A Series of Multicolor Electrochromic Ruthenium(II) Trisbipyridine Complexes: Synthesis and Electrochemistry

Article abstract of DOI:10.1021/jp991196w

A series of 5,5′-disubstituted 2,2′-bipyridines and their corresponding tris complexes with ruthenium(II) have been synthesized. The substituents used (ketone, ester, nitrile, imide, and two amides) are all electron withdrawing in nature and, with one exception, contain a carbonyl group in the position α to the bipyridine ring. The reduction potentials of the free ligands and ruthenium complexes have been determined by cyclic voltammetry and are correlated with the Hammett ρ constants of the substituents. Finally, the electron- withdrawing nature of these substituents shifts the reduction potentials of each complex sufficiently positive that up to six stable ligand-based reductions are observable. In these reduced oxidation states, all of the complexes display multicolor electrochromism.

Full text of DOI:10.1021/jp991196w

J. Phys. Chem. A 1999, 103, 6263-6267
6263
A Series of Multicolor Electrochromic Ruthenium(II) Trisbipyridine Complexes: Synthesis
and Electrochemistry
Franc¸ois Pichot,† Jeffrey H. Beck, and C. Michael Elliott*
Department of Chemistry, Colorado State UniVersity, Fort Collins, Colorado 80523
ReceiVed: April 12, 1999; In Final Form: May 28, 1999
A series of 5,5′-disubstituted 2,2′-bipyridines and their corresponding tris complexes with ruthenium(II) have
been synthesized. The substituents used (ketone, ester, nitrile, imide, and two amides) are all electron
withdrawing in nature and, with one exception, contain a carbonyl group in the position R to the bipyridine
ring. The reduction potentials of the free ligands and ruthenium complexes have been determined by cyclic
voltammetry and are correlated with the Hammett σ constants of the substituents. Finally, the electron-
withdrawing nature of these substituents shifts the reduction potentials of each complex sufficiently positive
that up to six stable ligand-based reductions are observable. In these reduced oxidation states, all of the
complexes display multicolor electrochromism.
Introduction
vacuum. Due to the reactive nature of the acid chloride, no
further purification was attempted and the acid chloride was
used for the next step immediately after drying.
Synthesis of 5,5′-Bis(ethoxycarbonyl)-2,2′-bipyridine: L1.
This ligand was prepared as previously described.4a
Trisbipyridineruthenium(II) complexes possess very rich
photochemical and electrochemical properties which have been
exploited in both applied and fundamental ways.1-3 In one
particular instance,4a the complex [Ru(L1)3]2+, where L1 is 5,5′-
bis(ethoxycarbonyl)-2,2′-bipyridine, displays distinctly different
absorption spectra for each accessible oxidation state (formal
states +2 through -4). The specific spectral changes result in
a vivid multicolor electrochromism which could potentially find
applications in display devices. Interestingly, while the complex
having analogous ester substitution in the 4- and 4′-positions
exhibits similar electrochemistry, it lacks the pleasing electro-
chromism.4b Also, electron-donating substituents (in either the
5- or 4-position of substitution) do not produce multicolor
electrochromism; moreover, they decrease the number of
reduced oxidation states accessible within the typical electro-
chemical solvent window.5 This study focuses on the effect that
several electron-withdrawing functional groups located in the
5- and 5′-positions of the bipyridine have on the electrochemistry
of the ligands and on the electrochromic properties of their tris
complexes with ruthenium. Furthermore, the relationship be-
tween the redox potentials of the ligand and the Hammet
constant for the substituents is considered.
Synthesis of 5,5′-Bis(diethylamido)-2,2′-bipyridine: L2. A
large excess (>20 equiv) of diethylamine was combined with
5COClBp in a refluxing mixture (1:1 V/V) of acetonitrile and
dichloromethane for 12 h under nitrogen. After rotary evapora-
tion of the solvent, the product was redissolved in dichlo-
romethane, washed with a solution of sodium carbonate in water,
and separated from impurities by liquid chromatography (silica
gel; 1:1 ethyl acetate/dichloromethane). The yield was nearly
1
quantitative. MS: m/z ) 354.3 (parent molecule). H NMR δ
in ppm from TMS, CDCl3 (multiplicity, integration): 1.1(bs,
3H); 1.3(bs, 3H); 3.3(bs, 2H); 3.6(bs, 2H); 7.9(d, 1H); 8.5(d,
1H); 8.7(s, 1H).
Synthesis of 5,5′-Di(N-methyl-N-phenylamido)-2,2′-bipy-
ridine: L3. The same reaction conditions as for L2, replacing
diethylamine with N-methylaniline, were used. The product, L3,
was separated from polar impurities by liquid chromatography
(silica gel, ethyl acetate) and recrystallized from ethyl acetate.
The yield was nearly quantitative. MS: m/z ) 422.2 (parent
molecule). 1H NMR δ in ppm from TMS, CDCl3 (multiplicity,
integration): 3.5(s, 3H); 7.1(m, 2H); 7.3(m, 3H); 7.7(d, 1H);
8.1(d, 1H); 8.5(s, 1H).
Experimental Section
Generalities about the Syntheses of Ligands and Com-
plexes. All the ligands prepared were 2,2′-bipyridines substituted
in the 5- and 5′-positions with electron-withdrawing functions.
The first step in each synthesis involved the transformation of
the 5,5′-dicarboxylicacid (5COOHBp) into the corresponding
diacid chloride (5COClBp). In a typical preparation, 1-2 g of
5COOHBp was combined with a large excess of thionyl chloride
(100 to 200 mL) and allowed to reflux under nitrogen until the
mixture became translucent. After an additional hour of reflux
ensuring completion of the reaction, the thionyl chloride was
removed by rotary evaporations and the product was dried under
Synthesis of 5,5′-Dicyano-2,2′-bipyridine: L4. In a typical
preparation, 200 mg of 5COClBp was dissolved in 200 mL of
warm dichloromethane. This solution was added dropwise over
a period of 30 min to a vigorously stirred aqueous solution (300
mL) of ammonium hydroxide (approximately 0.5 M). Upon the
addition, the primary amide formed precipitates. After the
addition was completed, the white precipitate was filtered,
washed with water, and dried overnight under vacuum. This
amide was then ground into a fine powder and allowed to react
with a large excess of refluxing thionyl chloride, which acts as
a dehydrating agent,6 under nitrogen for 24 h. After rotary
evaporation of the thionyl chloride, the product was purified
by liquid chromatography (silica gel, ethyl acetate). The yield
was typically less than 40%. MS: m/z ) 206.1 (parent
* Corresponding Author.
† Present address: National Renewable Energy Laboratory, 1617 Cole
Blvd., Golden, CO 80401.
10.1021/jp991196w CCC: $18.00 © 1999 American Chemical Society
Published on Web 07/16/1999

6264 J. Phys. Chem. A, Vol. 103, No. 31, 1999
Pichot et al.
Figure 1. Structures of the substituted 2,2′-bipyridines used in this study.
molecule). 1H NMR δ in ppm from TMS, CDCl3 (multiplicity,
intregation): 8.1(d, 1H); 8.6(d,1H); 9.0(s, 1H).
this precipitate, the desired complex was isolated by liquid
chromatography (silica gel, 1:1 acetonitrile/aqueous 0.02 M
KNO3). After rotary evaporation of the acetonitrile, the complex
was reprecipitated as the hexafluorophosphate salt, filtered,
washed with water, and dried under vacuum. The ruthenium
source was RuCl3‚3H2O. The reaction mixture was heated to
160 °C in a sealed Pyrex tube for 18 h.
Synthesis of [Ru(L5)3](PF6)2. The solvent used for this
synthesis was DMF, as ethylene glycol yielded a mixture of
products, presumably due to the formation of ketals and
hemiketals on the bipyridine substituents.
Synthesis of 5,5′-Di(1-ketobutyl)-2,2′-bipyridine: L5. This
compound was synthesized via an adapted literature prepara-
tion.7 Typically, 0.5 g of 5COClBp was added to 3 equiv of
(n-Bu)2CuLi in THF at -78 °C under argon. After an hour at
this temperature, the solution was slowly warmed to room
temperature and subsequently quenched with water. After rotary
evaporation of the THF, the product was extracted with ethyl
acetate and separated from impurities by liquid chromatography
(silica gel, ethyl acetate). The ketone was further purified by
recrystallization from ethyl acetate. Overall yield ) 15%. MS:
m/z ) 325.41 (protonated parent molecule). 1H NMR δ in ppm
from TMS, CDCl3 (multiplicity, integration): 1.0(t, 3H); 1.4-
(m, 2H); 1.8(m, 2H); 3.0(t, 2H); 8.3(d, 1H); 8.6(d, 1H); 9.2(s,
1H).
[Ru(L2)3](PF6)2. 1H NMR δ in ppm from TMS, CDCl3
(multiplicity, integration): 1.0(bs, 3H); 1.2(bs, 3H); 3.8(bs, 2H);
4.1(bs, 2H); 7.6 (s, 1H); 8.1(d, 1H); 8.6(d, 1H). UV-visible
(wavelength (nm), ꢀ/ꢀMLCT): 224 (2.3); 256 (2.8); 305 (5.7);
440 (0.8); 470 (1).
Synthesis of 5,5′-Di((N-(ethan-1-one)-N-butylimide)-2,2′-
bipyridine: L6. In a typical preparation, 200 mg of 5COClBp
was dissolved in 200 mL of warm dichloromethane. This
solution was added dropwise to a large excess of n-butylamine
dissolved in water (200 mL 0.5 M). The primary amide formed
immediately precipitated. This white precipitate was filtered,
washed with water, and dried overnight under vacuum. After it
was ground into a fine powder, this amide was allowed to react
with a large excess of refluxing acetyl chloride under nitrogen.
After solvent rotary evaporation, the final product was purified
by liquid chromatography (silica gel, 1:1 ethyl acetate/dichlo-
romethane) and recrystallized from cyclohexane. Yield ) 60%.
[Ru(L3)3](PF6)2. 1H NMR δ in ppm from TMS, CDCl3
(multiplicity, integration): 3.3(s, 3H); 7.0(m, 2H); 7.2(m, 3H);
7.5(s, 1H); 7.8(d, 1H); 8.2(d, 1H). UV-visible (wavelength
(nm), ꢀ/ꢀMLCT): 210 (8.3); 306 (7.1); 447 (0.7); 480 (1).
[Ru(L4)3](PF6)2. 1H NMR δ in ppm from TMS, CDCl3
(multiplicity, integration): 8.0(s, 1H); 8.2(d, 1H); 8.8(1H). UV-
visible (wavelength (nm), ꢀ/ꢀMLCT): 256 (5.2); 298 (10.4); 460
(0.8); 502 (1).
[Ru(L5)3](PF6)2. 1H NMR δ in ppm from TMS, CDCl3
(multiplicity, integration): 1.1(t, 3H); 1.5(m, 2H); 1.9(m, 2H);
2.9(t, 2H); 3.6(bm, 6H); 7.9(s, 1H); 8.2(d, 1H); 8.7(d, 1H). UV-
visible (wavelength (nm), ꢀ/ꢀMLCT): 226 (3.4); 268 (3.6); 310
(9); 465 (0.8); 502 (1).
1
MS: m/z ) 439.33 (protonated parent molecule). H NMR δ
in ppm from TMS, CDCl3 (multiplicity, integration): 0.7(t, 3H);
1.2(m, 2H); 1.5(m, 2H); 2.2(s, 3H); 3.7(t, 2H); 8.0(d, 1H); 8.4-
(d, 1H); 8.7(s, 1H).
Attempted Synthesis of [Ru(L6)3](PF6)2. Attempts at syn-
thesizing this complex under a variety of conditions were
unsuccessful.
The structures of the previously synthesized ligand (L1) and
of the newly synthesized ligands (L2, L3, L4, L5, and L6) are
presented in Figure 1.
Electrochemistry of the Free Ligands and Corresponding
Ruthenium Complexes. Tetra-n-butylammonium hexafluoro-
phosphate (TBAPF6) was prepared by metathesis of tetra-n-
butylammonium iodide with ammonium hexafluorophosphate
and was recrystallized three times from ethanol. “Distilled in
glass” acetonitrile was purchased from Burdick and Jackson and
was used without further purification.
All electrochemical experiments were carried out with a
EG&G PAR model 173 potentiostat/galvanostat in conjunction
with a EG&G PAR model 175 programmer. Cyclic voltammo-
grams were recorded on a Yokogama 3023 X-Y recorder.
Cyclic voltammetry was performed under nitrogen in a two-
compartment cell using a 2 mm diameter platinum disk as the
working electrode, a coiled platinum wire as the counter
electrode, and an SSCE as the reference electrode. Bulk
Synthesis of [Ru(L1)3](PF6)2, [Ru(L2)3](PF6)2, [Ru(L3)3]-
(PF6)2, and [Ru(L4)3](PF6)2. These complexes were all syn-
thesized in the same way. Between 30 and 70 mg of Ru-
(DMSO)4Cl2 was added to 2-4 mL of ethylene glycol in a 25
mL round-bottom flask and quickly brought to reflux until the
solution became orange. The solution was then immediately
cooled to 120 °C, whereupon 3.3 equiv of ligand (L1, L2, L3,
or L4) were added. After completion of the reaction (as
determined by TLC), the solution was cooled to room temper-
ature and diluted with 10-15 mL of water. Upon addition of
1-2 mL of saturated aqueous NH4PF6, the final product
precipitates as the hexafluorophosphate salt. After filtration of

Electrochromic Ruthenium(II) Trisbipyridine Complexes
J. Phys. Chem. A, Vol. 103, No. 31, 1999 6265
Figure 2. Cyclic voltammogram of L5, taken in 0.1 M TBAPF6/
acetonitrile; platinum working electrode; scan rate ) 50 mV/s.
Figure 3. Cyclic voltammogram of Ru(L5)3(PF6)2, taken in 0.1 M
TBAPF6/acetonitrile; platinum working electrode; scan rate ) 50 mV/
s.
electrolysis experiments employed a three-compartment cell,
using a 1 cm2 platinum grid in the middle compartment as the
working electrode and a platinum grid (2 cm2) counter electrode
and an SSCE reference electrode in each of the other two
compartments. The bulk electrolysis experiments were per-
formed to determine chromophoric properties of the complexes
in each oxidation state (2+, 1+, 0, 1-, 2-, etc.). Obtaining each
ligand-based reduction state in a completely pure form is
impossible due to the proximity in potential of adjacent redox
processes for these complexes (the difference in E1/2 between
successive reductions is less than 150 mV for each complex of
this series, thus significant disproportionation occurs; cf. the
cyclic voltammogram in Figure 3). The predominance of a redox
species (g80%) is, however, obtained by applying a potential
midway between consecutive formal potentials. Consequently,
applying these potentials to vigorously stirred solutions of each
complex allowed for rapid and optimized generation of the max-
imum relative amounts of each formal oxidation state species.
The visual observation of the colored solutions was performed
after the solutions had reached thermodynamic equilibrium
(electrochemical currents less than 1% of the initial current).
TABLE 1: Electrochemical Data and Hammett Constant
Values for a Selection of Substituted 2,2′-Bypyridines
E1/2(L0/-1
V vs SSCE
)
E1/2(L-1/-2
)
V vs SSCE
a
a
ligand
σmeta
σpara
2,2′-Bpy
-2.10
-2.29
-1.28
-1.81
-1.71
-1.30
-1.39
-1.27
0.00
-0.07
0.37
0.00
-0.17
0.45
5,5′-DMBpy
L1
L2
L3
L4
L5
L6
-1.67
0.35
0.36
0.56
0.38
0.66
0.50
-1.62
a Taken from ref 21.
the para position or the meta position) with the Ka of unsub-
stituted benzoic acid.9 Many substituents have been assigned
two parameters σmeta and σpara, which are defined by σmeta
)
log(KXmeta/K0) and σpara ) log(KXpara/K0), where KXmeta and KXpara
are the acid dissociation constants of the meta-substituted and
para-substituted benzoic acid, respectively, and K0 is the
dissociation constant of unsubstituted benzoic acid. Electron-
withdrawing substituents increase the dissociation constant and
thus have positive σ values, while electron-donating ones
decrease the dissociation constant and have negative σ values.
For these bipyridine ligands and complexes, in the electro-
chemical context, the distinction between para and meta
positions is not as significant as it is in the case of benzoic acid
dissociation. Even with the substituents in the 5- and 5′-
positions, the reaction of interest (injection of an electron in
the π* orbital of the bipyridine) is not localized at a specific
site of the ring system, as it is in the benzoic acid ionization
case. In an attempt at drawing a linear correlation between
Results and Discussion
Preparation of the Complexes. The synthesis and charac-
terization of the ligands was straightforward. Ru(L1)3 , Ru(L2)3,
Ru(L3)3, Ru(L4)3, and Ru(L5)3 also were prepared and isolated
in pure form as their PF6- salts. The lack of success in preparing
Ru(L6)3 is, in retrospect, not surprising. Noncyclic imides are
notoriously reactive, particularly to nucleophilic attack at the
nitrogen. Under conditions that are required to initiate ligand
exchange at the inert Ru(II) center, L6 likely degrades.
Electrochemistry of the Free Ligands and Corresponding
Ruthenium Complexes. The E1/2(L0/-1) values obtained from
cyclic voltammetry experiments were used to assess the electron-
withdrawing strength of each substituent on the bipyridine.
Under the conditions used (acetonitrile, 0.1 M TBAPF6;
platinum working electrode; scan rate of 50 mV/s), the series
of compounds displayed good chemical reversibility. The cyclic
voltammograms of L5 and Ru(L5)3(PF6)2 are displayed in Figures
2 and 3, respectively. The cyclic voltammograms of the other
compounds of this study are included as Supporting Information
(except for Ru(L1)3 which has been reported previously4a). In
the case of L1 and L5, the second one-electron reduction of the
bipyridine was observed, whereas it was not observed within
the solvent window for the other ligands. The reduction
potentials are reported in Table 1. Also presented in this table
are the Hammett constant values for these substituents.8 The
electronic influence of a substituent X can be estimated by
comparing the Ka of substituted benzoic acid (with X either in
E
1/2(L0/1-) and σ, a slightly better regression was obtained when
using σp (r2 ) 0.901) as compared to σm (r2 ) 0.845) (see Figure
4). However, in both cases, the fits are fairly poor and it is
more appropriate to look at the trends in a qualitative way.
Considering the σ values for the amide substituents, it is
surprising that they are much harder to reduce than the ester
analogue [E1/2(L20/1-) ) -1.81 V vs E1/2(L10/1-) ) -1.28 V].
The difference between the diethylamide and the methylphe-
nylamide was expected and can be explained as follows. The
π-system of the phenyl ring provides some additional delocal-
ization pathway for the amide nitrogen lone pair in L3. This, in
turn, leaves the amide carbonyl with more electron-withdrawing
strength toward the pyridine ring, thus making it easier to reduce.
The same effect can be observed with the imide-substituted
bipyridine (L6), where the additional carbonyl function bound
to the nitrogen makes the carbonyl R to the pyridine ring even
more electron withdrawing. This imide ligand is, within

6266 J. Phys. Chem. A, Vol. 103, No. 31, 1999
Pichot et al.
Figure 5. Correlation between the first reduction potential of selected
free ligands with the first ligand-based reduction potential of the
corresponding ruthenium tris complex. The ligands included are 5,5′-
dimethyl-2,2′-bipyridine, 2,2′-bipyridine, L1, L2, L3, L4, and L5.
Figure 4. Correlation between the first reduction potential of the free
ligands and the Hammett σ constants. Regression coefficients: σmeta
,
r2 ) 0.845; σpara, r2 ) 0.901.
) 0.947, slope ) 0.69). This correlation is greatly improved
(r2 ) 0.989, slope ) 0.76) if the data point of the nitrile-
substituted ligand (L4) is not taken into account. The reduction
potentials of the complexes are more positive than those of their
respective free ligands. This indicates that, in each complex,
the reduced radical anion is a better ligand and thus bound more
tightly than its neutral counterpart.4a The larger the difference
between the Ru(L)32+/1+ and L0/1- reduction potential, the larger
is the binding energy preference of Ru(II) for L-• over L. The
slope of the line in Figure 4 is less than unity, and this fact
provides some subtle insight into the mode of binding in these
complexes. The main contribution comes from σ donation from
the bipyridines’ nitrogen lone pairs, but also involved is the
TABLE 2: Electrochemical Data for a Series of Ruthenium
Trisbipyridine Complexes
E1/2(L0/-1
V vs SSCE
)
E1/2[Ru2+/1+(L)3]
V vs SSCE
|∆(E1/2s)|
ligand
V
2,2′-Bpy
5,5′-DMBpy
L1
L2
L3
L4
L5
-2.10
-2.29
-1.28
-1.81
-1.71
-1.30
-1.39
-1.27
-1.41
-0.66
-0.98
-0.93
-0.81
-0.70
0.83
0.88
0.62
0.83
0.78
0.49
0.69
experimental error, as easy to reduce as the ester analogue, and
its corresponding ruthenium complex, had it been successfully
prepared, would have been interesting to characterize. The fact
that the imide ligand is easier to reduce than the ketone probably
originates from the fact that the alkyl substituent â to the
pyridine ring (in L5) is σ-donating, whereas in L6, the “amide”
group â to the pyridine ring is σ-accepting.
In view of the σ values of cyano groups, the nitrile-substituted
bipyridine (L4) is surprisingly “hard” to reduce. The cyano group
is a very strong σ acceptor and somewhat of a weaker π
acceptor.10,11 These observations in conjunction with the better
regression using σpara tend to demonstrate that the dominant
factor on the value of E1/2(L0/1-) is more a delocalization effect
than an inductive one.
As for the free ligands, the redox potentials of the ligand-
based reductions of the ruthenium complexes, E1/2[Ru(L)32+/1+],
were determined by cyclic voltammetry. These values, as well
as the E1/2 values of the free ligands reduction, are listed in
Table 2. Ru(L1)3, Ru(L2)3, Ru(L3)3, Ru(L4)3 and Ru(L5)3 display
reversible electrochemistry, with only the first set of three one-
electron reductions being accessible for Ru(L2)3 and Ru(L3)3
because of the potential window of acetonitrile. The six ligand-
based one-electron reductions (two sets of three) are observable
for Ru(L1)3, Ru(L4)3, and Ru(L5)3. Noticeably, there is a small
potential separation between the two sets of three waves for
Ru(L5)3. Most probably, this is caused by the small potential
separation between the two reductions of the free ketone ligand
(i.e., 230 mV for L5 compared to 390 mV for L1).
π-back-bonding provided by the π-antibonding orbital of the
2+/1+
ligands. The fact that the difference between E1/2[Ru(L)3
]
and E1/2[L0/1-] decreases as E1/2[L0/1-] becomes more positive
tends to prove that the injection of an electron in the bound
ligand decreases its π-back-bonding ability more than it
improves its σ-donating ability.4a This trend is even more
accentuated in the case of the cyano-substituted bipyridine (L4).
As can be seen in Table 2, ∆E1/2s for L4 (i.e., the difference
between E1/2[Ru(L4)32+/1+] and E1/2[L40/1-]) is smaller than
expected from the linear trend defined by the other entries in
the table. This is consistent with the fact that the cyano group
is much more of a σ-acceptor than a π-acceptor. The added
electron in the bipyridine does not improve the σ-donating ability
of this ligand as much as it does for the other bipyridines
considered, as a result of the strong attractive inductive effect
of the cyano group which lowers, to some extent, the electron
density on the ring.
Electrochromic Behavior of Ru(L1)3, Ru(L2)3, Ru(L3)3, Ru-
(L4)3, and Ru(L5)3. The next goal of this study was to evaluate
the effect that different substitution in the 5- and 5′ -positions
of the bipyridine had on the electrochromic properties of the
corresponding ruthenium complexes. Bulk electrolysis was, thus,
performed on each of the compounds. By applying potentials
which maximized the concentration of each single redox species
and letting the solution reach equilibrium, one can visually
observe the color of the species in solution. As displayed in
Table 3, all complexes synthesized displayed some pleasant
electrochromism. The color ranges of the new complexes,
however, are not as wide as that for Ru(L1)3 and the colors are
not as sharply defined.4a The 2- and 3- oxidation states of
As with the free ligands, attempts at drawing a linear
correlation between E1/2[Ru(L)32+/1+] and the Hammett constants
of the substituents on the bipyridines yielded poor results (r2 )
0.851). However, as can be seen in Figure 5, there is a good
linear correlation between E1/2[Ru(L)32+/1+] and E1/2(L0/1-) (r2

Electrochromic Ruthenium(II) Trisbipyridine Complexes
J. Phys. Chem. A, Vol. 103, No. 31, 1999 6267
TABLE 3: Visual Observation of the Electrochromic
Behavior of Selected Ruthenium Complexes
synthesized display some multicolor electrochromism, but none
were superior in this regard to the previously studied complex
Ru(L1)3.
formal
RuL3
redox state
L1
L2
L3
L4
L5
Acknowledgment. Support of this work by the National
Science Foundation (CHE-971408) is gratefully acknowledged.
+2
+1
0
-1
-2
-3
orange orange orange
purple wine red gray-blue purple
red-orange
red-orange
red-brown
purple-brown
gray-blue
green
blue
green blue
brown
purple
turquoise blue
green turquoise
aquamarine
Supporting Information Available: Cyclic voltammograms
of selected compounds shown in Figure 1 and their ruthenium
complexes. This material is available free of charge via the
Internet at http://pubs.acs.org.
red
brown-green purple
Ru(L2)3 and Ru(L3)3 could not be accessed, once again, because
of the potential window of the solvent.
References and Notes
(1) Juris, A.; Balzani, V.; Barigeletti, F.; Campagna, S.; Belser, P.;
Von Zelewsky, A. Coord. Chem. ReV. 1988, 84, 85.
Conclusion
Five new and one previously known bipyridine ligands
containing electron-withdrawing groups in the 5- and 5′-
positions of the rings have been synthesized. With one exception,
their corresponding ruthenium tris complexes were also syn-
thesized. A rough linear correlation between the reduction
potentials of the free ligands and the Hammett constants of the
electron-withdrawing groups has been established. A much
better linear correlation between the potential of the first
reduction of the free ligands and the first ligand-based reduction
potential of the corresponding ruthenium tris complexes was
found. The cyano-substituted ligand fits this correlation much
more poorly than the rest. Most likely, this difference arises
from the fact that cyano groups are much better σ-acceptors
than they are π-acceptors; thus, their effect on the ligand π*-
orbital is smaller and less direct than those of the other
substituents considered here. All of the ruthenium complexes
(2) Abrun˜a, H. D.; Bard, A. J. J. Am. Chem. Soc. 1982, 104, 2641.
(3) Elliott, C. M.; Pichot, F.; Bloom, C. J.; Rider, L. S. J. Am. Chem.
Soc. 1998, 120, 6781.
(4) (a) Elliott, C. M.; Hershenhart, E. J. J. Am. Chem. Soc. 1982, 104,
7519. (b) While their electrochemistry is similar, the spectral behaviors of
Ru(L1)3 and its analogue substituted in the 4- and 4′-positions are quite
different. Complete spectra of these two complexes are presented in ref 4a
along with a discussion of their differences.
(5) Elliott, C. M. J. Chem. Soc., Chem. Commun. 1980, 261.
(6) March, J. AdVanced Organic Chemistry, 3rd ed.; John Willey &
Sons: New York, 1985; pp 932-933.
(7) Posner, G. H.; Whitten, C. E.; McFarland, P. E. J. Am. Chem. Soc.
1972, 94, 5106.
(8) Hansch, C.; Leo, A.; Unger, S. H.; Kim, K. H.; Nikaitani, D.; Lien,
E. J. J. Med. Chem. 1973, 16, 1207.
(9) Hammett, L. P. J. Am. Chem. Soc. 1937, 59, 96.
(10) Swain, C. G.; Lupton, E. C. J. Am. Chem. Soc. 1968, 90, 4328.
(11) Bromilow, J.; Brownlee, R. T. C.; Lopez, V. O.; Taft, R. W. J.
Org. Chem. 1979, 44, 4766.