A solution- and solid-state investigation of medium effects on charge separation in metastable photomerocyanines

Published on Web 08/23/2010

A Solution- and Solid-State Investigation of Medium Effects

on Charge Separation in Metastable Photomerocyanines

†

‡

†

Dinesh G. Patel, Michelle M. Paquette, Roni A. Kopelman, Werner Kaminsky,

§

,†,‡

Michael J. Ferguson, and Natia L. Frank*

Department of Chemistry, UniVersity of Washington, P.O. Box 351700, Seattle, Washington 98195,

Department of Chemistry, UniVersity of Victoria, P.O. Box 3065, Victoria, BC V8W 3V6, Canada,

and Department of Chemistry, UniVersity of Alberta, Edmonton, AB T6G 2G2, Canada

Received January 11, 2010; E-mail: nlfrank@uvic.ca

Abstract: The effects of solution-state dielectric and intermolecular interactions on the degree of charge

separation in metastable spirooxazine photomerocyanines (PMCs) is investigated. We report the ﬁrst X-ray

diffraction (XRD) analyses of an open form, a metastable photomerocyanine, of the spirooxazine class of

photochromic molecules in two derivatives: spiro[azahomoadamantane-isoquinolinoxazine] (1) and spiro-

[azahomoadamantane-phenanthrolinoxazine] (2). Using the results of XRD analysis of the open photomero-

cyanine forms, in conjunction with computation, solvatochromism, and solution NMR studies, we have investigated

the effect of the medium on the ground-state structure of these photomerocyanines. Solvatochromism and NMR

chemical shift studies of 1 and 2 support the assignment of a quinoidal structure in nonpolar solvents and a

zwitterionic structure in high-polarity solvents. The effect of azahomoadamantyl substitution is explored by

comparing 1 and 2 with the analogous indolyl derivatives, spiro[indoline-isoquinolinoxazine] (3) and spiro[indoline-

phenanthrolinoxazine] (4) through XRD analysis of the closed spirooxazine (SO) forms, solution-state kinetic

experiments, solvatochromism, and NMR studies. Longer Cspiro-O bond lengths in the SO form and slower

rates of thermal PMC f SO isomerization for the azahomoadamantyl derivatives are associated with greater

zwitterionic character in the PMC form, as found in the solvatochromism studies. XRD analysis of photomero-

cyanines 1 and 2 indicate a greater contribution from the canonical zwitterionic resonance form relative to the

quinoidal form in the solid state. Structural differences observed in two pseudopolymorphs of 2-PMC suggest

that the degree of charge-separated character is inﬂuenced by the crystal packing environment. These results

provide direct structural evidence for the effects of the medium polarity on charge-separated states of

photomerocyanines.

4

Introduction

responses, and a tendency to organize into ordered aggregates

1

2

sensitive to external electric ﬁelds. The spirooxazine class in

particular is technologically attractive due to its high fatigue

Photochromic molecules that undergo photoisomerization

between two optically and chemically distinct states

great interest for reversible memory photodevices, optical

switching elements in molecular electronics, and multifunc-

1

-3

are of

1

4

resistance, which has led to the successful incorporation of

1

3

5

spirooxazines into light-sensitive ophthalmic lenses. In addi-

tion, spirooxazines are under current investigation for optical

6

-9

tional materials.

Of the known classes of photochromic

molecules, spiropyrans and spirooxazines are distinguished by

their merocyanine photoproducts which exhibit unique transi-

1

0,11

(8) Mikami, R.; Taguchi, M.; Yamada, K.; Suzuki, K.; Sato, O.; Einaga,

tion-metal binding modes,

strong nonlinear optical (NLO)

Y. Angew. Chem., Int. Ed. 2004, 43, 6135–6139.

(

9) B e´ nard, S.; L e´ austic, A.; Rivi e` re, E.; Yu, P.; Cl e´ ment, R. Chem.

Mater. 2001, 13, 3709–3716.

†

University of Washington.

University of Victoria.

University of Alberta.

‡

(10) Kume, S.; Nishihara, H. Metal-based photoswitches derived from

photoisomerization. In Struct. Bonding (Berlin); Springer-Verlag:

Berlin, 2007; Vol. 123, pp 79-112.

(11) Kopelman, R. A.; Snyder, S. M.; Frank, N. L. J. Am. Chem. Soc.

2003, 125 (45), 13684–13685.

(12) Krongauz, V. A.; Parshutkin, A. A. Photochem. Photobiol. 1972, 15

(5), 503–507.

§

(

1) Bertelson, R. C. Chapter III: Photochromic Processes Involving

Heterolytic Cleavage. In Techniques of Chemistry Vol. III: Photo-

chromism; Brown, G. H., Ed.; Wiley-Interscience: New York, 1971.

2) Crano, J. C.; Guglielmetti, R. J. Organic Photochromic and Ther-

mochromic Compounds; Kluwer Academic Publishers: New York,

(

2

002; Vol. 1, p 2.

(13) Crano, J. C.; Flood, T.; Knowles, D.; Kumar, A.; Van Gemert, B.

Pure Appl. Chem. 1996, 68 (7), 1395–1398.

(14) Kawata, S.; Kawata, Y. Chem. ReV. 2000, 100, 1777–1788.

(15) Yuan, W. F.; Sun, L.; Tang, H. H.; Wen, Y. Q.; Jiang, G.; Huang,

W.; Jiang, L.; Song, Y. L.; Tian, H.; Zhu, D. B. AdV. Mater. 2005,

17 (2), 156–160.

(16) Cheng, T.; Lin, T.; Brady, R.; Wang, X. Fibers Polym. 2008, 9 (5),

521–526.

(17) Mennig, M.; Fries, K.; Lindenstruth, M.; Schmidt, H. Thin Solid Films

1999, 351 (1-2), 230–234.

(

3) D u¨ rr, H.; Bouas-Laurent, H. Photochromism: Molecules and Systems;

Elsevier: New York, 2003.

(

4) Berkovic, G.; Krongauz, V.; Weiss, V. Chem. ReV. 2000, 100 (5),

1

741–1753.

(

5) Feringa, B. L. Molecular Switches; Wiley-VCH: New York, 2001.

6) Einaga, Y.; Taguchi, M.; Li, G. M.; Akitsu, T.; Gu, Z. Z.; Sugai, T.;

Sato, O. Chem. Mater. 2003, 15 (1), 8–10.

(

7) Taguchi, M.; Yamada, K.; Suzuki, K.; Sato, O.; Einaga, Y. Chem.

Mater. 2005, 17 (17), 4554–4559.

1

2568 9 J. AM. CHEM. SOC. 2010, 132, 12568–12586

10.1021/ja100238h  2010 American Chemical Society

Charge Separation in Metastable Photomerocyanines

A R T I C L E S

Scheme 1

described by the relative contributions of the quinoidal (A,

Scheme 1) and zwitterionic (B, Scheme 1) resonance forms to

the ground-state electronic structure, remains ambiguous. Com-

putational studies of photochromic spirooxazine PMCs predict

2

9,30

25

both zwitterionic

and quinoidal ground-state structures.

Solvatochromism studies suggest either a quinoidal or zwitte-

rionic ground state depending on substituent and solvation

3

1,32

effects.

Nucleophilic trapping studies of metastable pho-

tomerocyanines suggest a dominant zwitterionic structure in

3

solution. Structural analysis in the solid state by X-ray

diffraction (XRD) through crystallization of metastable pho-

3

4,35

tomerocyanines has been impeded by their transient nature

and the spontaneous formation of molecular aggregates upon

36,37

irradiation.

In contrast, reports of structural characterization

of the metastable photomerocyanine form of the related class

of spiropyrans exist, but are rare. XRD analyses of a thiospiro-

4

,14,15

16,17

13

38

electronic devices and memories,

displays, and switches.

coatings,

ﬁlters,

pyran PMC form and a small number of spiropyran PMC

4

18,19

39-43

forms

suggest that spiropyran photomerocyanines with

Spirooxazines undergo photochemical or thermal isomeriza-

tion between a spirooxazine (SO) and a photomerocyanine

electron-withdrawing nitro and pyridinium substituents on the

oxazine moiety exhibit dominant zwitterionic character. XRD

analyses of two permanently open (i.e., nonphotochromic)

spirooxazines indicate that the merocyanines exhibit a predomi-

(

PMC) form (Scheme 1). In the SO form, the azaheterocyclic

and oxazine moieties are held orthogonal through a spirocyclic

linkage, which leads to electronic transitions in the near-UV

region (λmax ≈ 350 nm). UV-light-induced cleavage of the

4

4,45

nantly quinoidal structure.

To date, the structural charac-

terization of the open form of the photochromic spirooxazine

class of photochromic molecules has not been carried out by

XRD analysis. Consequently, controversy persists over the

relative contributions of the quinoidal and zwitterionic resonance

forms to the ground-state electronic structure of spirooxazine

C

spiro-O bond in the SO form leads to ring-opening and

formation of the metastable PMC species, which exhibits

extended conjugation through the central azomethine bridge and

an intense electronic absorption at λmax ≈ 600 nm. Ring closure

can be induced photochemically by irradiation into the low-

energy PMC π-π* transition with visible light. The forward

and reverse isomerization reactions can also occur thermally

2

5,31,33

PMCs and the explicit role of medium and substitution.

Herein we investigate medium and substitution effects on the

charge-separated character of photomerocyanines through a

series of solution-state studies focused on bicyclic azahomoada-

mantyl spirooxazines. These spirooxazine derivatives exhibit

3

along the ground-state potential energy surface.

The rates of thermal isomerization, photocolorability, and

fatigue resistance critical for current applications of spiroox-

azines are controlled by subtle changes in the electronic structure

(

26) Perrier, A.; Maurel, F.; Perp e` te, E. A.; Wathelet, V.; Jacquemin, D.

2

,3

of the SO and PMC forms. While optical coatings and ﬁlters

J. Phys. Chem. A 2009, 113 (46), 13004–13012.

1

7,20

require rapid thermal relaxation,

supported by decreased

(27) Hobley, J.; Malatesta, V. Phys. Chem. Chem. Phys. 2000, 2 (1), 57–

5

9.

stability of the PMC form, nonvolatile memories require stability

(

28) Delbaere, S.; Bochu, C.; Azaroual, N.; Buntinx, G.; Vermeersch, G.

4

of the photogenerated form. The structure and reactivity of

J. Chem. Soc., Perkin Trans. 2 1997, (8), 1499–1501.

metastable photomerocyanines dictate their utility in information

storage and molecular electronics applications and depend

critically on the degree of charge separation and polarizability

(29) Pommier, H.; Samat, A.; Guglielmetti, R. Mol. Cryst. Liq. Cryst.

Sci. Technol., Sect. A 1994, 246, 241–246.

(

30) Malatesta, V.; Ranghino, G.; Romano, U.; Allegrini, P. Int. J.

Quantum Chem. 1992, 42 (4), 879–887.

2

1-24

of the ground and excited states.

Our understanding of the

(31) Pozzo, J.-L.; Samat, A.; Guglielmetti, R.; De Keukeleire, D. J. Chem.

Soc., Perkin Trans. 2 1993, (7), 1327–1332.

electronic and structural factors governing spirooxazine pho-

toisomerization dynamics and charge distribution is limited,

however, by a lack of structural characterization of the

metastable photomerocyanine form in the solid state. Compu-

tational and spectroscopic studies suggest that the trans-trans-

cis (TTC) geometry of photomerocyanines is generally the most

(

32) Metelitsa, A. V.; Lokshin, V.; Micheau, J. C.; Samat, A.; Guglielmetti,

R.; Minkin, V. I. Phys. Chem. Chem. Phys. 2002, 4 (18), 4340–

4

345.

(33) Malatesta, V.; Neri, C.; Wis, M. L.; Montanari, L.; Millini, R. J. Am.

Chem. Soc. 1997, 119 (15), 3451–3455.

(

34) Minkin, V. I. Chem. ReV. 2004, 104 (5), 2751–2776.

35) Tamai, N.; Miyasaka, H. Chem. ReV. 2000, 100 (5), 1875–1890.

25-28

stable conformation,

while the degree of charge separation,

(36) Onai, Y.; Mamiya, M.; Kiyokawa, T.; Okuwa, K.; Kobayashi, M.;

Shinohara, H.; Sato, H. J. Phys. Chem. 1993, 97 (37), 9499–9505.

(

37) Krongauz, V. A.; Fishman, S. N.; Goldburt, E. S. J. Phys. Chem.

1978, 82 (23), 2469–2474.

(

18) Cheong, I. W.; Wang, S.; Ki, H. S.; Kim, S. H. Curr. Appl. Phys.

2

009, 9 (6), 1269–1271.

(38) Hirano, M.; Osakada, K.; Nohira, H.; Miyashita, A. J. Org. Chem.

2002, 67 (2), 533–540.

19) Paquette, M. M.; Kopelman, R. A.; Beitler, E.; Frank, N. L. Chem.

Commun. 2009, (36), 5424–5426.

(39) Naumov, P.; Yu, P.; Sakurai, K. J. Phys. Chem. A 2008, 112 (26),

5810–5814.

20) Hou, L.; Schmidt, H. Mater. Lett. 1996, 27 (4-5), 215–218.

(

21) Marder, S. R.; Beratan, D. N.; Cheng, L.-T. Science 1991, 252 (5002),

(40) Artemova, N. K.; Smirnov, V. A.; Rogachev, B. G.; Shilov, G. V.;

Aldoshin, S. M. Russ. Chem. Bull. 2006, 55 (9), 1605–1611.

(41) Hobley, J.; Malatesta, V.; Millini, R.; Montanari, L.; Parker, W. O. N.

Phys. Chem. Chem. Phys. 1999, 1 (14), 3259–3267.

1

03–106.

22) Marder, S. R.; Perry, J. W.; Bourhill, G.; Gorman, C. B.; Tiemann,

B. G.; Mansour, K. Science 1993, 261 (5118), 186–189.

23) Marder, S. R.; Gorman, C. B.; Meyers, F.; Perry, J. W.; Bourhill,

G.; Br e´ das, J.-L.; Pierce, B. M. Science 1994, 265 (5172), 632–635.

24) Teshome, A.; Kay, A. J.; Woolhouse, A. D.; Clays, K.; Asselberghs,

I.; Smith, G. J. Opt. Mater. 2009, 31 (4), 575–582.

(42) Aldoshin, S. M. Russ. Chem. ReV. 1990, 59 (7), 663–685.

(43) Aaker o¨ y, C. B.; Hurley, E. P.; Desper, J.; Natali, M.; Douglawi, A.;

Giordani, S. CrystEngComm 2010, 12 (4), 1027–1033.

(44) Lar e´ ginie, P.; Lokshin, V.; Samat, A.; Guglielmetti, R.; P e` pe, G.

J. Chem. Soc., Perkin Trans. 2 1996, (1), 107–111.

25) Nakamura, S.; Uchida, K.; Murakami, A.; Irie, M. J. Org. Chem.

1

993, 58 (20), 5543–5545.

(45) Song, H. F.; Chen, K. C.; Tian, H. Dyes Pigm. 2005, 67 (1), 1–7.

J. AM. CHEM. SOC. 9 VOL. 132, NO. 36, 2010 12569

A R T I C L E S

Patel et al.

Chart 1

for their efﬁcacy in reproducing structure-property relationships

in metastable photomerocyanines, and the role of speciﬁc

intermolecular interactions in both the solution and solid states

found to be quite signiﬁcant in these systems, leading to poor

agreement between theory and experimental geometries for the

PMC forms.

This work provides an analysis of structure-property rela-

tionships in spirooxazines and highlights the use of azaho-

moadamantyl substitution as a powerful and unexplored ap-

proachtooptimizingphotochromicﬁguresofmeritinspirooxazines

for optical applications. Further, the direct structural evidence

for the charge-separated character of metastable photomero-

cyanines and its sensitivity to medium polarity in both the

solution and solid states is relevant to furthering our understand-

ing of medium effects in organic systems involving charge

separation. Discussions of the relative contributions of quinoidal

and zwitterionic character in a wide range of organic systems

have existed for some time. Investigations in spiroheterocy-

attractive properties such as crystalline-state photochromism,

exceptional stability of the PMC form, high photocolorabilities,

and slow thermal relaxation rates relative to their highly studied

46

indolyl-based counterparts. We report the ﬁrst crystalline-state

structural analysis for metastable photomerocyanines of the

spirooxazine class of photochromic molecules in two derivatives:

spiro[azahomoadamantane-isoquinolinoxazine] (1) and spiro[aza-

homoadamantane-phenanthrolinoxazine] (2) (Chart 1). The

molecular structures of 1-PMC and 2-PMC are interpreted in

the context of their relative quinoidal and zwitterionic resonance

contributions. The inﬂuence of the crystalline lattice on the

charge-separated character of 2-PMC is analyzed in two

34

48-50

51

clics, nonlinear optical materials,

electro-optic materials,

52

53

conjugated polymers, donor-acceptor chromophores, enol-

54,55

keto tautomerizations in the solid state,

antiaromatic π-sys-

5

6

57-60

tems, and C-O rotational barriers in esters and amides

have focused on either computation, single-crystal X-ray

analysis, or solution-based studies. While it has been argued

that there is a dependence on solvent polarity, direct structural

evidence for medium effects in the crystalline state are valuable

toward discussions of ground-state electronic structures of

organics in which some degree of charge separation exists.

47

crystalline pseudopolymorphs, in which the degree and nature

of solvation differ: 2-PMC-I and 2-PMC-II (a trihydrate). The

molecular structures of the photomerocyanines 1-PMC and

Results

2

-PMC in the single crystalline state are compared with their

solution-state structures as inferred from analysis of the

Synthesis of Spirooxazines. The general methodology for the

synthesis of spirooxazines involves the condensation of an

vibrational structure and solvatochromism of the PMC π-π*

1

13

absorption bands combined with H and C NMR studies in a

wide range of solvent polarities. The effect of azahomoada-

mantyl substitution on charge separation and associated proper-

ties is explored by complementary studies on the analogous

indolyl derivatives, spiro[indoline-isoquinolinoxazine] (3) and

spiro[indoline-phenanthrolinoxazine] (4) (Chart 1). Geometrical

parameters in crystal structures of 1-SO, 3-SO, and 4-SO are

compared within the context of relationships between oxazine

geometry and photoisomerization dynamics, and are found to

be consistent with that known for the wider class of spiroox-

azines. Solution-state isomerization kinetics and solvatochromic

behavior of the azahomoadamantyl and indolyl derivatives are

contrasted to examine the effect of the amine functionality on

electronic structure. The stability of the metastable PMC form

observed in the azahomoadamantyl derivatives correlates with

the high degree of zwitterionic character. Finally, density

functional theory (DFT) geometry optimizations are evaluated

enamine with an o-hydroxy-nitroso arene under dehydrating

34,61,62

conditions.

The o-hydroxy-nitroso heterocycles required

for the preparation of spirooxazines 1-4 were obtained from

(48) Cross, G. H.; Hackman, N.-A.; Thomas, P. R.; Szablewski, M.;

Pålsson, L.-O.; Bloor, D. Opt. Mater. 2002, 21 (1-3), 29–37.

(

49) Szablewski, M.; Thomas, P. R.; Thornton, A.; Bloor, D.; Cross, G. H.;

Cole, J. M.; Howard, J. A. K.; Malagoli, M.; Meyers, F.; Br e´ das,

J. L.; Wenseleers, W.; Goovaerts, E. J. Am. Chem. Soc. 1997, 119

(

13), 3144–3154.

(

50) Cole, J. M.; Copley, R. C. B.; McIntyre, G. J.; Howard, J. A. K.;

Szablewski, M.; Cross, G. H. Phys. ReV. B 2002, 65 (12), 125107.

(51) Wang, Y. L.; Frattarelli, D. L.; Facchetti, A.; Cariati, E.; Tordin, E.;

Ugo, R.; Zuccaccia, C.; Macchioni, A.; Wegener, S. L.; Stern, C. L.;

Ratner, M. A.; Marks, T. J. J. Phys. Chem. C 2008, 112 (21), 8005–

8

015.

(

52) Duan, X. F.; Dudis, D. S.; Yeates, A. T. Synth. Met. 2001, 116 (1-

3), 285–288.

53) Jarowski, P. D.; Wu, Y.-L.; Boudon, C.; Gisselbrecht, J.-P.; Gross,

M.; Schweizer, W. B.; Diederich, F. Org. Biomol. Chem. 2009, 7

(

7), 1312–1322.

(

46) Patel, D. G.; Benedict, J. B.; Kopelman, R. A.; Frank, N. L. Chem.

Commun. 2005, 2208–2210.

(54) Ogawa, K.; Harada, J.; Tamura, I.; Noda, Y. Chem. Lett. 2000, (5),

528–529.

47) There is an ongoing debate regarding the justiﬁcation and use of the

term “pseudopolymorph”. Nangia, A. Cryst. Growth Des. 2006, 6

(55) Mikami, M.; Nakamura, S. Phys. ReV. B 2004, 69, 134205.

(56) Braunstein, P.; Siri, O.; Taquet, J.-P.; Rohmer, M.-M.; B e´ nard, M.;

Welter, R. J. Am. Chem. Soc. 2003, 125 (40), 12246–12256.

(57) Wiberg, K. B.; Laidig, K. E. J. Am. Chem. Soc. 1987, 109 (20), 5935–

5943.

(

1), 2–4. Bernstein, J. Cryst. Growth Des. 2005, 5 (5), 1661–1662.

Bernstein, J. Polymorphism in Molecular Crystals; Oxford University

Press: Oxford, 2002; 410 pp. While the term “solvate” is an

appropriate term for 2-PMC-II, we use the term “pseudopolymorph”

to compare and contrast the two structures of 2-PMC within the

discussion, and utilize the deﬁnition that “Pseudopolymorphs are solid

crystalline phases of a compound formed by the same substance

crystallized with different amounts or types of solvent molecules.”

Nangia, A. Nomenclature in crystal engineering. In Encyclopedia of

Supramolecular Chemistry; Atwood, J. L., Steed, J. W., Eds.; Marcel

Dekker: New York, 2004; Vol. 2, pp 967-972. We thank a reviewer

for insightful and helpful comments with regard to this issue.

(58) Wiberg, K. B.; Rablen, P. R. J. Am. Chem. Soc. 1995, 117 (8), 2201–

2209.

(59) Wiberg, K. B.; Rush, D. J. J. Am. Chem. Soc. 2001, 123 (9), 2038–

2046.

(60) Wiberg, K. B.; Rush, D. J. J. Org. Chem. 2002, 67 (3), 826–830.

(61) Lokshin, V.; Samat, A.; Guglielmetti, R. Tetrahedron 1997, 53 (28),

9669–9678.

(62) Chamontin, K.; Lokshin, V.; Samat, A.; Guglielmetti, R.; Dubest,

R.; Aubard, J. Dyes Pigm. 1999, 43, 119–125.

1

2570 J. AM. CHEM. SOC. 9 VOL. 132, NO. 36, 2010

Charge Separation in Metastable Photomerocyanines

A R T I C L E S

Scheme 2. Synthesis of o-Hydroxy-Nitroso Isoquinoline and Phenanthroline Derivatives

Scheme 3. Synthesis of Azahomoadamantyl Spirooxazines

63-66

commercially available starting materials (Scheme 2).

The

Scheme 4. Synthesis of Indolyl Spirooxazines

isoquinoline derivative (7) was obtained by the condensation

of 3-benzyloxybenzaldehyde (5) with aminoacetaldehyde di-

methylacetal, followed by treatment with triﬂuoroacetic anhy-

dride and boron triﬂuoride-etherate to afford 7-hydroxyiso-

63

quinoline (6) as a white powder in 60% yield. The nitrosylation

of 6 with sodium nitrite under acidic conditions gave 7-hydroxy-

64

8

-nitrosoisoquinoline (7) as a pale brown powder in 81% yield.

The phenanthroline derivative (10) was obtained from 1,10-

6

5

phenanthroline-5,6-dione (9), which was prepared by the

oxidation of 1,10-phenanthroline (8) with potassium bromide

under acidic conditions to give 9 as yellow needles in 60% yield.

The base-catalyzed condensation of 9 with hydroxylamine

hydrochloride gave 5-hydroxy-6-nitroso-1,10-phenanthroline

6

yladamant-2-ol (11) with sodium azide to give 12 as a pale

(

10) as a yellow powder in 98% yield.

The azahomoadamantyl spirooxazines 1 and 2 were synthe-

yellow oil in 92% yield, followed by treatment with methyl

6

7

iodide to afford 13 as a white powder in 88% yield. The

condensation of o-hydroxy-nitroso derivatives 7 and 10 with

enamine 14 in dichloromethane under dehydrating conditions

gave 1 and 2 as magenta and iridescent green crystalline solids

in 36 and 32% yields, respectively. The indolyl derivatives 3

and 4 were obtained by the condensation of commercially

available 1,3,3-trimethyl-2-methyleneindole (15) with o-hy-

droxy-nitroso heterocycles 7 and 10, respectively (Scheme 4).

The condensation of 15 with 7-hydroxy-8-nitrosoisoquinoline

sized by the condensation of the o-hydroxy-nitroso arenes 7

and 10, respectively, with 4-methyl-5-methylene-4-azahomoada-

mantane (14), generated in situ by the reaction of 4,5-dimethyl-

4

-azahomoadamant-4-enium iodide (13) with triethylamine at

62

low temperature (Scheme 3). The required azahomoadamantyl

iodide 13 was synthesized by the ring expansion of 2-meth-

(

63) Kucznierz, R.; Dickhaut, J.; Leinert, H.; von der Saal, W. Synth.

Commun. 1999, 29 (9), 1617–1625.

64) Stankyavichyus, A. P.; Terent’ev, P. B.; Solov’ev, O. A. Khim.

Geterotsikl. Soedin. 1989, (9), 1243–1247.

(7) in 2-propanol yielded spiro[indoline-isoquinolinoxazine] (3)

(

65) Paw, W.; Eisenberg, R. Inorg. Chem. 1997, 36 (11), 2287–2293.

66) Terent’ev, P. B.; Stankyavichyus, A. P. Chem. Heterocycl. Compd.

(67) Sasaki, T.; Eguchi, S.; Toi, N. J. Org. Chem. 1978, 43 (20), 3810–

1

988, 24 (11), 1258–1262.

3813.

J. AM. CHEM. SOC. 9 VOL. 132, NO. 36, 2010 12571

A R T I C L E S

Patel et al.

a

b

[s^-¹× 10^-³], of 1-4 in

Table 1. Thermal Equilibrium Constants, K

T

, % PMC Values, and Thermal Isomerization Rate Constants, k

1

and k

2

Toluene, CH

3

CN, and MeOH at 300 K

toluene

CH

3

CN

CH

3

OH

K

T

(% PMC)

k

1

k

2

K

T

(% PMC)

k

1

k

2

K

T

(% PMC)

k

1

k

2

c

d

1

2

3

4

<0.01 (<1)

1.2 (55)

<0.01 (<1)

0.02 (2)

-

0.11 (10)

20 (95)

<0.01 (<1)

0.07 (6)

0.033

23_f

0.30

1.1

120

190

0.84 (46)

25 (96)

<0.01 (<1)

0.07 (6)

0.059

8.7 _f

<0.1

6.5

0.071

0.35

15

82

68

e

d

c

g

c

g

-

<1

4.5

230

13

93

a

1

b

c

Determined by H NMR; K

T

) [PMC]/[SO]. Determined by electronic absorption spectroscopy. The concentration of the PMC form was below

1

the limit of detection by H NMR, although the solution was faintly colored, which indicates the presence of some concentration of PMC form with a

high extinction coefﬁcient. Not evaluated. The concentration of the PMC form was below the limit of detection by H NMR, and the solution was

yellow, indicating no detectable PMC form in solution. Upper limit calculated with K

d

e

1

f

g

T

) 0.01. Calculated assuming k

2

1

. k .

as a yellow powder in 35% yield. Spiro[indoline-phenanthroli-

noxazine] (4) was prepared by the condensation of 15 with

5

4

-hydroxy-6-nitroso-1,10-phenanthroline (10) in toluene to give

as yellow crystals in 25% yield.

Solution-State Isomerization Properties. Thermal isomeriza-

19

tion between the SO and PMC forms of spirooxazines estab-

lishes a thermal equilibrium that can be shifted upon steady-

state UV or visible irradiation to a photostationary state (Scheme

1

). The thermal equilibrium constants, K

T

(K

T

) [PMC]/[SO]),

CN, and

3

OH at 300 K from the H NMR spectra of the compounds.

of spirooxazines 1-4 were obtained in toluene, CH

CH

3

1

This was done by calculating the ratio of the peak areas for the

azomethine proton resonance of the PMC (δ ≈ 10 ppm) and

SO (δ ≈ 8 ppm) forms (Table 1). The K

T

values vary

signiﬁcantly, and range from <0.01 (<1% PMC) for the indolyl

isoquinoline derivative 3 to >20 (>95% PMC) for the azaho-

moadamantyl phenanthroline derivative 2 in polar solvents. It

is evident that the stability of the PMC isomer is greater in the

phenanthroline derivatives relative to the isoquinoline derivatives

and in the azahomoadamantyl derivatives relative to the indolyl

derivatives. The equilibrium constants depend strongly on

solvent, and become larger with increasing solvent polarity. This

effect is more pronounced for the azahomoadamantyl derivatives

Figure 1. Electronic absorption spectra of 2 during steady-state visible

irradiation (a) and of 3 during steady-state UV irradiation (b) in CH CN at

00 K.

3

than the indolyl derivatives. The value of K

from <0.01 (<1% PMC) in toluene to 0.84 (46% PMC) in

CH OH, and for 2, it increases from 1.2 (55% PMC) in toluene

to 25 (96% PMC) in CH OH. In contrast, the value of K only

CN and CH OH

T

for 1 increases

with the photoinduced conversion from the PMC form to the

SO form (Figure 1). In contrast, the thermal equilibrium of the

indolyl derivatives 3 and 4 lies far toward the SO form, which

prevents signiﬁcant visible-light-induced photoisomerization.

The kinetics of thermal isomerization from either of the UV-

or visible-light-induced photostationary states were monitored

by electronic absorption spectroscopy, and the observed thermal

rate constants (k_o_b_s) were obtained for the PMC f SO or SO

f PMC thermal isomerizations, respectively. The thermal rate

3

T

increases from 0.02 in toluene to 0.07 in CH

3

for 4.

The photochromic behavior of compounds 1-4 is consistent

with that of the larger class of spirooxazines and can be

monitored by the intense absorption band in the visible region

(

>400 nm) associated solely with the PMC form. In nonpolar

constants for SO f PMC conversion (k ) and PMC f SO

1

solvents such as pentane and toluene, the four spirooxazines

are photoresponsive to UV light and exhibit an increase in the

intensity of the PMC-based π-π* absorption band upon

irradiation, consistent with SO f PMC conversion (Figure 1).

Once a photostationary state is established with steady-state

irradiation, thermal reversion to the original SO/PMC equilib-

rium proceeds in the absence of light, and the process of

irradiation/thermal relaxation can be repeated over several

photocycles. This same behavior is observed for 1, 3, and 4 in

conversion (k ) (Scheme 1) were then calculated from k ) k₁

2

obs

+ k and K ) k /k , where K ) [PMC]/[SO] is the equilibrium

2

T

1

2

T

1

1,68

constant obtained by H NMR spectroscopy.

These rate

constants were determined in a nonpolar solvent (toluene), a

moderately polar solvent (CH CN), and a strongly polar solvent

3

(CH OH) at 300 K, and are summarized in Table 1. The values

3

of k are of the same order for the two phenanthroline derivatives

1

69

2 and 4, but are orders of magnitude lower for the isoquinoline

derivatives 1 and 3. Thus the type of substitution at the oxazine

moiety clearly dominates the rates of ring-opening, where

3 3

more polar solvents (CH CN and CH OH). For the azaho-

moadamantyl phenanthroline derivative 2 however, the thermal

equilibrium lies far toward the PMC isomer in polar solvents,

and UV irradiation leads to an irreversible decrease in the PMC

band, which suggests that the PMC form is susceptible to UV-

induced photodecomposition in these solvents. Visible irradia-

tion of the azahomoadamantyl derivatives 1 and 2 leads to a

decrease in the intensity of the PMC absorption band, consistent

(

68) Metelitsa, A. V.; Micheau, J. C.; Voloshin, N. A.; Voloshina, E. N.;

Minkin, V. I. J. Phys. Chem. A 2001, 105 (37), 8417–8422.

1

(69) Because the k value for 3 cannot be more precisely determined due

to a lack of a more precise K

T

value, it is not possible to compare its

value for 1 and 3. The value reported for 3 merely represents an

upper limit calculated assuming K

that of 1 than it appears.

T

) 0.01, and it could be closer to

1

2572 J. AM. CHEM. SOC. 9 VOL. 132, NO. 36, 2010

Charge Separation in Metastable Photomerocyanines

A R T I C L E S

isoquinoline substitution decreases the k

phenanthroline substitution. The rate constants for PMC f SO

isomerization (k ) in a given solvent are sensitive to both oxazine

substitution and amine functionality. The values of k are 2 to

orders of magnitude lower for the azahomoadamantyl deriva-

1

values relative to

2

3

tives relative to the indolyl derivatives, and in general an order

of magnitude higher for the phenanthroline derivatives relative

to the isoquinoline derivatives. This suggests that the effect of

amine substitution (azahomoadamantyl vs indolyl) on the rates

of ring closing is substantial and more signiﬁcant than the effect

of substitution at the oxazine moiety. The rate constant for SO

1

f PMC isomerization, k , deviates only slightly with solvent

and in general remains of the same order of magnitude. The k

2

values, however, exhibit a more signiﬁcant solvent dependence,

and decrease with increasing solvent polarity for all of the

compounds studied. This is consistent with an increase in the

energy barrier to thermal ring closure in more polar solvents.

Vibrational Structure and Solvatochromism of the Electronic

Absorption Spectra. The vibrational structure and solvato-

chromism of the PMC-based π-π* absorption band of 1, 2,

and 4 were examined in a series of 22 solvents. The features of

the absorption bands were correlated with the Dimroth-Reichardt

7

0

T

E solvent polarity scale, which takes into account solvation

effects arising from both nonspeciﬁc (e.g., dipole-dipole,

induced dipole-induced dipole, etc.) and speciﬁc (e.g., hydrogen

bonding) interactions (see Supporting Information for a list of

N

71

Figure 2. PMC π-π* absorption band shape at 300 K for 2 (a) and 4 (b)

in a representative selection of solvents illustrating shifts in λmax and changes

in band structure with solvent.

solvents used and their corresponding E

T

values) which has

been found to provide reasonable correlations in solvatochromic

7

2,73

studies of the photomerocyanines.

Representative PMC π-π* absorption band shapes are shown

in a selection of solvents for 2 and 4 in Figure 2 (also shown

for 1 in the Supporting Information). All of the derivatives

display a shoulder on the high-energy side of the absorption

band as well as a third less prominent feature at even higher

energies (between 400 and 500 nm). For the indolyl phenan-

throline derivative 4, a bathochromic shift in λmax is observed

scattering arising from the complex solute-solvent interactions

in these photomerocyanines.

The relative peak areas of ν(1), ν(2), and ν(3) for 2 and 4

are compared in Figure 3 as a function of solvent. The data

74

were ﬁt by locally weighted least-squares regression methods.

It is evident that for the azahomoadamantyl derivative 2, the

relative peak area of ν(1) increases as a function of solvent

N

with a concomitant decrease in the shoulder when going from

polarity until E_T≈ 0.5-0.6 before decreasing in higher polarity

N

low- to moderate-E

T

-value solvents. However, in solvents with

solvents, while the relative peak areas for ν(2) and ν(3) exhibit

the opposite behavior [Figure 3(a)]. In contrast, for the indolyl

derivative 4, the relative peak area of ν(1) increases as a function

N

E

T

values greater than ∼3-4, the position and shape of the

absorption band do not change signiﬁcantly [Figure 2(b)]. For

N

the azahomoadamantyl derivatives 1 and 2, the PMC π-π*

of solvent until E_T≈ 0.4-0.5 before leveling off at E

≈

T

N

absorption bands behave similarly to that of 4 in lower E

T

-

0.8, with ν(2) and ν(3) decreasing with higher solvent polarities

[Figure 3(b)].

value solvents, with a bathochromic shift in λmax and decrease

in shoulder intensity with increasing solvent polarity. However,

As regards the λ_m_a_xvalues of ν(1)-ν(3) for 2 and 4, the trends

observed for ν(1) and ν(2) as a function of solvent are essentially

the same, while those for ν(3) differ (Supporting Information).

The λ_m_a_xvalues of the experimental π-π* absorption bands

are nearly identical to the absolute values for ν(1) (Supporting

Information), and are shown graphically as a function of solvent

N

in high E

T

-value solvents, a hypsochromic shift for λmax is

observed along with a broadening and loss of ﬁne structure in

the absorption band [Figure 2(a)].

The PMC absorption bands for 2 and 4 were deconvoluted

into three peaks with good ﬁts using Lorentzian functions

N

(

Supporting Information). The three peaks were tentatively

E_Tvalues for 1, 2, and 4 in Figure 4, with the data again ﬁt by

74

assigned to vibrational structure (vide infra); the lowest energy

peak, labeled ν(1), accounts for 40-60% of the total peak area,

the midenergy peak, labeled ν(2), accounts for 30-55% of the

total peak area, and the highest energy peak, labeled ν(3),

accounts for 1-15% of the total peak area. Trends are evident

in both the relative peak areas and λmax values for the three

peaks as a function of solvent, despite a fairly high degree of

locally weighted least-squares regression methods (λ_m_a_xvalues

for both the experimental and deconvoluted peaks are tabulated

numerically in the Supporting Information).

For the experimental λ_m_a_xvalues, representative of the

behavior of both the ν(1) and ν(2) bands, the azahomoadamantyl

derivatives 1 and 2 exhibit weak solvatochromism at low

N

T

(<0.2-0.3) E values and negative solvatochromism at higher

N

(

>0.3-0.4) E

T

values. In contrast, the indolyl phenanthroline

derivative 4 exhibits more pronounced positive solvatochromism

in solvents with low (<0.3-0.4) E

solvatochromism in solvents with higher E

(

70) Reichardt, C. Chem. ReV. 1994, 94 (8), 2319–2358.

N

71) E

scale.

72) Keum, S.-R.; Hur, M.-S.; Kazmaier, P. M.; Buncel, E. Can. J. Chem.

991, 69 (12), 1940–1947.

73) Jacques, P. J. Phys. Chem. 1986, 90 (22), 5535–5539.

T

is the normalized dimensionless version of Reichardt’s E

T

(30)

T

values and relatively weak

N

T

values. In the low

(

1

(74) Cleveland, W. S. J. Am. Stat. Assoc. 1979, 74 (368), 829–836.

J. AM. CHEM. SOC. 9 VOL. 132, NO. 36, 2010 12573

A R T I C L E S

Patel et al.

Table 2. ¹H and ¹³C NMR Shifts of 2-SO, 2-PMC, 4-SO, and

,

a b

4-PMC in a Selection of Solvents

2-SO

2-PMC

N

T

solvent

E

H-1

H-3

H-1

H-3

C-2

C-3

C-5

toluene-d

benzene-d

8

0.099 2.21 7.53 2.69 10.07

0.111 2.20 7.59 2.61 10.20

123.5

6

THF-d

8

0.207 2.49 7.75 3.5

0.259 2.48 7.71 3.51

0.309 2.48 7.73 3.52 10.03

10.03 175.4 125.8 179.3

9.98 174.6 126.2 179.8

CDCl

CD Cl

3

2

acetone-d

DMSO-d

6

0.355 2.50 7.81 3.58 10.11

0.444 2.40 7.85 3.54 10.14 172.0 127.3 180.3

0.460 2.44 7.75 3.51 10.12

6

CD

3

CN

OD

3

0.762 2.45 7.75 3.62 10.16

4-SO

4-PMC

N

T

solvent

E

H-1

H-3

H-1

H-3

C-2

C-3

C-5

toluene-d

benzene-d

8

0.099 2.40 7.57 2.74 10.03

0.111 2.38 7.60 2.62 10.10

6

THF-d

8

0.207 2.78 7.93 3.66 10.05 174.6

177.8

CDCl

CD Cl

3

0.259 2.76 7.83 3.65 10.01 174.5 122.5 177.4

0.309 2.78 7.87 3.70 10.06

2

acetone-d

DMSO-d

6

0.355 2.80 7.99 3.75 10.10

6

0.444 2.71 8.03 3.72 10.08

0.460 2.74 7.91 3.64 10.06

0.762 2.76 7.95 3.80 10.14

122.3

CD

3

CN

OD

3

a

b

Spectra acquired at 300 K. See Figures 5 and 7 for atom

Figure 3. Relative peak areas of the three Lorentzian deconvolution peaks

numbering.

of the PMC π-π* absorption band, ν(1), ν(2), and ν(3), for 2 (a) and 4 (b)

N

as a function of the Dimroth-Reichardt E

T

solvent polarity scale at 300

negative solvatochromism over the full range of solvent

polarities (Supporting Information).

K (shown as the percent of the total peak area of the sum of the deconvoluted

peaks).

1

13

1

13

H and C NMR Spectroscopy. H and C NMR studies were

performed on 2 and 4 in a range of solvents to obtain additional

spectroscopic evidence for the molecular structures of the PMC

1

form as a function of solvent polarity. Experimental H NMR

spectra of 2 and 4 were obtained in eight solvents ranging in

polarity from toluene-d

8

to CD

OD. ¹³C NMR spectra were

3

obtained in toluene-d , THF-d

8

, CDCl , and DMSO-d . Experi-

3 6

1

mental H NMR resonances were assigned on the basis of

coupling constants and 2D correlation spectroscopy. Experi-

mental ¹³C NMR shifts were assigned on the basis of H- C

1

13

correlation spectroscopy for all CH, CH

2 3

, and CH signals, but

were assigned on the basis of the calculations for the quaternary

1

13

signals. Experimental H and C NMR shifts are given in Table

2

. A signiﬁcant downﬁeld shift from ∼2.7 ppm in toluene-d

8

/

N

benzene-d

THF-d

6

(E

T

≈ 0.1) to ∼3.6 ppm in solvents ranging from

N

4 T

8

to methanol-d

(E

≈ 0.2-0.8) is observed for the

Figure 4. PMC π-π* absorption band λmax values at 300 K for 1 (9), 2

N

N-methyl (H-1) resonance of both 2-PMC and 4-PMC (see

Figures 5 and 7 for atom numbering). As this change is greater

than that expected due to solvent effects alone (0.3-0.4 ppm

for 2-SO and 4-SO), this may indicate a change in molecular

structure upon going from very low to moderate/high polarity

solvents. In contrast, the azomethine (H-3) resonance does not

shift signiﬁcantly with solvent polarity for 2-PMC and 4-PMC.

The experimental ¹³C NMR signals of 2-PMC and 4-PMC

exhibit shifts with solvent polarity that differ slightly between

the two structures. For 2-PMC, a slight upﬁeld shift is observed

for C-2 with increasing solvent polarity (175 ppm in THF vs

(

1), and 4 (b) as a function of the Dimroth-Reichardt E

T

solvent polarity

scale.

polarity region, the greatest change in λmax, ∆λ, for 2 is 16 nm

559 nm in both 1,2-dichlorobenzene and nitrobenzene - 543

(

nm in pentane) vs 34 nm for 4 (599 nm in nitrobenzene - 565

nm in pentane), which evidences the more pronounced positive

solvatochromism for the indolyl derivative. In the high polarity

region, it is not possible to assign a positive or negative

N

solvatochromism for 4 as the slope of the λmax vs E

T

curve is

nearly zero. For 2, however, the ∆λ is signiﬁcant at -32 nm

527 nm in H O - 559 nm in 1,2-dichlorobenzene). To compare

the relative negative solvatochromism of 1 and 2, the ∆λ

evaluated as the difference between λmax in 1,3-propanediol

and 1,2-dichlorobenzene since 1 was not sufﬁciently soluble to

obtain a spectrum in H O) was found to be -0.33 nm for 1 and

0.22 nm for 2, which is consistent with greater negative

solvatochromism in 1. For both 2 and 4, the ν(3) band exhibits

1

72 in DMSO), with subtle downﬁeld shifts observed at C-3.

(

2

No apparent shift in C-5 is observed with changes in solvent

polarity. For 4-PMC, there is essentially no change in C NMR

13

(

shifts for C-2, C-3, and C-5 as a function of solvent polarity,

although experimental values were only obtained in a narrow

range of solvent polarities for this derivative. Chemical shifts

were obtained in the most complete series of solvent polarities

for the C-3 signal of 2-PMC. In this case, the slight shifts

2

-

1

2574 J. AM. CHEM. SOC. 9 VOL. 132, NO. 36, 2010

Charge Separation in Metastable Photomerocyanines

A R T I C L E S

for limiting quinoidal and zwitterionic structures, with slightly

more zwitterionic character predicted. For 4-PMC, however, the

experimental H-3 shift (10.0-10.1 ppm) falls much closer to

that predicted for the model quinoidal form A (9.9-10.1 ppm)

than the zwitterionic form B (11.3-11.5 ppm). As regards the

13

C shifts for 4-PMC, the C-2 and C-5 shifts are similarly closer

to those expected for form B, though the shifts for C-3 are ∼5

ppm more upﬁeld than for 2-PMC, which suggests a more

quinoidal structure for this species. Caution in interpretation of

the absolute theoretical chemical shift values (δtheor) is necessary

as the single point geometries may not be minima on the

potential energy surface surface. In addition, structural effects

such as methyl group free rotation and speciﬁc solvent effects

are not accounted for by the computations. Nevertheless, the

overall direction and trends in NMR chemical shifts suggest a

more zwitterionic structure for 2-PMC relative to 4-PMC and

mirror the structural changes of the PMC with increasing polarity

toward a more zwitterionic structure.

X-ray Diffraction Analysis of the Spirooxazine Forms. The

X-ray structural analysis of 1-SO has been previously reported

4

6

in the context of its crystalline-state photoreactivity. Yellow

prisms of 3-SO and 4-SO suitable for X-ray analysis were

obtained by the slow evaporation of ethyl acetate (see ref 76

for crystal data and reﬁnement parameters). Spirooxazine 3-SO

Figure 5. Molecular structures of spirooxazines 3-SO (top) and 4-SO

(

bottom) with thermal ellipsoids shown at the 50% probability level.

Disorder in 3-SO is omitted for clarity (Supporting Information).

j

crystallized in the triclinic P1 space group with dimers of two

crystallographically independent molecules [deﬁned as 3-SO(a)

and 3-SO(b)] making up the unit cell (Z ) 4). The crystal

structure of 3-SO exhibits disorder at the indolyl moiety resulting

from a mixture of (R) and (S) enantiomers in the crystalline

lattice (Supporting Information). Spirooxazine 4-SO crystallized

observed may be due to a combination of subtle changes in

molecular structure and speciﬁc solvation effects.

Computational NMR shift correlations are known to repro-

7

5

duce experimental shifts with high accuracies. Theoretical

NMR shifts were calculated for model quinoidal (A) and

zwitterionic (B) geometries of 2-PMC and 4-PMC as a function

of solvent in order to predict the chemical shifts of the canonical

forms at their structural limits. Theoretical NMR shifts of the

canonical forms were calculated using the GIAO method at the

DFT/B3LYP level of theory with the 6-31G(d,p) basis set and

1

in the monoclinic P2 /c space group with four molecules in the

unit cell. Packing interactions for both spirooxazines involve

primarily intermolecular π-π edge-to-face interactions between

the orthogonal indoline and heteroaromatics in addition to weak

π-stacking interactions between the heteroaromatics. No sig-

niﬁcant close contacts were noted. Packing views for 3-SO and

3

the IEFPCM solvation model (for toluene, CHCl , and DMSO)

as implemented in Gaussian 03. The agreement of the predicted

chemical shifts with experiment was evaluated for 4-SO and

corrected for systematic error (Supporting Information). Sig-

niﬁcant downﬁeld shifts of ∼0.3 and ∼1.5 ppm are predicted

4

-SO are provided in the Supporting Information.

Geometrical parameters of interest for the closed forms

relevant to the kinetics and thermodynamics of isomerization

include the degree of planarity of the oxazine ring, the Cspiro-O

bond length, and the degree of planarity about the amine

nitrogen. The structures for 3-SO and 4-SO are shown in Figure

1

for the N-methyl (H-1) and azomethine (H-3) H resonances,

respectively, and downﬁeld shifts of 10-15 ppm are predicted

1

3

for the C-2, C-3, and C-5 C resonances upon formation of a

zwitterionic structure from the quinoidal form (Supporting

Information). The changes in chemical shift for forms A and B

as a function of solvent in 2-PMC and 4-PMC suggest downﬁeld

shifts as a function of increasing solvent polarity.

4

6

5

, and selected geometric parameters for 1-SO, 3-SO, 4-SO,

as well as the known spiro[azahomoadamantane-naphthoxazine]

_O_._F_W_:₃₃₃_.₄₂_g_m_o_l-1. Crystal system:

/c. a: 6.6340(2) Å. b: 15.3270(4) Å. c:

(

76) 1-SO. Formula: C21

monoclinic. Space group: P2

6.2050(4) Å. ꢀ: 100.5580(12)°. V: 1619.82(9) Å . Z: 4. Fcalc: 1.367

23 3

H N

Comparison of experimental shifts to theoretical shifts may

be useful in estimating absolute structures of 2-PMC and 4-PMC

1

3

1

-

3

-1

1

mg mm . µ: 0.086 mm . T: 130(2) K. F: 0.71073 Å. R : 0.0365.

2 2 2

1

in solution. In both compounds, the H-1 H NMR resonance is

wR

2

: 0.1038. R

1

) ∑||F

o

| - |F

c

||/∑|F

o

|; wR

2

) [∑w(F

o

c

- F ) /

2

1/2

2

-1

more upﬁeld than that predicted for the quinoidal form (A) in

low polarity solvents (δtheor 2.79 ppm for 2-PMC/toluene,

Supporting Information) and more downﬁeld than that predicted

for the zwitterionic form (B) in high polarity solvents (δtheor

∑w(F )] ; w ) [σ (F ) + (0.0655P) + 0.3377P] where P )

o

2

-1

[

F

o

+ 2F

c

19 3

]/3). 3-SO. Formula: C21H N O. FW: 329.39 g mol .

Crystal system: triclinic. Space group: P1. a: 11.0123(6) Å. b:

j

1

1.1746(6) Å. c: 14.6641(8) Å. R: 72.493(3)°. ꢀ: 89.767(3)°. F:

3

. Z: 4. F : 1.281 mg mm-3. µ: 0.081

83.308(3)°. V: 1708.31(16) Å

calc

-

1

3

.26 ppm for 2-PMC/DMSO, Supporting Information). This

mm . T: 100(2) K. F: 0.71073 Å. R

1

: 0.0507. wR

2

: 0.1273. R

1

)

2

1/2

2

∑

+

||F

o

| - |F

c

||/∑|F

o

|; wR

2

) [∑w(F

o

- F

c

) /∑w(F

o

)] ; w ) [σ (F )

suggests that both 2-PMC and 4-PMC evolve from a quinoidal

to zwitterionic structure with increasing solvent polarity. The

remainder of the theoretical chemical shifts for 2-PMC (H-3,

C-2, C-3, and C-5) are all intermediate between those predicted

2

-1

2

(0.0589P) + 0.0199P] where P ) [F

o

c

+ 2F ]/3). 4-SO.

-1

Formula: C24

H

20

N

4

O. FW: 380.44 g mol . Crystal system: mono-

/c. a: 14.5008(5) Å. b: 8.9511(3) Å. c:

clinic. Space group: P2

4.8771(6) Å. ꢀ: 107.092(2)°. V: 1845.73(12) Å . Z: 4. Fcalc: 1.369

1

3

1

-

3

-1

mg mm . µ: 0.087 mm . T: 100(2) K. F: 0.71073 Å. R

1

: 0.0450.

2

wR

2

: 0.1236. R

1

) ∑||F

o

| - |F

c

||/∑|F

o

|; wR

2

) [∑w(F

o

- F

c

) /

2

1/2

2

-1

2

(

75) Bagno, A.; Rastrelli, F.; Saielli, G. Chem.sEur. J. 2006, 12 (21),

514–5525.

∑w(F

o

)] ; w ) [σ (F

o

) + (0.077P) + 0.3835P] where P ) [F

o

2

5

c

+ 2F ]/3).

J. AM. CHEM. SOC. 9 VOL. 132, NO. 36, 2010 12575

A R T I C L E S

Patel et al.

a

Table 3. Selected Geometric Parameters [Å, deg] for the Closed Forms of Azahomoadamantyl and Indolyl Spirooxazines

b

c

d

e

1-SO

16-SO

3-SO(a)

3-SO(b)

4-SO

17-SO

C(1)-N(1)

1.463(2)

1.431(2)

1.490(2)

1.521(2)

1.279(2)

1.406(2)

1.377(2)

1.368(2)

108.02(11)

105.09(11)

112.15(11)

108.16(12)

0.121

1.454(1)

1.441(6)

1.439(7)

1.4585(16)

1.5094(19)

1.2805(18)

1.4073(18)

1.3783(18)

1.3630(15)

110.24(11)

105.7(3)

111.6(3)

109.3(3)

0.113

1.448(3)

1.454(4)

1.457(7)

1.522(6)

1.2742(19)

1.4106(17)

1.3785(18)

1.3655(16)

109.4(4)

112.1(3)

106.2(4)

109.3(3)

0.135

1.4610(11)

1.4454(11)

1.4643(9)

1.5121(12)

1.2809(11)

1.4058(10)

1.3665(11)

1.3594(9)

110.14(7)

106.26(6)

110.01(6)

110.38(7)

0.094

1.442(4)

1.436(3)

1.454(3)

1.504(4)

1.267(3)

1.416(3)

1.366(3)

1.362(3)

110.6(2)

109.6(2)

106.9(2)

110.4(2)

0.064

N(1)-C(2)

1.4311(8)

1.4767(7)

1.5209(9)

1.2744(9)

1.4070(9)

1.375(1)

C(2)-O(1)

C(2)-C(3)

C(3)-N(2)

N(2)-C(4)

C(4)-C(5)

C(5)-O(1)

1.3600(8)

108.09(5)

104.38(5)

112.65(5)

107.82(5)

O(1)-C(2)-C(3)

O(1)-C(2)-C(6)

N(1)-C(2)-O(1)

N(1)-C(2)-C(3)

f

j

ꢁ

/Å

0.117

g

ꢂ /Å

0.605

0.255

351.0

0.572

0.263

350.3

0.558

0.319

345.7

-0.483

0.285

348.4

0.277

0.298

347.3

h

Ψ /Å

i

ω /deg

349.0

a

b

c

d

See ref 76 for crystal data and reﬁnement parameters. Reference 46. Reference 77. Spirooxazine 3-SO crystallized with two crystallographically

e

f

independent molecules in the asymmetric unit, denoted here as (a) and (b). Reference 78. ꢁ ) rms deviation of the oxazine ring atoms from the plane

g

h

i

of the oxazine ring. ꢂ ) distance of C(2) from the plane of the arene ring systems. Ψ ) distance of N(1) from the C(1)-C(2)-C(9) mean plane.

ω

j

77

)

∑[C(1)-N(1)-C(2), C(1)-N(1)-C(9), C(2)-N(1)-C(9)]. Chamontin et al. cite a value of 0.010 Å for ꢁ, but from the available structural data

we calculated a value of 0.117 Å.

7

78

(

16-SO) and spiro[indoline-naphthoxazine] (17-SO), are

naphthalene derivative 16-SO [C(2)-O(1) ) 1.4767(7) Å]. In

the indolyl series, a slight increase in the Cspiro-O bond length

from 1.454(3) Å for the naphthalene derivative 17-SO to

1.4585(16)/1.457(7) Å for the isoquinoline derivative 3-SO (a/

b) to 1.4643(9) Å for the phenanthroline derivative 4-SO is

observed. Additionally, the Cspiro-O bond lengths for the

naphthalene and isoquinoline azahomoadamantyl derivatives are

0.02-0.03 Å longer than those for the analogous indolyl

derivatives.

summarized in Table 3. In the closed forms of the spiroxazines,

the naphthoxazine moiety is held nearly orthogonal to the

azaheterocyclic moiety. The oxazine ring is puckered, and its

deviation from planarity can be parametrized by the root-mean-

square (rms) deviation of the oxazine ring atoms from the

oxazine ring mean plane (ꢁ). The values of ꢁ range from 0.064

Å for 17-SO to 0.135 Å for 3-SO(b). The oxazine ring puckering

in spirooxazines is such that the ring is folded into an envelope

conformation along the O(1) · · ·C(3) axis, and the oxazine

fragment either bends toward the N-Me group or away from

it. The direction and extent of bending can be described

geometrically by the distance of the C(2) atom from the mean

plane of the arene ring system (ꢂ), where a positive number

indicates bending toward the N-Me group. Of the structures

presented here, all have the oxazine fragment bent toward the

N-Me group (ꢂ ) 0.605 to 0.277 Å), with the exception of

one of the two distinct 3-SO molecules, namely 3-SO(b), in

which the isoquinoline moiety bends away from the N-Me

group (ꢂ ) -0.483 Å). Folding of the pyran ring in both

directions has been reported in the literature for several closed-

form spiropyran derivatives, including within the same crystal

for crystallographically independent molecules. This geometric

One of the parameters that has been used to describe the

geometry of the indoline moiety, and which can be extended to

the azahomoadamantyl moieties, is the planarity of the N(1)R

group, parametrized either by the distance of the N(1) atom

from the R plane (Ψ) or by the total sum of the three

3

R-N(1)-R angles (ω) (Table 3). The values of Ψ range from

0.255 Å for 1 to 0.319 Å for 3-SO(a), and the values of ω range

from 345.7° for 3-SO(a) to 351.0° for 1-SO. The results suggest

a greater degree of planarity about the amine moiety in the

azahomoadamantyl derivatives relative to the indolyl derivatives.

Typically, spirooxazines have N(1)-C(2) bond lengths which

are shorter than those expected for a C-N bond in a ﬁve-

4

3,79

membered heterocycle (1.47-1.48 Å).

This is the case in

all of the spirooxazines discussed here, in which the N(1)-C(2)

bond lengths range from 1.4311(8) Å for 16-SO to 1.454(4) Å

for 3-SO(b).

42

feature may therefore be largely inﬂuenced by packing effects.

Correlations have been found between the long Cspiro-O bond

lengths (1.45-1.50 Å) of spirooxazines [relative to the typical

C-O bond length of an oxygen-containing six-membered

heterocycle (1.41-1.43 Å)] and their degree of photo-

colorability.

X-ray Diffraction Analysis of the Photomerocyanine

Forms. Dark purple single crystals of 1-PMC were obtained

upon continuous UV irradiation (250 < λ < 400 nm, 6 mW) of

a dilute hexane solution of 1 for 1 h (see ref 81 for crystal data

and reﬁnement parameters). The photomerocyanine crystallized

7

9

4

2,68,80

The effect of aromatic ring structure on the

C

spiro-O bond length of the oxazine can be seen across the series

of compounds. The Cspiro-O bond of the azahomoadamantyl

isoquinoline derivative 1-SO [C(2)-O(1) ) 1.490(2) Å] is

greater by 0.013 Å than that of the azahomoadamantyl

in the monoclinic P2 /c space group with four molecules in the

1

unit cell (Figure 6). The molecules pack in head-to-tail dimeric

units (as in H aggregation) with parallel isoquinoline groups

and large intermolecular mean plane separations of 4.62 Å

(Figure 6, bottom). These dimeric units assemble lengthwise

to form a series of perpendicularly arranged chains running along

the c axis. The chains are arranged such that additional close

intermolecular N(1) · · ·O(1) contacts of 3.39 Å occur between

chains, slightly greater than the expected van der Waal’s

atom-atom contact of 3.2 Å.

(

77) Chamontin, K.; Lokshin, V.; Guglielmetti, R.; Samat, A.; P e` pe, G.

Acta Crystallogr., Sect. C: Cryst. Struct. Commun. 1998, 54, 670–

6

72.

78) Millini, R.; Del Piero, G.; Allegrini, P.; Crisci, L.; Malatesta, V. Acta

Crystallogr., Sect. C: Cryst. Struct. Commun. 1991, 47, 2567–2569.

(

79) Allen, F. H.; Kennard, O.; Watson, D. G.; Brammer, L.; Orpen, A. G.;

Taylor, R. J. Chem. Soc., Perkin Trans. 2 1987, (12), S1–S19.

(

80) Millini, R.; Del Piero, G.; Allegrini, P.; Malatesta, V.; Castaldi, G.

Dark purple single crystals of 2-PMC were obtained as two

pseudopolymorphssa crystalline lattice without solvent inclu-

Acta Crystallogr., Sect. C: Cryst. Struct. Commun. 1993, 49, 1205–

207.

1

2576 J. AM. CHEM. SOC. 9 VOL. 132, NO. 36, 2010

Charge Separation in Metastable Photomerocyanines

A R T I C L E S

Figure 6. Molecular structure of 1-PMC with thermal ellipsoids shown at

the 50% probability level (top) and crystal packing viewed along the a

axis (bottom).

Figure 7. Molecular structure of 2-PMC-I with thermal ellipsoids shown

at the 50% probability level (top) and crystal packing viewed along the b

axis (bottom).

sion (2-PMC-I) and a trihydrate of the PMC form (2-PMC-

II)sfrom the slow evaporation of ethyl acetate (see ref 81 for

crystal data and reﬁnement parameters). The unsolvated

pseudopolymorph of azahomoadamantyl phenanthroline pho-

tomerocyanine, 2-PMC-I, crystallized in the monoclinic C2/c

space group with eight molecules in the unit cell (Figure 7).

The molecules pack as slipped head-to-tail dimeric units with

large intermolecular mean plane separations of 4.64 Å (Figure

7

, bottom). These dimeric units assemble into chains running

along the c axis in which each dimeric unit is tilted with respect

to the last. Intermolecular N(1) · · ·O(1) contacts of 4.49 Å are

observed between chains.

The solvate of azahomoadamantyl phenanthroline photomero-

1

cyanine, 2-PMC-II, crystallized in the P2 /n space group with

four molecules in the unit cell (Figure 8). The presence of

solvent molecules in the lattice caused signiﬁcant disorder, and

the structure was solved as a trihydrate with three water

molecules per photomerocyanine molecule (see Supporting

Information). Although the quality of the data for this structure

(

81) 1-PMC. Formula: C21

H

23

N

3

O. FW: 333.42 g mol^-¹. Crystal system:

/c. a: 9.2890(10) Å. b: 15.728(2) Å. c:

monoclinic. Space group: P2

1

3

1

2.7750(9) Å. ꢀ: 111.391(5)°. V: 1737.8(3) Å . Z: 4. Fcalc: 1.274 mg

-

3

-1

mm . µ: 0.080 mm . T: 130(2) K. F: 0.71073 Å. R

1

: 0.0622. wR

2

:

;

Figure 8. Molecular structure of 2-PMC-II with thermal ellipsoids shown

at the 50% probability level (top) and crystal packing viewed along the a

axis (bottom).

2

1/2

0

.1604. R

1

o

) ∑||F |-|F

c

||/∑|F

o

|; wR

2

) [∑w(F

o

- F

c

) /∑w(F

o

)]

2

-1

2

²]/

w ) [σ (F

o

) + (0.0000P) + 0.0000P] where P ) [F

o

+ 2F

c

-

1

3

). 2-PMC-I. Formula: C24

H

24

N

4

O. FW: 384.47 g mol . Crystal

system: monoclinic. Space group: C2/c. a: 18.7530(18) Å. b:

was high [R(int) ) 0.0661 on 4020 reﬂections], the solvent-

3

8

F

R

.7476(8) Å. c: 24.797(3) Å. ꢀ: 108.605(5)°. V: 3855.3(7) Å . Z: 8.

-3 -1

1

induced disorder contributed to high R indices [R ) 0.0906,

calc: 1.325 mg mm . µ: 0.083 mm . T: 110(2) K. F: 0.71073 Å.

2

wR ) 0.2565]. The unit cell contains four molecules that are

1

: 0.0469. wR

2

: 0.1174. R

1

) ∑||F

o

| - |F

) + (0.0735P) + 2.1923P] where

]/3). 2-PMC-II. Formula: C24 O·3H O. FW:

/n (an

/c). a: 6.6536(12) Å. b: 16.165(3) Å. c:

c

||/∑|F

o

|; wR

2

) [∑w(F

o

-

2

1/2

2

-1

F

c

) /∑w(F

o

)] ; w ) [σ (F

o

oriented about six water molecules to give a water cavity formed

2

P ) [F

o

+ 2F

c

H N

24 4

2

-

1

in the b/c plane. Strong hydrogen bonds exist between the

heterocyclic ring oxygen, O(1), and closest H O molecule,

2

O(1S), [H(1S) · · ·O(1) ) 1.880 Å] (Figure 8, bottom) with

additional hydrogen bonding (H · · ·O ) 1.900 Å) between

remaining water molecules. The water molecules form a

hydrogen-bonded chain along the a axis, leading to a water

4

38.52 g mol . Crystal system: monoclinic. Space group: P2

alternate setting of P2

0.809(4) Å. ꢀ: 93.277(2)°. V: 2234.4(7) Å . Z: 4. Fcalc: 1.304 mg

1

3

2

-

3

-1

1 2

mm . µ: 0.090 mm . T: 193 K. F: 0.71073 Å. R : 0.0906. wR :

2 2 4 1/2

2

0.2565. R

1

) ∑||F

o

| - |F

c

||/∑|F

o

|; wR

2

) [∑w(F

o

- F

c

o

) /∑w(F )] ;

2

-1

w ) [σ (F

+

o

) + (0.1002P) + 4.1726P] where P ) [Max(F , 0)

o

2

c

2F ]/3).

J. AM. CHEM. SOC. 9 VOL. 132, NO. 36, 2010 12577

A R T I C L E S

Patel et al.

Table 4. Selected Bond Lengths [Å] and Angles [deg] for 1-PMC

a

and 2-PMC (I and II)

1-PMC

2-PMC-I

2-PMC-II

C(1)-N(1)

1.485(9)

1.337(9)

1.445(9)

1.319(8)

1.374(8)

1.447(11)

1.266(10)

173.55

1.478(5)

1.327(2)

1.427(2)

1.326(2)

1.346(2)

1.442(2)

1.249(2)

179.17

1.473(6)

1.313(5)

1.448(5)

1.315(5)

1.359(5)

1.423(6)

1.245(5)

170.85

N(1)-C(2)

C(2)-C(3)

C(3)-N(2)

Figure 9. Torsional angles R, ꢀ, and γ about the azomethine bridge

illustrated for 1-PMC.

N(2)-C(4)

C(4)-C(5)

C(5)-O(1)

Table 5. Selected Bond Lengths [Å] Calculated at the DFT/B3LYP

Level with the 6-31G(d,p) Basis Set for 1-SO, 2-SO, and 4-SO

N(1)-C(2)-C(3)-N(2) [R]

C(2)-C(3)-N(2)-C(4) [ꢀ]

C(3)-N(2)-C(4)-C(5) [γ]

175.43

-4.99

-174.93

1.45

176.41

-13.29

1-SO

4-SO

a

b

a

calcd

exptl

2-SO

calcd

exptl

See ref 81 for crystal data and reﬁnement parameters.

C(1)-N(1)

N(1)-C(2)

C(2)-O(1)

C(2)-C(3)

C(3)-N(2)

N(2)-C(4)

C(4)-C(5)

C(5)-O(1)

1.462

1.429

1.488

1.527

1.280

1.401

1.389

1.357

1.463(2)

1.431(2)

1.490(2)

1.521(2)

1.279(2)

1.406(2)

1.377(2)

1.368(2)

1.462

1.422

1.495

1.532

1.278

1.403

1.375

1.358

1.453

1.449

1.467

1.514

1.282

1.400

1.373

1.357

1.4610(11)

1.4454(11)

1.4643(9)

1.5121(12)

1.2809(11)

1.4058(10)

1.3665(11)

1.3594(9)

channel as a result of cooperative hydrogen bonding (Supporting

Information).

The experimental bond lengths and angles deﬁning the central

conjugated azomethine bridge are tabulated in Table 4 for

1-PMC and 2-PMC. Prediction of the theoretical limit for a pure

canonical quinoidal form (Scheme 1, A) using average bond

7

9

a

b

lengths leads to a predicted geometry along the azomethine

bridge with bond lengths of N(1)-C(2) ) 1.355(14) Å,

C(2)-C(3) ) 1.360(20) Å, C(3)-N(2) ) 1.376(11) Å,

N(2)-C(4) ) 1.279(8) Å, and C(4)-C(5) ) 1.478(11). A

similar prediction of the structure of the zwitterionic form

Calculated with the 6-31G(d) basis set. Reference 46.

retains predominantly quinoidal character with a length of

.249(2) Å in 2-PMC-I and 1.245(5) Å in 2-PMC-II.

1

The torsional angles R, ꢀ, and γ [R ) N(1)-C(2)-C(3)-N(2),

) C(2)-C(3)-N(2)-C(4), and γ ) C(3)-N(2)-C(4)-C(5);

(

Scheme 1, B) leads to bond lengths of N(1)-C(2) ) 1.316(9)

ꢀ

Å, C(2)-C(3) ) 1.460(15) Å, C(3)-N(2) ) 1.279(8) Å,

N(2)-C(4) ) 1.376(11) Å, and C(4)-C(5) ) 1.364(14) Å. For

Figure 9] relevant to the cisoid or transoid character about the

azomethine bridge are tabulated in Table 4. For each of the

three PMC structures, these angles deviate slightly from those

expected for a fully planar trans-trans-cis (or “TTC”) conﬁg-

uration (180°, 180°, and 0°), which suggests that this isomer is

1

-PMC, the three bonds that make up the azomethine bridge

[

C(2)-C(3), C(3)-N(2), and N(2)-C(4)] exhibit bond-length

alternation with predominantly single-bond [1.445(9) Å], double-

bond [1.319(8) Å], and single-bond [1.374(8) Å] character,

respectively. The direction of bond-length alternation and

absolute bond lengths are close to those expected for the

zwitterionic form B. The C(5)-O(1) bond length [1.266(10)

Å] falls between those expected for a quinoidal CdO bond

the most stable in the crystalline state, consistent with results

observed for similar molecules.

42,44,45

The total deviation from

planarity, as determined by the sum of the deviation from 0° or

1

80° for each of the three angles, differs considerably for the

three structures, where 1-PMC, 2-PMC-I, and 2-PMC-II have

total deviations of 16.01°, 7.35°, and 26.03°, respectively.

7

9

[

1.222(13) Å] and a zwitterionic phenoxyl C-O bond

79

1.362(15) Å], although it is closer to that expected for a CdO

Computational Results. Geometry optimization calculations

for the SO and PMC forms of 1, 2, and 4 were performed at

the DFT/B3LYP level with the 6-31G(d,p) basis set using the

Gaussian 03 suite of software. Selected parameters are sum-

marized in Tables 5 and 6. Additional calculations with the

-31+G(d,p), 6-311G(d,p), and 6-311+G(d,p) basis sets were

performed on 2-PMC (Supporting Information); however, larger

quinoidal bond.

The availability of two pseudopolymorphs for 2-PMC pro-

vides insight into the effects of chemical environment on the

structure of the PMC form. Although the reﬁned structure for

2

-PMC-II has a high degree of disorder and lower quality

reﬁnement parameters (R ) 0.0906, wR ) 0.2565) relative to

-PMC-I (R ) 0.0469, wR ) 0.1174), differences between

6

1

2

1

2

basis sets did not signiﬁcantly affect the optimized geometries

variations in bond lengths were e0.005 Å), which is consistent

several of the bond lengths for the two structures are equal to

or greater than 2σ (Table 4, Supporting Information), which

suggests that they are crystallographically signiﬁcant. Notable

differences are observed along the azomethine bridge. In

(

82,83

with previous computational studies.

of the phenanthroline spirooxazine derivatives, 2-PMC and

-PMC, geometry optimizations were also performed in toluene

ε ) 2.379), DMSO (ε ) 46.7), and water (ε ) 78.39) using

the Onsager model, although the results obtained in water were

nearly identical to those obtained in DMSO, and therefore only

the results obtained for DMSO are tabulated here.

Comparison of the computational geometries for the closed

forms of the azahomoadamantyl derivatives allows reasonable

prediction regarding their molecular structure, as is evident by

the good correlations between calculated and experimental

geometries. Optimized geometries (Table 5) of the isoquinoline

For the PMC forms

4

(

2-PMC-I, the bond lengths along the bridge [C(2)-C(3) )

1.427(2) Å, C(3)-N(2) ) 1.326(2) Å, and N(2)-C(4) )

1.346(2) Å] fall between those expected for a quinoidal form

(A) and a zwitterionic form (B), though they lean closer to those

expected for the zwitterionic form. In contrast, the corresponding

bond lengths in 2-PMC-II [C(2)-C(3) ) 1.448(5) Å, C(3)-N(2)

)

1.315(5) Å, and N(2)-C(4) ) 1.359(5) Å] exhibit the same

bond-length alternation pattern, but with a shifting of bond

lengths toward those expected for the zwitterionic resonance

form by 0.011-0.021 Å, which suggests that 2-PMC-II has more

zwitterionic character. Additionally, a decrease in length of the

adjacent C(4)-C(5) bond by 0.019 Å is observed in 2-PMC-II

relative to 2-PMC-I, as expected with a shift to a more

zwitterionic structure. In both moleules, the C(5)-O(1) bond

(

82) Sheng, Y. H.; Leszczynski, J.; Garcia, A. A.; Rosario, R.; Gust, D.;

Springer, J. J. Phys. Chem. B 2004, 108 (41), 16233–16243.

83) Maurel, F.; Aubard, J.; Rajzmann, M.; Guglielmetti, R.; Samat, A.

J. Chem. Soc., Perkin Trans. 2 2002, (7), 1307–1315.

1

2578 J. AM. CHEM. SOC. 9 VOL. 132, NO. 36, 2010

Charge Separation in Metastable Photomerocyanines

A R T I C L E S

Table 6. Selected Geometric Parameters (Bond Lengths [Å] and Angles [deg]) and Dipole Moments [D] Calculated at the DFT/B3LYP Level

with the 6-31G(d,p) Basis Set for 1-PMC, 2-PMC, and 4-PMC with and without Solvation Using the Onsager Model

a

1-PMC

2-PMC

4-PMC

b

no medium

exptl

no medium

toluene

DMSO

exptl

no medium

toluene

DMSO

ε

0

2.379

46.7

0

2.379

46.7

C(1)-N(1)

N(1)-C(2)

C(2)-C(3)

C(3)-N(2)

N(2)-C(4)

C(4)-C(5)

C(5)-O(1)

R

1.468

1.357

1.405

1.344

1.325

1.488

1.244

-179.99

180

-0.01

5.8789

1.09

0.0005

5.98

1.485(9)

1.337(9)

1.445(9)

1.319(8)

1.374(8)

1.447(11)

1.266(10)

173.55

1.467

1.358

1.403

1.346

1.322

1.480

1.241

-179.20

-179.19

0.18

1.469

1.353

1.409

1.341

1.326

1.474

1.242

-179.16

-179.30

0.14

1.472

1.343

1.422

1.331

1.335

1.464

1.244

-179.04

-179.53

0

1.478(5)

1.327(2)

1.427(2)

1.326(2)

1.346(2)

1.442(2)

1.249(2)

179.17

1.454

1.369

1.385

1.351

1.320

1.484

1.239

-179.97

179.99

-0.01

7.00

1.455

1.363

1.391

1.345

1.324

1.478

1.240

-179.98

179.99

0.01

1.460

1.348

1.409

1.331

1.339

1.462

1.245

-179.98

179.99

-0.00

18.2626

2.01

ꢀ

γ

175.43

-4.99

-174.93

-1.45

dipole (x)

dipole (y)

dipole (z)

total dipole

9.0405

1.68

0.0062

9.20

11.5458

1.98

0.019

11.71

16.1694

2.44

0.041

16.35

10.0077

1.66

0.0009

10.14

1.43

0.0008

7.15

0.0009

18.37

a

b

Optimized using initial coordinates from the X-ray structure geometry. The experimental geometry for 2-PMC-I was arbitrarily chosen.

and phenanthroline derivatives, 1-SO and 2-SO, reveal subtle

differences in geometry. The most signiﬁcant effects predicted

upon heterocyclic ring expansion are an increase in the

C(2)-O(1) bond length by 0.007 Å and a decrease in the

N(1)-C(2) bond length by 0.007 Å, consistent with the observed

increase in photocolorability for 2-SO. A predicted decrease in

the C(4)-C(5) bond length by 0.014 Å is consistent with greater

double-bond character of the 5,6-positions of the phenanthroline

ring. More signiﬁcant differences in bond lengths are predicted

for the oxazine and amine moiety of the azahomoadamantyl

and indolyl phenanthroline derivatives 2-SO and 4-SO, as

expected. Substitution of azahomoadamantyl for indolyl leads

to a predicted increase in the N(1)-C(2) bond length by 0.027

Å and decrease in the C(2)-O(1) bond length by 0.028 Å,

consistent with the lower photocolorability of 4-SO.

to the experimental structures, consistent with the tendency of

DFT to favor delocalized geometries.

Optimization of the geometries in solvents with varying

dielectric constants using the Onsager model has a signiﬁcant

effect on the resulting structure for 2-PMC and 4-PMC. In both

derivatives, the N(1)-C(2), C(2)-C(3), C(3)-N(2), N(2)-C(4),

and C(4)-C(5) bond lengths deviate by 0.002-0.006 Å in

toluene or 0.010-0.020 Å in DMSO relative to those predicted

in the gas phase, in each case becoming closer to those expected

for the zwitterionic resonance form A. The calculated C(5)-O(1)

bond length, however, was found to be only 0.001 and 0.003-

0.006 Å longer in toluene and DMSO, respectively.

The calculations predict large dipole moments along the long

axis of the molecule for all of the PMC forms. The total dipole

moments were found to be 5.98, 9.20, and 7.15 D for 1-PMC,

2-PMC, and 4-PMC, respectively, with the major contribution

being along the x axis (long axis). For both 2-PMC and 4-PMC,

the dipole moment increases signiﬁcantly for calculations

performed in higher dielectric media: for 2-PMC, the total dipole

moment increased to 11.71 and 16.35 D in toluene and DMSO,

respectively, and for 4-PMC, it similarly increased to 10.14 and

18.37 D in toluene and DMSO, respectively. This suggests an

increase in charge-separated state with increasing polarity of

the media.

DFT predicts very similar structures for 1-PMC and 2-PMC

with regard to the bond lengths and angles in the azomethine

bridge (Table 6). The optimized bond lengths in the bridge

[

C(2)-C(3), C(3)-N(2), and N(2)-C(4)] for both molecules

lie nearly exactly between those expected for resonance forms

A and B, which would indicate approximately equal contribu-

tions from the two forms. This result is consistent with previous

83

calculations on spirooxazines. The optimized bond lengths for

the C(1)-N(1), N(1)-C(2), and C(4)-C(5) bonds are within

0

.005 Å of those expected for the quinoidal resonance form A.

The predicted C(5)-O(1) bond lengths of 1.244 Å in 1-PMC

Discussion

and 1.241 Å in 2-PMC are only slightly longer than a quinoidal

7

9

Vibrational Structure and Solvatochromism of the Electronic

Absorption Spectra of the Indolyl and Azahomoadamantyl

Photomerocyanines. The molecular structure of merocyanines

can be probed indirectly by electronic absorption spectroscopy

as a function of solvent polarity by analyzing either (a) the

relative geometries of the ground and excited states inferred

from the relative intensities of the vibrational bands of the π-π*

CdO double bond (1.221 Å) and signiﬁcantly shorter than a

7

9

C-O single bond (1.362 Å). For the indoline phenanthroline

derivative, 4-PMC, the C(1)-N(1) and N(1)-C(2) bond lengths

are predicted to be shorter and longer, respectively, than those

expected for resonance form A. Of the remaining bond lengths,

all are within 0.005 Å of those obtained for 2-PMC except for

the C(2)-C(3) bond length which is 0.018 Å shorter. This result

indicates that the DFT-optimized geometries are quite similar

for both the indolyl and azahomoadamantyl derivatives as

regards the structure of the central bridge. However, the

deviations observed, particularly for the C(2)-C(3) bond,

suggest that the structure for 4-PMC has a slightly greater

contribution from the quinoidal resonance form A. For all of

the molecules, the optimized dihedral angles R, ꢀ, and γ are

very close to 180°, 180°, and 0°, respectively, with the total

deviation from planarity for the three angles not exceeding 2°.

DFT calculations thus predict a highly planar structure relative

84,85

electronic transition

or (b) the solvent-induced stabilization

of the ground and excited states inferred from the shift in energy

86

of the absorption band, i.e., its solvatochromism. The inferred

structural changes may be interpreted in the context of the

relative contributions of the canonical quinoidal (A) and

(84) Hiramatsu, T.; Yoshida, H.; Sato, N. J. Phys. Chem. A 2009, 113

(

32), 9174–9179.

(

85) Mustroph, H.; Reiner, K.; Mistol, J.; Ernst, S.; Keil, D.; Hennig, L.

ChemPhysChem 2009, 10 (5), 835–840.

(86) Buncel, E.; Rajagopal, S. Acc. Chem. Res. 1990, 23 (7), 226–231.

J. AM. CHEM. SOC. 9 VOL. 132, NO. 36, 2010 12579

A R T I C L E S

Patel et al.

zwitterionic (B) resonance forms to the ground-state molecular

structure (Scheme 1).

quinoidal character in the gas phase or nonpolar solvents to

greater charge-separated zwitterionic character in polar sol-

8

4

9

1,93–97

Vibronic progressions are commonly observed in the π-π*

charge transfer bands of merocyanines, where, in addition to a

ν ) 0 f ν′ ) 0 transition, higher energy (ν ) 0 f ν′ ) 1, ν

vents.

Recent electro-optical absorption studies have

conﬁrmed that such structural changes do indeed result from

8

polarization effects in polar solvents. Several authors have

proposed alternative explanations for the experimentally ob-

served solvatochromic behaviors of merocyanines including

interactions between permanent solute dipoles and induced

)

0 f ν′ ) 2, etc.) transitions are evident as coarse band

85,87,88

structure.

The major vibrational mode associated with the

electronic transition in typical merocyanines is a CdC sym-

metric stretch, in which the CdC bonds are expected to lengthen

8

4

solvent dipoles, as well as solvent-induced changes in mo-

8

5

98

87

in the excited state. The spirooxazine-based photomerocya-

nines display an analogous vibronic progression in their

lecular aggregation, vibrational structure, or distribution of

9

cis/trans isomers in solution. It has been suggested that

analysis of the vibronic intensity proﬁle is a more reliable

method than analysis of the solvatochromism of the absorption

8

9,90

electronic absorption spectra,

presumably associated with

similar symmetric modes involving the conjugated bridge. For

a fully delocalized conjugated structure resulting from equal

contributions of resonance forms A and B, small changes in

geometry would be expected between ground and excited states,

and the Franck-Condon overlap model dictates that the ν ) 0

f ν′ ) 0 transition should dominate the charge transfer band.

For predominantly quinoidal (A) or zwitterionic (B) structures

with localized bond alternation throughout the conjugated bridge,

more signiﬁcant differences in geometry are expected between

the ground and excited states, and transitions to higher ν′

vibrational states should intensify relative to the ν ) 0 f ν′ )

84

band toward predicting molecular structure. Indeed, interpreta-

tion of solvatochromic effects may be case-dependent, and one

must proceed with caution, particularly when interpreting

solvatochromic trends in electronic absorption spectra exhibiting

coarse structural features where effects associated with these

features can be convoluted with solvatochromic effects.

In order to thoroughly interpret the solvatochromism of the

spirooxazines under investigation, we examined both the

vibrational structure and λmax values of their PMC π-π*

absorption bands in a large selection of solvents. We observed

different vibrational intensity proﬁles and solvatochromic

behavior for the azahomoadamantyl and indolyl PMCs, but in

0

transition.

As regards the solvatochromism of photomerocyanines, three

both types of derivatives a qualitative reversal in behavior was

scenarios can be identiﬁed based on the solvatochromic behavior

N

9

1

observed in solvents with moderate-polarity values of E

.3-0.6.

The Lorentzian deconvoluted ν(1), ν(2), and ν(3) bands of

the PMC π-π* band can be assigned to ν ) 0 f ν′ ) 0, ν )

T

≈

of typical merocyanine dyes: (1) for a quinoidal (A) ground-

state structure, a more dipolar excited state should exhibit greater

stabilization with increasing solvent polarity, and a bathochromic

shift should result (i.e., positive solvatochromism); (2) for a

zwitterionic (B) ground-state structure, a more dipolar ground

state should exhibit greater stabilization with increasing solvent

polarity, and a hypsochromic shift should be observed (i.e.,

negative solvatochromism); and (3) for a hybrid structure

possessing equal contributions from the resonance forms A and

B, the moderately dipolar ground and excited states should be

equally stabilized with increasing solvent polarity, and no

0

f ν′ ) 1, and ν ) 0 f ν′ ) 2 vibrational transitions,

respectively, in accordance with the spectroscopic features of

typical merocyanines. For the azahomoadamantyl derivative,

2

-PMC, an increase in intensity of ν(1) with a corresponding

decrease in intensity of ν(2) and ν(3) is observed in lower

polarity solvents, followed by a decrease in intensity of ν(1)

with a corresponding increase in intensity of ν(2) and ν(3) in

8

6

higher polarity solvents [Figure 3(a)], where the change in

signiﬁcant solvatochromic shift should result.

N

behavior is observed at E

T

≈ 0.5-0.6. This is consistent with

The solvatochromic behavior of merocyanines has been

9

2

the ground- and excited-state molecular structures being most

heavily investigated and remains controversial. Key points of

contention have been the relationship of solvatochromic behavior

to molecular structure, as well as the origin of the experimentally

observed phenomenon of “inverted solvatochromism”, where

a switch in overall solvatochromic behavior is observed with a

similar in moderate polarity solvents when the ν ) 0 f ν′ )

0

transition dominates; a more quinoidal structure is expected

in lower polarity solvents, and a more zwitterionic structure is

expected in higher polarity solvents. In contrast, for the indolyl

derivative 4-PMC, an increase in intensity of ν(1) with a

corresponding decrease in intensity of ν(2) and ν(3) is likewise

observed in lower polarity solvents, but this is followed by a

7

3,86,93

change in solvent polarity.

Theoretical studies have

correlated experimentally observed solvatochromic effects with

molecular structure; calculations at various levels of theory

predict signiﬁcant changes in dipole moment and molecular

structure upon increasing the dielectric constant of the sur-

rounding medium, which indicate a clear transition from greater

leveling off of the intensities of the three bands, with a possible

N

reversal in behavior observed at E

T

≈ 0.8 [Figure 3(b)]. This

behavior is consistent with the existence of a quinoidal structure

in low polarity solvents, and a hybrid A/B structure in moderate-

to high-polarity solvents.

(

87) Catal a´ n, J.; Mena, E.; Meutermans, W.; Elguero, J. J. Phys. Chem.

As regards solvatochromic behavior, the azahomoadamantyl

derivatives, 1-PMC and 2-PMC, exhibit a change from weakly

1

992, 96 (9), 3615–3621.

(88) W u¨ rthner, F.; Archetti, G.; Schmidt, R.; Kuball, H.-G. Angew. Chem.,

Int. Ed. 2008, 47 (24), 4529–4532.

(

89) Pottier, E.; Dubest, R.; Guglielmetti, R.; Tardieu, P.; Kellmann, A.;

(94) Benson, H. G.; Murrell, J. N. J. Chem. Soc., Faraday Trans. 2 1972,

68 (590), 137–143.

Tﬁbel, F.; Levoir, P.; Aubard, J. HelV. Chim. Acta 1990, 73 (2), 303–

3

15.

(95) Albert, I. D. L.; Marks, T. J.; Ratner, M. A. J. Phys. Chem. 1996,

100 (23), 9714–9725.

(

90) Ernsting, N. P.; Arthen-Engeland, T. J. Phys. Chem. 1991, 95 (14),

5

502–5509.

(96) Morley, J. O.; Morley, R. M.; Docherty, R.; Charlton, M. H. J. Am.

Chem. Soc. 1997, 119 (42), 10192–10202.

(

91) Botrel, A.; Lebeuze, A.; Jacques, P.; Strub, H. J. Chem. Soc., Faraday

Trans. 2 1984, 80, 1235–1252.

(97) Baraldi, I.; Momicchioli, F.; Ponterini, G.; Vanossi, D. Chem. Phys.

1998, 238 (3), 353–364.

(

92) Mishra, A.; Behera, R. K.; Behera, P. K.; Mishra, B. K.; Behera,

G. B. Chem. ReV. 2000, 100 (6), 1973–2011.

(98) Niedbalska, M.; Gruda, I. Can. J. Chem. 1990, 68 (5), 691–695.

(99) Tsukada, M.; Mineo, Y.; Itoh, K. J. Phys. Chem. 1989, 93 (24), 7989–

7992.

93) da Silva, L.; Machado, C.; Rezende, M. C. J. Chem. Soc., Perkin

Trans. 2 1995, (3), 483–488.

1

2580 J. AM. CHEM. SOC. 9 VOL. 132, NO. 36, 2010

Charge Separation in Metastable Photomerocyanines

A R T I C L E S

positive to strongly negative solvatochromism with increasing

solvent polarity (Figure 4), which suggests that their ground-

state structure progresses from quinoidal to zwitterionic in

character. The negative solvatochromism of 1-PMC is more

pronounced than that of 2-PMC, which suggests that 1-PMC

has more zwitterionic character. In contrast, the indolyl phenan-

throline derivative, 4-PMC, exhibits a change from relatively

strong positive solvatochromism in low-polarity solvents to little

or no solvatochromism in high-polarity solvents (Figure 4). This

behavior suggests that the ground-state structure of 4-PMC

progresses from predominantly quinoidal in character to a hybrid

A/B form with approximately equal contributions from the

quinoidal and zwitterionic resonance structures with increasing

E

T

^N-valuesolvents,ashasbeenfoundforsimilarsystems.⁷⁰^,⁷³^,⁹⁶^,¹⁰⁰

Because there is no obvious steric reason for which the indolyl

derivatives would not exhibit the same H-bonding effects as

the azahomoadamantyl derivatives, this may suggest that

solvation through H-bonding serves to stabilize the charge-

separated character of molecules which have an intrinsically

more stable zwitterionic structure rather than inducing a more

zwitterionic structure through H-bonding interactions alone.

In summary, the degree of quinoidal or zwitterionic character

of the ground-state PMC species in solution was found to

dominate the observed solvatochromic behavior for the inves-

tigated spiroooxazines. The azahomoadamantyl derivatives

clearly exhibit a reversal of the vibrational band intensity proﬁle

and a transition to negative solvatochromism in high polarity

solvents that are absent in the analogous indolyl derivative. This

suggests that the two families of spirooxazines possess funda-

mentally different ground-state molecular structures in which

the degree of quinoidal or zwitterionic character in a given

medium differs with increasing polarity of the solvent. The

bicyclic derivatives progress from slightly quinoidal to zwitte-

rionic in character, while the indolyl derivatives progress from

quinoidal in character to a hybrid A/B form with increasing

polarity. The overall possible range of quinoidal-to-zwitterionic

character accessible by changing the solvent polarity is unique

in the two classes, in which the bicyclic derivatives access a

greater range of structures.

solvent polarity. The positive solvatochromism observed for

N

4

-PMC in solvents with low to moderate E

T

values is consistent

3

1

with the observations previously reported by Pozzo et al,

although the authors did not highlight the absence of a

N

solvatochromic shift at higher E

T

values. For both 2-PMC and

4

-PMC, the solvatochromism of the experimental PMC π-π*

absorption band mirrors that of both the deconvoluted ν(1) and

ν(2) bands; however, the deconvoluted ν(3) band exhibits

negative solvatochromism over the full range of solvent

polarities. This observation can be rationalized by the possibility

of different solvatochromic responses for different vibrational

bands, as has been experimentally demonstrated by W u¨ rthner

et al., who have associated the effect with a change in molecular

Analysis of Solution-State Structures by H and ¹³C NMR.

1

8

1

13

dipole of a given vibrational level in the excited state.

H and C NMR studies were performed on 2 and 4 to obtain

However, if this interpretation is correct, it begs the question

of how the behavior of each vibrational band ultimately

corresponds to the molecular structure of the merocyanine.

additional spectroscopic evidence for the molecular structures

of the PMC forms with regard to both (a) the relative changes

in PMC structure as a function of solvent polarity and (b) the

absolute structures of 2-PMC and 4-PMC, whether they be more

quinoidal (A) or zwitterionic (B) in character. To address (a),

spectra of the spirooxazines were obtained in a range of solvents

It is of interest that the qualitative reversal in behavior of the

vibrational band intensity proﬁle and solvatochromism as a

function of solvent occurs at different solvent polarities, where

the former is observed at slightly higher polarities. This

observation is consistent with studies by Hiramatsu et al., who

observe a similar effect and posit that the reversal of the

vibrational band intensity proﬁle behavior may be a more

accurate indicator of the polarity at which the limiting structure

possesses equal contributions of quinoidal and zwitterionic

(

4

Supporting Information). Solvent-dependent NMR studies on

have been previously reported by Pozzo et al., although in

3

1

this study the authors used a more limited series of solvents.

Limited success has been achieved in unambiguously correlating

1

13

H or C NMR shifts to solvent-induced structural changes of

31,91,96

merocyanines,

although good correlations between chemi-

cal shifts or JHH coupling constants for merocyanines with

8

4

resonance character. However, in their case, negative solva-

tochromism was not observed even in high polarity solvents as

it was for the molecules described here, and our data may

highlight the fact that molecular structure may still be inferred

from solvatochromic behavior, with the knowledge that the

limiting A/B form exists at slightly higher polarities than

predicted by the solvatochromic reversal. This is supported by

theoretical work by Botrel et al. who show that the limiting

A/B form exists at slightly higher polarities than indicated by

85,101

extended polymethine chains have been reported.

To

address (b), experimental NMR shifts were compared to

theoretical shifts of model A and B forms.

As regards solvent polarity effects, a signiﬁcant downﬁeld

shift in the N-methyl H-1 resonance of ∼1 ppm is observed

N

when going from aromatic hydrocarbon solvents (E

T

≈ 0.1)

N

to higher polarity solvents (E

T

≈ 0.2-0.8) for both 2-PMC

and 4-PMC (Table 2). This may indicate a change in molecular

structure upon going from very low to moderate/high polarity

solvents, or could be explained by speciﬁc solvation effects (or

lack thereof) in aromatic hydrocarbons. As regards absolute

structures, comparison of the experimental and calculated H-3

shifts for 2-PMC suggests that its structure can be approximated

by nearly equal contributions of resonance forms A and B.

Comparison of the experimental and calculated H-3 shifts for

9

1

the solvatochromic data.

It is unlikely that aggregation effects play a major role in the

observed spectral shifts of 1-PMC, 2-PMC, and 4-PMC as their

π-π* transitions obey the Beer-Lambert law in both nonpolar

and polar solvents, and no change in the appearance of their

absorption bands is observed over a wide concentration range

4

-PMC, in contrast, suggests that its structure is closer to that

(

Supporting Information). The occurrence of solute-solvent

of the quinoidal form A. Comparison of the experimental and

calculated shifts for C-3 and C-5 predicts that the structures for

hydrogen bonding interactions in the azahomoadamantyl deriva-

tives 1-PMC and 2-PMC in strong hydrogen-bond donating

solvents is possible. We did not attempt to quantify the

contributions of speciﬁc and nonspeciﬁc interactions to PMC

solvation, though we expect that both polarization and H-

bonding effects play comparable solvating roles in the high-

(

100) Abdel-Halim, S. T. J. Chem. Soc., Faraday Trans. 1993, 89 (1), 55–

5

7.

101) Kulinich, A. V.; Derevyanko, N. A.; Ishchenko, A. A. J. Photochem.

Photobiol., A 2007, 188 (2-3), 207–217.

J. AM. CHEM. SOC. 9 VOL. 132, NO. 36, 2010 12581

A R T I C L E S

Patel et al.

2

-PMC and 4-PMC fall between those of resonance forms A

roborated by the relatively short N-Cspiro bond lengths in 1-SO

[1.431(2) Å] and 16-SO [1.4311(8) Å] relative to those in 3-SO

[1.439(7) Å for (a) and 1.454(4) Å for (b)] and 17-SO [1.436(3)

and B, with more zwitterionic character overall and more

zwitterionic character for 2-PMC relative to 4-PMC. The C-2

shifts predict signiﬁcant zwitterionic character for both mol-

ecules. In summary, the NMR results do not emphatically

support the change in molecular structure with solvent polarity

suggested by the solvatochromism studies; however, the results

do provide some evidence that the absolute structure of 2-PMC

is more zwitterionic than that of 4-PMC, as inferred from the

solvatochromism results, and that subtle shifts from quinoidal

to more zwitterionic occur in the PMC forms in solution. The

discrepancies between the solvatochromism and NMR shift

results may be attributed to strong speciﬁc solvation effects that

are expected to contribute signiﬁcantly to NMR chemical shifts

and to changes in molecular geometry. In general, deconvolution

of the two effects is difﬁcult for structures in which several

geometries with similar energies can exist in solution.

Structures of the Closed Forms. The most important structural

parameter in spirooxazines is undoubtedly their long Cspiro-O

bond length, a feature which is intimately related to their

photochromic properties and typically associated with ease of

bond cleavage. The long Cspiro-O bond length in the azaho-

moadamantyl isoquinoline derivative 1-SO [1.490(2) Å] is

consistent with that observed in a previously characterized

3

Å] [where an average C(sp )-N(sp ) bond length is 1.47-1.48

79

Å]. The long Cspiro-O bond lengths and correspondingly short

N-Cspiro bond lengths in the azahomoadamantyl derivatives

relative to the indolyl derivatives are consistent with the greater

4

6

photocolorability observed for this family of spirooxazines.

The availability of the N lone pair has been correlated with

the degree of planarity of the NR group (as parametrized here

by Ψ and ω) for various indolyl derivatives.

azahomoadamantyl derivatives, however, the NR

3

1

07,108

In the

3

group is

actually closer to planarity than in any of the indolyl derivatives

Table 3). While there is little structural characterization for

(

3

,8

the 4-azahomoadamantane skeleton (4-azatricyclo[4.3.1.1 ]-

undecane) beyond that published for the 4-azahomoadamantane-

naphthalene spirooxazine discussed in this work, it is clear

from numerous reactivity studies of the 4-azahomoadamantane

7

3

,8

skeleton that 4-azatricyclo[4.3.1.1 ]undec-4-enes are in fact

109

2

very stable. This suggests that formation of an sp -hybridized

nitrogen at the 4-position is thermodynamically favorable,

distinct from formation of an anti-Bredt bridgehead imine, which

in most bicyclics and tricyclics has not been isolated due to

instability associated with Baeyer strain. In addition, crystal

structure analysis of the corresponding lactone reveals a highly

planar nitrogen, consistent with the observed planarity in the

present work.

3

4

1

02

bicyclic derivative, and is on the order of the Cspiro-O bond

4

2

103

lengths found in nitro- and pyridinium-substituted indolyl

spiropyrans (1.48-1.50 Å). The Cspiro-O bond lengths in the

indolyl isoquinoline and phenanthroline derivatives 3-SO and

Correlations between the degree of oxazine ring puckering

and photochromic parameters have been proposed. Explicitly,

4

-SO (∼1.46 Å) are crystallographically equivalent to those

found in previously characterized indolyl spirooxazine deriva-

tives containing heterocyclic aromatics.

a higher degree of ring puckering has been correlated with higher

1

04-106

102,105

photocolorabilities.

However, this has not always been

7

The unusually long Cspiro-O bond length in spirooxazines

can be explained by speciﬁc orbital interactions facilitated by

the hybridization and geometry about the spiro carbon. The

N-Cspiro and Cspiro-O bonds are in the appropriate relative

orientations to interact via the anomeric effect, where

n(N)-σ*(Cspiro-O) donation is expected to lead to a decrease

in the N-Cspiro bond length and a corresponding increase in

the Cspiro-O bond length. An n(O)-π*(N-Cspiro) interaction is

also expected to be in effectsalbeit to a lesser extentsand would

have the opposite effect on the N-Cspiro and Cspiro-O bond

found to be the case. No strong correlation between ring

puckering and photochromic parameters is apparent for the

derivatives investigated here. It has alternatively been suggested

that the conformation of the oxazine ring is determined by

molecular sterics and crystal packing rather than electronic

effects, which would make correlation of this crystalline state

structural feature with solution-state photochromic parameters

challenging. The data presented here may corroborate the latter

conclusion when one considers the relatively large difference

in deviation from planarity of the oxazine ring (ꢁ) found

between the two crystallographically independent molecules in

3-SO [0.113 Å for (a) and 0.135 Å for (b)].

42

4

2,107

lengths.

The increase in the Cspiro-O bond length by

0

.02-0.03 Å for the azahomoadamantyl isoquinoline (1-SO)

and naphthalene (16-SO) derivatives relative to the respective

indolyl derivatives (3-SO and 17-SO) can be rationalized by a

greater n(N)-σ*(Cspiro-O) interaction in the azahomoadamantyl

compounds. The nitrogen lone pair is expected to be more

available in the saturated bicyclic derivatives than in the

conjugated indolyl derivatives. A stronger n(N)-σ*(Cspiro-O)

interaction in the azahomoadamantyl derivatives is also cor-

Structures of the Photomerocyanine Forms and Effect of

the Crystalline Environment. The molecular structures for

-PMC and 2-PMC-II exhibit signiﬁcant contributions from the

1

zwitterionic resonance form B, as evidenced by a comparison

of the experimental bond lengths with those expected for the

canonical resonance forms A and B (Scheme 1). The nature of

the zwitterionic bond-length alternation observed is consistent

4

2

with that found in nitro-substituted indolyl spiropyrans. The

short C(5)-O(1) bond length and the relatively long C(4)-C(5)

bond length observed in the experimentally characterized

photomerocyanines (1-PMC, 2-PMC-I, and 2-PMC-II) suggest

that these molecules possess contributions from the zwitterionic

resonance form C in addition to B, in which a portion of the

negative charge on the oxygen atom is delocalized onto the

conjugated bridge and the C-O bond takes on greater quinoidal

character (Scheme 5). This delocalization pattern is consistent

(

102) P e` pe, G.; Lar e´ ginie, P.; Samat, A.; Zaballos, E. Acta Crystallogr.,

Sect. C: Cryst. Struct. Commun. 1995, 51, 1617–1619.

103) Aldoshin, S. M.; Nikonova, L. A.; Smirnov, V. A.; Shilov, G. V.;

Nagaeva, N. K. Russ. Chem. Bull. 2005, 54 (9), 2113–2118.

104) Osano, Y. T.; Mitsuhashi, K.; Maeda, S.; Matsuzaki, T. Acta

Crystallogr., Sect. C: Cryst. Struct. Commun. 1991, 47, 2137–2141.

105) Reboul, J.-P.; Samat, A.; Lar e´ ginie, P.; Lokshin, V.; Guglielmetti,

R.; P e` pe, G. Acta Crystallogr., Sect. C: Cryst. Struct. Commun. 1995,

5

1, 1614–1617.

(

106) Clegg, W.; Norman, N. C.; Flood, T.; Sallans, L.; Kwak, W. S.;

Kwiatkowski, P. L.; Lasch, J. G. Acta Crystallogr., Sect. C: Cryst.

Struct. Commun. 1991, 47, 817–824.

107) Aldoshin, S. M.; Chuev, I. I.; Filipenko, O. S.; Utenyshev, A. N.;

Lokshin, V.; Lar e´ genie, P.; Samat, A.; Guglielmetti, R. Russ. Chem.

Bull. 1998, 47 (6), 1089–1097.

(108) Anisimov, V. M.; Aldoshin, S. M. THEOCHEM 1997, 419, 77–84.

(109) Averina, N. V.; Borisova, G. S.; Zeﬁrov, N. S. Russ. J. Org. Chem.

2001, 37 (7), 901–934.

1

2582 J. AM. CHEM. SOC. 9 VOL. 132, NO. 36, 2010

Charge Separation in Metastable Photomerocyanines

A R T I C L E S

Scheme 5. Resonance Forms of 1-PMC

electrostatic neutrality. By contrast, in the crystalline lattice of

2-PMC-II, the molecules are more disordered and more loosely

packed, which suggests that “self-solvation” may be less

important. The presence of polar water molecules contributes

to an increase in the effective dielectric of the lattice as well as

the clear presence of hydrogen bonding of water molecules to

the “phenoxyl” oxygen. Both contributions would be likely to

stabilize the zwitterionic resonance contribution, leading to

greater stability of the open form and greater zwitterionic

character. This result is signiﬁcant, as it provides the ﬁrst direct

structural evidence for the shift in charge-separated character

associated with environment in the crystalline state.

Comparison of the DFT-Optimized Geometries with

Experiment. The DFT/B3LYP level of theory has been found

to reproduce the experimental geometries of many organic

molecules with reasonable accuracies while using a practical

1

10,111

level of computational resources,

and this treatment has

been used by several authors in the study of spiropyran and

8

2,83,112

spirooxazine systems.

A comparison of bond lengths

predicted by computation to those determined experimentally

suggests that DFT reasonably predicts the geometries of the

closed forms. DFT geometry optimization calculations for 4-SO

produces bond lengths with an average deviation of 0.006 Å

and a maximum deviation of 0.02 Å from those determined

experimentally (Table 5, Supporting Information). In particular,

the experimental Cspiro-O bond lengths are very close to the

calculated lengths, which has not always been the case in studies

with the behavior observed in many of the structurally charac-

38,42

83,112

terized zwitterionic photomerocyanines

predictions of the electrostatic charge distributions for related

classes of photochromic molecules.

as well as theoretical

of other spirooxazines.

39

For the open forms 1-PMC and 2-PMC, the calculated bond

lengths differ more signiﬁcantly from experiment. Bond lengths

in the conjugated azomethine bridge deviate by 0.03-0.05 Å

from experiment (Table 6), whereas those in the azahomoada-

mantyl and phenanthroline groups deviate by 0.01-0.02 Å. The

differences between the calculated and experimental bond

lengths for the azomethine bridge are greater than the experi-

mental error for crystallographic analysis (the value of 2σ for

the XRD measurements is on the order of 0.004-0.020 Å).

Deviations of 0.03-0.05 Å are quite signiﬁcant when one

considers that the canonical quinoidal and zwitterionic resonance

forms differ only by bond lengths on the order of 0.05-0.10

Å. Importantly, the calculations consistently underestimate the

contribution from the zwitterionic resonance form. Furthermore,

the DFT-optimized structures possess nearly completely planar

bridges (i.e., the R, ꢀ, and, γ dihedral angles are ≈0° or 180°),

which does not accurately reﬂect the slight twisting observed

in the crystal structures. From the calculations alone, one would

predict a higher contribution from the quinoidal resonance form

relative to the zwitterionic form for 1-PMC and 2-PMC, which

would be misleading in light of the crystallographic data.

Although we cannot compare the calculated geometry for

4-PMC with experimental crystalline-state structural data, we

can analyze the computational results in the context of the

solution-state solvatochromism studies. The relative solvato-

chromic behaviors of 2-PMC and 4-PMC suggest that the

azahomoadamantyl and indolyl derivatives possess fundamen-

tally different electronic structures, and that the zwitterionic

resonance contribution is more substantial in the azahomoada-

Comparison of the two pseudopolymorphs of 2-PMC suggests

that 2-PMC-I has a greater contribution from the quinoidal

resonance form A than does 2-PMC-II, though both structures

have more zwitterionic than quinoidal character in general. The

differences between the pseudopolymorphs can be rationalized

by a combination of molecular structure and crystal packing

effects. There is a correlation between the degree of twisting

about the three partial double bonds of the azomethine bridge

and the degree of charge separation in the crystalline-state PMC

structures: 2-PMC-II displays a greater deviation from planarity

and a correspondingly greater degree of zwitterionic-like bond-

length alternation than does 2-PMC-I. Whether this twisting is

sterically or electronically imposed is not clear, although the

existence of hydrogen bonding to water solvate molecules in

2

-PMC-II is expected to have a signiﬁcant effect on the

electronic structure of the PMC form through stabilization of

the partial negative charge on the oxygen atom.

As regards possible electronic effects, each photomerocyanine

has a large dipole moment directed along the long axis of the

molecule, and the crystalline lattice itself can effectively act as

a strong dielectric medium. This idea is corroborated by the

fact that the solvatochromic behavior of 2-PMC suggests that

the photomerocyanine is slightly quinoidal in nonpolar solvents,

while its crystal structures provide evidence for signiﬁcant

zwitterionic character. The molecular structure of the PMC form

in the crystalline state may therefore be more similar to that

expected in polar solvents than in nonpolar ones. The orientation

of molecules in the highly ordered and unsolvated crystal of

2

-PMC-I facilitates charge compensation, a behavior similar to

(110) Ziegler, T. Chem. ReV. 1991, 91 (5), 651–667.

that of typical merocyanine aggregates. The packing is such

that the molecular dipole moments in relatively closely as-

sembled head-to-tail dimeric PMC units cancel, while short

(111) El-Azhary, A. A.; Suter, H. U. J. Phys. Chem. 1996, 100 (37), 15056–

1

5063.

(

112) Minkin, V. I.; Metelitsa, A. V.; Dorogan, I. V.; Lukyanov, B. S.;

Besugliy, S. O.; Micheau, J.-C. J. Phys. Chem. A 2005, 109 (42),

9605–9616.

δ+

δ-

interchain N · · ·O contacts help to maintain additional local

J. AM. CHEM. SOC. 9 VOL. 132, NO. 36, 2010 12583

A R T I C L E S

Patel et al.

mantyl PMCs. The DFT-optimized geometries for 2-PMC and

It is clear from the above discussion that DFT displays serious

shortcomings in predicting absolute PMC structures as a function

of substitution and medium. Discrepancies between calculated

and experimental geometries are highly relevant to deriving

structure-property relationships in PMCs, because even small

changes in electronic structuresas manifested by small changes

in molecular geometriesscan have signiﬁcant consequences on

the NLO and photochromic properties of these materials.

However, calculated and experimental geometries are much

more consistent for the SO forms, and DFT may therefore

remain a useful tool for predicting structure-property trends

for spirooxazines based on SO geometries.

4

-PMC are, however, nearly the same in both the gas phase

and more polar media, which does not accurately reﬂect the

differences in electronic structure suggested by the solvato-

chromism studies.

Thus, both crystalline- and solution-state experiments suggest

that the DFT/B3LYP level of theory does not accurately predict

the geometries of metastable photomerocyanines. The deviation

of the computational geometries from the experimental geom-

etries can be understood within the context of the signiﬁcant

effects of molecular environments on the geometry of the PMC

forms. The effect of solvents on solute molecules may be

categorized into speciﬁc and nonspeciﬁc classes, in which

nonspeciﬁc interactions arise from dispersion and electrostatic

interactions between the charge distribution of the solute and

solvent molecules. In crystalline 2-PMC, the crystal packing

of the photomerocyanines can be characterized as governed by

electrostatic interactions between the charged groups. In 2-PMC-

I, the polar groups are compensated by the close vicinity of the

opposite polar groups of neighboring molecules, and in 2-PMC-

II, where the charge compensation is expected to be weaker,

by appropriately oriented water molecules leading to stabiliza-

tion by hydrogen bonding. While electrostatic interactions

between solute and solvent can be described by a moment

expansion such as that given by Onsager reaction ﬁeld theory,

the addition of an effective dielectric using the Onsager model,

Effects of Molecular Structure on the Charge-Separated

Character and Photochromic Properties in Photomerocya-

nines. To date, the only metastable photomerocyanines that have

been structurally characterized are a handful of nitro- and

pyridinium-substituted spiropyran or spirothiopyran derivatives,

which were all shown to exhibit a predominantly zwitterionic

3

8–43

structure in the crystalline state.

These molecules also

display negative solvatochromism, which corroborates their

41,72,119

zwitterionic structure in solution.

Only a few permanent

spirooxazine-based PMCs have been characterized, and these

have been described as predominantly quinoidal in character,

4

4,45

in some cases on the basis of the CdO bond length alone.

Solvatochromic studies of spirooxazine-based PMCs have

indicated that these exhibit positive solvatochromic behavior,

and it was concluded that they possess a predominantly

2

even in high dielectric solvents such as DMSO and H O, was

3

1,34,68,89

not found to be sufﬁcient to reproduce experimental geometries

of the PMC forms, most likely because of speciﬁc and

nonspeciﬁc intermolecular interactions in the crystalline state.

It is well-known that intermolecular interactions may cause the

molecular geometries in the crystalline state to be quite different

quinoidal structure.

Computational studies have also

25

predicted a quinoidal structure for spirooxazine-based PMCs.

The interpretation of these observationsslimited to speciﬁc

derivatives and conditionssled in some instances to ambiguity

regarding the nature of the molecular structure of spiropyrans

113,114

33,120

from those of the free molecules,

e.g., structures of isolated

and spirooxazines and,

in some cases, to the belief that

molecules of pyridinium betaines optimized using DFT were

found to be signiﬁcantly different from those observed in the

crystals.

spirooxazines as a class are quinoidal while spiropyrans are

3

4

zwitterionic. Recently it has been emphasized that the degree

of quinoidal or zwitterionic character in these photomerocya-

1

15

8

2

nines is a function of substitution pattern and medium. For

spirooxazines, recent computational and solvatochromism stud-

ies have predicted zwitterionic character in spirooxazine deriva-

tives with electron-withdrawing groups on the oxazine moiety

and a dependence of the degree of zwitterionic character on

medium polarity. Here, for the ﬁrst time, we have presented

X-ray structural characterization for PMCs of the spirooxazines

class, namely, azahomoadamantyl isoquinoline and phenanthro-

line spirooxazines, and have shown that these derivatives exhibit

a predominantly zwitterionic structure in both the crystalline

state and in polar solvents.

In addition to effects due to the speciﬁc electronic structures

of the PMC forms, there are generalized effects due to

computational methodology. Inconsistencies between calculated

and experimental geometries are not unexpected given the

difﬁculties in reproducing ground states, excited states, charge-

separated states, and acceptor-donor excitations of fully delo-

3

2

1

16,117

calized organic molecules by DFT.

DFT schemes based

on conventional exchange-correlation (XC) functionals have

been found to lead to catastrophic values of polarizability and

other electronic sructure factors for push-pull conjugated

systems. The overly large µ values are associated with an

overestimation of the charge transfer between the donor and

the acceptor, and incomplete screening of the external electric

ﬁeld is responsible for the large discrepancies. Current XC

functionals were found to incorrectly describe the polarization

of conjugated systems when the polarization is due to

Trends in the properties of spiropyrans and spirooxazines can

be evaluated in terms of the substituents on the amine and

3

4

oxazine functionalities. Authors have discussed these trends

8

2

in terms of “donor-acceptor” effects, “push-pull” interac-

3

2

121

tions, or chemical hardness. One of the trends that can be

understood in this way is the relative contribution of the charge-

separated zwitterionic resonance form to the PMC molecular

structure. Because charge separation in these systems results in

a greater distribution of positive charge at the amine moiety

and of negative charge at the oxazine moiety, the presence of

1

18

donor-acceptor substitution or an external ﬁeld or both.

(

113) Hargittai, I. Pure Appl. Chem. 1989, 61 (4), 651–660.

114) Chiang, J. F.; Song, J. J. J. Mol. Struct. 1982, 96 (1-2), 151–162.

115) Szafran, M.; Dega-Szafran, Z.; Katrusiak, A.; Buczak, G.; Głlowiak,

T.; Sitkowski, J.; Stefaniak, L. J. Org. Chem. 1998, 63, 2898–2908.

116) Choi, C. H.; Kertesz, M.; Karpfen, A. J. Chem. Phys. 1997, 107

(

17), 6712–6721.

(119) Wojtyk, J. T. C.; Wasey, A.; Kazmaier, P. M.; Hoz, S.; Buncel, E.

J. Phys. Chem. A 2000, 104 (39), 9046–9055.

(120) Chibisov, A. K.; G o¨ rner, H. J. Phys. Chem. A 1999, 103 (26), 5211–

5216.

117) Masunov, A. M.; Tretiak, S. J. Phys. Chem. B 2004, 108 (3), 899–

9

07.

118) Champagne, B.; Perp e` te, E. A.; Jacquemin, D.; van Gisbergen,

S. J. A.; Baerends, E.-J.; Soubra-Ghaoui, C.; Robins, K. A.; Kirtman,

B. J. Phys. Chem. A 2000, 104 (20), 4755–4763.

(121) Dom ´ı nguez, M.; Rezende, M. C. J. Phys. Org. Chem. 2010, 23 (2),

156–170.

1

2584 J. AM. CHEM. SOC. 9 VOL. 132, NO. 36, 2010

Charge Separation in Metastable Photomerocyanines

A R T I C L E S

electron-withdrawing groups (EWGs) on the amine would be

expected to destabilize the zwitterionic form, while their

presence on the oxazine would be expected to stabilize it.

azahomoadamantyl derivatives 1 and 2 are no exception to this

trend, and exhibit particularly long Cspiro-O bond lengths.

However, whereas for previously reported spirooxazines, these

long bonds could be explained by a stronger anomeric effect

caused by a greater removal of electron density from the oxygen

atom due to EWGs on the oxazine moiety, for the azahomoada-

mantyl derivatives they can instead be explained by the greater

availability of the lone pair on the nitrogen atom as a result of

the electron-donating properties of the amine functionality. As

regards photochromic properties, long spirooxazine Cspiro-O

bond lengths are expected to lead to lower activation energies

for the initial bond cleavage step, both thermally and photo-

chemically, though this may not be the most important step in

determining forward thermal isomerization rates or photocol-

3

4,82

To date, all of the spiropyran or spirooxazine systems which

have been found to exhibit zwitterionic charactersas determined

by XRD, solvatochromism, or computational studiessare in-

dolyl-based and possess strong EWGs on the oxazine function-

ality. The substitution of EDGs on the amine does not have a

strong enough electronic effect to signiﬁcantly stabilize the

zwitterionic resonance form, at least as predicted by solvato-

chromism or computational studies (no such PMC derivatives

have been structurally characterized). The changes in the

electronic structure of the PMC form resulting from the

incorporation of the azahomoadamantyl group are apparently

much more signiﬁcant than those resulting from the substitution

of EDGs on the amine moiety for typical indolyl derivatives.

Whereas the stabilization of the zwitterionic resonance form is

typically explained by the delocalization of negative charge onto

an electron-withdrawing nitro or pyridinium substituent on the

oxazine functionality, this is clearly not the case in the

azahomoadamantyl derivatives, in which the zwitterionic form

appears to be favored through the stabilization of positive charge

by the strongly electron-donating azahomoadamantyl group.

8

2,122

oration quantum yields.

Conclusions

We have presented investigations of medium and substitution

effects on the molecular structure of photomerocyanines of the

spirooxazines class. The structural features examined here

suggest that bicyclic azahomoadamantyl substitution leads to

signiﬁcant stabilization of the zwitterionic resonance form

without destroying photochromic activity. A comparison of the

vibrational structure and solvatochromism of the PMC π-π*

absorption band of the indolyl and azahomoadamantyl deriva-

tives suggests that the two families have fundamentally different

ground-state structures, where the indolyl derivatives have more

quinoidal character in a medium of given polarity. The two

spirooxazine families exhibit different ranges of accessible

quinoidal/zwitterionic resonance contributions; even in very high

polarity solvents, the indolyl derivatives do not show negative

solvatochromic evidence of zwitterionic character, while the

azahomoadamantyl derivatives exhibit negative solvatochromism

in both hydrogen-bond donating and non-hydrogen-bond donat-

ing polar solvents. The electronic properties observed in the

azahomoadamantyl derivatives are consistent with the amine

functionality acting as a very strong electron-donating group.

Azahomoadamantyl derivatization is therefore a synthetically

feasible and potentially powerful approach to designing spiroox-

azines in which the PMC isomer has signiﬁcant charge-separated

character and associated photochromic properties such as slow

thermal back reactions.

The XRD analysis of metastable photomerocyanines from

the spirooxazine class of photochromic molecules in azaho-

moadamantyl isoquinoline and phenanthroline derivatives in-

dicates a substantial contribution from the canonical zwitterionic

resonance form in the crystalline state. A comparison of the

XRD and solvatochromism for the azahomoadamantyl PMCs

suggests that because a quinoidal structure is inferred from

solution studies in nonpolar solvents, yet a predominantly

zwitterionic structure is evident in the crystalline state, the crystal

lattice may act as a strong dielectric medium. A shift in the

bond-length alternation of the azomethine bridge in two

pseudopolymorphs of 2-PMC suggests that changes in speciﬁc

intermolecular interactions and dielectric of the crystalline lattice

affect the relative contributions of the quinoidal and zwitterionic

resonance forms. These results have important implications for

the sensitivity of charge-separated states to their environment

in constrained media. Finally, the direct XRD evidence for the

The photochemical and thermal isomerization mechanisms

and rates in spirooxazines can be interpreted in the context of

their structural features. For the ground-state thermal isomer-

ization process, general consensus is that the SO f PMC

isomerization mechanism involves at least a C-O bond cleavage

step to form a cis intermediate, and a cis-to-trans isomerization

3

4,82,83

step to form the trans PMC isomer.

For the forward SO

f PMC thermal isomerization process, the transition states and

energy barriers for these steps may be molecular-structure- and

medium-dependent. However, the reverse PMC f SO thermal

isomerization process is more straightforward to analyze as the

rate-limiting step is necessarily trans-to-cis isomerization, and

the rate constant for this process, k

the activation energy for bond rotation.

lower k values for the azahomoadamantyl spirooxazines relative

2

, can be directly related to

8

2,83,119

We observed

2

to the indolyl spirooxazines. This is consistent with the

azahomoadamantyl derivatives having greater zwitterionic

character, and therefore a higher bond order for the central C-N

bond of the azomethine bridge, which would impede trans-to-

cis isomerization and lead to higher energy barriers for this

8

2,119

process.

2

The k values were also found to be lower in the

isoquinoline spirooxazines relative to the phenanthroline spiroox-

azines, which suggests that the prior have more zwitterionic

character. Finally, the lower k values observed in more polar

2

solvents can be rationalized by a greater stabilization of the PMC

form and of the zwitterionic resonance form with increasing

solvent polarity. The relative energy difference of the PMC and

SO forms (∆E ) ESO - EPMC) as derived from computation

[

3

DFT B3LYP/6-31G(d)] for compounds 1, 2, and 4 are 20.4,

.8, and 1.4 kcal/mol, respectively, consistent with stabilization

of the open form and increasing rates of PMC f SO conversion

) along the series. All of the structural features highlighted

(k

2

here are consistent with the XRD and solvatochromism studies

described above.

The degree of quinoidal or zwitterionic character in PMCs

can also be correlated with the structure of the SO form. Those

structurally characterized PMCs which have strong EWGs on

the oxazine moiety and exhibit zwitterionic character also tend,

in the SO form, to have Cspiro-O bond lengths at the upper

(

122) Maurel, F.; Aubard, J.; Millie, P.; Dognon, J. P.; Rajzmann, M.;

Guglielmetti, R.; Samat, A. J. Phys. Chem. A 2006, 110 (14), 4759–

4771.

4

2

limit of those found for spirooxazines in general. The

J. AM. CHEM. SOC. 9 VOL. 132, NO. 36, 2010 12585

A R T I C L E S

Patel et al.

structures of the metastable spirooxazine-based PMCs reported

here, in combination with the investigation of medium polarity

effects on PMC structure in both the crystalline and solution

states, may lead to a better understanding of medium effects

and structural features governing the degree of charge-separation

in organic systems.

J ) 2.1 Hz, 1H, H-6), 3.51 (s, 3H, H-1), 2.20-1.75 (m, 12H); SO

form: δ 9.14 (dd, J ) 4.5, 2.1 Hz, 1H, H-24), 8.86 (dd, J ) 8.1,

1

.4 Hz, 1H, H-19), 7.63 (dd, J ) 8.1, 4.4 Hz, 1H, H-23), 7.61 (dd,

J ) 8.1, 4.4 Hz, 1H, H-20), 7.70 (s, 1H, H-3), 2.47 (s, 3H, H-1),

.20-1.75 (m, 12H) ppm (the remaining signals for the SO form

2

13

are not distinguishable from those of the PMC form). C NMR

125 MHz, CDCl ) PMC and SO forms: δ 179.5 (C), 174.5 (C),

57.36 (CH), 152.6 (CH), 151.2 (C), 150.4 (CH), 147.8 (CH), 146.8

(

1

3

Experimental Section

(

CH), 146.4 (C), 143.5 (C), 140.1 (C), 135.2 (CH), 133.1 (C), 130.9

CH), 130.4 (CH), 129.9 (CH), 129.7 (C), 127.8 (C), 127.1 (C),

Spectroscopic Methods. All electronic absorption spectroscopy

was performed with an Agilent 8453 spectrometer. For the

photoisomerization experiments, solutions were prepared at con-

1

(

3

(

25.9 (CH), 123.5 (CH), 123.3 (CH), 122.9 (CH), 122.7 (CH), 122.1

C), 119.4 (C), 95.2 (C), 77.2 (C/CH), 64.5 (CH), 58.9 (CH), 42.9

CH), 39.8 (CH ), 39.7 (CH ), 39.4 (CH ), 38.4 (CH ), 36.0 (CH ),

4.0 (CH ), 33.9 (CH ), 33.6 (CH ), 32.3 (CH ), 30.2 (CH ), 30.0

CH ), 29.3 (CH/CH ), 26.7 (CH/CH ), 26.7 (CH/CH ), 26.2 (CH/

) ppm. EI-MS: m/z (%) 384 (91) [M] , 163 (100). Anal. Calcd

: C, 74.98; H, 6.29; N, 14.57. Found: C, 74.71; H,

.17, N, 14.36.

,3-Dihydro-1,3,3-trimethylspiro[2H-indole-2,3′-[3H]pyrido-

3,4-f][1,4]benzoxazine] {Spiro[indoline-isoquinolinoxazine] (3)}.

-5

-6

centrations of 10 -10 M in spectrograde solvents. The samples

were exposed to steady-state irradiation from a high-pressure Hg

arc lamp operating at 60 W with glass cutoff ﬁlters to obtain the

desired wavelength range (280 < λ < 400 nm for UV irradiation,

and λ > 500 nm for visible irradiation). The observed rate constants

of thermal reversion were determined in the absence of light by

following the absorbance kinetics at λmax after generating a

photostationary state, and ﬁtting the data to monoexponential rate

functions by linear least-squares methods. Runs were repeated

several times and averaged for accuracy.

2

3

+

CH

3

for C24H24ON

4

6

1

[

126

7

-Hydroxy-8-nitrosoisoquinoline (7) (206 mg, 1.18 mmol) and

1

,3,3-trimethyl-2-methylene indoline (15) (0.209 mL, 1.18 mmol)

Computational Methods. Geometry optimization calculations

were heated at reﬂux in 2-propanol (30 mL) over 4 Å molecular

sieves for 3.5 h. The solvent was removed by rotary evaporation,

and the material was puriﬁed by column chromatograph on silica

were performed using density functional theory (DFT) with the

123

hybrid Becke-style three-parameter exchange functional and the

124

Lee-Yang-Parr correlation functional (B3LYP) with the Gauss-

1

25

gel ﬁrst with CH

2 2 3

Cl /CH OH (95:5) and then with EtOAc/hexanes

ian 03 software package.

The optimization for 3-PMC was

(

40:60). Recrystallization from EtOAc yielded yellow crystals (136

performed from a starting geometry based on the X-ray diffraction

analysis atomic coordinates. Solvent effects were modeled in toluene

1

mg, 35% yield). Mp: 163-164 °C. H NMR (360 MHz, CDCl

δ 9.95 (t, J ) 0.8 Hz, 1H, H-19), 8.49 (d, J ) 5.7 Hz, 1H, H-20),

.80 (s, 1H, H-3), 7.62 (br d, J ) 8.9 Hz, 1H, H-16), 7.53 (dd, J

5.7, 0.8 Hz, 1H, H-21), 7.21 (td, J ) 7.8, 1.3 Hz, 1H, H-13),

3

):

(

ε ) 2.379), DMSO (ε ) 46.7), and H

2

O (ε ) 78.39) using the

7

)

Onsager model with a values of 5.64 Å for 2-PMC and 5.95 Å

o

for 4-PMC as determined by a volume calculation in Gaussian.

Stability calculations were performed on all optimized geometries

obtained from gas-phase calculations, and structures were found

to be minima on the potential energy surfaces.

7

.20 (d, J ) 8.8 Hz, 1H, H-15), 7.08 (dd, J ) 7.3, 1.1 Hz, 1H,

H-11), 6.90 (td, J ) 7.3, 0.8 Hz, 1H, H-12), 6.57 (br d, J ) 7.8

Hz, 1H, H-14), 2.75 (s, 3H, H-1), 1.36 (s, 3H, H-7/H-8), 1.34 (s,

1

3

4

6

19 63 64 65

3H, H-7/H-8) ppm. C NMR (90 MHz, CDCl

147.3 (Ar CH), 144.5 (Ar C), 141.9 (Ar CH), 135.6 (Ar C), 131.4

Ar C), 128.7 (Ar CH), 128.1 (Ar CH), 125.1 (Ar C), 123.0 (Ar

C), 121.6 (Ar CH), 121.5 (Ar CH), 120.0 (Ar CH), 119.9 (Ar CH),

3

): δ 152.4 (Ar CH),

Synthesis of Spirooxazines. Compounds 1, 4, 6, 7, 9,

66

67

1

0, 12, and 13 were prepared according to previously reported

(

1

procedures and characterized by H NMR spectroscopy and/or

1

melting points (Supporting Information). The H NMR spectra of

1

07.2 (CH), 99.0 (C), 77.2 (C), 52.0 (C), 29.6 (CH

3

), 25.4 (CH

3

),

) ppm. EI-MS: m/z (%) 329 (57) [M] , 314 (60), 159

O: C, 76.57; H, 5.81; N,

2.76. Found: C, 76.81; H, 5.91, N, 12.50.

2

and 3 were assigned with COSY and NOESY spectroscopy and

are labeled according to the connecting carbon atoms (see Figures

and 7 for atom numbering).

-Methylspiro[4-azahomoadamantane-5,2′-[2H-1,4]oxazino-

2,3-f][1,10]phenanthroline] {Spiro[azahomoadamantane-phen-

anthrolinoxazine] (2)}. Et N (1.60 mL, 11.5 mmol) was added to

a stirring solution of 4,5-dimethyl-4-azahomoadamant-4-enium

iodide (13) (1.72 g, 5.62 mmol) in CH Cl

0 min, 4 Å molecular sieves (∼3 g) and 5-hydroxy-6-nitroso-

,10-phenanthroline (10) (1.27 g, 5.63 mmol) were added to the

+

2

0.7 (CH

3

(

19 3

100), 144 (41). Anal. Calcd for C21H N

5

1

4

[

Acknowledgment. The authors thank the American Chemical

3

Society Petroleum Research Fund (ACS-PRF), the British Columbia

Knowledge Development Fund (BCKDF), the Canada Foundation

for Innovation (CFI), the Natural Sciences and Engineering Council

of Canada (NSERC), the University of Victoria, and the University

of Washington for ﬁnancial support of this research. M.M.P. thanks

NSERC for an NSERC-CGS fellowship and PEO International for

a PEO Scholar Award.

2

(150 mL) at 0 °C. After

3

1

mixture, and the solution was slowly warmed to rt and then heated

at reﬂux for 3 h. The solution was ﬁltered, washed with NaHCO

O, and brine, and then dried over MgSO , before removal of the

solvent by rotary evaporation to yield a dark purple solid (1.4 g,

5% yield) of reasonable purity. The material was further puriﬁed

by column chromatography on silica gel with CH Cl /CH OH (90:

0) followed by recrystallization from EtOAc to yield iridescent

3

,

H

2

4

Supporting Information Available: Characterization of com-

pounds by NMR, packing diagrams and geometric parameters

from XRD analysis, crystallographic data and cif ﬁles, electronic

absorption spectroscopy of 1 in representative solvents, details

of spectral deconvolution for 2 and 4, concentration dependence

studies for 2 and 4, theoretical and experimental determination

6

2

3

1

green crystals (0.69 g, 32% yield). Mp: 200-201 °C. H NMR

300 MHz, CDCl ) PMC form: δ 9.97 (s, 1H, H-3), 9.03 (dd, J )

.5, 1.7 Hz, 1H, H-24), 8.75 (dd, J ) 7.9, 1.7 Hz, 1H, H-22), 8.74

(

3

4

(

dd, J ) 4.4, 2.1 Hz, 1H, H-21), 8.64 (dd, J ) 8.1, 2.2 Hz, 1H,

1

13

of H and C NMR shifts of 2 and 4, complete ref 125, basis

set comparison, and Gaussian outputs. This material is available

free of charge via the Internet at http://pubs.acs.org.

H-19), 7.48 (dd, J ) 8.1, 4.4 Hz, 1H, H-23), 7.39 (dd, J ) 8.0, 4.4

Hz, 1H, H-20), 5.11 (tt, J ) 5.9, 1.9 Hz, 1H, H-10), 3.74 (septet,

(

123) Becke, A. D. J. Chem. Phys. 1993, 98 (7), 5648–5652.

JA100238H

124) Lee, C. T.; Yang, W. T.; Parr, R. G. Phys. ReV. B 1988, 37 (2),

7

85–789.

(

125) Frisch, M. J.; et al. Gaussian 03, Revision E.01; Gaussian, Inc.:

(126) Salemi-Delvaux, C.; Giusti, G.; Guglielmetti, R.; Dubest, R.; Aubard,

Wallingford CT, 2004.

J. J. Chim. Phys. Phys.-Chim. Biol. 1998, 95 (9), 2001–2014.

1

2586 J. AM. CHEM. SOC. 9 VOL. 132, NO. 36, 2010

Article Doi

A solution- and solid-state investigation of medium effects on charge separation in metastable photomerocyanines

DOI: 10.1021/ja100238h

Source and publish data:

Authors:

Article abstract of DOI:10.1021/ja100238h

Full text of DOI:10.1021/ja100238h

Products guided by the article

R&D Labs maybe for 7651-83-4

Relevant to this article

Hot Product