Phosphorus-phosphorus coupling constants in mixed-phosphine tricarbonyl iron complexes, Fe(CO)3LL′. Crystal structure of trans-Fe(CO)3(PEt3) (PPh3)

Article abstract of DOI:10.1021/om9609224

Mixed-ligand complexes, trans-Fe(CO)₃LL' (L = PPh₂Me, L' = PPh₂Me, PPhMe₂, PMe₃, PPh₂-Et, PEt₃, PPh₂CH=CH₂, PPh₂H, AsPh₃, P(OPh)₃; L = PMe₃, L' = PEt₃, PPh₂Et, PCy₃, PPh₂Me, PPhMe; L = PEt₃, L' = PPh₂Me; L = PPh₂H, L' = PPh₂CH=CH₂, PPh₂Et; L = AsPh₃, L' = PPhMe₂, P(OPh)₃, P(OMe)3, P(OEt)₃), have been obtained from the stepwise reaction of phosphines with Fe(CO)₃(BDA) (BDA = benzylideneacetone) or Fe(CO)₃(AsPh₃)₂ and from the reaction of phosphine with Fe(CO)₄PPh₃ in the presence of base. A strong negative correlation exists between ²Jpp coupling constant values and the sum of the phosphine pK_a values. By application of quantitative analysis of ligand effects, it has been shown that ²Jpp for the mixed-ligand complexes correlates strongly with both γ and E_ar, but not with θ. Although a near perfect fit is obtained from the three-parameter equation, a statistical analysis suggests that for this small data set there are no predictive advantages over the one-parameter pK_a model. It is possible to calculate reliable ²Jpp values for transFe(CO)₃L₂ complexes with either model. An X-ray structure of solid-state frans-Fe(CO)₃-(PEt₃)(PPh₃) shows equal Fe-PEt₃ and Fe-PPh₃ bond distances, implying that bond strength equalization may occur when two rather different phosphines occupy trans coordination sites.

Full text of DOI:10.1021/om9609224

2246
Organometallics 1997, 16, 2246-2253
P h osp h or u s-P h osp h or u s Cou p lin g Con sta n ts in
Mixed -P h osp h in e Tr ica r bon yl Ir on Com p lexes,
F e(CO)3LL′. Cr ysta l Str u ctu r e of
tr a n s-F e(CO)3(P Et3)(P P h 3)
Richard L. Keiter,* J ohn William Benson,† Ellen A. Keiter, Travis A. Harris,
Matthew W. Hayner, Laura L. Mosimann, Eric E. Karch, Carol A. Boecker,
Donna M. Olson, J ennaver VanderVeen, and Douglas E. Brandt
Department of Chemistry, Eastern Illinois University, Charleston, Illinois 61920
Arnold L. Rheingold* and Glenn P. A. Yap
Department of Chemistry, University of Delaware, Newark, Delaware 19716
Received October 31, 1996X
Mixed-ligand complexes, trans-Fe(CO)3LL′ (L ) PPh3, L′ ) PPh2Me, PPhMe2, PMe3, PPh2-
Et, PEt3, PPh2CHdCH2, PPh2H, AsPh3, P(OPh)3; L ) PMe3, L′ ) PEt3, PPh2Et, PCy3,
PPh2Me, PPhMe; L ) PEt3, L′ ) PPh2Me; L ) PPh2H, L′ ) PPh2CHdCH2, PPh2Et; L )
AsPh3, L′ ) PPhMe2, P(OPh)3, P(OMe)3, P(OEt)3), have been obtained from the stepwise
reaction of phosphines with Fe(CO)3(BDA) (BDA ) benzylideneacetone) or Fe(CO)3(AsPh3)2
and from the reaction of phosphine with Fe(CO)4PPh3 in the presence of base. A strong
2
negative correlation exists between J PP coupling constant values and the sum of the
phosphine pKa values. By application of quantitative analysis of ligand effects, it has been
2
shown that J PP for the mixed-ligand complexes correlates strongly with both ø and Ear, but
not with θ. Although a near perfect fit is obtained from the three-parameter equation, a
statistical analysis suggests that for this small data set there are no predictive advantages
2
over the one-parameter pKa model. It is possible to calculate reliable J PP values for trans-
Fe(CO)3L2 complexes with either model. An X-ray structure of solid-state trans-Fe(CO)3-
(PEt3)(PPh3) shows equal Fe-PEt3 and Fe-PPh3 bond distances, implying that bond strength
equalization may occur when two rather different phosphines occupy trans coordination sites.
In tr od u ction
of the possible disubstituted products resulted and
product separation was not achieved. It is noteworthy
that no crystal structures of mixed-ligand phosphine
tricarbonyl complexes have been reported.
Although reliable methods recently have become
available for the synthesis of Fe(CO)3L2 (L ) PR3)
complexes,1 preparation of disubstituted Fe(CO)3LL′ (L
) PR3; L′ ) PR3′) complexes remains difficult, as shown
by the scarcity of literature examples.2 The most
systematic approach, reported by Cardaci’s group, uti-
lized a hydrosilyl derivative, Fe(CO)3L(H)(SiPh3), from
which HSiPh3 is displaced by L′:2c
In the course of their investigations, Cardaci et al.
obtained phosphorus-phosphorus coupling constants
for several mixed-phosphine-ligand complexes. Par-
ticularly intriguing to us was the linear relationship
2
that was shown to exist between J PP and pKa values
of L in Fe(CO)3(PMe3)L. If the validity of the relation-
ship could be established for a wider range of com-
pounds, it would be possible to determine phosphorus-
phosphorus coupling constants for complexes containing
two identical phosphines [e.g., Fe(CO)3(PPh3)2], for
which direct determination is experimentally difficult.
In this work, we have investigated several synthetic
approaches for production of Fe(CO)3LL′ and have
established that Cardaci’s coupling constant/pKa cor-
relation holds even when L′ is sterically demanding. The
results have been further analyzed within the context
of Giering and Prock’s quantitative analysis of ligand
effects (QALE).3 In addition, we have succeeded in
obtaining the crystal structure of trans-Fe(CO)3(PEt3)-
(PPh3), which permits assessment of the bonding ca-
Fe(CO)4L + HSiPh3 f Fe(CO)3L(H)(SiPh3) + CO
(1)
Fe(CO)3L(H)(SiPh3) + L′ f Fe(CO)3LL′ + HSiPh3
(2)
Highest yields (35%) were obtained when L ) PMe3 and
L′ ) PPh3, but for L′ ) PMe2Ph, PMePh2, or PEt3, all
* To whom correspondence should be addressed. E-mail:
cfrlk@eiu.edu; arnrhein@udel.edu.
† Current address: Department of Chemistry, Rockford College,
Rockford, IL 61108.
X Abstract published in Advance ACS Abstracts, April 1, 1997.
(1) (a) Keiter, R. L.; Keiter, E. A.; Hecker, K. H.; Boecker, C. A.
Organometallics 1988, 7, 2466. (b) Therien, M. J .; Trogler, W. C. Inorg.
Synth. 1990, 28, 173. (c) Bellachioma, G.; Cardaci, G.; Colomer, E.;
Corriu, R. J . P.; Vioux, A. Inorg. Chem. 1989, 28, 519. (d) Keiter, R.
L.; Keiter, E. A.; Boecker, C. A.; Miller, D. R. Synth. React. Inorg. Met.-
Org. Chem. 1991, 21 (3), 473. (e) Sowa, J . R.; Zanotti, V.; Facchin, G.;
Angelici, R. J .; J . Am. Chem. Soc. 1992, 114, 160. (f) Brunet, J .-J .;
Commenges, G.; Kindela, F.-B.; Neilbecker, D. Organometallics 1992,
11, 1343. (g) Brunet, J .-J .; Kindela, F. B.; Neibecker, D. Inorg. Synth.
1992, 29, 151. (h) Luo, L.; Nolan, S. P. Inorg. Chem. 1993, 32, 2410.
(i) Keiter, R. L.; Keiter, E. A.; Boecker, C. A.; Miller D. R.; Hecker, K.
H. Inorg. Synth. 1997, 31, 210.
(2) (a) Ogilvie, F. B.; Keiter, R. L.; Wulfsberg, G.; Verkade, J . G.
Inorg. Chem. 1969, 8, 2346. (b) Allison, D. A.; Clardy, J .; Verkade, J .
G. Inorg. Chem. 1972, 11, 2804. (c) Bellachioma, G.; Cardaci, G.;
Macchioni, A.; Reichenbach, G. J . Organomet. Chem. 1990, 391, 367.
(3) (a) Wilson, M. R.; Woska, D. C.; Prock, A.; Giering, W. P.
Organometallics 1993, 12, 1742. (b) Bartholomew, J .; Fernandez, A.
L.; Lorsbach, B. A.; Wilson, M. R.; Prock, A.; Giering, W. P. Organo-
metallics 1996, 15, 295.
S0276-7333(96)00922-3 CCC: $14.00 © 1997 American Chemical Society

Mixed-Phosphine Tricarbonyl Iron Complexes
Organometallics, Vol. 16, No. 11, 1997 2247
in which the monosubstituted complex is soluble.6 The undis-
solved portion was recrystallized from a CH2Cl2/CH3OH solu-
tion to give the disubstituted product (10.1 g, 40%). Substi-
tuting sodium hydroxide1d for sodium borohydride did not
improve the yield. (b) The triphenylarsine displacement of
BDA from Fe(CO)3(BDA) provided a more selective synthesis
of this compound.1e,7 A solution of Fe(CO)3(BDA) (6.11 g, 21.4
mmol) and AsPh3 (13.1 g, 42.8 mmol) in CH2Cl2 (55 mL) was
stirred at room temperature for 24 h. The tan precipitate was
collected by filtration and recrystallized from CH2Cl2/CH3OH
to give pure product (9.66 g, 60%; dec 200-202 °C6a). IR
pacities of two rather different phosphines in a com-
petitive environment.
Exp er im en ta l Section
Gen er a l Con sid er a tion s. All reactions were carried out
under dry, oxygen-free nitrogen with standard Schlenk tech-
niques. Phosphines were obtained from Aldrich and used
without further purification. Solvents were purged with N2
prior to use. Starting materials, Fe(CO)4(PPh3) and Fe(CO)3-
(BDA) (BDA ) benzylideneacetone), were prepared as de-
scribed in the literature.4,5 Infrared spectra were recorded on
a Nicolet 20 DX-B FT spectrometer. Phosphorus-31 NMR
spectra were obtained from CD2Cl2 or CDCl3 solutions with a
General Electric QE-300 FT NMR spectrometer. Microanaly-
ses were performed at the University of Illinois Microanalytical
Laboratory, Urbana, IL.
(toluene): νCO 1886 cm-1
.
Syn th esis of tr a n s-F e(CO)3(P P h 3)2. A toluene solution
(25 mL) of trans-Fe(CO)3(AsPh3)2 (0.500 g, 0.665 mmol) and
PPh3 (0.349 g, 1.33 mmol) was heated under reflux for 12 h.
The solvent was removed under vacuum, and the product was
recrystallized from CH2Cl2/CH3OH to give 0.38 g (86%) of
spectroscopically pure trans-Fe(CO)3(PPh3)2. IR (CHCl3): νCO
1884 cm-1 31P{1H} NMR (CDCl3): 82.5 ppm.1a
.
St ep w ise R ea ct ion s of P h osp h in es w it h F e(CO)3-
(BDA). Syn th esis of tr a n s-F e(CO)3(P Et3)(P P h 3). To a
solution of Fe(CO)3(BDA) (0.319 g, 1.12 mmol) in toluene (15
mL, distilled from CaH2 under N2) chilled to ice temperature
was added dropwise over 20 min a toluene solution (10 mL) of
PEt3 (0.165 mL, 1.12 mmol). The ice bath was removed after
4 h, and PPh3 (0.291 g, 1.11 mmol) dissolved in toluene (10
mL) was added dropwise over 20 min. After the solution was
stirred for 64 h, the solvent was removed and the residue was
chromatographed on alumina (Brockman, Activity I from
Aldrich, deactivated with 5% w/w water) using a 3:1 mixture
of hexanes/CH2Cl2. The first light yellow band was collected
and found to be a mixture of Fe(CO)3(PPh3)2, Fe(CO)3(PEt3)-
(PPh3), and Fe(CO)3(PEt3)2 by its 31P{1H} NMR spectrum.
Evaporation of the eluate and recrystallization from CH2Cl2/
C5H12 (1:10) gave large, feathery, yellow crystals and small,
yellow blocks. The feathery crystals were separated manually
and found to be pure trans-Fe(CO)3(PEt3)(PPh3) (0.339 g,
58.4%; mp 160-161 °C). IR (hexane): νCO 1886 cm-1. Anal.
Calcd for C27H30FeO3P2: C, 62.33; H, 5.81. Found: C, 62.41;
H, 5.83.
Step w ise Rea ction s of P h osp h in es w ith tr a n s-F e(CO)3-
(AsP h 3)2. Syn th esis of tr a n s-F e(CO)3(AsP h 3)(P P h 3) a n d
tr a n s-F e(CO)3(P P h 3)(P P h Me2). The formation of the arsine
intermediate was maximized when the reaction was carried
out by dropwise addition (85 min) of PPh3 (0.174 g, 0.663
mmol) in toluene (10 mL) to a refluxing toluene solution (20
mL) containing trans-Fe(CO)3(AsPh3)2 (0.500 g, 0.664 mmol).
The total reaction time was 4 h. Analysis of the crude reaction
mixture by 31P{1H} NMR showed formation of trans-Fe(CO)3-
(AsPh3)(PPh3) and trans-Fe(CO)3(PPh3)2 (2:1). Separation of
the two compounds by recrystallization and by column chro-
matography was attempted but not achieved. No improve-
ment in mixed-ligand product formation was noted when
reaction times were increased to 8 h. When the reaction was
carried out in lower boiling solvents (cyclohexane, acetone,
isopropyl alcohol, or THF), trans-Fe(CO)3(PPh3)2 formed ex-
clusively. The 4 h preparation described above was repeated,
and to the resulting crude reaction mixture, PPhMe2 (0.091
g, 0.664 mmol) was added. After the solution was heated at
reflux temperature for 4 h, the solvent was removed and the
crude reaction mixture was shown by 31P{1H} NMR analysis
to contain all of the possible disubstituted products, including
trans-Fe(CO)3(PPh3)(PPhMe2). Also prepared by this method
were trans-Fe(CO)3(PPh3)(PPh2H) and trans-Fe(CO)3(PPh3)-
[P(OPh)3].
Rea ction s of P (OR)3 (R ) Me, Et, P h ) w ith tr a n s-F e-
(CO)3(AsP h 3)2. A toluene solution (25 mL) of trans-Fe(CO)3-
(AsPh3)2 (0.50 g, 0.66 mmol) and P(OMe)3 (0.08 g, 0.66 mmol)
was refluxed (oil-bath temperature thermostated at 120 °C)
for 8 h. The solvent was removed from the solution, and the
residue was placed on a silica gel column and eluted with
toluene. The first fraction collected was the mixed-ligand
product, trans-Fe(CO)3(AsPh3)[P(OMe)3], followed by trans-Fe-
(CO)3[P(OMe)3]2, and a third fraction consisting of AsPh3. The
mixed-ligand product was further purified by recrystallization
from CH2Cl2/CH3OH to give 0.025 g (6.5%) of yellow trans-
Fe(CO)3(AsPh3)[P(OMe)3]; mp 201-202 °C, dec. Anal. Calcd
for C24H24AsFeO6P: C, 50.56; H, 4.24. Found: C, 50.35; H,
4.12. By the same procedure, trans-Fe(CO)3(AsPh3)[P(OEt)3]
and trans-Fe(CO)3(AsPh3)[P(OPh)3] were prepared.
Mixed-ligand complexes trans-Fe(CO)3(PMe3)(PPh3), trans-
Fe(CO)3(PPh2H)(PPh2CHdCH2), trans-Fe(CO)3(PEt3)(PPh2-
Me), trans-Fe(CO)3(PMe3)(PPh2Et), trans-Fe(CO)3(PMe3)(PCy3),
trans-Fe(CO)3(PPh2H)(PPh2Et), and trans-Fe(CO)3(PMe3)-
(PEt3) were prepared by similar procedures and identified by
their 31P{1H} NMR spectra.
Rea ction s of P h osp h in es w ith F e(CO)4(P P h 3) in th e
P r esen ce of Na OH. tr a n s-F e(CO)3(P P h 3)(P P h 2Et). To 50
mL of 1-butanol was added Fe(CO)4(PPh3) (1.00 g, 2.33 mmol),
NaOH (0.22 g, 5.3 mmol), and PPh2Et (0.50 mL, 2.3 mmol).
The solution was heated at reflux for 2 h, cooled to room
temperature, and filtered to give a solid product shown by 31P-
{1H} NMR spectral integration to consist of trans-Fe(CO)3-
(PPh3)(PPh2Et) (70%), trans-Fe(CO)3(PPh2Et)2 (19%), and
trans-Fe(CO)3(PPh3)2 (11%). Attempts to separate these com-
plexes by chromatography were not successful.
The reaction of Fe(CO)4(PPh3) with PPh2Me, prepared as
described above, gave trans-Fe(CO)3(PPh3)(PPh2Me), trans-Fe-
(CO)3(PPh2Me)2, and trans-Fe(CO)3(PPh3)2 in a spectral ratio
of 4.4:1.8:1.0 and with PPh2CHdCH2 gave trans-Fe(CO)3-
(PPh3)(PPh2CHdCH2), trans-Fe(CO)3(PPh3)2, and trans-Fe-
(CO)3(PPh2CHdCH2)2 in a ratio of 12:7.0:1.0.
A 1:5 molar ratio of trans-Fe(CO)3(AsPh3)2 (0.5 g, 0.66 mmol)
to P(OR)3 (R ) Me, Et, Ph) (3.2 mmol) in 25 mL of toluene
was refluxed (oil-bath temperature thermostated at 120 °C)
for 24 h. The solvent was removed, and integral product
percentages were obtained from the 31P{1H} NMR spectra
(Table 1). Carbonyl stretching frequencies of the trans-Fe-
(CO)3(AsPh3)[P(OR)3] and trans-Fe(CO)3[P(OR)3]2 complexes,
separated by chromatography as described above, are given
in Table 2.
Syn th esis of tr a n s-F e(CO)3(AsP h 3)2. (a) The reaction of
Fe(CO)5 (6.5 g, 33 mmol) with AsPh3 (20.4 g, 66.6 mmol) and
NaBH4 (1.26 g, 33.3 mmol) in refluxing 1-butanol (100 mL)
for 2 h gave a precipitate consisting of Fe(CO)4AsPh3 and trans-
Fe(CO)3(AsPh3)2.1a Separation of the two complexes was
achieved by extracting the solid residue with boiling heptane,
(4) Albers, M. O.; Singleton, E.; Coville, N. J . Inorg. Synth. 1990,
28, 168.
(5) Howell, J . A. S.; J ohnson, B. F. G.; J osty, P. L.; Lewis, J . J .
Organomet. Chem. 1972, 39, 329.
(6) (a) Davison, A.; McFarlane, W.; Pratt, L.; Wilkinson, G. J . Chem.
Soc. 1962, 3653. (b) Modi, S. P.; Atwood, J . D. Inorg. Chem. 1983, 22,
26.
(7) Li, C.; Nolan, S. P. Organometallics 1995, 14, 1327.

2248 Organometallics, Vol. 16, No. 11, 1997
Keiter et al.
Ta ble 2. In fr a r ed Da ta (CO Str etch in g
Ta ble 1. Rela tive Yield s (%) fr om Rea ction s of
a
tr a n s-F e(CO)3(AsP h 3)2 w ith P (OR)3
F r equ en cies) for P h osp h ite Com p lexes of
Ir on Tr ica r bon yla
reactant
ratiob
Fe(CO)3-
Fe(CO)3-
Fe(CO)2-
ligand
(AsPh3)[P(OR)3] [P(OR)3]2 [P(OR)3]3
trans-
trans-
L
Fe(CO)3AsPh3L
Fe(CO)3L2
ref
P(OMe)3
P(OEt)3
P(OPh)3
1:1
1:1
1:1
74.2
80.2
87.6
25.8
19.8
12.4
0
0
0
P(OMe)3
P(OEt)3
P(OPh)3
1900(s), 1910(s)
1898(s), 1910(s)
1909(s)
1913(s), 1922(s)
1907(s), 1917(s)
1926(s)
30
30a
30, 6b
P(OMe)3
P(OEt)3
P(OPh)3
1:5
1:5
1:5
53.7
43.8
66.0
42.9
53.2
33.0
3.4
3.0
1.0
In hexane, cm-1
.
a
Ta ble 3. Cr ysta l Da ta a n d Str u ctu r e Refin em en t
a
The reactions with a 1:1 mole ratio were carried out for 8.0 h,
and the reactions with a 1:5 mole ratio were carried out for 24.0
for tr a n s-F e(CO)3(P Et3)(P P h 3)
b
h, both in refluxing toluene. Complex/phosphite.
empirical form
fw
C27H30FeO3P2
520.3
temp
296(2) K
Cr ysta llogr a p h ic Str u ctu r a l Deter m in a tion of F e-
(CO)3(P Et3)(P P h 3). A suitable crystal of trans-Fe(CO)3-
(PEt3)(PPh3) was obtained by diffusion of a layer of pentane
into a CH2Cl2 solution of Fe(CO)3(PEt3)(PPh3) at -18 °C under
nitrogen. Crystallographic data are given in Table 3. Photo-
graphic evidence and systematic absences in the diffraction
data uniquely identified the space group. No correction for
absorption was required. The structure was refined with all
non-hydrogen atoms anisotropic and all hydrogen atoms
idealized. All computations used SHELXTL, version 5, soft-
ware (G. Sheldrick, Siemans XRD, Madison, WI).
wavelength
cryst syst
space group
unit cell dimens
0.710 73 Å
monoclinic
P21/n
a ) 11.314(3) Å, R ) 90°
b ) 16.042(4) Å, â ) 93.58(2)°
c ) 14.813(4) Å, γ ) 90°
2683(1) Å3, 4
volume, Z
density (calcd)
abs coeff
F(000)
cryst size
θ range for data collection
limiting indices
1.288 g/cm3
0.706 mm-1
1088
0.40 × 0.37 × 0.35 mm
2.20-25.00°
-13 e h e 12, 0 e k e 19,
-17 e l e 0
no. reflns collected
no. indep reflns
refinement method
data/restraints/
4611
Resu lts a n d Discu ssion
4443 (Rint ) 0.0439)
full-matrix least-squares on F2
4400/0/298
Syn th etic Asp ects. Vessieres and Dixneuf found
that Fe(CO)3(BDA) reacts with 1 mol of PPhMe2 to give
Fe(CO)3(BDA)(PPhMe2), a complex in which BDA is
bound as a monodentate ligand.8 In addition, Cardaci
and Sorriso were able to isolate Fe(CO)3(BDA)(SbPh3).9
However, the reaction of PPh3 (less basic than PMe2-
Ph) with Fe(CO)3(BDA) gave a mixture of Fe(CO)3-
(BDA)(PPh3) and Fe(CO)2(BDA)(PPh3).10 More recently,
it has been demonstrated that 2 mol of PR3 displace
BDA from Fe(CO)3(BDA) rapidly and cleanly to give
trans-Fe(CO)3(PR3)2.1e,h,7 These results suggested a
potential method for the preparation of mixed-ligand
phosphine complexes from Fe(CO)3(BDA) via a two-step
reaction:
parameters
goodness-of-fit on F2
final R indices [I > 2σ(I)]
final R indices (all data)
largest diff. peak and hole
1.064
R1 ) 0.660, wR2 ) 0.1626
R1 ) 0.1285, wR2 ) 0.2246
0.846 and -0.373 e Å-3
Identification of the Fe(CO)3LL′, Fe(CO)3L2, and Fe-
(CO)3L′2 complexes by 31P{1H} NMR spectroscopy is
straightforward. The mixed-ligand complexes give the
2
expected two doublets with J PP values ranging from
22.9 to 33.2 Hz (Table 4). When all three complexes
are present, the two doublets tend to lie between
singlets arising from the two symmetrical products. For
example, the 31P chemical shifts of trans-Fe(CO)3(PPh3)2
and trans-Fe(CO)3(PMe3)2 are singlets at 82.6 and 39.3
ppm, respectively, while the chemical shifts of trans-
Fe(CO)3(PMe3)(PPh3) are 81.8 and 42.9 ppm. The signal
for PPh3 shifts to a lower frequency when the more basic
PMe3 occupies the trans position, while the signal for
PMe3, when it is trans to the less basic PPh3, shifts to
higher frequency.
Our second approach for preparing mixed-ligand
complexes was based on a synthesis developed for Fe-
(CO)3(PR3)2 complexes. When Fe(CO)5 is heated under
reflux in 1-butanol with 2 mol of PR3 in the presence of
NaBH4, trans-Fe(CO)3(PR3)2 is formed selectively.1a,i
When Fe(CO)4PPh3 is substituted for Fe(CO)5, but
conditions are otherwise unchanged, no reaction occurs.
However, if sodium hydroxide is used in place of sodium
borohydride, conversion of Fe(CO)4PPh3 to trans-Fe-
(CO)3(PPh3)2 takes place in high yield.1a,d Thus, it
appeared that it would be possible to prepare Fe(CO)3-
LL′ complexes from Fe(CO)4L and L′ by this method:
Even without isolation of the intermediate Fe(CO)3L-
(BDA) in these reactions, the major product for eight
reactions was the mixed-ligand derivative. In all cases,
the symmetrical derivatives, Fe(CO)3L2 and Fe(CO)3L′2,
were present as minor products. Although traces of Fe-
(CO)2L(BDA) were present in product mixtures, its
formation was minimized by adding the more nucleo-
philic phosphine first (e.g., PMe3 before PPh3) at 0 °C.
Fe(CO)4(PR3) + OH- f [HFe(CO)3PR3]- + CO2
(5)
[HFe(CO)3(PR3)]- + PR′3 f
(8) Vessieres, A.; Dixneuf, P. Tetrahedron Lett. 1974, 16, 1499.
(9) Cardaci, G.; Sorriso, S. Inorg. Chem. 1976, 15, 1242.
(10) Cardaci, G.; Concetti, G. J . Organomet. Chem. 1974, 90, 49.
trans-Fe(CO)3(PR3)(PR′3) + H2 (6)

Mixed-Phosphine Tricarbonyl Iron Complexes
Ta ble 4. 31P {1H} NMR (CD2Cl2) a n d p Ka Da ta for Ca r bon yl Com p lexes of Ir on a
trans-Fe(CO)3LL′
δL (ppm)
Organometallics, Vol. 16, No. 11, 1997 2249
L
L′
δ
L′ (ppm)
2J PP (Hz)
(pKa + pKa′)a
PPh3
PPh3
PPh3
PPh3
PPh3
PPh3
PPh3
PPh2Et
PPh2CHdCH2
PPh2Me
PPh2Et
PPh2Me
PPhMe2
PEt3
PPh2CHdCH2
PPh2Me
PPh2H
PPh2Et
PPhMe2
PMe3
83.1
82.5
82.1
82.9
82.0
81.8
82.5
77.1
76.3
67.0
76.4
70.6
52.1
72.1
45.0
79.3
76.7
66.3
55.6
77.8
52.2
42.9
71.4
55.4
55.1
44.2
42.3
65.4
43.2
43.9
88.1
186.0
55.6
85.3
185.6
192.2
185.3
33.2
31.8
31.1
30.9
30.0
28.7
27.8
29.4
28.8
26.8b
26.7
25.9
24.5b
22.9
23.5
86.3
2.73 + (?)
2.73 + 4.57
2.73 + 0.03
2.73 + 4.9
2.73 + 6.50
2.73 + 8.65
2.73 + 8.69
4.9 + 0.03
(?) + 0.03
PEt3
PPh2H
PPh2H
PMe3
4.57 + 8.65
4.9 + 8.65
4.57 + 8.69
6.50 + 8.65
8.69 + 8.65
8.65 + 9.7
2.73 + (-2.0)
PMe3
PEt3
PMe3
PMe3
PMe3
PPh3
PCy3
P(OPh)3
PPhMe2
PPh3
AsPh3
AsPh3
AsPh3
P(OPh)3
P(OMe)3
P(OEt)3
AsPh3
AsPh3
trans-Fe(CO)3L2
L
δ (ppm)
ref
L
δ (ppm)
ref
P(OMe)3
P(OEt)3
P(OPh)3
PPh3
PPhMe2
PPh2CHdCH2
190.7
185.0
182.0
82.6
PPh2Me
PPh2H
PPh2Et
PMe3
65.4
54.3
77.3
39.3
70.5
1a
1a
1a
1b
1a
1f
1a
50.3
PEt3
76.7
Fe(CO)2L3
L
δ (ppm)
ref
L
δ (ppm)
ref
31
P(OPh)3
P(OEt)3
165.8
182.8
31
31
P(OMe)3
188.6
a
b
pKa values were taken from ref 15. Data taken from ref 2c.
The harsh reaction conditions required (strong base,
>90 °C) led to some product scrambling such that, in
each case, both symmetrical products were produced
along with the mixed-ligand complex (60-70%). At
temperatures lower than 90 °C, essentially no reaction
took place over the 2 h reaction time employed. Sub-
stitution of [Et4N]OH for NaOH did not lead to greater
product selectivity.
Kinetic studies of the reaction of trans-Fe(CO)3-
(AsPh3)2 with CO, reported by Modi and Atwood,
showed that AsPh3 is replaced by CO to give Fe(CO)4-
(AsPh3) (k ) 4.57 × 10-5 s-1 at 90 °C).6b Furthermore,
their work revealed that the reaction of trans-Fe(CO)3-
(PPh3)2 with CO is much slower and requires a much
higher temperature in order to achieve a rate compa-
rable to that of the arsine reaction (k ) 2.64 × 10-5 s-1
at 136 °C). These reactions proceed by a dissociative
mechanism and indicate that the Fe-AsPh3 bond is
significantly weaker than the Fe-PPh3 bond, a conclu-
sion verified by Luo and Nolan from thermodynamic
studies.1h These and other results11 indicated that
stepwise replacement of AsPh3 from trans-Fe(CO)3-
(AsPh3)2 by phosphines may provide easy access to
mixed-ligand complexes:
(CO)3(PPh3)2 in essentially quantitative yield, 1 mol of
PPh3 gives a 2:1 mixture of trans-Fe(CO)3(AsPh3)(PPh3)
and trans-Fe(CO)3(PPh3)2. The latter reaction was
performed under a variety of conditions in an attempt
to optimize the formation of the mixed-ligand product.
The ratio was not improved in favor of the mixed-ligand
complex by changing reaction times from 4 h to 2, 6, or
8 h. On the premise that lower temperatures might
discourage displacement of both arsine ligands, lower
boiling solvents (cyclohexane, isopropyl alcohol, THF,
and acetone) were substituted for toluene. However,
only trans-Fe(CO)3(PPh3)2 was produced from these
lower temperature reactions. Although these observa-
tions could be explained in terms of ligand scrambling
occurring at higher temperatures but not lower, sepa-
rate experiments showed that a mixture of trans-Fe-
(CO)3(AsPh3)2 and trans-Fe(CO)3(PPh3)2 does react to
give mixed-ligand complexes. Furthermore, no reaction
occurs when trans-Fe(CO)3(PPh3)2 and AsPh3 are heated
under similar conditions. We conclude that the activa-
tion energy for the conversion of trans-Fe(CO)3(AsPh3)-
(PPh3) to trans-Fe(CO)3(PPh3)2 is lower than that for
the conversion of trans-Fe(CO)3(AsPh3)2 to trans-Fe-
(CO)3(AsPh3)(PPh3), but as the temperature is in-
creased, the slower reaction becomes relatively more
favorable, which accounts for the appearance of the
mixed-ligand complex in the crude reaction mixtures
produced at higher temperatures.
trans-Fe(CO)3(AsPh3)2 + PR3 f
trans-Fe(CO)3(AsPh3)(PR3) + AsPh3 (7)
trans-Fe(CO)3(AsPh3)(PR3) + PR′3 f
trans-Fe(CO)3(PR3)(PR′3) + AsPh3 (8)
A number of reactions were carried out between trans-
Fe(CO)3(AsPh3)2 and P(OMe)3, P(OEt)3, or P(OPh)3.
(11) (a) Wovkulich, M. J .; Atwood, J . D. Organometallics 1982, 1,
1316. (b) Fontaine, X. L. R.; Fowles, E. H.; Layzell, T. P.; Shaw, B. L.;
Thornton-Pett, M. J . Chem. Soc., Dalton Trans. 1991, 1519.
Although 2 mol of PPh3 displace AsPh3 from trans-
Fe(CO)3(AsPh3)2 in refluxing toluene to give trans-Fe-

2250 Organometallics, Vol. 16, No. 11, 1997
Keiter et al.
F igu r e 1. Plot of experimentally determined phosphorus-
phosphorus coupling constants vs the sum of the pKa values
for the two phosphines in trans-Fe(CO)3LL′.
F igu r e 2. Three-dimensional plot showing correlations
between J PP, pKa, and phosphine cone angle, θ, for trans-
Fe(CO)3LL′.
2
Whereas phosphines displace arsines quantitatively,
phosphites do not. Even when trans-Fe(CO)3(AsPh3)2
was heated under reflux in toluene for 24 h with a 5-fold
excess of phosphite, formation of trans-Fe(CO)3[(P(OR)3]2
was incomplete (Table 1). When 1:1 mole ratios of the
reactants were heated under reflux in toluene for 8 h,
mixtures of trans-Fe(CO)3[P(OR)3](AsPh)3 and trans-Fe-
(CO)3[P(OR)3]2 were obtained (Table 1). Although the
reactions most likely proceed primarily by AsPh3 dis-
sociation, the data in Table 1 suggest that the incoming
ligands play a role in the substitution process (cone
angles12 for P(OMe)3, P(OEt)3, and P(OPh)3 are 107°,
109°, and 128°, respectively, and pKa values13 are 2.60,
3.31, and -2.0, respectively). The better nucleophiles,
P(OMe)3 and P(OEt)3 (and sterically less hindered), lead
to higher conversions to di- and trisubstituted products
than does P(OPh)3.
2
2
F igu r e 3. Plot of 2J PP (calculated) vs J PP (experimental)
Rela tion sh ip betw een J P P a n d p Ka . It has been
noted that absolute values of the phosphorus-phospho-
rus coupling constants for the complexes trans-Fe(CO)3-
(PMe3)L (L ) PMe3, PEt3, PMe2Ph, PMePh2, PPh3) are
linearly related to the pKa values of the ligands, L,2c
with less basic ligands associated with larger values of
2J PP. This correlation is in agreement with the general
for trans-Fe(CO)3LL′ complexes (eq 9).
containing a sterically demanding phosphine (cone angle
) 170°), does not depart significantly from the other
data collected. Correlations with pKas and cone angles
are displayed in three dimensions in Figure 2. The two-
dimensional contour reveals a strong negative correla-
tion of J PP with pKa, but an insignificant correlation
with phosphine cone angles. Equation 9 allows the
2
rule that the magnitude of J PP increases as the elec-
tronegativities of groups attached to phosphorus in-
crease14 and that the affinities of phosphines for protons
lessen (and pKa values decrease) as substituent elec-
tronegativities increase.15
2
calculation of J PP coupling constants to within 2% of
2
We have examined the J PP/pKa relationship for a
the experimental values (shown graphically in Figure
3). Data collected for phosphite and secondary phos-
phine complexes deviated substantially from those for
mixed-ligand tertiary phosphine complexes and were
not included in Figures 1, 2, or 3. In addition, complexes
of PPh2CHdCH2 were omitted because experimental
values of pKa are not available.
range of Fe(CO)3LL′ complexes in which both phos-
phines are allowed to vary. To take into account these
2
ligand combinations, we have plotted J PP vs the sum
of the two pKa values, pKa + pKa′ (Figure 1, Table 4).
2
The plot shows that the strong J PP/pKa correlation
remains intact for the much larger and more diverse
data set. The equation describing the relationship is
Qu a n tita tive An a lysis of Liga n d Effects. The pKa
values used in eq 9 are based on proton dissociation
from R3PH+ in polar aprotic media and are referenced
to aqueous solution. Evidence has been presented by
Giering16 and more recently by Poe¨17 that these values
have a steric component. Modified pKa values, from
which the steric contributions have been deducted, are
thought to better account for the nucleophilic behavior
2J PP ) (-0.80 ( 0.05)(pKa + pKa′) + (37.3 ( 0.6)
(9)
n ) 11 R2 ) 0.968
Notably, even trans-Fe(CO)3(PMe3)(PCy3), a complex
(12) (a) Tolman, C. A. Chem. Rev. 1977, 77, 313. (b) Rahman, M.;
Liu, H. Y.; Prock, A.; Giering, W. P. Organometallics 1987, 6, 650.
(13) Streuli, C. A. Anal. Chem. 1960, 32, 985.
(14) J ameson, C. J . In Phosphorus-31 NMR Spectroscopy in Stere-
ochemical Analysis; Verkade, J . G., Quin, L. D., Eds.; VCH: Deerfield
Beach, FL, 1993.
(16) Liu, H.-Y.; Eriks, K.; Prock, A. Giering, W. P. Organometallics
1990, 9, 1758.
(17) Hudson, R. H. E.; Poe¨, A. J . Organometallics 1995, 14, 3238.
(15) Bush, R. C.; Angelici, R. J . Inorg. Chem. 1988, 27, 681.

Mixed-Phosphine Tricarbonyl Iron Complexes
Organometallics, Vol. 16, No. 11, 1997 2251
Ta ble 5. P a r a m eter s Used for Deter m in a tion of
Equ a tion 11
estimate of their uncertainty. Examining Giering and
Prock’s method of determining Ear,3a it does not seem
unreasonable and seems even conservative to estimate
an uncertainty of (0.3 in the values of Ear. A reason-
able estimate of the uncertainty in ø is 0.1. If the error
in each variable is assumed to be independent and
normally distributed, a calculation of the expected sum
of square residuals for eq 11 results in a value of about
7 Hz. The actual sum of square residuals for the fitted
equation is 0.7 Hz. Why is the difference between the
expected sum of square residuals and the actual sum
of square residuals important? The fitting process
cannot distinguish between two different models when
it yields for both a sum of square residuals that is less
than the expected sum for each model. Any sum of
square residuals that is less than the expected value
for a perfect model is less only because of an accidental
fit of the particular data set. As an example, consider
the model
L
L′
2J PP (Hz)
ø
θ
Ear
PPh3
PPh3
PPh3
PPh3
PPh2Me
PPh2Et
PPhMe2
PMe3
PEt3
PMe3
PMe3
PEt3
31.8
30.9
30.0
28.7
27.8
26.8
26.7
25.9
24.5
22.9
23.5
25.35
24.55
23.85
21.80
19.55
20.65
19.85
18.4
281
285
267
263
277
254
258
268
240
250
288
4.7
4.7
3.7
2.7
2.7
2.0
2.0
2.0
1.0
0.0
0.0
PPh3
PPh2Me
PPh2Et
PPh2Me
PPhMe2
PEt3
PMe3
PMe3
PCy3
19.15
14.85
9.95
PMe3
of phosphines in nonpolar solvents. When the scaled
values from Poe¨’s work were used in eq 9, a straight
line with a R2 value of 0.869 was obtained. This poorer
correlation may in fact demonstrate that the better
correlation given by traditional pKa values results from
an inherent steric component that is built into tradi-
tional pKas.
2J PP ) AEar + B
(12)
Giering and Prock have shown that many physio-
chemical properties can be related by the equation
a linear correlation using only the Ear parameter.
2
Linear regression applied to our J PP data results in
property ) aø + b(θ - θst) + cEar + d
(10)
2J PP ) (1.79 ( 0.09)Ear + (23.1 ( 0.3)
n ) 11 R2 ) 0.976
(13)
in which ø is an electronic factor, θ is Tolman’s cone
angle, θst is the steric threshold, and Ear is an aryl effect
parameter.3b Analysis of data from many kinetic stud-
ies suggests that the steric term is of no consequence
until the cone angle reaches the steric threshold, at
which point it turns on and has a significant bearing
on the observed rate.
A calculation of the expected sum of square residuals
results in 2.1 Hz, which is identical to the actual sum
(2.1 Hz). Thus, the model accounts for all of the
variations in the data except that expected due to
uncertainties inherent in the data itself. Even though
the QALE model results in a higher correlation coef-
ficient than the simpler one (eqs 12 and 13), there is no
statistical basis for using the correlation coefficient to
accept one model over the other. This in no way implies
that the QALE model is inappropriate, only that this
particular set of data cannot be used to distinguish
between the complete QALE model and some other
simpler model with fewer parameters by simply com-
paring the coefficient of correlation between the experi-
mental values and model predicted values. It only
points out the importance of considering more than
correlation coefficients in deciding between competing
models. A more complete analysis is possible using
analysis of variance techniques, but that would require
more detailed information about the uncertainties in the
independent variables.
Nolan18 has applied the QALE analysis to heats of
reaction obtained for the formation of Fe(CO)3(PR3)2
from phosphine and Fe(CO)3(BDA). Excellent correla-
tions were found between ∆H and both the electronic
and steric factors, with a steric threshold of 135°. It
was of interest to determine if eq 10 could be also
2
applied to our J PP data.
The following equation was determined by multiple
linear regression using the experimental phosphorus-
phosphorus coupling values, 2J PP, and literature values
of ø, θ, and Ear (Table 5):
2J PP ) (2.15 ( 0.58)ø + (0.286 ( 0.076)θ -
(4.80 ( 1.8)Ear - (80.2 ( 28) (11)
n ) 11 R2 ) 0.992
This appears to be an unusually excellent fit of the
QALE model to the data. Caution must be used,
however, before declaring this a meaningful result. The
standard linear regression procedures used to fit the
model equation to the data assume no uncertainty in
the independent variables. However, the uncertainties
The QALE analysis of our data does not reveal a steric
threshold, perhaps surprising in view of Nolan’s ther-
mochemical results, but our largest cone angle sum is
288°, or 144°/ligand, which is not an extreme value. It
is possible that a threshold is present at some greater
cone angle sum. Cone angle changes could influence
2J PP if these changes were accompanied by changes in
the R-P-R bond angle (and consequently changes in
the s character present in the metal-phosphorus bond)
or if steric repulsion prevented maximum overlap of the
phosphorus and iron bonding orbitals. As seen in the
crystallographic section which follows, the Fe-P dis-
tances and R-P-R angles change minimally at best as
the substituents on phosphorus are changed, and there-
fore, one is led to expect that cone angle changes do not
2
in the experimental values of J PP and in the indepen-
dent variables will lead to an expected sum of square
residuals, even for an exact model. To estimate this
value requires a knowledge of the uncertainties in the
2
2
values of J PP, ø, θ, and Ear. Our J PP values have an
uncertainty of (0.1 Hz, and θ has an uncertainty of at
least (2°.12a A value of (3° or more may be more
reasonable for some of the ligands. Unfortunately, the
values of Ear are reported in the literature with no
2
have much bearing on J PP within the range of ligands
(18) Li, C.; Stevens, E. D.; Nolan, S. P. Organometallics 1995, 14,
3791.
employed in this study.

2252 Organometallics, Vol. 16, No. 11, 1997
Keiter et al.
2
a
Ta ble 6. J P P for F e(CO)3L2 Ca lcu la ted fr om p Ka
Ta ble 7. Selected Bon d Len gth s (Å) a n d An gles
Va lu esa a n d QALE F a ctor sb
(d eg) for tr a n s-F e(CO)3(P Et3)(P P h 3)
calcd
calcd
Distances
complex
Fe(CO)3(PPh3)2
exptl
30.4c
(eq 9)
(eq 11)
Fe-P(1)
2.204(2)
2.201(2)
1.137(8)
1.153(8)
1.151(8)
1.781(10)
1.866(11)
Fe-C(41)
Fe-C(42)
Fe-C(43)
P(1)-C(6)
P(1)-C(16)
P(1)-C(26)
P(2)-C(36)
1.773(8)
1.768(8)
1.767(8)
1.836(7)
1.834(7)
1.836(7)
1.843(12)
Fe-P(2)
32.9
31.1
30.0
29.5
26.9
23.5
23.4
21.8
33.6
C(41)-O(41)
C(42)-O(42)
C(43)-O(43)
P(2)-C(32)
P(2)-C(34)
Fe(CO)3[η1-PPh2CH2CH2PPh2]2
Fe(CO)3(PPh2Me)2
Fe(CO)3(PPh2Et)2
Fe(CO)3(PPhMe2)2
Fe(CO)3(PMe3)2
30.4
29.3
25.6
24.1
22.4
23.0
23.5d
Fe(CO)3(PEt3)2
Fe(CO)3(PCy3)2
Angles
P(1)-Fe-P(2)
177.34(9)
123.6(4)
117.(3)
92.1(2)
88.8(2)
C(43)-Fe-C(42)
Fe-C(41)-O(41)
C(41)-Fe-P(1)
C(43)-Fe-P(1)
C(42)-Fe-P(2)
Fe-C(42)-O(42)
118.6(3)
178.0(7)
88.7(2)
91.9(2)
89.7(2)
178.6(7)
C(43)-Fe-C(41)
C(42)-Fe-C(41)
C(42)-Fe-P(1)
C(41)-Fe-P(2)
C(43)-Fe-P(2)
Fe-C(43)-O(43)
a
b 2
2J PP ) (-0.80 ( 0.05)(pKa + pKa′) + (37.3 ( 0.6).
J
) (2.15
PP
( 0.58)ø + (0.286 ( 0.076)θ - (4.8 ( 1.8)Ear - (80.2 ( 28).
c Analysis of second-order proton spectrum, ref 19. Analysis of
d
second-order phosphorus spectrum, ref 20.
88.9(2)
177.8(7)
example of an Fe(CO)3 fragment bound to two different
phosphine ligands, and it provides a comparison to
recently reported structures of complexes with two
identical phosphines, (Fe(CO)3(PPh3)2,21 Fe(CO)3(PPh2-
Cy)2,18 and Fe(CO)3(PPh2Me)2.22
The geometry about iron is approximately trigonal
bipyramidal, with equatorial CO ligands and axial
phosphine ligands (P-Fe-P angle ) 177.34(9)°). Two
of the C-Fe-C angles (118.6(3)°, 117.7(3)°) are nearly
equivalent, while the third is substantially larger (123.6-
(4)°), a trend observed previously in bis(phosphine)
derivatives and ascribed to packing effects.21,22
The Fe atom and the three CO groups are nearly
coplanar, but the three CO groups are bent slightly
toward the triethylphosphine ligand (average C-Fe-P
angle ) 89.1(2)°) and away from the triphenylphosphine
ligand (average C-Fe-P angle ) 90.9(2)°), suggesting
a slight mutual steric awareness of the two phosphines.
The most unusual feature of the structure is the Fe-P
bond distances, which are identical within experimental
error (2.204(2) and 2.201(2) Å).24 If bond length cor-
relates with bond strength, a concept which has been
recently questioned,23 the Fe-PPh3 and Fe-PEt3 bonds
are of equal strength. Support for equal strength is
provided by the positive correlation between the Fe-P
bond length and bond energy data.18 When PPh3, PPh2-
Cy, and PPh2Me react with Fe(CO)3(BDA) to give
disubstituted trans-Fe(CO)3(PR3)2 complexes, 26.9, 27.5,
and 34.1 kcal are liberated, respectively, with Fe-P
bond lengths in the resulting complexes of 2.2173(9),
2.218(1), and 2.2059(14) Å, respectively. In other words,
the significantly shorter Fe-P bond in Fe(CO)3(PPh2-
Me)2 is associated with a significantly larger Fe-P bond
energy.
F igu r e 4. Molecular structure of trans-Fe(CO)3(PEt3)-
(PPh3).
Equations 9 and 11 allow one to calculate three
2
important types of data: (1) J PP for trans-Fe(CO)3L2
complexes in which the phosphines are equivalent; (2)
2J PP values for mixed-phosphine complexes not yet
synthesized; and (3) phosphine pKa values not yet
2
experimentally measured. Whereas J PP has been
determined for trans-Fe(CO)3(PMe3)2 by proton NMR
analysis of its X9AA′X′9 spin system2c,19 and for trans-
Fe(CO)3(η1-PPh2CH2CH2PPh2)2 by phosphorus NMR
analysis of its XAA′X′ spin system,20 obtaining values
by similar methods for more complex systems such as
trans-Fe(CO)3(PPh3)2 is not straightforward. Values of
2J PP for several bis(phosphine) complexes, calculated
from eqs 9 and 11, are shown in Table 6. Similar
Increasing the basicities of ligands (L) trans to triph-
enylphosphine in Fe(CO)3(PPh3)L would be expected to
increase the Fe-PPh3 bond strength. Substituting the
CO group trans to PPh3 in trans-Fe(CO)4PPh3 with a
second PPh3 ligand gives trans-Fe(CO)3(PPh3)2, in which
the average Fe-P distance (2.2173(9) Å) is significantly
2
calculations enable one to obtain J PP for mixed-ligand
complexes not yet synthesized. An example of the pKa
determination is provided by PPh2CHdCH2, for which
2
the estimated value (eq 9) is 2.4 based on the J PP for
trans-Fe(CO)3(PPh3)(PPh2CHdCH2) (33.2 Hz) and the
pKa for PPh3 (2.73).
St r u ct u r e of tr a n s-F e(CO)3(P E t 3)(P P h 3). The
structure of trans-Fe(CO)3(PEt3)(PPh3) is shown in
Figure 4; selected structural parameters are given in
Table 7. This is the first structurally characterized
(21) The structure of Fe(CO)3(PPh3)2 has been reported by two
groups (the latter one is reported with crystallized ether): (a) Glaser,
R.; Yoo, Y.-H.; Chen, G. S.; Barnes, C. L. Organometallics 1994, 13,
2578. (b) Lane, H. P.; Godfrey, S. M.; McAuliffe, C. A.; Pritchard, R.
G. J . Chem. Soc., Dalton Trans. 1994, 3249.
(22) Glaser, R.; Haney, P. E.; Barnes, C. L. Inorg. Chem. 1996, 35,
1758.
(23) (a) Alyea, E. C.; Song, S. Inorg. Chem. 1995, 34, 3864. (b) Ernst,
R. D.; Freeman, J . W.; Stahl, L.; Wilson, D. R.; Arif, A. M.; Nuber, B.;
Ziegler, M. L. J . Am. Chem. Soc. 1995, 117, 5075.
(19) Harris, R. K. Can. J . Chem. 1964, 42, 2275.
(20) Keiter, R. L.; Rheingold, A. L.; Hamerski, J . J .; Castle, C. K.
Organometallics 1983, 2, 1635.
(24) Riley, P. E.; Davis, R. E. Inorg. Chem. 1980, 19, 159.

Mixed-Phosphine Tricarbonyl Iron Complexes
Organometallics, Vol. 16, No. 11, 1997 2253
shorter than the Fe-P distance in Fe(CO)4PPh3 (2.244-
(1) Å).24 Replacing one PPh3 of trans-Fe(CO)3(PPh3)2
with the more basic PEt3 shortens the Fe-PPh3 distance
even further. Thus, the equal Fe-P distances in trans-
Fe(CO)3(PEt3)(PPh3) may be the result of a significantly
enhanced π-acceptance by PPh3 induced by the very
basic trans PEt3. This is not the first time that
equalization of metal-phosphorus bond lengths of dis-
similar phosphines trans to one another has been
observed: the Cr-P bond distances in trans-Cr(CO)4-
(PBu3)(PPh3) are and 2.344(4) and 2.349(4) Å, respec-
tively.25 An exception to bond length equalization is
found, however, in trans-[Ru(NO2)(terpy)(PMe3)(PPh3)]+,
where the Ru-PMe3 distance (2.347(2) Å) is substan-
tially shorter than the Ru-PPh3 distance (2.458(2) Å).26
Perhaps the difference is associated with the fact that
when phosphines are coordinated to ruthenium in the
+2 oxidation state, they receive only minimal π-dona-
tion from ruthenium.
interactions between the PR3 and CO ligands, cannot
be attributed to weak bonding interactions because
there are no close contacts that support the argument.
Structures of the similar complexes Fe(CO)3(PPh3)2,21
Fe(CO)3(trans-PPh2CHdCHPPh2)2,28 and Ru(CO)3-
29
(PPh3)2 also have an eclipsed/gauche arrangement of
phosphine ligands.
Su m m a r y. In this work we have shown that a good
2
negative correlation exists between J PP coupling con-
stants in Fe(CO)3LL′ complexes and the sum of the pKa
values for the free phosphine ligands. An even better
2
correlation is obtained when J PP is related to the QALE
factors, ø, θ, and Ear, but statistical analysis leads us
to conclude that this model is no more reliable for
estimating phosphorus-phosphorus coupling constants
in trans-Fe(CO)3L2 than the simpler one-parameter pKa
approach. The QALE model is potentially useful for
determining the relative importance of the ø, θ, and Ear
2
contributions to J PP but, because these parameters
The P-C distances (average 1.830(9) Å) of the PPh3
ligand found in Fe(CO)3(PPh3)(PEt3) are virtually un-
changed compared to those in the free ligand (average
1.831(2) Å) and those in Fe(CO)4PPh3 (average 1.831-
(3) Å). Likewise, the average C-P-C bond angles
(102.9(3)°) for PPh3 in our mixed-ligand complex are
identical to those in the free ligand (102.767(924)°) and
only slightly less than those (103.9(5)°) in Fe(CO)4PPh3.
The invariance of P-C distances and C-P-C angles
for free PPh3 compared to transition-metal-coordinated
PPh3 has been well-documented in the comprehensive
study by Orpen’s group.27
The range of Fe-C distances in the mixed-ligand
complex is 1.767(8)-1.773(8) Å (average ) 1.769 Å),
almost identical to that in the bis(triphenylphosphine)
derivative at 1.765(4)-1.776(4) Å (average ) 1.770 Å).
Replacement of PPh3 with the more basic PEt3 would
be expected to add electron density and lead to longer
CO bonds. However, the CO bond lengths in trans-Fe-
(CO)3(PEt3)(PPh3) range from 1.137(8) to 1.153(8) Å
(average 1.147(8) Å) compared to 1.132(5)-1.154(5) Å
(average ) 1.140(5) Å) in Fe(CO)3(PPh3)2,21 a change
that is not significant. Moreover, variation in the CO
stretching frequencies (νCO(hexane) ) 1885 cm-1 for Fe-
(CO)3(PPh3)2; 1886 cm-1 for Fe(CO)3(PEt3)(PPh3)) is
insignificant, further revealing an insensitivity of the
carbonyl groups to the nature of the phosphine.
themselves are strongly correlated for the particular
phosphines chosen for our study, this data set is of
limited value for drawing fundamental conclusions. It
2
is apparent from these studies, however, that J PP is
strongly dependent on the electronic properties of the
phosphine ligands but rather insensitive to their cone
angles.
The molecular structure of trans-Fe(CO)3(PEt3)(PPh3)
has revealed two Fe-P bonds of equal length, implying
equal metal-phosphorus bond strengths for the two
different phosphines. Equalization of bond strengths
may occur because the electron density provided to iron
by PEt3 enhances the π interaction between Fe and
PPh3.
Ackn owledgm en t. We thank the Camille and Henry
Dreyfus Foundation for a Scholar/Fellow grant, the
Research Corporation for support of this work, and the
National Science Foundation (CHE 8851619) which
made possible the purchase of a General Electric QE300
NMR spectrometer. We also thank Warren Giering for
stimulating our interest in the QALE model and J eremy
Wheeler for technical assistance.
Su p p or tin g In for m a tion Ava ila ble: Tables of crystal
data and structure refinement, final atomic positional param-
eters, interatomic distances and angles, anisotropic displace-
ment parameters, and hydrogen coordinates (5 pages). Or-
dering information is given on any current masthead page.
The ipso carbons of the PPh3 group and the methylene
carbons of the PEt3 group are eclipsed, giving an
approximate D3h symmetry. Viewed along the P-Fe-P
axis, these C atoms are in a gauche conformation
relative to the three CO ligands, with torsional angles
of 77° and 43° on average. The deviation from the
staggered (60°/60°) structure, expected based on steric
OM9609224
(28) Luh, L.-S.; Liu, L.-K Inorg. Chim. Acta 1993, 89, 206.
(29) Dahan, F.; Sabo, S.; Chaudret, B. Acta Crystallogr. 1984, C40,
786.
(25) The Cr-P bond distances in trans-Cr(CO)4(PBu3)(PPh3) are
2.349(4) and 2.344(4) Å. Wovkulich, M. J .; Atwood, J . L.; Canada, L.;
Atwood, J . D. Organometallics 1985, 4, 867.
(26) Szczepura, L. F.; Kubow, S. A.; Leising, R. A.; Perez, W. J .;
Huynh, M. H. V.; Lake, C. H.; Churchill, D. G.; Churchill, M. R.;
Takeuchi, K. J . J . Chem. Soc., Dalton Trans. 1996, 14, 2527.
(27) Dunne, B. J .; Morris, R. B.; Orpen, A. G. J . Chem. Soc., Dalton
Trans. 1991, 653.
(30) (a) Battaglia, L. P.; Boselli, T.; Chiusoli, G. P.; Nardelli, M.;
Pelizzi, C.; Predieri, G. Gazz. Chim. Ital. 1985, 115, 395. (b) Darens-
bourg, D. J .; Nelson, H. H., III; Hyde, C. L. Inorg. Chem. 1974, 13,
2135. (c) Chiang, T. H.; Zink, J . I. Inorg. Chem. 1985, 24, 4016. (d)
Reckziegel, A.; Bigorgne, M. J . Organomet. Chem. 1965, 3, 341. (e)
Whitmire, K. H.; Lee, T. R. J . Organomet. Chem. 1985, 282, 95.
(31) Brunet, J .-J .; Commenges, G.; Kindela, F.-B.; Neibecker, D.
Organometallics 1992, 11, 3023.