Electronic properties of furyl substituents at phosphorus and their influence on 31P NMR chemical shifts

Article abstract of DOI:10.1021/ja057085u

The electronic properties of 2-furyl and 3-furyl substituents attached to phosphanes and phosphonium salts were studied by means of IR spectroscopy and experimental and computational ³¹P NMR spectroscopy. The heteroaromatic systems proved to be electron withdrawing with respect to phenyl substituents. However, phosphorus atoms with attached furyl substituents are strongly shielded in NMR. The reason for this phenomenon was studied by solid state ³¹P MAS NMR experiments. The chemical shift tensor was extracted, and the orientation within the molecules was determined. The tensor component σ33, which is effected the most by furyl systems, is oriented perpendicular to the P-C bonds of the substituents. P-furyl bonds are shorter than P-phenyl bonds. We assume therefore a lower ground-state energy of the molecules, because of the electron withdrawing properties of the 2-furyl systems. The σ^para component of the ³¹P NMR magnetic shielding is therefore smaller, which results in an overall increase of the magnetic shielding.

Full text of DOI:10.1021/ja057085u

Published on Web 06/10/2006
Electronic Properties of Furyl Substituents at Phosphorus
and Their Influence on 31P NMR Chemical Shifts
Marco Ackermann,† Aurelia Pascariu,† Thomas Ho¨cher,† Hans-Ullrich Siehl,‡ and
Stefan Berger*,†
Contribution from the Institut fu¨r Analytische Chemie, Fakulta¨t fu¨r Chemie und Mineralogie,
UniVersita¨t Leipzig, Linne´str. 3, D-04103 Leipzig, Germany and Abteilung fu¨r Organische
Chemie I, Fakulta¨t fu¨r Naturwissenschaften, UniVersita¨t Ulm, D-89069 Ulm, Germany
Received October 25, 2005; E-mail: stberger@rz.uni-leipzig.de
Abstract: The electronic properties of 2-furyl and 3-furyl substituents attached to phosphanes and
phosphonium salts were studied by means of IR spectroscopy and experimental and computational 31P
NMR spectroscopy. The heteroaromatic systems proved to be electron withdrawing with respect to phenyl
substituents. However, phosphorus atoms with attached furyl substituents are strongly shielded in NMR.
The reason for this phenomenon was studied by solid state 31P MAS NMR experiments. The chemical
shift tensor was extracted, and the orientation within the molecules was determined. The tensor component
σ33, which is effected the most by furyl systems, is oriented perpendicular to the P-C bonds of the
substituents. P-furyl bonds are shorter than P-phenyl bonds. We assume therefore a lower ground-state
energy of the molecules, because of the electron withdrawing properties of the 2-furyl systems. The σpara
component of the 31P NMR magnetic shielding is therefore smaller, which results in an overall increase of
the magnetic shielding.
Introduction
ties of the 2-furyl rings might be surprising. But the oxygen
within the ring has a strong -I and a +M effect (Figure 1),
which could be responsible for the assumed electron withdrawal
of furyl rings.
For 3-furyl systems one could construct similar models; but
experimental data are scarce in this case. The -I effect should
be smaller since the oxygen phosphorus distance is increased,
whereas the +M effect is still present.
In recent years tris(2-furyl)phosphane (TFP) has become an
important ligand in transition metal catalysis.1 TFP was first
identified as an exceptional well ligand system for Stille type
cross-couplings by Farina et al.2 This discovery triggered many
other investigations concerning the use of TFP in cross-coupling
reactions, e.g., in Negishi cross-couplings,3 Heck reactions,4
Suzuki reactions,5 nickel-catalyzed reactions,6 rhodium-catalyzed
reactions,7 and ruthenium-catalyzed reactions.8 TFP is highly
advantageous especially in Stille type cross-coupling reactions
because often a significant rate acceleration is observed over
traditional PPh3 based catalysts. This observation is attributed
to the different electronic properties of TFP over PPh3. TFP is
believed to be less Lewis basic and therefore represents a poor
σ-donor ligand. In other words, the 2-furyl rings of TFP should
be electron withdrawing with respect to phenyl rings, leaving
less electron density at the phosphorus nucleus. The furyl
substituents of TFP are five-membered aromatic rings with 6
π-electrons and therefore can be regarded as π-excess aromatic
systems. From this point of view, electron withdrawing proper-
However, these ideas are so far only insufficiently verified
since the electronic properties of tertiary phosphanes are difficult
to quantify. For example, the well-known Tolman angles
correlate the steric size of phosphanes with their electronic
properties.9 However the Tolman angle reported for PPh3 (145°)
is only slightly bigger than that for TFP (133°) and therefore
does not explain the significant differences in chemical behavior
as well as their ligand properties. Another approach to quantify-
ing phosphane basicity was reported by Allen and Taylor.10 They
1
measured the J(31P-77Se) coupling constants of phosphorus
selenides. The coupling constant for Ph3PdSe was reported with
732 Hz compared to the TFP based system with 793 Hz. An
increase in this coupling constant is attributed to an increased
s-character of the phosphorus lone pair. This is the only work
reported in the literature, which quantifies the electron with-
drawing character of TFP with respect to PPh3. Many other
methods have been suggested to additionally support these
findings, including IR stretching frequencies of phosphane
transition-metal carbonyl complexes,11 31P NMR chemical shift
† Universita¨t Leipzig.
‡ Universita¨t Ulm.
(1) Andersen, N. G.; Keay, B. A. Chem. ReV. 2001, 101, 997.
(2) Farina, V.; Baker, S. R.; Benigni, D. A.; Sapino, C. Tetrahedron Lett. 1988,
29, 5739.
(3) Klement, I.; Rottla¨nder, M.; Tucker, C. E.; Majid, T. N.; Knochel, P.;
Venegas, P.; Cahiez, G. Tetrahedron 1996, 52, 7201.
(4) Herrmann, W. A.; Brossmer, S.; O¨ fele, K.; Beller, M.; Fischer, H. J. Mol.
Catal. A: Chem. 1995, 103, 133.
(5) Anderson, J. C.; Namli, H.; Roberts, C. A. Tetrahedron 1997, 53, 15123.
(6) Shirakawa, E.; Yamasaki, K.; Hiyama, T. Synthesis 1998, 1544.
(7) Sakai, M.; Hayashi, H.; Miyaura, N. Organometallics 1997, 16, 4229.
(8) Trost, B. M.; Rhee, Y. H. J. Am. Chem. Soc. 1999, 121, 11680.
(9) (a) Tolman, C. A. Chem. ReV. 1977, 77, 313. (b) Tolman, C. A. J. Am.
Chem. Soc. 1970, 92, 2953.
(10) Allen, D. W.; Taylor, B. F. J. Chem. Soc., Dalton Trans. 1982, 51.
9
8434
J. AM. CHEM. SOC. 2006, 128, 8434-8440
10.1021/ja057085u CCC: $33.50 © 2006 American Chemical Society

Influence of Furyl Rings on 31P NMR Chemical Shifts
A R T I C L E S
furyl substituents attached to phosphorus nuclei are believed to
act electron withdrawing, with respect to phenyl rings. The
question arose how the postulated electron withdrawing proper-
ties of the furyl groups should lead to higher shielded phos-
phorus nuclei.
Although furyl systems are believed to act electron withdraw-
ing, there is still a considerable lack of experimental data,
supporting and quantifying this assumption. Additional data can
be obtained from IR investigations.9,11,12 We therefore prepared
carbonyl complexes from the 2-furyl and 3-furyl group bearing
phosphanes 1a to 1g with Ni(CO)4. From the resulting carbonyl
complexes of the general formula PL3Ni(CO)3 (L ) Ph, 2-furyl,
3-furyl) the IR stretching frequencies of the carbonyl groups
were measured. The CO stretching frequency depends on the
ligand pattern at the phosphorus nuclei, since the electronic
properties of the phosphorus substituents are transmitted through
the chemical bonds to the carbonyl groups. The carbonyl
stretching frequencies extracted from the nickel complexes
bearing phosphanes 1a to 1g are presented in Table 2.
Both 2-furyl and 3-furyl substituents increase the CO-IR
stretching frequency although to different extents. For 2-furyl
groups the CO-IR stretching frequency is increased by 3.2 cm-1
per additional 2-furyl ligand, and for 3-furyl groups the average
increase in CO stretching frequency is only 0.9 cm-1. Increased
CO-IR stretching frequencies indicate stronger CO bonds
caused by furyl substituents. σ-Donor ligands, which increase
the electron density at the phosphorus center, weaken the CO
bond, as more electrons can be donated into antibinding
π*-orbitals. The observed increased CO bond strength, however,
indicates less electrons in the antibinding π*-orbitals, thus poor
σ-donor capabilities of the furyl bearing phosphanes, which are
the result of electron withdrawing properties of the 2-furyl
groups. The IR stretching frequency of CO groups in nickel-
carbonylphosphanes can be calculated according to Tolman
using the formula:9b
Figure 1. (a) Mesomeric core structures of 2-furyl bearing phosphanes.
(b) Mesomeric core structures of 3-furyl bearing phosphanes.
correlation,12 1J(31P-195Pt) coupling constants,13 lone pair
ionization potential,14 and gas-phase basicity.15 However, none
of these methods has been applied to TFP or other phosphanes
bearing furyl substituents. This work presents some of the
lacking experimental data, which should help to understand the
electronic properties of TFP. Additionally, the influence of
2-furyl moieties on the 31P NMR chemical shift will be studied
by means of MAS NMR experiments, to extract information
about the chemical shift tensor components. Ab initio calcula-
tions will clarify the orientation of these tensor components
within the molecules, providing an explanation for the observed
31P NMR chemical shift.
Results and Discussion
Recently we were interested in the use of heteroaromatic
substituents at the phosphorus atom of phosphonium salts in
Wittig reactions.16 We prepared (2-furyl)diphenylphosphane 1b,
bis(2-furyl)phenylphosphane 1c, TFP 1d, and the corresponding
phosphanes bearing 3-furyl groups 1e-1g. These phosphanes
were quarternized with ethyliodide yielding the corresponding
phosphonium salts 2a-2g. The phosphonium salts were depro-
tonated using either NaHMDS or n-BuLi, and the resulting
ylides 3 were reacted with benzaldehyde in Wittig reactions.
Figure 2 shows all intermediates during Wittig reactions, which
were observed by 31P NMR spectroscopy.
31P NMR chemical shift data for the intermediates 1a-1g to
5a-5g are presented in Table 1.
The 31P NMR resonances of all intermediates are shifted to
lower frequencies, when phenyl rings are replaced by 2-furyl
or 3-furyl groups. A linear correlation is observed between the
number of furyl rings present in the molecule and the extent of
the chemical shift movement to lower frequencies. Comparing
the influence of 2-furyl and 3-furyl rings on the chemical shift,
we conclude that in the case of phosphonium salts and
phosphane oxides the 2-furyl rings lead to lower overall
chemical shift frequencies (2d vs 2g and 5d vs 5g), whereas in
the case of phosphanes and oxaphosphetanes the opposite is
true (1d vs 1g and 4d vs 4g). Overall, the influence of the furyl
substituents on the 31P NMR chemical shift is most pronounced
in the phosphanes, since the phosphorus atom in these com-
pounds lacks any other substituents, which could lower the
influences of furyl groups.
3
νCO ) 2056.1 + ø [cm-1]
(1)
∑
i
i)1
The value øi represents the contribution of every substituent
at the phosphorus atom to the overall IR stretching frequency
shift. For 2-furyl one can calculate a øi value of 7.4 cm-1. Phenyl
substituents possess a øi value of 4.3 cm-1. 2-Furyl rings are
therefore more electron withdrawing to phosphorus atoms than
phenyl substituents, for example, ethoxy groups (øi ) 6.8 cm-1),
but less so than methoxy substituents with a øi value of 7.7
cm-1. For 3-furyl groups a øi value of 5.1 cm-1 can be
calculated. This is lower than the value of 7.4 cm-1 for 2-furyl
rings but still electron withdrawing with respect to the phenyl
substituent. With respect to Tolmans list of øi values,9b the
3-furyl system is stronger electron withdrawing than p-fluo-
rophenyl substituents with øi ) 5.0 cm-1 and less withdrawing
Lower NMR resonance frequencies as shown in Table 1
correspond to higher shielding. As outlined in the Introduction,
than p-chlorophenyl with øi ) 5.6 cm-1
.
From these IR investigations we conclude that 2-furyl and
3-furyl substituents in phosphanes are indeed electron withdraw-
ing in comparison to phenyl rings in triphenylphosphane.
Although the electron density on the phosphorus nuclei is
reduced, when furyl systems are attached, the phosphorus atoms
are shielded. This means that furyl systems cause higher shielded
phosphorus nuclei, although less electron density is present. To
(11) Strohmeier, W.; Mu¨ller, F. J. Chem. Ber. 1974, 96, 53.
(12) Derencse´nyi, T. T. Inorg. Chem. 1981, 20, 665.
(13) Cobley, C. J.; Pringle, P. G. Inorg. Chim. Acta 1997, 265, 107.
(14) Berners-Price, S. J.; Sadler, P. J. Structure and Bonding; Springer-Verlag:
Berlin, Heidelberg, 1998; Vol. 70, pp 33-35.
(15) Ikuta, S.; Kebarle, P.; Bancroft, G. M.; Chan, T.; Puddephatt, R. J. J. Am.
Chem. Soc. 1982, 104, 5899.
(16) (a) Schro¨der, U.; Berger, S. Eur. J. Org. Chem. 2000, 2601. (b) Appel,
M.; Berger, S. Eur. J. Org. Chem. 2002, 1143. (c) Ackermann, M.; Berger,
S. Tetrahedron 2005, 61, 6764.
9
J. AM. CHEM. SOC. VOL. 128, NO. 26, 2006 8435

A R T I C L E S
Ackermann et al.
Figure 2. Starting materials, intermediates, and reaction products observed by 31P NMR spectroscopy during Wittig reactions (a, triphenyl; b-d, series of
2-furyl bearing compounds; e-g, series of 3-furyl systems).
Table 1. 31P NMR Chemical Shift of Intermediates Detected during Wittig Reactions
δ
31P [ppm]
δ
31P [ppm]
δ
31P [ppm]
ylideb
3
δ
31P [ppm]
δ
31P [ppm]
substituents
phosphanea
phosphonium salta
oxaphosphetaneb,c
phosphane oxideb
at phosphorus
1
2
4
5
Ph3 a
-6.0
-26.3
-50.3
-76.8
-31.5
-58.0
-82.3
26.9
16.3
4.4
-9.6
17.2
7.7
14.8
2.6
-60.7d
24.7
16.4
2.3
-12.9
20.8
12.7
4.1
(2-furyl)Ph2 b
(2-furyl)2Ph c
(2-furyl)3 d
(3-furyl)Ph2 e
(3-furyl)2Ph f
(3-furyl)3 g
-73.7 (-73.8)
-87.0 (-88.9)
-101.4 (-102.0)
-75.7 (-75.8)
-91.1 (-91.2)
-103.2 (-103.6)
-12.1
-30.9
-2.6
-15.7
-31.2
-1.8
a
b
c
31P NMR chemical shift in CDCl3. 31P NMR chemical shift in [d8]THF. 31P NMR chemical shift for the cis-oxaphosphetane. In parentheses the 31
P
NMR chemical shift of the trans-isomer. d Isomers not resolved.
were the same as those with 2-furyl groups. The 31P MAS NMR
spectra were recorded on a 600 MHz NMR spectrometer (243
MHz for 31P) at room temperature with spinning frequencies
of 4000, 3200, 2300, 1400, and 700 Hz. The measurement time
varied between 30 min and 5 h, depending on the spinning
frequency and the number of spinning sidebands which had to
be resolved. As an example, the recorded 31P MAS NMR spectra
of ethyltris(2-furyl)phosphonium iodide 2d at spinning frequen-
cies of 4000 and 700 Hz are shown in Figure 3.
Table 2. IR Stretching Frequencies of Nickelcarbonylphosphane
Complexes of the General Formula P(2- or 3-furyl)nPh3-nNi(CO)3
1
number n of 2-furyl substituents
CO stretching frequency [cm-
]
n ) 0 (prepared from 1a)
n ) 1 (prepared from 1b)
n ) 2 (prepared from 1c)
n ) 3 (prepared from 1d)
2068.9
2071.9
2075.4
2078.4
1
number n of 3-furyl substituents
CO stretching frequency [cm-
]
n ) 1 (prepared from 1e)
n ) 2 (prepared from 1f)
n ) 3 (prepared from 1g)
2069.8
2070.5
2071.5
(17) (a) Dusold, S.; Maisel, H.; Sebald, A. J. Magn. Reson. 1999, 141, 78. (b)
Ashbrook, S. E.; Wimperis, S. High-resolution NMR of quadrupolar nuclei
in solids: The satellite-transition magic angle spinning (STMAS) experi-
ment. Prog. Nucl. Magn. Reson. Spectrosc. 2004, 45, 53. (c) Kohler, S. J.;
Ellett, D. J., Jr.; Klein, M. P. J. Chem. Phys. 1976, 64, 4451. (d) Kohler,
S. J.; Klein, M. P. Biochemistry 1976, 15, 967. (e) Kohler, S. J.; Klein, M.
P. J. Am. Chem. Soc. 1977, 99, 8290. (f) Herzfeld, G.; Griffin, R. G.;
Haberkorn, R. A. Biochemistry 1978, 17, 2711. (g) Tutunjian, P. N.; Waugh,
J. S. J. Chem. Phys. 1982, 76, 1223. (h) Tutunjian, P. N.; Tropp, J.; Waugh,
J. S. J. Am. Chem. Soc. 1983, 105, 4848. (i) Naito, A.; Sastry, D. L.;
McDowell, C. A. Chem. Phys. Lett. 1985, 115, 19. (j) Hauser, H.; Radloff,
C.; Ernst, R. R.; Sundell, S.; Pascher, I. J. Am. Chem. Soc. 1988, 110,
1054. (k) McDowell, C. A.; Naito, A.; Sastry, D. L.; Takegoshi, K. J. Magn.
Reson. 1988, 78, 498. (l) Ermark, F.; Topic, B.; Haeberlen, U.; Blinc, R.
J. Phys.: Condens. Matter 1989, 1, 5489. (m) Eichele, K.; Wasylishen, R.
E. J. Phys. Chem. 1994, 98, 3108. (n) Eichele K.; Wu, G.; Wasylishen, R.
E.; Britten, J. F. J. Phys. Chem. 1995, 99, 1030. (o) Lumsden, M. D.;
Wasylishen, R. E.; Britten, J. F. J. Phys. Chem. 1995, 99, 16602. (p) Eichele
K.; Wasylishen, R. E.; Maitra, K.; Nelson, J. H.; Britten, J. F. Inorg. Chem.
1997, 36, 3539. (q) Jensen, T. R.; Hazell, R. G.; Vosegaard, T.; Jakobsen,
H. J. Inorg. Chem. 2000, 39, 2026. (r) Gee, M.; Wasylishen, R. E.; Eichele,
K. J. Phys. Chem. A 2000, 104, 4598. (s) Grossmann, G.; Scheller, D.;
Malkina, O. L.; Malkin, V. G.; Zahn, G.; Schmidt, H.; Haeberlen, U. Solid
State Nucl. Magn. Reson. 2000, 17, 22.
(18) (a) Dusold, S.; Klaus, E.; Sebald, A.; Bak, M.; Nielsen, N. C. J. Am. Chem.
Soc. 1997, 119, 7121. (b) Dusold, S.; Milius, W.; Sebald, A. J. Magn.
Reson. 1998, 135, 500. (c) Dusold, S.; Maisel, H.; Sebald, A. J. Magn.
Reson. 1999, 141, 78. (d) Bechmann, M.; Dusold, S.; Forster, H.; Haeberlen,
U.; Lis, T.; Sebald, A.; Stumber, M. Mol. Phys. 2000, 98, 605. (e) Eichele,
K.; Wasylishen, R. E.; Corrigan, J. F.; Taylor, N. J.; Carty, A. J.; Feindel,
K. W.; Benard, G. M. J. Am. Chem. Soc. 2002, 124, 1541. (f) Power, W.
P.; Wasylishen, R. E. Inorg. Chem. 1992, 31, 2176. (g) Eichele, K.;
Ossenkamp, G. C.; Wasylishen, R. E.; Cameron, T. S. Inorg. Chem. 1999,
38, 639. (h) Bernard, G. M.; Wu, G.; Lumsden, M. D.; Wasylishen, R. E.;
Maigrot, N.; Charrier, C.; Mathey, F. J. Phys. Chem. A 1999, 103, 1029.
(i) Bechmann, M.; Dusold, S.; Geipel, F.; Sebald, A.; Sellmann, D. J. Phys.
Chem. A 2005, 109 (24), 5275.
solve this apparent contradiction we initiated solid phase MAS
NMR studies to have a closer look of the 31P NMR chemical
shift tensors.
Solid Phase 31P NMR Investigations
Generally, the shielding of a NMR active nucleus from the
outer magnetic field B0 is generated by electron density around
that nucleus. The question arose how electron withdrawing
substituents, which decrease the electron density around the
phosphorus nucleus, also increase the shielding. To understand
the direction in which the 31P NMR chemical shift is moved by
different substituents, information on the spatial components
of the 31P NMR chemical shift is necessary. Either 31P NMR
investigations on oriented single crystals17 or 31P MAS NMR
on polycrystalline powders18 are suitable to obtain such data.
We performed MAS NMR studies on the polycrystalline
phosphonium salts 2a, 2b, 2c, and 2d. Phosphonium salts were
chosen for these investigations, since the phosphorus atom lacks
the free electron pair. Free electron pairs on phosphorus are
known to have multiple influences on the chemical shift, which
are difficult to predict and to quantify. 3-Furyl bearing phos-
phonium salts were not implemented in the 31P MAS NMR19
investigation, since the overall effects on the chemical shifts
(19) (a) Andrew, E. R.; Bradbury, A.; Eades, R. G. Nature 1958, 182, 1659.
(b) Lowe, I. J. Phys. ReV. Lett. 1959, 2, 285.
9
8436 J. AM. CHEM. SOC. VOL. 128, NO. 26, 2006

Influence of Furyl Rings on 31P NMR Chemical Shifts
A R T I C L E S
Figure 3. 31P MAS NMR spectra of ethyltris(2-furyl)phosphonium iodide 2d at MAS frequencies of 4000 Hz (upper picture) and 700 Hz (lower picture).
Table 3. Experimentally Determined Main Axis Chemical Shift
Tensor Elements δ11, δ22, δ33 for the Four Investigated
Phosphonium Salts 2a to 2d Together with the Overall Change of
the 31P NMR Chemical Shift ∆δ
phosphonium salt
δiso [ppm]
δ11 [ppm]
δ22 [ppm]
δ
33 [ppm]
EtP+Ph3 I- 2a
EtP+Ph2Fu I- 2b
EtP+PhFu2 I- 2c
EtP+Fu3 I- 2d
∆δ
26.9
16.3
4.4
-9.6
-36.5
42.7
39.2
34.2
24.8
30.9
14.7
14.7
-5.0
7.1
-5.0
-35.7
-48.6
-55.7
-17.9
-35.9
It was possible to extract the main axis tensor components
of the chemical shift vector δ11, δ22, and δ33 from the observed
spinning sidebands by a moment analysis known from Maricq
and Waugh.20 The results of these calculations are shown in
Table 3.
To visualize the data, these chemical shifts are drawn in
Figure 4 versus the number of 2-furyl groups in the phospho-
nium salts.
The dotted line in Figure 4 represents the values of the
isotropic 31P NMR chemical shift in solution. As discussed
above, the isotropic chemical shift for the phosphonium salts
in solution is shifted by a total of -36.5 ppm to lower
frequencies, if the three phenyl groups in 2a are replaced by
2-furyl systems leading to 2d. The three main axis chemical
shift tensor elements δ11, δ22, and δ33 generally show the same
trend to lower resonance frequencies, although to different
extents. δ11 shows the smallest dependence on the 2-furyl
systems and is only shifted by a total of -17.9 ppm to lower
frequencies. The strongest influence is visible for δ33, which is
shifted by a total of -55.7 ppm to lower frequencies if we
Figure 4. Chemical shift in solution (dotted line) together with the shift
of the three main axis components vs the number of 2-furyl groups in the
phosphonium salts 2a to 2d.
compare 2a with 2d. We conclude that the chemical shift
component δ33 provides the main contribution for the significant
shift of δiso to lower resonance frequencies. To understand the
unexpected drift of δiso in solution, we need information on the
orientation of δ33 within the molecular structure of the phos-
phonium salts. This question could have been answered if the
31P NMR investigation would have been performed on oriented
single crystals. However such information is not accessible with
the 31P MAS NMR approach on polycrystalline samples used
here. We therefore performed accompanying ab initio quantum
chemical calculations of chemical shieldings to clarify why δ33
is effected the most in the presence of 2-furyl groups.
(20) Maricq, M. M.; Waugh, J. S. J. Chem. Phys. 1979, 70, 3300.
9
J. AM. CHEM. SOC. VOL. 128, NO. 26, 2006 8437

A R T I C L E S
Ackermann et al.
Table 5. Calculated Absolute Shieldings σ(S)g for the Phosphorus
Atoms in the Four Phosphonium Cations of 2a-2d and for the
Chloride Salt of 2d, with the Corresponding Calculated Relative
31P NMR Chemical Shifts δ(S)g in the Gas Phase and the
Experimental 31P NMR Chemical Shifts for the Iodide Salts of
a
2a-2d in the Liquid Phase in CDCl3
basis set
6-311G(d,p)
TZ2P
G3large
expmt
EtP+Ph3 2a
σ(S)g [ppm]
δ(S)g [ppm]
σ(S)g [ppm]
δ(S)g [ppm]
σ(S)g [ppm]
δ(S)g [ppm]
σ(S)g [ppm]
δ(S)g [ppm]
σ(S)g [ppm]
δ(S)g [ppm]
287.6
40.8
303.5
24.9
320.5
7.8
332.5
-4.1
329.4
-1.1
299.3
29.0
314.3
14.1
331.2
-2.9
341.1
-12.7
337.4
-9.1
265.9
62.4
280.8
47.6
294.8
33.6
309.1
19.3
26.9
16.3
4.4
EtP+Ph2Fu 2b
EtP+PhFu2 2c
EtP+Fu3 2d
-9.6
EtP+Fu3 Cl-
306.1
22.2
a σ(S)g values calculated with 6-311G(d,p), (TZ2P),22c and [G3large]22d
basis sets, respectively
Calculations of magnetic shieldings were performed using the
Gaussian program suite.22 We used the B3LYP hybrid density functional
method to describe the N-electron space.22b The one-electron space was
explored using different basis set models, the standard 6-311G(d,p)
basis set, a TZ2P22c basis set, and the G3large basis set used in G3
theory. The relative calculation times on dual CPU Opteron 248
workstations were 1: 5: 28.
The atomic coordinates from the crystal structures were used to
generate the geometries for the input of the NMR calculations. We
calculated the magnetic shieldings for the isolated cation structures 2a-
2d without the iodide anion. Comparative calculations performed for
the salt 2d where the iodide anion was substituted by a chloride anion
revealed that the anion exhibits negligible influence on absolute isotropic
shielding of the phosphorus atom (around -3 ppm, see Table 5). The
absolute shieldings were transformed into chemical shifts referenced
to aqueous H3PO4 using the equation:23
Figure 5. Crystal structure of ethyltris(2-furyl)phosphonium iodide 2d.
Table 4. Bond Lengths of the Phosphorus-Carbon Bonds in the
Crystal Structures. the Average Bond Length Is Given in Cases, in
Which More than One Bond of This Type Was Present
P
−C(ethyl)
P
−C(phenyl)
P−C(2-furyl)
phosphonium salt
bond [pm]
bond [pm]
bond [pm]
EtP+Ph3 I- 2a
EtP+Ph2Fu I- 2b
EtP+PhFu2 I- 2c
EtP+Fu3 I- 2d
181.0
179.4
179.6
178.8
179.9
178.9
179.3
176.5
176.3
176.6
Theoretical Calculations
For theoretical calculations of the 31P NMR chemical shift tensors
we needed the structures of the phosphonium salts in the solid phase.
From those structures we should be able to isolate the atomic
coordinates, which are required for the ab initio calculations. For this
purpose we initiated X-ray crystallographic studies to record the solid
phase structures. From the four phosphonium salts under investigation
only the ethyltriphenylphosphonium cation X-ray structure is known.
For our studies we used the X-ray structure of ethyltriphenylphospho-
nium triiodide,21 since the difference in the cation structure whether
iodide or triiodide is present as anion should be small. This compound
crystallizes in a monoclinic crystal system with the space group P2-
(1)/n. The crystal structures of 2-furyl bearing phosphonium salts 2b
to 2d have not been published in the literature and therefore had to be
determined. The crystals were grown in a solution of ethanol with ethyl
acetate at a temperature of -27 °C. In Figure 5 the structure of ethyltris-
(2-furyl)phosphonium iodide 2d is shown. The crystal structures of 2b
and 2c are given in the Supporting Information.
This molecule crystallizes in a monoclinic crystal system with the
space group P2(1)/c. If one compares the bond lengths of the
phosphorus-carbon(substituent) bonds in the crystal structures of the
cations in 2a to 2d an interesting trend is revealed. The average bond
lengths in the investigated crystal structures are shown in Table 4.
The bond length of the phosphorus-carbon(ethyl) bond decreases
as the number of 2-furyl groups present in the molecule increases.
Furthermore, the P-C(substituent) bonds become longer by an average
of 2.7 pm, when 2-furyl groups instead of phenyl rings are attached to
the phosphorus. This is an indication, that under the presence of 2-furyl
groups the orbital overlap is better, leading to shorter bonds and
presumably more stable molecules.
δ(S) ) σ(H3PO4) - σ(S) with σ(H3PO4) ) 328.35 ppm (2)
Table 5 shows the calculated absolute isotropic shieldings and the
relative chemical shifts according to eq 2 for the isolated phosphonium
cations 2a-2d without anion and for the salt 2d with chloride as anion.
The calculated absolute shieldings are highly dependent on the basis
set employed (Table 5). The gas-phase isotropic 31P NMR chemical
shifts δ(S)g calculated using b3lyp/6-311g(d,p) deviate by an average
of 7.8 ppm from the experimental chemical shift in solution. The data
obtained with the TZ2P basis set are in reasonable agreement and
deviate only by 3.7 ppm in the average, whereas the data from the
G3large basis set are off by 31.2 ppm in the average. The differences
are in part due to the model used to describe the one electron space.
Basis sets are optimized for modeling valence electrons better than
core electrons, whereas NMR experiments also probe the electron
density closer to the nuclei. Discrepancies between experimental
chemical shifts in solution and the calculated data are also expected
(22) (a) Gaussian 03, Revision D.01/D02, Frisch, M. J. et al. Gaussian, Inc,
Wallingford CT, 2004. For full citation, see Supporting Information. (b)
Becke, A. D. J. Chem. Phys. 1993, 98, 5648. (c) The TZ2P basis set was
constructed from the Dunning triple-ú set (Dunning, T. H., Jr. J. Chem.
Phys. 1971, 55, 716. available from http://www.emsl.pnl.gov/forms/
basisform.html. For phosphorus and chlorine the 6s5p contraction of the
Huzinaga 12s9p scheme was used as reported in McLean, A. D.; Chandler,
G. S. J. Chem. Phys. 1980, 72 (10), 5639. The standard pairs of polarization
function were added by C. Janssen April 5, 1991 as suggested by Brian
Yates, see: http://zopyros.ccqc.uga.edu/psi/Psiman/basis_set_info.txt. (d)
The G3large basis set corresponds to the G3large basis set used in the
MP2(full) calculation in G3 theory (Curtiss, L. A.; Raghavachari, K.;
Redfern, P. C.; Rassolov, V.; Pople, J. A. J. Chem. Phys. 1998, 109, 7764).
See: http://chemistry.anl.gov/compmat/G3basis/hkrg3l.txt
(23) Jameson, C. J.; de Dios, A.; Jameson, A. K. Chem. Phys. Lett. 1990, 167,
575.
(21) Tebbe, K.-F.; Farida, T. Z. Naturforsch. 1995, 50b, 1685.
9
8438 J. AM. CHEM. SOC. VOL. 128, NO. 26, 2006

Influence of Furyl Rings on 31P NMR Chemical Shifts
A R T I C L E S
Table 6. Calculated (B3LYP/TZ2P) Main Axis Components of the
Chemical Shift Tensor in the Gas Phase δ11,g, δ22,g, δ33,g Together
with the Experimental Values δ11,s to δ33,s Determined by 31P MAS
NMR from Table 3
δ11,g
δ11,s
δ22,g
δ22,s
δ33,g
δ33,s
[ppm]
[ppm]
[ppm]
[ppm]
[ppm]
[ppm]
EtP+Ph3 2a
38.6
37.0
26.6
18.4
24.1
42.7
39.2
34.2
24.8
30.7
9.3
7.2
-10.5
-4.7
30.9
14.7
14.7
-5.0
17.7
-4.1
-42.5
-46.0
-46.6
7.1
EtP+Ph2Fu 2b
EtP+PhFu2 2c
EtP+Fu3 2d
-5.0
-35.7
-48.6
EtP+Fu3 Cl-
because solute-solvent interactions are not considered in the model
calculations. Despite these shortcomings of the calculation model, it is
of utmost importance for the context of these investigations that the
calculations with all three basis sets are very consistent by predicting
10 to 16 ppm higher shielding for each added furyl ring. The
experimental result that the resonance frequencies in solution are shifted
to lower frequencies when phenyl substituents are replaced by 2-furyl
groups is confirmed by the calculated gas-phase chemical shifts.
We have further calculated the main axis components of the 31P
NMR chemical shift in the gas phase. Table 6 shows a comparison of
the calculated values (B3LYP/TZ2P) of the main axis components of
the 31P NMR chemical shift tensor with the experimentally determined
values. The individual tensor values obtained for the 6-311G(d,p) and
G3large basis sets are very similar and show consistency. They are
given in the Supporting Information.
The similarities between the experimental results and the quantum
chemical calculations are striking. The three main axis components of
the 31P NMR chemical shift are shifted to lower resonance frequencies
under the influence of 2-furyl substituents. Consistent with the
experimental results the calculated tensor component δ33 shows the
largest effect. The shielding difference of δ33 going from compound
2a to 2d is 63.6 ppm according to the calculations and 55.7 ppm from
experimental results. Of highest interest is the orientation of this
component within the molecular geometry. This information is available
from the eigenvector also calculated by Gaussian.
Figure 6. Orientation of the main axis shielding components within the
crystal structure of ethyltris(2-furyl)phosphonium cation as calculated with
the tz2p basis set. 2d. Bottom: View in direction of the negative z-axis of
the molecule coordinate system. Top: View in direction of the negative
x-axis.
shorter bond length can be attributed to better orbital overlap and finally
more stable molecules, with a lower ground state energy. A lower
ground state energy leads generally to higher activation energies ∆E.
According to the Pople equation:24
The calculated orientation of the main axis shielding tensor
components within the molecule is shown in Figure 6 for the ethyltris-
(2-furyl)phosphonium cation 2d. Results for the compounds 2a to 2c
are given in the Supporting Information.
µ0µ2B
σ
) -
r-3 [Qi +
Qj]
(3)
para
iso
∑
2π∆E
i*j
The main attention should be drawn to the component σ33, the
shielding value of the chemical shift component δ33, which was
identified above as the most sensitive tensor component. σ33 is almost
parallel to the z-axis of the molecule coordinate system, with an angle
between the z-axis and σ33 of 7.2°. The angular difference between the
direction of the P-C(ethyl) bond and the component σ33 is even lower
(5.2°). Comparable orientations for the component σ33 and the
P-C(ethyl) bond could be observed in every compound investigated.
The angles between these two vectors were 39.3° in the cation of 2a,
37.8° in 2b, and 23.9° in 2c as calculated with the tz2p basis set. The
angles observed with the other employed basis sets are very similar.
These results verify that the most sensitive shielding component σ33 is
oriented almost parallel to the P-C(ethyl) bond and therefore at the
same time almost perpendicular to the P-C(substituent) bonds in the
four phosphonium salt structures. As a consequence σ33 is most sensitive
concerning changes in the electron density of the P-C(substituent)
bonds.
At this point, the main axis shielding component σ33 has been
identified as the most sensitive shielding component, and it could be
rationalized that this sensitivity has its origin in the orientation of that
component σ33 within the molecule. Still unclear is why are all
components shifted to lower resonance frequencies, meaning that the
phosphorus nucleus gets better shielded from the outer magnetic field.
This question might be answered when taking into account that the
P-C(phenyl) bonds are longer than P-C(2-furyl) bonds (Table 4). A
∆E is inversely proportional to σpara, which is one component of the
overall magnetic shielding of any nucleus given by eq 4:
dia
iso
para
iso
σiso ) σ + σ
+
σj
(4)
∑
i*j
The component σiso arises from core near s-electrons and always
increases the overall shielding, whereas σpara has its origin in the
p-electron density and always acts as a deshielding influence. According
to the Pople equation, a higher activation energy ∆E results in a
decreased σpara term. Due to the deshielding influence of the σpara term,
the final concequence is an increase in the overall shielding of the
nucleus of interest. This explains why the exchange of phenyl ligands
by more electron withdrawing 2-furyl groups leads to lower 31P NMR
resonance frequencies. Electron density is withdrawn from the p-
electrons, the influence of the deshielding σpara term is reduced, leading
to a higher overall shielding of the phosphorus nuclei, which results in
a significant shift of all 31P NMR resonances to lower frequencies under
the presence of 2-furyl groups.
Conclusion
In this work the electronic influences of 2- and 3-furyl
substituents at phosphorus atoms on the 31P NMR chemical shift
(24) Karplus, M.; Pople, J. A. J. Chem. Phys. 1963, 38, 2803.
9
J. AM. CHEM. SOC. VOL. 128, NO. 26, 2006 8439

A R T I C L E S
Ackermann et al.
were studied. An extensive shift of 31P NMR resonances to
lower frequencies is observed, when phenyl groups attached to
phosphorus are replaced by 2-furyl or 3-furyl rings indicating
an increased overall shielding of phosphorus nuclei under these
conditions. IR investigations confirmed that 2- and 3-furyl
substituents act electron withdrawing on phosphorus nuclei with
respect to standard phenyl rings. The question arose why the
shielding of phosphorus atoms under the presence of furyl
systems is increased, although the electron density is reduced
by these furyl systems. The principal chemical shift tensor
components were extracted from solid state 31P MAS NMR
spectra, showing that all three components are shifted to lower
NMR frequencies, but the component δ33 to the largest extent.
The orientation of this shift component within the molecules
was calculated using ab initio methods. δ33 stands almost
perpendicular to the P-C(phenyl or furyl) bonds in the
investigated molecules, proving the reason for the observed shift
of NMR resonances has its origin in changed electronic
properties of the P-C(furyl) bonds. Furthermore, X-ray crystal-
lography studies showed that P-C(furyl) bonds are significantly
shorter than the bonds to phenyl groups. This indicates better
orbital overlap, leading to lower ground-state energies of
molecules containing 2-furyl groups. These facts result in an
increased activation energy of the phosphonium salts, which
lowers the deshielding contribution of the σpara term to the
overall magnetic shielding. We therefore observe an increased
overall magnetic shielding of the 31P nuclei under the presence
of 2-furyl groups, bringing this trend in harmony with the
electron withdrawing properties of the furyl substituents.
nickelcarbonylphosphane complexes were prepared by shaking
a 0.2 M Ni(CO)4 solution in CH2Cl2 with a 0.2 M phosphane
solution in CH2Cl2 at room temperature. The solution was
directly measured as film between KBr plates.
X-ray Crystallography: Suitable crystals for crystal structure
analysis were obtained by carefully recrystallizing the solids in
a mixture of ethanol and ethyl acetate. The crystals were grown
for several days at a temperature of -27 °C. A CCD X-ray
diffractometer from BRUKER was used to determine the crystal
structures. The structures were recorded at 220 K. Ethyl(2-furyl)-
diphenylphosphonium iodide (C18H18IOP, 408,2 g/mol) grew
in a triclinic unit cell with the space group P1h. Ethylbis(2-furyl)-
phenylphosphonium iodide (C16H16IO2P, 398,18 g/mol) had an
orthorhombic unit cell with the space group Pca2 (1). Ethyltris-
(2-furyl)phosphonium iodide (C14H14IO3P, 388.14 g/mol) grew
in a monoclinic crystal system with the space group P2(1)/c.
Quantum chemical calculations were performed at Ulm
university using the Gaussian program suite22 on a SUSE
LINUX AMD Athlon and Opteron cluster.
Acknowledgment. We thank the Verband der Chemischen
Industrie for a Chemiefonds-Stipendium Grant Number 168291
(University of Leipzig) and for financial support for computer
equipment (University of Ulm). Furthermore we thank Prof. Dr.
W. Petz (Marburg) for the generous support with Ni(CO)4. We
thank Prof. Dr. H. Krautscheid (Leipzig) and Dr. K. Eichele
(Tu¨bingen) for very helpful discussions.
Supporting Information Available: Detailed moment analy-
sis; crystal structures of ethyl(2-furyl)diphenylphosphonium
iodide 2b and ethylbis(2-furyl)phenylphosphonium iodide 2c;
cif files for the crystal structures of 2b, 2c, and 2d; graphical
representations for the main axis shielding components and their
orientation within the molecules 2a, 2b, and 2c; chemical shift
tensor values obtained for the 6-311G(d,p) and G3large basis
sets. This material is available free of charge via the Internet at
http://pubs.acs.org.
Experimental Section
NMR Investigations: 31P NMR spectra in liquids and all
spectra measured at low temperatures were recorded on a
BRUKER DRX-400 spectrometer in either CDCl3 or [d8]THF
as solvent. The solid phase 31P MAS NMR spectra were
recorded on a BRUKER DRX-600 with a CP/MAS NMR Probe
DOTY DSI 750.
IR Investigations: IR spectra were recorded on an Avatar
360 FT-IR ESR spectrometer from THERMO NICOLET. The
JA057085U
9
8440 J. AM. CHEM. SOC. VOL. 128, NO. 26, 2006