Synthesis, structure, strain energy, and excess strain of a phospha[3]triangulane

Article abstract of DOI:10.1021/ja9905896

(7-Phenyl-7-phosphadispiro[2.0.2.1]heptane)pentacarbonyltungsten (8), a phospha[3]triangulane, was synthesized from bicyclopropylidene. Its single-crystal X-ray structure determination is reported. Comparison of the crystal structure data with those of the related phosphaspiropentane 7 and phosphirane 6 complexes suggests that the phosphirane ring tightens when the number of spiro atoms is increased. This is supported by the B3LYP and MP2/6-31G* computed geometries of the uncomplexed parent systems. Ab initio calculated heats of formation and strain energies (SE) are reported for the parent phosphirane 11, phosphaspiropentane 12, and phospha[3]triangulane 13 using both G2MP2 theory and ring separation reactions. Our best estimates for the ΔH_f of 11, 12, and 13 are 18.3, 48.4, and 78.2 kcal/mol, respectively, with corresponding SE values of 21.3, 54.7, and 87.9 kcal/mol. For comparison, the slightly modified G2MP2 method was also applied to cyclopropane 1, spiropentane 2, and [3]triangulane 3 to give respective ΔH_f values of 12.6, 44.3, and 75.3 kcal/mol, with corresponding SEs of 28.0, 64.6, and 100.5 kcal/mol, all of which are in excellent agreement with reported experimental data. These strain energies suggest that the excess strain per spiro atom is 5.3 kcal/mol for phospha[n]triangulanes, which is smaller than the 8.6 kcal/mol determined from the heat of combustion measurements for the [n]triangulanes.

Full text of DOI:10.1021/ja9905896

11650
J. Am. Chem. Soc. 1999, 121, 11650-11655
Synthesis, Structure, Strain Energy, and Excess Strain of a
Phospha[3]triangulane
Koop Lammertsma,*,†,‡ Bing Wang,† Jui-Te Hung,† Andreas W. Ehlers,‡ and
Gary M. Gray†
Contribution from the Faculty of Chemistry, Vrije UniVersiteit, De Boelelaan 1083,
1081 HV Amsterdam, The Netherlands, and Department of Chemistry, UniVersity of Alabama at
Birmingham, UAB Station, Birmingham, Alabama 35294-1240
ReceiVed February 24, 1999. ReVised Manuscript ReceiVed September 23, 1999
Abstract: (7-Phenyl-7-phosphadispiro[2.0.2.1]heptane)pentacarbonyltungsten (8), a phospha[3]triangulane, was
synthesized from bicyclopropylidene. Its single-crystal X-ray structure determination is reported. Comparison
of the crystal structure data with those of the related phosphaspiropentane 7 and phosphirane 6 complexes
suggests that the phosphirane ring tightens when the number of spiro atoms is increased. This is supported by
the B3LYP and MP2/6-31G* computed geometries of the uncomplexed parent systems. Ab initio calculated
heats of formation and strain energies (SE) are reported for the parent phosphirane 11, phosphaspiropentane
12, and phospha[3]triangulane 13 using both G2MP2 theory and ring separation reactions. Our best estimates
for the ∆Hf of 11, 12, and 13 are 18.3, 48.4, and 78.2 kcal/mol, respectively, with corresponding SE values
of 21.3, 54.7, and 87.9 kcal/mol. For comparison, the slightly modified G2MP2 method was also applied to
cyclopropane 1, spiropentane 2, and [3]triangulane 3 to give respective ∆Hf values of 12.6, 44.3, and 75.3
kcal/mol, with corresponding SEs of 28.0, 64.6, and 100.5 kcal/mol, all of which are in excellent agreement
with reported experimental data. These strain energies suggest that the excess strain per spiro atom is 5.3
kcal/mol for phospha[n]triangulanes, which is smaller than the 8.6 kcal/mol determined from the heat of
combustion measurements for the [n]triangulanes.
Three-membered ring structures are of broad interest and have
oxaspiropentanes,4 dioxaspiropentanes,5 and azaspiropentanes6
were reported some three decades ago. Also, dispirocyclohep-
been subject to intense study.1 Spiro compounds that connect
such rings by joint carbons are even more intriguing, but much
less is known about them. Several groups have excelled in the
studies on hydrocarbons and have reported an array of exotic
polyspiro condensed rings, such as the linear [n]triangulanes
and the cyclic [n]rotanes, where n indicates the number of
cyclopropane (1) rings.2 Hetero[n]triangulanes have received
only modest attention. Whereas spiropentane, the parent [2]-
triangulane 2,3 has been known for nearly a century, the
tane, the parent [3]triangulane 3,7 was synthesized in that period,
while the substituted oxa-8 and sila[3]triangulanes9 are of more
recent vintage. Building on the recently reported di- and
tetraphosphaspiropentanes 410 and 511 and our earlier work on
* Address correspondence to this author at the Vrije Universiteit.
‡ Vrije Universiteit.
(5) (a) Crandall, J. K.; Batal, D. J. J. Org. Chem. 1988, 53, 1340. (b)
Crandall, J. K.; Machleder, W. H.; Thomas, M. J. J. Am. Chem. Soc. 1968,
90, 7346.
† University of Alabama at Birmingham.
(1) See, for example: (a) Greenberg, A.; Liebman, J. F. Strained Organic
Molecules; Academic Press: New York, 1978. (b) Wiberg, K. B. Angew.
Chem., Int. Ed. Engl. 1986, 25, 317. (c) The Chemistry of the Cyclopropyl
Group; Rappoport, Z., Ed.; Wiley: Chichester, 1987; two volumes. (d)
Strained Organic Compounds. Chem. ReV. 1989, 89 (5), special issue. (e)
Pihlaja, K.; Taskinen, E. Physical Methods in Heterocyclic Chemistry;
Academic Press: New York, 1978; Vol. 6, p 199. (f) Mathey, F. Chem.
ReV. 1990, 90, 997. (g) Houben-Weyl; de Meijere, A., Ed.; Thieme:
Stuttgart, 1997; Vol. E, 17a-c.
(2) For a recent review, see: (a) de Meijere, A.; Kozhushkov, S. In
AdVances in Strain in Organic Chemistry; Halton, B., Ed.; JAI Press:
Greenwich, CT, 1995; Vol. 4, pp 225-282. See also: (b) de Meijere, A.;
Kozhushkov, S. I.; Khlebnikov, A. F. Zh. Org. Khim. 1996, 32, 1607; Russ.
J. Org. Chem. (Engl. Transl.) 1996, 32, 1555. (c) Zefirov, N. S.; Kuznetsova,
T. S.; Eremenko, O. V.; Kokoreva, O. V.; Zatonsky, G.; Ugrak, B. I. J.
Org. Chem. 1994, 59, 4087. (d) Eaton, P. E.; Lukin, K. A. J. Am. Chem.
Soc. 1993, 115, 11370. (e) Fitjer, L.; Conia J. M. Angew. Chem., Int. Ed.
Engl. 1973, 12, 761.
(6) (a) Crandall, J. K.; Conover, W. W. J. Org. Chem. 1974, 39, 63. (b)
Crandall, J. K.; Conover, W. W. J. Chem. Soc., Chem. Commun. 1973, 33.
(7) (a) Dolbier, W. R., Jr.; Akiba, K.; Riemann, J. M.; Harman, C. A.;
Bertrand, M.; Bezaquet, A.; Santelli, M. J. Am. Chem. Soc. 1971, 93, 3933.
(b) Fitjer, L.; Conia, J.-M. Angew. Chem., Int. Ed. Engl. 1973, 12, 334.
(8) (a) Salau¨n, Cf. J. R.; Conia, J. M. Chem. Commun. 1971, 1579. (b)
Salau¨n, Cf. J. R.; Champion, J.; Conia, J. M. Org. Synth. 1977, 57, 36. (c)
Pacifici, J. G.; Diebert, C. J. Am. Chem. Soc. 1969, 91, 4595. (d) Aue, D.
H.; Meshishnek, M. J.; Shellhamer, D. F. Tetrahedron Lett. 1973, 4799.
(e) Pacifici, J. G.; Diebert, C. J. Am. Chem. Soc. 1969, 91, 4595. (f) Slafer,
W. D.; English, A. D.; Harris, D. O.; Shellhamer, D. F.; Meshishnek, M.;
Aue, D. H. J. Am. Chem. Soc. 1975, 97, 6638.
(9) (a) Lambert, R. L.; Seyferth, D. J. Am. Chem. Soc. 1972, 94, 9246.
(b) Mollere, P. D.; Hoffmann, R. J. Am. Chem. Soc. 1975, 97, 3680. (c)
Seyferth, R. L.; Lambert, R. L.; Annarell, D. C. J. Organomet. Chem. 1976,
122, 311. (d) Seyferth, R. L.; Haas, C. K.; Lambert, R. L.; Annarell, D. C.
J. Organomet. Chem. 1978, 152, 131. (e) Eckert-Maksic´, M.; Kovacˇevic´,
K.; Maksic´, Z. B. J. Organomet. Chem. 1979, 168, 295. (f) Jones, P. R.;
White, D. D. J. Organomet. Chem. 1979, 181, 399. (g) Delker, G. L.; Wang,
Y.; Stucky, G. D.; Lambert, R. L.; Haas, C. K.; Seyferth, D. J. Am. Chem.
Soc. 1973, 98, 1779. (h) Delker, G. L.; Wang, Y.; Stucky, G. D.; Lambert,
R. L.; Hass, C. K.; Seyferth, D. J. Am. Chem. Soc. 1976, 98, 1179.
(3) Gustavson, G. J. Prakt. Chem. 1896, 54, 104.
(4) (a) Crandall, J. K.; Paulson, D. R. J. Org. Chem. 1968, 33, 991. (b)
Crandall, J. K.; Paulson, D. R. Tetrahedron Lett. 1969, 2751. (c) Trost, B.
M.; Bogdanowicz, M. J. Am. Chem. Soc. 1973, 95, 5311. (d) Hofland, A.;
Steinberg, H.; de Boer, Th. J. Recl. TraV. Chim. Pays-Bas 1985, 104, 350.
10.1021/ja9905896 CCC: $18.00 © 1999 American Chemical Society
Published on Web 12/02/1999

Synthesis and Structure of a Phospha[3]triangulane
J. Am. Chem. Soc., Vol. 121, No. 50, 1999 11651
the complexed phosphirane 612 and phospha[2]triangulane 7,12
we here report the first synthesis of the remarkably stable
phospha[3]triangulane complex 8 by employing the carbene-
like reactivity of the in situ generated terminal phosphinidene
complex PhPW(CO)5.13 As olefin we chose the well-character-
Figure 1. ORTEP presentation scaled to 50% probability ellipsoids
of 8.
ized bicyclopropylidene 9,14 which has also been used for the
synthesis of [3]triangulanes, their oxa derivatives, and Ti and
Co complexes.15 Because of the importance of strain energies
(SE) for small rings and catenated rings, we also report
computed SE values for the parent phosphirane 11, phospha-
[2]triangulane 12, and phospha[3]triangulane 13.
planes containing the two cyclopropyl rings is normal and
illustrates the absence of any distortion. The P-phenyl is
orthogonal (86.4(5)°) to the phosphirane ring but slightly rotated
(9.5°) from bisecting it. Likewise, the W(CO)5 group is rotated
by 9.0° from eclipsing the P-phenyl group. The phosphirane
ring of 8 is very similar to that of 6 and derivatives thereof.
Only 7 shows P-C bonds of quite different lengths. This
differentiation in bond lengths is common in hetero[2]-
triangulanes, which are also thermally more labile and sensitive
to acidic rearrangements.8,9,15a,b We note that 8 is stable in
toluene for 48 h at 60°.
Do cyclopropyl ring substituents increase the strain in the
phosphirane ring? Comparison of the X-ray structure of 8 with
those of 7 and 6 does not provide a clear answer. We therefore
resorted to ab initio theory to compute their structures and strain
energies. Parent structures 11, 12, and 13 were computed without
the W(CO)5 group and without P substituents (i.e., only P-H)
to maintain the calculations within manageable proportions.
Their B3LYP and MP2(full)/6-31G* geometries are shown in
Figure 2, which also contains the geometries of the hydrocarbons
1, 2, and 3. B3LYP/6-31G* structural parameters of the
phosphirane rings are summarized in Table 3, together with
those determined by microwave spectroscopy for 11. These data
illustrate that the stabilizing Ph-P-W(CO)5 substituents (Table
1) shorten the C-P bonds by ∼0.05 Å, lengthen the C-C bond
by ∼0.01-0.03 Å, and widen the CPC bond angle by ∼2°.
Comparison between the B3LYP structures (Table 3) shows a
decrease both in C-P and C-C bond lengths and in the CPC
angle with each cyclopropyl increment (an average CP bond
length is used for 12). Thus, the cyclopropyl groups appear to
Results and Discussion
Catalyzed by CuCl, the addition reaction of the complexed
phosphinidene precursor 10 with an excess of 9 in toluene at
55 °C for 2 h in a pressure chamber gives the desired product
8 in 70% yield as colorless, air-stable crystals after workup.
While we concentrate on the structural properties and strain
energies, it is of interest to note that the 31P NMR resonance of
8 at δ -129.4 ppm is at surprisingly low field as compared to
the δ -154.8 ppm for phospha[2]triangulane complex 7 and
the δ -187.6 ppm for phosphirane complex 6.
X-ray and Ab Initio Structures. The results of the single-
crystal X-ray structure analysis are shown in Figure 1, with
selected bond lengths and angles summarized in Table 1. Table
2 compares the bond lengths and angles of the phosphirane ring
of 8 with those of 6 and 7. The P-C(1) and P-C(2) bonds of
8 are equal to each other within experimental error limits, which
is in line with the orthogonality of the phosphirane ring with
both cyclopropyl rings (angles of 89.5(6)° and 89.7(7)°) and
with the WPC(7) plane. The intercept of 72.0(7)° between the
(13) Mathey, F. Angew. Chem., Int. Ed. Engl. 1987, 26, 275. Marinetti,
A.; Mathey, F. Organometallics 1984, 3, 456. Marinetti, A.; Mathey, F.;
Fischer, J.; Mitschler, A. J. Am. Chem. Soc. 1982, 104, 4484. Marinetti,
A.; Mathey, F. Organometallics 1982, 1, 1488. Lammertsma, K.; Chand,
P.; Yang, S.-W.; Hung, J.-T. Organometallics 1988, 7, 1875. Hung, J.-T.;
Yang, S.-W.; Chand, P.; Gray, G. M.; Lammertsma, K. J. Am. Chem. Soc.
1994, 116, 10966. Lammertsma, K.; Hung, J.-T. J. Org. Chem. 1992, 57,
6557. Lammertsma, K.; Hung, J.-T. J. Org. Chem. 1993, 58, 1800. Wang,
B.; Lake, C. H.; Lammertsma, K. J. Am. Chem. Soc. 1996, 118, 1690.
(14) de Meijere, A.; Kozhushov, S. I.; Spaeth, T.; Zefirov, N. S. J. Org.
Chem. 1993, 58, 502.
(10) Weber, L.; Luker, E.; Boese, R. D. Organometallics 1988, 7, 978.
(11) Baudler, M.; Leonhardt, W. Angew. Chem., Int. Ed. Engl. 1983,
22, 632.
(12) Hung, J.-T.; Yang, S.-W.; Gray, G. M.; Lammertsma, K. J. Org.
Chem. 1993, 58, 6786.
(15) (a) Erden, I.; de Meijere, A. Tetrahedron Lett. 1980, 21, 2501. (b)
de Meijere, A.; Erden, I.; Weber, W.; Kaufmann, D. J. Org. Chem. 1988,
53, 2501. (c) Foerstner, J.; Kozhushkov, S.; Binger, P.; Wedemann, P.;
Noltemeyer, M.; de Meijere, A.; Butenscho¨n, H. Chem. Commun. 1998,
239.

11652 J. Am. Chem. Soc., Vol. 121, No. 50, 1999
Lammertsma et al.
Table 1. Selected Crystallographic Bond Distances (in Angstroms), Angles (in Degrees), and Dihedral Angles (in Degrees) with ESDs of 8
atoms
W-P
P-C(1)
P-C(2)
P-C(7)
distance/angle
atoms
distance/angle
atoms
distance/angle
2.495(2)
1.807(8)
1.820(8)
1.833(8)
C(1)-C(2)
C(1)-C(3)
C(1)-C(4)
C(3)-C(4)
1.48(1)
1.48(1)
1.49(1)
1.51(1)
C(2)-C(5)
C(2)-C(6)
C(5)-C(6)
1.50(1)
1.51(1)
1.53(1)
WPC(7)
122.3(3)
48.1(4)
61.2(7)
61.4(6)
WPC(1)C(2)
WPC(2)C(1)
C(7)PC(1)C(2)
107.9(4)
-109.4(4)
-97.6(5)
C(7)PC(2)C(1)
C(3)C(1)C(2)C(5)
C(4)C(1)C(2)C(6)
97.5(5)
2.6(1.7)
3.7(1.6)
C(1)PC(2)
C(3)C(1)C(4)
C(5)C(2)C(6)
Table 2. Comparison of Selected Bond Lengths (in Angstroms)
and Angles (in Degrees) for the X-ray Crystal Structures of 8, 7,
and 6
Table 3. Selected B3LYP/6-31G* Bond Lengths (in Angstroms)
and Angles (in Degrees) for 13, 12, and 11
bond/angle
13
12
11a
bond/angle
8
7a
6a
C(1)-C(2)
C(1)-P
1.461
1.866
1.866
46.1
66.9
1.475
1.905
1.849
46.3
68.9/64.9
1.493 (1.502)
1.886 (1.867)
1.886 (1.867)
46.6 (47.48)
66.7 (66.26)
C(1)-C(2)
C(1)-P
1.48(1)
1.820(8)
1.807(8)
2.495(2)
48.1(4)
1.508(9)
1.855(7)
1.794(6)
2.500(2)
48.6(3)
1.50(2)
1.83(2)
1.80(2)
2.504(2)
48.6(7)
122.9(3)
C(2)-P
C(2)-P
C(1)-P-C(2)
P-C(2)-C(1)
P-W
C(1)-P-C(2)
W-P-Ph
a Experimental values from ref 16 are given in parentheses.
122.3(3)
123.2(2)
a Reference 10.
cyclopropyl rings. X-ray structure 8 only suggests this effect,
with a distal C(3)-C(4) bond length of 1.51(1) Å and C(1)-
C(3) and C(1)-C(4) proximal bonds of 1.48(1) and 1.49(1) Å,
respectively, because the differences are within experimental
uncertainties; the C(2)C(5)C(6) cyclopropyl ring has a distal
bond of 1.53(1) Å, with proximal bonds of 1.50(1) and 1.51(1)
Å.
Heats of Formation. With decreasing phosphirane bond
lengths on cyclopropyl substitution, an increase in strain energy
(SE) may be expected, i.e., tighter rings are more strained. For
the parent phosphirane 11, we report its heat of formation (∆Hf)
and SE at the G2MP2 level of theory. After validation of this
methodology on the hydrocarbons 1, 2, and 3, ∆Hf and SEs
are given for the larger systems 12 and 13. Throughout, we
employ homodesmotic reactions (eqs 1, 2, 7-10) to calculate
the SEs. Ring separation reactions (eqs 3-6) are also used as
a simple means to estimate ∆Hf values for the spiro compounds.
(i) Phosphirane. It is informative to review first available
SE values of various three-ring structures. The established
(experimental) SE for cyclopropane of 27.5 kcal/mol is based
on its heats of formation of 12.74 kcal/mol18sin the present
work we use eq 1. Similar SEs have been reported for aziridine
(27.1 kcal/mol) and oxirane (27.2 kcal/mol).19 Bachrach20
reported earlier a SE for phosphirane of 20.1 kcal/mol computed
at the HF/6-31G* level of theory (+ZPE at 3-21G*) using the
homodesmotic reaction shown in eq 2. The reported SE of 19.8
kcal/mol for thiirane is of a similar magnitude,21 while the ab
Figure 2. B3LYP and MP2/6-31G* optimized structures for 1-3 and
11-13. Upper values are the B3LYP bond lengths (in angstroms) and
angles (in degrees), while the lower italic values are those at MP2.
(17) (a) Boese, R.; Miebach, T.; de Meijere, A. J. Am. Chem. Soc. 1991,
113, 1743. (b) Zo¨llner, S.; Buchholz, H.; Boese, R.; Gleiter, R.; de Meijere,
A. Angew. Chem., Int. Ed. Engl. 1991, 30, 1518. (c) Boese, R.; Micbach,
T.; de Meijere, A. Liebigs Ann. 1996, 913. (d) Boese, R.; Blaser, D.;
Gomann, K.; Brinker, U. H. J. Am. Chem. Soc. 1989, 111, 1501.
(18) (a) Wiberg, K. B. Angew. Chem., Int. Ed. Engl. 1986, 25, 312. (b)
The reported SE value of 28.1 kcal/mol is based on the group increment
(GI) method: Schleyer, P. v. R.; Williams, J. E.; Blanchard, K. R. J. Am.
Chem. Soc. 1970, 92, 2377. Wu¨rthwein, E.-U.; Chandrasekhar, J.; Jemmis,
D. E.; Schleyer, P. v. R. Tetrahedron Lett. 1981, 22, 843.
(19) Pihlaja, K.; Taskinen, E. In Physical Methods in Heterocyclic
Chemistry; Kratitzky, A. R., Ed.; Academic Press: New York, 1974; Vol.
6, p 199.
(20) Bachrach, S. M. J. Phys. Chem. 1989, 93, 7780.
(21) Stirling, C. J. M. Tetrahedron 1985, 41, 1613.
(22) Boatz, J. A.; Gordon, M. S.; Hilderbrandt, R. L. J. Am. Chem. Soc.
1988, 110, 352. See also: Kitchen, D. B.; Jackson, J. E.; Allen, L. C. J.
Am. Chem. Soc. 1990, 112, 3414 (this work reports a SE for silirane of
40.2 kcal/mol).
tighten the phosphirane ring. The X-ray structures suggest a
similar effect, but those differences are well within experimental
error limits.
Boese and de Meijere showed by low-temperature X-ray
structure determinations the distal C-C bonds (the bonds away
from the spiro carbon) of [n]triangulanes to be slightly longer
than their proximal bonds (the bonds connected to the spiro
carbon).17 For example, for [3]triangulane 3 these are 1.531(1)
and 1.485(1) Å, respectively, and the ab initio bond lengths
show the same effect. The calculated phosphirane geometry of
13 (but also of 12) displays the same behavior for the
(16) Bowers, M.; Baudet, R. A.; Goldwhite, H.; Tang, R. J. Am. Chem.
Soc. 1969, 91, 17.

Synthesis and Structure of a Phospha[3]triangulane
J. Am. Chem. Soc., Vol. 121, No. 50, 1999 11653
initio estimate for silirane is much higher (43.4 kcal/mol).22
of formation for cyclopropane (12.74 kcal/mol) and methane
(-17.89 kcal/mol) results in estimated ∆Hf values of 43.37 kcal/
mol for 2 and 74.00 kcal/mol for 3, which are in excellent
agreement with the corresponding experimental values of 44.23
and 72.27 kcal/mol.27
1 + 3 CH3CH3 f 3 CH3CH2CH3
11 + 2 CH3PH2 + CH3CH3 f
(1)
CH3PHCH3 + 2 CH3CH2PH2 (2)
2 + CH4 f 2 1
(3)
(4)
In the absence of experimental data, such as its heat of
formation, a more rigorous theoretical treatment seemed desir-
able to ascertain the SE of phosphirane. We estimate a ∆Hf
value of 20.2 kcal/mol on the basis of the G2MP2 calculated
heat of atomization. An SE of 21.3 kcal/mol is obtained using
G2MP2 calculated ∆Hf values for all the fragments23 of eq 2.
Using experimental ∆Hf data instead gives a larger SE of 25.5
kcal/mol, but uncertainties in the experimental data render this
a less reliable estimate. The accuracy of the G2MP2 method
was verified for cyclopropane, giving calculated ∆Hf and SE
values (eq 1) of 14.5 and 28.0 kcal/mol, respectively. These
are in good agreement with the corresponding experimental
values of 12.7 and 27.5 kcal/mol. The SE value of 21.3 kcal/
mol for the low-coordinate phosphirane is similar in magnitude
to that of thiirane and expectedly smaller than those of the three
rings of the first row of the periodic table. Bachrach20 explained
this to result from the small CPC angle, which reduces the
degree of C-C bending and thereby creates a stronger C-C
bond. The rather large SE for silirane has been traced to the
preference of silicon for angles near 90°.22
3 + 2 CH4 f 3 1
Applying the same approach to the phospha[n]triangulanes,
by using eqs 5 and 6 and the G2MP2 calculated ∆Hf value of
20.2 kcal/mol for phosphirane 11, gives estimated ∆Hf values
of 50.7 kcal/mol for 12 and of 81.2 kcal/mol for 13.
12 + CH4 f 11 + 1
(5)
(6)
13 + 2 CH4 f 11 + 2 1
(b) G2MP2 Heats of Formation. To alleviate concerns about
the accuracy of the heats of formation of the organophosphorus
compounds, we computed them from atomization energies using
the G2MP2 method. At this level of theory, the ∆Hf values are
51.2 kcal/mol for 12 and 82.0 kcal/mol for 13, both of which
compare remarkably well with those resulting from the above-
discussed ring separation method.
To verify proper performance of the G2MP2 method, we also
computed the heats of formation for spiropentane 2 and [3]-
triangulane 3. Sizable deviations from experimental data were
found, much to our surprise. For example, the G2MP2 calculated
value of 47.08 kcal/mol for 2 is already larger than the
experimentally determined value by a noticeable 2.85 kcal/mol.
However, for 3 the difference amounts to a very significant 6.8
kcal/mol, well outside the range of <2 kcal/mol established for
small systems. This suggests that the G2MP2 estimated ∆Hf-
(3) of 79.06 kcal/mol is too large, since the differences for the
smaller alicyclic hydrocarbons, used in the homodesmotic
reaction, are modest and within the expected range.23 It appears
that the G2MP2 method works well for small molecules but
that its performance is less satisfactory for the larger ones, and
in particular for [3]triangulane, the system of interest. The origin
of this deviation lies mainly in the empirical higher level
correction (HLC) that the G2MP2 method uses. This HLC
correction becomes substantial for large systems because it is
related to the number of valence electron pairs, E(HLC) )
(-0.19nR - 4.81nâ) × 10-3. Arbitrarily increasing E(HLC)
by 10%, which we denote as G2MP2′,28 results in the more
reasonable ∆Hf values of 44.26 kcal/mol for 2 and 75.30 kcal/
mol for 3, while those for the smaller hydrocarbons are less
affected. We note that the similarly computed ∆Hf of 12.57
kcal/mol for cyclopropane also agrees better with the experi-
mental value.
If the ∆Hf values of 1, 2, and 3 require an HLC adjustment
in the G2MP2 approach, we must assume that the same applies
for the organophophorus compounds. Indeed, the resulting
G2MP2′ calculated ∆Hf values are likewise smaller, although
the effect is less than that for 1-3, i.e., 18.3 kcal/mol for
phosphirane 11, 48.4 kcal/mol for 12, and 78.2 kcal/mol for 13
(see Table 4).
G2MP2 Strain Energies. First we evaluate the performance
of G2MP2 theory by determining the strain energies for 2 and
3 using only calculated G2MP2 heats of formations for all the
molecules of eqs 7 and 8.
(ii) Phospha[n]triangulanes. The total ring strain of [n]-
triangulanes is not a simple addition of the SEs of the number
of cyclopropane rings. For example, based on experimental heats
of formation, the SEs are estimated to be 65 kcal/mol for [2]-
triangulane 2 and 99 kcal/mol for [3]triangulane 3,25 which are
more than twice and three times, respectively, that of cyclo-
propane 1. De Meijere showed for a series of [n]triangulanes
(n e 5) that the excess SE per cyclopropane increment is 8.6
kcal/mol.25 Whether the same effect applies to phospha[n]-
triangulanes is not easily established due to the noted lack of
experimental heats of formation. We therefore resort to two
computational approaches, i.e., ring separation reactions and
G2MP2 theory, to obtain estimated ∆Hf values. The methods
are also applied to the related hydrocarbons for reference.
(a) Ring Separation Reactions. This approach to obtain ∆Hf
values for saturated spiro hydrocarbons, including 2 and 3, was
successfully employed earlier by Radom26 and by Ru¨chard and
de Meijere.25 Thus, using eqs 3 and 4 and the experimental heats
(23) The G2MP2′ ∆Hf298 enthalpies (in kcal/mol), with G2MP2 (first)
and reported experimental values24 (second) in parentheses, are as follow:
-18.94 (-17.99, -17.80) for methane, -21.29 (-19.72, -20.1) for ethane,
-26.45 (-24.25, -25.0) for propane, -42.31 (-38.86, -40.0) for
neopentane, -5.37 (-3.80, -4) for methylphosphine, -14.51, (-12.31,
-14) for dimethylphosphine, -10.25 (-8.05, -9.7 (est.)) for ethylphos-
phine, and -24.57 (-21.11, -24) for tert-butylphosphine. The experimental
∆Hf of 9.7 for EtPH2 is an estimate based on the G2MP2 calculated
difference of 4.3 kcal/mol with the isomeric Me2PH.
(24) Lias, S. G.; Bartmess, J. E.; Liebman, J. F.; Holmes, Levin, R. D.;
Mallard, W. G. J. Phys. Chem. Ref. Data 1988, 17, Suppl. 1 (Gas-Phase
Ion and Neutral Thermochemistry).
(25) Beckhaus, H.-D.; Ru¨chardt, C.; Kozhushkov, S. I.; Belov, V. N.;
Verevkin, S. P.; de Meijere, A. J. Am. Chem. Soc. 1995, 117, 11855.
(26) Kao, J.; Radom, L. J. Am. Chem. Soc. 1978, 100, 760.
(27) Luk’yanova, V. A.; Pimenova, S. M.; Kolesov, V. P.; Kuznetsova,
T. S.; Kokoreva, O. V.; Kozhushkov. S. I.; Zefirov, N. S. Zh. Fiz. Khim.
1993, 67, 1148; Russ. J. Phys. Chem. 1993, 67, 1023.
(28) The recently developed G3 method uses four molecule-independent
parameters for the higher level coorection, but this method is too expensive
for the molecules of the present study. G3: Pople, J. A. (a) Angew. Chem.
1999, 111, 2015. (b) Curtiss, L. A.; Raghavachari, K.; Redfern, P. C.;
Rassolov, V.; Pople, J. A. J. Chem. Phys. 1998, 109, 7764. G3(MP2):
Curtiss, L. A.; Redfern, P. C.; Raghavachari, K.; Rassolov, V.; Pople, J. A.
J. Chem. Phys. 1999, 109, 4703.
2 + 6 CH3CH3 f 4 CH3CH2CH3 + C(CH3)4
(7)

11654 J. Am. Chem. Soc., Vol. 121, No. 50, 1999
3 + 9 CH3CH3 f 5 CH3CH2CH3 + 2 C(CH3)4 (8)
Lammertsma et al.
Table 4. Summary of ∆Hf and SE (in kcal/mol) from Ring
Separation (RS) Reactions and G2MP2 Estimates
∆Hf/SE method 11 12 13
∆Hf RS 50.69 81.19
We note that the SE values are not influenced by the discussed
HLC correction because of the balance of valence electron pairs
in the homodesmotic equations. The resulting SEs are 64.6 kcal/
mol for 2 and 100.5 kcal/mol for 3, which are 1.0 and 4.2 kcal/
mol larger than those based on experimental data. This is a
remarkably good comparison considering that eq 4 (for 3)
contains 17 molecules. The 4.2 kcal/mol difference in SE values
for 3 results largely from the 3.8 kcal/mol difference between
its G2MP2′ and experimental ∆Hf values. It then appears that
G2MP2 theory performs very satisfactorily in determining strain
energies for the hydrocarbons.
1
2
3
43.37
74.00
79.06
75.30
G2MP2 20.19 51.24 81.99 14.46 47.08
G2MP2′ 18.31 48.41 78.22 12.57 44.26
exp(g)
G2MP2 21.3 54.7 87.9 28.0
exp
27.5a 63.6d,e 96.4d,e
excess SE G2MP2 8.6 16.5
12.74a 44.23b 72.27c
SE
64.6
100.5
5.4 10.6
a Reference 18. b Fraser, F. M.; Prosen, E. J. J. Res. Nat. Bur. Stand.
1955, 54, 143. c Reference 27. d Determined from eq 7 using experi-
mental ∆Hf values. e Reference 29.
For the organophosphorus compounds, using calculated ∆Hf
values only for all the molecules of eqs 9 and 10 results in SE
values of 54.7 kcal/mol for 12 and 87.9 kcal/mol for 13. These
are, as expected, considerably lower than those of the hydro-
carbons 2 and 3.
the organophosphorus analogues, for which no experimental data
are available, the G2MP2′ calculated ∆Hf values are 18.3 kcal/
mol for 11, 48.4 kcal/mol for 12, and 78.2 kcal/mol for 13.
Strain energies for the parent organophosphorus and hydro-
carbon compounds were obtained from G2MP2 calculated ∆Hf
values. These theoretical values are 21.3 kcal/mol for 11, 54.7
kcal/mol for 12, and 87.9 kcal/mol for 13.
The excess strain in the phospha[n]triangulanes (n e 2),
which is the strain for spiro carbons in excess of that of the
three-membered rings, amounts to 5.3 kcal/mol per spiro carbon.
This is ca. 40% less than the excess strain in linear [n]triangulane
hydrocarbons.
12 + 2 CH3PH2 + 4 CH3CH3 f
CH3PHCH3 + CH3CH2PH2 + 2 CH3CH2CH3 +
CH3C(CH3)2PH2 (9)
13 + 2 CH3PH2 + 7 CH3CH3 f
CH3PHCH3 + 4 CH3CH2CH3 + 2 CH3C(CH3)2PH2 (10)
Excess Strain. From the computed SE (and ∆Hf) data, excess
strain per cyclopropyl increment on the cyclopropane and
phosphirane rings can be determined. Based on experimental
heats of formation only, de Meijere showed, as already noted,
that the excess SE per cyclopropyl is 8.6 kcal/mol for the
hydrocarbon [n]triangulanes (n e 5).25 Using SE values based
on G2MP2 calculated heats of formations for the more limited
set 1, 2, and 3 (n e 2), we obtain, gratifyingly, a similar excess
SE per cyclopropyl group of 8.4 kcal/mol.
Computational Section
All electronic structure calculations were carried out using the
GAUSSIAN 94 suite of programs.30 For the density functional theory
(DFT) calculations, Becke’s three-parameter hybrid exchange functional
was used in combination with the Lee-Yang-Parr correlation
functional,21-33 denoted as B3LYP. The 6-31G* basis set was employed
throughout for the geometry optimizations. The B3LYP and MP2(full)
structures of 1-3 and 11-13 were characterized as minima by analysis
of their Hessian matrixes.
Heats of formation (∆Hf298) of methane, ethane, propane, neopentane,
1, 2, 3, methylphosphine, dimethylphosphine, ethylphosphine, tert-
butylphosphine, 11, 12, and 13 were estimated from their heats of
atomization calculated at G2MP2.23 This method34 uses MP2/6-31G-
(d) geometries and obtains its energies from the QCISD(T) method,35
using the 6-311G(d,p) basis set,36 with basis set additivity corrections
at the MP2 level of theory. The combination of basis set and correlation
corrections and two empirical corrections yields where the empirical
Substitution of the phosphirane ring results in a smaller excess
SE per cyclopropyl increment of 5.3 kcal/mol, which is based
on the similarly determined SE values for 1, 11, 12, and 13.
This ca. 40% reduction in excess SE reflects the lesser degree
of rehybridization needed for the spiro atoms of the organo-
phosphorus compounds, which is in line with phosphirane’s 6.7
kcal/mol smaller SE than that of cyclopropane.
Conclusions
The first phospha[3]triangulane has been synthesized as a
W(CO)5 complex. This remarkably stable compound 8, carrying
only a Ph-P substituent, has been fully characterized by its
X-ray crystallographic structure. Comparison with the X-ray
structures of the similarly Ph-P substituted phosphaspiropentane
(7) and phosphirane (6) complexes suggests a denser structure,
but the geometrical differences are within experimental uncer-
tainties.
Comparison of the B3LYP/6-31G* and MP2/6-31G* geo-
metrical parameters of the parent compounds phosphirane (11),
phosphaspiropentane (12), and phospha[3]triangulane (13)
shows, indeed, a tightening of the [3]triangulane structure.
Heats of formation were computed by ring separation
reactions and by G2MP2 theory for the parent organophosphorus
compounds and their hydrocarbon analogues. A small modifica-
tion in the HLC correction used in the G2MP2 method was
needed to obtain agreement between the G2MP2 calculated ∆Hf
values and the experimental values reported in the literature
for cyclopropane (1), spiropentane (2), and [3]triangulane (3).
The resulting G2MP2′ calculated ∆Hf values are 12.57 kcal/
mol for 1, 44.26 kcal/mol for 2, and 75.30 kcal/mol for 3. For
E(G2MP2) ) E(QCISD(T)/6-311G(d,p) +
E(MP2/6-311+G(3df,2p) - E(MP2/6-311G(d,p) +
E(HLC) +E(ZPE)
“higher level correction” is given by E(HLC) ) (-0.19nR - 4.81nâ)
(29) Using the GI method gives SE values of 65.1 and 98.5 kcal/mol
for 2 and 3, respectively.25
(30) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Gill, P. M. W.;
Johnson, B. G.; Robb, M. A.; Cheeseman, J. R.; Keith, T.; Petersson, G.
A.; Montgomery, J. A.; Raghavachari, K.; Al-Laham, M. A.; Zakrzewski,
V. G.; Ortiz,J. V.; Foresman, J. B.; Cioslowski, J.; Stefanov, B. B.;
Nanayakkara, A.; Challacombe, M.; Peng, C. Y.; Ayala, P. Y.; Chen, W.;
Wong, M. W.; Andres, J. L.; Replogle, E. S.; Gomperts, R.; Martin, R. L.;
Fox, D. J.; Binkley, J. S.; Defrees, D. J.; Baker, J.; Stewart, J. P.; Head-
Gordon, M.; Gonzalez, C.; Pople, J. A. GAUSSIAN94, Revision B.1;
Gaussian, Inc.: Pittsburgh, 1994.
(31) Becke, A. D. J. Chem. Phys. 1993, 98, 5648.
(32) Becke, A. D. Phys. ReV. A 1988, 38, 3098.
(33) Lee, C.; Yang, W.; Parr, R. G. Phys. ReV. B 1988, 37, 785.
(34) Curtiss, L. A. R., K.; Pople, J. A. J. Chem. Phys. 1993, 98, 1293.
(35) Raghavachari, K.; Trucks, G. W.; Pople, J. A.; Replogle, E. S. Chem.
Phys. Lett. 1989, 158, 207.
(36) Krishnan, R.; Binkley, J. S.; Seeger, R.; Pople, J. A. J. Chem. Phys.
1980, 72, 650.

Synthesis and Structure of a Phospha[3]triangulane
J. Am. Chem. Soc., Vol. 121, No. 50, 1999 11655
Table 5. Crystal Data and Data Collection Procedures for 8
determined by 31P NMR. The reaction mixture was filtered, evaporated
to dryness, and chromatographed on silica gel with hexanes to yield
0.18 g (70%) of 8, which after fractional crystallization from hexanes
gave colorless crystals: mp 80-81 °C; 31P NMR (C6D6) δ -129.4
(1J(31P-183W) ) 170.6 Hz); 13C NMR (C6D6) δ 7.96 (s, syn-CH2), 9.67
formula
molecular weight
crystal system
space group
a (Å)
C19H13O5PW
512
orthorhombic
P212121
6.7571(5)
15.6324(44)
16.8693(36)
1781.9(6)
4
1
(s, anti-CH2), 26.0 (d, J(P-C) ) 27.7 Hz), 128.6-133.5 (Ph), 195.7
b (Å)
(cis CO); 1H NMR (C6D6) δ ) 0.97 (q, 3J(P-H) ) 4.5 Hz, syn-CH2),
3
3
c (Å)
1.18 (q, J(P-H) ) 4.5 Hz, syn-CH2), 1.33 (dt, J(P-H) ) 7.5 Hz,
V (Å3)
3
anti-CH2), 1.45 (dt, J(P-H) ) 7.5 Hz, anti-CH2), 7.56 (m, Ph); MS
(
184W) m/e (relative intensity) 512 (M+, 20), 428 (M+ - 3CO, 30),
Z
F
calc (g/cm3)
1.909
404 (Ph-P(W(CO)4, 20), 376 (Ph-P(W(CO)3, 100), 348 (Ph-
P(W(CO)2, 40). Anal. Calcd for C17H13O5PW: C, 39.84; H, 2.54.
Found: C, 39.59; H, 2.57.
crystal dimensions (mm)
0.14 × 0.18 × 0.08
132.329
Cu KR (1.5418)
5.4130 × 10-6
0.1 < θ < 148
ω/2θ
abs coeff (cm-1
radiation (Å)
)
X-ray Structure Determination of 8. Diffraction data of a single
crystal, mounted on a glass fiber with epoxy cement, were collected at
room temperature on an Enraf-Nonius CAD4 diffractometer using Ni-
filtered Cu KR radiation. Standard peak search and automatic indexing
routines, followed by least-squares fits of 25 centered reflections,
yielded the lattice constants for the crystal. Three reflections were
measured periodically to monitor decay, and linear decay corrections
were applied. Intensities were corrected for Lorentz and polarization
effects; linear decay and numerical absorption corrections were applied.
Variances were assigned on the basis of standard counting statistics,
extinction coeff
2θ limits (deg)
scan type
scan width (Å
1.64
no. of unique data
no. of data with I > 3σ
no. of parameters
largest parameter shift/error
R (%)
2088
2040
257
0.01
3.43
4.72
Rw (%)
GOF
1.507
0.719
-0.538
0%
2
with the addition of an instrumental uncertainty term, 0.04Fo . Lorentz,
max peak in diff Fourier (e/Å3)
min peak in diff Fourier (e/Å3)
decay
polarization, and absorption corrections (using the DIFABS program)
were made to the Is and σ2 terms. The structure was solved by standard
Patterson and difference Fourier techniques and refined by weighted
full-matrix least squares. The final difference Fourier map contained
no interpretable peaks near the W atom as extra atoms. All computations
used the MolEN software. Details of the crystal data and structure
solution procedures are summarized in Table 5.
max abs corrn
min abs corrn
1.513
0.707
× 10-3 au. E(ZPE) is obtained by scaling the SCF/6-31G(d) harmonic
frequencies by 0.8929.
Acknowledgment is made to the National Science Founda-
tion (CHE-9500344) for support and to the Netherlands
Organization for Scientific Research (NCF/NWO) for providing
a grant for supercomputer time. A.W.E. thanks the European
Union for a Marie Curie postdoctoral fellowship.
Experimental Section
NMR spectra were recorded on a Bruker NT-300, wide-bore FT-
NMR spectrometer. Chemical shift are referenced in ppm to internal
(CH3)4Si for the 1H and 13C NMR spectra and external 85% H3PO4 for
the 31P NMR spectra. Downfield shifts are reported as positive. The
elemental analysis was performed by Atlantic Microlab, Inc., Norcross,
GA.
(7-Phenyl-7-phosphadispiro[2.0.2.1]heptane)pentacarbonyl-
tungsten (8). Reaction of complex 1013 (0.33 g, 0.5 mmol) with 10
equiv of bicyclopropylidene 914 (0.45 g, 5 mmol) and ca. 10% CuCl
(50 mg, 0.05 mmol) was carried out in toluene in a pressure chamber
at 55 °C for 0.5-1 h until complex 10 was fully converted, as
Supporting Information Available: Tables of bond dis-
tances, angles, positional parameters, and displacement param-
eters for 8 (PDF). This material is available free of charge via
the Internet at http://pubs.acs.org.
JA9905896