Steroids as photonic wires. Z → E olefin photoisomerization involving ketone singlet → triplet switches by through-bond energy transfer

Article abstract of DOI:10.1021/ja963261a

A series of steroids has been studied in which there is a dimethylphenylsiloxy (DPSO) group at the 3 position to serve as a light-absorbing antenna, a ketone group at C6 or C11 to serve as an energy 'relay' or as a 'singlet-triplet switch', and an olefin

Full text of DOI:10.1021/ja963261a

J. Am. Chem. Soc. 1997, 119, 7945-7953
7945
Steroids as Photonic Wires. Z f E Olefin Photoisomerization
Involving Ketone Singlet f Triplet Switches by Through-Bond
Energy Transfer1,2
Joseph K. Agyin, Larry D. Timberlake, and Harry Morrison*
Contribution from the Department of Chemistry, Purdue UniVersity,
West Lafayette, Indiana 47907
ReceiVed September 16, 1996X
Abstract: A series of steroids has been studied in which there is a dimethylphenylsiloxy (DPSO) group at the 3
position to serve as a light-absorbing antenna, a ketone group at C6 or C11 to serve as an energy “relay” or as a
“singlet-triplet switch”, and an olefin at C17 to serve as a triplet energy acceptor. These include 3R-
(dimethylphenylsiloxy)-17-(Z)-ethylidene-5R-androstan-11-one (2), its C6 carbonyl analog (3), and the C6 ketone
3â isomer (4). The nonketonic steroid 3R-(dimethylphenylsiloxy)-17-(Z)-ethylidene-5R-androstane (1) serves as a
reference. Excitation of all four compounds with light absorbed by the DPSO chromophore leads to Z f E
isomerization of the C17 ethylidene group. For the ketonic steroids this isomerization involves intramolecular singlet-
singlet energy transfer (intraSSET) from the DPSO group to the carbonyl group, intersystem crossing to the carbonyl
triplet, and intramolecular triplet-triplet energy transfer (intraTTET) to the alkene. For compound 1 there is modestly
efficient (through-bond) intraTTET directly from C3 to C17. For 2-4 intraSSET is ca. 75-90% efficient and occurs
with rate constants of (1.1-1.7) × 109 s-1. The C3 DPSO antenna transfers both singlet and triplet energy more
efficiently when it is R (axial) than when it is â (equatorial). IntraTTET from C6 and C11 to C17 is ca. 80%
efficient; kC6f17Z, determined from triplet quenching experiments, equals 8.3 × 108 s-1. The C11 ketone has an
unusually short singlet lifetime (0.4 ns), which can be attributed to an enhanced rate of radiationless decay caused
by the proximity of the C19 angular methyl group.
Introduction
properties as outlined earlier)9 and the ketone functionality as
the potential recipient of intramolecular singlet-singlet and
triplet-triplet energy (intraSSET and intraTTET, respectively).
The antenna and the ketone have been mounted on the 5R-
androstane steroidal framework. We have shown that excitation
of the DPSO chromophore in 3R-(dimethylphenylsiloxy)-5R-
androstane-11,17-dione (3RDPSO/11/17) leads to triplet pho-
As part of the surge of interest in the general area of
nanotechnology3 there has been developing attention to the
potential use of molecular dimension “wires” to transmit
photonic energy.4 Examples of frameworks shown to be capable
of through-bond transmission of electronic excitation include
bicyclo[2.2.1]heptanes,5a-c bicyclo[2.2.2]octanes,5d oligo[1.1.1]-
propellanes,5e,f polyenes and polyynes,5g,h polyphenylenes,5i
oligothiophenes,5j polypeptides,5k-m polyethers,5n and block
polymers.5o,6 Steroids represent an additional group of transmit-
ters, with recent work from our own1,2,7 and other laboratories8
providing ample confirmation that these rigid tetracyclics are
excellent “photonic wires”. These molecules provide a well-
defined scaffold upon (and within) which one can insert different
moieties to fulfill various electronic functions. The latter may
include an “antenna” to absorb incident irradiation, relays and
gates which facilitate and impede energy migration (the latter
potentially multiplicity specific),1 and multiplicity switches
which convert singlet to triplet energy.
(5) Representative examples: (a) Kroon, J.; Oliver, A. M.; Paddon-Row,
M. N.; Verhoeven, J. W. J. Am. Chem. Soc. 1990, 112, 4868-4873. (b)
Shephard, M. J.; Paddon-Row, M. N.; Jordan, K. D. J. Am. Chem. Soc.
1994, 116, 5328-5333. (c) Ren, Y.; Wang, Z.; Zhu, H.; Weininger, S. J.;
McGimpsey, W. G. J. Am. Chem. Soc. 1995, 117, 4367-4373. (d)
Zimmerman, H. E.; Goldman, T. D.; Hirzel, T. K.; Schmidt, S. P. J. Org.
Chem. 1980, 45, 3933-3951. (e) Kszynski, P.; Friedl, A. C.; Michl, J. J.
Am. Chem. Soc. 1992, 114, 601-620. (f) Paddon-Row, M. N.; Jordan, K.
D. J. Am. Chem. Soc. 1993, 115, 2952-2960. (g) Osuka, A.; Tanabe, N.;
Kawabata, S.; Yamazaki, I.; Noshimura, Y. J. Org. Chem. 1995, 60, 7177-
7185. (h) Holl, N.; Port, H.; Wolf, H. C.; Strobel, H.; Effenberger, F. Chem.
Phys. 1993, 176, 215-220. (i) Helms, A.; Heiler, D.; McLendon, G. J.
Am. Chem. Soc. 1992, 114, 6227-6228. (j) Wu¨rthner, F.; Vollmer, M. S.;
Effenberger, F.; Emele, P.; Meyer, D. U.; Port, H.; Wolf, H. C. J. Am.
Chem. Soc. 1995, 117, 8090-8099. (k) Schanze, K. S.; Sauer, K. J. Am.
Chem. Soc. 1988, 110, 1180-1186. (l) Vassilian, A.; Wishart, J. F.; van
Hemelryck, B.; Schwarz, H.; Isied, S. S. J. Am. Chem. Soc. 1990, 112,
7278-7286. (m) Voyer, N.; LaMothe, J. Tetrahedron 1995, 51, 9241-
9284. (n) Benniston, A. C.; Harriman, A.; Lynch, V. M. J. Am. Chem. Soc.
1995, 117, 5275-5291. (o) Watkins, D. M.; Fox, M. A. J. Am. Chem. Soc.
1994, 116, 6441-6442.
In our previous reports we have utilized the dimethylphe-
nylsiloxy (DPSO) group as the antenna (because of its favorable
X Abstract published in AdVance ACS Abstracts, August 15, 1997.
(1) Organic Photochemistry. Part 113. Part 112: Jiang, S. A.; Xiao, C.;
Morrison, H. J. Org. Chem. 1996, 61, 7045-7055.
(2) (a) Preliminary communication: Agyin, J. K.; Morrison, H.; Siemi-
arczuk, A. J. Am. Chem. Soc. 1995, 117, 3875-3876. (b) Abstracted from
the Doctoral Dissertation of J. K. Agyin, Purdue University, August, 1996.
(3) For an overview, see: Tolles, W. M. In Nanotechnology. Molecularly
Designed Materials; Chow, G.-M., Gonsalves, K. E., Eds.; ACS Symp. Ser.
No. 622; American Chemical Society: Washington, DC, 1996; Chapter 1,
pp 1-19. See also: Balzani, V.; Scandola, F. Supramolecular Photochem-
istry; Ellis Horwood: Chichester, Great Britain, 1991.
(4) Examples with leading references: Wagner, R. A.; Lindsey, J. S.;
Seth, J.; Palaniappan, V.; Bocian, D. F. J. Am. Chem. Soc. 1996, 118, 3996-
3997. Zhou, Q.; Swager, T. M. J. Am. Chem. Soc. 1995, 117, 12593-
12602.
(6) Recent reviews see: Paddon-Row, M. N.; Jordan, K. D. In Modern
Models of Bonding Delocalization; Liebman, J. F., Greenberg, A., Eds.;
VCH: New York, 1988; Vol. 6, ,Chapter 3. Jordan, K. D.; Paddon-Row,
M. N. Chem. ReV. 1992, 92, 395-410. Verhoeven, J. W.; Kroon, J.; Paddon-
Row, M. N.; Warman, J. M. In Supramolecular Chemistry; Balzani, V.,
De Cola, L., Eds.; Kluwer Academic Publishers: Dordrecht, 1992; pp 181-
200.
(7) (a) Wu, Z.-Z.; Morrison, H. J. Am. Chem. Soc. 1989, 111, 9267-
9269. (b) Wu, Z.-Z.; Morrison, H. Tetrahedron Lett. 1990, 31, 5865-5868.
(c) Wu, Z.-Z.; Morrison, H. J. Am. Chem. Soc. 1992, 114, 4119-4128. (d)
Wu, Z.-Z.; Nash, J. J.; Morrison, H. J. Am. Chem. Soc. 1992, 114, 6640-
6648.
S0002-7863(96)03261-1 CCC: $14.00 © 1997 American Chemical Society

7946 J. Am. Chem. Soc., Vol. 119, No. 34, 1997
Agyin et al.
tochemistry at C17, although the C17 ketone itself intersystem
crosses with very low efficiency; by contrast, direct photolysis
of the carbonyl groups in this steroid primarily leads to singlet-
derived epimerization in ring D (cf. eq 1).7c We proposed that
Scheme 1
OAc
OH
1. NBS, (PhCO)2O2, CCl4
2. conc. HCl, MeOH
3. p-TsOH
O
O
O
O/p-TsOH
Et
O
Et
O
OH
activation of C17 in 3RDPSO/11/17 occurs by intraSSET from
the DPSO chromophore to C11, intersystem crossing of this
ketone, and intraTTET from C11 to C17. To verify and expand
this proposed role for the C11 carbonyl group as a “singlet f
triplet switch”, we have investigated energy migration in the
trifunctional steroids 3R-(dimethylphenylsiloxy)-17-(Z)-ethylidene-
5R-androstan-11-one (2), its C6 carbonyl analog (3), and the
C6 ketone 3â isomer (4), (3RDPSO/11/17-Z, 3RDPSO/6/17-Z,
3âDPSO/6/17-Z, respectively). 3R-(Dimethylphenylsiloxy)-17-
(Z)-ethylidene-5R-androstane (3RDPSO/17-Z, 1) serves as a
nonketonic reference. Compounds 2-4 were chosen to remove
any ambiguity as to the source of excitation on the ultimate
ring D acceptor, i.e. by contrast with a ketone, an olefin can
only be activated by ketone or aryl donors through intraTTET.
Our results demonstrate that excitation of the DPSO antenna
with 266-nm light does indeed result in enhanced Z f E
isomerization of the C17 ethylidene group in the ketonic
steroids, thus confirming our proposal of the ketone as a
multiplicity switch.
RuO4, CCl4
O
O
O
O
O
O
O
O
Ph3P(CH2CH3)Br,
KOtBu, THF
PPTS, wet acetone
O
O
O
O
O
O
K-Selectride
DPSCl, Et3N, DMF
3 + 4
Results
HO
Syntheses of the DPSO-Derivatized Steroids 1-4. The
C17 olefin, 3R-hydroxy-17-(Z)-ethylidene-5R-androstane, was
prepared from androsterone by a Wittig reaction and silylated
O
with chlorodimethylphenylsilane (DPSCl) to afford 1. The
reaction produced predominantly the Z isomer (97%), consistent
with literature precedent10 and with an X-ray analysis of 3 (see
below). Epiandrosterone was used to prepare the 3â isomer of
1 (as the Z olefin) in like manner. The alcohol precursor for 2
(3R-hydroxy-5R-androstan-11,17-dione) was prepared by olefin
reduction of ∆4-androstene-3,11,17-trione with lithium in am-
monia followed by further regiospecific reduction at C3 with
K-Selectride (the 3â hydrogen is evident as a singlet at δ 4.00;
the C17 and C11 carbonyl carbons are seen as resonances at δ
217.81 and 209.32, respectively). A subsequent Wittig reaction
is specific to C17 due to the steric encumbrance of C11. Again,
the Z isomer predominates (97%). Compounds 3 and 4 were
prepared from the commercially available testosterone acetate
in 8 steps (Scheme 1). The two isomers were separated by flash
chromatography. The structure assignment for 3 was confirmed
by X-ray analysis (the X-ray structure and a table summarizing
the crystal data and data collection parameters are provided as
Supporting Information).
(8) (a) Zhu, Y.; Schuster, G. B. J. Am. Chem. Soc. 1993, 115, 2190-
2199. (b) Cvitas, Y.; Kovac, B.; Pasa-Tolic, L.; Ruscic, B.; Klasinc, L.;
Knop, J. V.; Bhacca, N. S.; McGlynn, S. P. Pure Appl. Chem. 1989, 61,
2139-2150. (c) Klasinc, L.; Pasa-Tolic, Lj.; Spiegl, H.; Knop, J. V.;
McGlynn, S. P. Int. J. Quantum Chem.: Quantum Biol. Symp. 1993, 20,
191-198. (d) Tong, Z.-H.; Yang, G.-Q.; Wu, S.-K. Acta Chim. Sin. (Engl.
Ed.) 1989, 450. Tong, Z.-H.; Yang, G.-Q.; Wu, S.-K. Chinese J. Chem.
1990, 61. For related studies involving electron transfer, see: Closs, G. L.
J. Phys. Chem. 1986, 90, 3673-3683. Closs, G. L.; Piotrowiak, P.; Miller,
J. R. In Photochemical Energy ConVersion; Norris, J. R., Jr., Meisel, D.,
Eds.; Elsevier: New York, 1989; p 23. Liang, N.; Miller, J. R.; Closs, G.
L. J. Am. Chem. Soc. 1990, 112, 5353-5354. The seminal papers involving
the study of energy transfer in steroids are the following: Latt, S. A.;
Cheung, H. T.; Blout, E. R. J. Am. Chem. Soc. 1965, 87, 995-1003. Keller,
R. A.; Dolby, L. J. J. Am. Chem. Soc. 1967, 89, 2768-2770. Keller, R. A.;
Dolby, L. J. J. Am. Chem. Soc. 1969, 91, 1293-1299.
Spectroscopy. The UV absorption spectra of 2-4 in
cyclohexane reflect the component aryl (λmax ca. 258 nm) and
carbonyl (λmax 290-300 nm) chromophores and show no
evidence of significant interaction between them. Compound
2 shows only aryl emission. Compounds 3 and 4 show dual
emission with the aryl and ketone fluorescence maxima at ca.
285 and 430 nm, respectively. Compound 3 shows significantly
(10) (a) Wovkulich, P. M.; Barcelos, F.; Batcho, A. D.; Sereno, J. F.;
Baggliolini, E. G.; Hennessy, B. M.; Uskokovic, M. R. Tetrahedron 1984,
40, 2283-2296. (b) Baggiolini, E. G.; Iacobelli, J. A.; Hennessy, B. M.;
Uskokovic, M. R. J. Am. Chem. Soc. 1982, 104, 2945-2949.
(9) Morrison, H.; Agyin, K.; Jiang, A.; Xiao, C. Pure Appl. Chem. 1995,
67, 111-116.

Steroids as Photonic Wires
J. Am. Chem. Soc., Vol. 119, No. 34, 1997 7947
Table 3. Estimated Singlet and Triplet Energiesa
DPSO C6/C11 ketones
107 84
C17 olefin
ES
ET
135b
77c
84 76
a Units of kcal/mol. b Estimated from ref 13a. c Estimated from ref
13b.
Photochemistry. Photolysis of 2 in Cyclohexane. Irradia-
tion of an argon-degassed 1.1 × 10-2 M solution of 2 in
cyclohexane at 254 nm produced a single product (39.5%) (eq
2). The photoproduct has a GLC retention time identical with
Figure 1. Normalized fluorescence spectra for 3 and 4 in cyclohexane.
Table 1. Fluorescence Quantum Efficiencies (φf) and Lifetimes
(τf) for Compounds 1, 2, 3, and 4 Using 254-nm Excitationa,b
compd
φf
τf (ns)
that of the small amount of E isomer formed as a side product
1
2
3
4
0.013
2.63
0.55
0.33
0.67
in the synthesis of 2. GC/MS with EI on the photolysis solution
0.0024
0.0010
0.0027
1
gave a parent ion at m/e 450. The H NMR spectrum of the
reaction mixture further confirmed the presence of the E isomer
with resonances at δ 5.02 (1 H, q) corresponding to the H-20
vinyl proton and at δ 1.53 (3 H, d, J ) 6 Hz) for the allylic
H-21 protons. The corresponding resonances for the Z isomer
are at δ 5.11 and 1.66, respectively. The quantum efficiency
for formation of the E isomer (φZfE) was determined at 266
nm with a Nd:YAG laser and found to be 0.035.
In a separate photolysis with excitation at 308 nm with an
excimer laser, the 17-E isomer was again observed as the only
photoproduct. The quantum efficiency for formation of the E
isomer at 308 nm was determined to be 0.041. The photosta-
tionary state was determined in cyclohexane with 254-nm light.
The [E]pss/[Z]pss ratio is 1.38.
a Solutions in cyclohexane with toluene emission in cyclohexane as
the reference.11 b φf and τf values estimated to be (10% and (0.1 ns,
respectively.
Table 2. Fluorescence Quantum Efficiencies and Lifetimes Using
300-nm Excitationa
compd
φf(300 nm)
τf
3
4
1.82 × 10-3
1.49 × 10-3
1.43 × 10-3
4.5 × 10-4
4.0 ( 0.2
3.6 ( 0.1
2.9 ( 0.1
0.4 ( 0.1
2-methylcyclohexanone
3,17-diethylenedioxy-5R-
androstan-11-one (5)
2
none detected
Photolysis of 2 at 308 nm in the Presence of Triethylamine
(TEA). It has been found that the reduction of the carbonyl
group by TEA provides an effective means of testing for the
creation, and involvement, of ketone triplet states in our steroid
studies.7c,d Photolysis of an argon degassed cyclohexane
solution of 1.1 × 10-2 M 2 and 12.6 × 10-2 M TEA with 308-
nm light for 20 min showed a minimal effect of the amine, i.e.
5.8 and 6.3% of Z f E photoisomerization was observed with
and without TEA, respectively.
After prolonged photolysis (190 min) a mixture of two new
products was formed (6.6% by GLC). Replacing TEA by tert-
butyl alcohol in an analogous experiment gave 5.3% of the new
product mixture after 190 min. The products were isolated by
flash chromatography and identified as a 2:1 mixture of the E/Z
cyclobutanol isomers 6 and 7 (cf. eq 3). Though the two
a Solutions in cyclohexane; acetone emission (φf ) 9.3 × 10-4) as
the reference.12
more ketone emission than compound 4, and its aryl fluores-
cence is considerably diminished relative to 4 (cf. Figure 1).
The 0-0 transition energy for the DPSO S0 f S1 transition is
estimated from the onset at 270 nm to be 107 kcal/mol.
Fluorescence quantum efficiencies (φf) and lifetimes (τf) for
compounds 1-4 were determined in cyclohexane at room
temperature by using excitation of the DPSO antenna (254 nm)
with toluene emission in cyclohexane (φf ) 0.14)11 as the
reference. The data are presented in Table 1.
Fluorescence quantum efficiencies and lifetimes were also
measured for compounds 3, 4, 3,17-diethylenedioxy-5R-an-
drostan-11-one (5), and 2-methylcyclohexanone in cyclohexane,
using 300-nm excitation and with acetone emission in cyclo-
hexane (φf ) 9.3 × 10-4 12
)
as the reference. The results are
summarized in Table 2.
Phosphorescence spectra for compounds 1 and 2 were
obtained at 77 K in an ether/methylcyclohexane glass. An
emission onset at 340 nm provided a triplet energy for the DPSO
group of ca. ET ) 84 kcal/mol.
Table 3 summarizes the relevant singlet and triplet energies
of the chromophores by using literature data (including our
previous paper) and the current work.
products could not be isolated in pure form, the 1H NMR
spectrum of the mixture gave evidence of the cyclization through
the absence of the characteristic resonance for the C19 angular
methyl group. The C20 vinyl proton resonances at δ 5.07 (6)
and 4.99 (7) and the C18 resonances at δ 0.99 (6) and 0.88 (7)
were still evident. The IR spectrum of the mixture showed a
broad peak at 2900 cm-1 indicative of the presence of an
alcohol.
(11) Birks, J. B. Photophysics of Aromatic Molecules; John Wiley and
Sons, Ltd.: New York, 1970; p 126.
(12) Halpern, A. M.; Ware, W. R. J. Chem. Phys. 1971, 54, 1271-1276.
(13) (a) Cf. The S1 energy of 2-butene in: Murov, S. L. Handbook of
Photochemistry; Marcel Dekker: New York, 1973; p 4. (b) Cf. Triplet
energies for 2-butene and 2,3-dimethyl-2-butene in: Turro, N. J. Modern
Molecular Photochemistry; University Science Books: Mill Valley, CA,
1991; p 292.
Photolysis of 1. Irradiation of an argon-degassed cyclohex-
ane solution of 1 at 254 nm produced a single photoproduct

7948 J. Am. Chem. Soc., Vol. 119, No. 34, 1997
(cf. eq 4).
1H NMR spectrum of the product indicated
Agyin et al.
Table 4. Efficiencies and Rate Constants for Intramolecular
A
Singlet-Singlet Energy Transfer
from φf
from τf
kintraSSET
kintraSSET
compd
φintraSSET
(×109 s-1
)
φintraSSET
(×109 s-1
)
2
3
4
0.82
0.92
0.79
1.5
2.8
1.2
0.79
0.88
0.75
1.4
2.7
1.1
primary changes at the vinyl proton (to δ 5.02) and allylic
methyl (to δ 1.54) resonances. Assignment as the 3RDPSO/
17-E isomer was confirmed by comparison of the mass spectrum
with an independently synthesized sample, prepared by silylation
after p-tert-butyltoluene sensitized photolysis of the alcohol
precursor of 1. The quantum efficiency for Z f E isomerization
of 1 was determined to be 0.043 with use of 266-nm light.
Discussion
Photochemistry. As shown in eqs 2-4, excitation of the
antenna chromophore in each of the C3 DPSO steroids results
in Z f E olefin isomerization at C17. The lack of quenching
by external cis-2-heptene confirmed that energy transfer in these
steroids is intramolecular, i.e. as with earlier cases, energy is
being transmitted from the antenna in the A ring to the olefin
in the ring D. Under normal conditions there is no observable
chemistry at the C11 carbonyl group in 2 or at the C6 carbonyl
in 3. However, prolonged photolysis of 2 in cyclohexane
containing TEA or tert-butyl alcohol produces, in addition to
Z f E olefin isomerization, ca. 9% of a 2:1 mixture of Z and
E isomers of product attributable to the C11 cyclobutanol.
Energy Migration in Compounds 2, 3, and 4. Intramo-
lecular Singlet/Singlet Energy Transfer (intraSSET). It is
clear from the fluorescence data in Table 1 that there is highly
efficient intraSSET from the DPSO group to the carbonyl groups
in compounds 2-4. As one would expect, the aryl fluorescence
efficiency (0.013) and lifetime (2.6 ns) in the bifunctional
substrate, 1, are typical of those observed for monofunctional
DPSO substrates.7 The presence of a carbonyl group at C11
(2) reduces the aryl φf to 0.0024, and the more proximal ketone
at C6 further diminishes φf to 0.001. There is a corresponding
reduction in the DPSO singlet lifetime to 0.55 and 0.33 ns,
Photolysis of 3âDPSO/17-Z. Photolysis of the 3â isomer
of 1 with 254-nm light produced a single product that was
identified as 3âDPSO/17-E by coinjection with a synthetic
mixture of the two olefin isomers or with an independently
prepared sample by using toluene sensitization of the precursor
alcohol. The quantum efficiency for isomerization was deter-
mined by using 1 as a secondary actinometer and found to be
0.036. The 17-E isomer can also be formed by toluene
sensitization of the 3âDPSO/17-Z substrate itself.
Photolysis of 3. Irradiation of an argon-degassed cyclohex-
ane solution of 3 at 254 nm produced the 17-E isomer as a
1
single product (cf. eq 5). A H NMR spectrum of a photosta-
respectively. The quantum efficiency for intraSSET, φintraSSET
-
(i), from the DPSO singlet to the ketone group in a substrate i
can be calculated from either the φf or τf values by using the
fluorescence data for 1 as a reference for the intrinsic photo-
physical properties of a DPSO group (cf. eqs 6 and 7).7 The
corresponding rate constants (kintraSSET(i)) can then be calculated
from the usual expression, φintraSSET ) kintraSSETτf.14 The
calculated values are given in Table 4.
tionary state mixture of the two olefins showed new resonances
at δ 5.02 and 1.52 for the H-20 vinyl hydrogen and the allylic
methyl group, respectively. The quantum efficiency for isomer-
ization at 266 nm was found to be 0.36. The isomerization
quantum efficiency was also measured at 308 nm and deter-
mined to be 0.46. The isomerization quantum efficiency at 254
nm was also determined for the 3â isomer, 4, and found to be
0.23. The photostationary state was determined in cyclohexane
with 266-nm light and found to be [E]pss/[Z]pss ) 1.08. An
analogous experiment with 300-nm light gave a ratio of 1.13.
φintraSSET(i) ) [φf(1) - φf(i)]/φf(1)
φintraSSET(i) ) [τf(1) - τf(i)]/τr(1)
(6)
(7)
These results are consistent with our earlier report of φintraSSET
) 0.68 for energy transfer from the C3 DPSO antenna to the
C11 carbonyl in 3R-DPSO-17-â-hydroxy-5R-androstan-11-one
and 0.73 for φintraSSET from the C3 DPSO group to a C6
ketone.7c,d Likewise, the rate constants in Table 4 match well
with the values of 2.2 × 109 and 1.7 × 109 s-1 we have
previously observed for C3 DPSO energy transfer to C11 and
C6, respectively.7c,d
The energy transfer parameters for C3 f C6 in the 3â DPSO
isomer, 4, are also given in Table 4 and clearly indicate that, as
a donor, the 3â antenna is some 15% less efficient and greater
than 2-fold slower relative to the 3R isomer (3). These relative
energy transfer efficiencies are reflected in the quantum ef-
ficiencies for antenna initiated Z f E olefin isomerization in 3
vs 4, 0.36 and 0.23 respectively. We have seen comparable
Effect of Added cis-2-Heptene on the Photoisomerization
of 1 and 3. cis-2-Heptene was used to test for the possible
involvement of intermolecular TTET by noting whether the
added olefin quenched Z f E isomerization. For neither 1 nor
3 was any quenching observed when comparable concentrations
of the olefin were present. Thus, for Z f E isomerization in
the presence and absence of the heptene, at ca. 1 × 10-2 M,
concentrations of steroid and olefin were as follows: 1, 16.9
and 16.7%; 3, 36.1 and 36.5%, respectively.
Effect of Added Triethylamine (TEA) on Z f E Photoi-
somerization of 3. In a preliminary experiment 12.6 × 10-2
M TEA was added to 1.02 × 10-2 M 3 in cyclohexane and the
solution photolyzed with 266-nm light. The extent of isomer-
ization was reduced from 16.2% without the amine to 7.7%
with TEA. The quenching was studied in greater detail by using
0.032-0.126 M TEA with 1.02 × 10-2 M 3. A Stern-Volmer
plot of the data gave a slope and intercept of kqτ ) 8.5 M-1
and 1.0, respectively, with a correlation coefficient of 0.995.
(14) We note that there is an error in eq 9 of ref 7c. The denominator
has the labels interchanged; the denominator should read as the multiple of
φf(6)τf(i). Rate constants in parentheses in Table VII for compounds 3, 4,
and 1 in this reference are 1.2, 19, and 25 (×108 s-1), respectively.

Steroids as Photonic Wires
J. Am. Chem. Soc., Vol. 119, No. 34, 1997 7949
Scheme 2
O
O
1O*
intra-SSET
hν/266 nm
*
DPSO
1DPSO
DPSO
isc
3
*
O
O
3O*
intra-TTET
Z
E
DPSO
DPSO
DPSO
levels of reduced intraSSET efficiency for the equatorial 3â
antenna in previous studies involving DPSO sensitization of C6
and C17 ketones.7c,d Our observation that intraTTET from the
3R DPSO group to the C17 olefin in 1 leads to a φZfE of 0.043
vs a value of 0.036 for the 3â isomer further demonstrates the
“3â/3R effect” (see additional discussion below).
transfer observed by olefin photoisomerization is also consis-
tent with the observation of significant DPSO phosphorescence
from 1.
The C6 and C11 Carbonyl Groups as Singlet f Triplet
Switches. As noted in the Introduction, a primary objective
for the current study was to confirm our proposal that a C11
ketone can function as a singlet-triplet switch. This must
certainly be the case, though at first glance this may not be
obvious from the DPSO sensitized olefin isomerization efficien-
cies, φZfE: 0.035 for 2 (with a C11 ketone) vs 0.043 for 1 (no
ketone). Recall, however, that intraSSET from the DPSO group
to the carbonyl group in 2 is highly efficient (ca. 80%, cf. Table
4). Clearly, olefin isomerization could only be occurring in 1
and 2 at a comparable efficiency if a portion of the DPSO singlet
energy transferred to C11 ends up as triplet energy at C17.
The argument is even more cogent for compounds 3 and 4,
where the insertion of a C6 carbonyl group raises φ(ZfE) with
266-nm excitation to 0.36 and 0.23, respectively, several fold
higher than the value of 0.043 without a ketone present. For
these compounds, the efficiencies of intraSSET from the C3
DPSO group to C6 are as high as ca. 90% (Table 4). The
complete mechanism (illustrated for antenna initiated isomer-
ization in 2) is presented in Scheme 2.
We note that these observations are contrary to the general
conclusions of Closs et al.,8a-c,15 using cyclohexane and trans-
decalin as spacers and the 2-naphthyl and 4-benzophenonyl
groups as intraTTET acceptor and donor, respectively. In these
studies, groups occupying an equatorial position were found to
be more efficient as energy transfer donors. In the Closs
systems the triplet energy migration proceeds through a Dexter-
exchange, through-bond mechanism.16 This is likewise the only
feasible mechanism for migration of triplet energy from C3 to
C17 in 1, where the interchromophore distance (i.e., silicon
bonded aryl carbon to C17) from X-ray analysis of the crystal
is 9.85 Å.17 However, intraSSET from an aryl donor to a ketone
acceptor can proceed through both the exchange and resonance
energy transfer mechanisms.7c,d The possibility that the latter
participates in, e.g., 3 and 4, and therefore contributes a through-
space component to the energy migration process, prevents us
from rationalizing the 3R/3â ratio in 3 and 4 solely in terms of
a TBI interaction. Studies are in progress which will address
the equatorial/axial issue for intraSSET in the steroid wire.
Intramolecular Triplet/Triplet Energy Transfer. Energy
Migration from C3 to C17. The DPSO initiated isomerization
of the C17 ethylidene group is unambiguously a consequence
It is possible to calculate the efficiency by which singlet
energy at C6 or C11 ultimately surfaces as triplet energy at
C17 by using eq 9. Here φCnf17Z (Cn ) C6 or C11) actually
φZfE(266 nm) ) φDPSOfCnφCnf17ZFDfE
(9)
of intraTTET. The efficiency of energy migration (φintraTTET
)
from the C3 DPSO group to C17 can be calculated with eq 8.
represents the product of φisc for the C6 or C11 ketones ×
φintraTTET (this is discussed further below). The quantum
efficiencies for intraSSET (φDPSOfCn), and for the overall
isomerization (φZfE(266 nm)), have been previously discussed.
This missing parameter is the fraction of the relaxed olefin triplet
diradical that decays to the E isomer (FDfE). This value is
obtainable from the photostationary state by using eq 10. For
φZfE ) φiscφintraTTETFDfE
(8)
Thus, we have elsewhere shown that φisc for a steroidal DPSO
chromophore is 0.19.7c If one assumes a decay ratio for the
orthogonal alkylidene triplet (“D”) such that FDfE ) 0.58 (as
it does in 2 (see below)), then a φZfE of 0.043 for 1 leads to a
calculated value for φintraTTET of 0.39. This is similar to the
value of 0.21 measured earlier for φintraSSET from the R C3 DPSO
donor to a C17 ketone.7c,18 The less than complete triplet energy
[E]pss/[Z]pss ) (φCnf17Z)/(φCnf17E)(FDfE/FDfZ
) (10)
compound 2, the [E]pss/[Z]pss was found to be 1.38. Assuming
that (φC11f17Z)/(φC11f17E) ) 1, and that FDfE + FDfZ ) 1, the
calculated value of FDfE was 0.58. Using an averaged value
for φDPSOfC11 of 0.80 (cf. Table 4), one calculates an oVerall
efficiency for energy transfer from C11 to C17 of 0.075. (Note
that this value would decrease slightly to the extent that there
is direct TTET by the residual DPSO triplets formed by the
short-lived DPSO singlet.)
(15) Closs, G. L.; Piotrowiak, P.; MacInnis, J. M.; Fleming, G. R. J.
Am. Chem. Soc. 1988, 110, 2652-2653. Closs, G. L.; Johnson, M. D.;
Miller, J. R.; Piotrowiak, P. J. Am. Chem. Soc. 1989, 111, 3751-3753.
(16) Dexter, D. L. J. Chem. Phys. 1953, 21, 836-850. Katz, J. L.; Jortner,
J.; Choi, S. I.; Rice, S. A. J. Chem. Phys. 1963, 39, 1897-1900.
(17) The X-ray analysis of 3 shows the R DPSO aromatic ring to be
tucked under the steroid skeleton. This is in contrast to what was observed
in the X-ray structure of the analogous 3R DPSO C17 ketone, where the
DPSO group extends away from the steroid skeleton, and is therefore 12.85
Å from C17.7d
One can do the same analysis by using the data for direct
excitation of the ketone group with 308-nm light. Equation 9
then simplifies to eq 11. Insertion of the measured φZfE(308
(18) We have elsewhere shown that singlet energy transfer from a C3
ketone to a C17 ketone is highly efficient.1

7950 J. Am. Chem. Soc., Vol. 119, No. 34, 1997
φZfE(308 nm) ) φC11f17ZFDfE
Agyin et al.
are effectively deactivated by C-H bonds, an observation
explained by the proposal that the reactants return to starting
materials via an avoided crossing on the reaction surface which
precedes the actual formation of a product radical pair.24
With these considerations in mind, we revisit the value for
φCnf17E, which we noted above represents the product of
intersystem crossing at C6 or C11 and the efficiency of triplet-
triplet energy transfer to C17 (φintraTTET). For C6, 77-87% of
the singlet excitation at C3 ends up as triplet energy at C17,
i.e., if φisc at C6 is essentially unity, these percentages are truly
the corresponding values of φintraTTET. However, the shortened
singlet lifetime at C11 clearly will impact on the triplet yield at
this position. An inefficient φisc would then be reflected in the
7% calculated overall efficiency of energy transferred and
φintraTTET might also be quite high for C11 to C17. In fact, if
one uses the φZfE of 0.041 observed for direct excitation of 2
with 308-nm light and assumes that φintraTTET ) 1.0,25 one can
use eq 8 to calculate a φisc for the C11 ketone of 0.07. This,
combined with a plausible value for kisc of 2.5 × 108 to 3.0 ×
108 s-1 (values obtained from the data for 3 and 4 in Table 2
assuming isc occurs in these compounds with unit efficiency),
affords a singlet lifetime of 0.19-0.28 ns. Such a short lifetime
would account for the failure to detect ketone fluorescence from
2 upon 300-nm irradiation. We thus conclude that C11, like
C6, efficiently transfers triplet energy to an olefin at C17, but
the oVerall extent of energy transferred to the olefin suffers from
a reduced triplet yield at C11.
Conclusions. The trifunctional steroids 3RDPSO/6/17-Z,
3âDPSO/6/17-Z, and 3RDPSO/11/17-Z exhibit a rich array of
photophysical and photochemical properties as a consequence
of intramolecular singlet and triplet energy transfer. Of
particular interest are the activation of the C17 ethylidene group
through TBI-mediated triplet-triplet energy transfer and the
ability of the C6 and C11 keto groups to act as “singlet f triplet
switches” in relaying excitation energy from the DPSO antenna
to the C17 alkene. This work, together with our previous
reports,1,7d shows that groups at C6 can act as efficient (1) donors
or relays of singlet energy, (2) singlet-triplet switches, and (3)
transmitters of triplet energy. By contrast, the C11 ketone
excited singlet state exhibits rapid radiationless decay, which
prevents its function as a singlet energy relay and reduces its
efficiency to act as a singlet-triplet switch + triplet energy
relay.26
(11)
nm) ) 0.041 yields a value for energy transfer from C11 to
C17 of 0.071, consistent with the number calculated from the
266-nm data. The analogous analysis for energy transfer from
C6, employing the measured [E]pss/[Z]pss of 1.08 for compound
3 (i.e. with FDfE ) 0.52), and φDPSOfC6 ) 0.90 (Table 4), gives
values of φC6f17Z of 0.77 and 0.87 (for 266 and 308 nm initiated
isomerization, respectively). These values confirm that triplet
energy transfer from C6 to C17 is highly efficient. The data
obtained from the TEA quenching of alkene isomerization in 3
allow one to calculate the rate constant for C6 to C17 triplet
energy transfer. Assuming that triplet quenching by the amine
is diffusion controlled19a,b (in cyclohexane, kdiff ) 7 × 109 M-1
s-1),19c and that the ketone triplet lifetime (1.2 ns) calculated
from the slope of the Stern-Volmer plot is entirely due to
intraTTET, the calculated rate of intraTTET, kC6f17Z, is 8.3 ×
108 s-1 20
.
This rate agrees well with the value of 1.3 × 109
s-1 determined for trans-1,4-(4-benzophenonyl-2-naphthyl)-
cyclohexane in which the donor and acceptor are also separated
by five bonds.15
The observation that triplet energy is efficiently transferred
from C6 to C17 is consistent with our earlier observation that
singlet energy also passes efficiently and rapidly between these
two positions, i.e. φintraSSET ) 0.85 and kintraSSET g 4 × 108 s-1
for transfer between the C6 and C17 carbonyl groups.7d In
effect, the C6 ketone serves as an excellent energy relay to C17
when an appropriate acceptor is at that site, and as an efficient
singlet-triplet switch/triplet donor when there is no singlet
acceptor available. The results at C11 (i.e. where φC11f17Z
)
ca. 7%) are strikingly different from that at C6, but again
consistent with our earlier studies of the C11/C17 dione.7c There
was also noted that only a small fraction (in this case, ca. 4%)
of the singlet energy transferred to C11 from the C3 DPSO
chromophore surfaced as C17 carbonyl triplet.
Some insight as to the origin of this energy loss at C11 is
provided by the data in Table 2. There we note that the
fluorescence quantum efficiencies (λexc 300 nm) are (×10-4):
18, 15, 14, and 4.5 for 3, 4, 2-methylcyclohexanone, and 5,
respectively. The corresponding lifetime data are 4.0, 3.6, 2.9,
and 0.4 ns, respectively. Clearly, both φf and τf are markedly
reduced for the ketone at C11 in 5 relative to the C6 keto steroids
and a simplified model, 2-methylcyclohexanone. We attribute
the diminution to deactivation of the C11 carbonyl singlet
excited state by the proximate, axial C19 methyl group. In fact,
the C11 carbonyl is known to form a cyclobutanol involving
hydrogen abstraction from the C19 angular methyl group,
reported, for example, upon photolysis of 5 in ethyl alcohol.21
No quantum efficiency was reported but the reaction is very
inefficient in our hands, i.e., φdis ) 0.02 in acetonitrile at 308
nm.22,23 Thus, involvement of the methyl group would have to
be primarily through an interaction with the ketone excited
singlet state that was chemically unproductive, i.e., through
reversible hydrogen abstraction or enhanced radiationless decay.
In fact, recent reports confirm that n,π* excited singlet states
A complete summary of our observations to date on the
steroid photonic wire is shown in Figure 2.
Experimental Section
Only the salient experimental features are presented here; complete
details may be found in ref 2b.
Chemicals. The following chemicals were obtained from Aldrich
Chemical Company. Except where noted, they were used as received
and stored at room temperature: adrenosterone; androsterone; chlo-
(22) Photolysis of 5 for 48 min with 300 nm light in a Rayonet Reactor,
using t-butyl alcohol or triethylamine to facilitate ring closure,23 led to a
mixture of three products by glc which we attribute to R-cleavage products.
Prolonged photolysis gave a fourth product at a longer retention time than
5 which we believe is the cyclobutanol. None of the fourth product was
observable in the absence of added TEA or t-butyl alcohol.
(23) Wagner, P. J. J. Am. Chem. Soc. 1967, 89, 5898-5901. Wagner,
P. J. Tetrahedron Lett. 1967, 1753-1756. Wagner, P. J. J. Am. Chem. Soc.
1968, 90, 5385-5388.
(24) Nau, W. M.; Cozens, F. L.; Scaiano, J. C. J. Am. Chem. Soc. 1996,
118, 2275-2282. Nau, W. M.; Cozens, F. L.; Adam, W.; Scaiano, J. C.
Book of Abstracts; XVIth IUPAC Symposium on Photochemistry, July 21-
26, 1996, Helsinki, Finland, pp 67 and 68.
(19) (a) Inbar, S.; Linschitz, H.; Cohen, S. G. J. Am. Chem. Soc. 1980,
102, 1419-1421. (b) Simon, J. D.; Peters, K. S. J. Am. Chem. Soc. 1981,
103, 6403-6406. (c) Turro, N. J. Modern Molecular Photochemistry;
University Science Books: Mill Valley, CA, 1991; p 314.
(20) A referee has noted that the comparable rate constants may be
fortuitous given the difference in triplet energies represented by the
naphthalene/benzophenone and ketone/olefin donor/acceptor pairs. Unfor-
tunately, we had insufficient material to do a similar analysis of 4, and the
classic steric impediments provided by neighboring axial methyl groups
prevent TEA quenching at the C11 ketone in 2.
(25) A referee has noted that an interaction with the angular methyl group
(21) Wehrli, H.; Heller, M. S.; Schaffner, K.; Jeger, O. HelV. Chim. Acta
1961, 44, 2162-2173. Iriate, J.; Jeger, O. HelV. Chim. Acta 1963, 46, 1599-
1609.
may diminish φintraTTET
.
(26) The Z f E quantum efficiencies reported for compounds 1, 2, and
3 correct and replace those reported in the preliminary communication.

Steroids as Photonic Wires
J. Am. Chem. Soc., Vol. 119, No. 34, 1997 7951
tive photochemical experiments employing 254- (quartz tubes) and 300-
nm (Pyrex tubes) light were performed in a Model RPR-100 Rayonet
Reactor available from Southern New England Ultraviolet Company.
The reactor was equipped with a merry-go-round turntable that
positioned photolysis tubes approximately 2 cm from the lamps.
Chromatography. Analytical GLC was performed on a capillary
gas chromatograph (Varian model 3700) equipped with a flame
ionization detector utilizing a Hewlett-Packard 3390A integrator. The
following J & W Scientific DB-1 capillary columns, 0.25 µm film
thickness, were used in this work: A, 15.0 m × 0.25 mm i.d.; B, 12.5
m × 0.25 mm i.d. Analytical thin layer chromatography utilized silica
gel 60A, 0.25 mm, coated on glass support (Whatman) visualized by
using anisaldehyde or potassium permanganate spray reagent and a
hand-held UV lamp. Flash chromatography was performed with silica
gel 60 (E. Merck 9285, 230-400 mesh). Preparative separations were
performed on a Chromatotron Model 7924T made by Harrison
Research.
Figure 2. Summary diagram showing efficient (solid arrow, >50%
ET) and inefficient (broken arrow, <50% ET) energy transfer pathways
through the steroid photonic wire. Donors for intraSSET and intraTTET
are aryl and keto chromophores. The acceptor for intraSSET is the keto
group. Acceptors for intraTTET are both the keto and alkene moieties.
rodimethylphenylsilane; epiandrosterone; ethyltriphenylphosphonium
bromide; 2-ethyl-2-methyl-1,3-dioxolane; K-Selectride, 1.0 M solution
in tetrahydrofuran (THF); 2-methylcyclo-hexanone, distilled; palladium,
10% on activated carbon; pyridinium-p-toluene sulfonic acid (PPTS);
ruthenium(IV) oxide hydrate; silica gel, Merck, grade 10181; and
sodium periodate. Testosterone acetate was from Sigma.
Syntheses. 3r-(Dimethylphenylsiloxy)-17-(Z)-ethylidene-5r-an-
drostane (1). To a slurry of ethyltriphenylphosphonium bromide (3.83
g, 10.33 mmol), dry THF (27 mL), and t-BuOK (1.2 g, 10.45 mmol)
under an atmosphere of nitrogen in a 50-mL round-bottom flask was
added 3R-hydroxy-5R-androstane-17-one (0.65 g, 2.15 mmol) dissolved
in 2.5 mL of dry THF. After the mixture was stirred for 3 h at room
temperature, additional ethyltriphenylphosphonium bromide (3.83 g,
10.33 mmol) and tBuOK (1.2 g, 10.45 mmol) were added. The mixture
was stirred at room temperature for 60 h, poured into 100 mL of ice-
water, and extracted 3 times with 25 mL of EtOAc. The combined
extracts were washed with 3 × 25 mL of water and 25 mL of brine
Spectrograde solvents were used in the photochemical and spectro-
scopic studies without further purification: acetonitrile (Fisher and
Baxter); cyclohexane (Aldrich); toluene (Fisher); acetone (Fisher).
Tetrahydrofuran (THF, Mallinckrodt) and diethyl ether (Mallinckrodt)
used in syntheses were distilled under nitrogen from sodium benzophe-
none ketyl. N,N-Dimethylformamide (Mallinckrodt) was treated with
4-Å molecular sieves followed by distillation under reduced pressure.
and dried over anhydrous Na
2SO4. Filtration, solvent removal, and
Instrumentation. Melting points were determined with a Fisher-
Johns melting point apparatus and are uncorrected. Infrared spectra
were recorded on a Perkin-Elmer Model 1800 Fourier transform infrared
spectrophotometer and a Perkin-Elmer Model 1420 ratio recording
spectrophotometer. 1H NMR spectra were obtained in CDCl3 with a
GE 300-MHz NMR spectrometer with chemical shifts recorded in ppm
relative to TMS at 0.0 ppm or relative to residual chloroform at 7.26
ppm. 13C and attached proton test (APT) spectra were also acquired
on the GE spectrometer operating at 75.6 MHz. Carbon chemical shifts
were recorded relative to CDCl3 at 77.0 ppm. Mass spectra were
recorded on Finnigan 4000 mass spectrometers, operating with a source
temperature of 250 °C, interfaced to a gas chromatograph containing
either packed or capillary columns. Electron impact (EI) and chemical
ionization (CI) mass spectra were recorded at 70 eV, the latter with
isobutane at a pressure of 0.30 Torr. High-resolution mass spectra were
recorded on a Kratos Model MS-50 instrument operated at 70 eV for
EI spectra. Ultraviolet absorption spectra were recorded on a Perkin-
Elmer model Lambda 3B spectrophotometer interfaced to a Zenith
microcomputer (Z-386/20) controlled by Perkin-Elmer computerized
spectroscopy software (PECSS). Steady state fluorescence spectra were
recorded on an SLM Aminco SPF-500C spectrofluorometer, using the
A/B mode in all experiments. Fluorescence quantum efficiencies were
obtained with reference to toluene or acetone and corrected for
differences in absorbance, detector gain, and solvent refractive index;
the spectral areas were numerically integrated by using SLM software
(“500”). Fluorescence lifetimes were obtained with PTI Model 100
or LaserStrobe lifetime spectrophotometers at room temperature with
the solutions purged with argon for at least 15 min prior to measure-
ment. Phosphorescence spectra were run in an ether-methylcyclo-
hexane glass at 77 K, which was degassed by using at least 3 freeze-
pump-thaw cycles at 10-4 Torr.
Photochemical Apparatus. All photochemical studies were con-
ducted at room temperature with solutions purged with argon for at
least 15 min prior to use. Laser photolyses at 266 nm were conducted
with a Continuum NY-61 Nd:YAG laser equipped with a frequency
quadrupler (10 Hz, 3.0-3.3 mJ/pulse). A 2× beam enlarger was used
in front of the photolysis cell to avoid cell damage. Photolyses with
308-nm light utilized a minics EX-700 Pulse Master Excimer Laser
(10 Hz, 3.5 mJ/pulse) equipped with a Pyrex glass beam splitter and a
laser beam mask with an open area 7 × 9 mm for photolysis of small
solution volumes. All sample solutions for photolysis were purged
with argon for at least 15 min prior to use. All photochemical reactions
were run at room temperature. The power meters used in the photolyses
were OPHIR Models 3A-P-CAL-S or 30A-P. Qualitative and quantita-
chromatography on silica gel (hexane-ether, 2:1) afforded 0.73 g (71%
yield) of 3R-hydroxy-17-(Z)-ethylidene-5R-androstane as white crystals
(GLC analysis of the product on column B indicated ca. 3% of the E
isomer was present), mp 158-160 °C. 1H NMR (CDCl3, 300 MHz) δ
5.12 (q, 1 H, H-20 vinyl H), 4.05 (s, 1H, 3â-H), 2.42-0.93 (m, 26 H),
0.87 (s, 3H, 19-CH3), 0.79 (s, 3H, 18-CH3); 13C NMR (CDCl3, 75.6
MHz) δ 150.61 (17-CdC), 113.26 (20-CdC), 66.70, 56.40, 54.46,
44.20, 39.22, 37.35, 36.26, 36.02, 35.16, 32.26, 31.96, 31.52, 29.14,
28.64, 24.45, 21.07, 17.01, 13.20, 11.25.
The olefinic alcohol (0.51 g, 1.69 mmol) was dissolved in 5 mL of
DMF in a 15-mL round-bottom flask under nitrogen. The flask was
flushed with nitrogen for 10 min, and freshly distilled triethylamine
(0.5 mL) was added by syringe under ice-salt cooling and constant
stirring. Chlorophenyldimethylsilane (0.34 mL, 2.02 mmol) was added
dropwise. After 1 h TLC analysis indicated that the reaction was
complete. The mixture was diluted with 20 mL of benzene and then
quickly washed with 2 × 4.0 mL of cold 5% NaHCO3, 4.0 mL of cold
2% aqueous HCl, and again with 4.0 mL of NaHCO3. The separated
organic layer was dried with anhydrous MgSO4. The crude product
was purified by silica gel chromatography (5% EtOAc/95% hexane)
to afford 0.59 g (80% yield) of the pure product as white crystals. These
were further recrystallized from acetonitrile to afford the analytically
pure title compound (GLC analysis indicated 1.1% of the E isomer to
be present), mp 88-90 °C. 1H NMR (CDCl3, 300 MHz) δ 7.63-7.37
(m, 5 H, arom), 5.14 (q, 1H, 20-vinyl H), 4.05 (s, 1H, 3â-H), 2.37-
2.15 (m, 3 H), 1.69 (d, J ) 7.2 Hz, 3H, 21-allylic CH3), 1.67-0.97
(m, 22 H), 0.89 (s, 3H, 19-CH3), 0.78 (s, 3H, 18-CH3), 0.38 (s, 3H,
SiCH3), 0.37 (s, 3H, SiCH3); 13C NMR (CDCl3, 75.6 MHz) δ 150.73,
139.27, 133.56, 129.40, 127.81, 113.25, 67.61, 56.46, 54.38, 44.54,
39.04, 37.40, 36.61, 36.14, 35.20, 32.48, 32.08, 31.62, 29.70, 28.69,
24.53, 21.12, 17.06, 13.27, 11.50, -0.77, -0.91. MS (EI) m/e 436
(M+), 269 (100); high-resolution MS (EI) m/e calcd for C29H44SiO
436.3161, found 436.3148.
3r-(Dimethylphenylsiloxy)-17-(Z)-ethylidene-5r-androstan-11-
one (2). This compound was prepared in a manner virtually identical
with that described above. The crude product was purified by silica
gel chromatography (5% EtOAc/95% hexane) to afford 2 in 69% yield
as a white solid. GLC analysis indicated 2.7% of the E isomer was
present, mp 117-119 °C. 1H NMR (CDCl3, 300 MHz) δ 7.60-7.35
(m, 5 H, arom), 5.18 (q, 1H, 20-vinyl H), 4.00 (s, 1H, 3â-H), 2.74 (d
× d, 2H, 12-CH2), 2.46-1.66 (m, 9 H), 1.61 (d, J ) 7.0 Hz, 3H, 21-
allylic CH3), 1.60-1.13 (m, 9 H), 0.98 (s, 3H, 19-CH3), 0.83 (s, 3H,
18-CH3), 0.36 (s, 3H, SiCH3), 0.35 (s, 3H, SiCH3); 13C NMR (CDCl3,

7952 J. Am. Chem. Soc., Vol. 119, No. 34, 1997
Agyin et al.
75.6 MHz) δ 211.38 (11-CdO), 147.61, 139.18, 133.52, 129.42, 127.80,
114.85, 67.35, 64.96, 56.19, 55.77, 47.62, 38.92, 36.16, 35.99, 35.80,
32.59, 31.82, 31.17, 29.39, 28.03, 24.20, 18.16, 13.02, 11.27, -0.78,
-1.00. MS (EI) m/e 450 (M+), 137 (100); high-resolution MS (EI)
m/e calcd for C29H42SiO2 450.2954, found 450.2949.
3r-(Dimethylphenylsiloxy)-17-(Z)-ethylidene-5r-androstan-6-
one (3) and 3â-(Dimethylphenylsiloxy)-17-(Z)-ethylidene-5r-an-
drostan-6-one (4). The alcohol mixture (0.72 g, 2.29 mmol), DMF
(8.0 mL), and TEA (0.9 mL) was added to a dry three-neck flask
equipped with a magnetic stir bar, rubber septa, and a nitrogen inlet.
Stirring was commenced under nitrogen, and chlorodimethylphenyl-
silane (0.47 g, 2.75 mmol) was syringed into the reaction mixture at 0
°C. The mixture was stirred for 1 h at 0 °C. The mixture was then
diluted with benzene (40 mL) and washed successively with cold 5%
NaHCO3 (5 mL), 1 N HCl (5 mL), and 5% NaHCO3 (5 mL). The
organic layer was dried with anhydrous NaSO4 and evaporated to give
a crude product that was purified by silica gel chromatography (5%
EtOAc, 95% hexane) to give 142 mg of crude 3â-(dimethylphenylsi-
loxy)-17-(Z)-ethylidene-5R-androstan-6-one (4) and the desired 3R
isomer, 3, as a white solid. Compound 3 was further purified by two
recrystallizations from acetonitrile to afford 500 mg of product as white
needles (100% GLC pure), mp 145-146 °C. 1H NMR (CDCl3, 300
MHz) δ 7.58-7.36 (m, 5 H, arom), 5.15 (q, 1H, 20-CH), 4.11 (s, 1H,
3â-H), 2.75-1.21 (m, 23 H), 0.87 (s, 3H, 19-CH3), 0.71 (s, 3H, 18-
CH3), 0.35 (s, 6H, Si(CH3)2). MS (EI) m/e (rel intensity): 451 (M +
1), 450 (M+, 50), 435 (M - CH3, 33), 135 (100); MS (CI) m/e 451 (M
+ 1, 100), 299 (89); high-resolution MS (EI) m/e calcd for C29H42-
SiO2 450.2954, found 450.2945.
3r-(Dimethylphenylsiloxy)-17-(Z)-ethylidene-5r-androstan-6-
one (3) and 3â-(Dimethylphenylsiloxy)-17-(Z)-ethylidene-5r-an-
drostan-6-one (4): 17-(Z)-Ethylidene-3,6-diethylenedioxyandros-
tane. To a slurry of ethyltriphenylphosphonium bromide (5.2 g, 14.0
mmol), dry THF (46 mL), and tBuOK (1.62 g, 14.5 mmol) under an
atmosphere of nitrogen in a 250-mL round-bottom flask was added
3.0 mL of dry THF containing 1.82 g (4.67 mmol) of 3,6-diethylene-
dioxyandrostan-17-one (from 17â-hydroxy-5R-androstan-3,6-dione which
had been prepared from testosterone acetate). After the mixture was
stirred for 3 h at room temperature, additional ethyltriphenylphospho-
nium bromide (5.2 g, 14.0 mmol) and t-BuOK (1.62 g, 14.5 mmol)
was added. The mixture was stirred at room temperature for 60 h and
then poured into 400 mL of ice-water and extracted with ethyl acetate
(3 × 100 mL). The combined extract was washed with 3 × 100 mL
of water and 100 mL of brine and dried over anhydrous sodium sulfate.
Filtration, evaporation of solvent, and chromatography on silica gel
(40% EtOAc, 60% hexane) afforded 1.6 g (84% yield) of pure product
as a white solid. A sample was recrystallized from acetone to afford
white needles, mp 184-186 °C. 1H NMR (CDCl3, 300 MHz) δ 5.10
(q, 1 H), 3.94-3.73 (m, 8 H), 2.37-1.02 (m, 23 H), 0.95 (s, 3H, 19-
CH3), 0.88 (s, 3H, 18-CH3); 13C NMR (CDCl3, 75.6 MHz) δ 150.206,
113.522, 109.779, 65.535, 64.320, 55.849, 53.464, 49.653, 44.487,
41.193, 37.319, 37.125, 37.002, 33.062, 31.434, 31.144, 29.390, 24.453,
21.309, 16.975, 13.620, 13.245, 0.122. MS (EI) m/e 402 (M+); MS
(CI) 403 (M + 1); high-resolution MS (EI) m/e calcd for C25H38O4
402.2770, found 402.2766.
17-(Z)-Ethylidene-5r-androstan-3,6-dione. A solution of 3,6-
diethylenedioxy-17-(Z)-ethylidene-5R-androstane (1.54 g, 3.84 mmol)
in wet acetone (40.0 mL) containing pyridinium tosylate (0.579 g, 2.3
mmol) was refluxed for 6 h. Excess solvent was then removed in
Vacuo, ether (100 mL) was added, and the mixture was washed with
saturated NaHCO3 (3 × 20 mL) and once with 20 mL of brine. The
organic phase was dried with anhydrous Na2SO4 and the solvent
removed in Vacuo to give the diketone. GLC analysis indicated 95.7%
C17 (Z) ethylidene and 3.2% of the corresponding E isomer. After
the solution was dried under high vacuum, 0.9 g (75%) of product was
obtained as a white solid. A sample was recrystallized from acetone,
mp 174-176 °C. 1H NMR (CDCl3, 300 MHz) δ 5.14 (q, 1H, 20-
CH), 2.60-1.72 (m, 14 H), 1.65 (d, 3H, 21-CH3, J ) 7.2 Hz), 1.58-
1.20 (m, 6 H), 0.96 (s, 3H, C-19 CH3), 0.89, (s, 3H, C-18 CH3). 13C
NMR (CDCl3, 75.6 MHz) δ 211.269, 209.069, 149.075, 114.177,
57.583, 56.372, 53.555, 46.532, 44.717, 41.367, 38.129, 37.655, 37.467,
37.079, 36.705, 31.199, 24.319, 21.916, 16.949, 13.226, 12.642. MS
(EI) m/e 314 (M+), 299 (M - CH3); MS (CI) m/e 315 (M + 1); high-
resolution MS (EI) m/e calcd for C21H30O2 314.2246, found 314.2259.
Compound 4 (142 mg) was also obtained from the silica gel
chromatography and was recrystallized from acetonitrile, mp 115-
116 °C; MS (EI) m/e 450 (M+, 7), 135 (100); high-resolution MS (EI)
m/e calcd for C29H42SiO2 450.2954, found 450.2948.
Crystals of 3 were prepared by recrystallization from acetonitrile.
A colorless plate of crystal having approximate dimensions of 0.25 ×
0.20 × 0.13 mm was mounted on a glass fiber in a random orientation.
Preliminary examination and data collection were performed with Cu
KR radiation (λ ) 1.54184 Å) on an Unroof-Nonius CAD4 computer
controlled kappa axis diffractometer equipped with a graphite crystal,
incident beam monochromator. Cell constants and an orientation matrix
for data collection were obtained for least-squares refinement by using
the setting angles of 25 reflections in the range 17 < θ < 43° measured
by the computer-controlled diagonal slit method of centering. The data
were collected at a temperature of 295 ( 1 K by using the ω - 2θ
scan technique. Data were collected to a maximum of 2θ of 136.3°.
Photolysis of 3 with 266-nm Light. A degassed solution of 3 in
cyclohexane (10.46 mM, 1.0 mL) was irradiated for 2 min with the
266-nm laser at a power of 4 mJ/pulse. GLC analysis on column A at
240 °C indicated a new peak at tR ) 8.12 min, which corresponded to
1
the E isomer (tR for the Z isomer ) 8.78 min). A H NMR spectrum
of the reaction mixture indicated an upfield shift of the C20 vinyl proton
resonance from δ 5.18 to δ 5.13 and an upfield shift of the allylic C21
methyl resonance from δ 1.66 to δ 1.52 (d, J ) 6.9 Hz) for the E
isomer. Compound 1, 2, and 4 were similarly photolyzed in cyclo-
hexane and analyzed by GLC. In all cases the E isomer was the only
photoproduct detected.
Photolysis of 3 with 300-nm Light. A degassed solution of 3 in
cyclohexane (13.1 mM, 3.0 mL) was irradiated for 26.0 min with 300-
nm light. GLC analysis on column A at 240 °C indicated the formation
of a new product with a retention time identical with that obtained
3-Hydroxy-17-(Z)-ethylidene-5r-androstan-6-one. A 50-mL round-
bottom flask with a side arm capped with a rubber septum was fitted
with a magnetic stir bar and a condenser with a nitrogen inlet tube.
The apparatus was flame dried under nitrogen and a solution of 3,6-
dioxo-17-(Z)-ethylidene-5R-androstane (0.85 g, 2.71 mmol) in dry THF
(12 mL) was syringed into the flask. The solution was stirred and
cooled in a dry ice bath to -78 °C. A solution of K-Selectride (3.0
mL, 1 M in THF) was then added dropwise to the flask, and the
yellowish mixture was stirred at -78 °C for 3 h. The mixture was
warmed to room temperature, stirred for 20 min, and hydrolyzed by
the addition of 8 mL of 50% ethanol in water. The organoborane was
oxidized by the addition of 1.0 mL of 6 N sodium hydroxide and 3.2
mL of 30% hydrogen peroxide. The reaction mixture was saturated
with anhydrous K2CO3 and the layers were separated. The aqueous
layer was extracted with ether/THF (1:1). The organic layers were
combined, washed with 1 N HCl solution and brine, and dried over
anhydrous Na2SO4. After removal of the solvent in Vacuo, the crude
product was dried under high vacuum overnight to afford 0.84 g (98%)
of product as a white solid. GLC analysis of the product indicated a
mixture of 78% C-3 R-alcohol and 22% of the corresponding â-isomer.
The crude product was used for the next reaction without further
purification.
1
from the 266-nm irradiation. Likewise, the H NMR spectrum of the
product mixture matched that obtained from the 266-nm irradiation.
Photolysis of 3âDPSO/17Z with 254-nm light. Irradiation of a
5.0-mL solution of argon degassed 11.0 mM 3âDPSO/17-Z in
cyclohexane with 254-nm light for 1 produced a single product (16.7%)
by GLC analysis on column B. The product was identified as 3âDPSO/
17-E by coinjection on the gas chromatograph with the starting material
(which contains ca. 3% of the E isomer).
Determination of the Photostationary State for 3. A degassed
solution of 3 in cyclohexane (13 mM, 3.0 mL) was photolyzed at
various time intervals with use of 308-nm light and analyzed by GLC
to determine the amount of the E isomer formed. The results are
summarized in Table 5. A photostationary state ratio (E/Z) of 1.2 was
obtained. The same experiment repeated with 254-nm lamps gave a
photostationary state ratio of 1.08.
Determination of the Photostationary State for 2. A degassed
solution of 2 in cyclohexane (6.61 mM, 5.0 mL) was photolyzed at
various time intervals with 254-nm light and analyzed by GLC on

Steroids as Photonic Wires
J. Am. Chem. Soc., Vol. 119, No. 34, 1997 7953
Table 5. Photostationary State Data for 3 at 266 nm
Table 7. Effect of Added Triethylamine (TEA) on Z f E Olefin
Photoisomerization of 3
photolysis
time (min)
% E isomer
% Z isomer
[TEA] (mM)
φ
φ0/φ
1.5
2.5
3.5
4.5
5.5
9.5
11.5
30.9
39.4
44.3
48.1
48.6
51.9
51.9
67.7
59.5
54.8
49.9
51.4
48.1
48.1
0.0
32.1
64.7
96.8
125.5
0.55
0.46
0.39
0.34
0.29
1.00
1.31
1.54
1.78
2.09
Photolysis of 2 and 3 with TEA. Two degassed cyclohexane
solutions of 2 (2 mL, 10.7 mM), with and without TEA (35 µL, 125.5
mM), were irradiated in quartz cells at room temperature for 10 min
with the 308-nm excimer laser (3.4 mJ/pulse). The solutions were
analyzed by GLC on column A at 240 °C. The E isomer was formed
to the extent of 5.8 and 6.3% with and without TEA, respectively. The
experiment was repeated with compound 3 (1 mL, 10.6 mM), with
and without TEA (35 µL, 125.5 mM), to give 7.7 and 16.2% E isomer
with and without TEA, respectively. An analysis with varying
concentrations of TEA gave the data shown in Table 7.
Table 6. Photostationary State Data for 2 at 266 nma
photolysis
time (min)
% E isomer
% Z isomer
40.0
80.0
100.0
130.0
160.0
35.4
50.4
54.4
58.2
58.4
64.6
48.6
45.6
41.8
41.6
a In cyclohexane.
Acknowledgment. We are grateful to Phil Fanwick, Ken
Haber, and Changhe Xiao, all of Purdue University, for their
assistance with various portions of this study. We also thank
Alex Siemiarczuk of Photon Technology International for
obtaining some of the lifetime data reported herein. We are
grateful to the National Science Foundation (Grant CHE
9311828) for support of this research.
column B at 240 °C. The photostationary state ratio (E/Z) was
determined to be 1.38. The data are presented in Table 6.
Quantum Efficiency Determinations. A quartz cuvette containing
1.0-2.0 mL of argon degassed sample solution of ca. 7-15 mM was
placed in a sample holder and irradiated with the 266- or 308-nm laser
for 3-9 min. Dark control experiments were carried out at the same
temperature for each of the quantum efficiency measurements.
Photolysis of 1 and 3 in the Presence of cis-2-Heptene. Degassed
cyclohexane solutions of 3 (15.34 mM, 3 mL) with and without 15
mM cis-2-heptene were photolyzed in matched quartz tubes with 254-
nm light. GLC analysis of the photolysates indicated 36.1 and 36.5%
formation of the E isomer with and without cis-2-heptene, respectively.
The same experiment conducted with 1 (11 mM, 3 mL) and 11 mM
2-heptene in cyclohexane gave 16.9% and 16.7% Z f E olefin
isomerization with and without cis-2-heptene, respectively.
Supporting Information Available: The X-ray structure and
a table summarizing the crystal data and data collection
parameters for compound 3 (4 pages). See any current masthead
page for ordering and Internet access instructions.
JA963261A