The ozonation of silanes and germanes: An experimental and theoretical investigation

Article abstract of DOI:10.1021/ja058065v

Ozonation of various silanes and germanes produced the corresponding hydrotrioxides, R₃-SiOOOH and R₃GeOOOH, which were characterized by ¹H, ¹³C, ¹⁷O, and ²⁹Si NMR, and by infrared spectroscopy in a two-pronged approach based on measured and calculated data. Ozone reacts with the E-H (E = Si, Ge) bond via a concerted 1,3-dipolar insertion mechanism, where, depending on the substituents and the environment (e.g., acetone-d₆ solution), the H atom transfer precedes more and more E-O bond formation. The hydrotrioxides decompose in various solvents into the corresponding silanols/germanols, disiloxanes/digermoxanes, singlet oxygen (O₂(¹Δ _g)), and dihydrogen trioxide (HOOOH), where catalytic amounts of water play an important role as is indicated by quantum chemical calculations. The formation of HOOOH as a decomposition product of organometallic hydrotrioxides in acetone-d₆ represents a new and convenient method for the preparation of this simple, biochemically important polyoxide. By solvent variation, singlet oxygen (O₂(¹Δ _g)) can be generated in high yield.

Full text of DOI:10.1021/ja058065v

Published on Web 03/02/2006
The Ozonation of Silanes and Germanes: An Experimental
and Theoretical Investigation
Janez Cerkovnik,† Tell Tuttle,‡,§ Elfi Kraka,‡ Nika Lendero,†
Bozˇo Plesnicˇar,*,† and Dieter Cremer*,‡
Contribution from the Department of Chemistry, Faculty of Chemistry and Chemical
Technology, UniVersity of Ljubljana, P.O. Box 537, 1000 Ljubljana, SloVenia, and the
Departments of Chemistry and Physics, UniVersity of the Pacific, 3601 Pacific AVenue,
Stockton, California 95211, USA
Received November 28, 2005; E-mail: bozo.plesnicar@fkkt.uni-lj.si; dcremer@pacific.edu
Abstract: Ozonation of various silanes and germanes produced the corresponding hydrotrioxides, R3-
SiOOOH and R3GeOOOH, which were characterized by 1H, 13C, 17O, and 29Si NMR, and by infrared
spectroscopy in a two-pronged approach based on measured and calculated data. Ozone reacts with the
E-H (E ) Si, Ge) bond via a concerted 1,3-dipolar insertion mechanism, where, depending on the
substituents and the environment (e.g., acetone-d6 solution), the H atom transfer precedes more and more
E-O bond formation. The hydrotrioxides decompose in various solvents into the corresponding silanols/
germanols, disiloxanes/digermoxanes, singlet oxygen (O2(1∆g)), and dihydrogen trioxide (HOOOH), where
catalytic amounts of water play an important role as is indicated by quantum chemical calculations. The
formation of HOOOH as a decomposition product of organometallic hydrotrioxides in acetone-d6 represents
a new and convenient method for the preparation of this simple, biochemically important polyoxide. By
solvent variation, singlet oxygen (O2(1∆g)) can be generated in high yield.
Introduction
triethylsilyl hydrotrioxide (Et3SiOOOH), generated by the low-
temperature ozonation of triethylsilane, is an excellent source
Reactions of ozone with saturated organic substrates leading
to organic hydrotrioxides and dihydrogen trioxide are at the
focus of current research, especially because of their relevance
to the toxicity of ozone for the human body.1,2 Although the
involvement of hydrotrioxides, ROOOH, in the reaction of
ozone with C-H bonds has already been well documented, the
ozonation of organometallic hydrides, such as silanes and
germanes, and the potential involvement of the corresponding
hydrotrioxides in these reactions were much less explored. These
polyoxides have already been proposed in the past as unstable
intermediates in the reaction of ozone with various silanes.3
More recently, Corey et al. have presented evidence that
of singlet oxygen (O2(1∆g)).4 Et3SiOOOH was also demonstrated
to react with electron-rich olefins to give dioxetanes that react
intramolecularly with a keto group in the presence of tert-
butyldimethylsilyl triflate to afford 1,2,4-trioxanes, some of them
with a very potent antimalarial activity (Posner trioxane
synthesis).5 However, no unambiguous direct spectroscopic
evidence for the existence of this hydrotrioxide was presented
at that time.
We have already characterized dimethylphenylsilyl hydro-
trioxide as a reactive intermediate in the low-temperature
ozonation of dimethylphenylsilane and suggested that dihydro-
gen trioxide (HOOOH) might be one of the products formed
during its decomposition.6 Here we now report that the formation
of HOOOH (among other products) is indeed the characteristic
feature of the decomposition of various silyl as well as germyl
hydrotrioxides, generated by the low-temperature ozonation of
the corresponding silanes and germanes. This represents still
another convenient method for the preparation of HOOOH.2
The results of detailed experimental and theoretical studies of
the NMR and IR spectral characterization of the polyoxides
† University of Ljubljana.
‡ University of the Pacific.
§ Current address: Max-Planck-Institut fu¨r Kohlenforschung, Kaiser-
Wilhelm-Platz 1, Mu¨lheim an der Ruhr, 45470 Germany.
(1) (a) Plesnicˇar, B,; Tuttle, T.; Cerkovnik, J.; Koller, J.; Cremer, D. J. Am.
Chem. Soc. 2003, 125, 11553. (b) Nyffeler, P. T.; Boyle, N. A.; Eltepu,
L.; Wong, C.-H.; Eschenmoser, A.; Lerner, R. A.; Wentworth, P., Jr. Angew.
Chem., Int. Ed. 2004, 43, 4656. (c) Tuttle, T.; Cerkovnik, J.; Plesnicˇar, B.;
Cremer, D. J. Am. Chem. Soc. 2004, 126, 16093.
(2) (a) Engdahl, A.; Nelander, B. Science 2002, 295, 482. (b) Wentworth, P.,
Jr.; Jones, L. H.; Wentworth, A. D.; Zhu, X. Y.; Larsen, N. A.; Wilson, I.
A.; Xu, X.; Goddard, W. A., III; Janda, K. D.; Eschenmoser, A.; Lerner,
R. A. Science 2001, 293, 1806. (c) Suma, K.; Sumiyoshi, Y.; Endo, Y. J.
Am. Chem. Soc. 2005, 127, 14998. (d) For a review on HOOOH, see:
Plesnicˇar, B. Acta Chim. SloV. 2005, 52, 1.
(4) Corey, E. J.; Mehrotra, M. M.; Khan, A. U. J. Am. Chem. Soc. 1986, 108,
2472.
(3) (a) Spialter, L.; Pazdernik, L.; Bernstein, S.; Swansiger, W. A.; Buell, G.
R.; Freeburger, M. E. J. Am. Chem. Soc. 1971, 93, 5682. (b) Spialter, L.;
Swansiger, W. A.; Pazdernik, L.; Freeburger, M. E. J. Organomet. Chem.
1971, 27, C25. (c) Spialter, L.; Swansiger, W. A. J. Am. Chem. Soc. 1968,
90, 2187. (d) Ouellette, R. J.; Marks, D. L. J. Organomet. Chem. 1968,
11, 407. (e) Spialter, L.; Pazdernik, L.; Bernstein, S.; Swansiger, W. A.;
Buell, G. R.; Freeburger, M. E. AdV. Chem. Ser. 1972, 112, 65. (f)
Alexandrov, Yu. A.; Tarunin, B. I. J. Organomet. Chem. 1982, 238, 125.
(5) (a) Posner, G. H.; Weitzberg, M.; Nelson, W. M.; Murr, B. L.; Seliger, H.
H. J. Am. Chem. Soc. 1987, 109, 278. (b) Posner, G. H.; Webb, K. S.;
Nelson, W. M.; Kishimoto, T.; Seliger, H. H. J. Org. Chem. 1989, 54,
3252. (c) Posner, G. H.; Oh, C. H.; Milhous, W. K. Tetrahedron Lett. 1991,
32, 4235. (d) Hassner, A.; Stumer, C. Organic Syntheses Based on Name
Reactions, 2nd ed.; Pergamon: Amsterdam, Boston, 2002; p 293.
(6) Plesnicˇar, B.; Cerkovnik, J.; Koller, J.; Kovacˇ, F. J. Am. Chem. Soc. 1991,
113, 4946.
9
4090
J. AM. CHEM. SOC. 2006, 128, 4090-4100
10.1021/ja058065v CCC: $33.50 © 2006 American Chemical Society

The Ozonation of Silanes and Germanes
A R T I C L E S
Scheme 1. Possible Ozonation Mechanisms for Silanes (1a,b,e-g) and Germanes (1c,d,h,i)
under investigation as well as of the mechanism of their
formation and decomposition are presented.
mixtures in various solvents (acetone-d6, methyl acetate, tert-
butyl methyl ether, methylene chloride-d2, toluene-d8) at -78
°C produced the corresponding hydrotrioxide (2), characterized
by the OOOH 1H NMR absorption at δ 13.0 ( 1.0 ppm
downfield from Me4Si in yields of 70-90%, as determined by
Experimental Section
Instrumentation, materials (see also Scheme 1), ozonation procedures,
product analysis, and methodology of kinetic studies are collected in
the section Supporting Information.
1
1H NMR spectroscopy. The H, 13C, 17O, and 29Si NMR data
of 2 are, together with those of some other silicon homologues,
collected in Table 1. For selected segments of NMR and IR
spectra of triethylsilyl hydrotrioxide (2e), see Figures 1 and 2
(see also Figure S1 for deuterated analogues in Supporting
Information).
The theoretical investigation focused on the study of the model
systems: silane (1a), trimethylsilane (1b), germane (1c), and trimeth-
ylgermane (1d) (see Scheme 1) and was based on a variety of methods,
including density functional theory (DFT),7 Many Body Perturbation
theory using the Møller Plesset Hamiltonian at second and fourth order
(MP2 and MP4)8 (for energy, geometry, and other property calcula-
tions), the polarizable continuum descriptions COSMO and PISA (for
solvent effect calculations),9 sum-over-states density functional per-
turbation theory (SOS-DFPT) based on the “individual gauge for
localized orbitals” (IGLO) scheme10 (NMR shift calculations), and the
adiabatic vibrational mode analysis11 (calculation and analysis of
infrared spectra). Calculations were performed using the program
packages COLOGNE 200512a and Gaussian 03.12b Details of the
methods used, the calculation protocol, and the full references for the
programs applied can also be found in the Supporting Information.
Similar observations were also made by studying low-
temperature ozonation of triethylgermane (1h) and dimeth-
ylphenylgermane (1i) (0.01-0.05 M) in acetone-d6, methyl
acetate, and tert-butyl methyl ether at -78 °C. The correspond-
ing hydrotrioxides with characteristic OOOH 1H NMR absorp-
tion at δ 13.0 ( 0.5 ppm downfield from Me4Si were formed
1
in yields of 30-50%. The H, 13C, and 17O NMR data of 2i,h
are collected in Table 1.
Still another OOOH absorption at δ 13.2 ( 0.3 ppm,
corresponding to another polyoxide species with exchangeable
protons (as determined by a relatively fast exchange with DOD
at -60 °C), was observed in the 1H NMR spectra. This
absorption was assigned to dihydrogen trioxide (HOOOH), on
the basis of 17O NMR spectra of polyoxides that were highly
enriched with 17O.1a
Results and Discussion
Ozonation of Silanes and Germanes (Experimental Re-
sults). Ozonation of trimethylsilane (1b), triethylsilane (1e),
dimethylphenylsilane (1f), or triphenylsilane (1g) (1, 0.1 (
0.05M) (see Scheme 1) with ozone-oxygen or ozone-nitrogen
Ozonation of Silanes and Germanes (Theoretical Results).
The basic mechanism proposed for the ozonation of compounds
1a-d (Scheme 1) followed the mechanism found for the
ozonation of hydrocarbons.13 However, contrary to the hydro-
carbons, in the ozonation of silanes and germanes, the formation
of complexes vdW1a-d (see Scheme 1) could not be con-
firmed. At the DFT level, the vdW structures found (listed in
Tables 2 and 3) were no longer stable after BSSE (basis set
superposition error) corrections had been included into the
quantum chemical description. Also at the MP2 level of theory,
the stabilization of potential vdW complexes was smaller than
the changes caused by vibrational corrections. Considering the
fact that the H atoms directly bonded to Si or Ge are negatively
charged and that ozone has a dipole moment of just 0.53 (0.7
calculated) Debye,14 other than weak dispersion interactions,
(7) (a) Kohn, W.; Sham, L. Phys. ReV. A 1965, 140, 1133. For reviews on
DFT, see for example: (b) Parr, R. G.; Yang, W. International Series of
Monographs on Chemistry 16: Density Functional Theory of Atoms and
Molecules; Oxford University Press: New York, 1989. (c) Theoretical and
Computational Chemistry, Vol. 2, Modern Density Functional Theory/A
Tool for Chemistry; Seminario, J. M., Politzer, P., Eds.; Elsevier: Am-
sterdam, 1995.
(8) For a recent review, see: (a) Cremer, D. In Encyclopedia of Computational
Chemistry; Schleyer, P. v. R., Allinger, N. L., Clark, T., Gasteiger, J.,
Kollman, P. A., Schaefer, H. F., III., Schreiner, P. R., Eds.; Wiley:
Chichester, 1998; Vol. 3, p 1706. See also: (b) Møller, C.; Plesset, M. S.
Phys. ReV. 1934, 46, 618.
(9) (a) Miertus, S.; Scrocco, E.; Tomasi, J. J. Chem. Phys. 1981, 55, 117. (b)
Barone, V.; Cossi, M.; Tomasi, J. J. Chem. Phys. 1997, 107, 3210. (c)
Cammi, R.; Cossi, M.; Tomasi, J. J. Chem. Phys. 1996, 104, 4611. (d)
Mennucci, B.; Tomasi, J. J. Chem. Phys. 1997, 106, 5151. (e) Tomasi, J.;
Mennucci, B. In Encyclopedia of Computational Chemistry; Schleyer, P.
v. R., Allinger, N. L., Clark, T., Gasteiger, J., Kollman, P. A., Schaefer,
H. F., III, Schreiner, P. R., Eds.; Wiley: Chichester, 1998; Vol. 1, p 2547.
(10) (a) Olsson, L.; Cremer, D. J. Chem. Phys. 1996, 105, 8995. (b) Olsson, L.;
Cremer, D. J. Phys. Chem. 1996, 100, 16881.
(11) Konkoli, Z.; Cremer, D. Int. J. Quantum Chem. 1998, 67, 1.
(12) (a) Kraka, E.; Gra¨fenstein, J.; Filatov, M.; He, Y.; Gauss, J.; Wu, A.; Polo,
V.; Olsson, L.; Konkoli, Z.; He, Z.; Cremer, D. COLOGNE 2005; University
of the Pacific: Stockton, CA, 2005. (b) Frisch, M. J.; et al. Gaussian 03,
revision C.01; Gaussian, Inc.: Wallingford, CT, 2004.
(13) Wu, A.; Cremer, D.; Plesnicˇar, B. J. Am. Chem. Soc. 2003, 125, 9395.
(14) CRC Handbook of Chemistry and Physics on CD-ROM, 2000 version; Lide,
D. R., Ed.; CRC Press LLC: Boca Raton, FL, 2000.
9
J. AM. CHEM. SOC. VOL. 128, NO. 12, 2006 4091

A R T I C L E S
Cerkovnik et al.
Table 1. Comparison of Selected 1H, 13C, 29Si, and 17O NMR Chemical Shifts for Some Hydrotrioxides of Silanes and Germanes (2), and
Some of Their Analogues in Acetone-d6 at -30 °Ca
Species
R3E
Xb
1H NMR
13C NMR
29Si NMR
17O NMR
−
δCH3
δ
E
−X
δCH3
δ
C(1)c
δ
Si
δO
trimethylsilane
(1b)
(2b)
X ) H
OH
0.074
0.058
0.15
3.94
3.5-5.0
10.4
2.00
1.50
-1.48
-2.44
-16.12
12.57
24.85
30.95
12.6
OOH
O3O2O1Hd
0.23
13.48
305 (O1)
434 (O2)
318 (O3)
triethylsilanee
(1e)
(2e)
X ) H
0.70
0.75
0.82
0.95
4.20
3.5-5.0
10.3
15.0
13.4
11.2
10.5
-15.24
10.50
23.45
29.64
OH
12.7
OOH
O3O2O1H
13.48
305 (O1)
433 (O2)
319 (O3)
dimethylphenylsilane
(1f)
(2f)
X ) H
0.30
0.32
0.43
4.40
4.5-5.5
10.4
-1.1
140.6
143.5
139.6
-16.95
3.30
14.31
OH
3.1
2.8
22.0
O2O1H
213 (O1)
265 (O2)
303 (O1)
419 (O2)
320 (O3)
O3O2O1H
0.55
13.45
2.7
138.5
17.31
triethylgermanee
(1h)
(2h)
X ) H
1.06
1.07
1.09
3.66
4.5-5.0
10.2
8.2
8.4
OH
13.5
O2O1H
221 (O1)
250 (O2)
306 (O1)
435 (O2)
335 (O3)
O3O2O1H
1.11
12.95
7.0
dimethylphenylgermane
(1i)
(2i)
X ) H
0.35
0.55
0.70
3.80
5.0-5.5
10.6
4.7
1.4
-1.7
140.2
138.9
138.1
OH
24.0
O2O1H
216 (O1)
264 (O2)
305 (O1)
420 (O2)
337 (O3)
O3O2O1H
0.84
13.06
-2.6
a 2 ) (0.07 ( 0.02) M. Chemical shifts in parts per million relative to the internal standard: tetramethylsilane (1H, 13C, and 29Si NMR) and H2O (17O
NMR). b E ) Si, Ge. c Chemical shift for C(1) atom in the phenyl ring. d Chemical shifts for 2b were obtained at -50 °C. e 1H NMR signals for the CH2
and CH3 groups appear as quartets and triplets, respectively.
binding between ozone and a silane (germane) molecule is either
unlikely or unfavorable.
reaction path of 1a were reoptimized using the more complete
6-311++G(3df,3pd) basis set. The larger basis set led to little
change in the reaction energetics (Table 2, values in parenthe-
ses); the activation enthalpy was increased relative to the 6-31G-
(d,p) results by 1.4 kcal/mol (∆Hact ) 12.1 kcal/mol)), whereas
the reaction enthalpy decreased by 1.8 to -71.0 kcal/mol. Since
the structural and electronic changes are similar for all molecules
investigated in this work, we can conclude that the moderate
enthalpic differences and the conservation of the preferred
reaction mechanism obtained with the larger basis set confirm
the reliability of the smaller, 6-31G(d,p), basis set for the
description of the reactions investigated in this work.
According to the calculated TS geometries TS1(1a-d) shown
in Figure 3, the formation of the silyl hydrotrioxide (2a) occurs
in a concerted fashion with the 1,3-insertion of ozone into the
Si-H bond. This is obvious for TS1(1a) because both terminal
O atoms of ozone are close to either H(Si) or Si atom (1.54
and 2.35 Å, respectively; Figure 3a). Also, the Si-H distance
is lengthened from 1.48 to 1.62 Å, indicating the breaking of
this bond. For the purpose of confirming the pericyclic insertion
mechanism and distinguishing it from a H abstraction reaction
caused by ozone, we carried out IRC (intrinsic reaction
coordinate) calculations, which led indeed to H3SiOOOH (2a)
as the product. The activation enthalpy for the insertion reaction
is fairly modest (∆Hact(1a) ) 10.7 kcal/mol), although due to
the large loss in entropy relative to the separated products, the
Gibbs activation energy is twice as large (∆Gact(1a) ) 21.7 kcal/
mol). The product is formed in a strongly exothermic reaction
(∆Hreact(2a) ) -69.1 kcal/mol).
The substitution of silane to form trimethylsilane (1b) results
in some changes in the TS of the insertion reaction (Figure 3b).
The Si-H bond is lengthened to 1.69 Å. There is a closer
contact with one of the terminal ozone oxygen atoms (1.29 Å),
whereas the other terminal O atom is separated by 3.5 Å from
the Si atom. This clearly reflects the steric screening of the Si
atom by the methyl groups, which makes a two-pronged attack
of ozone more difficult. One could assume that because of the
steric screening of the Si atom the barrier increases or the
For the purpose of testing the adequacy of the basis set in
describing the ozone-silane reaction, all structures along the
9
4092 J. AM. CHEM. SOC. VOL. 128, NO. 12, 2006

The Ozonation of Silanes and Germanes
A R T I C L E S
Figure 3. TS structures for the ozonation of compounds 1a-d. Optimized
gas-phase structures using the B3LYP functional and the 6-31G(d,p) basis
set for all atoms except Ge, which was described with the SDD pseudo-
potential and associated basis set. Values in parentheses correspond to the
distances for the structures optimized using the CPCM solvent model with
acetone as the solvent (ꢀ ) 20.7). Distances in Å. For details of the
calculations, see the Supporting Information.
1
Figure 1. (A) H and (B) 17O NMR spectra of triethylsilyl hydrotrioxide
(2e) at -30 °C, generated by the low-temperature ozonation of 1e in acetone-
d6.
simultaneous to Si-O bond formation or a delayed Si-O bond
formation. In the latter case, the barrier decreases rather than
increases because Si-H bond cleavage is supported by the
hyperconjugative and inductive effects of the three methyl
groups that all withdraw electron density from the Si-H bond.
These substituents are responsible for the magnitude of the
reaction barrier because Si-O bond formation occurs relatively
late in the exit channel of the concerted reaction.
Although in the gas phase the same type of product is always
obtained, the difference in the insertion mechanisms may matter
in solution (see below) considering the fact that the Me3Si•/
•OOOH radical pair has an energy of just 2.3 kcal/mol relative
to the reactants (∆Hrad(1b) ) 3.4 kcal/mol), and by this it is
6.1 kcal/mol below the TS energy of 8.4 kcal/mol (Table 2).16
The product of this reaction is again strongly stabilized below
the reactants by -76.2 kcal/mol since the rather labile ozone
π-bonds are converted into the polar (and thereby stabilized)
Si-O and O-H single bonds.
Figure 2. Segments of IR spectra of triethylsilyl hydrotrioxide (2e), and
lower homologues (triethylsilanol and triethylsilyl hydroperoxide) in tert-
butyl methyl ether at - 65 °C (c(2e) ) 1.0 ( 0.2 M).
The ozonation of both germane (1c) and trimethylgermane
(1d) also follows an insertion mechanism with early O-H and
late Ge-O bond formation as described for 1b (∆Hact(1c) )
10.9; ∆Hact(1d) ) 6.4 kcal/mol; Table 2). Germane 1c behaves
differently than silane 1a, which is not caused by steric screening
but by the different polarity of the Ge-H bond compared to
that of the Si-H bond. It has been demonstrated17 that, contrary
insertion mechanism changes to a H abstraction reaction leading
to the radical pair Me3Si•/•OOOH. However, IRC calculations
reveal that the final product is still the hydrotrioxide (2b), simply
because in the gas phase the homolytic splitting of the Si-H
bond triggers the Si-O bond formation.15
Results indicate that, depending on the Si substituents, ozone
insertion proceeds in a different fashion, with a H transfer either
(15) Pryor et al. have already discussed the possibility that in the ozonation of
various saturated organic substrates (hydrocarbons, alcohols, ethers) a
concerted insertion mechanism with the transition state having contributions
from radical or ionic resonance forms might be operative in these oxidations.
(a) Giamalva, D. H.; Church, D. F.; Pryor, W. A. J. Am. Chem. Soc. 1986,
108, 7678. (b) Giamalva, D. H.; Church, D. F.; Pryor, W. A. J. Org. Chem.
1988, 53, 3429. See also: (c) Hellman, T. M.; Hamilton, G. A. J. Am.
Chem. Soc. 1974, 96, 1530.
(16) (a) The calculated activation energy of 8.4 kcal/mol for the ozonation of
trimethylsilane (1b) is in excellent agreement with that reported for the
ozonation of triethylsilane (1e) in CCl4 (8.0 kcal/mol).3f (b) Considerable
Si-H bond breaking in the transition state is indicated by a large kinetic
isotope effect; i.e., kH/kD ) 6.9, for the ozonation of (n-C4H9)3SiH(D).3b
(c) The Hammett F value of -1.43 was reported for the ozonation of
dimethylphenylsilanes.3d
(17) Giju, K. T.; De Proft, F.; Geerlings, P. J. Phys. Chem. A 2005, 109, 2925.
9
J. AM. CHEM. SOC. VOL. 128, NO. 12, 2006 4093

A R T I C L E S
Cerkovnik et al.
Table 2. Reaction Energies for the Gas-Phase-Optimized Ozonation of Silanes and Germanes (1a-d)a
molecule
∆E
∑
ZPE
∆H
∑
S
∆G
µ
1a
vdW1a
-517.29447
0.25 (0.01)
11.14 (12.74)
174.77 (155.22)
0.49 (1.09)
-72.9 (-74.77)
-635.28675
0.54
24.23
-517.24795
0.91 (1.19)
10.70 (12.06)
175.10 (155.35)
1.19 (1.71)
-69.14(-70.97)
-635.14764
1.71
110.72
-517.30056
7.60 (3.26)
21.68 (22.37)
173.26 (153.24)
-0.64 (-0.24)
-58.52(-60.49)
-635.21336
7.78
0, 0.62
24.79 (24.48)
24.72 (24.28)
23.82 (23.49)
24.52 (24.35)
28.78 (28.69)
79.57
88.27 (103.68)
73.87 (76.02)
116.89 (117.69)
116.86 (117.15)
75.07 (75.45)
138.32
0.96 (0.66)
0.73 (1.06)
0, 0.60 (0, 0.98)
0.03, 1.22 (0.06, 1.13)
2.90 (2.71)
0, 0.62
TS1(1a)
H3Si+ + HOOO-
H3Si• + HOOO•
2a
1b
vdW1b
79.98
117.96
0.16
TS1(1b)
8.42
131.73
2.26
-79.48
-231.59673
3.74 (0.01)
11.93 (9.25)
161.36 (139.57)
-9.37 (-8.27)
-57.6 (-62.68)
-349.56875
3.77
78.57
78.58
80.25
83.00
7.34
131.81
3.36
-76.18
-231.55251
4.37 (0.60)
10.95 (8.25)
162.07 (140.36)
-8.30 (-7.26)
-54.01(-58.97)
-349.43064
4.95
104.45
153.12
146.04
101.46
17.44
127.40
1.06
-65.19
-231.60660
10.58 (3.91)
20.84 (18.05)
160.90 (138.29)
-9.48 (-8.56)
-43.72(-48.73)
-349.49883
11.69
2.64
0, 0.60
0.67, 1.22
3.59
0, 0.62
0.90 (0.66)
1.46 (1.70)
0, 0.60 (0, 0.98)
0.06, 1.22 (0.15, 1.13)
1.59 (1.56)
0, 0.62
Me3Si+ + HOOO-
Me3Si• + HOOO•
2b
1c
vdW1c
22.73
113.85
23.17 (22.90)
22.27 (22.24)
22.72 (22.73)
23.41 (23.30)
26.92(26.96)
78.56
93.02 (102.68)
80.68 (80.91)
117.76 (120.72)
117.80 (118.14)
79.35(79.42)
143.53
TS1(1c)
H3Ge+ + HOOO-
H3Ge• + HOOO•
2c
1d
vdW1d
79.03
120.92
0.18
TS1(1d)
7.40
122.28
-8.93
-67.20
77.57
78.49
79.67
82.01
6.44
123.13
-7.45
-63.65
111.02
157.93
150.82
109.74
16.13
118.840
-9.62
-53.58
2.74
0, 0.60
0.46, 1.22
2.03
Me3Ge+ + HOOO-
Me3Ge• + HOOO•
2d
a The energies, enthalpies, and free energies are given in kcal mol-1. The entropy is in e.u., and the dipole moment is in debye. All calculations were
performed with B3LYP. The 6-31G(d,p) basis set was used for all atoms (values in parentheses calculated with 6-311++G(3df,3pd) basis set) except Ge,
which was treated with the SDD pseudopotential (for details, see Supporting Information).
Table 3. Reaction Energies for the Solvent-Phase-Optimized
to the normal behavior of atoms in a group of the periodic table,
Ozonation of Silanes and Germanes (1a-d)a
Ge as the element with the larger atomic number is somewhat
molecule
∆E
∆
G
µ
less electropositive (more electronegative) than Si, which leads
to a reduced charge transfer from Ge to H (see NBO charges in
Figure 4). The Ge-H bond is less polar and by this weaker
than the Si-H bond. This effect is enhanced by the smaller
overlap between a 4pσ(Ge) and the 1s(H) orbital compared to
that between the 3pσ(Si) and the 1s(H) orbital. Hence, Ge-H
is more prone to homolytic (or heterolytic) bond cleavage than
Si-H (see Table 2). This is confirmed by the activation
enthalpies calculated with the larger basis set18 (1a, 12.1; 1c,
8.2 kcal/mol; the smaller basis set is not sensitive to this
difference) and by the geometry of TS1(1c), which identified
early Ge-H bond breakage. Methyl screening of Ge leads to
another reduction of the activation enthalpy (6.4 kcal/mol; Table
2). Ge-H bond weakening is also responsible for the fact that,
for the germanes, the formation of the radical pair occurs in an
exothermic rather than slightly endothermic process (∆Hrad(1c)
) -8.3; ∆Hrad(1d) ) -7.5 kcal/mol; Table 2). Again, this
observation is only relevant to the solution rather than the gas
phase. The final products (2c and 2d, respectively) are formed
in a less exothermic reaction (∆Hreact(1c) ) -54.0; ∆Hreact(1d)
) -63.7 kcal/mol) than the corresponding Si products because
of the reduced polarity of the Ge-O bond caused by the higher
electronegativity of Ge.
1a
vdW1a
-517.29707
0.47
11.22
34.66
-0.99
-77.11
-635.29002
0.95
6.97
8.81
-517.28655
1.31
0, 0.62
1.03
TS1(1a)
9.82
34.90
-0.88
1.28
H3Si+ + HOOO-
0, 1.46
0.07, 1.47
3.73
0, 0.62
0.20
H3Si• + HOOO•
2a
1b
vdW1b
-78.36
-635.27341
10.75
TS1(1b)
16.09
7.15
-3.11
5.90
Me3Si+ + HOOO-
0.01, 1.46
0.87, 1.77
4.61
0, 0.62
0.96
Me3Si• + HOOO•
-2.84
2b
1c
vdW1c
TS1(1c)
-83.50
-231.59925
3.98
-75.52
-231.58842
4.97
11.33
21.30
-10.92
-61.94
-349.57192
4.35
6.26
-1.12
-13.96
-70.77
10.50
21.46
2.67
H3Ge+ + HOOO-
H3Ge• + HOOO•
2c
0, 1.46
0, 1.47
2.14
0.65, 0.73
0.23
-10.85
-62.94
-349.55470
3.60
1d
vdW1d
TS1(1d)
4.99
-2.84
4.69
Me3Ge+ + HOOO-
Me3Ge•+ HOOO•
2d
0, 1.46
0, 1.77
2.72
-14.24
-72.95
a The energies and free energies are given in kcal mol-1. The dipole
moment is in debye. All calculations were performed with B3LYP, and
the 6-31G(d,p) basis set was used for all atoms except Ge, which was treated
with the SDD pseudopotential. The CPCM solvent model was used with
united force field atomic radii and the dielectric constant of acetone (ꢀ )
20.7). For details, see the Supporting Information.
In the gas phase, heterolytic Si-H (Ge-H) bond cleavage
followed by the formation of the HOOO anion and a silylium
or germylium cation is highly endothermic (Table 2). The
corresponding data are included into Table 3 to obtain an
estimate on solvent effects with regard to heterolytic bond
cleavage. The effect of the solvent on the reaction mechanism
was investigated by using a continuum dielectric solvent
approach. The optimization of the structures, applying the
dielectric constant of acetone (ꢀ ) 20.7), did not result in a
change in the mechanism for the parent silane, 1a. However,
the TS geometry for the ozonation of 1a changed from
simultaneous H transfer and Si-O bond formation to late Si-O
bond formation as observed already in the gas phase for 1b,
1c, and 1d. This change is accompanied by a significant
(18) The SDD 28-electron pseudopotential and associated (4s,2p)/[3s,2p] basis
set for the valence shell are used to describe the Ge atom; all other atoms
are described with the 6-311++G(3df,3pd) basis set.
9
4094 J. AM. CHEM. SOC. VOL. 128, NO. 12, 2006

The Ozonation of Silanes and Germanes
A R T I C L E S
Figure 4. NBO partial charge distribution for the complexes involved in the ozonation of compounds 1a-d. Charges for compounds 1a and 1c have been
calculated with the 6-311++G(3df,3pd) basis set, whereas 1b and 1d were done with the 6-31G(d,p) basis set. All calculations employed the B3LYP hybrid
functional. For structures containing the methyl groups, the charges of the methyl H atoms have been summed into the carbon. For details of the calculations,
see the Supporting Information.
solv
reduction of the free energy of activation leading to ∆Gact
9.8 kcal/mol (Table 3).
)
calculations confirm the existence of hidden intermediates,
however, they do not give the conditions under which experi-
mental conditions of a solvent-stabilized radical pair are
generated.
The change in the TS geometry opens the possibility of a
solvent-supported shift from the concerted mechanism to a two-
step process leading to an intermediate radical pair. Specific
solvation of the reaction complex by one or more solvent
molecules can separate H3Si• and •OOOH so that they exist
for a certain lifetime outside the solvent cage before abstracting
H atoms from another partner, such as unreacted silane. At least
part of HOOOH might be formed (beside the recombination
product H3SiOOOH) in this way. The experimental results
appear to suggest this possibility (observation of small amounts
of HOOOH even before the decomposition of R3EOOOH),
although our continuum description (specific solvation in the
continuum) is not suitable for proving this reaction product.
We note that the concerted insertion mechanism when
investigated with the URVA (unified reaction valley approach)
of Kraka and co-workers19 indicates for 1b, 1c, and 1d in the
gas phase a “hidden intermediate”. Such an intermediate does
not correspond to a minimum (stationary point) on the potential
energy surface, but is associated with an area of the reaction
path distinctively differing from other path areas (TS area, vdW
area, reactant area, product area). As found for symmetry
forbidden addition reactions, such hidden intermediates indicate
a change in the reaction mechanism from a concerted to a
nonconcerted (two-step) mechanism under the impact of envi-
ronmental factors (solvent polarity, added salt effect, etc.). Our
Contrary to other ozone reactions,13 the formation of an ion
pair requires, even in acetone, a higher energy than the formation
of a radical pair (in the case of 1a, ∆Gradsolv ) -0.9 kcal/mol,
solv
∆Gion
) 34.9 kcal/mol; Table 3), where differences reach
from 11 kcal/mol (methyl-substituted systems 1b and 1d) to
more than 30 kcal/mol (1c, 32.3; 1a, 35.8 kcal/mol; Table 3),
despite the stabilizing influence of the solvent on the ionic pair.
In the methyl-substituted ions, charge polarization is much larger
than in the parent compounds (as is already reflected by the
NBO charges of the TSs given in Figure 4), and accordingly,
the relative energies of the ions in acetone are closer to those
of the radicals. Considering that heterolytic cleavage is facilitated
by acetone by more than 100 kcal/mol, one can predict that a
more polar solvent than acetone can stabilize a potential ion
pair to such an extent that its stability becomes comparable with
that of a radical pair.
The overall reaction mechanism for the germanes remains
unaffected by the inclusion of solvent effects, although there is
a shift in the relative energies. The activation free energy is
solv
reduced in both 1c and 1d by more than 10 kcal/mol (∆Gact
-
(1c) ) 10.5; ∆Gactsolv(1d) ) 5.0 kcal/mol; Table 3). We note
that a potential generation of radical pairs is more exothermic
in acetone than in the gas phase (∆Gradsolv(1c) ) -10.9;
∆Gradsolv(1d) ) -14.2 kcal/mol; Table 3).
(19) Cremer, D.; Wu, A.; Kraka, E. Phys. Chem. Chem. Phys. 2001, 3, 674.
9
J. AM. CHEM. SOC. VOL. 128, NO. 12, 2006 4095

A R T I C L E S
Cerkovnik et al.
Table 4. Kinetic and Activation Parameters for the Decomposition of Hydrotrioxides (ROOOH, 2) and Dihydrogen Trioxide (HOOOH),
Formed in the Low-Temperature Ozonation of Trimethylsilane (1b), Triethylsilane (1e), Dimethylphenylsilane (1f), Triphenylsilane (1g),
Triethylgermane (1h), and Dimethylphenylgermane (1i) in Various Solventsa
ROOOH
HOOOH
1
1
δ
, ppm
k
×
104, s-
k
×
104, s-
c
1
1
1
solvent
T,
°C
ROOOH
OOOHb
CH3
Ea, kcal mol-
log A
δ, ppm
k
×
104, s-
Ea, kcal mol-
log A
(2b) acetone-d6
methyl
acetate
-40
-40
-20
-40
13.43
12.88
12.70
13.30
12.3
0.47
4.46
1.10
(10.0 ( 1.0)b
(11.9 ( 1.2)b
(11.9 ( 1.2)c
(11.7 ( 1.3)b
(6.3 ( 0.7)b
(7.0 ( 0.8)b
(6.9 ( 0.7)c
(7.3 ( 0.8)b
13.15
12.63
12.45
12.95
1.45
0.88
4.57
(11.0 ( 2.0)
(13.8 ( 1.5)
(6.1 ( 1.0)
(8.7 ( 0.9)
0.46
4.34
tert-butyl
methyl ether
(2e) acetone-d6
-40
-20
-40
-20
-40
-20
-70
-60
-70
-60
-40
-40
-20
-40
13.57
13.39
13.02
12.78
13.16
13.02
13.84
13.72
12.10
11.35
13.56
13.24
13.02
13.30
1.80
7.33
0.37
4.94
0.32
1.94
3.53
22.5
0.30
1.24
3.70
2.70
17.0
1.45
2.40
6.80
0.34
3.86
(9.3 ( 1.1)b
(8.9 ( 1.0)c
(14.5 ( 1.2)b
(14.4 ( 1.2)c
(10.8 ( 1.1)b
(5.0 ( 0.6)b
(4.6 ( 0.6)c
(9.3 ( 0.9)b
(9.1 ( 0.9)c
(5.7 ( 0.6)b
13.27
13.10
methyl
acetate
tert-butyl
methyl ether
methylene
chloride-d2
toluene-d8
3.37
19.3
0.24
(12.8 ( 1.1)b
(12.7 ( 1.1)c
(14.2 ( 1.2)b
(14.3 ( 1.2)c
(9.5 ( 1.0)b
(10.0 ( 1.1)b
(10.4 ( 0.9)b
(10.3 ( 0.9)c
(10.7 ( 1.0)b
(10.7 ( 1.0)c
(5.5 ( 0.7)b
(5.9 ( 0.7)b
0.97
(2f) acetone-d6
methyl
acetate
13.16
12.85
12.63
12.79
0.56
(11.0 ( 2.0)
(6.1 ( 1.0)
tert-butyl
(12.6 ( 1.1)b
(8.2 ( 0.8)b
methyl ether
(2g) acetone-d6
tert-butyl
-40
-40
14.00
13.49
4.91
12.37
(11.8 ( 1.1)b
(10.9 ( 1.2)b
(7.8 ( 0.8)b
(7.4 ( 0.8)b
12.75
methyl ether
(2h) acetone-d6
-40
10
-40
-20
-20
10
-40
10
-40
10
13.05
0.41
(10.5 ( 1.2)b
(10.6 ( 1.2)b
(12.7 ( 1.1)b
(5.4 ( 0.7)b
(5.8 ( 0.7)b
(7.1 ( 0.6)b
13.31
12.84
12.88
12.67
12.64
12.74
13.29
12.83
12.72
12.27
12.78
12.39
6.24
0.88
8.10
(16.1 ( 1.0)
(13.8 ( 1.5)
(9.4 ( 0.7)
(8.7 ( 0.7)
methyl
acetate
tert-butyl
methyl ether
(2i) acetone-d6
12.65
12.42
12.57
0.91
3.75
1.17
1.57
2.70
1.07
0.71
(12.0 ( 1.6)
(11.9 ( 1.5)
(12.6 ( 1.6)
(17.2 ( 0.5)
(5.4 ( 0.6)
(5.6 ( 0.6)
(5.9 ( 0.6)
(8.9 ( 0.8)
13.16
12.56
12.86
3.24
0.74
3.52
3.00
0.62
(9.0 ( 1.5)b
(8.9 ( 1.5)c
(12.7 ( 1.3)b
(13.2 ( 1.2)c
(7.4 ( 2.0)b
(5.0 ( 0.6)b
(4.9 ( 0.7)c
(7.8 ( 0.6)b
(8.2 ( 0.7)c
(3.2 ( 1.2)b
methyl
acetate
tert-butyl
methyl ether
-40
20
a 1 ) (0.10 ( 0.05) M, 2 ) (0.09 ( 0.05) M, HOOOH ) (0.05 ( 0.02) M. Standard deviation (10%. b Following decay of the OOOH absorption by
1
1H NMR spectroscopy. c Following decay of the CH3 absorption by H NMR spectroscopy.
The Decomposition of Silyl and Germyl Hydrotrioxides.
Hydrotrioxides of silanes (2e-g) decomposed in the temperature
range of -85 to -10 °C to produce the corresponding silanols
(94 ( 3%, yield/mol of 2), disiloxanes (4 ( 2%), and singlet
oxygen, O2(1∆g) (70 ( 10%), as revealed by GC/MS method,
1H NMR spectroscopy, and 9,10-dimethylanthracene as singlet
oxygen acceptor.1a,4,20 Hydrogen peroxide was also detected as
a minor product in the decomposition mixture (5 ( 2% yield
by iodometry) after the warm-up procedure. It is interesting to
mention that smaller amounts of trimethylsilanol (40 ( 10%,
yield/mol of 1b) and larger amounts of hexamethyldisiloxane
(60 ( 10%) were detected after the decomposition of trimeth-
ylsilyl hydrotrioxide (2b). Similarly, the corresponding diger-
moxanes were found as the main decomposition products (85
( 5%) after the decomposition of germyl hydrotrioxides (2h-
i).
mol of 2) during the decomposition of all investigated hydro-
trioxides (2). We also noticed that the amount of the HOOOH
formed was larger when the warm-up was slower. Additionally,
we observed significantly larger amount of HOOOH formed
(80 ( 10%, yield/mol of 2) when water was added to the
solutions of 2 in all solvents investigated. A relatively faster
decomposition of 2f in acetone-d6 with added water was
observed (compare with Tables 4 and S3 in Supporting
Information). Similarly, the addition of CH3OH or (CH3)3SiOH
to the ozonized solution of silanes (1b,e,f) in acetone-d6 leads
to the formation of larger amounts of HOOOH and, at the same
time, the formation of either methoxy-substituted silane or mixed
disiloxane, respectively.
Kinetic Studies. The kinetics of decomposition of the
hydrotrioxides 2 was measured by following the decay of
OOOH and CH3 absorptions by 1H NMR spectroscopy and was
found to obey first-order kinetics in all solvents investigated.
Both absorptions decayed by approximately the same rate. The
kinetic and activation parameters for the decomposition of 2 in
various solvents are summarized in Table 4 (see also Tables
S1 and S2 in Supporting Information). The disappearance of
the O-H and O-O stretching frequencies in the IR spectra of
Besides the decomposition products already mentioned,
dihydrogen trioxide (HOOOH) was formed (40 ( 20%, yield/
(20) It should be pointed out that singlet oxygen is formed in the ozonation of
various organic compounds. See, for example: (a) Munoz, F.; Mvula, E.;
Braslavsky, S. E.; von Sonntag, C. J. Chem. Soc., Perkin Trans. 2 2001,
1109, and references therein. (b) For older literature, see: Plesnicˇar, B. In
Organic Peroxides; Ando, W., Ed.; Wiley & Sons: New York, 1992; p
479.
9
4096 J. AM. CHEM. SOC. VOL. 128, NO. 12, 2006

The Ozonation of Silanes and Germanes
A R T I C L E S
triethylsilyl hydrotrioxide (2e) is shown in Figure S2 (Supporting
Information).
however, the calculated activation enthalpy for this reaction is
too large (∆Hact(2a) ) 29.2 kcal/mol) to warrant further
investigation of this mechanism for the other species.
The effect of solvent polarity and basicity on the decomposi-
tion of the hydrotrioxides 2 was also examined. We found that
the rate of decomposition of 2 is only slightly sensitive to solvent
polarity, the rate increasing with the solvent polarity (2e: krel
(-40 °C) ) 1.80, 0.37, and 0.32 for acetone-d6, methyl acetate,
and tert-butyl methyl ether, respectively) and the amount of
water present in the reaction mixture (molar ratio 2/H2O ) 1:2
in acetone-d6 (by 1H NMR)). Namely, water was always present
in the reaction mixture in sufficient amounts to participate in
the reaction. Furthermore, the addition of water (2 vol %) to
acetone-d6 solutions did not change significantly the activation
parameters of the decomposition of 2 (compare Tables 4 and
S3).
Path B involves the use of a silanol/germanol molecule that
results from an alternative decomposition mechanism. As in the
case of path A, the calculated activation enthalpies for the parent
compounds 2a and 2c are both within the experimental limit
(∆Hact(2a) ) 13.7; ∆Hact(2c) ) 11.9 kcal/mol), whereas the
activation enthalpies for 2b (∆Hact ) 29.1 kcal/mol) and 2d
(∆Hact ) 19.8 kcal/mol) are again too large. All path B reactions
are calculated to be mildly exothermic (∆Hreact ≈ -5 kcal/mol).
As in the case of path A, it is the steric bulk of the methyl
groups that hinders the approach of the silanol/germanol and
consequently results in larger activation barriers.
In the case of path C, first a single water variant was tested;
however, the activation barriers for all four species (2a-d) were
too large (e.g., ∆Hact(2b) ) 29.2 kcal/mol; ∆Hact(2d) ) 21.4
kcal/mol) to explain the experimentally observed formation of
HOOOH. The high barriers result primarily from the strain in
the four-membered cyclic TS involving the Si-O(OOH) bond
and the H-O(H) bond of the water molecule. Therefore, we
tested the two-water mechanism shown as path C. This leads
to activation enthalpies of 10.5, 22.6, 13.1, and 20.0 kcal/mol
respectively, reflecting the steric advantages of a six-membered
pericyclic arrangement. Despite the favorable reduction in the
activation enthalpy, the barriers calculated for 2b and 2d are
still too large, again due to the steric repulsion with the methyl
groups, to explain the observed HOOOH formation.
The involvement of water in the decomposition was also
confirmed by the magnitude of the solvent isotope effect for
the decomposition of 2f, that is, kH O/kD O ) 1.65 ( 0.20, and
2
2
HOOOH (kH O/kD O ) 4.5 ( 0.7), which indicated a considerable
2
2
water-assisted proton transfer in the transition state of this
reaction (see also Table S3 in Supporting Information).
Only small amounts (sometimes below the detection limit21)
of HOOOH was formed during the decomposition of hydro-
trioxides 2 in methylene chloride-d2 and toluene-d8, most
probably due to the low solubility of water in these solvents.
Again, this observation supports the involvement of water in
proton transfer processes during the formation of dihydrogen
trioxide.
Mechanism of the Decomposition Reactions. The primary
focus of the theoretical investigation on the decomposition of
2a-d was to establish plausible mechanisms that describe this
process and result in the experimentally observed products. As
observed in the experimental studies, the final products of these
reactions should include dihydrogen trioxide (HOOOH), sil-
anols/germanols (5a-d), and disiloxanes/digermoxanes (4a-
d). The generation of HOOOH was detected in the reaction
mixture in appreciable amounts, and therefore, we considered,
beside the direct formation in the first step via radical formation,
also other mechanisms, by which the formation of this species
could occur (paths A, B, and C, Scheme 2). Paths A, B, and C
would also lead to disilyl/digermyl trioxide (3a-d), disiloxane/
digermoxane (4a-d) and silanol/germanol (5a-d) depending
on what reaction partner 2a-d would find in the reaction
mixture (a second hydrotrioxide, silanol (germanol), and water).
In each of these cases, water would play a catalytic role.
Calculated results describing the energetics associated with paths
A, B, and C are listed in Table 5.
Three alternative decomposition pathways that may also
account for this formation are shown, namely, paths D, E, and
F (Scheme 2). Path D involves the possible decomposition of
two hydrotrioxide species to two silanol/germanol molecules
and O2(1∆g) (6a-d, Scheme 2) in an exothermic reaction
(∆Hreact(6a) ) -22.9; ∆Hreact(6b) ) -25.8; ∆Hreact(6c) ) -9.9;
∆Hreact(6d) ) -15.0 kcal/mol; Table 5). The activation enthal-
pies for the four compounds are similar (∆Hact ) 16-18 kcal/
mol), which reflects the fact that the interaction between the
two hydrotrioxide molecules involves only their OOOH units.
The activation enthalpy of >15 kcal/mol is too high to explain
the occurrence of either R3EOH or O2(1∆g) under the experi-
mental conditions.
Path E involves the intramolecular transfer of the hydrotri-
oxide H atom from O3 to O1, resulting in the formation of O2-
(1∆g) and the silanol/germanol (7a-d). The investigation of this
reaction path for the parent silane (2a) and germane (2c) reveals
this mechanism to be an unlikely contributor to the decomposi-
tion of 2a-d (∆Hact(2a) ) 39.2; ∆Hact(2c) ) 38.5 kcal/mol).
The large activation energy for this process is not unprecedented,
with a similar mechanism for the parent dihydrogen trioxide
species reportedly requiring an activation energy of 48 ( 2 kcal/
mol.1c
As for path A, both 2a and 2c are found to react with
moderate activation enthalpies (∆Hact(2a) ) 16.5; ∆Hact(2c) )
13.9 kcal/mol). However, the TSs for the substituted compounds
(TS(2b-3b) and TS(2d-3d)) could not be located on the
potential energy surface probably because steric hindrance
excludes an energetically favorable approach of the reaction
partners. Therefore, it is not likely that this mechanism could
be active for the more highly substituted compounds tested
experimentally. A mechanistic modification of this reaction
without the additional water molecule was also tested for 2a;
The final mechanism (path F), again, similar to path C,
involves two water molecules as a catalyst for the formation of
the silanol/germanol; however, in path F, both water molecules
are retained and one molecule of O2(1∆g) is formed (8a-d,
Scheme 2). Initially, the single water analogue of path F was
tested, resulting in moderate activation enthalpies, ranging from
17 to 20 kcal/mol across the series. However, these activation
enthalpies were still too high, compared to that which was
obtained experimentally. Thus, we considered path F (the two-
(21) The OOOH absorption of all polyoxides under investigation is very broad
in methylene chloride and toluene. Addition of small amounts of an oxygen
base (acetone, ethers) to such solutions considerably sharpens this absorp-
tion.
9
J. AM. CHEM. SOC. VOL. 128, NO. 12, 2006 4097

A R T I C L E S
Cerkovnik et al.
Scheme 2. Possible Decomposition Paths for 2a, 2b, 2c, and 2d (E ) Si, Ge; R ) H, Me)
water mechanism) to test whether a lowering of the strain in
the cyclic TS (Scheme 2) would help to lower the activation
enthalpy. Indeed, this was observed with a decrease in all cases
by about 5 kcal/mol. This leads to an activation enthalpy of
12.7 kcal/mol for 2b and 14.4 kcal/mol for 2d, which are both
comparable with the measured activation enthalpy for the
decomposition of these species. Additionally, the products 8a-d
were found to be generated in an exothermic reaction, where a
reaction enthalpy of -12 to -15 kcal/mol (∆Hreact(2b) ) -13.4;
∆Hreact(2d) ) -12.5 kcal/mol) was calculated.
Thus, we conclude that the formation of the silanol and
germanol species occurs via the two-water-catalyzed mechanism
(path F) and results in the formation of O2(1∆g), which was
also detected experimentally. This mechanism of decomposition
is expected to be the primary source of the O2(1∆g) and the
R3EOH detected in the reaction mixture.
quantum chemical calculations.22 All silyl and germyl hydro-
trioxides under investigation (except for dimethylphenylsilyl
hydrotrioxide) were characterized by IR and NMR spectroscopy
in this work for the first time, and our calculations in
combination with the measured data (Table 1) clearly reveal
that 17O NMR spectroscopy is the appropriate tool to detect
and identify these hydrotrioxides (see part 4 of the Supporting
Information). The IR spectra of the hydrotrioxides 2 (in
solutions) are reported for the first time, whereas their analysis
based on the adiabatic mode approach provides an insight into
the electronic structure of these compounds (see Supporting
Information). The previous matrix IR study of the reaction of
ozone with the parent silane (1a) and trimethylsilane (1b),
reported by Andrews et al.,23 did not reveal the presence of the
corresponding hydrotrioxides 2a and 2b. For a discussion of
their IR spectra, see part 4 of the Supporting Information.
Spectroscopic Characterization of Silyl Hydrotrioxides.
Dihydrogen trioxide (HOOOH)2d and alkyl hydrotrioxides have
been previously characterized by both NMR spectroscopy and
(22) Wu, A.; Cremer, D.; Gauss, A. J. Phys. Chem. A 2003, 107, 8737.
(23) Withnall, R.; Andrews, L. J. Phys. Chem. 1988, 92, 594.
9
4098 J. AM. CHEM. SOC. VOL. 128, NO. 12, 2006

The Ozonation of Silanes and Germanes
A R T I C L E S
Table 5. Reaction Energies for the Decomposition of Silyl and Germyl Hydrotrioxides (2a-d)a
molecule
∆E
∑
ZPE
∆H
∑
S
∆G
µ
B.E.
Path A
2a
-1111.27411
19.44
-0.19
-539.82166
17.38
74.32
73.13
74.38
70.66
68.86
70.87
-1111.13854
16.46
-0.19
-539.69115
13.91
137.18
115.43
131.40
145.26
123.40
138.23
-1111.20372
22.95
1.53
-539.76017
20.43
4.17
2.70
3.73
4.24
3.27
5.39
-13.47
TS(2a-3a)
3a
2c
TS(2c-3c)
3c
-13.14
-14.50
-2.35
-2.24
-0.14
-16.57
-13.81
Path B
-960.88562
13.67
-5.83
-1196.70913
29.05
-5.17
-389.43543
11.91
-5.77
-625.22392
19.78
2a
-961.01205
16.83
-6.46
-1197.01811
31.18
-6.23
-389.55710
15.61
-6.02
-625.53184
20.59
70.01
68.67
70.66
178.44
177.19
179.07
66.68
64.27
67.10
176.81
176.56
177.53
123.64
103.38
121.49
166.56
155.38
168.76
132.27
115.65
125.29
177.31
168.93
176.33
-960.94437
19.70
-5.19
-1196.78827
32.38
-5.82
-389.49828
16.86
-3.69
-625.30816
22.27
3.09
3.02
1.59
3.65
2.24
2.10
2.67
4.12
2.72
3.28
2.96
3.11
TS(2a-4a)
4a
2b
TS(2b-4b)
4b
2c
TS(2c-4c)
4c
2d
TS(2d-4d)
4d
-11.28
-14.07
-12.92
-14.84
-17.76
-15.30
-4.73
-3.84
-3.55
-19.58
-16.49
Path C
-670.17079
10.50
-11.28
-788.08381
22.59
-6.84
-384.44614
13.09
-4.26
-502.34241
19.99
2a
-670.27984
14.01
-11.40
-788.28412
26.14
-7.10
-384.55242
16.79
60.98
58.88
61.04
114.32
112.61
114.82
58.86
56.37
58.93
113.57
111.62
114.18
102.12
90.65
-670.21931
13.92
-11.91
-788.14634
27.27
-7.35
-384.49679
15.82
2.11
3.61
0.53
2.84
3.35
2.72
2.02
3.82
2.80
3.19
4.45
2.93
TS(2a-5a)
5a
2b
TS(2b-5b)
5b
2c
TS(2c-5c)
5c
2d
TS(2d-5d)
5d
104.24
131.61
115.91
133.32
106.59
97.43
112.48
136.61
125.89
137.83
-18.52
-13.52
-15.78
-15.31
-4.50
-502.54218
23.35
-6.01
-502.40732
23.19
-16.18
-16.16
-4.89
-4.59
-4.96
-18.78
-11.15
Path D
-1034.74222
15.87
-22.90
-1270.56717
17.45
-25.79
-463.30006
17.84
-9.94
-699.08926
18.44
2a
-1034.85001
21.42
-21.26
-1270.85766
22.33
-25.30
-463.40309
22.88
-9.07
-699.37983
22.92
-14.07
59.08
53.94
56.48
167.34
53.94
56.48
55.85
53.94
56.48
166.19
53.94
56.48
115.94
112.15
124.26
162.40
112.15
124.26
120.14
112.15
124.26
174.99
112.15
124.26
-1034.79730
16.99
-25.38
-1270.64433
16.57
-31.24
-463.35714
18.25
-13.72
-699.17240
19.21
-18.97
0
0
TS(2a-6a)
6a
2b
TS(2b-6b)
6b
2c
TS(2c-6c)
6c
2d
TS(2d-6d)
6d
1.44
0.01
0.00
1.44
0.00
0.00
1.44
0.01
0.00
1.44
-17.57
-12.38
-8.03
-14.47
-7.14
-15.93
-15.02
-15.02
Path E
-517.35814
39.22
-19.07
-231.63857
38.49
2a
-517.41073
43.16
-18.66
-231.68862
42.07
-14.44
28.78
24.89
27.18
26.92
23.51
25.48
75.07
74.98
91.69
79.35
78.37
95.12
-517.39381
39.25
-24.02
-231.67627
38.78
-19.46
2.90
3.07
1.90
1.59
4.86
1.88
TS(2a-7a)
7a
2c
TS(2c-7c)
7c
-1.76
-14.76
-1.26
Path F
-670.18972
11.27
-14.50
-788.10079
12.72
-13.42
-384.46421
13.48
-13.62
-502.35810
14.37
2a
-670.29860
16.38
-13.98
-788.30149
18.22
-12.87
-384.57074
19.16
-13.31
-502.55841
20.42
-12.07
60.84
56.71
59.64
115.02
110.53
113.86
59.16
54.53
58.39
114.34
109.32
113.36
102.69
95.61
-670.23851
13.38
-16.48
-788.16224
14.81
-15.39
-384.51511
16.09
-14.62
-502.42273
16.89
-14.45
1.72
3.63
2.19
2.14
2.67
1.43
2.58
2.77
1.78
3.07
3.12
2.05
-22.81
TS(2a-8a)
8a
2b
TS(2b-8b)
8b
2c
TS(2c-8c)
8c
2d
TS(2d-8d)
8d
109.32
129.34
122.34
135.95
107.13
98.39
110.47
136.04
127.61
142.53
-21.63
-22.64
-20.90
-22.44
-22.07
-22.71
-12.51
-21.89
a The energies, enthalpies, free energies, and binding energies (B.E. ) basis set superposition error is corrected) are given in kcal mol-1. The entropy is
in e.u., and the dipole moment is in debye. All calculations were performed with B3LYP, and the 6-31G(d,p) basis set was used for all atoms except Ge,
which was treated with the SDD pseudopotential. For details of the calculations, see the Supporting Information.
9
J. AM. CHEM. SOC. VOL. 128, NO. 12, 2006 4099

A R T I C L E S
Cerkovnik et al.
Conclusions
Competing with this path is the cleavage of the R3E-OOOH
bond to yield R3EOH and HOOOH, which is also assisted by
two molecules of water. Because water was always present in
the reaction mixture, this mechanistic pathway appears to be
more important in acetone-d6 (and methyl acetate or tert-butyl
methyl ether) than in methylene chloride-d2, due to the different
solubility of water in these solvents. A relatively high activation
enthalpy of the reactions, as compared to the experimentally
obtained values, appears to indicate that even greater assemblies
of water molecules (most likely assisted by one or more
molecules of R3EOH, and perhaps acetone as well) might be
involved in these decompositions.
The formation of dihydrogen trioxide (HOOOH) in the
decomposition of organometallic hydrotrioxides in acetone-d6
represents a new and convenient method for the preparation of
this simplest of polyoxides. When run in methylene chloride
as solvent, these reactions (for example, the decomposition of
Et3SiOOOH) are at present the method of choice for the
generation of singlet oxygen (O2(1∆g)).
In the present investigation, we have demonstrated that the
ozonation of various silanes and germanes produced the
corresponding hydrotrioxides, R3SiOOOH and R3GeOOOH,
which were characterized by H, 13C, 17O, and 29Si NMR, and
(for the first time for any polyoxide) by IR spectroscopy.
Calculated NMR chemical shifts and characteristic IR frequen-
cies in acetone-d6 as solvent agree well with the experimentally
obtained values.
1
A theoretical investigation of the first step of the ozonation
reaction (i.e., the formation of the hydrotrioxides) revealed that
a concerted 1,3-dipolar insertion mechanism, proposed in the
past to accommodate the experimental data, is operative for all
compounds investigated in the gas phase. However, in substi-
tuted silanes and all germanes as well as in acetone-d6 solutions,
the H atom transfer increasingly precedes the E-O formation.
Considering dynamic effects (not investigated in this work), one
cannot exclude that the formation of a radical pair, R3E• •OOOH,
encompassed in a solvent cage is possible. Collapsing of the
radical pair would lead to a hydrotrioxide (R3EOOOH), whereas
diffusion out of the solvent cage would be the first step for
HOOOH generation.
The hydrotrioxides under investigation decompose in acetone-
d6, methyl acetate, and tert-butyl methyl ether into the corre-
sponding silanols/germanols, disiloxanes/digermoxanes, singlet
oxygen (O2(1∆g)), and dihydrogen trioxide (HOOOH). The
activation parameters for the decomposition of the hydrotri-
oxides, together with the magnitude of the kinetic solvent effect
indicate the involvement of water in these reactions. Several
mechanistic possibilities for the formation of all these products
were tested. Theoretical results indicate that there are at least
two competing reactions, both first-order processes, which can
accommodate all the decomposition products. The first one,
involving the cleavage of the R3EO-OOH bond with the
formation of R3EOH and O2(1∆g), is assisted by two molecules
of water. This mechanistic pathway has the lowest activation
energy and is most likely operative in “inert” solvents, such as
CH2Cl2, in which water is only sparingly soluble.
Acknowledgment. We dedicate this work to the memory of
Professor Christopher (“Chris”) S. Foote, one of the pioneers
in the field of singlet oxygen chemistry. At Ljubljana, this work
was financially supported by the Slovenian Research Agency.
D.C. and E.K. thank the University of the Pacific for generous
support of their work. J.C. and B.P. thank Dr. Janez Plavec
(Slovenian NMR Centre, National Institute of Chemistry,
Ljubljana) for running 17O NMR spectra on the 600 MHz Varian
spectrometer, and Prof. Janez Levec for enabling the low-
temperature IR measurements.
Supporting Information Available: Instrumentation, materi-
als, ozonation procedures, product analysis, methodology of
kinetic studies, energy data, and the analysis of the calculated
and measured NMR and IR spectra are provided (38 pages,
PDF). The geometries of all calculated structures, reported in
this work, are also available in xyz format (TXT). This material
is available free of charge via the Internet at http://pubs.acs.org.
JA058065V
9
4100 J. AM. CHEM. SOC. VOL. 128, NO. 12, 2006