Delocalized carotenoid cations in relation to the soliton model

Article abstract of DOI:10.1021/ja0492541

A series of charge-delocalized carotenoid mono- and dications have been prepared by treatment of selected carotenoids with Bronsted and Lewis acids. The detailed structures of the carbocations were established by NMR studies in the temperature range from

Full text of DOI:10.1021/ja0492541

Published on Web 07/02/2004
Delocalized Carotenoid Cations in Relation to the Soliton
Model
Bjart Frode Lutnaes, Geir Kildahl-Andersen, Jostein Krane, and
Synnøve Liaaen-Jensen*
Contribution from the Department of Chemistry, Norwegian UniVersity of Science and
Technology (NTNU), NO-7491 Trondheim, Norway
Received February 11, 2004; E-mail: slje@chem.ntnu.no
Abstract: A series of charge-delocalized carotenoid mono- and dications have been prepared by treatment
of selected carotenoids with Brønsted and Lewis acids. The detailed structures of the carbocations were
established by NMR studies in the temperature range from -10 to -20 °C. The general strategy for structure
elucidation by NMR of several cationic components in a mixture is outlined. Bond type and regions of bond
inversion were established, as well as the charge distribution, which was determined from the difference
in 13C chemical shift at each carbon. This method gave a more accurate estimate for the partial charges
than by using the Spiesecke-Schneider relationship. The resulting charge distribution was used as models
for the structure of charged solitons. These carotenoid cations have the most delocalized charge so far
determined, and the monocations represent the first experimental structure determination of positively
charged solitons. The soliton width determined here is in good agreement with the results of previous AM1
calculations.
units have been determined by X-ray crystallography.9,16 The
carotenoids in these photosystems serve for light harvesting in
Introduction
Carotenoids are isoprenoid polyene pigments widely distrib-
uted in nature.1,2 Around 800 naturally occurring carotenoids
are known.3 The isolation,1 structure determination by spectro-
scopic methods,4 biosynthesis and metabolism,5 and total
synthesis6 of carotenoids are treated in recent monographs.
A general structural feature of the carotenoids is the iso-
prenoid polyene chain, responsible for the yellow-orange-red
colors.4 Carotenoproteins exhibit purple or blue colors.7 Electron-
deficient carotenoids, including radical cations, monocations,
and dications, have received increasing attention during the past
decade, theoretically and experimentally. Methods mainly
employed for studies of charged carotenoid species include NIR,
EPR, ENDOR, and resonance Raman spectroscopy, cyclic
voltammetry, and AM1 calculations.8
the 450-570 nm region where the absorption of chlorophyll is
negligible, for quenching of chlorophyll triplet states, singlet
oxygen scavenging, excess energy dissipation, and structure
stabilization.9,17 Photosystem II (PS II), which oxidizes water
to oxygen in photosynthesis, is the best studied of these
photosystems, and here the functions of carotenoid cation
radicals have been studied by resonance Raman, IR, EPR,
ENDOR, and ESEEM spectroscopy.14,15,18-21 In addition to the
already mentioned roles of carotenoids in PS II, also the function
as a redox intermediate in the electron-transfer pathway within
PS II has been suggested.13,18
Treatment of carotenoids and other polyenes with electron
acceptors, for example, iodine, in a process known as doping,
Carotenoids have been encountered in both Photosystems
I9-11 and II12-15 reaction centers. Structures of the photosynthetic
(9) Jordan, P.; Fromme, P.; Witt, H. T.; Klukas, O.; Saenger, W.; Krauss, N.
Nature 2001, 411, 909.
(10) Bialek-Bylka, G. E.; Fujii, R.; Chen, C.-H.; Oh-Oka, H.; Kamiesu, A.; Satoh,
K.; Koike, H.; Koyama, Y. Photosynth. Res. 1998, 58, 135.
(11) Bialek-Bylka, G. E.; Hiyama, T.; Yumoto, K.; Koyama, Y. Photosynth.
Res. 1996, 49, 245.
(12) Bialek-Bylka, G. E.; Tomo, T.; Satoh, K.; Koyama, Y. FEBS Lett. 1995,
363, 137.
(13) Vasil’ev, S.; Brudvig, G. W.; Bruce, D. FEBS Lett. 2003, 543, 159.
(14) Telfer, A.; Frolov, D.; Barber, J.; Robert, B.; Pascal, A. Biochemistry 2003,
42, 1008.
(15) Tracewell, C. A.; Cua, A.; Stewart, D. H.; Bocian, D. F.; Brudvig, G. W.
Biochemistry 2001, 40, 193.
(16) Zouni, A.; Witt, H.-T.; Kern, J.; Fromme, P.; Krauss, N.; Saenger, W.;
Orth, P. Nature 2001, 409, 739.
(17) Frank, H. A.; Cogdell, R. J. Photochem. Photobiol. 1996, 63, 257.
(18) Tracewell, C. A.; Vrettos, J. S.; Bautista, J. A.; Frank, H. A.; Brudvig, G.
W. Arch. Biochem. Biophys. 2001, 385, 61.
(19) Faller, P.; Maly, T.; Rutherford, A. W.; MacMillan, F. Biochemistry 2001,
40, 320.
(1) Goodwin, T. W. The Biochemistry of the Carotenoids Vol. 1. Plants, 2nd
ed.; Chapman and Hall: London, 1980.
(2) Goodwin, T. W. The Biochemistry of the Carotenoids Vol. 2. Animals, 2nd
ed.; Chapman and Hall: London, 1984.
(3) Britton, G.; Liaaen-Jensen, S.; Pfander, H. Carotenoids: Handbook;
Birkha¨user: Basel, 2004.
(4) Britton, G.; Liaaen-Jensen, S.; Pfander, H. Carotenoids Vol. 1B: Spec-
troscopy; Birkha¨user: Basel, 1995.
(5) Britton, G.; Liaaen-Jensen, S.; Pfander, H. Carotenoids Vol. 3: Biosynthesis
and Metabolism; Birkha¨user: Basel, 1998.
(6) Britton, G.; Liaaen-Jensen, S.; Pfander, H. Carotenoids Vol. 2: Synthesis;
Birkha¨user: Basel, 1996.
(7) Britton, G.; Armitt, G. M.; Lau, S. Y. M.; Patel, A. K.; Shone, C. C. In
Carotenoid Chemistry & Biochemistry, Proc. Int. Symp. Carotenoids, 6th
ed.; Britton, G., Goodwin, T. W., Eds.; Pergamon Press: Oxford, 1982;
pp 237-251.
(8) Liaaen-Jensen, S.; Lutnaes, B. F. Charged Cartenoid Species. In Studies in
Natural Products Chemistry (BioactiVe Natural Products); Atta-ur-Rahman,
Ed.; Elsevier: Amsterdam, 2004; Vol. 30, pp 515-557.
(20) Barber, J. Curr. Opin. Struct. Biol. 2002, 12, 523.
(21) Deligiannakis, Y.; Hanley, J.; Rutherford, A. W. J. Am. Chem. Soc. 2000,
122, 400.
9
10.1021/ja0492541 CCC: $27.50 © 2004 American Chemical Society
J. AM. CHEM. SOC. 2004, 126, 8981-8990
8981

A R T I C L E S
Lutnaes et al.
Chart 1
causes an extreme increase in the electrical conductivity of the
doped polyene.22-26 The use of doped carotenoids in batteries
has been investigated.27 At present, the charge in conducting
polyenes is considered to be carried in “electron holes” called
positive solitons in the polyene chain, the Su-Schrieffer-
Heeger (SSH) model.24,28-30 This model describes the charge
as a charge-density wave, with the charge located on every other
carbon, and with a concentration of the charge density wave in
the center of the polyene chain. This phenomenon is ac-
companied by a decrease in the difference between single and
double bonds. The model has later been modified, showing that
there might be a small negative charge on the carbons not
carrying the positive charge, and that the width and position of
the charge density wave and the bond-length distortions need
not be the same.31-34 NMR studies on dienes and polyenes,
including solitons, have recently been reviewed,35 but apart from
NMR studies of polyene anions,36-38 only short-chain ions have
been examined. The studies on the structures of the solitons
have otherwise been centered on electron absorption spectra and
theoretical calculations.31-34,39,40
Scheme 1
Another area receiving interest concerning the structure of
charged carotenoids is in the field of molecular design.41-52 This
includes molecular wires41-47 that are able to transport electrons
across or through, for example, a lipid membrane, molecular
switches,43 and push-pull polyenes48-50 with nonlinear optical
properties. Carotenoids and structurally related compounds that
have been purpose-designed, especially the caroviologens,41,44,45
(22) Sen, S.; Pal, P.; Misra, T. N. J. Mater. Sci. 1993, 28, 1367.
(23) Huggins, C. M.; LeBlanc, O. H. J. Nature 1960, 186, 552.
(24) Ghosh, D.; Hazra, S.; Pal, P.; Misra, T. N. Bull. Mater. Sci. 1993, 16, 127.
(25) Ehrenfreund, E.; Moses, D.; Heeger, A. J.; Cornil, J.; Bredas, J. L. Chem.
Phys. Lett. 1992, 196, 84.
(26) Shirakawa, H. ReV. Mod. Phys. 2001, 73, 713.
(27) Shirakawa, H.; MacDiarmid, A.; Heeger, A. Chem. Commun. 2003, 1, 1.
(28) Su, W. P.; Schrieffer, J. R.; Heeger, A. J. Phys. ReV. Lett. 1979, 42, 1698.
(29) Su, W. P.; Schrieffer, J. R.; Heeger, A. J. Phys. ReV. B: Condens. Matter
1980, 22, 2099.
have been extensively studied in this respect and were shown
to be promising model compounds.
We have reported the first detailed studies on charged,
spinless, carotenoids by NMR spectroscopy, resulting in com-
plete structure determinations of the â,â-carotene (1) dication
(2).53,54 Regions of bond inversion, where the bond orders have
changed from single and double to an intermediate type, are
shown by dotted bonds, and distribution of the positive charge
in 2 is illustrated by the diameter of the filled circles in Chart
1.
More recently, we have conducted NMR studies of protonated
â,â-carotene-4,4′-dione (3, canthaxanthin) as models for electron-
deficient 3,3′-dihydroxy-â,â-carotene-4,4′-dione (4, astaxanthin)
in the blue carotenoprotein crustacyanin,55 in light of the recent
X-ray structure of 4 bound in â-crustacyanin.56
Whereas dication 2 contains 20 π electrons, it was considered
of interest to study â,â-carotene (1) related monocations and
dications containing 22 π electrons. Removal of the hydroxy
group from the allylic carotenol â,â-caroten-4-ol (5, isocryp-
toxanthin) provided the desired monocation 6, while â,â-
carotene-4,4′-diol (7, isozeaxanthin) in principle could furnish
the target dication 8, Scheme 1.
(30) Heeger, A. J.; Kivelson, S.; Schrieffer, J. R.; Su, W. P. ReV. Mod. Phys.
1988, 60, 781.
(31) Villar, H. O.; Dupuis, M.; Clementi, E. Phys. ReV. B: Condens. Matter
1988, 37, 2520.
(32) Craw, J. S.; Reimers, J. R.; Bacskay, G. B.; Wong, A. T.; Hush, N. S.
Chem. Phys. 1992, 167, 101.
(33) Reimers, J. R.; Craw, J. S.; Hush, N. S. J. Phys. Chem. 1993, 97, 2778.
(34) Broszeit, G.; Diepenbrock, F.; Graf, O.; Hecht, D.; Heinze, J.; Martin, H.
D.; Mayer, B.; Schaper, K.; Smie, A.; Strehblow, H. H. Liebigs Ann. 1997,
11, 2205.
(35) Takeuchi, Y.; Takayama, T. In The Chemistry of Dienes and Polyenes;
Rappoport, Z., Ed.; Wiley: Chichester, 2000; Vol. 2, pp 59-198.
(36) Tolbert, L. M.; Ogle, M. E. J. Am. Chem. Soc. 1990, 112, 9519.
(37) Tolbert, L. M.; Ogle, M. E. Mol. Cryst. Liq. Cryst. 1990, 189, 279.
(38) Tolbert, L. M.; Ogle, M. E. J. Am. Chem. Soc. 1989, 111, 5958.
(39) Painelli, A.; Girlando, A.; Del Freo, L.; Soos, Z. G. Phys. ReV. B: Condens.
Matter 1997, 56, 15100.
(40) An, Z.; Wong, K. Y. J. Chem. Phys. 2003, 119, 1204.
(41) Li, J.; Tomfohr, J. K.; Sankey, O. F. Physica E 2003, 19, 133.
(42) Ramachandran, G. K.; Tomfohr, J. K.; Li, J.; Sankey, O. F.; Zarate, X.;
Primak, A.; Terazono, Y.; Moore, T. A.; Moore, A. L.; Gust, D.; Nagahara,
L. A.; Lindsay, S. M. J. Phys. Chem. B 2003, 107, 6162.
(43) Porres, L.; Alain, V.; Thouin, L.; Hapiot, P.; Blanchard-Desce, M. Phys.
Chem. Chem. Phys. 2003, 5, 4576.
(44) Slama-Schwok, A.; Blanchard-Desce, M.; Lehn, J. M. J. Phys. Chem. 1992,
96, 10559.
(45) Maerkl, G.; Poell, A.; Aschenbrenner, N. G.; Schmaus, C.; Troll, T.;
Kreitmeier, P.; Noeth, H.; Schmidt, M. HelV. Chim. Acta 1996, 79, 1497.
(46) Gonzalez, C.; Morales, R. G. E. Chem. Phys. 1999, 250, 279.
(47) Lehn, J. M. Angew. Chem. 1988, 100, 91.
(48) Laage, D.; Thompson, W. H.; Blanchard-Desce, M.; Hynes, J. T. J. Phys.
Chem. A 2003, 107, 6032.
(53) Lutnaes, B. F.; Bruas, L.; Krane, J.; Liaaen-Jensen, S. Tetrahedron Lett.
2002, 43, 5149.
(49) Alain, V.; Blanchard-Desce, M.; Ledoux-Rak, I.; Zyss, J. Chem. Commun.
2000, 5, 353.
(54) Lutnaes, B. F.; Bruas, L.; Kildahl-Andersen, G.; Krane, J.; Liaaen-Jensen,
S. Org. Biomol. Chem. 2003, 1, 4064.
(50) Blanchard-Desce, M.; Ledoux, I.; Lehn, J. M.; Malthete, J.; Zyss, J. J.
Chem. Soc., Chem. Commun. 1988, 11, 737.
(55) Kildahl-Andersen, G.; Lutnaes, B. F.; Liaaen-Jensen, S. Org. Biomol. Chem.
2004, 2, 489.
(51) Heinze, J.; Tschuncky, P.; Smie, A. J. Solid State Electrochem. 1998, 2,
102.
(56) Cianci, M.; Rizkallah, P. J.; Olczak, A.; Raftery, J.; Chayen, N. E.; Zagalsky,
P. F.; Helliwell, J. R. Proc. Natl. Acad. Sci. U.S.A. 2002, 99, 9795.
(52) Mori, Y. J. Raman Spectrosc. 2001, 32, 543.
9
8982 J. AM. CHEM. SOC. VOL. 126, NO. 29, 2004

Delocalized Carotenoid Cations
A R T I C L E S
Scheme 2
The structure of monocation 6, prepared by treatment of 5
with CF3COOH in CH2Cl2 at -10 °C, has been reported in a
preliminary communication.57 This provided experimental evi-
dence for a cation intermediate during the elimination of allylic
hydroxy groups by 0.03 N HCl in CHCl3. This reaction, which
has long been a standard reaction in the carotenoid field, has
been considered to proceed via a hypothetical cation intermedi-
ate to products with prolonged chromophore.58,59
The mono- and dications prepared in this work are the longest
conjugated polyenyl cations that have been fully characterized.
The studies of these carotenoid cations may be useful for the
understanding of the structures of charged polyenes and solitons
in general, as well as for the functions of carotenoids in the
systems mentioned above.
Materials and Methods
Materials. Synthethic â,â-carotene-4,4′-dione (canthaxanthin, 3),
3,3′-dihydroxy-â,â-carotene-4,4′-dione (astaxanthin, 4), â,â-caroten-
4-one (echinenone, 9), 4′,5′-didehydro-4,5′-retro-â,â-carotene (isocaro-
tene, 10), ψ,ψ-carotene (lycopene, 11), and 2,6,11,15-tetramethylhexa-
deaheptaene-1,16-dial (crocetindial, 12) were obtained from Hoffmann-
La Roche, Basel. Commercially available CF3COOH from Acros and
Merck and CF3COOD and BF3-diethyl etherate from Acros were used.
CF3SO3H, CF3SO3D, and THF-d8 were supplied from Aldrich.
Instrumentation. Visible light (VIS) and near-infrared (NIR) spectra
were recorded on a Varian Cary 50 UV-vis spectrophotometer (190-
1100 nm).
ROESY spectrum. This gave the 1H chemical shifts of the entire
polyene chain from H-7 to H-7′. ROESY cross-peaks between
methyl group protons H-16, H-17, and H-18 and in-chain
protons H-7 and H-8 connected the polyene chain to the end
groups and established the bond geometry of the C-6,7 bond.
The spin systems in the shielded part of the 1H spectrum,
corresponding to H-2, H-3, and H-4, were correlated to the H-18
methyl protons through long-range coupling in the COSY
spectrum, as well as ROESY interactions to methyl groups H-16
and H-17. Inequality of the H-16 and H-17 shifts shows that
the end group is unsymmetrically substituted.
NMR spectra were obtained on a Bruker Avance DRX 500
instrument, equipped with a 5 mm inverse probe (TXI). CD2Cl2 or
CDCl3 was used as the solvent. Chemical shifts were cited relative to
1
TMS with calibration against CHDCl2 at 5.32 and 53.8 ppm for H
and 13C, respectively, or with calibration against CHCl3 at 7.27 ppm
1
for H.
General Method for Preparation and NMR Analysis of Caro-
tenoid Cations. Carotenoid substrate (1.0-1.4 mg) was dissolved in a
mixture of CD2Cl2 (0.75 mL) and Brønsted acid (in specified amounts).
The mixture was transferred to a chilled NMR tube and analyzed by
The further assignment of 13C chemical shifts for carbons
with protons attached is available from HSQC or the direct
coupling processed HMBC spectrum. By experience, HSQC and
HMBC experiments on carotenoid cations in various stages of
decomposition give spectra that are difficult to interpret.
Quaternary carbons will in most cases be detected either by
2JC,H couplings (e.g., H-16/H-17 - C-1,H-18 - C-5,H-19 -
C-9 and H-20 - C-13) or by 3JC,H couplings (e.g., H-16/H-17/
H-18 - C-6) in the HMBC spectrum. In systems with a higher
level of degeneracy, such as the symmetrical dication 8, where
1
1
1
500 MHz NMR (1D H, 2D H-1H gs-COSY, 2D ROESY, H-13C
gs-HSQC, and 1H-13C gs-HMBC). Experiments with Lewis acids were
performed by dissolving the carotenoid substrate (1-2 mg) in CD2Cl2
(0.4 mL) in an NMR tube at -20 °C and then adding chilled BF3-
etherate (0.2 mL). The spectra were recorded as described previously.55
Decomposition of the sample was observed during experiments,
restricting the acquisition time. The available acquisition time, generally
above 6 h, was dependent on the temperature, the substrate and acid
used, and the concentration of acid.
For a detailed description of the experiments, experimental data not
given in the Results, and additional carotenoid structures, see the
Supporting Information.
1
the two rotamers give a theoretical total of six different H
chemical shifts corresponding to the in-chain methyl groups
H-19 and H-20, all apparently confined to a narrow window
between 2.23 and 2.24 ppm, alternative approaches are needed.
In this special case, the assignments of C-9 and C-13 were aided
by the detection of the 3JC,H couplings in the HMBC spectrum
between H-7 and C-9, as well as H-11 and C-13.
Results
Analysis of NMR Spectra. The H and 13C chemical shifts
1
and coupling constants for the carotenoid cations were estab-
lished from 1D H, 2D H-1H COSY, 2D ROESY, H-13C
1
1
1
HSQC, and H-13C HMBC experiments run at low tempera-
1
Structure of Monocations 6, Prepared from â,â-Caroten-
4-ol (5). The above-described approach resulted in the structure
elucidation of monocation 6, prepared from â,â-caroten-4-ol
(5), as the two diastereomers 6a and 6b, as reported in a
preliminary communication.57 Bond inversion in the central
region was inferred from the coupling constants, and the
distribution of the charge was demonstrated. The two isomers
6a and 6b differ in the stereochemistry of the C-6,7 double bond,
Scheme 2. The presence of the conformers 6c and 6d was also
tures (-10 to -20 °C). A common strategy was employed for
analysis of all carotenoid cations. First, the different spin
subsystems (2-spin, 3-spin, or 4-spin) in the polyene chain (or
chains) were identified from the COSY spectrum. They then
were linked to in-chain methyl groups by correlations in the
(57) Kildahl-Andersen, G.; Lutnaes, B. F.; Krane, J.; Liaaen-Jensen, S. Org.
Lett. 2003, 5, 2675.
(58) Karrer, P.; Leumann, E. HelV. Chim. Acta 1951, 34, 445.
(59) Entschel, R.; Karrer, P. HelV. Chim. Acta 1958, 41, 402.
9
J. AM. CHEM. SOC. VOL. 126, NO. 29, 2004 8983

A R T I C L E S
Lutnaes et al.
1
A complete determination of all H and 13C chemical shifts
for dication 13 was not possible from the experiments per-
formed. No further attempts were made on the complete
assignments of all proton and carbon chemical shifts of 13
because of the close similarity between this cation and the
monocation prepared from mono-ol 5. Its structure would not
provide new charge-distribution data.
Scheme 3
Elevating the temperature is known to cause the elimination
of protonated hydroxy groups that will not eliminate at lower
temperature.61 From our previous experience with carotenoid
cations, increasing the temperature will result in a loss of
stability, which would render the structure determination of the
cation by NMR impossible. The allylic hydroxy groups were
therefore acetylated in an attempt to prepare a better leaving
group. However, treatment of 4,4′-diacetoxy-â,â-carotene (14)
with CF3COOH resulted in the formation of the same dication
13. To further enhance the leaving group, the preparation of
4,4′-di(trifluoroacetoxy)-â,â-carotene (15) was attempted. This
resulted in decomposition of the carotenoid, probably by
spontaneous elimination of trifluoroacetate.
We therefore resorted to use of the stronger Brønsted acid
CF3SO3H. Treatment of â,â-carotene-4,4′-diol (7) with this acid
immediately caused the formation of a cation with a λmax of
1022 nm at -15 °C. Analysis by NMR as described above
resulted in the structure elucidation of the target molecule, 4,4′-
didehydro-â,â-carotenyl dication (8) with 22 π electrons.
Formally, this is the dication derived from isocarotene (10) by
removal of two electrons. Dication 8 may therefore be referred
to as the isocarotene (10) cation.
documented. Further experimental details on the monocations
(6a, b, c, d), including coupling constants, are given in the
present paper.
Due to the size of the molecules, the configuration of the
C-6,7 double bond does not affect the 13C chemical shifts beyond
the central carbon atoms of the polyene chain. Therefore, the
NMR spectra show two different sets of signals for the end
groups of the molecule, which converge to only one set of
signals when approaching the center of the polyene chain. The
ratio 3:2 of the two end groups was determined from the
intensities of the H-18 methyl protons, indicating a 44:45:11
ratio of the three diastereomeric dications 8c, 8b, and 8a,
respectively; see Scheme 3. NMR data for the two (6-E) and
(6-Z) isomeric half molecules are given in Table 1.
Structures of Cations Prepared from Echinenone (9). A
model for protonated astaxanthin (4) in the blue carotenoprotein
crustacyanin was recently presented, based on protonated
canthaxanthin (3).55 In a recent paper, the presence of 3′-
hydroxy-â,â-caroten-4-one (3′-hydroxyechinenone, 16) in an-
other carotenoprotein has been demonstrated by X-ray crystal-
lography.62 In addition to the general interest in comparing the
structure of cations of echinenone (9) to that of the structurally
related carotenoid cations, it was therefore also of interest to
see whether protonated echinenone (9) could be used as a model
for 16 bound in this carotenoprotein.
Structures of the Dications Prepared from â,â-Carotene-
4,4′-diol (7). The first attempt to prepare the symmetrical
dicarbocation 8 was performed by treatment of â,â-carotene-
4,4′-diol (7) with CF3COOH, by analogy to the preparation of
monocation 6 from â,â-caroten-4-ol (5). The expected color
change upon preparation of a cation was immediately noted,
and VIS/NIR spectroscopy showed a λmax of 1028 nm at -20
°C.
Subsequent analysis of the cation obtained revealed the
presence of a monocarbocation with the charge delocalized in
the center of the polyene chain. One of the hydroxy groups was
still present, as evidenced by chemical shifts and coupling
pattern of the â end group. A more thorough investigation of
the NMR data for the hydroxylated end group showed that the
hydroxyl moiety had been protonated by the acid, but was still
1
present. This was evident from the H chemical shift of H-4
(δH ) 5.43 ppm), which, in comparison with â,â-carotene-4,4′-
diol (7), corresponded to a downfield shift of 1.42 ppm.
Comparable downfield shifts are observed for a proton in a
similar neighboring position in a protonated adamantyladaman-
tane ether (∆δH 1.41 ppm)60 and a series of protonated
alcohols,61 although the effect on the 13C chemical shift was
smaller than that for the adamantane system.60 CF3COOH is
therefore not a strong enough acid to induce the elimination of
both allylic hydroxy groups at -20 °C. The cation prepared
from â,â-carotene-4,4′-diol (7) by treatment with CF3COOH
was thus identified as dication 13, Scheme 3.
Treatment of echinenone (9) in CH2Cl2 with CF3COOH
provided the formation of cationic species, as expected, with a
λmax of 950 nm at -15 °C. NMR experiments at -15 °C, as
described above, were used for a complete structure determi-
nation of the cations prepared. The NMR spectra revealed a
complex mixture of several cations. Surprisingly, however, it
(60) Laali, K. K.; Okazaki, T.; Takeuchi, K.; Ogawa, K.; Bennet, A. J. J. Chem.
Soc., Perkin Trans. 2 2002, 6, 1105.
(61) Olah, G. A.; Sommer, J.; Namanworth, E. J. Am. Chem. Soc. 1967, 89,
3576.
(62) Kerfeld, C. A.; Sawaya, M. R.; Brahmandam, V.; Cascio, D.; Ho, K. K.;
Trevithick-Sutton, C. C.; Krogmann, D. W.; Yeates, T. O. Structure 2003,
11, 55.
9
8984 J. AM. CHEM. SOC. VOL. 126, NO. 29, 2004

Delocalized Carotenoid Cations
A R T I C L E S
Table 1. NMR Data (500 MHz, CD2Cl2, -15 °C) for the Two
Diastereomeric Half Molecules of Dication 8
Scheme 4
(6-E) half molecule (8c)
(6-Z) half molecule (8a)
δH
δC
3JH,H
1JC,H
δH
δC
3JH,H
1JC,H
1
2
3
38.5
39.0
29.8
40.0
35.2
a
1.71
2.53
1.75
2.60
4
5
7.05 156.7
139.2
6.91 156.6
136.6
6
185.3
188.1
7
8
9
7.00 126.9 13.8
8.48 166.6 14.3
145.3
158 7.17 127.1 14.1
157 8.39 168.9 13.8
145.3
10
11
12
13
14
15
7.86 169.1 12.5
7.31 131.2
7.59 163.2
147.1
7.26 151.4
7.41 141.5
157 7.83 169.3 12.3
156 7.30 131.1
156 7.56 163.1
147.1
7.25 151.4
7.41 141.5
16/17 1.48
29.6
128 1.29
128 2.31
28.2
125
129
18
19
20
2.10
2.23
2.24
21.0 H-4b
25.7 H-4b
12.1 H-8/H-10b 129 2.23
12.3 H-8/H-10b 129
13.2 H-14b
129 2.24
13.2 H-14b
129
a Chemical shifts could not be determined from the spectra recorded.
b Long-range correlations obtained from the COSY spectrum.
was found that the main carotenoid monocation had not been
protonated by the acid, but that the cation was formed by
elimination of a hydride from the 4′-position. The formation of
hydrogen gas was not investigated. However, an analogous
mechanism involving the expulsion of molecular hydrogen has
been elaborated for the conversion of carotenoid furanoxides
to blue oxonium ions using CF3COOH.63 The monocation (17)
obtained existed in both the (6′-Z) (17a) and the (6′-E) (17b)
upon treatment with the Lewis acid, but none of them provided
a stable dication whose structure could be determined by NMR
analysis. Isorenieratene (20), however, provided monocation 21,
containing an extra hydrogen in the 7-position; see Scheme 5.
Such hydrogen substituents have been observed also in other
carotenoid cations, and the origin of this hydrogen atom, as a
proton, a hydride, or a hydrogen radical, will be treated
elsewhere. Complete 1H and 13C chemical shifts for monocation
21 are given in Table 2. This product only counted for a minor
amount of the carotenoid species in the reaction mixture. Some
signals in the COSY and ROESY spectra were also compatible
with the expected dication, when compared to chemical shifts
and coupling patterns of other carotenoid cations.
1
configurations; see Scheme 4. H and 13C chemical shifts of
this cation are given in Table 2. The conformation at C-6 could
not be determined because the overlap of the chemical shifts
of H-7 and H-8 made it impossible to distinguish between the
ROESY signals from these protons to the methyl protons on
the end group.
In addition to the E/Z isomeric monocations 17a and 17b,
the third major carotenoid was identified as the 5′-protonated
cation 18. For 18, and for the minor 5-protonated cation 19,
Scheme 4, full assignment of the chemical shifts was not
possible because of the strong overlap of signals and low
intensity. For the same reason, the positions of the double bonds
were determined only from the long-range couplings in the
COSY spectrum. The three major cations 17a, 17b, and 18,
accounted for 75% of the total amount of carotenoid present in
the reaction mixture.
Preparation of Cations by Treatment of Carotenoids with
BF3-Etherate. Using the same approach as for the preparation
of the first carotenoid dication by treatment of â,â-carotene (1)
with BF3-etherate,54 the preparation of dications from several
other carotenoids was attempted to determine the effects of the
end groups on the charge distribution. The carotenoids inves-
tigated include isorenieratene (20), lycopene (11), canthaxanthin
(3), and astaxanthin (4). Previous problems with the NMR
analyses due to the strong signals from the etherate reagent were
avoided by using BF3-THF-d8 as reagent, which could be easily
distilled from a mixture of BF3-diethyl etherate and THF-d8.
The absorption maxima of all of these carotenoids im-
mediately shifted from the 450 nm region to the 1000 nm region
The special property of â,â-carotene (1) that the oxidation
of the second electron is at a lower energy level than the first
electron64,65 might provide an explanation for the ease of â,â-
carotene (1) to form a dication upon treatment with the Lewis
acid BF3. Canthaxanthin (3), on the other hand, with a
significantly higher oxidation potential for the second oxidation
step in CH2Cl2, does not. Simulations in DigiSim have shown
that there is a small difference in the oxidation levels for
lycopene (11),64 as for â,â-carotene (1), but the aliphatic 11 is
known to undergo cyclizations when treated with acid. No
electrochemical data are available for isorenieratene (20).
NMR Assignments for Isocarotene (10) and Isorenieratene
1
(20). A literature survey revealed that only partial H and 13C
NMR chemical shift data for isocarotene (10)66,67 and isoreni-
(64) Liu, D.; Gao, Y.; Kispert, L. D. J. Electroanal. Chem. 2000, 488, 140.
(65) Hapiot, P.; Kispert, L. D.; Konovalov, V. V.; Saveant, J.-M. J. Am. Chem.
Soc. 2001, 123, 6669.
(66) Andrewes, A. G.; Englert, G.; Borch, G.; Strain, H. H.; Liaaen-Jensen, S.
Phytochemistry 1979, 18, 303.
(67) Englert, G. In Carotenoids Vol. 1B: Spectroscopy; Britton, G., Liaaen-
Jensen, S., Pfander, H., Eds.; Birkha¨user: Basel, 1995; pp 147-260.
(63) Haugan, J. A.; Liaaen-Jensen, S. Acta Chem. Scand. 1994, 48, 152.
9
J. AM. CHEM. SOC. VOL. 126, NO. 29, 2004 8985

A R T I C L E S
Lutnaes et al.
Table 2. NMR Data (500 MHz, CD2Cl2, -15 °C) for the C-6 E/Z
Isomeric Monocations 17a and 17b, and for the 7-Protonated
Isorenieratene Cation (21)
(2,3,6-methylbenzyl)triphenylphosfinyl bromide (22) and cro-
cetindial (12); see Scheme 5. NMR data for isocarotene (10)
and isorenieratene (20) are given in the Supporting Information.
17b
17a
21
Discussion
δH
δC
δH
δC
δH
δC
(ppm)
(ppm)
(ppm)
(ppm)
(ppm)
(ppm)
Structures of the Carotenoid Cations. No in-chain proto-
nation of â,â-caroten-4-ol (5) or â,â-carotene-4,4′-diol (7) was
observed by treatment with CF3COOH, or of the diol 7 treated
with CF3SO3H, demonstrating that protonation of the hydroxy
group is greatly favored, and that these acids are not strong
enough to protonate these polyene mono- and dications.
Previously, it has been shown that â,â-carotene (1) suffers in-
chain protonation with CF3COOH68 and that canthaxanthin (3)
is protonated with CF3COOH or CF3SO3H in the polyene chain
as well as on the keto groups.55
It is also interesting to note that all of the carotenoid cations,
with the exception of the bond connecting the end groups to
the polyene chain, seem to exist in the all-E configuration, as
no signs of Z isomers have been detected in the NMR
experiments. A similar observation has previously been made
in studies of R,ω-diphenylpolyenyl anions.36,38 For the charged
carotenoids, also the 6-E configuration, which is not present in
neutral carotenoids,4 has been observed.
4-Dehydro-â,â-Carotenyl Monocation (6) and the Soliton
Structure. It may be concluded from basic NMR theory that
the 13C chemical shift is correlated to the electron density on
the carbon atom. In a pioneering work, Spiesecke and Schneider69
found that there is a linear relationship between 13C chemical
shift and charge ratio in charged monocyclic aromatics.
O’Brien69 extended this to sp2-hybridized systems, in general,
however still restricted to planar, unbridged and unsubstituted
hydrocarbons. The correlation between 13C chemical shift, δC,
and π-charge density, F, is given in eq 1,36
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
1′
2′
3′
4′
5′
6′
7′
8′
9′
10′
11′
12′
13′
14′
15′
16′
17′
18′
19′
20′
37.1
39.7
23.7
38.4
35.9
24.6
135.8
134.7
128.4
127.9
134.3
136.4
31.0
147.4
137.2
156.7
127.9
159.1
140.9
166.4
133.2
15.4
1.65
2.35
6.51
1.65
2.4
6.34
6.92
6.89
144.3
136.5
169.8
123.2
155.9
136.6
165.3
128.1
163.9
142.3
161.1
132.8
29.4
29.4
21.1
12.4
12.5
36.7
35.9
32.9
206.4
129.6
172.2
128.1
143.7
139.6
136.3
135.3
140.0
156.0
136.3
151.1
27.1
144.1
133.5
172.1
122.8
157.8
136.2
166.7
128.1
164.4
6.77
7.97
6.86
7.91
3.66
6.00
7.61
7.15
7.55
7.59
7.15
7.54
7.05
6.85
7.24
7.46
7.06
1.42
1.42
2.02
2.15
2.19
7.63
7.10
2.15
2.21
2.22
2.08
2.11
1.23
1.23
2.22
2.16
27.9
27.9
25.9
12.4
20.5
20.0
12.2
12.1
136.3
135.4
130.3
128.1
134.7
136.4
134.4
139.0
157.2
134.7
148.1
138.9
172.0
138.5
160.8
17.0
1.91
2.74
7.01
6.96
6.60
6.60
7.35
6.62
6.56
7.32
6.73
6.72
8.03
6.91
6.77
7.67
1.24
1.24
1.92
2.14
2.27
6.99
8.07
2.24
2.23
2.28
2.36
2.46
δC ) RF + δ0
(1)
27.1
13.7
12.4
13.8
20.2
20.8
13.6
14.6
where R is the total change in 13C chemical shift and δ0 is the
average 13C chemical shift of all sp2 carbons in the uncharged
molecule. O’Brien69 found the best fit with δC ) 156.3F +
133.2. Surveying five other studies, Tolbert and Ogle36 found
that δC ) 156.2F + 131.9 was the best equation to fit these
data, while they in their study of a series of homologue R,ω-
diphenylpolyenyl anions found somewhat higher values, R )
187.3 and δ0 ) 133.7. A significantly lower change in δC per
charge has been found for aromatic di- and tetraanions. This
was attributed to anisotropic ring current effects.71
Scheme 5
In our studies of positively charged carotenoids, we have
found a much higher change in chemical shifts, ca. 250 ppm
per charge,54,55 and this seems to be a general value for these
compounds. Fitting the data obtained for monocation 6 with eq
1, we obtain eq 2.
δC ) 249.1F + 132.7
(2)
eratene (20)67 are available. To provide the best neutral reference
Solving this equation for F, the π-charge density at each carbon
may be calculated. The result is plotted in Figure 1. The plot
data possible for determination of the charge distribution of the
1
cations 6, 8, 13, and 17, all H and 13C chemical shifts were
(68) Mortensen, A.; Skibsted, L. H. J. Agric. Food Chem. 2000, 48, 279.
(69) Spiesecke, H.; Schneider, W. G. Tetrahedron Lett. 1961, 3, 468.
(70) O’Brien, D. H.; Hart, A. J.; Russell, C. R. J. Am. Chem. Soc. 1975, 97,
4410.
(71) Eliasson, B.; Edlund, U.; Muellen, K. J. Chem. Soc., Perkin Trans. 2 1986,
7, 937.
determined for both carotenoids. A synthetic sample of iso-
carotene (10) was obtained, which contained both the 6-Z and
the 6-E isomeric half-molecules, present in a 1:6 ratio. Isore-
nieratene (20) was prepared in a double Wittig reaction from
9
8986 J. AM. CHEM. SOC. VOL. 126, NO. 29, 2004

Delocalized Carotenoid Cations
A R T I C L E S
using eq 4 gives a charge distribution that is much more
symmetrically located in the center of the polyene chain, as
compared to eq 2, as is expected for a soliton.
From the soliton theory,28-30 it may be deduced that the
accumulated charge, Qi, on a polyene chain can be expressed
by eq 5.32,33
Qi ) Qtot/2 * tanh[(i - ic)/l]
(5)
Here, l is the soliton charge-wave half-width, and i is the number
of carbons, when the carbon number ic gives the center of the
soliton charge-wave. In Figure 2 is plotted the accumulated
charge, Qi, for 6, against the charge-carrying carbons i, setting
ic ) C-14, where Qi ) 0.5. The charge for the charge-carrying
carbons has been renormalized to 1 (89% of downfield shift is
located on these 10 carbons), so that Qtot can be set to 1, in
accordance with the charge of +1. Curve fitting with eq 5 then
gives l ) 7.8, in good agreement with l ) 7-9, found by
Reimers et al.33 using AM1 calculations.
Figure 1. Charge distribution for monocation 6, from eq 2 (red) and eq 4
(black).
The deviations between the fitted curve and the experimental
curve toward the ends of the plot, taken together with the charge
localized on carbon atoms not included in the plot, may indicate
that the polyene chain of monocation 6 is not long enough to
avoid effects from the ends of the polyene chain, thus
underestimating the half-width of the charge-wave. Nevertheless,
these results indicate that the value l ) 14 for the soliton half-
width found by Tolbert and Ogle36 using O’Briens hypothesis70
is an overestimate as previously pointed out by Reimers et al.33
In a recent paper treating the protonation of carbonyl groups
Figure 2. Accumulated charge (Qi) for monocations 6 versus carbon
number i (setting middle of soliton to i ) 0 for C-14) (black) and regression
curve for Qi ) 0.5 tanh(i/7.8), R2 ) 0.997 (red).
1
in canthaxanthin (3), we also prepared and determined the H
clearly shows that the resulting charge distribution does not
resemble the expected distribution of a soliton, with a concen-
tration of the charge in the center of the polyene chain, consistent
with the warnings by Reimers et al.33 that using O’Briens
postulate70 for determining the charge on individual carbons may
give quantitative errors.
A better model has been devised by employing the change
in carbon chemical shift at each position, ∆δC,i, and correlating
this to the total change in chemical shift ∑∆δC. This leads to
eq 3,
and 13C chemical shifts of a monocation protonated in the C-7
position (23). In 23, the charge may be delocalized over 19 sp2
carbon atoms (not including the carbonyl), as compared to the
23 sp2 hybridized carbon atoms available for delocalization in
6. Applying eqs 1 and 3 for determination of the charge
distribution in 23 again shows that eq 3 gives a better estimate;
see Figure 3. The deviation of the alternating charge distribution
to the right side of the molecule is ascribed to the carbonyl
moiety. Some of the charge is being delocalized into the
carbonyl moiety (including the â-carbon), forcing the main
charge away from the carbonyl-containing end group. This also
results in a compression of the soliton.
F ) ∆δ / ∆δ * Q
tot
(3)
∑
i
C,i
C
1
where Qtot is the total charge on the polyene. The H and 13C
chemical shift changes of 6a and 6b relative to the isocarotene
(10)/â,â-carotene (1) neutral model described previously57 are
given in the Supporting Information, showing small deviations
in some positions when using the new data obtained here for
isocarotene (10), as compared to the previously published shift
changes.57 The total shift change of 6a was found to be 13.8
A plot of the accumulated charge for 23, according to eq 5,
is also shown in Figure 3, together with the regression curve,
giving the charge-wave half-width l ) 6.2, shorter than that
found for 6. This shows that the number of carbon atoms
available for charge delocalization in 23 is too small to represent
a free soliton.
The SSH soliton model28-30 also allows for the determination
of the area of bond-length alternations, expressed as
ppm for H and 254.2 ppm for 13C. Of the 13C chemical shift
1
change, 249.1 ppm, or 98%, was associated with the polyene
chain. Only minor differences are seen in the results for the
isomer 6b, and the results for the major cation 6a are therefore
used for the treatment of 6 below. Inserting the values found
for 6a into eq 3 leads to eq 4 for calculation of the charge
distribution of 6.
∆ri ) ∆r∞ tanh[(i - ic)/l]
(6)
where ∆r∞ is the difference between single and double bond
lengths in an infinite chain, and l is the soliton half-width.33 In
principle, the bond lengths of the E and s-trans bonds in
polyenes may be determined from the 3JH,H coupling constant.72
For the carotenoid monocations, the methyl groups on the
polyene chain complicate this determination (although simplify-
Fi ) ∆δC,i/249.1
(4)
The charge distribution calculated for 6 by using eq 4 is shown
in Figure 1. It may immediately be seen from Figure 1 that
(72) Friebolin, H. Basic One- and Two-Dimensional NMR Spectroscopy; VCH
Verlagsgesellschaft: Weinheim, 1993.
9
J. AM. CHEM. SOC. VOL. 126, NO. 29, 2004 8987

A R T I C L E S
Lutnaes et al.
Figure 5. Charge distribution for the 7-hydro-φ,φ-carotenyl cation (21)
(top), and for the 4-dehydro-4′-oxo-â,â-carotenyl cation (17) (bottom),
calculated using eq 1 (red) and eq 3 (black). For 21, the charge in the ortho
and meta positions in the aromatic end group has been added (bottom).
Figure 3. Charge distribution for monocation 23, calculated from eq 1
(red) and eq 3 (black) (top). Accumulated charge, Qi, plotted again carbon
atoms along the polyene chain for 23 (black), and curve fitting (red), giving
Qi ) 0.5 tanh(i/6.2), R2 ) 0.998 (bottom).
13C chemical shifts for echinenone (9)67 and 1H chemical shifts
for canthaxanthin (3)74 were used as a model for the right half
in the neutral reference for monocation 17, the cation formed
by elimination of hydride from the 4′-position of echinenone
(9). For the isorenieratene monocation (21), 7,8-dihydro-8,7′-
retro-â,â-carotene-4,4′-dione (24)67 was used as neutral model
for C-7 - C-11,C-19,H-7 - H-10 and H-19 and isorenieratene
(20) for C-12 - C-15,C-20,C-1′ - C-20′,H-11 - H-15,H-20
and for H-3′ - H-20. No suitable model for the nonconjugated
aromatic end group of monocation 21 was available, but as this
end group is not carrying the charge, these data were not
necessary for the determination of the charge distribution. The
change in chemical shifts at each position is given in Table B
in the Supporting Information, and the charge at each position
is shown in Figure 5 and by the diameter of the filled circles
on the structures in Chart 2. For comparison, the charge
distribution calculated using eq 1 is also shown in Figure 5.
Figure 4. Correlation between coupling constants and experimental (red)
and calculated (black) bond lengths for â,â-carotene (1).
ing the determination of the cation structure itself) by affecting
the coupling constants through substituent effects, and by
3
reducing the number of JH,H coupling constants available for
Also, for these carotenoid cations, it may be seen that eq 3
gives a better estimate for the charge distribution than does eq
1. When the charge distribution of monocation 17 is compared
to that of monocation 23, prepared from canthaxanthin (3), it
is interesting to note the strong charge repulsion from the
carbonyl moiety. Some charge is stored in the carbonyl moiety,
as evidenced from the downfield shifts of the C-6′ carbon in
the â-position to the carbonyl. The chemical shift difference of
the carbonyl carbon does not provide a good measure for the
charge on this carbon, as charge delocalization to the carbonyl
will lead to a change in CdO bond hybridization, which will
also influence the chemical shift.55,75,76 However, in comparison
with carotenoid cations previously analyzed,55 it is clear that
bond-length determination. The coupling constants of â,â-
carotene (1) did not give a good correlation with the bond
lengths determined by X-ray crystallography;73 see Figure 4. A
better correlation was obtained by correlation to bond lengths
determined using AM1 calculations.65 A reason for this differ-
ence may be a difference in structure in the solid state, but a
higher resolution X-ray structure might have given better
correlation. This approach for determination of the soliton width
was abandoned due to the low correlation between the bond
length and the determined coupling constants combined with
the few coupling constants available for the cations.
Charge Distribution of Other Carotenoid Monocations.
The charge distribution for monocations 17 and 21, prepared
from echinenone (9) and isorenieratene (20), respectively, was
determined from the change in chemical shifts according to eq
3. Isocarotene (10) was used as model for the left half, while
(74) Schwieter, U.; Englert, G.; Rigassi, N.; Vetter, W. Pure Appl. Chem. 1969,
20, 365.
(75) Olah, G. A.; Berrier, A. L.; Prakash, G. K. S. J. Am. Chem. Soc. 1982,
104, 2373.
(76) Olah, G. A.; Burrichter, A.; Rasul, G.; Gnann, R.; Christe, K. O.; Prakash,
G. K. S. J. Am. Chem. Soc. 1997, 119, 8035.
(73) Sterling, C. Acta Crystallogr. 1964, 17, 1224.
9
8988 J. AM. CHEM. SOC. VOL. 126, NO. 29, 2004

Delocalized Carotenoid Cations
A R T I C L E S
Chart 2
Figure 6. Charge distribution on the polyene chain as determined by eq 3
for the all-E (8c) (black) and 6,6′-di-Z (8a) (red) dication of isocarotene
(10).
the data obtained for monocation 17 are not compatible with a
protonation of the carbonyl moiety.
Using eq 5 for the analysis of the solitons in dication 8 is
not straightforward. In principle, the dication could be treated
as the sum of two solitons. However, the shape of the charge
distribution shows large deviations from the shape of a soliton
due to the strong Coulombic repulsion between the two charges,
and in the center of the polyene chain the contributions from
the two solitons are overlapping.
For monocation 21, prepared from isorenieratene (20), the
most important feature is the negligible contribution to the
charge delocalization of the aromatic end group. There is a
strong driving force to avoid a reduction in electron density in
the aromatic ring, which would cause a reduction in the Hu¨ckel
(4n + 2) aromaticity.
Both the carotenoid monocations 17 and 21 have soliton
widths that are smaller than that of monocation 6, providing no
further insight into the width of a free soliton. The total 13C
chemical shift change for 17 and 21, 253.2 and 240.8 ppm (not
including the nonconjugated end group), respectively, is con-
sistent with the carotenoid cations studied previously.53-55,57
Structures of Dications 8 and 13, and the Structural
Relationship between Dications 2 (20 πe-) and 8 (22 πe-).
By the determination of all 1H and 13C chemical shifts of
dication 8 (Scheme 3) and of isocarotene (10), it became
possible to calculate the chemical shift change of all of the sp2-
hybridized carbon atoms for the isocarotene dication (8) relative
to neutral 10. The chemical shift differences for all positions in
the two E/Z isomeric half-molecules of 8 are shown in Table C
in the Supporting Information.
Some information may be gained, however, by comparison
of the isocarotene dication (8) with the â,â-carotene dication
(2) previously reported,53,54 which is 2 carbon units shorter, and
contains 20, rather than 22 π-electrons. The most interesting
feature is the decreasing charge on the terminal carbon atoms
in 8. In the â,â-carotene dication (2), the Coulombic repulsion
is high enough to completely erase the wave-shape of the soliton,
forcing the highest charge density to the terminal carbons.
Opposing the Coulombic repulsion force is the energy gain in
storing the charge in a wave-shaped soliton. The longer polyene
chain in dication 8 allows one more carbon for delocalization
of each charge, and the lower charge on the terminal carbons
may be an indication that the Coulombic repulsion is reduced
sufficiently to make the wave-shape of the soliton apparent.
However, some caution must be exercised not to interpret these
data too far. The stabilization of the charge by the tertiary
terminal carbons in 2, as compared to the secondary terminal
carbons in 8, may play a role, and a longer conjugated dication
would be needed to confirm this hypothesis.
The total 13C downfield shift for dication 8 is 240 ppm per
charge, in good agreement with the 252 ppm 13C downfield
shift previously found per charge for the â,â-carotene dication
(2),53 and the value found for monocation 6 (259.8 ppm). The
1
total downfield chemical shift for H upon the formation of a
The charge distribution for dication 13, containing a proto-
nated hydroxy group, Scheme 3, could not be completely
determined from the recorded NMR data. As judged by the
carotenoid cation is also surprisingly uniform in the range 12.4-
13.8 ppm per charge.54,55,57
1
From the chemical shift change data, the charge distribution
for the E/E (8c) and Z/Z (8a) isomers was calculated, as shown
in Figure 6, where two solitons are apparent. Due to the
Coulombic repulsion of the two charges, they are located toward
the ends of the polyene chain, as predicted by the SSH theory,
rather than being centered in the middle, or being evenly
distributed. For the Z/Z isomer (8a), it may be seen that the
charge delocalization to the C-4,5 double bond is smaller than
that for the E/E isomer (8c), as may be expected from the lower
overlap of the π-orbitals of the terminal C-4,5 double bonds.
The lower charge on the terminal double bond is compensated
by a somewhat higher charge on the carbons next to the terminal
double bonds in the Z/Z isomer (8a). The charge density on
each carbon for the all-E isomer 8c is illustrated by the diameter
of the filled circles on the structure in Chart 2.
available H and 13C chemical shift differences, the charge
distribution of 13, with one charge on the polyene chain, is very
similar to that of monocation 6, as expected. However, the center
of the soliton is shifted somewhat away from the end of the
molecule carrying the protonated hydroxy group. This is
interpreted as a result of the Coulombic repulsion from the
charge on the protonated hydroxy group and is another piece
of indirect evidence for the presence of charge on an oxonium
moiety in one end group.
NIR Absorption and Stability of Carotenoid Cations.
Comparison of the NIR spectra of the carotenoid carbocations
studied revealed increasing λmax for the isorenieratene mono-
cation (21, 871 nm, Figure 7) < â,â-carotene dication (2, 925
nm54) < echinenone carbocation (17, 953 nm, Figure 7) <
dication 13 (982 nm, Figure 7) < isocarotene dication 8 (1022
9
J. AM. CHEM. SOC. VOL. 126, NO. 29, 2004 8989

A R T I C L E S
Lutnaes et al.
were elucidated by NMR spectroscopy in the temperature range
from -10 to -20 °C.
1
The NMR analysis was based on 1D H, and a series of 2D
(1H-1H COSY, 2D ROESY, 1H-13C HSQC, and 1H-13C
HMBC) experiments. A general strategy used for the determi-
nation of mixtures with up to four stereoisomers is outlined.
For determination of the charge distribution of the carotenoid
monocations 6, 17, 21, and 23, it was found that correlation of
the change in 13C chemical shift of the charged versus the neutral
carotenoid model gave a better result than correlation of the
charge against 13C chemical shift, in accordance with the
O’Brien implementation70 of the Spiesecke-Schneider relation-
ship.69 It may be inferred that this result is generally valid for
sp2-hybdridized polyene cations, while for a nonsubstituted
polyene (or polyacetylene) cation of infinite chain length, the
result from using this method and the results from using the
O’Brien postulate70 should be equal, because the chemical shifts
of all carbon atoms are equal. The disadvantage of the method
employed here is the necessity of knowing the chemical shifts
of a relevant neutral model.
Figure 7. VIS/NIR spectra of isocarotene dicarbocation (8), dication 13,
monocation 17, and the isorenieratene monocation (21).
nm, Figure 7) < monocation 6 (1028 nm57). A rationalization
of these data is not simple because the method of preparation/
counterion, size of the charge, length of the polyene chain,
solvent, and temperature have to be considered.
The charge distribution determined by the change in 13C
chemical shifts was used to calculate the soliton width. For
monocation 6, with charge delocalized over 23 carbon atoms,
the soliton half-width was found to be 7.8 carbons, in good
agreement with previous AM1 calculations for a free soliton.
For the shorter-chain monocation 23, with the charge delocalized
on 19 sp2 carbon atoms, the soliton half-width was found to be
only 6.2 carbons, indicating that cations with longer delocal-
ization of the charge than over 23 carbon atoms may be
necessary for unequivocal determination of the half-width of a
free soliton.
In the VIS/NIR spectra of the carotenoid cations, the charged
state seems to be more important than the length of the polyene
chain. This may explain why the stability measured for the
cations is much higher when determined from the NIR spectra
as compared to what is found in the NMR experiments. Even
though there are reactions going on, which makes the reaction
mixture more and more complex, and thereby making it
impossible to perform NMR experiments on the sample, there
are still cations present that will show an absorption around
900-1000 nm in the VIS/NIR spectrum. This also provides an
explanation as to why the color of the cation solution may retain
its black color for weeks on the lab bench.
The charge distribution of the 22 π electron dication 8 showed
that upon increasing the number of carbons available for charge
delocalization from 22 (20 π electron cation 2) to 24, the
Coulombic repulsion is no longer strong enough to force the
charge to the far ends of the polyene chain. Thus, the appearance
of two solitons can be seen in the isocarotene dication 8.
Conclusion
The stereoisomeric monocations 6a, 6b, 6c, and 6d with 22
delocalized π-electrons were prepared from â,â-caroten-4-ol (5)
by reaction with CF3COOH. Treatment of â,â-carotene-4,4′-
diol (7) with the same acid resulted in a dication (13) with the
protonated hydroxy group intact.
Three stereoisomeric dicarbocations 8a, 8b, and 8c also with
22 delocalized π-electrons were obtained from â,â-carotene-
4,4′-diol (7) by reaction with the stronger Brønsted acid CF3-
SO3H.
Another four carotenoid cations were prepared and completely
characterized by treatment of echinenone (9) with CF3COOH
(17, 18, 19) and by treatment of isorenieratene (20) with BF3-
THF-d8 etherate (21).
Detailed structures of these delocalized cations including the
charge distribution, bond types, and regions of bond inversion
Uncritical use of NIR spectra for the identification and
determination of the stability of carotenoid cations has been
discussed.
Acknowledgment. Hoffmann-La Roche, Basel, is gratefully
acknowledged for a research grant to S.L.-J., as well as for gifts
of synthetic carotenoids.
Supporting Information Available: Detailed experimental
procedures and additional spectral data. This material is available
free of charge via the Internet at http://pubs.acs.org.
JA0492541
9
8990 J. AM. CHEM. SOC. VOL. 126, NO. 29, 2004