Syntheses and properties of two-dimensional charged nonlinear optical chromophores incorporating redox-switchable cis-tetraammineruthenium(II) centers

Article abstract of DOI:10.1021/ja0424124

In this article, we describe a series of new complex salts in which electron-donating cis-{Ru^II(NH₃)₄}²⁺ centers are connected to two electron-accepting N-methyl/aryl-pyridinium groups. These V-shaped complexes contain either monodentate 4,4′- bipyridyl-derived ligands or related chelates based on 2,2′:4,4″: 4′,4?-quaterpyridyl and have been characterized by using various techniques including electronic absorption spectroscopy and cyclic voltammetry. Molecular quadratic nonlinear optical (NLO) responses β have been determined by using hyper-Rayleigh scattering at 800 nm and also via Stark (electroabsorption) spectroscopic studies on the intense, visible d → π* metal-to-ligand charge-transfer bands. These experiments reveal that these dipolar pseudo-C_2v chromophores exhibit two substantial components of the β tensor, β_zzz and β_zyy, with the difference between them being most marked for the nonchelated systems. Time-dependent density-functional theory and finite field calculations serve to further illuminate the electronic structures and associated linear and NLO properties of the new chromophoric salts.

Full text of DOI:10.1021/ja0424124

Published on Web 03/10/2005
Syntheses and Properties of Two-Dimensional Charged
Nonlinear Optical Chromophores Incorporating
Redox-Switchable cis-Tetraammineruthenium(II) Centers
Benjamin J. Coe,*,† James A. Harris,†, Lathe A. Jones,† Bruce S. Brunschwig,‡,
Kai Song,§ Koen Clays,§ Javier Gar´ın,| Jesu´s Orduna,| Simon J. Coles,# and
Michael B. Hursthouse#
Contribution from the School of Chemistry, UniVersity of Manchester, Oxford Road,
Manchester, U.K. M13 9PL, Chemistry Department, BrookhaVen National Laboratory,
P.O. Box 5000, Upton, New York 11973-5000, Laboratory of Chemical and Biological
Dynamics, Center for Research on Molecular Electronics and Photonics, UniVersity of LeuVen,
Celestijnenlaan 200D, B-3001 LeuVen, Belgium, Departamento de Qu´ımica Orga´nica, ICMA,
UniVersidad de Zaragoza-CSIC, E-50009 Zaragoza, Spain, and EPSRC National
Crystallography SerVice, School of Chemistry, UniVersity of Southampton,
Highfield, Southampton, U.K. SO17 1BJ
Received December 17, 2004; E-mail: b.coe@man.ac.uk
Abstract: In this article, we describe a series of new complex salts in which electron-donating cis-
{RuII(NH3)4}2+ centers are connected to two electron-accepting N-methyl/aryl-pyridinium groups. These
V-shaped complexes contain either monodentate 4,4′-bipyridyl-derived ligands or related chelates based
on 2,2′:4,4′′:4′,4′′′-quaterpyridyl and have been characterized by using various techniques including electronic
absorption spectroscopy and cyclic voltammetry. Molecular quadratic nonlinear optical (NLO) responses â
have been determined by using hyper-Rayleigh scattering at 800 nm and also via Stark (electroabsorption)
spectroscopic studies on the intense, visible d f π* metal-to-ligand charge-transfer bands. These
experiments reveal that these dipolar pseudo-C2v chromophores exhibit two substantial components of the
â tensor, âzzz and âzyy, with the difference between them being most marked for the nonchelated systems.
Time-dependent density-functional theory and finite field calculations serve to further illuminate the electronic
structures and associated linear and NLO properties of the new chromophoric salts.
Introduction
for practical applications, derive at the molecular level from
first hyperpolarizabilities â. The majority of chromophores
Nonlinear optical (NLO) materials based on molecular
compounds are of considerable current interest because they
hold promise for applications in advanced optoelectronic and
all-optical data processing technologies.1 Fundamental research
in this area has recently featured an increasing number of studies
on organotransition metal complexes which offer greater scope
for the creation of multifunctional NLO materials when
compared with purely organic compounds.2 Quadratic (second-
order) NLO properties, which are of the most immediate interest
studied to date, whether metal-containing or purely organic,
contain an electron donor group (D) connected to an electron
acceptor (A) via a polarizable π-conjugated bridge. The mo-
lecular optical nonlinearities of such simple dipolar species are
largely one-dimensional (1D) in nature, i.e. dominated by one
â tensor component. Introducing multiple D and/or A groups
increases the number of significant components of â, which then
acquires a greater degree of multidimensional (MD) character.
(2) (a) Nalwa, H. S. Appl. Organomet. Chem. 1991, 5, 349-377. (b) Marder,
S. R. In Inorganic Materials, 2nd ed.; Bruce, D. W., O’Hare, D., Eds.;
Wiley: Chichester; 1992; pp 121-169. (c) Kanis, D. R.; Ratner, M. A.;
Marks, T. J. Chem. ReV. 1994, 94, 195-242. (d) Long, N. J. Angew. Chem.,
Int. Ed. Engl. 1995, 34, 21-38. (e) Whittall, I. R.; McDonagh, A. M.;
Humphrey M. G.; Samoc, M. AdV. Organomet. Chem. 1998, 42, 291-
362. (f) Whittall, I. R.; McDonagh, A. M.; Humphrey M. G.; Samoc, M.
AdV. Organomet. Chem. 1998, 43, 349-405. (g) Heck, J.; Dabek, S.;
Meyer-Friedrichsen, T.; Wong, H. Coord. Chem. ReV. 1999, 190-192,
1217-1254. (h) Gray, G. M.; Lawson, C. M. In Optoelectronic Properties
of Inorganic Compounds; Roundhill, D. M.; Fackler, J. P., Jr., Eds.;
Plenum: New York, 1999; pp. 1-27. (i) Shi, S. In Optoelectronic
Properties of Inorganic Compounds; Roundhill, D. M.; Fackler, J. P., Jr.,
Eds.; Plenum: New York, 1999; pp. 55-105. (j) Le Bozec, H.; Renouard,
T. Eur. J. Inorg. Chem. 2000, 229-239. (k) Barlow, S.; Marder, S. R.
Chem. Commun. 2000, 1555-1562. (l) Lacroix, P. G. Eur. J. Inorg. Chem.
2001, 339-348. (m) Di Bella, S. Chem. Soc. ReV. 2001, 30, 355-366. (n)
Coe, B. J. In ComprehensiVe Coordination Chemistry II; McCleverty J.
A., Meyer, T. J., Eds.; Elsevier Pergamon: Oxford, 2004; Vol. 9, pp 621-
687.
† University of Manchester.
‡ Brookhaven National Laboratory.
§ University of Leuven.
| Universidad de Zaragoza.
# University of Southampton.
Present address: Molecular Materials Research Center, Beckman
Institute, MC 139-74, California Institute of Technology, 1200 East
California Blvd., Pasadena, CA 91125.
(1) (a) Nonlinear Optical Properties of Organic Molecules and Crystals;
Chemla, D. S., Zyss, J., Eds.; Academic Press: Orlando, FL, 1987; Vols.
1 and 2. (b) Zyss, J. Molecular Nonlinear Optics: Materials, Physics and
DeVices; Academic Press: Boston, 1994. (c) Organic Nonlinear Optical
Materials; Bosshard, Ch., Sutter, K., Preˆtre, Ph., Hulliger, J., Flo¨rsheimer,
M., Kaatz, P., Gu¨nter, P., Eds.; Advances in Nonlinear Optics, Vol. 1;
Gordon & Breach: Amsterdam, The Netherlands, 1995. (d) Nonlinear
Optics of Organic Molecules and Polymers; Nalwa, H. S., Miyata, S., Eds.;
CRC Press: Boca Raton, FL, 1997.
9
10.1021/ja0424124 CCC: $30.25 © 2005 American Chemical Society
J. AM. CHEM. SOC. 2005, 127, 4845-4859
4845

A R T I C L E S
Coe et al.
Various organic MD NLO compounds have recently emerged
as candidates for quadratic applications,3 offering potential
advantages over 1D chromophores such as increased â responses
without undesirable losses of transparency in the visible region.
Of particular relevance to the present report, it has been shown
that off-diagonal tensor components of â can become significant
in dipolar molecules with C2V symmetry that display electronic
transitions for which the direction of the transition dipole
moment µ12 is perpendicular to the C2 axis.3a Examples of such
molecules include (dicyanomethylene)pyran derivatives3d,i or
asymmetrically substituted 1,3,5-tris(phenylethynyl)benzenes;3f,h
compounds of this type are attractive for electrooptic applica-
tions and also offer new possibilities for achieving phase-
matched second harmonic generation. A relatively small number
of neutral dipolar 2D metal complexes have also been studied,2l,4
in particular Schiff base derivatives, but to our knowledge no
related charged metallochromophores have yet been investigated
for their NLO properties.
in the creation of two series of novel dipolar 2D C2V chro-
mophores, each of which features a cis-{RuII(NH3)4}2+ D center
connected to two pyridinium A groups. A number of salts of
related trans complexes have also been prepared; although such
centrosymmetric or pseudocentrosymmetric species are not
expected to possess substantial â responses, they are of interest
largely for comparison of their spectroscopic and electrochemi-
cal properties with those of the cis species. The molecular
quadratic NLO properties of our new compounds have been
assessed directly by using the hyper-Rayleigh scattering (HRS)
technique7 and also indirectly via electronic Stark effect
(electroabsorption) spectroscopy.8 In addition, time-dependent
density functional theory (TD-DFT) and finite field calculations
have been carried out to further illuminate the electronic
structures of the complexes studied. Some of the experimental
results described herein for the nonchelated cis-{RuII(NH3)4}2+
complex salts have been previously communicated.9
Experimental Section
We have previously described systematic studies on 1D
ruthenium ammine complex salts which possess very large NLO
responses that are tunable by ligand-based changes5 and also
reversibly switchable via RuIII/II redox.6 The latter general
observation is particularly significant because it provides a key
incentive for incorporating RuII ammine centers into NLO
chromophores. We have now employed similar design strategies
Materials and Procedures. All reactions were performed under an
Ar atmosphere and in Ar-purged solvents. All reactions and chromato-
graphic purifications involving complex salts were performed in the
dark. The compounds cis-[RuIIICl2(NH3)4]Cl,10 trans-[RuIICl(NH3)4-
(SO2)]Cl,11 [RuII(NH3)5(H2O)][PF6]2,11 [MeQ+]I,5a [PhQ+]Cl‚2H2O,5c
[4-AcPhQ+]Cl‚2H2O,5c [2-PymQ+]Cl,5d and N′′,N′′′-dimethyl-2,2′:4,4′′:
4′,4′′′-quaterpyridinium hexafluorophosphate ([Me2Qpy2+][PF6]2)12 were
synthesized by previously published methods. All other reagents were
obtained commercially and used as supplied. Products were dried
overnight in a vacuum desiccator (CaSO4) prior to characterization.
General Physical Measurements. 1H NMR spectra were recorded
on a Varian Gemini 200 spectrometer, and all shifts are referenced to
TMS. The fine splitting of pyridyl or phenyl ring AA′BB′ patterns is
ignored, and the signals are reported as simple doublets, with J values
referring to the two most intense peaks. Elemental analyses were
performed by the Microanalytical Laboratory, University of Manchester
and UV/VIS spectra were obtained by using a Hewlett-Packard 8452A
diode array spectrophotometer. Infrared spectroscopy was performed
on nondiluted samples using an Excalibur BioRad FT-IR spectrometer
equipped with a Golden Gate attenuated total reflectance accessory.
Cyclic and differential pulse voltammetric measurements were
performed on a BAS CV50W voltammetric analyzer. A single-
compartment cell was used with a silver/silver chloride reference
electrode separated by a salt bridge from a Pt disk working electrode
and Pt wire auxiliary electrode. Acetonitrile was freshly distilled (from
(3) Selected examples: (a) Wortmann, R.; Kra¨mer, P.; Glania, C.; Lebus S.;
Detzer, N. Chem. Phys. 1993, 173, 99-108. (b) Zyss, J. J. Chem. Phys.
1993, 98, 6583-6599. (c) Zyss J.; Ledoux, I. Chem. ReV. 1994, 94, 77-
105. (d) Moylan, C. R.; Ermer, S.; Lovejoy, S. M.; McComb, I.-H.; Leung,
D. S.; Wortmann, R.; Krdmer P.; Twieg, R. J. J. Am. Chem. Soc. 1996,
118, 8, 12950-12955. (e) Wortmann, R.; Glania, C.; Kra¨mer, P.;
Matschiner, R.; Wolff, J. J.; Kraft, S.; Treptow, B.; Barbu, E.; La¨ngle D.;
Go¨rlitz, G. Chem. Eur. J. 1997, 3, 1765-1773. (f) Wolff, J. J.; La¨ngle, D.;
Hillenbrand, D.; Wortmann, R.; Matschiner, R.; Glania, C.; Kra¨mer, P.
AdV. Mater. 1997, 9, 138-143. (g) Tomonari, M.; Ookubo N.; Takada, T.
Chem. Phys. Lett. 1997, 266, 488-498. (h) Wolff J. J.; Wortmann, R. AdV.
Phys. Org. Chem. 1999, 32, 121-217. (i) Liu, Y.-J.; Liu, Y.; Zhang, D.-
J.; Hu, H.-Q.; Liu, C.-B. J. Mol. Struct. 2001, 570, 43-51. (j) Ostroverkhov,
V.; Petschek, R. G.; Singer, K. D.; Twieg, R. J. Chem. Phys. Lett. 2001,
340, 109-115. (k) Yang M.-L.; Champagne, B. J. Phys. Chem. A 2003,
107, 3942-3951. (l) Wortmann, R.; Lebus-Henn, S.; Reis, H.; Papadopou-
los, M. G. J. Mol. Struct. (THEOCHEM) 2003, 633, 217-226.
(4) Selected examples: (a) Lacroix, P. G.; Di Bella, S.; Ledoux, I. Chem. Mater.
1996, 8, 541-545. (b) Di Bella, S.; Fragala`, I.; Ledoux, I.; Diaz-Garcia,
M. A.; Marks, T. J. J. Am. Chem. Soc. 1997, 119, 9550-9557. (c) Averseng,
F.; Lacroix, P. G.; Malfant, I.; Lenoble, G.; Cassoux, P.; Nakatani, K.;
Maltey-Fanton, I.; Delaire, J. A.; Aukauloo, A. Chem. Mater. 1999, 11,
995-1002. (d) Briel, O.; Su¨nkel, K.; Krossing, I.; No¨th, H.; Schma¨lzlin,
E.; Meerholz, K.; Bra¨uchle, C.; Beck, W. Eur. J. Inorg. Chem. 1999, 483-
490. (e) Hilton, A.; Renouard, T.; Maury, O.; Le Bozec, H.; Ledoux, I.;
Zyss, J. Chem. Commun. 1999, 2521-2522. (f) Di Bella, S.; Fragala`, I.;
Ledoux, I.; Zyss, J. Chem. Eur. J. 2001, 7, 3738-3743. (g) Di Bella, S.;
Fragala`, I. New J. Chem. 2002, 26, 285-290. (h) Di Bella, S.; Fragala`, I.
Eur. J. Inorg. Chem. 2003, 2606-2611. (i) Maya, E. M.; Garc´ıa-Frutos,
E. M.; Va´zquez, P.; Torres, T.; Mart´ın, G.; Rojo, G.; Agullo´-Lo´pez, F.;
Gonza´lez-Jonte, R. H.; Ferro, V. R.; Garc´ıa de la Vega, J. M.; Ledoux, I.;
Zyss, J. J. Phys. Chem. A 2003, 107, 2110-2117. (j) Di Bella, S.; Fragala`,
I.; Guerri, A.; Dapporto, P.; Nakatani, K. Inorg. Chim. Acta 2004, 357,
1161-1167.
CaH2), and [NBun ]PF6, twice recrystallized from ethanol, and dried
4
in vacuo, was used as the supporting electrolyte. Solutions containing
ca. 10-3 M analyte (0.1 M electrolyte) were deaerated by purging with
N2. All E1/2 values were calculated from (Epa + Epc)/2 at a scan rate of
200 mV s-1
.
Synthesis of 2,2′:4,4′′:4′,4′′′-Quaterpyridyl (Qpy). A mixture of
4,4′-bipyridyl (20 g, 128 mmol) and 10% Pd on charcoal (4 g) was
heated at ca. 250 °C for 48 h and then cooled to room temperature.
(5) (a) Coe, B. J.; Chamberlain, M. C.; Essex-Lopresti, J. P.; Gaines, S.; Jeffery,
J. C.; Houbrechts, S.; Persoons, A. Inorg. Chem. 1997, 36, 3284-3292.
(b) Coe, B. J.; Essex-Lopresti, J. P.; Harris, J. A.; Houbrechts S.; Persoons,
A. Chem. Commun. 1997, 1645-1646. (c) Coe, B. J.; Harris, J. A.;
Harrington, L. J.; Jeffery, J. C.; Rees, L. H.; Houbrechts S.; Persoons, A.
Inorg. Chem. 1998, 37, 3391-3399. (d) Coe, B. J.; Harris, J. A.;
Asselberghs, I.; Persoons, A.; Jeffery, J. C.; Rees, L. H.; Gelbrich T.;
Hursthouse, M. B. J. Chem. Soc., Dalton Trans. 1999, 3617-3625. (e)
Coe, B. J.; Harris, J. A.; Brunschwig, B. S. J. Phys. Chem. A 2002, 106,
897-905. (f) Coe, B. J.; Jones, L. A.; Harris, J. A.; Brunschwig, B. S.;
Asselberghs, I.; Clays, K.; Persoons, A. J. Am. Chem. Soc. 2003, 125, 862-
863. (g) Coe, B. J.; Jones, L. A.; Harris, J. A.; Sanderson, E. E.; Brunschwig,
B. S.; Asselberghs, I.; Clays, K.; Persoons, A. Dalton Trans. 2003, 2335-
2341. (h) Coe, B. J.; Jones, L. A.; Harris, J. A.; Brunschwig, B. S.;
Asselberghs, I.; Clays, K.; Persoons, A.; Gar´ın, J.; Orduna, J. J. Am. Chem.
Soc. 2004, 126, 3880-3891.
(7) (a) Terhune, R. W.; Maker, P. D.; Savage, C. M. Phys. ReV. Lett. 1965,
14, 681-684. (b) Clays, K.; Persoons, A. Phys. ReV. Lett. 1991, 66, 2980-
2983. (c) Clays, K.; Persoons, A. ReV. Sci. Instrum. 1992, 63, 3285-3289.
(d) Hendrickx, E.; Clays, K.; Persoons, A. Acc. Chem. Res. 1998, 31, 675-
683.
(8) (a) Liptay, W. In Excited States; Lim, E. C., Ed.; Academic Press: New
York, 1974; Vol. 1, pp 129-229. (b) Bublitz, G. U.; Boxer, S. G. Annu.
ReV. Phys. Chem. 1997, 48, 213-242.
(9) Coe, B. J.; Harris, J. A.; Brunschwig, B. S. Dalton Trans. 2003, 2384-
2386.
(10) Boggs, S. E.; Clarke R. E.; Ford, P. C. Inorg. Chim. Acta 1996, 247, 129-
130.
(6) (a) Coe, B. J.; Houbrechts, S.; Asselberghs I.; Persoons, A. Angew. Chem.,
Int. Ed. 1999, 38, 366-369. (b) Coe, B. J. Chem. Eur. J. 1999, 5, 2464-
2471.
(11) Curtis, J. C.; Sullivan, B. P.; Meyer, T. J. Inorg. Chem. 1983, 22, 224-
236.
(12) Morgan R. J.; Baker, A. D. J. Org. Chem. 1990, 55, 1986-1993.
9
4846 J. AM. CHEM. SOC. VOL. 127, NO. 13, 2005

Synthesis of Charged 2D Dipolar NLO Chromophores
A R T I C L E S
HCl (2 M, 500 mL) was added, and the resulting black suspension
was filtered. Dropwise addition of aqueous NaOH (8 M) to the stirred
filtrate gave a cream precipitate which was filtered off, washed with
water, and dried. This solid was then loaded into a Soxhlet and extracted
with acetone (4 × 400 mL portions) over several days. Each portion
was treated as follows: any precipitate was filtered off and dried, and
the filtrate was also evaporated to dryness. Both of the resulting solids
C5H4N), 8.31 (2 H, dd, J ) 1.9, 5.1 Hz, C5H3N), 8.08 (2 H, t, J ) 4.9
Hz, C4H3N2). Anal. Calcd (%) for C28H20F12N8P2: C, 44.34; H, 2.66;
N, 14.77. Found: C, 44.67; H, 2.59; N, 14.50.
Synthesis of cis-[RuII(NH3)4(MeQ+)2][PF6]4 (1). A solution of cis-
[RuIIICl2(NH3)4]Cl (55 mg, 0.200 mmol) in degassed water (5 mL)
acidified with trifluoroacetic acid (3 drops) was reduced over zinc
amalgam (5 lumps) for 15 min with argon agitation. The solution was
filtered under argon into a flask containing [MeQ+]I (298 mg, 1.00
mmol), and the reaction was stirred in the dark under argon for 5 h.
Acetone (850 mL) was added, and the crude halide salt product was
collected by filtration, washed with acetone, and dried. This material
was metathesized using water/aqueous NH4PF6 to yield a dark-purple
solid. Purification was achieved by precipitation from acetone/aqueous
NH4PF6: 109 mg, 49%; δH (CD3COCD3) 9.12 (4 H, d, J ) 6.9 Hz,
2C5H4N), 8.88 (4 H, d, J ) 6.9 Hz, 2C5H4N), 8.66 (4 H, d, J ) 7.0
Hz, 2C5H4N), 7.95 (4 H, d, J ) 6.9 Hz, 2C5H4N), 4.61 (6 H, s, 2Me),
3.39 (6 H, s, 2NH3), 2.99 (s, 6 H, 2NH3). Anal. Calcd (%) for
C22H34F24N8P4Ru‚1.5H2O: C, 23.62; H, 3.33; N, 10.02. Found: C,
23.69; H, 3.19; N, 9.63.
1
were then analyzed by H NMR, and materials with similar spectra
were combined. Compounds were extracted in the order of their acetone
solubility, i.e. unreacted 4,4′-bpy first, followed by Qpy, and then
higher-order oligomers. The combined Qpy crops were heated at reflux
in CHCl3 (300 mL) in the presence of activated charcoal (3 g) and
concentrated HCl (5 drops) for 1 h. The resulting mixture was cooled
to room temperature, filtered though Celite 545, and evaporated to
dryness to give a white powder: 7.48 g, 37%; δH (CDCl3) 8.82 (2 H,
d, J ) 5.1 Hz, C5H3N), 8.79-8.76 (6 H, m, C5H3N + C5H4N), 7.68 (4
H, d, J ) 6.1 Hz, C5H4N), 7.60 (2 H, dd, J ) 1.8, 5.1 Hz, C5H3N).
Anal. Calcd (%) for C20H14N4‚0.5H2O: C, 75.22; H, 4.73; N, 17.54.
Found: C, 75.03; H, 4.42; N, 17.38.
Synthesis of cis-[RuII(NH3)4(PhQ+)2][PF6]4 (2). This was prepared
in manner similar to 1 by using [PhQ+]Cl‚2H2O (305 mg, 1.00 mmol)
in place of [MeQ+]I. Acetone (150 mL) was added to precipitate the
crude chloride product. Purification was achieved by sequential
precipitations from acetone/diethyl ether to afford a dark-blue solid:
120 mg, 49%; δH (CD3COCD3) 9.42 (4 H, d, J ) 7.2 Hz, 2C5H4N),
8.95 (4 H, d, J ) 6.9 Hz, 2C5H4N), 8.84 (4 H, d, J ) 7.2 Hz, 2C5H4N),
8.05 (4 H, d, J ) 7.0 Hz, 2C5H4N), 8.01-7.96 (4 H, m, 2Ph), 7.85-
7.79 (6 H, m, 2Ph), 3.45 (6 H, s, 2NH3), 3.04 (6 H, s, 2NH3). Anal.
Calcd (%) for C32H38F24N8P4Ru‚0.5H2O: C, 31.38; H, 3.21; N, 9.15.
Found: C, 31.34; H, 3.24; N, 9.01.
Synthesis of cis-[RuII(NH3)4(4-AcPhQ+)2][PF6]4 (3). This was
prepared and purified in manner similar to 1 by using [4-AcPhQ+]Cl‚
2H2O (347 mg, 1 mmol) in place of [MeQ+]I. Acetone (100 mL) was
added to precipitate the crude chloride product. A dark-blue solid was
obtained: 121 mg, 46%; δH (CD3COCD3) 9.47 (4 H, d, J ) 6.9 Hz,
2C5H4N), 8.95 (4 H, d, J ) 6.8 Hz, 2C5H4N), 8.87 (4 H, d, J ) 6.8
Hz, 2C5H4N), 8.36 (4 H, d, J ) 8.8 Hz, 2C6H4), 8.15 (4 H, d, J ) 8.8
Hz, 2C6H4), 8.05 (4 H, d, J ) 6.8 Hz, 2C5H4N), 3.46 (6 H, s, 2NH3),
3.05 (6 H, s, 2NH3), 2.72 (6 H, s, 2Me). ν(CdO) 1686 cm-1. Anal.
Calcd (%) for C36H42F24N8O2P4Ru‚H2O: C, 32.81; H, 3.37; N, 8.50.
Found: C, 32.92; H, 2.87; N, 8.26.
Synthesis of cis-[RuII(NH3)4(2-PymQ+)2][PF6]4 (4). This was
prepared and purified in manner similar to 3 by using [2-PymQ+]Cl
(271 mg, 1 mmol) in place of [4-AcPhQ+]Cl‚2H2O. Further purification
was achieved by precipitation from acetone/aqueous NH4PF6 to afford
a dark-blue solid: 64 mg, 26%; δH (CD3COCD3) 10.26 (4 H, d, J )
7.3 Hz, 2C5H4N), 9.27 (4 H, d, J ) 4.8 Hz, 2C4H3N2), 9.01-8.92 (8
H, m, 4C5H4N), 8.11 (4 H, d, J ) 7.0 Hz, 2C5H4N), 8.04 (2 H, t, J )
4.9 Hz, 2C4H3N2), 3.53 (6 H, s, 2NH3), 3.09 (6 H, s, 2NH3). Anal.
Calcd (%) for C28H34F24N12P4Ru‚H2O: C, 27.17; H, 2.93; N, 13.58.
Found: C, 27.13; H, 2.64; N, 13.13.
Synthesis of trans-[RuII(NH3)4(MeQ+)2][PF6]4 (5). A mixture of
trans-[RuIICl(NH3)4(SO2)]Cl (100 mg, 0.329 mmol) and [MeQ+]I (196
mg, 0.657 mmol) was dissolved in water (5 mL) and heated at ca. 50
°C under argon for 30 min. Acetone (30 mL) was added to the brown
solution, and a brown precipitate was filtered off, washed with acetone,
and dried. This material (a mixture of chloride or iodide salts of trans-
[RuII(NH3)4(SO2)(MeQ+)]3+ and unreacted [MeQ+]I) was dissolved in
water (5 mL) and oxidized by the addition of a 1:1 mixture of 30%
aqueous H2O2/2M HCl (3 mL). After 5 min at room temperature, a
brown precipitate was filtered off, washed with water, and discarded.
Acetone (200 mL) was added to the filtrate, and a golden-yellow
precipitate was filtered off, washed with acetone, and dried to afford
trans-[RuIII(SO4)(NH3)4(MeQ+)]Cl2. This product was dissolved in
water (5 mL) and reduced over zinc amalgam (5 lumps) with argon
agitation for 15 min. The deep-blue solution was then filtered under
Synthesis of N′′,N′′′-Diphenyl-2,2′:4,4′′:4′,4′′′-quaterpyridinium
Hexafluorophosphate [Ph2Qpy2+][PF6]2. A mixture of Qpy‚0.5H2O
(100 mg, 0.31 mmol) and 2,4-dinitrochlorobenzene (607 mg, 3.00
mmol) in ethanol (25 mL) was heated at reflux for 48 h. Addition of
diethyl ether produced a gray precipitate which was filtered off, washed
with diethyl ether, and dried to yield crude N′′,N′′′-di(2,4-dinitrophenyl)-
2,2′:4,4′′:4′4′′′-quaterpyridinium chloride ([(2,4-DNPh)2Qpy2+]Cl2). This
material was dissolved in DMSO (5 mL), and aniline (0.3 mL, 3.30
mmol) was added. The resulting solution was heated at 70 °C for 3 h,
cooled, and then filtered. Acetone was added to the filtrate, and a white
precipitate was filtered off, washed with acetone, and dried. The solid
was dissolved in water, and aqueous NH4PF6 was added to give a white
precipitate. The product as filtered off, washed with water, and dried:
154 mg, 66%; δH (CD3COCD3) 9.61 (4 H, d, J ) 7.0 Hz, C5H4N),
9.15 (2 H, d, J ) 1.2 Hz, C5H3N), 9.11 (2 H, d, J ) 5.1 Hz, C5H3N),
9.00 (4 H, d, J ) 7.0 Hz, C5H4N), 8.25 (2 H, dd, J ) 1.9, 5.1 Hz,
C5H3N), 8.09-8.04 (4 H, m, Ph), 7.87-7.84 (6 H, m, Ph). Anal. Calcd
(%) for C32H24F12N4P2: C, 50.94; H, 3.21; N, 7.43. Found: C, 50.40;
H, 2.91; N, 7.24.
Synthesis of N′′,N′′′-Di(4-acetylphenyl)-2,2′:4,4′′:4′,4′′′-quater-
pyridinium Hexafluorophosphate [(4-AcPh)2Qpy2+][PF6]2. This was
prepared in a manner similar to [Ph2Qpy2+][PF6]2 by using 4-aceto-
phenone (406 mg, 3.00 mmol) instead of aniline, with heating at reflux
in ethanol (50 mL) for 50 h. The resulting solution was reduced to a
small volume, acetone was added, and a brown precipitate was filtered
off. The solid was dissolved in water, and aqueous NH4PF6 was added
to give a light-brown precipitate. The product was filtered off, washed
with water, and dried. Further purification was achieved by reprecipi-
tation from acetone/diethyl ether: 71 mg, 26%; δH (CD3COCD3) 9.67
(4 H, d, J ) 7.0 Hz, C5H4N), 9.16 (2 H, d, J ) 1.2 Hz, C5H3N), 9.12
(2 H, d, J ) 5.1 Hz, C5H3N), 9.04 (4 H, d, J ) 7.0 Hz, C5H4N), 8.41
(4 H, d, J ) 8.8 Hz, C6H4), 8.27-8.20 (6 H, m, C5H3N + C6H4), 2.74
(6 H, s, Me). ν(CdO) 1686 cm-1. Anal. Calcd (%) for C36H28F12N4O2P2‚
1.5H2O: C, 49.95; H, 3.61; N, 6.47. Found: C, 49.45; H, 3.27; N,
6.76.
Synthesis of N′′,N′′′-Di(2-pyrimidyl)-2,2′:4,4′′:4′,4′′′-quaterpyri-
dinium Hexafluorophosphate [(2-Pym)2Qpy2+][PF6]2. A mixture of
Qpy‚0.5H2O (100 mg, 0.31 mmol) and 2-chloropyrimidine (359 mg,
3.13 mmol) was heated quickly to 120 °C to produce a white solution
which turned green after 10 min. Ethanol (3 mL) was added and the
suspension heated at reflux for 4 h. Diethyl ether (50 mL) was added,
and a white precipitate was filtered off, washed with diethyl ether, and
dried. The solid was dissolved in water, and aqueous NH4PF6 was added
to give a white precipitate. The product as filtered off, washed with
water, and dried: 226 mg, 96%; δH (CD3COCD3) 10.44 (4 H, d, J )
7.2 Hz, C5H4N), 9.33 (4 H, d, J ) 4.8 Hz, C4H3N2), 9.21 (2 H, s,
C5H3N), 9.15 (2 H, d, J ) 5.1 Hz, C5H3N), 9.09 (4 H, d, J ) 7.3 Hz,
9
J. AM. CHEM. SOC. VOL. 127, NO. 13, 2005 4847

A R T I C L E S
Coe et al.
argon into a flask containing [MeQ+]I (196 mg, 0.657 mmol), and the
solution was stirred at room temperature in the dark under argon for
24 h. The addition of acetone (100 mL) to the deep-blue solution gave
a dark precipitate which was filtered off, washed with acetone, and
dried. This crude material was reprecipitated from water/aqueous
NH4PF6. Further purification was effected by precipitation from acetone/
aqueous NH4PF6 and then from acetone/diethyl ether to afford a dark-
purple solid: 188 mg, 52%; δH (CD3COCD3) 9.21 (4 H, d, J ) 6.9
Hz, 2C5H4N), 9.19 (4 H, d, J ) 7.2 Hz, 2C5H4N), 8.73 (4 H, d, J )
7.0 Hz, 2C5H4N), 8.16 (4 H, d, J ) 6.9 Hz, 2C5H4N), 4.65 (6 H, s,
2Me), 2.90 (12 H, s, 4NH3). Anal. Calcd (%) for C22H34F24N8P4Ru:
C, 24.21; H, 3.14; N, 10.27. Found: C, 24.34; H, 3.27; N, 10.11.
Synthesis of trans-[RuII(NH3)4(PhQ+)2][PF6]4 (6). This was pre-
pared and purified in manner similar to 5, by using [PhQ+]Cl‚2H2O
(200 mg, 0.656 mmol) instead of [MeQ+]I. Following the oxidation
step, no brown precipitate was produced and the trans-[RuIII(SO4)-
(NH3)4(PhQ+)]Cl2 intermediate was precipitated with acetone (100 mL).
A dark-blue solid was obtained: 159 mg, 39%; δH (CD3COCD3) 9.49
(4 H, d, J ) 7.1 Hz, 2C5H4N), 9.27 (4 H, d, J ) 6.8 Hz, 2C5H4N),
8.92 (4 H, d, J ) 7.1 Hz, 2C5H4N), 8.26 (4 H, d, J ) 6.8 Hz, 2C5H4N),
8.04-7.98 (4 H, m, 2Ph), 7.84-7.81 (6 H, m, 2Ph), 2.94 (12 H, s,
4NH3). Anal. Calcd (%) for C32H38F24N8P4Ru‚H2O: C, 31.16; H, 3.27;
N, 9.08. Found: C, 31.05; H, 3.25; N, 8.76.
Synthesis of trans-[RuII(NH3)4(4-AcPhQ+)2][PF6]4 (7). This was
prepared and purified in manner similar to 6, by using [4-AcPhQ+]Cl‚
2H2O (228 mg, 0.657 mmol) instead of [PhQ+]Cl‚2H2O. A dark-blue
solid was obtained: 123 mg, 29%; δH (CD3COCD3) 9.54 (4 H, d, J )
7.0 Hz, 2C5H4N), 9.27 (4 H, d, J ) 6.7 Hz, 2C5H4N), 8.94 (4 H, d, J
) 7.0 Hz, 2C5H4N), 8.38 (4 H, d, J ) 8.8 Hz, 2C6H4), 8.27 (4 H, d, J
) 6.7 Hz, 2C5H4N), 8.17 (4 H, d, J ) 8.7 Hz, 2C6H4), 2.95 (12 H, s,
4NH3), 2.86 (6 H, s, 2Me). Anal. Calcd (%) for C36H42F24N8O2P4Ru:
C, 33.27; H, 3.26; N, 8.62. Found: C, 33.27; H, 3.36; N, 8.25.
Synthesis of trans-[RuII(NH3)4(2-PymQ+)2][PF6]4 (8). This was
prepared and purified in manner similar to 6, by using [2-PymQ+]Cl
(178 mg, 0.658 mmol) instead of [PhQ+]Cl‚2H2O. A dark-blue solid
was obtained: 176 mg, 43%; δH (CD3COCD3) 10.32 (4 H, d, J ) 7.3
Hz, 2C5H4N), 9.32 (4 H, d, J ) 6.9 Hz, 2C5H4N), 9.29 (4 H, d, J )
4.9 Hz, 2C4H3N2), 9.00 (4 H, d, J ) 7.3 Hz, 2C5H4N), 8.30 (4 H, d, J
) 6.9 Hz, 2C5H4N), 8.04 (2 H, t, J ) 4.9 Hz, 2C4H3N2), 2.97 (12 H,
s, 4NH3). Anal. Calcd (%) for C28H34F24N12P4Ru‚2H2O: C, 26.78; H,
3.05; N, 13.39. Found: C, 26.82; H, 2.79; N, 13.27.
Synthesis of trans-[RuII(NH3)4(MeQ+)(PhQ+)][PF6]4 (9). This was
prepared and purified in manner similar to 5, with the second portion
of [MeQ+]I replaced by [PhQ+]Cl‚2H2O (200 mg, 0.656 mmol). A
dark-purple solid was obtained: 130 mg, 34%; δH (CD3COCD3) 9.49
(2 H, d, J ) 7.1 Hz, C5H4N), 9.27-9.18 (6 H, m, 3C5H4N), 8.91 (2 H,
d, J ) 7.2 Hz, C5H4N), 8.74 (2 H, d, J ) 7.0 Hz, C5H4N), 8.24 (2 H,
d, J ) 6.9 Hz, C5H4N), 8.18 (2 H, d, J ) 6.8 Hz, C5H4N), 8.04-7.99
(2 H, m, Ph), 7.84-7.81 (3 H, m, Ph), 4.66 (3 H, s, Me), 2.92 (12 H,
s, 4NH3). Anal. Calcd (%) for C27H36F24N8P4Ru: C, 28.11; H, 3.15;
N, 9.71. Found: C, 28.09; H, 3.10; N, 9.50.
Synthesis of trans-[RuII(NH3)4(MeQ+)(4-AcPhQ+)][PF6]4 (10).
This was prepared and purified in manner similar to 5, with the second
portion of [MeQ+]I replaced by [4-AcPhQ+]Cl‚2H2O (228 mg, 0.657
mmol). A dark-purple solid was obtained: 124 mg, 32%; δH (CD3-
COCD3) 9.53 (2 H, d, J ) 7.2 Hz, C5H4N), 9.27-9.17 (6 H, m,
3C5H4N), 8.93 (2 H, d, J ) 7.1 Hz, C5H4N), 8.74 (2 H, d, J ) 7.0 Hz,
C5H4N), 8.38 (2 H, d, J ) 8.1 Hz, C6H4), 8.24 (2 H, d, J ) 7.0 Hz,
C5H4N), 8.18-8.14 (4 H, m, C5H4N + C6H4), 4.65 (3 H, s, Me), 2.92
(12 H, s, 4NH3), 2.73 (3 H, s, C(O)Me). Anal. Calcd (%) for
C29H38F24N8OP4Ru: C, 29.13; H, 3.20; N, 9.37. Found: C, 28.84; H,
3.32; N, 9.23.
10.31 (2 H, d, J ) 7.1 Hz, C5H4N), 9.31-9.18 (8 H, m, 3C5H4N +
C4H3N2), 9.00 (2 H, d, J ) 7.3 Hz, C5H4N), 8.74 (2 H, d, J ) 6.9 Hz,
C5H4N), 8.27 (2 H, d, J ) 7.0 Hz, C5H4N), 8.19 (2 H, d, J ) 6.8 Hz,
C5H4N), 8.05 (1 H, t, J ) 4.7 Hz, C4H3N2), 4.66 (3 H, s, Me), 2.94 (12
H, s, 4NH3). Anal. Calcd (%) for C25H34F24N10P4Ru: C, 25.99; H, 2.97;
N, 12.12. Found: C, 26.25; H, 2.79; N, 11.82.
Synthesis of [RuII(NH3)4(Me2Qpy2+)][PF6]4 (12). A mixture of
[RuII(NH3)5(H2O)][PF6]2 (200 mg, 0.405 mmol) and [Me2Qpy2+][PF6]2
(255 mg, 0.405 mmol) in acetone (30 mL) was heated at reflux for 6
h in the dark under argon, cooled to room temperature, and filtered to
remove unreacted ligand. The resulting filtrate was treated as outlined
in purification method 1 below (three oxidation/reduction cycles), with
further purification via precipitation from acetone/diethyl ether. A dark-
purple solid was obtained: 71 mg, 16%; δH (CD3COCD3) 9.67 (2 H,
d, J ) 6.2 Hz, C5H3N), 9.17-9.14 (6 H, m, C5H3N + C5H4N), 8.82 (4
H, d, J ) 6.9 Hz, C5H4N), 8.22 (2 H, dd, J ) 2.2, 6.3 Hz, C5H3N),
4.67 (6 H, s, 2Me), 3.86 (6 H, s, 2NH3), 2.49 (6 H, s, 2NH3). Anal.
Calcd (%) for C22H32F24N8P4Ru: C, 24.25; H, 2.96; N, 10.29. Found:
C, 23.82; H, 2.73; N, 10.04.
Synthesis of [RuII(NH3)4(Ph2Qpy2+)][PF6]4 (13). This was pre-
pared and purified in manner similar to 12 by using [Ph2Qpy2+][PF6]2
(305 mg, 0.404 mmol) instead of [Me2Qpy2+][PF6]2 and with only two
oxidation/reduction cycles. A dark-purple solid was obtained: 123 mg,
25%; δH (CD3COCD3) 9.72 (2 H, d, J ) 6.3 Hz, C5H3N), 9.44 (4 H,
d, J ) 7.1 Hz, C5H4N), 9.29 (2 H, d, J ) 1.8 Hz, C5H3N), 8.98 (4 H,
d, J ) 6.7 Hz, C5H4N), 8.33 (2 H, dd, J ) 1.9, 6.3 Hz, C5H3N), 8.03-
7.98 (4 H, m, Ph), 7.85-7.81 (6 H, m, Ph) 3.91 (6 H, s, 2NH3), 2.54
(6 H, s, 2NH3). Anal. Calcd (%) for C32H36F24N8P4Ru: C, 31.67; H,
2.99; N, 9.23. Found: C, 31.79; H, 2.97; N, 8.75.
Synthesis of [RuII(NH3)4{(4-AcPh)2Qpy2+}][PF6]4 (14). This was
prepared in manner similar to 12 by using [(4-AcPh)2Qpy2+][PF6]2‚
1.5H2O (351 mg, 0.406 mmol) instead of [Me2Qpy2+][PF6]2. The initial
filtrate was treated as outlined in purification method 2 below, to afford
a dark-purple solid: 224 mg, 41%; δH (CD3COCD3) 9.73 (2 H, d, J )
6.4 Hz, C5H3N), 9.50 (4 H, d, J ) 7.0 Hz, C5H4N), 9.30 (2 H, d, J )
1.6 Hz, C5H3N), 9.01 (4 H, d, J ) 7.1 Hz, C5H4N), 8.41-8.32 (6 H,
m, C5H3N + C6H4), 8.16 (4 H, d, J ) 8.8 Hz, C6H4), 3.93 (6 H, s,
2NH3), 2.74 (6 H, s, 2Me), 2.56 (6 H, s, 2NH3). ν(CdO) 1686 cm-1
.
Anal. Calcd (%) for C36H40F24N8O2P4Ru‚2H2O C, 32.42; H, 3.33; N,
8.40. Found: C, 32.10; H, 3.29; N, 8.43.
Synthesis of [RuII(NH3)4{(2-Pym)2Qpy2+}][PF6]4 (15). This was
prepared in manner similar to 12 by using [(2-Pym)2Qpy2+][PF6]2 (307
mg, 0.405 mmol) instead of [Me2Qpy2+][PF6]2. The initial filtrate was
treated as outlined in purification method 2 below, to afford a dark-
purple solid: 89 mg, 18%; δH (CD3COCD3) 10.27 (4 H, d, J ) 7.3
Hz, C5H4N), 9.76 (2 H, d, J ) 6.3 Hz, C5H3N), 8.31-9.28 (6 H, m,
C5H3N + C4H3N2), 9.13 (4 H, d, J ) 7.2 Hz, C5H4N), 8.38 (2 H, dd,
J ) 1.8, 6.5 Hz, C5H3N), 8.05 (2 H, t, J ) 4.9 Hz, C4H3N2), 3.97 (6
H, s, 2NH3), 2.63 (6 H, s, 2NH3). Anal. Calcd (%) for C28H32F24N12-
P4Ru‚0.5Et2O: C, 28.72; H, 2.97; N, 13.40. Found: C, 28.56; H, 2.82;
N, 13.14.
Purification Method 1. A solution of [NBun ]I (0.15 M) was added
4
dropwise to an acetone solution of the crude PF6- salt of the complex,
and a dark purple precipitate was filtered off, washed with acetone,
and dried. The resulting I- salt was dissolved in a minimum of water,
and a solution of 1:1 2 M HCl/30% aqueous H2O2 (25 mL) was added.
After 1 h at room temperature, a brown precipitate was filtered off,
and acetone (1.5 L) was added to the filtrate. After standing for 24 h,
the dark precipitate was filtered off, washed with acetone, and dried.
This crude RuIII salt was dissolved in water (10 mL) and reduced over
zinc amalgam (5 lumps) for 30 min with stirring. The zinc amalgam
was removed by filtration and aqueous NH4PF6 was added dropwise
to the filtrate to give a dark precipitate which was filtered off, washed
Synthesis of trans-[RuII(NH3)4(MeQ+)(2-PymQ+)][PF6]4 (11). This
was prepared and purified in manner similar to 5, with the second
portion of [MeQ+]I replaced by [2-PymQ+]Cl (178 mg, 0.658 mmol).
A dark-purple solid was obtained: 104 mg, 27%; δH (CD3COCD3)
1
with water, and dried. The H NMR spectrum of the product at this
stage was used as an indication of purity, and if necessary the oxidation/
reduction cycle was repeated.
9
4848 J. AM. CHEM. SOC. VOL. 127, NO. 13, 2005

Synthesis of Charged 2D Dipolar NLO Chromophores
A R T I C L E S
Purification Method 2. An acetone solution of the crude PF6- salt
of the complex was loaded onto a silica gel column (230-400 mesh,
5 cm × 58 cm, 1 L dry gel) and eluted with NH4PF6 in acetonitrile
(0.1 M). Fractions were collected and compared using thin-layer
chromatography. Similar fractions were combined, reduced to dryness,
and precipitated from acetone/aqueous NH4PF6. Further purification
was achieved by precipitation from acetone/diethyl ether.
fit the Stark spectra with Liptay’s equation.8 The dipole moment change
∆µ12 associated with each of the optical transitions considered in the
fit was then calculated from the coefficient of the second derivative
component. For all of the complexes two or three Gaussian functions
were necessary to fit the absorption spectrum. In each case, only one
or two of the Gaussian curves contributed the majority of the Stark
signal, so the data yielded by the insignificant components have been
neglected.
X-ray Crystallography. Crystals of the pro-ligand salt [Ph2Qpy2+]-
[PF6]2‚Me2CO were obtained by slow evaporation of an acetone
solution. The pale-yellow plate crystal chosen for diffraction study had
approximate dimensions of 0.50 mm × 0.30 mm × 0.06 mm. Data
were collected on a Nonius Kappa CCD area-detector diffractometer
controlled by the Collect software package.13 The data were processed
by Denzo14 and corrected for absorption by using the empirical method
employed in Sortav15 from within the MaXus suite of programs.16 The
structure was solved by direct methods and refined by full-matrix least-
The Liptay equation is:8
Bx
Cx
30h2
∂(ꢀ(ν)/ν)
∂ν
∂2(ꢀ(ν)/ν)
∂ν2
∆ꢀ(ν)/ν ) Axꢀ(ν)/ν +
+
F2int (1)
[
]
15h
where ν is the frequency of the light in Hz and the internal electric
field is related to the applied external field by Fint ) fintFext. ∆µ12 is
related to the coefficient of the second derivative component by
2
squares on all F0 data using SHELXS-9717 and SHELXL-97.18 All
Cø ) 3∆µ12(2 - cos2 ø) + 3(m∆µ12)2(3 cos2 ø - 1)
(2)
non-hydrogen atoms were refined anisotropically with hydrogen atoms
included in idealized positions with thermal parameters riding on those
of the parent atom. Crystallographic data and refinement details are
presented in Table 3.
where m∆µ12 is the dipole moment change along the direction of the
transition dipole moment µ12. Butyronitrile was used as the glassing
medium, for which the local field correction fint is estimated as 1.33.5e,24
A two-state analysis of the MLCT transitions gives
Hyper-Rayleigh Scattering. General details of the hyper-Rayleigh
scattering (HRS) experiment have been discussed elsewhere,7 and the
experimental procedure and data analysis protocol used were as
previously described.19,20 Measurements were carried out in acetone
∆µ2ab ) ∆µ212 + 4µ212
(3)
or acetonitrile, with crystal violet as an external reference (âxxx,800
)
where ∆µab is the dipole moment change between the diabatic states
and ∆µ12 is the observed (adiabatic) dipole moment change. The value
of µ12 can be determined from the oscillator strength fos of the transition
by
500 × 10-30 esu in acetonitrile; from the value of 340 × 10-30 esu in
methanol,20 corrected for local field factors at optical frequencies). All
measurements were performed by using the 800 nm fundamental of a
regenerative mode-locked Ti3+:sapphire laser (Spectra Physics, model
Tsunami, 100 fs pulses, 1 W, 80 MHz). Dilute solutions (10-5-10-6
|µ12| ) [fos/(1.08 × 10-5Emax)]1/2
(4)
2
M) were used to ensure a linear dependence of I2ω/Iω on solute
concentration, precluding the need for Lambert-Beer correction factors.
Samples were filtered (Millipore, 0.45 µm), and none showed any
fluorescence. A relative error of (15% is estimated for the â values
determined via these measurements. HRS depolarization ratios21 were
determined at 800 nm in acetonitrile according to a published
methodology.22,23
where Emax is the energy of the MLCT maximum (in wavenumbers).
The degree of delocalization cb2 and electronic coupling matrix element
Hab for the diabatic states are given by
1/2
∆µ212
1
c2b ) 1 -
(5)
(6)
2
2
(
)
2
[
]
∆µ12 + 4µ12
Stark Spectroscopy. The Stark apparatus, experimental methods
and data collection procedure were as previously reported,5e,24 except
that a Xe arc lamp was used as the light source in the place of a W
filament bulb. The Stark spectrum for each compound was measured
a minimum of three times using different field strengths, and the signal
was always found to be quadratic in the applied field. To fit the Stark
data, the absorption (ꢀ/ν vs ν) spectrum was modeled with a sum of
Gaussian curves that reproduce the data and separate the peaks. The
first and second derivatives of the Gaussian curves were then used to
Emax(µ12)
∆µab
|Hab| )
|
|
If the hyperpolarizability tensor â0 has only nonzero elements along
the CT direction, then this quantity is given by
3∆µ12(µ12)2
â0 )
(7)
2
(Emax
)
(13) Hooft, R. Collect, Data collection software; Nonius BV: Delft, The
Netherlands, 1998.
(14) Otwinowski, Z.; Minor, W. Methods Enzymol. 1997, 276, 307-326.
(15) (a) Blessing, R. H. Acta Crystallogr., Sect. A 1995, 51, 33-37. (b) Blessing,
R. H. J. Appl. Crystallogr. 1997, 30, 421-426.
(16) Mackay, S.; Gilmore, C. J.; Edwards, C.; Tremayne, M.; Stewart, N.;
Shankland, K. MAXUS, a computer program for the solution and refinement
of crystal structures from diffraction data; University of Glasgow: Glasgow,
U.K., Nonius BV, Delft, The Netherlands and MacScience Co. Ltd.,
Yokohama, Japan, 1998.
A relative error of (20% is estimated for the â0 values derived from
the Stark data and using eq 7.
Computational Procedures. All theoretical calculations were
performed by using the Gaussian 0325 program. The molecular
geometries were optimized assuming C2V symmetry (and also with no
restrictions for the cation in 1) by using the hybrid functional B3P8626
and the LanL2DZ27 basis set. It should be noted that quantum chemistry
(17) Sheldrick, G. M. Acta Crystallogr., Sect. A 1990, 46, 467-473.
(18) Sheldrick, G. M. SHELXL 97, Program for crystal structure refinement;
University of Go¨ttingen: Go¨ttingen, Germany, 1997.
(19) Clays, K.; Olbrechts, G.; Munters, T.; Persoons, A.; Kim O.-K.; Choi, L.-
S. Chem. Phys. Lett. 1998, 293, 337-342.
(25) Frisch, M. J. et al. Gaussian 03, revision B.05; Gaussian, Inc.: Pittsburgh,
PA, 2003.
(26) The B3P86 Functional consists of Becke’s three parameter hybrid functional
(Becke, A. D. J. Chem. Phys. 1993, 98, 5648-5652) with the nonlocal
correlation provided by the Perdew 86 expression: Perdew, J. P. Phys.
ReV. B 1986, 33, 8822-8824.
(27) D95 on first row: Dunning, T. H.; Hay P. J. In Modern Theoretical
Chemistry; Schaefer, H. F. III, Ed.; Plenum: New York, 1976, Vol. 3, p 1.
Los Alamos ECP plus DZ on Na-Bi: (a) Hay, P. J.; Wadt, W. R. J. Chem.
Phys. 1985, 82, 270-283. (b) Wadt, W. R.; Hay, P. J. J. Chem. Phys.
1985, 82, 284-298. (c) Hay, P. J.; Wadt, W. R. J. Chem. Phys. 1985, 82,
299-310.
(20) Olbrechts, G.; Strobbe, R.; Clays K.; Persoons, A. ReV. Sci. Instrum. 1998,
69, 2233-2241.
(21) Heesink, G. J. T.; Ruiter, A. G. T.; van Hulst, N. F.; Bo¨lger, B. Phys. ReV.
Lett. 1993, 71, 999-1002.
(22) Hendrickx, E.; Boutton, C.; Clays, K.; Persoons, A.; van Es, S.; Biemans,
T.; Meijer, B. Chem. Phys. Lett. 1997, 270, 241-244.
(23) Boutton, C.; Clays, K.; Persoons, A.; Wada, T.; Sasabe, H. Chem. Phys.
Lett. 1998, 286, 101-106.
(24) Shin, Y. K.; Brunschwig, B. S.; Creutz, C.; Sutin, N. J. Phys. Chem. 1996,
100, 8157-8169.
9
J. AM. CHEM. SOC. VOL. 127, NO. 13, 2005 4849

A R T I C L E S
Coe et al.
The new complex salts 1-15 (Figure 1) were prepared by
using slight modifications of established coordination chemistry
procedures based on the precursors cis-[RuIIICl2(NH3)4]Cl,10
trans-[RuIICl(NH3)4(SO2)]Cl,11 or [RuII(NH3)5(H2O)][PF6]2.11
Five equivalents of the pro-ligand salts were used in the
preparations of 1-4, the unreacted excesses being readily
separated by precipitation of the crude products from aqueous
solutions with acetone. In contrast, only one equivalent of the
required pro-ligand salt was used in the preparations of 12-15
because these reactions are relatively inefficient and substantial
amounts of unreacted pro-ligand salt remain. Purification of 12-
15, largely involving removal of the unreacted pro-ligand salts,
also proved rather more demanding than was the case for 1-11.
Two methods were adopted, one involving cycles of oxidation
and reduction in which the remaining pro-ligand is converted
into a relatively insoluble N-oxide, and the other using column
chromatography on silica gel. Although both methods were
attempted in each case, those described above gave the highest
purified yields.
Electronic Spectroscopy Studies. Electronic absorption
spectra for the new complex salts 1-15 were recorded in
acetonitrile, and the results are presented in Table 1. As
expected, all of the complexes show intense d(RuII) f π*(L)
(L ) pyridinium ligand) metal-to-ligand charge-transfer (MLCT)
bands in the visible region, together with UV bands due to
intraligand π f π* excitations.
Salts 1-4 each show two very close MLCT bands (Table 1,
Figure 2), as found in other related complexes such as
cis-[RuII(NH3)4(L1)2]2+ [L1 ) pyridine (py), isonicotinamide
(isn)].30 A simple MO analysis has attributed these bands to
orthogonally polarized excitations from primarily the HOMO
(a 4d Ru orbital in the plane defined by the Ru-N(pyridyl)
bonds) into two L-based LUMOs which are combinations of
π* MOs of the individual pyridyl ligands.30b Note, however,
that in these previous studies the x and y axes were taken as
lying along the Ru-N(pyridyl) bonds, with the z axis along
the two mutually trans Ru-NH3 bonds.30b Here we use a
different axis labeling convention, as defined by the computa-
tional procedure and shown in Figure 1. For 1-4, the lower-
energy absorption is more intense than the higher-energy band
in all cases, and the energy gap between these bands is constant
at ca. 0.3 eV. As observed previously in related 1D complexes,5d
the Emax values for both MLCT bands decrease as L changes in
the order MeQ+ > PhQ+ > 4-AcPhQ+ > 2-PymQ+, due to the
steadily increasing acceptor strength of the pyridinium groups.
Simple MO theory also predicts that symmetric trans complexes
such as those found in 5-8 should exhibit two MLCT
transitions, but the higher-energy one is parity forbidden.30b This
is consistent with the observed spectra, which again show the
trend of decreasing Emax as the pyridinium groups become more
electron deficient. The asymmetric trans complexes in 9-11
would be expected to show two close MLCT bands, those at
higher energy corresponding to d(RuII) f π*(MeQ+) processes,
but these bands are unsurprisingly not resolved in the spectra
obtained.
Figure 1. Chemical structures of the complex salts investigated, showing
the axis convention used in the theoretical calculations.
programs use the so-called “standard orientation” that for a C2V group
assigns the z axis to the C2 axis of symmetry and the molecule is placed
on the zy plane with the x axis orthogonal to the molecule (as shown
in Figure 1). The same model chemistry was used for properties
calculations. Electronic transitions were calculated by means of the
TD-DFT method, and the excited-state dipole moments were calculated
by using the one particle RhoCI density. The default Gaussian 03
parameters were used in every case. Molecular orbital contours were
plotted by using Molekel 4.3.28
Results and Discussion
Synthetic Studies. The compound 2,2′:4,4′′:4′,4′′′-quater-
pyridyl (Qpy) was obtained by a modification of a literature
method which involves heating at 125 °C for 72 h with 10%
Pd on charcoal.12 We have found that heating a 4,4′-bipyridyl
(4,4′-bpy) melt at ca. 250 °C for 48 h with the same coupling
reagent results in an approximately 30% increase in the yield
of Qpy when compared with the published procedure,12 and
some higher-order oligopyridine byproducts are also generated.
Qpy [also known as 4-(4-pyridyl)-2,2′:4′,4′′-terpyridyl] has also
been prepared in 26% yield via reaction of 4,4′-bpy with lithium
diisopropylamide as the coupling reagent in diethyl ether at -78
°C, with subsequent warming to room temperature.29 The new
Qpy-based pro-ligand salts, [R2Qpy2+][PF6]2, (R ) Ph, 4-AcPh
or 2-Pym) were synthesized by using methods which we have
previously developed to prepare analogous 4,4′-bpy-based
ligands (Scheme 1).5c,d The compound [(2,4-DNPh)2Qpy2+]Cl2
(2,4-DNPh ) 2,4-dinitrophenyl) was prepared as an intermedi-
ate, but not purified or characterized.
(30) (a) Ford P. C.; Sutton, C. Inorg. Chem. 1969, 8, 1544-1546. (b) Zwickel
A. M.; Creutz, C. Inorg. Chem. 1971, 10, 2395-2399. (c) Pavanin, L. A.;
Giesbrecht E.; Tfouni, E. Inorg. Chem. 1985, 24, 4444-4446. (d) Pavanin,
L. A.; Novais da Rocha, Z.; Giesbrecht E.; Tfouni, E. Inorg. Chem. 1991,
30, 2185-2190.
(28) Portmann, S.; Lu¨thi, H. P. Chimia 2000, 54, 766-770.
(29) Hayes, M. A.; Meckel, C.; Schatz E.; Ward, M. D. J. Chem. Soc., Dalton
Trans. 1992, 703-708.
9
4850 J. AM. CHEM. SOC. VOL. 127, NO. 13, 2005

Synthesis of Charged 2D Dipolar NLO Chromophores
A R T I C L E S
Scheme 1. Syntheses of the Qpy-Based Pro-Ligands
Table 1. UV-Visible Data for Complex Salts 1-15 in Acetonitrilea
1
salt
λ
max, nm (
ꢀ
, M-1 cm-
)
E
max (eV)
assignment
1 cis-[RuII(NH3)4(MeQ+)2][PF6]4
570 (17 500)
502 (15 100)
262 (33 300)
606 (20 400)
528 (16 500)
284 (31 200)
620 (22 600)
536 (17 500)
288 (41 800)
644 (21 500)
558 (17 200)
286 (44 600)
595 (27 700)
259 (35 200)
628 (29 400)
277 (31 700)
641 (31 100)
283 (43 600)
673 (33 700)
282 (53 500)
611 (29 600)
268 (31 700)
620 (30 700)
269 (35 800)
639 (28 400)
272 (38 900)
510 (13 000)
320 (12 300)
260 (35 300)
540 (25 000)
264 (35 500)
548 (26 900)
288 (42 700)
568 (24 800)
286 (44 100)
2.18
2.47
4.73
2.05
2.35
4.37
2.00
2.31
4.31
1.93
2.22
4.34
2.08
4.79
1.97
4.48
1.93
4.38
1.84
4.40
2.03
4.63
2.00
4.61
1.94
4.56
2.43
3.88
4.77
2.30
4.70
2.26
4.31
2.18
4.34
d f π*
d f π*
π f π*
d f π*
d f π*
π f π*
d f π*
d f π*
π f π*
d f π*
d f π*
π f π*
d f π*
π f π*
d f π*
π f π*
d f π*
π f π*
d f π*
π f π*
d f π*
π f π*
d f π*
π f π*
d f π*
π f π*
d f π*
π f π*
π f π*
d f π*
π f π*
d f π*
π f π*
d f π*
π f π*
2 cis-[RuII(NH3)4(PhQ+)2][PF6]4
3 cis-[RuII(NH3)4(4-AcPhQ+)2][PF6]4
4 cis-[RuII(NH3)4(2-PymQ+)2][PF6]4
5 trans-[RuII(NH3)4(MeQ+)2][PF6]4
6 trans-[RuII(NH3)4(PhQ+)2][PF6]4
7 trans-[RuII(NH3)4(4-AcPhQ+)2][PF6]4
8 trans-[RuII(NH3)4(2-PymQ+)2][PF6]4
9 trans-[RuII(NH3)4(MeQ+)(PhQ+)][PF6]4
10 trans-[RuII(NH3)4(MeQ+)(4-AcPhQ+)][PF6]4
11 trans-[RuII(NH3)4(MeQ+)(2-PymQ+)][PF6]4
12 [RuII(NH3)4(Me2Qpy2+)][PF6]4
13 [RuII(NH3)4(Ph2Qpy2+)][PF6]4
14 [RuII(NH3)4{(4-AcPh)2Qpy2+}][PF6]4
15 [RuII(NH3)4{(2-Pym)2Qpy2+}][PF6]4
a Solutions ca. 3-8 × 10-5 M.
The complex cis-[RuII(NH3)4(2,2′-bpy)]2+ (bpy ) bipyridyl)
was first reported in 1974,31 and quite a number of studies into
its electronic structure have since been carried out.32 This species
has two MLCT bands, with λmax values of 526 and 365 nm (for
the PF6- salt in acetonitrile),32c and INDO/S calculations predict
a number of closely spaced transitions, polarized in the bpy
9
J. AM. CHEM. SOC. VOL. 127, NO. 13, 2005 4851

A R T I C L E S
Coe et al.
Table 2. Electrochemical Data for Complex Salts 1-15 and [Me2Qpy2+][PF6]2 in Acetonitrile
E1 2, V vs Ag
−
AgCl (
∆
Ep, mV)a
E (V vs Ag
−
AgCl)a,b
L0/-/L-/
/
/
II
+/
0
2
/
2-
+/+
2-
salt
RuIII
L+/0/L2
L0/-/L-/ -/L0
L+/0/L2
1 cis-[RuII(NH3)4(MeQ+)2][PF6]4
0.79 (85)
0.79 (70)
-0.81 (110)
-0.66 (110)
-1.42 (65)
-1.55 (70)
-1.26 (70)
-1.38 (65)
-0.79
-0.64
-1.40
-1.54
-1.23
-1.36
2 cis-[RuII(NH3)4(PhQ+)2][PF6]4
3 cis-[RuII(NH3)4(4-AcPhQ+)2][PF6]4
4 cis-[RuII(NH3)4(2-PymQ+)2][PF6]4
5 trans-[RuII(NH3)4(MeQ+)2][PF6]4
0.80 (70)
0.82 (100)
0.74 (100)
-0.55 (120)
-0.42d
-0.83 (130)
-1.09 (290)c
-1.38d
-1.45d
-1.21d
-1.25d
-1.36d
-1.09d
6 trans-[RuII(NH3)4(PhQ+)2][PF6]4
0.76 (90)
-0.66 (100)
7 trans-[RuII(NH3)4(4-AcPhQ+)2][PF6]4
8 trans-[RuII(NH3)4(2-PymQ+)2][PF6]4
9 trans-[RuII(NH3)4(MeQ+)(PhQ+)][PF6]4
0.77 (90)
0.76 (80)
0.72 (100)
-0.56 (100)
-0.44d
-0.66 (70)
-0.86 (70)
-0.58 (80)
-0.84 (80)
-0.41 (90)
-0.85 (70)
-0.82 (160)
-1.27 (80)
-1.11 (150)
-1.16 (100)
10 trans-[RuII(NH3)4(MeQ+)(4-AcPhQ+)][PF6]4
11 trans-[RuII(NH3)4(MeQ+)(2-PymQ+)][PF6]4
12 [RuII(NH3)4(Me2Qpy2+)][PF6]4
0.74 (90)
0.74 (90)
0.79 (85)
0.81 (90)
-1.39 (80)
-1.61 (95)
-1.21 (75)
-1.40 (70)
-0.74
-0.84
-0.55
-0.65
-1.37
-1.57
-1.19
13 [RuII(NH3)4(Ph2Qpy2+)][PF6]4
-0.64 (130)
-1.38e
14 [RuII(NH3)4{(4-AcPh)2Qpy2+}][PF6]4
15 [RuII(NH3)4{(2-Pym)2Qpy2+}][PF6]4
[Me2Qpy2+][PF6]2
0.82 (70)
0.83 (80)
-0.56d
-0.37d
-0.91 (110)
-1.57 (85)
a Measured in solutions ca. 10-3 M in analyte and 0.1 M in [NBun ]PF6 at a platinum bead/disk working electrode with a scan rate of 200 mV s-1
.
4
Ferrocene internal reference E1/2 ) 0.44 V, ∆Ep ) 70 mV. b Potential increment ) 4 mV, amplitude ) 50 mV, pulse width ) 0.05 s. c Shoulders evident.
d Epc for an irreversible reduction process. e Potential increment ) 1 mV.
electronic structures, and hence absorption spectra, of these cis-
{RuII(NH3)4}2+ complexes. However, the overall decreased
extent of visible absorption (observed for example on moving
from 2 to 13) contrasts with the behavior of the simpler
complexes cis-[RuII(NH3)4(py)2]2+ and cis-[RuII(NH3)4(2,2′-
bpy)]2+. The former is yellow, with a lowest-energy MLCT band
at λmax ) 409 nm (for the PF6- salt in acetonitrile),33 while the
latter displays a dark red-purple color.
Electrochemical Studies. The new complex salts 1-15 were
studied by cyclic voltammetry in acetonitrile, and the results
are presented in Table 2. All of the complexes show reversible
or quasi-reversible RuIII/II oxidation waves, together with
reversible, quasi-reversible, and/or irreversible ligand-based
reduction processes. It is worth noting that such species
exhibiting several reversible redox processes offer intriguing
Figure 2. UV-visible absorption spectra of 2 (red) and 13 (blue) at 295
K in acetonitrile.
possibilities for multistate switching of NLO responses, exploit-
plane, as well as weak perpendicularly polarized nonbonding d
ing not only RuIII/II oxidations6 but also ligand-based reductions.
f π* transitions.32b The related complexes in 12-15 exhibit
Representative cyclic voltammograms of the salts 1 and 12 are
only broad MLCT bands in the visible region which result from
two or more transitions. In each case the absorption profile is
asymmetrical, with a tail on the lower energy side indicating
the presence of a less intense low energy transition. The absence
shown in Figure 3.
Literature RuIII/II E1/2 values for the salts cis-[RuII(NH3)4(py)2]-
(PF6)233 and cis-[RuII(NH3)4(2,2′-bpy)](PF6)211 are respectively
0.54 and 0.53 V vs Ag-AgCl, ca. 200-300 mV lower than
the corresponding potentials observed for 1-15. It is therefore
apparent that the electron-withdrawing influence of the two
pyridinium substituents in the new complexes causes substantial
stabilization of the Ru-based HOMOs when compared with the
model compounds. The accompanying red-shifting of the MLCT
bands in 1-15 (see above) is therefore indicative of very
pronounced stabilization of the ligand-based LUMOs which
more than offsets that of the HOMOs. As noted previously,5c,d
the RuIII/II E1/2 values of complexes of N-R-4,4′-bipyridinium
of MLCT absorptions in the UV is consistent with stabilization
of the ligand π* orbitals when compared with those of 2,2′-
bpy in cis-[RuII(NH3)4(2,2′-bpy)]2+. Representative UV-visible
spectra of 2 and 13 are shown in Figure 2. As expected, the
formation of a planar chelating system by linking the two
coordinated 4,4′-bipyridinium units dramatically alters the
(31) Alvarez, V. E.; Allen, R. J.; Matsubara T.; Ford, P. C. J. Am. Chem. Soc.
1974, 96, 7686-7692.
(32) Selected examples: (a) Mines, G. A.; Roberts J. A.; Hupp, J. T. Inorg.
Chem. 1992, 31, 125-128. (b) Streiff, J. H.; Edwards W. D.; McHale, J.
L. Chem. Phys. Lett. 1999, 312, 369-375. (c) Endicott, J. F.; Schlegel, H.
B.; Uddin Md. J.; Seniveratne, D. S. Coord. Chem. ReV. 2002, 229, 95-
106.
(33) Ando, I.; Ishimura, D.; Ujimoto K.; Kurihara, H. Inorg. Chem. 1996, 35,
3504-3508.
9
4852 J. AM. CHEM. SOC. VOL. 127, NO. 13, 2005

Synthesis of Charged 2D Dipolar NLO Chromophores
A R T I C L E S
Table 3. Crystallographic Data and Refinement Details for Salt
[Ph2Qpy2+][PF6]2‚Me2CO
formula
M
C35H30F12N4OP2
812.57
crystal system
space group
monoclinic
C2/c
a/Å
b/Å
c/Å
24.1853(7)
10.3325(3)
14.4450(2)
110.375(2)
3383.88(15)
4
â/deg
U/Å3
Z
T/K
120(2)
0.233
19666
3742 (0.0795)
0.0405, 0.1044
0.0528, 0.1124
µ/mm-1
reflections collected
independent reflections (Rint
final R1, wR2 [I > 2σ(I)]a
(all data)
)
a The structure was refined on Fo using all data; the value of R1 is
given for comparison with older refinements based on Fo with a typical
threshold of Fo > 4σ(Fo).
2
220, and 190 mV for 1, 2, 12, and 13, respectively. These
observations show that the extent of electronic coupling between
the pyridyl ligands increases upon reduction, and the Qpy-based
systems again show greater coupling compared with their
nonchelated analogues. In contrast, the pro-ligand salt [Me2Qpy2+]-
[PF6]2 exhibits only one broad reversible Me2Qpy2+/0 wave and
one Me2Qpy0/2- wave, showing that metal coordination is
required to allow significant communication between the
pyridinium groups. This is entirely logical because the 2,2′-
bpy unit in Qpy derivatives would be expected to adopt a
transoid geometry unless chelated, as confirmed by the crystal
structure of [Ph2Qpy2+][PF6]2‚Me2CO (see Figure 4 and Table
3). All of the trans complexes show only irreversible or quasi-
irreversible L0/- processes, with the exception of 9.
Figure 3. Cyclic (red) and differential pulse (blue) voltammograms of 1
(a) and 12 (b) in acetonitrile.
ligands are relatively insensitive to changes in R. However,
within each of the three series 1-4, 5-8, and 12-15, the RuIII/II
E
1/2 increases slightly, as the HOMO becomes stabilized by the
Crystallographic Study. A single-crystal X-ray structure was
obtained for the pro-ligand salt [Ph2Qpy2+][PF6]2‚Me2CO. A
representation of the molecular structure of this salt is shown
in Figure 4, and selected interatomic distances and angles are
presented in Table 4. To our knowledge, this is the first reported
structure of a compound containing a Qpy unit.
As expected, the 2,2′-bpy unit in [Ph2Qpy2+][PF6]2‚Me2CO
adopts a planar, transoid geometry to minimize steric repulsions,
with a center of symmetry in the middle of the C-C bond
between the pyridyl rings. The structure is otherwise heavily
twisted, with dihedral angles of 34.65° between the two
4-pyridyl rings and 49.35° between the pyridyl and attached
phenyl rings.
increasing electron-accepting ability of the pyridinium substitu-
ent. The RuIII/II potentials for the trans complex salts are a little
lower (by 30-80 mV) than those of their cis analogues, but
the latter and their Qpy-based counterparts have essentially
identical RuIII/II E1/2 values, showing that chelate formation does
not significantly affect the HOMO energy.
The complexes in salts 4, 8, 14, and 15 exhibit only one
irreversible reduction process, but 1-3 show especially interest-
ing ligand-based electrochemistry, with broad, quasi-reversible
(presumably) two-electron L+/0 reductions. The E1/2 values of
these waves increase as R changes in the order Me < Ph <
4-AcPh, reflecting the increasing acceptor strength of the 4,4′-
bipyridinium ligands which is mirrored in the red-shifting of
the MLCT bands. Similar first-ligand reduction waves are also
seen in the salts 5-7 and 12 and 13, but in the latter two
compounds the two-electron L2+/0 waves show slight shoulders,
indicating two separate processes occurring at similar potentials.
Differential pulse voltammetry experiments resolve these pro-
cesses for 12 and 13, but not for 1 and 2 (Table 2, Figure 3),
showing that with the Qpy-based ligands there is a detectable
degree of electronic communication between the pyridinium
moieties. The absence of a direct linkage between the coordi-
nated pyridyl rings logically results in a smaller extent of
coupling in 1 and 2.
Hyper-Rayleigh Scattering Studies. The â values of 1-4
and 12-15 were measured in acetone solutions by using the
HRS technique7,19,20 with an 800 nm Ti3+:sapphire laser, and
the results are collected in Table 5. This laser fundamental was
chosen because none of the complexes absorb strongly at either
800 nm or at the second harmonic (SH) of 400 nm, their MLCT
bands falling conveniently between these wavelengths (Figure
2). Other available laser fundamentals, such as 1064 or 1300
nm, would be unsuitable due to very strong absorption at their
SH wavelengths. Using an 800 nm laser also has the advantage
of avoiding any possibility of fluorescence contributions at the
SH, which can cause problems in HRS studies.20 1-4 were also
studied in acetonitrile. Because the symmetric trans complexes
would be expected to possess â values of zero and their
Each of the cis complexes with N-Me and N-Ph substituents
also shows two separate, reversible L0/- or L-/2- reduction
processes with differences between the E1/2 values of 130, 120,
9
J. AM. CHEM. SOC. VOL. 127, NO. 13, 2005 4853

A R T I C L E S
Coe et al.
-
Figure 4. Representation of the molecular structure of the salt [Ph2Qpy2+][PF6]2‚Me2CO (only one PF6 is shown because the asymmetric unit contains
-
only half a PF6 and the drawing software therefore automatically generates only one ion).
Table 4. Selected Interatomic Distances (Å) and Angles (deg) for
Salt [Ph2Qpy2+][PF6]2‚Me2COa
Table 6. HRS Depolarization Ratios and â Components for
Complex Salts 1-4
b
c
âzzz
âzyy
C1-N1
C1-C2
C1-C1ii
C2-C3
C3-C4
C3-C6
C4-C5
C5-N1
1.343(2)
1.395(2)
1.494(3)
1.391(2)
1.388(2)
1.487(2)
1.389(2)
1.334(2)
C6-C7
C6-C10
C7-C8
C8-N2
C9-N2
C9-C10
C11-N2
1.391(2)
1.397(2)
1.374(2)
1.353(2)
1.347(2)
1.379(2)
1.454(2)
salt
Fa
k
(10-30 esu) (10-30 esu)
1 cis-[RuII(NH3)4(MeQ+)2][PF6]4
2 cis-[RuII(NH3)4(PhQ+)2][PF6]4
2.32 ( 0.05 -0.36
2.07 ( 0.06 -0.43
142
126
141
143
-51
-54
-61
-60
3 cis-[RuII(NH3)4(4-AcPhQ+)2][PF6]4 2.09 ( 0.05 -0.43
4 cis-[RuII(NH3)4(2-PymQ+)2][PF6]4 2.11 ( 0.06 -0.42
a Depolarization ratio determined in acetonitrile by using a 800 nm Ti3+
:
sapphire laser fundamental. b,c Hyperpolarizability tensor components de-
rived from the HRS intensity and depolarization ratio measurements by
using eqs 8-10.
N1-C1-C2
N1-C1-C1ii
N1-C5-C4
N2-C8-C7
N2-C9-C10
C16-C11-N2
122.95(14)
116.19(17)
123.66(15)
120.70(15)
120.58(15)
117.97(14)
C12-C11-N2
C5-N1-C1
C9-N2-C8
C9-N2-C11
C8-N2-C11
119.47(15)
117.34(14)
120.51(14)
120.58(13)
118.67(13)
a small degree of enhancement of â800 for the complexes of
N-aryl ligands when compared with their N-Me analogues, the
data for 1-4 in acetonitrile do not reinforce this putative trend.
Although the â800 data given in Table 5 are based on the
assumption of a single â component, the hyperpolarizabilities
of 1-4 and 12-15 are expected to display substantial 2D
character. For a chromophore with C2V symmetry, there are five
nonzero components of the â tensor, âzzz, âzyy, âzxx, âyyz, and
a Symmetry transformation used to generate equivalent atoms: (ii)
-x,-y+1,-z+1.
Table 5. MLCT and HRS Data for Complex Salts 1-4 and 12-15
a
λ
max[MLCT]
â800
(nm)
(10-30 esu)
âxxz. Assuming Kleinman symmetry, âzyy ) âyyz and âzxx ) âxxz.
salt
Me2CO
MeCN
Me2CO
MeCN
Furthermore, if we assume an essentially 2D structure, then âzxx
) âxxz ) 0. Hence, only the components âzzz and âzyy are
significant. The lack of existing NLO chromophores structurally
similar to 1-4 and 12-15 renders comparisons with published
cases difficult. In previous HRS studies on the complex fac-
ReIBr(CO)3(L2)2 [L2 ) trans-4-(4-(dimethylamino)-phenyl-4-
buta-1,3-dienyl)pyridine] it was assumed that the chromophore
consists of two noninteracting 1D dipolar subunits, separated
by an angle of 90°.4d In such a case, âzzz and âzyy are equal. A
similar approach has been applied to neutral, purely organic
C2V chromophores.22,23 However, the observation of multiple
MLCT bands clearly shows that to treat the cations in 1-4 and
12-15 as containing noninteracting 1D dipolar subunits is rather
an oversimplification. Hence, to obtain further information about
the importance of off-diagonal tensor components in these
chromophores, we have measured HRS depolarization ratios F
for 1-4 (Table 6). The quantity F is defined as the ratio of the
intensities of the scattered SH light polarized parallel and
perpendicular to the polarization direction of the fundamental
light.21 Unfortunately, the relatively low harmonic scattering
intensities for 12-15 preclude the determination of F values
for these compounds.
1 cis-[RuII(NH3)4(MeQ+)2][PF6]4
586
508
624
536
638
540
672
568
514
544
552
574
570
502
606
528
620
536
644
558
510
540
548
568
135
142
135
160
139
2 cis-[RuII(NH3)4(PhQ+)2][PF6]4
3 cis-[RuII(NH3)4(4-AcPhQ+)2][PF6]4
4 cis-[RuII(NH3)4(2-PymQ+)2][PF6]4
127
142
143
12 [RuII(NH3)4(Me2Qpy2+)][PF6]4
13 [RuII(NH3)4(Ph2Qpy2+)][PF6]4
32
48
40
42
-
-
-
-
14 [RuII(NH3)4{(4-AcPh)2Qpy2+}][PF6]4
15 [RuII(NH3)4{(2-Pym)2Qpy2+}][PF6]4
a First hyperpolarizability measured by using a 800 nm Ti3+:sapphire
laser fundamental ((15% error), assuming a single major tensor component.
The quoted cgs units (esu) can be converted into SI units (C3 m3 J-2) by
dividing by a factor of 2.693 × 1020.
asymmetric counterparts should display only small responses,
5-11 were not investigated via HRS.
The following conclusions can be drawn from the data in
Table 5: (i) the â800 values for 1-4 are all similar and do not
show a significant variation in the two solvents chosen, and
(ii) the â800 values for 12-15 are also similar, but substantially
smaller than those of their nonchelated counterparts 1-4. In
our previous studies with related 1D dipolar complex chro-
mophores of N-R-4,4′-bipyridinium ligands, we have noted that
substantial increases in â0 (the estimated static first hyper-
polarizability) are observed on changing R from a Me to an
aryl substituent.5b-d Although the data in acetone may indicate
A F value of 5 is the upper limit for purely dipolar symmetry,
i.e. a single dipolar â tensor component or for C2V symmetry
with identical âzzz and âzyy tensor components, and in the limit
of ideal experimental conditions, while a F of 1.5 is the lower
9
4854 J. AM. CHEM. SOC. VOL. 127, NO. 13, 2005

Synthesis of Charged 2D Dipolar NLO Chromophores
A R T I C L E S
Table 7. Visible Absorption and Stark Spectroscopic Data for Complex Salts 1-4 and 12-15a
b
b
c
d
e
f
g
i
λmax
Emax
µ12
∆µ12
(D)
∆µab
(D)
r12
rab
Hab
â0j
b
2h
1
salt
(nm)
(eV)
fos
(D)
(Å)
(Å)
cb
(cm-
)
(10-30 esu)
k
1 cis-[RuII(NH3)4(MeQ+)2][PF6]4
628
518
632
512
684
558
731
589
547
497
570
581
608
1.98
2.39
1.96
2.42
1.81
2.22
1.97
2.10
2.27
2.49
2.17
2.13
2.04
0.34
0.24
0.41
0.20
0.40
0.26
0.40
0.27
0.18
0.20
0.22
0.24
0.26
6.7
5.2
7.4
4.6
7.6
5.6
7.9
5.9
4.5
4.6
5.2
5.4
5.8
10.1
11.3
13.8
9.9
12.1
13.0
11.7
12.3
13.0
18.4
12.9
12.9
12.9
16.8
15.3
20.3
13.5
19.5
17.1
19.7
16.9
15.8
20.5
16.6
16.9
17.3
2.1
2.4
2.9
2.1
2.5
2.7
2.4
2.6
2.7
3.9
2.7
2.7
2.7
3.5
3.2
4.2
2.8
4.1
3.6
4.1
3.5
3.3
4.4
3.5
3.5
3.6
0.20
0.13
0.16
0.13
0.19
0.12
0.20
0.14
0.09
0.05
0.11
0.12
0.13
6400
6500
5800
6700
5700
5800
5500
5900
5200
4500
5500
5500
5500
137
62
231
42
248
95
298
110
60
72
86
97
120
k
2 cis-[RuII(NH3)4(PhQ+)2][PF6]4
k
3 cis-[RuII(NH3)4(4-AcPhQ+)2][PF6]4
k
4 cis-[RuII(NH3)4(2-PymQ+)2][PF6]4
12 [RuII(NH3)4(Me2Qpy2+)][PF6]4
13 [RuII(NH3)4(Ph2Qpy2+)][PF6]4
14 [RuII(NH3)4{(4-AcPh)2Qpy2+}][PF6]4
15 [RuII(NH3)4{(2-Pym)2Qpy2+}][PF6]4
a Measured in butyronitrile glasses at 77 K. b For the two fitted Gaussian components. c Calculated from eq 4. d Calculated from fint∆µ12 using fint ) 1.33.
f
e Calculated from eq 3. Delocalized electron-transfer distance calculated from ∆µ12/e. g Effective (localized) electron-transfer distance calculated from
∆µab/e. h Calculated from eq 5. i Calculated from eq 6. j Calculated from eq 7. k Data taken in part from ref 9.
limit for complete octopolar symmetry. The value of 3.4
determined for the reference compound Disperse Red 1 (DR1)
is a good indication of a single major tensor component in DR1
and an appropriate experimental setup. The low F values
obtained for 1-4 clearly indicate that the hyperpolarizabilities
of these chromophores have substantial 2D character, and as
with the MLCT spectra, are not consistent with the presence of
two noninteracting chromophores. The values of âzzz and âzyy
can be experimentally determined from two observables, the
total HRS intensity â2HRS and F as follows:
majority of the Stark signal, but in 13-15 only one of the curves
contributed most of the signal, and thus the data yielded by the
insignificant curves have been neglected. Although the absorp-
tion profiles for 13-15 do indicate the presence of two MLCT
bands, the Stark fitting approach does not allow fits with two
reasonable Gaussian curves in these cases. The fitting invariably
affords one large Gaussian with a reasonable ∆µ12 (as given in
Table 7) and one small but very wide Gaussian with an
unrealistically large ∆µ12 for the types of complex chromophore
under study. For example, the data obtained for the small curve
in 14 gives a ∆µ12 of 7.42 eÅ or 35 D. Note that the ∆µ12
quoted for the second band in 12 (18.4 D) is also relatively
large, but perhaps not implausible. The Stark results for 1-4
are therefore probably more reliable than those for 12-15.
Gaussian-fitting has previously been adopted in Stark analyses
of purely organic34 and metal-containing chromophores,35 but
has only rarely been used to derive NLO responses.36 Repre-
sentative absorption spectra and electroabsorption spectra for
the salts 1 and 14 are shown in Figure 5.
â2HRS ) â2ZZZ + âY2ZZ
â2ZZZ
â2YZZ
(8)
F )
{
The HRS intensity with parallel polarization for fundamental
and SH wavelength, â2ZZZ , and for perpendicular polarization,
â2YZZ , are given in terms of the molecular tensor components
âzzz and âzyy according to
As observed previously in related compounds,5e,24 the MLCT
bands of 1-4 and 12-15 undergo large red-shifts when moving
from 295 K acetonitrile solution to a 77 K butyronitrile glass
(Tables 1 and 7). No clear trends are observed in the µ12 or
∆µ12 values, but the ∆µab values are considerably larger than
their adiabatic counterparts, as also observed in related
compounds.5e,24 The magnitudes of cb2 and Hab found for 1-4
and 12-15 are similar to those for related complex chromo-
phores,5e,24 and are consistent with the assignment of the visible
absorptions as arising from MLCT transitions in which the
degree of delocalization is limited. For the Qpy-based chro-
mophores in 12-15, this conclusion contrasts with the results
of previous electrochemical32a and vibrational spectroscopic32a,c
studies which indicate substantial ground-state charge delocal-
1
â2ZZZ ) â2zzz
7
6
35
9
35
+
âzzzâzyy
+
â2zyy
(9)
1
2
11
105
2
{
âYZZ
)
â2zzz
-
âzzzâzyy
+
â2zyy
35
105
and F can be expressed in terms of the parameter k ) âzyy/âzzz
by
15 + 18k + 27k2
3 - 2k + 11k2
F )
(10)
Application of eqs 8-10 to our experimental data in acetonitrile
affords the âzzz and âzyy values shown in Table 6. These data,
when compared with those in Table 5, show that F is a very
sensitive function of off-diagonal tensor components and
confirm the 2D nature of the molecular NLO responses of 1-4.
Stark Spectroscopic Studies. Complex salts 1-4 and 12-
15 were studied by Stark spectroscopy8 in butyronitrile glasses
at 77 K, and the results are presented in Table 7. Unsurprisingly,
Gaussian-fitting of the absorption spectra was necessary to
successfully model the Stark data, using two or three curves in
each case. For 1-4 and 12, two of these curves contributed the
ization in the parent complex cis-[RuII(NH3)4(2,2′-bpy)]2+
.
However, as noted earlier, it is apparent that introduction of
(34) Bublitz, G. U.; Ortiz, R.; Marder S. R.; Boxer, S. G. J. Am. Chem. Soc.
1997, 119, 3365-3376.
(35) (a) Karki L.; Hupp, J. T. Inorg. Chem. 1997, 36, 3318-3321. (b) Karki,
L.; Williams, R. D.; Hupp, J. T.; Allan, C. B.; Spreer, L. O. Inorg. Chem.
1998, 37, 2837-2840. (c) Bublitz, G. U.; Laidlaw, W. M.; Denning R. G.;
Boxer, S. G. J. Am. Chem. Soc. 1998, 120, 6068-6075. (d) Walters, K.
A.; Premvardhan, L. L.; Liu, Y.; Peteanu L. A.; Schanze, K. S. Chem.
Phys. Lett. 2001, 339, 255-262.
(36) Vance F. W.; Hupp, J. T. J. Am. Chem. Soc. 1999, 121, 4047-4053. Note,
however, that in this case the chromophore is a 3D pseudo-octopolar
derivative of [RuII(2,2′-bpy)3]2+ which has a complex electronic structure
in which â is dominated by intra-ligand CT, rather than MLCT processes.
9
J. AM. CHEM. SOC. VOL. 127, NO. 13, 2005 4855

A R T I C L E S
Coe et al.
Figure 5. Stark spectra and calculated fits for 1 and 14 in external electric fields of 3.26 and 2.79 × 107 V m-1, respectively. (Top) Absorption spectrum
illustrating Gaussian curves utilized in data fitting (numbers signify components which contribute significantly to the Stark signal; the data for which are
reported). (Bottom) Electroabsorption spectrum, experimental (blue) and fits (red) according to eq 1.
the 4-pyridinium substituents at the 4,4′-positions of the bpy
ligand markedly alters the overall electronic structures of these
complexes.
and temperature, which may affect the molecular NLO proper-
ties. Regarding the magnitudes of the observed â responses for
1-4 and 12-15, a factor which can be expected to limit these
is the decreased electron-donating strength of the RuII centers,
as evidenced by their RuIII/II E1/2 values of ca. 0.80 V vs Ag-
AgCl, which are ca. 300 mV higher than those of their
previously studied [RuII(NH3)5L]3+ analogues.5c,d
Theoretical Studies. To shed further light on the molecular
electronic structures of 1-4 and 12-15, we have performed
theoretical calculations on the representative model complex
chromophores in salts 1 and 12 using Gaussian 03. The
geometries were optimized at the B3P86/LanL2DZ model
chemistry, and both Finite Field (FF) and TD-DFT calculations
used the same level of theory.
For the complex in 1, there is an important question as to
the orientations of the coordinated pyridyl rings. In the only
reported X-ray crystal structure for a related compound, cis-
[RuII(NH3)4(isn)2][ClO4]2,38 these rings lie at angles of 37.2 and
38.6° with respect to the yz plane as defined here. Whether a
similar situation also pertains in a frozen glass or solution is
unclear, but some degree of rotation about the Ru-N(pyridyl)
bonds is obviously to be expected. Within the restriction of a
Although it is likely that application of the traditional two-
state model to the electronic transitions of 1-4 and 12-15 is
somewhat of an approximation,3g,k such a simplified approach
has been applied previously to purely organic C2V dipolar
chromophores, giving reasonable results.3a,d,f,j We have therefore
used eq 7 to estimate the magnitudes of the two orthogonal â0
components for our complexes (Table 7).37 As expected, the
results show that the molecular quadratic NLO responses of
1-4 and 12 have considerable 2D nature, in agreement with
the HRS data for 1-4. It is notable that the â0 values quoted in
Table 7 compare favorably with most of those previously
reported for C2V dipolar organic chromophores3a,d-l and neutral
metal complexes,2l,4 and are by no means yet optimized.
Comparison of the present results with the data available for
the related 1D complexes [RuII(NH3)5L]3+ [L(â0 × 1030 esu):
MeQ+ (120); PhQ+ (186); 4-AcPhQ+ (229)]5e affords the
following general conclusions: (i) the â0 values for both types
of chromophore are similarly large, and (ii) N-arylation causes
substantial increases in â0 when compared with the MeQ+
complexes in both 1D and 2D cases (although the data for 2
are perhaps anomalous).
C2V symmetric system, there are two extreme possibilities which
involve having the coordinated rings being either coplanar and
Broadly speaking, these Stark-derived â0 values are consistent
with the HRS â800 data (Table 5) inasmuch as the NLO
responses of the nonchelated chromophores are found to be
larger than those of their Qpy-based analogues. Furthermore,
the derivation of two substantial â components for 1-4 via both
techniques (Tables 6 and 7) verifies the 2D nature of the NLO
responses. However, any more detailed comparisons between
the two sets of data should be made only with caution. While
the Stark analysis indirectly affords estimated directional
components of â associated with the MLCT transitions, the HRS
measurements are direct but a function of all nonzero tensor
components. Furthermore, the Stark and HRS experiments are
carried out under quite different physical conditions of medium
(37) Note that in our previous communication,9 (somewhat smaller) â0 values
for 1-4 were derived by using µ12 values which were adjusted to take into
account the angles θ (ca. 20-50°) between µ12 and ∆µ12 estimated from
the Stark measurements. However, the theoretical analyses (see below)
indicate that for the low-energy transitions ∆µ12 lies along the z axis, whilst
µ12 is perpendicular and along the y axis. For the high-energy transitions,
∆µ12 and µ12 are parallel and lie along the z axis. The angles between µ12
and ∆µ12 should therefore be 90° and 0° for the low- and high-energy
transitions, respectively. Use of the component of µ12 along the ∆µ12
directions in the two-state model will necessarily give a â0 of 0 for a
transition where µ12 is perpendicular to ∆µ12, which is evidently nonsensical.
Furthermore, we can find no literature precedent for using directionally
adjusted µ12 values in the two-state approximation (see, for example: ref
3a, eqs 40 and 41 in which µy is used with ∆µz). We have therefore used
the as-measured µ12 values here to derive â0 values which should be viewed
as replacements for those previously reported.
(38) Richardson, D. E.; Walker, D. D.; Sutton, J. E.; Hodgson, K. O.; Taube,
H. Inorg. Chem. 1979, 18, 2216-2221.
9
4856 J. AM. CHEM. SOC. VOL. 127, NO. 13, 2005

Synthesis of Charged 2D Dipolar NLO Chromophores
A R T I C L E S
Table 8. Results of B3P86/LanL2DZ and Finite Field Calculations for the Electronic Transitions of the Complex Cations in Salts 1 and 12
cation structure
E
max (eV)
∆µ12 (D)
µ12 (D)
fos
symmetry
major contributions
âzyyc (10-30 esu)
âzzzc (10-30 esu)
1a
2.42
14.7
5.4a
0.27
B2
HOMO-1 f LUMO+1 (59%)
HOMO f LUMO (37%)
HOMO f LUMO+1 (56%)
HOMO-1 f LUMO (31%)
HOMO f LUMO (93%)
HOMO f LUMO+1 (86%)
HOMO-1 f LUMO+1 (48%)
HOMO f LUMO (44%)
HOMO f LUMO+1 (62%)
HOMO-1 f LUMO (20%)
HOMO-1 f LUMO+1 (92%)
HOMO-1 f LUMO+2 (3%)
HOMO f LUMO+1 (37%)
HOMO-1 f LUMO (32%)
HOMO-1 f LUMO+2 (8%)
72
2.66
14.3
3.1b
0.10
A1
17
19
18
1b
1c
2.43
2.77
2.62
13.5
12.9
13.6
6.7a
3.7b
6.4a
0.41
0.14
0.40
B2
A1
91
70
3.00
2.65
2.91
14.1
18.0
14.2
3.9b
3.9a
5.3b
0.17
0.15
0.31
12
B2
A1
54
36
a Transition dipole moment along the y axis. b Transition dipole moment along the z axis. c Obtained from FF calculations.
Figure 6. Conformations 1a (left), 1b (center) and 1c (right) of the cation
1 used in the theoretical calculations.
lying in the yz plane (denoted 1a) or both lying perpendicular
to this plane (denoted 1b). The latter option corresponds to that
assumed in the earlier MO treatment of Zwickel and Creutz.30b
In both structures, the pyridinium rings are coplanar with the
adjacent pyridyl rings. To present a complete picture, we have
therefore conducted calculations on both of these alternate
conformations for the cation in 1, and also for the fully
optimized structure without any restrictions (denoted 1c). The
latter has an almost C2 geometry, with the coordinated pyridyl
rings lying at an angle of 45° (in the same direction) with respect
to the yz plane, in fashion similar to that for the related crystal
structure,38 while the pyridinium rings are rotated by 77° out
of the yz plane (again in the same direction). The three
conformations of the cation in 1 are shown in Figure 6.
According to the TD-DFT calculations, the complex cation
in 1 has several absorption bands below 3 eV, but only the two
of these which have substantial intensity will be considered here
(Table 8). The predicted Emax and ∆µ12 values are generally
somewhat larger than the corresponding experimental data
(Tables 1 and 7), but no consistent pattern is observed in the
12 values. More importantly, but not unexpectedly, our calcula-
Figure 7. Illustration of the 0.06 contour surface diagrams of the molecular
orbitals of 1 (conformation 1a) involved in the two lowest-energy transitions
µ
tions conclude that the model previously proposed for related
complex chromophores cis-[RuII(NH3)4(L1)2]2+ (L1 ) py, etc.)
by Zwickel and Creutz30b is only valid for the 1b conformation.
For conformations 1a and 1c, a more complex model is required
since the HOMO is not the only occupied orbital involved in
the low-energy electronic transitions; the HOMO-1 is very
close in energy to the former and also contributes to these
transitions. The molecular orbitals involved in these low-energy
transitions in the complex cation in salt 1 are shown in Figure
7 (for simplicity, only the 1a conformation is depicted). It can
be seen that the HOMO and HOMO-1 are both ruthenium d
orbitals, while the LUMO and LUMO+1 are π orbitals located
on the 4,4′-bipyridinium ligands and consequently these transi-
tions are MLCT processes.
Although conformational changes in the cation in 1 modify
the contribution of different orbitals to the MLCT transitions
and hence their energies and oscillator strengths, their sym-
9
J. AM. CHEM. SOC. VOL. 127, NO. 13, 2005 4857

A R T I C L E S
Coe et al.
Figure 8. Illustration of the 0.06 contour surface diagrams of the molecular orbitals of 12 involved in the two lowest-energy transitions (structure restricted
to having all four rings coplanar)
metries are unchanged: the lowest-energy transition is polarized
along the y axis, while the second lowest-energy transition is
polarized along the z axis. Considering that the dipole moments
of ground and excited states are directed along the z axis, the
lowest-energy transition should be associated with the âzyy
response and the higher-energy transition with âzzz. The â values
shown in Table 8 have been evaluated using the FF approach
which involves the numerical calculation of the second deriva-
tive of the dipole moment with respect to the electric field. Quite
surprisingly, âzzz is nearly insensitive to the molecular confor-
mation while âzyy does not change on passing from conformation
1a to 1c, but reaches a somewhat higher value in conformation
1b with the pyridyl rings perpendicular to the yz plane. Although
these calculated â values differ somewhat from those obtained
by applying eq 7 to the Stark data (Table 7), they do verify the
2D nature of the NLO responses.
Calculations on the complex cation in 12 were performed
restricted to the C2V symmetry39 with all four rings lying in the
yz plane, which can be directly compared to 1a. The results of
these calculations are shown in Table 8, and the molecular
orbitals involved in the low-energy transitions are depicted in
Figure 8. It should be noted that transitions to the LUMO+2
have an important contribution to the lowest-energy absorptions
in the case of the cation in 12, but not for that in 1. This effect
is due to the linking of the two 4,4′-bpy ligands in 12 that causes
a lowering in the energy of the LUMO-2, which is in this
instance only 0.47 eV above the LUMO, compared to 1.45 eV
above the LUMO in 1. The calculated Emax values show that
these electronic transitions are hypsochromically shifted in 12
when compared with 1. On moving from 1 to 12, µ12 and fos
decrease for the lowest-energy transition and increase for the
high-energy one. In any case, the symmetries of these MLCT
transitions are the same in both 1 and 12: the lowest-energy
transition has B2 symmetry, while the second lowest-energy
transition is A1. The predicted differences (which mostly parallel
the experimental trends shown in Table 7) account for a
decreased âzyy and an increased âzzz for 12 when compared with
1, as confirmed by our FF calculations and the Stark measure-
ments. However, the latter data indicate that âzzz for 12 is
actually larger than âzyy, in contrast with the FF results. It is
worth noting that the results of our calculations are consistent
with those of previous ZINDO studies on the prototypical
complex cis-[RuII(NH3)4(2,2′-bpy)]2+ which predict only rela-
tively weak MLCT transitions polarized perpendicular to the
bpy plane.32b
Previous ZINDO computational studies have addressed the
effects on linear absorption and quadratic NLO properties of
the dihedral angles between aryl rings in purely organic
chromophores, as a potential approach to modulating optical
behavior at the molecular level.40 Similar studies have also been
carried out with the organometallic complex trans-[RuII(Ct
CC6H4C6H4NO2-4,4′)Cl(dppm)2] [dppm ) bis(diphenylphos-
(39) A full unrestricted geometry optimization of the cation in 12 resulted in
the pyridinium rings being rotated by ca. 38° with respect to the pyridyl
rings. However, the properties calculated for this geometry are similar to
those reported in Table 8 for the completely planar conformation: Emax
)
2.79 and 2.88 eV.; âzyy ) 42 × 10-30 esu; âzzz ) 30 × 10-30 esu.
9
4858 J. AM. CHEM. SOC. VOL. 127, NO. 13, 2005

Synthesis of Charged 2D Dipolar NLO Chromophores
A R T I C L E S
phino)methane].41 It is notable that, because our FF calculated
âzyy values for the complex chromophore in 1 show some
variation as the relative orientations of the pyridyl and pyri-
dinium rings are changed, such species correspond (at least in
theory) to systems in which the relative 2D nature of the â
response can be controlled.
By analogy with the published studies with BG,43 it seems
likely that a comparable situation may pertain for 1-4,
especially since our HRS measurements used an SH at 400 nm,
on the high-energy side of the MLCT bands. Because our
calculations were carried out at zero frequency where Kleinman
symmetry is exact, it is not wholly unexpected that their
conclusions should differ somewhat from those derived from
HRS measurements at an optical frequency where Kleinman
symmetry breaking may occur. The calculations indicate that
the higher-energy MLCT transition, which is closer to resonance
when using a 400 nm SH, is the main contributor to âzzz.
However, dispersion may change a measured â value by as
much as an order of magnitude when compared with â0, and
the dispersion effects may be different for the âzzz and âzyy
components because these are associated with transitions which
lie at different distances from resonance.
Further Comments on the NLO Properties of the Chro-
mophores in 1-4. Our theoretical analyses of the cation in 1
are consistent with the observed visible absorption and Stark
spectroscopic results for 1-4 in predicting two MLCT transi-
tions, the lower-energy one corresponding with the larger off-
diagonal â component (âzyy) and the higher-energy one corre-
sponding with the smaller â component (âzzz) (Tables 7 and 8).
Although the Stark data do not clearly identify which â
component corresponds with each transition,42 application of
the theoretical outcomes to the Stark results gives a âzyy/âzzz
ratio of 2-3 for 1, 3, and 4, while the corresponding FF-derived
ratios for the cation in 1 are ca. 4-5 for all three conformations.
This observed dominance of the NLO response by off-diagonal
â component(s) is very unusual and somewhat reminiscent of
the purely organic chromophore N,N′-dihexyl-1,3-diamino-4,6-
dinitrobenzene, for which âzyy and âyzy have been found to be
similar and about an order of magnitude larger than âzzz.3l
However, for 1-4 there is clearly a discrepancy between the
calculated and HRS-derived â components, because the latter
data indicate that âzzz is ca. 2-3 times larger than âzyy (Table
6). It appears likely that this apparent discrepancy may arise
from dispersion and resonance effects.
Conclusions
We have synthesized and characterized the first family of
charged 2D dipolar NLO chromophores based on redox-
switchable transition metal centers. These cis-{RuII(NH3)4}2+
compounds are found to display relatively large and 2D first
hyperpolarizabilities by HRS at 800 nm, and Stark spectroscopic
studies also show that the NLO responses have two substantial
orthogonal components. The orbital structures have been
elucidated by using TD-DFT which shows that the lowest-
energy MLCT transition is associated with the âzyy response
and the higher-energy transition with âzzz. The â values obtained
from FF calculations are somewhat smaller than those derived
from the Stark data, and âzyy shows some variation as the relative
orientations of the pyridyl and pyridinium rings are changed.
Further studies will involve the maximization of â responses
by modifications of the electron-accepting ligand structures and
co-ligands. We also intend to investigate the possibilities for
multistate redox-switching of the NLO properties by exploiting
reversible RuIII/II oxidation processes and ligand-based reduc-
tions.
Previous polarized HRS studies by Kaatz and Shelton
involving the triarylmethane dye Brilliant Green (BG)43 are
especially relevant to our studies. The C2V chromophore in BG
has an intramolecular charge-transfer absorption band at λmax
) 632 nm [in 1,4-dioxane/methanol (95:5)], and HRS measure-
ments were carried out with both 1319 and 1064 nm lasers,
giving SH wavelengths of 660 and 532 nm, on either side of
the resonance peak.43 These experiments yielded â and F values
similar to those which we have obtained for 1-4, and further
F measurements using elliptically polarized incident light were
employed to test the validity of some of the assumptions made
in the analysis of the polarization results. With the SH at 660
nm, a k value of -0.34 was obtained for BG (very similar to
that for 1), and the assumptions of 2D â and Kleinman symmetry
were found to be valid.43 In contrast, the elliptically polarized
data showed that Kleinman symmetry is broken when using
the 532 nm SH, with out-of-plane â components becoming
significant.43 Subsequent HRS studies with BG and the closely
related dye Malachite Green have concluded that measurements
of F alone can give a misleading indication of the NLO character
of such chromophores, if Kleinman-disallowed contributions are
not taken into consideration.3j Similar concerns have been noted
in relation to 2D Ni(II) Schiff base complexes4f and a Zn(II)
phthalocyanine derivative.4i
Acknowledgment. We thank the EPSRC for support (Grants
GR/M93864 and GR/R54293) and also the Fund for Scientific
Research-Flanders (FWO-V, G.0297.04), the University of
Leuven (GOA/2000/3), the Belgian Government (IUAP P5/3),
MCyT-FEDER (BQU2002-00219), and Gobierno de Aragon-
Fondo Social Europeo (P009-2001 and E39). This research was
partially carried out at Brookhaven National Laboratory under
Contract No. DE-AC02-98CH10886 with the U.S. Department
of Energy and supported by its Division of Chemical Sciences,
Office of Basic Energy Sciences. Dr. Kurt Wostyn is thanked
for his assistance with some of the HRS measurements, and
we are very grateful to Profs. David P. Shelton (UNLV) and
Santo Di Bella (Universita` di Catania) for some helpful
comments regarding the physical aspects of these studies.
(40) (a) Albert, I. D. L.; Marks, T. J.; Ratner, M. A. J. Am. Chem. Soc. 1997,
119, 3155-3156. (b) Albert, I. D. L.; Marks, T. J.; Ratner, M. A. J. Am.
Chem. Soc. 1998, 120, 11174-11181.
Supporting Information Available: Cartesian coordinates of
theoretically optimized geometries for the cation conformations
1a, 1b, 1c and for the cation in 12 (C2V and C1 forms); complete
ref 25. This material is available free of charge via the Internet
at http://pubs.acs.org.
(41) McDonagh, A. M.; Whittall, I. R.; Humphrey, M. G.; Skelton, B. W.; White,
A. H. J. Organomet. Chem. 1996, 519, 229-235.
(42) Note however that the observation that the Stark-derived θ value is, in
every case, considerably larger for the lower-energy transition than for the
higher-energy one9 is consistent with the assignment of the former transition
as corresponding with the âzyy component for which µ12 is perpendicular
to ∆µ12 (i.e. θ ) 90°).
(43) Kaatz, P.; Shelton, D. P. J. Chem. Phys. 1996, 105, 3918-3929.
JA0424124
9
J. AM. CHEM. SOC. VOL. 127, NO. 13, 2005 4859