca. 3060 cm−1), the PF6− anion
(840 and 560 cm−1) and the coordinated nitrile (w(CN)
at ca. 2200 cm−1) in all complexes. No significant
change was observed on w(CN) upon coordination of
the dimethylamino derivatives but in the case of p-
NCC6H4NO2, depending on the coordinated phosphine
coligand, a significant negative shift was found, up to
35 cm−1 for the DPPE derivative. As we described
before, this effect would result from enhanced p-back-
donation owing to p-bonding between the metal d
orbitals and the p* orbital of the CN group [6].
Chemical shifts for the cyclopentadienyl ring are in
the range usually observed for monocationic iron(II)
complexes and seem to be affected both by the phos-
phine and by the nature of the coordinated nitrile. In
fact, substitution of (S)-PROLOPHOS by the more
electron donating phosphines (R)-PROPHOS or
DPPE, has the expected effect on Cp resonances (a shift
downfield of 0.26 and 0.36 ppm respectively, in the
1H-NMR spectra) consistent with the increased Cp ring
currents owing to the increased electron density at the
metal center. Replacement of the donor group p-
(N(CH3)2) by the acceptor group (p-NO2) on the nitrile
benzene ring gives a deshielding effect on the Cp reso-
nance, as was observed previously for the analogous
(+)-DIOP complexes [6]. The presence of different
phosphine coligands is also confirmed by the signals of
the carbon backbone in the corresponding ranges of the
spectra. Conversely to the effect observed for Cp ring,
the phosphine signals were relatively insensitive to the
nature of the aromatic nitrile.
The effect on coordination of the nitriles to the
different iron(II) fragments is mainly found in the
shielding observed at the ortho (relative to the NC
group) protons of the benzene ring. This shielding was
found to vary from 0.05 ppm for [FeCp(S)-PRO-
LOPHOS)(p-NCC6H4N(CH3)2)][PF6] up to 1.16 ppm
for [FeCp(DPPE)(p-NCC6H4NO2)][PF6]. Although in
some cases this effect might be insignificant, a trend is
clearly observed. The highest shielding on ortho protons
was observed for the p-nitrobenzonitrile DPPE deriva-
tive, in which this effect is concomitantly combined
with a deshielding on protons of the Cp ring, indicating
clearly an electronic flow towards the aromatic protons
due to a p-backdonation effect involving the metal
center. These observed shifts are consistent with the
possibility of some contribution of a vinylidene form in
solution:
13C-NMR data of this family of compounds confirm
the evidence found for proton spectra. The Cp ring
seems to be sensitive both to the coordinated nitrile and
phosphine, being more shielded with the enhancement
of donor ability of the phosphine. For the same phos-
phine a deshielding is observed when the N(CH3)2
donor group is replaced by the NO2 acceptor group.
31P-NMR data of complexes reported in ppm
downfield from the external standard (85% H3PO4)
showed the expected deshielding upon coordination of
the phosphines. Accordingly this effect was enhanced
with the better donor character of the phosphine, sug-
gesting an easier release of electron density towards the
metal center.
2.2. Electronic spectra
Optical absorption spectra of complexes [FeCp(P
–
P)(p-NCR)][PF6] were recorded in ca. 1.5×10−4 M in
dichloromethane, in the range of wavelengths 230–1100
nm (Table 1). In order to study the solvatochromic
behavior
for
compound
[FeCp(DPPE)((E)-p-
NCC(H)ꢀC(H)C6H4NO2)][PF6], the spectra were ob-
tained in other solvents of increasing polarity (acetone
and DMSO) in the same range of wavelengths (Fig. 3).
The absorption spectra for all compounds are charac-
terized by one intense absorption band in the UV
region, attributed to electronic transitions occurring in
the aromatic ligands. In addition, the nitro derivatives
also present one extra broad band in the visible region
characterized as a metal to ligand charge transfer
(MLCT) band, since a comparable absorption maxi-
mum is not found either for the [FeCp(P P)(I)] or for
–
the uncoordinated nitrile precursors. In Fig. 2 spectra
recorded for the compounds [FeCp(DPPE)(p-
NCC6H4R)][PF6] (R=N(CH3)2 and NO2) typify the
electronic spectra of this family of compounds, ob-
tained in dichloromethane solution.
Chromophore chain lengthening leads to a slight
bathochromic shift on umax. (MLCT band), with
[FeCp(P P)(p-NCC(H)ꢀC(H)C6H4NO2)][PF6] (P P=
–
–
(+)-DIOP, DPPE) complexes containing the lowest
energy transitions, indicating that conjugation is
slightly increased. Nevertheless in the case of biphenyl
compounds, a hipsochromic shift of the low-energy
absorption band was observed, suggesting a breaking of
the conjugation through the second ring.

M.H. Garcia et al. / Journal of Organometallic Chemistry 619 (2001) 252–264
255
Table 1
was chosen because it yields molecular hyperpolariz-
abilities that are much easier to interpret in terms of
molecular structure–property relationships than Kurtz-
powder results, since the latter also depend very
strongly on bulk properties such as crystal packing,
grain size and phase-matching effects. The more tradi-
tional electric field induced second harmonic generation
(EFISHG) technique, in which a static electric field is
applied to break the macroscopic centrosymmetry of
the liquid and induce coherent frequency doubling, is
not applicable to the studied ionic compounds. Instead,
the HRS technique relies on the incoherent scattering of
second harmonic light which results from the micro-
scopic disorder (random orientation) of molecules in a
liquid solution [15]. Molecular i values are deduced
from the intensity of this second harmonic light as
detailed in Section 4.4.
Optical spectral data for complexes [FeCp(P P)(p-NCR)][PF6] in
–
dichloromethane solutions at room temperature, concentration ca.
1.5×10−4 mol dm−3
The experimental hyperpolarizabilities are shown in
Table 2 for six of the Fe(II) complexes with the accep-
tor group NO2, as well as for two of the ligands. Static
polarizabilities are also listed as estimated from the
two-level model using the lowest energy electronic tran-
sition corresponding to the long wavelength absorption
band at ueg. This procedure has to be considered with
caution, taking into account that near-resonance condi-
tions occur for several of the compounds. As in the case
of the ruthenium compounds, the hyperpolarizabilities
(both i and i0) of the complexes are much higher (one
to two orders of magnitude) compared to the conju-
gated ligands. Starting from the shortest ligand (R=
C6H4NO2), a strong increase of i is observed when the
conjugated chain length is increased by inserting a
vinylene unit, for both the DPPE and (+)-DIOP
derivatives. This is in good agreement with the spectro-
scopic data (e.g. red shift of the electronic transition)
and confirms the idea that the high hyperpolarizabilities
result from the extension of this conjugated path from
the iron center to the nitro group. On the contrary, i
decreases by 40% and 50% (for DPPE and (+)-DIOP
compounds, respectively) upon insertion of a second
phenyl ring. This can be explained by a breaking of the
conjugation due to a significant torsion angle in the
biphenyl linkage [11], which also fits with the blue-shift
of the absorption band. After reduction to static values
the trends in i are strongly reduced (see Table 2),
showing the importance of the near-resonance effect
between the electronic transition and the second har-
monic wavelength. In fact, from the two-level model
even the static i0 values are expected to increase with
increasing ueg, which is not observed in these com-
pounds. However, the variation in ueg is relatively small
and other factors, e.g. the change in dipole moment
upon excitation (also occurring in the two-level model
expression for i0), are likely to vary more among the
compounds.
The solvatochromic response of [FeCp(DPPE)(p-
NCC(H)ꢀC(H)C6H4NO2)][PF6] was studied and the re-
sults are shown in Fig. 3. A pronounced red shift in the
position of the MLCT absorption band occurs when
the solvent polarity is increased, with differences of ca.
40 nm, while the positions of the other bands remain
almost unchanged. The positive solvatochromic behav-
ior exhibited by the compound is characteristic of
MLCT transitions with an increase of the dipole mo-
ment upon photoexcitation.
2.3. Second-order NLO characterization
The molecular second-order polarizabilities i of
three DPPE complexes as well as their (+)-DIOP
analogs were determined by means of hyper-Rayleigh
scattering (HRS) in liquid solution [13,14] at a funda-
mental wavelength of 1064 nm. The HRS technique

256
M.H. Garcia et al. / Journal of Organometallic Chemistry 619 (2001) 252–264
Fig. 2. UV–Vis absorption spectra of [FeCp(DPPE)(p-NꢂCC6H4NO2)][PF6] (——) and [FeCp(DPPE)(p-NꢂCC6H4N(CH3)2)][PF6] (–
recorded in CH2Cl2 (ca. 1.5×10−4 mol dm−3).
– –)
For a series of similar complexes (including cobalt,
nickel and ruthenium analogs) studied previously [11], a
very clear correlation was found between increasing i
values and increasing wavelength and/or intensity of
the lowest energy absorption band, attributed to the
MLCT transition. This trend is continued in the present
iron complexes, although for quantitative analysis the
application of the two-level model must be questioned,
taking into account that these compounds have their
lowest energy transition due on resonance with the
second harmonic wavelength.
Each of the DPPE compounds is found to possess a
somewhat larger hyperpolarizability than the corre-
sponding (+)-DIOP analog, in good agreement with
the IR and NMR data, all together pointing out an
increased electron density on the metal center for the
DPPE compounds. However, the slight blue-shift in the
optical absorption spectra of [FeCp(DPPE)(p-
NC(C6H4)2NO2)][PF6] compared to its (+)-DIOP
analog is not well understood. This increase in i is not
affected by the resonance effect as it persists in the
static values i0, which indicates that more electronic
density at the metal center is effectively contributing to
the hyperpolarizability.
Extended Hu¨ckel MO calculations (EHMO) (see Sec-
tion 4.5) carried out on the model complexes
[FeCp(PH3)2(p-NCR)]+
(R=C6H4NO2 or
C6H4C6H4NO2) show that the larger values of experi-
mental hyperpolarizabilities correspond to lower
HOMO–LUMO gaps. A simplified scheme (Fig. 4) of
the interaction diagram between the metal center
[FeCp(PH3)2]+ and the p-NCC6H4NO2 ligand shows
that the HOMO of the complex is essentially localized
in the metal fragment while the LUMO is essentially a
pure ligand orbital. This scheme also depicts a p-back-
bonding interaction between the SHOMO (second
HOMO) of the metal center (Fig. 5b) and an empty
orbital of the ligand (Fig. 5a).
One interesting feature that can be observed is the
contribution of the in-plane p orbital of the aryl ortho
carbons perpendicular to the CC bond (Fig. 5a) and
2.4. Molecular orbital calculations
In order to understand the geometric and electronic
factors that may be conditioning the design of new
compounds showing large hyperpolarizabilities, MO
calculations were carried out on several model
complexes.
Fig. 3. Solvatochromic behavior of the MLCT band of
[FeCp(DPPE)((E)-p-NCC(H)ꢀC(H)–C6H4NO2)][PF6] in solvents of
different polarity.

M.H. Garcia et al. / Journal of Organometallic Chemistry 619 (2001) 252–264
257
Table 2
The experimental hyperpolarizability, i, measured by HRS at u0=1064 nm, for a selection of the ligands and the Fe(II) compounds; the static
hyperpolarizability i0 is obtained from the two-level model taking the transition corresponding to the long-wavelength absorption band, ueg
i (10−30 esu)
i0 (10−30 esu)
ueg (nm)
Fe(II) compounds
[FeCp((+)-DIOP)(p-NCC6H4NO2)]+
[FeCp(DPPE)(p-NCC6H4NO2)]+
380
395 a,b
190
86
115
78
108
105
112
454
460
403
372
466
484
[FeCp((+)-DIOP)(p-NCC6H4C6H4NO2)]+
[FeCp(DPPE)(p-NCC6H4C6H4NO2)]+
[FeCp((+)-DIOP)((E)-p-NCC(H)ꢀC(H)C6H4NO2)]+
[FeCp(DPPE)((E)-p-NCC(H)ꢀC(H)C6H4NO2)]+
240 a
520
570
Ligands
p-NCC6H4NO2
p-NCC6H4C6H4NO2
4.4 a
3.2
9.4
257
294
14.6 a
a Ref. [11].
b Corrected value: in ref. [11] a CH2Cl2 solvent molecule in the structure had been erroneously ommitted in the calculation of the molecular
weight.
NCC6H4C6H4NO2 which were fully optimized at STO-
3G level and then the static hyperpolarizabilities were
computed at STO-3G, 3-21G and 3-21G**. The results
presented in Table 3 show that for the more sophisti-
cated basis sets including contracted valence orbitals
(3-21G and 3-21G**) the theoretical values are in very
good agreement with the static hyperpolarizabilities
derived from the experiment. In fact they are less than
a factor two smaller, and the ratio between the values
for the two ligands is quite well reproduced. Extension
of the basis set to include polarization functions (3-
21G**) does not yield any further improvement. The
factor between the results with the STO-3G and the
more complete basis sets is nearly the same for both
conjugated ligands.
also, in the metallic fragment, the contribution of one
of the Cp p orbitals (Fig. 5b).
Due to the anti-bonding character of the (a) orbital
we can predict a lengthening of the NC bond with the
corresponding observed decrease in the IR stretching
frequency, when compared with the free ligand. Also
the increase in charge density of the ortho carbons will
induce the upfield shift in the NMR spectra leaving the
meta carbons almost unchanged. This increase in
charge density is derived from p-backdonation from the
(b) orbital, involving also a contribution of Cp carbons,
which could explain the deshielding observed for the
Cp carbons.
Once understood that the best values of hyperpolar-
izabilities can be achieved with lower HOMO–LUMO
gaps, and since these orbitals are nearly non-bonding
orbitals of the metal center and p-NCR ligand respec-
tively, we can improve the non-linear optics of our
compounds by fine tuning the metal center with differ-
ent stabilizing ligands (changing phosphines or replac-
ing the Cp by indenyl or pyrrolyl ligands) aiming to
raise the HOMO or by using other R groups leading to
lower LUMOs. Because lower LUMOs (or higher HO-
MOs) will be better p acceptors (p donors) we can use,
in this general family of compounds, the NMR chemi-
cal shifts of the ortho carbons in the R group as a probe
to find compounds with potentially improved NLO
properties.
After the method had been tested in the free ligands,
further calculations were carried out at HF-STO-3G
level in two model iron complexes [FeCp(PH3)2(p-
NCR)]+ (R=C6H4NO2 and C6H4C6H4NO2) for com-
parison with the experimental results (see Table 4).
To help in the design of new compounds, selecting
the directions which the time consuming synthetic work
should follow, we performed ab initio calculations to
compute the static hyperpolarizability tensor compo-
nents (see Section 4.5) of two model compounds and
the corresponding conjugated ligands.
Fig. 4. Simplified interaction diagram between the metal center and
the p-NCC6H4NO2 ligand, showing the p-backbonding interaction as
a dotted line.
The influence of the extension of the basis set was
tested with the free ligands p-NCC6H4NO2 and p-

258
M.H. Garcia et al. / Journal of Organometallic Chemistry 619 (2001) 252–264
Fig. 5. Fragment orbitals involved in the p-backbonding interaction: (a) p-NCC6H4NO2 ligand, (b) metal center.
Table 3
Comparison of calculated and experimental values of the static hyperpolarizability i0 for the free ligands; the ratios are given between calculated
results obtained from the largest and smallest basis sets
i0 (10−30 esu)
Factor
Exp.
STO-3G
3-21G
3-21G**
STO-3G“3-21G**
p-NCC6H4NO2
p-NCC6H4C6H4NO2
3.2
9.4
0.77
1.60
2.42
5.54
2.40
5.52
3.1
3.5
The results reproduce the trends in the experimental
i0 values quite well, apart from a systematic underesti-
mation by a factor between 4.6 and 8.6. The latter is
not unexpected, as even for the free ligands reducing
the basis set from 3-21G** to STO-3G reduces the i0
values by a factor of 3.1 to 3.5. One should also bear in
mind here that in the calculation the electron-rich phos-
phines are replaced by simple PH3 groups, which is a
drastic simplification taking into account that the sub-
stitution of (+)-DIOP by DPPE already yields a size-
able increase (about 20–30%) of the static
hyperpolarizabilities. Also rewarding is the observation
of the same trend in the calculated and experimental
values of decreasing i0 from the phenyl to the biphenyl
compound. This decrease is more important in the
calculation (30% versus about 10% in the experiment)
but the overall agreement is very good, taking into
account the simplifications and approximations
involved.
structural study of [FeCp(DPPE)(p-NCC6H4NO2)]-
[PF6]·CH2Cl2 (4) was carried out to afford bond length
data about the metal–nitrile linkage, to give solid-state
packing information and to generate structural parame-
ters (interatomic distances, angles) to be used in the
molecular orbital calculations of the model complexes
detailed above.
The molecular structure of [FeCp(DPPE)(p-
NCC6H4NO2)][PF6]·CH2Cl2 (4) is shown in Fig. 6,
along with the atom numbering scheme. The values of
selected bond lengths and bond angles are given in
Table 5. The structural study confirmed the presence of
the PF−6 anion and revealed the presence of a solvent
CH2Cl2 molecule of crystallization.
The metal is coordinated to the h5-cyclopentadienyl
ring, the two phosphorus atoms of the phosphine lig-
Table 4
Comparison of experimental static i0 values with the HF-STO-3G
results for the [FeCp(PH3)2(NCR)]+ model complexes
2.5. Crystallographic studies
i0 (10−30 esu)
While the focus of this work is the molecular optical
non-linearities of the nitrile complexes, it was also of
interest to examine crystal packing as an indicator of
bulk material response, since a centrosymmetric ar-
rangement in the lattice would spoil the bulk response
of the SHG effect. With this in mind, a single X-ray
Exp.
STO-3G
18.3
[FeCp((+)-DIOP)(p-NCC6H4NO2)]+
[FeCp(DPPE)(p-NCC6H4NO2)]+
86
115
78
[FeCp((+)-DIOP)(p-NCC6H4C6H4NO2)]+
[FeCp(DPPE)(p-NCC6H4C6H4NO2)]+
12.6
108

M.H. Garcia et al. / Journal of Organometallic Chemistry 619 (2001) 252–264
259
Table 5
,
Selected bond lengths [A] and angles [°] for FeCp(DPPE)(p-
NꢂCC6H4NO2)][PF6]·CH2Cl2 (4)
Fe(1)–N(1)
Fe(1)–C(13)
Fe(1)–C(14)
Fe(1)–C(12)
Fe(1)–C(15)
Fe(1)–C(11)
Fe(1)–P(1)
Fe(1)–P(2)
1.874(11)
2.063(13)
2.088(14)
2.068(14)
2.090(13)
2.078(13)
2.210(4)
N(1)–Fe(1)–C(13)
N(1)–Fe(1)–C(14)
N(1)–Fe(1)–C(12)
N(1)–Fe(1)–C(15)
N(1)–Fe(1)–C(11)
N(1)–Fe(1)–P(1)
N(1)–Fe(1)–P(2)
P(1)–Fe(1)–P(2)
157.1(5)
125.9(6)
126.1(6)
92.3(5)
93.3(5)
92.1(3)
86.6(3)
2.209(3)
87.70(12)
122.3(8)
126.7(7)
129.2(8)
122.0(4)
113.9(5)
108.0(4)
102.3(6)
107.4(7)
101.9(6)
105.1(6)
103.4(6)
102.4(6)
107.3(4)
120.7(4)
116.1(4)
176.6(11)
177.4(15)
139(3)
Fe(1)–Centroid 1.721(7)
Fe(1)–Cp
Centroid–Fe(1)–N(1)
−1.703(7) Centroid–Fe(1)–P(1)
P(1)–C(111)
P(1)–C(121)
P(1)–C(131)
P(2)–C(231)
P(2)–C(211)
P(2)–C(221)
N(1)–C(2)
C(2)–C(3)
C(3)–C(8)
C(3)–C(4)
C(4)–C(5)
C(5)–C(6)
C(6)–C(7)
C(6)–N(2)
C(7)–C(8)
N(2)–O(1)
N(2)–O(2)
C(131)–C(231)
1.816(13)
1.818(14)
1.866(13)
1.857(12)
1.829(11)
1.814(13)
1.129(14)
1.42(2)
1.35(2)
1.38(2)
1.39(2)
1.35(2)
1.33(2)
1.60(2)
1.37(2)
1.15(2)
1.16(2)
1.51(2)
Centroid–Fe(1)–P(2)
C(111)–P(1)–Fe(1)
C(121)–P(1)–Fe(1)
C(131)–P(1)–Fe(1)
C(111)–P(1)–C(131)
C(121)–P(1)–C(131)
C(111)–P(1)–C(121)
C(231)–P(2)–C(211)
C(231)–P(2)–C(221)
C(211)–P(2)–C(221)
C(231)–P(2)–Fe(1)
C(211)–P(2)–Fe(1)
C(221)–P(2)–Fe(1)
C(2)–N(1)–Fe(1)
Fig. 6. Diagram for 4, [FeCp(DPPE)(p-NꢂCC6H4NO2)][PF6], with
50% thermal ellipsoids, showing the labeling scheme. Hydrogen
atoms have been omitted for clarity.
and and the nitrile nitrogen atom of the p-NCC6H4NO2
ligand, showing the typical structure of cyclopentadi-
enyl complexes in pseudo-octahedral three-legged piano
stool geometry, on the assumption that the cyclopenta-
dienyl group takes up three coordination sites.
N(1)–C(2)–C(3)
O(1)–N(2)–O(2)
O(1)–N(2)–C(6)
O(2)–N(2)–C(6)
The C(131)–P(1)–Fe–P(2)–C(231) chelate ring is
not planar but exhibits an envelope conformation, i.e.
the atoms Fe, P(1), P(2) and C(131) are essentially
coplanar while atom C(231), the ‘flap’ atom of the
112(2)
109(2)
C(231)–C(131)–P(1)
C(131)–C(231)–P(2)
109.9(8)
108.6(8)
,
envelope, lies more than −0.63(2) A from this plane.
As shown in Fig. 6, the non-planarity of this conforma-
tion diminishes the contacts between the chelate ring
phenyl groups and the nitrile and cyclopentadienyl
ligands.
The structural data permit comparisons, although the
relative imprecision of the crystallographic determina-
tion of complex 4 necessitates caution in interpreting
these data. A comparison between the coordination
geometry about the Fe atom is given in Table 6 for
reported structures of [FeCpL2L%] (L2=DPPE, DPPM;
L=CO, P(C6H5)3, P(OCH3)3; L%=NCCH3, NCC6H5)
complexes containing Fe–P or Fe–NꢂC bonds. The
phosphine structural data of compound 4, i.e. Fe–P
not bring to evidence the existence of p-backdonation
in the solid state. However, the nitrile group shows an
almost linear geometry with Fe–N1–C2 and N1–C2–
C3 angles of 177°, which indicate that the iron atom
and the benzonitrile ligand are in the same plane and
could stimulate the metal–ligand p-backdonation sug-
gested by the spectroscopic data.
Unfortunatelly, the crystal packing of 4 is centrosym-
metric due to crystallization in the orthorhombic space
group Pbcn with eight independent molecules in the
unit cell, making this compound unsuitable for macro-
scopic NLO purposes.
,
bond lengths 2.209 and 2.210 A and P1–Fe–P2 angle
87.7° are similar to those in other FeCp derivatives
containing DPPE and acetonitrile bonded ligands.
,
The Fe–N bond length (1.87 A), being shorter than
3. Concluding remarks
the corresponding distance found for some iron(II)
complexes with acetonitrile ligand, is within crystallo-
graphic error similar to the Fe–N bond lengths of
other FeCp(DPPE) complexes containing nitriles. Dis-
tances and angles within the benzonitrile group are
consistent with the retention of aromaticity, in particu-
lar there is no obvious bond length alternation which
would be expected for appreciable quinoidal contribu-
tion as was suggested by the spectroscopic IR and
NMR data obtained for the complex. These features do
A series of iron(II) nitrile complexes, involving sys-
tematic variation in coligands, have been synthesized as
model donor–acceptor systems to probe effects of
structural variation on non-linear optical properties.
Trends in spectroscopic and structural data and molec-
ular quadratic hyperpolarizabilities have been exam-
1
ined. Spectroscopic data (IR, H and 13C-NMR) are
sensitive to phosphine replacement, introduction of dif-
ferent para-substituent (nitro or dimethylamino group)

260
M.H. Garcia et al. / Journal of Organometallic Chemistry 619 (2001) 252–264
Table 6
Structural data for FeCp derivatives containing ligands with nitrogen or phosphorus donor atoms
, , , , ,
Fe–Cp (A) Fe–P (A) P1–Fe–P2 (°) Fe–N (A) NꢂC (A) NC–C (A) Fe–N–C (°) NꢂC–C (°) Reference
Compound
[FeCp(CO)2(NCCH3)][BF4] 2.030
1.914
1.916
1.200
1.137
1.494
1.457
175.676
179.791
175.045
178.592
[16a]
[16b]
2.112
[FeCp(CO)(PPh3)(NCCH3)]- 2.061
[BF4]
2.227
2.182
2.124
[FeCp(P(OCH3)3)2-
(NCCH3)][PF6]
2.127
95.360
86.454
74.673
86.75
1.924
1.881
1.892
1.895
1.905
1.888
1.913
1.874
1.094
1.137
1.135
1.128
1.133
1.149
1.141
1.129
1.155
1.482
1.432
1.438
1.451
1.456
1.433
1.467
1.42
172.960
171.864
179.591
173.5
175.363
176.869
178.146
174.626
178.611
174.72
179.094
177.4
[16c]
[16d]
[16e]
[16f]
[16g]
[16h]
[16i]
2.076
2.115
2.175
2.205
[FeCp(DPPE)(NCCH3)]-
[BPh4]
2.077
2.046
2.194
2.196
[FeCp(DPPM)(NCCH3)]-
[PF6]
2.090
2.075
2.207
2.232
[Fe(AcCp)(DPPE)-
(NCCH3)][PF6]
2.103
2.103
2.207
2.218
[Fe(MeCp)(DPPE)-
(NCCH3)][PF6]
86.311
84.835
83.197
87.70
177.322
174.64
177.908
176.6
2.126
2.237
2.260
[Fe(DMPE)2(NCC6H5)2]-
[PF6]2
2.277
2.320
[Fe(DPPE)2(NCCH3)2]-
[BF4]·CH2Cl2
2.342
2.209
[FeCp(DPPE)-
(p-NCC6H4NO2)][PF6]
2.063
2.090
This
work
2.210
p-NCC6H4NO2
1.438
179.2
[16j]
on nitrile ligand and nitrile chain-lengthening. Analysis
of these data gives an insight in the role played by the
organometallic moiety and the direction and extension
of the p-delocalization on the complexes. The finding of
an MLCT band in the electronic spectra of all nitro
derivatives is in good agreement with the direction of
this p-delocalization.
nitrile ligand and a deshielding effect on Cp carbons.
Therefore, in this family of compounds, NMR chemical
shifts of the ortho positions can be used as a probe
to search for compounds with potentially improved
NLO properties as was suggested in our previous work
[6,7].
The good agreement between the i0 values derived
from the experiment and those calculated theoretically
at HF-STO-3G level encourages the future use of this
calculation for prediction of static hyperpolarizabilities.
Moreover, the good correlation found between the
spectroscopic studies and the experimental second-or-
der non-linearities for some of the studied complexes,
indicates that IR and NMR data can be used as a tool
to predict non-linear optical behavior in solution, and
therefore are useful as a probe to plan new synthetic
strategies and to tune the non-linear optical behavior of
these families of compounds.
Accordingly, all the compounds measured by hyper-
Rayleigh scattering, with DPPE or (+)-DIOP as phos-
phine ligand and an NO2 acceptor group, presented
significant NLO properties. It was found that i values
increase with replacement of the (+)-DIOP coligand
by the more donating DPPE bidentate phosphine and
also with chromophore chain-lengthening. [FeCp-
(DPPE)((E)-p-NCC(H)ꢀC(H)C6H4NO2)][PF6] is found
to have the highest i value (570×10−30 esu) in the
series. The reduction to static hyperpolarizabilities
shows that these values are resonance enhanced as
could be expected from the close proximity of the
MLCT band to the second harmonic wavelength.
Theoretical calculations depict a p-backbonding in-
teraction between the SHOMO of the metal center and
an empty orbital of the ligand. These results may also
explain some 1H and 13C-NMR data found for the
complexes: a shielding effect on the ortho carbons in the
X-ray structural study carried out for the promising
compound [FeCp(DPPE)(p-NCC6H4NO2)][PF6] gave
no evidence for a significant vinylidene contribution on
the ground state geometry as would be expected from
the spectroscopic data. Since this compound crystallizes
on a centrosymmetric space group Pbcn, it is not
suitable for application as an NLO crystal. However,

M.H. Garcia et al. / Journal of Organometallic Chemistry 619 (2001) 252–264
261
considering that the spectroscopic data indicate a simi-
lar electronic behavior for the DPPE and (R)-
PROPHOS complexes, the use of this chiral phosphine
can constitute a strategy for further studies on macro-
scopic NLO properties.
DPPE: 1H-NMR ((CD3)2CO): 2.46 (m, 2H, CH2);
2.67 (m, 2H, CH2); 7.35 (m, 4H, C6H5); 7.54 (q, 12H,
C6H5); 7.85 (m, 4H, C6H5). 13C{1H}-NMR (CDCl3):
22.71 (t, CH2, JCP=21.40); 130.97 (C-meta, C6H5);
2
131.23 (C-para, C6H5); 132.86 (C-ortho, C6H5, JCP
23.20); 136.59 (C-ipso, JCP=18.33).
=
4.2. Preparation of [FeCp(P P)(I)]
–
4. Experimental
[FeCp(S)-PROLOPHOS)(I)] was prepared following
the procedure described previously [19]. [FeCp-
(DPPE)(I)] and [FeCp((R)-PROPHOS)(I)] were pre-
pared following the procedure: a boiling mixture of the
iodide [FeCp(CO)2(I)] (1 mmol) and (R)-PROPHOS
(1.2 mmol) in benzene (50 ml) was irradiated, with
stirring, with UV light for 30 min. Then the solvent was
evaporated to dryness and the resulting solid washed
three times with 20 ml of light petroleum (b.p. 30–
40°C). The pure complex was obtained after a further
recrystallization from dichloromethane–light petroleum
(b.p. 60–80°C).
[FeCp((S)-PROLOPHOS)(I)]: dark-violet; 48% yield;
m.p. 140–142°C, Anal. Calc. for C34H34FeINOP2: C,
56.92; H, 4.77; N, 1.95. Found: C, 56.73; H, 4.87; N,
1.86%. 1H-NMR (CD2Cl2): 1.18 (2H, m, CH2); 1.81
(1H, m, CH2); 2.59 (1H, m, CH2); 3.03 (1H, m, CH2);
3.44 (1H, m, CH); 3.65 (1H, m, CH2); 3.85 (5H, s,
h5-C5H5), 4.56 (1H, m, CH2); 6.45 (2H, d, C6H5); 6.99
(2H, t, C6H5); 7.17 (4H, m, C6H5); 7.46 (10H, m, C6H5);
8.12 (1H, br.s, C6H5); 8.27 (1H, br.s, C6H5).
4.1. General procedures
All experiments were carried out under dinitrogen or
argon by use of standard Schlenk techniques. Solvents
were dried according to the usual published methods
[17]. (S)-PROLOPHOS was prepared following the lit-
erature [18]. DPPE and (R)-PROPHOS were used as
purchased from Aldrich Chemical Co.
Solid state IR spectra were recorded on a Perkin
Elmer 683 spectrophotometer in KBr pellets; only sig-
nificant bands are cited in the text. H, 13C and the
1
31P-NMR spectra were recorded on a Varian Unity 300
spectrometer at probe temperature. The UV–Vis spec-
tra were taken on a Shimadzu 1202 spectrophotometer.
Microanalyses were performed in our laboratories using
a Fisons Instruments EA1108 system. Data acquisition,
integration and handling were performed using a PC
with the software package Eager-200 (Carlo Erba In-
struments). Melting points were obtained on a Reichert
Thermovar. The molar conductivities of 10−3 mol l−1
solutions of the complexes in nitromethane were
recorded with a Schott CGB55 Konduktometer.
[FeCp((R)-PROPHOS)(I)]: anthracite; 80% yield;
m.p. 142°C (dec.) Anal. Calc. for C32H31FeIP2: C,
1
1
The H (acetone-d6 and chloroform-d) and 13C (chlo-
58.21; H, 4.73. Found: C, 57.91; H, 4.71%. H-NMR
roform-d) chemical shifts are reported in parts per
million downfield from internal Me4Si and the 31P-
NMR spectra are reported in parts per million
downfield from external 85% H3PO4. Spectral assign-
ments follow the numbering scheme shown in Fig. 7.
1H- and 13C-NMR data are similar for all the com-
plexes containing the same phosphine. (S)-PRO-
(CDCl3): 1.12 (3H, br.s, CH3); 1.96 (1H, m, CH2); 2.57
(1H, m, CH); 3.15 (1H, m, CH2); 4.16 (5H, s, h5-C5H5);
7.15 (2H, m, C6H5); 7.35 (14H, m, C6H5); 7.64 (2H, m,
C6H5); 8.06 (2H, m, C6H5).
4.3. Preparation of [FeCp(P P)(p-NCC6H4R)][PF6]
–
1
LOPHOS: H-NMR ((CD3)2CO): 1.49 (m, 1H, CH2);
[FeCp((+)-DIOP)(p-NCR)][PF6]
(R=C6H4N-
1.69 (m, 2H, CH2); 1.86 (m, 1H, CH2); 2.20 (m, 1H,
CH2); 2.74 (m, 1H, CH2); 3.17 (m, 1H, CH2); 3.62 (m,
1H, CH); 3.86 (m, 1H, CH2); 4.08 (m, 1H, CH2); 6.68
(t, 2H, C6H5); 7.21 (q, 3H, C6H5); 7.34 (t, 1H, C6H5);
7.61–7.90 (cm, 14H, C6H5); 8.09 (t, 2H, C6H5).
13C{1H}-NMR (CDCl3): 25.60 (CH2); 28.82 (CH2);
49.86 (CH2); 57.35 (CH); 70.14 (CH2); 127.84–133.40
(C6H5); 136.80 (C-ipso, C6H5); 139.80 (C-ipso, C6H5).
(R)-PROPHOS: 1H-NMR ((CD3)2CO): 1.20 and 1.23
(3H, dd, JHH=6.6, CH3); 2.29 (m, 1H, CH2), 2.44 (1H,
br.s, CH); 3.28 and 3.44 (1H, dm, CH2); 7.49–7.71
(16H, m, C6H5); 7.84 (2H, t, C6H5); 8.11 (2H, t, C6H5).
13C{1H}-NMR (CDCl3): 16.34 (CH3); 31.92 (m, CH2);
(CH3)2, C6H4NO2, C6H4C6H4NO2, (E)-C(H)ꢀC(H)-
C6H4NO2) synthesis and characterization was described
previously [6]. Some spectroscopic data of these com-
pounds are also introduced in order to make compari-
sons with the phosphine derivatives studied in this
work.
All the other complexes were prepared by the process
described below. To a solution of [FeCp(P P)(I)] (1
–
mmol) and the appropriate nitrile p-NCR (R=
C6H4N(CH3)2, C6H4NO2, (E)-C(H)ꢀC(H)C6H4NO2,
C6H4C6H4NO2) (2 mmol) in dichloromethane (40 ml)
was added TlPF6 (1 mmol) at room temperature. The
mixture was stirred at room temperature for 12–20 h.
A change was observed from dark-violet to orange or
red with simultaneous precipitation of thallium iodide.
34.76 (m, CH); 128.90–133.96 (C6H5); 135.96 (d, JCP
43.52, C-ipso, C6H5).
=

262
M.H. Garcia et al. / Journal of Organometallic Chemistry 619 (2001) 252–264
After filtration, the solution was evaporated under vac-
uum to dryness and washed several times with ether to
remove the excess of nitrile. The residue was recrystal-
lized from dichloromethane–ether.
[FeCp((S)-PROLOPHOS)(p-NCC6H4N(CH3)2)][PF6]
(1): dark orange; 68% yield, m.p. 198°C (dec.). Anal.
Calc. for C43H44F6FeN3OP3: C, 58.58; H, 5.03; N, 4.76.
Found: C, 58.67; H, 5.18; N, 4.62%. IR (KBr): w(CN)
[FeCp(DPPE)(p-NCC6H4NO2)][PF6] (4): violet; 52%
yield, m.p. 110°C (dec.). Anal. Calc. for C38H33F6Fe-
N2O2P3: C, 56.18; H, 4.09; N, 3.45. Found: C, 55.73; H,
4.54; N, 3.21%. IR (KBr): w(CN) 2205, w(NO2) 1520,
1340 cm−1 1H-NMR ((CD3)2CO): 4.72 (s, 5H, h5-
.
C5H5); 6.97 (d, 2H, JHH=9.0, H2,H6); 8.11 (d, 2H,
J
HH=9.3, H3,H5). 13C{1H}-NMR (CDCl3): 80.42 (h5-
C5H5); 116.93 (C1); 123.69 (C3,C5); 129.91 (CN);
131.01 (C2,C6); 149.05 (C4). 31P{1H}-NMR (CDCl3):
98.09.
2210 cm−1
.
1H-NMR ((CD3)2CO): 3.08 (s, 6H,
N(CH3)2); 4.23 (s, 5H, h5-C5H5); 6.78 (d, 2H, JHH
=
8.9, H3,H5); 7.41 (d, 2H, JHH=9.2, H2,H6). 13C{1H}-
NMR (CDCl3): 39.85 (N(CH3)2); 81.72 (h5-C5H5);
95.94 (C1); 111.83 (C3,C5); (C2,C6) (signal obscured by
the aromatic carbons of phosphine); (CN) (signal ob-
scured by the aromatic carbons of phosphine); 153.05
(C4).
[FeCp(DPPE)(p-NCC(H)ꢀC(H)C6H4NO2)][PF6] (5):
violet; 45% yield, m.p. 221°C (dec.). Anal. Calc. for
C40H35F6FeN2O2P3: C, 57.30; H, 4.21; N, 3.34. Found:
C, 56.94; H, 4.06; N, 3.16%. IR (KBr): w(CN) 2205,
1
w(NO2) 1520, 1345 cm−1. H-NMR ((CD3)2CO): 4.65
(s, 5H, h5-C5H5); 5.95 (d, 1H, JHH=17.1, H8); 6.52 (d,
1H, JHH=16.5, H7); 7.61 (d, 2H, JHH=8.4, H2,H6);
[FeCp((S)-PROLOPHOS)(p-NCC6H4NO2)][PF6] (2):
red; 59% yield, m.p. 208°C (dec.). Anal. Calc. for
C41H38F6FeN3O3P3.CH2Cl2: C, 52.08; H, 4.16; N, 4.34.
Found: C, 51.97; H, 4.63; N, 3.81%. IR (KBr): w(CN)
8.19 (d, 2H,
J
HH=8.4, H3,H5). 13C{1H}-NMR
(CDCl3): 80.12 (h5-C5H5); 99.75 (C8); 124.08 (C3,C5);
128.18 (C2,C6); 130.43 (CN); 139.17 (C1); 147.64 (C7);
148. 62 (C4). 31P{1H}-NMR (CDCl3): 97.64.
1
2210, w(NO2) 1525, 1345 cm−1. H-NMR ((CD3)2CO):
4.40 (s, 5H, h5-C5H5); 7.99 (d, 2H, JHH=8.6, H2,H6);
[FeCp(DPPE)(p-NCC6H4C6H4NO2)][PF6] (6): or-
ange; 84% yield, m.p. 220–222°C. Anal. Calc. for
C44H37F6FeN2O2P3·CH2Cl2: C, 55.52; H, 4.04; N, 2.88.
Found: C, 55.84; H, 3.95; N, 2.87%. IR (KBr): w(CN)
8.41 (d, 2H,
J
HH=8.6, H3,H5). 13C{1H}-NMR
(CDCl3): 83.14 (h5-C5H5); 118.80 (C1); 124.58 (C3,C5);
(CN) (signal obscured by the aromatic carbons of
phosphine); 133.69 (C2,C6); 149.50 (C4). 31P{1H}-
NMR (CDCl3): 111.40 (d, JPP=91.07, P(N)); 174.58
(d, JPP=91.7, P(O)).
1
2205, w(NO2) 1515, 1350 cm−1. H-NMR ((CD3)2CO):
4.69 (s, 5H, h5-C5H5); 6.83 (d, 2H, JHH=8.7, H2,H12);
7.73 (d, 2H, JHH=7.8, H3,H11); 7.90 (d, 2H, JHH=9.0,
H6,H10); 8.31 (d, 2H, JHH=9.0, H7,H9). 13C{1H}-NMR
(CDCl3): 79.75 (h5-C5H5); 111.16 (C1); 124.12 (C7,C9);
127.69 (C3,C11); 128.42 (C6,C10); (CN) (signal ob-
scured by the aromatic carbons of phosphine); 132.41
(C2,C12); 142.75 (C4); 144.95 (C5); 147.64 (C8).
31P{1H}-NMR (CDCl3): 96.02.
[FeCp(DPPE)(p-NCC6H4N(CH3)2)][PF6] (3): brick
orange; 92% yield, m.p. 208–210°C. Anal. Calc. for
C40H39F6FeN2P3: C, 59.27; H, 4.85; N, 3.45. Found: C,
58.72; H, 5.20; N, 3.47%. IR (KBr): w(CN) 2210 cm−1
.
1H-NMR ((CD3)2CO): 2.97 (s, 6H, N(CH3)2); 4.59 (s,
5H, h5-C5H5); 6.38 (d, 2H, JHH=9.0, H3,H5); 6.47 (d,
2H, JHH=9.3, H2,H6). 13C{1H}-NMR (CDCl3): 39.72
(N(CH3)2); 78.99 (h5-C5H5); 95.67 (C1); 111.00 (C3,C5);
130.52 (C2,C6); (CN) (signal obscured by the aromatic
carbons of phosphine); 152.29 (C4). 31P{1H}-NMR
(CDCl3): 96.31.
[FeCp((R)-PROPHOS)(p-NCC6H4N(CH3)2)][PF6](7):
red; 92% yield, m.p. 135–137°C. Anal. Calc. for
C41H41F6FeN2P3: C, 59.51; H, 4.99; N, 3.40. Found: C,
60.04; H, 5.22; N, 3.09%. IR (KBr): w(CN) 2210 cm−1
.
1H-NMR ((CD3)2CO): 2.99 (s, 6H, N(CH3)2); 4.49 (s,
5H, h5-C5H5); 7.48 (d, 2H, JHH=9.6, H3,H5); 7.64 (d,
2H, JHH=8.1, H2,H6). 13C{1H}-NMR (CDCl3): 39.78
(N(CH3)2); 79.22 (h5-C5H5); 95.69 (C1); 110.99 (C3,C5);
131.12 (C2,C6); (CN) (signal obscured by the aromatic
carbons of phosphine); 152.59 (C4). 31P{1H}-NMR
(CDCl3): 82.45 (d, JPP=44.4); 106.64 (d, JPP=45.1).
[FeCp((R)-PROPHOS)(p-NCC6H4NO2)][PF6] (8): vi-
olet; 96% yield, m.p. 236°C (dec.). Anal. Calc. for
C39H35F6FeN2O2P3: C, 56.47; H, 4.25; N, 3.38. Found:
C, 56.37; H, 4.22; N, 3.23%. IR (KBr): w(CN) 2210,
1
w(NO2) 1515, 1340 cm−1. H-NMR ((CD3)2CO): 4.65
(s, 5H, h5-C5H5); 7.10 (d, 2H, JHH=8.7, H2,H6); 8.12
(d, 2H, JHH=8.7, H3,H5). 13C{1H}-NMR (CDCl3):
80.48 (h5-C5H5); 116.98 (C1); 123.70 (C3,C5); (CN)
(signal obscured by the aromatic carbons of phos-
phine); 133.14 (C2,C6); 148.99 (C4). 31P{1H}-NMR
(CDCl3): 81.89 (d, JPP=43.6); 104.54 (d, JPP=43.1).
Fig. 7. Numbering scheme for NMR spectral assignments.

M.H. Garcia et al. / Journal of Organometallic Chemistry 619 (2001) 252–264
263
4.4. i Measurements
tions, and d orbitals were taken as contracted linear
combinations of two Slater type wave functions. Stan-
dard parameters were used for all atoms. All MO
drawings were made with CACAO [25]. The crystal
structure of the organometallic complex [FeCp(DPPE)-
(p-NCC6H4NO2)][PF6]·CH2Cl2 was used as starting
model for all calculations. In all ab initio calculations
the phosphines were replaced by PH3 groups.
Hyper-Rayleigh scattering (HRS) measurements were
performed using 70 ps pulses from a Nd:YAG regener-
ative amplifier (u=1064 nm) operating at a repetition
rate of 2 kHz. A detailed description of the experimen-
tal procedure has been given in refs. [11,20]. In this
setup, any erroneous signals due to the two photon
luminescence are adequately eliminated by systemati-
cally scanning a narrow region around the second
harmonic wavelength using a monochromator and to
some extent also by using a nanosecond electronic time
gating. Hyperpolarizabilities are determined from the
HRS intensity by internal reference relative to chloro-
form. The intensity of the HRS light is proportional to
the concentration and to the orientational averages
ꢀi2ZZZꢁ and ꢀiX2 ZZꢁ (upper case indices: laboratory
coordinates) of the molecular hyperpolarizability, for
the polarization components parallel and orthogonal to
the polarization of the incident laser beam, respectively.
For simplicity we assume that the i tensors of both
solvent and solute molecules are dominated by a single
(diagonal) tensor component izzz (lower case indices:
molecular coordinates). For the present NLO com-
pounds, this is expected to be a very good approxima-
tion, as previous measurements [11] on similar
complexes all showed depolarization ratios ꢀi2XZZꢁ/
ꢀi2ZZZꢁ very close to 0.2, and the theoretical prediction
ꢀi2XZZꢁ/ꢀiZ2 ZZꢁ=1/5 is also exactly reproduced by our
HF-STO-3G calculations on the Fe(II) model com-
pounds. Also, with the proper choice of the molecular
z-direction, the theoretical calculations izzz value is at
least 15 times larger than any other tensor component.
With this assumption, the orientational averages are
given by ꢀi2ZZZꢁ=71i2zzz and ꢀi2XZZꢁ=315i2zzz, and
therefore iHRS, defined as (ꢀi2ZZZꢁ+ꢀiX2 ZZꢁ)1/2 (8
total HRS intensity), is equal to ꢂ6/35izzz. In this way,
the ratio iHRS(solute)/iHRS(CHCl3) obtained from the
experiments can be directly interpreted as the ratio of
the molecular zzz-components and izzz(solute) is ob-
tained by adopting a value of 0.49×10−30 esu for
ꢃizzz(CHCl3)ꢃ (EFISHG, from ref. [21]). Although for
CHCl3 significant off-diagonal components may be ex-
pected, additional assumptions on the tensor compo-
nents would be needed to improve on this analysis.
The iijk components of the static hyperpolarizability
(10 independent components for this symmetrical ten-
sor) were determined from ab initio calculations, in
order to obtain the orientationally averaged observ-
ables ꢀi2ZZZꢁ and ꢀiX2 ZZꢁ, following the formulae de-
scribed by Cyvin et al. [26]. This eventually yields the
depolarization ratio as i2XZZ/iZ2 ZZ. The intensity of the
unpolarized second harmonic scattering is proportional
to iHRS=(ꢀi2ZZZꢁ+ꢀiX2 ZZꢁ)1/2, and the reported i0 is
an effective izzz value computed as i0=ꢂ35/6iHRS
.
4.6. X-ray structure of [FeCp(DPPE)-
(p-NꢂCC6H4NO2)][PF6]·CH2Cl2
Crystal data: [C38H31F6N2O2P3Fe.CH2Cl2], Mr
895.33, orthorhombic, space group Pbcn, a=12.928(2),
3
,
,
b=29.9820(10), c=21.725(2) A, V=8420.8 A ,
D
calc=1.412 Mg m−3, Z=8, u( Mo–Ka)=0.71069 A,
,
v(Mo–Ka)=0.661 mm−1. Data were collected at
293(2) K on an Enraf–Nonius TURBOCAD4 diffrac-
tometer. The unit cell dimensions and orientation ma-
trix were obtained by least-squares refinement of 25
centered reflections with 13BqB16°. Using the CAD4
software, data were corrected for Lorentz and polariza-
tion effects and empirically for absorption. Intensities
of 7416 reflections (7416 unique reflections) in the range
1.655q524.99° were measured by the ꢀ–2q scan
mode. The structure was solved with SHELXS-86 [27]
and refined by full-matrix least-squares method with
SHELX93 [28]. 5607 reflections with F\3|(F) were
used in the structure solution and the final R(F) value
was 0.1172, for 511 refined parameters.
Non-hydrogen atoms were anisotropically refined
and all hydrogen atoms were inserted in calculated
positions and refined isotropically with a thermal
parameter equal to 1.2 times those of the atoms to
which they are bonded. A solvent molecule, CH2Cl2,
was found to be disordered over two positions. The
occupation factors refined 0.75 and 0.25 values used in
the last cycle of refinement. In the disorder model the
4.5. Molecular orbital calculations
,
Combined ab initio and extended Hu¨ckel calculations
were carried out on model complexes. Ab initio calcula-
tions were performed at Hartree–Fock level using the
GAUSSIAN94/DFT program [22]. Extended Hu¨ckel
[23] type calculations with modified Hij values [24] were
also carried out. The basis set for the metal atoms
consisted of ns, np and (n−1)d orbitals. The s and p
orbitals were described by single Slater type wave func-
C–Cl distances were restrained to 1.75 A and the
thermal parameters were restrained to be approximately
isotropic with an e.s.d. of 0.005 (ISOR). Two half
molecules of PF6 were found to be in a two-fold axis.
The thermal parameters of the fluorine atoms in one of
them were also refined with ISOR (e.s.d. of 0.005). The
thermal parameters of the oxygen atoms of the nitrile
ligand were also restrained (e.s.d. 0.01). The largest

264
M.H. Garcia et al. / Journal of Organometallic Chemistry 619 (2001) 252–264
[7] A.R. Dias, M. Helena Garcia, J.C. Rodrigues, M.L.H. Green,
S.M. Klueber, J. Organomet. Chem. 475 (1994) 241.
[8] A.R. Dias, M.H. Garcia, P.J. Mendes, M.P. Robalo, J.C. Ro-
drigues, Trends Organomet. Chem. 1 (1994) 335.
[9] (a) M. Paula Robalo, PhD Thesis, FCUL, Lisbon, 1992. (b) J.
Manuel Rodrigues, PhD Thesis, FCUL, Lisbon, 1999.
[10] J.C. Calabrese, L.-T. Cheng, J.C. Green, S.R. Marder, W. Tam,
J. Am. Chem. Soc. 113 (1991) 7227.
peak in the final difference Fourier synthesis, 0.937 e.
A
−3, was found near the F atoms of one of the PF6
,
molecules. The illustrations were drawn with program
ORTEP-II [29]. The atomic scattering factors and anoma-
lous scattering terms were taken from international
tables [30].
[11] W. Wenseleers, A.W. Gerbrandij, E. Goovaerts, M.H. Garcia,
M.P. Robalo, P.J. Mendes, J.C. Rodrigues, A.R. Dias, J. Mater.
Chem. 8 (1998) 925.
5. Supplementary material
[12] W.J. Geary, Coord. Chem. Rev. 7 (1977) 81.
Crystallographic data for the structural analysis of
compound 4, [FeCp(DPPE)(p-NCC6H4NO2)][PF6] have
been deposited with the Cambridge Crystallographic
Data Centre, CCDC no. 145093. Copies of this informa-
tion may be obtained free of charge from The Director,
CCDC, 12 Union Road, Cambridge, CB2 1EZ, UK
(Fax: +44-1223-336033; e-mail: deposit@ccdc.cam.ac.
uk or www: http://www.ccdc.cam.ac.uk).
[13] K. Clays, A. Persoons, Phys. Rev. Lett. 66 (1991) 2980.
[14] K. Clays, A. Persoons, Rev. Sci. Instrum. 63 (1992) 3285.
[15] R.W. Terhune, P.D. Maker, C.M. Savage, Phys. Rev. Lett. 14
(1965) 681.
[16] (a) S. Fadel, K. Weidenhammer, M.L. Ziegler, Z. Anorg. Allg.
Chem. 453 (1979) 98. (b) A.G.M. Barrett, N.E. Carpenter, M.
Sabat, J. Organomet. Chem. 352 (1988) C8. (c) H. Schumann, L.
Eguren, J.W. Ziller, J. Organomet. Chem. 408 (1991) 361. (d) P.E.
Riley, C.E. Capshew, R. Pettit, R.E. Davis, Inorg. Chem. 17 (1978)
408. (e) J. Ruiz, M.-T. Garland, E. Roman, D. Astruc, J.
Organomet. Chem. 377 (1989) 309. (f) J. Ruiz, M.A. Gonzalez,
E. Roman, M.T. Garland, J. Chem. Soc. Chem. Commun. (1990)
21. (g) J. Ruiz, D. Astruc, J.P. Bideau, M. Cotrait, Acta Crystal-
logr. Sect. C 46 (1990) 2367. (h) A.V. George, L.D. Field, E.Y.
Malouf, A.E.D. McQueen, S.R. Pike, G.R. Purches, T.W. Ham-
bley, I.E. Buys, A.H. White, D.C.R. Hockless, B.W. Skelton, J.
Organomet. Chem. 538 (1997) 101. (i) A.J. Blake, A.J. Atkins,
R.O. Gould, M. Schroder, Z. Kristallogr. 198 (1992) 287. (j) T.
Higashi, K. Osaki, Acta Crystallogr. Sect. B 33 (1977) 2337.
[17] D.D. Perrin, W.L.F. Amarego, D.R. Perrin, Purification of
Laboratory Chemicals, Pergamon, New York, 2nd edn., 1980.
[18] E. Cesarotti, A. Chiesa, G. Ciani, A. Sironi, J. Organomet. Chem.
251 (1983) 79.
Acknowledgements
Financial support was given by Junta Nacional de
Investigac¸a˜o Cient´ıfica e Tecnolo´gica (JNICT) actually
Fundac¸a˜o para a Cieˆncia e Tecnologia (FCT) (PCEX/C/
QUI/96). The work in Antwerp is partly funded by the
Flanders Government in its action for the promotion of
participation in EU-research programs and also by the
Fund for Scientific Research of Flanders (F.W.O.). E.G.
is a Research Director of the F.W.O.
[19] G. Balavoine, M.L.H. Green, J.P. Sauvage, J. Organomet. Chem.
128 (1977) 247.
[20] M. Szablewski, P.R. Thomas, A. Thornton, D. Bloor, G.H. Cross,
J.M. Cole, J.A.K. Howard, M. Malagoli, F. Meyers, J.-L. Bre´das,
W. Wenseleers, E. Goovaerts, J. Am. Chem. Soc. 119 (1997) 3144.
[21] F. Kajzar, I. Ledoux, J. Zyss, Phys. Rev. A 36 (1987) 2210.
[22] M.J. Frisch, G.W. Trucks, H.B. Schlegel, P.M.W. Gill, B.G.
Johnson, M.A. Robb, J.R. Cheeseman, T. Keith, G.A. Petersson,
J.A. Montgomery, K. Raghavachari, M.A. Al-Laham, V.G.
Zakrzewski, J.V. Ortiz, J.B. Foresman, J. Cioslowski, B.B. Ste-
fanov, A. Nanayakkara, M. Challacombe, C.Y. Peng, P.Y. Ayala,
W. Chen, M.W. Wong, J.L. Andres, E.S. Replogle, R. Gomperts,
R.L. Martin, D.J. Fox, J.S. Binkley, D.J. Defrees, J. Baker, J.P.
Stewart, M. Head-Gordon, C. Gonzalez, J.A. Pople, GAUSSIAN
94, Revision C.3, Gaussian, Inc.; Pittsburgh, PA, 1995.
[23] (a) R. Hoffmann, J. Chem Phys. 39 (1963) 1397. (b) K. Tatsumi,
A. Nakamura, J. Am. Chem. Soc. 109 (1987) 3195. (c) R.
Hoffmann, Angew. Chem. Int. Ed. Engl. 21 (1982) 711.
[24] J.H. Ammeter, H.-B. Bu¨rgi, J.C. Thibeault, R. Hoffmann, J. Am.
Chem. Soc. 100 (1978) 3686.
References
[1] D.S. Chemla, J. Zyss (Eds.), Nonlinear Optical Properties of
Organic Molecules and Crystals, vols. 1 and 2, Academic Press,
Orlando, FL, 1987.
[2] S.R. Marder, J.E. Sohn, G.D. Stucky (Eds.), Materials for
Nonlinear Optics: Chemical Perspectives, American Chemical
Society Press, Washington, DC, 1991, ACS Symp. Ser. 455.
[3] D.J. Williams, Angew. Chem. Int. Ed. Engl. 23 (1984) 690.
[4] J. Zyss (Ed.), Molecular Nonlinear Optics: Materials, Physics, and
Devices, Academic Press, San Diego, CA, 1994.
[5] (a) H.S. Nalwa, Appl. Organomet. Chem. 5 (1991) 349. (b) I.R.
Whittall, A.M. McDonagh, M.G. Humphrey, M. Samoc, Adv.
Organomet. Chem. 42 (1998) 291 and refs. cited therein. (c) J. Qin,
D. Liu, C. Da´ı, C. Chen, B. Wu, C. Yang, C. Zhen, Coord. Chem.
Rev. 188 (1999) 23. (d) I.R. Whittall, M.P. Cifuentes, M.G.
Humphrey, B. Luther-Davies, M. Samoc, S. Houbrechts, A.
Persoons, G.A. Heath, D.C.R. Hockless, J. Organomet. Chem.
549 (1997) 127. (e) S. Houbrechts, C. Boutton, K. Clays, A.
Persoons, I.R. Whittall, R.H. Naulty, M.P. Cifuentes, M.G.
Humphrey, J. Nonlinear Opt. Phys. Mater. 7 (1998) 113. (f) E.
Goovaerts, W.E. Wenseleers, M.H. Garcia, G.H. Cross, Design
and characterisation of organic and organometallic molecules for
second order nonlinear optics. In: H.S. Nalwa (Ed.), Handbook
of Advanced Electronic and Photonic Materials, Academic Press,
New York, in press.
[25] C. Mealli, D.M. Proserpio, CACAO, Universita´ di Firenze, Italy,
1992.
[26] S.J. Cyvin, J.E. Rauch, J.C. Decius, J. Chem. Phys. 43 (1965) 4083.
[27] G.M. Scheldrick, in: G.M. Scheldrick, C. Kruger, R. Goddard
(Eds.), Crystallographic Computing 3, Oxford University Press,
Oxford, 1985, pp. 175–189.
[28] G.M. Scheldrick,
SHELX93, Crystallographic Calculation Pro-
gram, University of Cambridge, 1993.
[29] C.K. Johnson, ORTEP-II, Report ORNL-5138, Oak Ridge Na-
tional Laboratory, Park Ridge, TN, 1976.
[30] International Tables for X-Ray Crystallography, vol. IV, Kynoch
Press, Birmingham, 1974.
[6] A.R. Dias, M. Helena Garcia, M. Paula Robalo, M.L.H. Green,
K.K. Lai, A.J. Pulham, S.M. Klueber, G. Balavoine, J.
Organomet. Chem. 453 (1993) 241.