Ruthenium complexes with naphthyridine ligands. Synthesis, characterization and catalytic activity in oxidation reactions

Article abstract of DOI:10.1039/a702983g

New ruthenium complexes with 1,8-naphthyridine (napy) or derivatives thereof as ligands have been prepared and characterized. Three groups of complexes were obtained. The first consists of three dinuclear ruthenium complexes with two ligands (1,8-naphthyridine and pyridopyrazine) co-ordinated to two ruthenium ions in a bridging fashion. The second consists of two ruthenium dinuclear complexes having one ligand (2,7-dimethoxy-or 2,7-dichloro-1,8-naphthyridine, abbreviated to dmnapy and dcnapy respectively) co-ordinated to two ruthenium atoms. Proton NMR spectra for both complexes in aqueous solution and in acetonitrile revealed the conversion of a symmetrical form, suggesting dinucleating behaviour of the ligand, into an asymmetrical form, suggesting mononucleating behaviour of the ligand. The third group consists of a mono- and a di-nuclear complex with the ligand 2,7-di(phenylazo)-1,8-naphthyridine. The catalytic activity of the novel naphthyridine complexes in oxidation reactions has been studied. The catalytic oxidation of alcohols and the epoxidation of trans-stilbene were examined and the different reaction rates and selectivities are discussed in a comparative way. The active high-valent species resulting from the [Ru₂(napy)₂(H₂O)₄Cl(OH)] ⁴⁺ complex is discussed in more detail.

Full text of DOI:10.1039/a702983g

View Article Online / Journal Homepage / Table of Contents for this issue
Ruthenium complexes with naphthyridine ligands. Synthesis,
characterization and catalytic activity in oxidation reactions†
Alexandra E. M. Boelrijk, Thomas X. Neenan and Jan Reedijk*
Leiden Institute of Chemistry, Gorlaeus Laboratories, Leiden University, PO Box 9502,
2300 RA Leiden, The Netherlands
New ruthenium complexes with 1,8-naphthyridine (napy) or derivatives thereof as ligands have been prepared
and characterized. Three groups of complexes were obtained. The first consists of three dinuclear ruthenium
complexes with two ligands (1,8-naphthyridine and pyridopyrazine) co-ordinated to two ruthenium ions in a
bridging fashion. The second consists of two ruthenium dinuclear complexes having one ligand (2,7-dimethoxy-
or 2,7-dichloro-1,8-naphthyridine, abbreviated to dmnapy and dcnapy respectively) co-ordinated to two
ruthenium atoms. Proton NMR spectra for both complexes in aqueous solution and in acetonitrile revealed the
conversion of a symmetrical form, suggesting dinucleating behaviour of the ligand, into an asymmetrical form,
suggesting mononucleating behaviour of the ligand. The third group consists of a mono- and a di-nuclear complex
with the ligand 2,7-di(phenylazo)-1,8-naphthyridine. The catalytic activity of the novel naphthyridine complexes
in oxidation reactions has been studied. The catalytic oxidation of alcohols and the epoxidation of trans-stilbene
were examined and the different reaction rates and selectivities are discussed in a comparative way. The active
high-valent species resulting from the [Ru2(napy)2(H2O)4Cl(OH)]4ϩ complex is discussed in more detail.
When an alcoholic group is part of a polyfunctional mol-
ecule or a molecule that is sensitive towards acidic or basic
reagents the choice of effective and selective oxidants is rapidly
narrowed.1 Therefore, research leading to new mild and select-
ive oxidation systems is an important area. In addition to
mononuclear ruthenium polypyridyl complexes,2 dinuclear
complexes are of major interest as catalytic oxidants.3
to the formation3 of two molecules of cis-[Ru(H2O)2(bpy)2]2ϩ
.
Our aim has been to synthesize dinuclear complexes with a
ligand that bridges the two ruthenium centres, thereby improv-
ing the stability of the dinuclear complex, and further con-
straining the steric effects.
1,8-Naphthyridine (napy) and derivatives have been studied
before as ligands forming di-9 and mono-nuclear complexes.10,11
Interest in these ligands arose from the desire to study the
unusual manifestations caused by the formation of four-
membered chelate rings when the two nitrogen sites bind to one
central metal. The result was the characterization of complexes
with abnormally high co-ordination numbers, which are
favoured as a result of the small ‘bite’ of 2.2 Å for the 1,8-
naphthyridine ligands. An example of this kind of co-
ordination behaviour11 is the complex [Ru(napy)4]2ϩ. The
ligand has also been reported9 to be able to form dinuclear com-
plexes with the general formula [M2(µ-napy)2(µ-X)2Y2], where
Dinuclear ruthenium complexes of the type [{Ru(H2O)-
L2}2O], where L is 2,2Ј-bipyridine (bpy) or a ring-substituted
analogue, display unique catalytic capabilities for water oxid-
ation, both electrochemically and in reactions with strong
oxidants in homogeneous solution.4–6 These reactions have con-
siderable intrinsic interest, e.g. in understanding how redox
metal clusters can overcome kinetic barriers imposed by react-
ant non-complementarity.3 The presence of two ruthenium
centres with aqua ligands and multiple redox states with poten-
tials suitable for oxygen evolution makes these complexes
attractive candidates for water oxidation. Besides being of
interest for the oxidation of water, these ruthenium dimers are
also potential catalysts for oxidation of organic substrates.7
The [(bpy)2(H2O)RuIIIORuIII(H2O)(bpy)2]4ϩ dimer has es-
pecially received increasing attention as an oxidation catalyst in
this respect.2–7 This dinuclear complex has been shown to be an
effective oxidant towards a variety of organic substrates, at least
after being electrochemically oxidized to the RuIVRuV oxidation
state. The enhanced thermodynamic oxidizing strength of the
RuIVRuV dinuclear complex leads to accelerated rates of oxid-
ation of a series of organic substrates, when compared to the
monomeric analog [RuIVO(bpy)2(py)]2ϩ (py = pyridine). The
extent of the rate enhancement is substrate dependent, i.e. 300
times as fast for certain alcohols, 60 times for ethanol and 40
times for olefins.7 Quantitative catalytic conversion of organic
substrates, however, cannot be performed with this complex
because of its instability. Although the dinuclear ruthenium
cation [(bpy)2(H2O)RuIIIORuIII(H2O)(bpy)2]4ϩ is stable in basic
solution, its high-oxidation-state forms, RuIIIRuIV and RuIV-
RuV, are unstable at high pH, undergoing self-reduction by
oxidation of ligands.8 Under acidic conditions the dimer is
known to undergo reductive cleavage of the RuIII2 form leading
M is CuII and X and Y are ClϪ or CO3Ϫ, or [M2(µ-napy)4]nϩ
,
where M is FeII, CdII or HgII. Therefore, napy, with a basicity
nearly equivalent to that of pyridine, can be considered to be
both a potentially dinucleating ligand of the carboxylate type
and a mononucleating ligand.
We previously communicated12 the synthesis, crystal
structure and preliminary catalytic reactivity of a dinuclear
ruthenium() complex with 1,8-naphthyridine, [Ru2(napy)2-
(H2O)4Cl(OH)][ClO4]4 2. In this paper the syntheses and
full characterization (NMR, Fourier-transform IR, UV/VIS
spectroscopy and elemental analysis) of this and several other
novel ruthenium complexes of napy derivatives which contain
electron-withdrawing or -donating groups are described. The
performance of these complexes as oxidation catalysts is
investigated, and the different reactions are discussed in a com-
parative way. The active high-valent species resulting from the
[Ru2(napy)2(H2O)4Cl(OH)]4ϩ complex is discussed in significant
detail.
Results and Discussion
Synthesis of the ligands
The synthesis of 2,7-dichloro-1,8-naphthyridine followed the
† Supplementary data available (No. SUP 57294, 3 pp.): elemental
analyses and IR data. See J. Chem. Soc., Dalton Trans., 1997, Issue 1.
same general procedure reported earlier by Newkome et al.13
J. Chem. Soc., Dalton Trans., 1997, Pages 4561–4570
4561

View Article Online
Chromatography of the product on neutral alumina yielded
N
N
2,7-di(phenylazo)-1,8-naphthyridine as an orange powder in
22% yield from the diamino starting material.
N
N
N
napy
ppyz
HO
Synthesis and characterization of the ruthenium complexes
4
5
The syntheses of the ruthenium complexes were performed
in a straightforward way, although some problems were
encountered. In general the choice of reaction solvent proved to
be quite important for the ultimate purity of the complexes.
Ruthenium complexes with the ligands napy, ppyz and danapy
were prepared in methanol. Earlier attempts to perform the
complex syntheses in dimethylformamide instead of methanol
yielded irreproducible results or products that contain carbon
monoxide (CO co-ordinated to Ru, IR bands 1790 cmϪ1 for
the napy derivative and 1820 cmϪ1 for the ppyz derivative).
Elemental analyses of all complexes proved to be satisfactory
and are summarized in SUP 57294.
6
7
3
2
H2SO4/NaNO2
HO
N8
N1
NH2
N
N
OH
PCl5
Cl
Cl
N
N
dcnapy
NaOMe
NH3
1
The H NMR spectra of the complexes are characterized
by deshielding of the aromatic protons relative to the free
heterocycle. This change in the proton environment upon co-
ordination results from a decrease in electron density induced
by the positive ruthenium centre. Proton NMR data for the
complexes with napy and ppyz prepared in methanol showed
that highly symmetrical compounds were formed. For both
complexes only one set of ligand signals (two doublets and a
triplet for the napy complex, and five signals for the ppyz com-
H2N
N
N
N
NH2
MeO
N
N
OMe
dmnapy
nitrosobenzene
1
plex) was found in the H NMR spectra. Elemental analyses
agree with ruthenium complexes in which the atom ratios are in
accordance with Ru2(napy)2Cl4 1 and Ru2(ppyz)2(dmso)2Cl4 3
(dmso = dimethyl sulfoxide).
N
N
N
N9
N10
With the ligand danapy two different ruthenium complexes
could be obtained. After refluxing the ligand and RuCl3(H2O)3
in methanol a red complex precipitated which showed an
11
12
15
14
13
1
asymmetrical H NMR spectrum with an integral of 14 pro-
danapy
tons. Elemental analysis indicates that a mononuclear complex,
Ru(danapy)Cl2 6, was formed. The inequivalence of the H3, H4,
H5 and H6 protons in the NMR spectrum suggests that the
ruthenium atom co-ordinates to N1 or N8 of the naphthyridine
unit and one nitrogen of the azo-group (N10). If the ruthenium
co-ordinated to the two nitrogens of the naphthyridine one
would expect the equivalence of H3 and H6 and H4 and H5.
From the reaction filtrate a green complex could be isolated,
yielding also an asymmetric NMR spectrum, however with an
elemental analysis indicative for a complex with the ratio of
elements as in Ru2(danapy)Cl6 7. In this case no reduction to
RuII had occurred.
When dcnapy and RuCl3(H2O)3 were refluxed in methanol
two different complexes were formed in solution: a symmetrical
complex showing only two doublets at δ 8.49 and 7.76, and a
compound showing four doublets at δ 8.15, 8.06, 7.40 and 6.78
in the 1H NMR spectrum indicating lower symmetry. The latter
spectrum is characteristic of complexes wherein napy acts as a
monodentate ligand.11,15 The ratio of the asymmetric to the
symmetric complex was changed with reaction conditions. The
amount of asymmetric complex increased with longer reaction
times, while that of the symmetric complex could be enhanced
by increasing the ionic strength (i.e. addition of lithium
chloride) of the reaction solution. When the reaction was per-
formed in aqueous solution without LiCl only the asymmetric
complex was formed. Precipitation of the asymmetric com-
pound, or a mixture of asymmetric and symmetric complexes,
was troublesome due to their hygroscopic properties. However,
if the co-ordination reaction of dcnapy was carried out in ethyl
acetate only the symmetric product was formed which could be
isolated by filtration. Elemental analysis indicates that the
formed complex has a ratio of elements as in Ru2(dcnapy)Cl4 5.
The existence of a polymeric chain cannot be excluded. The
fact that only one dcnapy ligand is co-ordinated to two
ruthenium atoms must be due to steric constraints preventing
Scheme 1
It began with the preparation of 2-amino-7-hydroxy-1,8-naph-
thyridine (see Scheme 1); this material was converted into its
diazonium salt which was hydrolysed to form the desired
2,7-dihydroxy-1,8-naphthyridine. Our procedure differs from
Newkome’s in this step only in the work-up and recovery of the
product. Conversion of the hydroxy groups into the desired
2,7-dichloro-1,8-naphthyridine using a mixture of phosphorus
pentachloride and phosphorus trichloride oxide worked well,
even when the procedure was scaled up to twenty-fold over the
original report.13 The intermediates and the final product were
1
easily identified by their H NMR spectra; the unsymmetrical
shifts of the 2-amino-7-hydroxy at δ 7.76 (H4,5), 6.46 (H6) and
6.22 (H3) were converted into the two symmetrical doublets (δ
7.78 and 6.38) for the dihydroxy compound. Conversion into
the dichloro derivative caused these two doublets to shift to δ
8.15 and 7.61. The synthesis of the dimethoxy derivative was
accomplished by refluxing a suspension of the dichloro deriv-
ative with an excess of sodium methoxide in dry methanol. The
spectrum of the isolated product showed resonances for the
aromatic protons at δ 8.21 and 6.94, together with a large
singlet at δ 4.01 for the methoxy protons.
2,7-Di(phenylazo)-1,8-naphthyridine is accessible in two
steps from the readily prepared 2,7-dichloro-1,8-naphthyrid-
ine.13 Reaction of the dichloro compound with ammonia gas in
phenol at 170 ЊC for 20 h was carried out as described by Collin
and Youinou14 and yielded 2,7-diamino-1,8-naphthyridine,
albeit in significantly lower yield (20 versus 94%) than reported
by these authors. Part of the difficulty may originate from
the need of keeping a sufficiently high concentration of
ammonia in the reaction vessel. 2,7-Diamino-1,8-naphthyridine
was treated with nitrosobenzene as described by Collin and
Youinou, although our conditions were somewhat more drastic.
4562
J. Chem. Soc., Dalton Trans., 1997, Pages 4561–4570

View Article Online
Table 1 Electronic absorption spectral data at 298 K of the ruthenium complexes
Complex
Solvent
λmax/nm (ε/Ϫ1 cmϪ1
)
Ru2(napy)2Cl4(H2O)2
[Ru2(napy)2(H2O)4Cl(OH)][ClO4]4ؒ3H2O
dmso
Water
438 (3610), 270 (7610)
393 (9651), 305 (10 968), 258 (24 379)
368 (11 342), 313 (9993), 242 (19 573)
492 (10 930), 323 (12 420), 276 (16 510)
399 (16 651), 323 (34 341)
MeCN
dmso
MeOH
MeCN*
MeOH
MeCN
Water
MeCN
Water
MeCN
Ru2(ppyz)2(dmso)2Cl4(H2O)3
Ru2(dmnapy)Cl4
415 (14 210), 319 (16 210)
Ru2(dcnapy)Cl4
389 (13 844), 320 (30 271), 312 (25 339)
416 (8121), 319 (7212), 311 (6563)
612 (5340), 396 (15 720), 332 (13 810), 297 (8140)
604 (4060), 397 (14 310), 329 (11 520), 315 (10 800)
385 (23 000), 318 (26 200), 308 (17 500)
547 (4600), 385 (12 420), 326 (11 030)
Ru2(danapy)Cl6(H2O)4
Ru(danapy)Cl2(H2O)2
* After equilibration in acetonitrile at room temperature for 2 weeks.
conversion of the dcnapy complex is much faster (5 h for com-
plete conversion) than that of the dmnapy complex (3 d for
complete conversion) in D2O at 80 ЊC. The exact structural
formula for the asymmetric complexes could not be determined
because of problems with isolation as described above. Com-
parison of the NMR data with literature results,11,15 however,
suggests that the naphthyridine derivatives change from di- to
mono-nucleating ligands.
Although attempts were made to prepare the aquated deriv-
atives of complexes 1 to 7, the only reproducible result could be
obtained for the Ru2(napy)2Cl4 complex. The complex Ru2-
(napy)2Cl4 was converted into its aquated derivative, [Ru2-
(napy)2(H2O)4Cl(OH)][ClO4]4ؒ3H2O 2 by refluxing in acetone–
water with an excess of AgClO4. The resulting complex shows
only three signals in the 1H NMR spectrum (two doublets and a
triplet, at δ 8.92, 8.56 and 7.81, respectively) indicating the
existence of a highly symmetric complex. Crystals suitable for
X-ray analysis were obtained and the structure briefly com-
municated before12 (for schematic structure see Scheme 2 upper
left corner). The structure proves the dinuclearity of the com-
plex and explains the high symmetry as observed in the NMR
spectrum. Attempts to convert the other complexes into their
aquated forms have failed so far.
Electronic spectra
Fig. 1 Proton NMR spectra of the aromatic region of the dmnapy
complex 4 in D2O (4.5 m) showing the conversion of the symmetric
complex into the asymmetric complex: (a) starting solution; (b) after
24 h at 80 ЊC; (c) after 48 h at 80 ЊC; (d) after 72 h at 80 ЊC
Staniewicz and Hendricker11 discussed the synthesis and spec-
troscopic properties of a mononuclear tetrakis ruthenium()
complex with 1,8-naphthyridine, [Ru(napy)4][PF6]2. They con-
cluded that its ground state has significant charge-transfer
character and, therefore, metal-to-ligand π bonding contributes
to the stability of the complex. The considerably anodic
potentials of the complex demonstrated the ability of napy to
participate in π-back bonding, and thereby stabilize the
ruthenium() state in the tris(2,7-dimethyl-1,8-naphthyridine)
complex. They also concluded that the mononuclear [Ru-
(napy)4]2ϩ complex undergoes solvolysis upon dissolution in
MeCN with a corresponding change from red to yellow, in-
dicative of a shift in the metal-to-ligand charge-transfer
(MLCT) band. The replacement of napy by MeCN apparently
stabilizes the t2g level of RuII, thereby causing the energy of the
d RuII → π* (napy) transition to move to higher energy.
Table 1 presents the visible and UV maxima for the new
ruthenium complexes. The visible bands for the complexes
described in this paper exhibit high molar absorption and
are assigned as t2g–π* MLCT transitions in analogy with those
of previous ruthenium complexes with 1,8-naphthyridine
ligands.9,10 The spectra of the ruthenium complexes also display
intraligand bands in the UV region, which are shifted compared
to the bands of the free heterocycles. The complexes Ru2-
(napy)2Cl4 and Ru2(ppyz)2(dmso)2Cl4 were not soluble in any
other solvent than dmso and therefore the solvolysis by
acetonitrile could not be studied. The MLCT band of [Ru-
fitting of two chloride (trans oriented) groups of each dcnapy
ligand around the RuII. However, two such ligands in a cis
orientation would be possible.
The results with the ligand dmnapy are very similar to those
with dcnapy. Also with this ligand a mixture of both an asym-
metric and a symmetric complex is obtained in methanol or
water and also here only the symmetric compound is isolated
when the synthesis is done in ethyl acetate (1H NMR spectro-
scopy shows two doublets and a singlet at δ 8.50, 7.25 and 4.23,
respectively). Elemental analysis agrees with a dinuclear or
polymeric complex with the atom ratio Ru2(dmnapy)Cl4 4.
The symmetric complexes obtained in ethyl acetate can be
converted into asymmetric complexes by heating in aqueous
solution. For the dmnapy complex this is illustrated in Fig. 1.
The symmetric complex 4 (obtained in ethyl acetate) was dis-
solved in D2O (4.5 m) and heated at 80 ЊC for 3 d. Each day a
1H NMR spectrum was taken and the spectra clearly show the
clean conversion of the symmetric complex (two doublets) into
an asymmetric complex (four doublets of which two are over-
lapping). In the aliphatic region the intensity of the singlet
signal of the CH3 was also doubled (not shown in this figure). A
difference between the dmnapy and dcnapy complexes appears
to be the rate of conversion into the asymmetric isomer. The
J. Chem. Soc., Dalton Trans., 1997, Pages 4561–4570
4563

View Article Online
(napy)2(H2O)4Cl(OH)][ClO4]4 at 393 nm shifts upon dissolution
in acetonitrile to 368 nm, suggesting water replacement by
acetonitrile. The MLCT band of Ru2(dmnapy)Cl4, 399 nm in
methanol, shifts to 415 nm upon dissolution in acetonitrile
which would indicate replacement of ligands, or chloride, by
acetonitrile. However, this solvolysis occurs in two steps. Upon
dissolution in acetonitrile a new band appears in the spectrum
at 604 nm apart from the MLCT shift. After standing in
acetonitrile for extended time (1–2 weeks, room temperature)
the band at 604 nm disappears, while the MLCT band does not
change. The 1H NMR spectrum of Ru2(dmnapy)Cl4 in CD3CN
shows the slow conversion of the symmetric complex into the
asymmetric complex, already described before when water is
used as the solvent. Dissolution of Ru2(dcnapy)Cl4 in aceto-
nitrile shifts the MLCT band at 389 to 416 nm and has a large
decreasing effect on the intensity of all bands in the spectrum,
the ruthenium complex. The low-frequency spectra of Ru2-
(napy)2Cl4 and [Ru2(napy)(H2O)4Cl(OH)][ClO4]4 also display
new absorptions compared to free napy in this region. The
spectrum of Ru2(napy)2Cl4 shows four intense bands at 339,
306, 261 and 250 cmϪ1, which can be attributed to both Ru᎐N
or Ru᎐Cl stretches (νRu᎐Cl are usually found20 between 280
and 350 cmϪ1). Also five distinct absorbances are found for
[Ru2(napy)2(H2O)4Cl(OH)][ClO4]4 in the far-infrared region
[albeit much weaker than in the spectrum of Ru2(napy)2Cl4], at
224, 247, 253, 279 (which might be assigned to Ru᎐N stretches)
and 357 cmϪ1 (which might originate from Ru᎐Cl᎐Ru stretch),
respectively. To exclude displacement of the aqua groups by
iodide, the IR spectrum was also obtained in a polyethylene
pellet; this did not alter the already observed frequencies in CsI.
Krause and Krause21 assigned the Ru᎐N and Ru᎐Cl vi-
brations of several isomers of the ruthenium complexes with
the ligand 2-(phenylazo)pyridine (azpy), Ru(azpy)2Cl2. Bands
at 308 and 336 cmϪ1 were assigned to Ru᎐Cl stretching modes,
bands around 280, 304 and 376 cmϪ1 to Ru᎐N (azo) modes and
bands around 268 and 358 cmϪ1 to Ru᎐N (pyridine) modes. The
IR spectrum of Ru2(danapy)Cl6 displays bands at 327 and 318
cmϪ1, which might be Ru᎐Cl modes, and a weak band at 269
cmϪ1, which might be tentatively assigned to a Ru᎐N (pyridine)
mode. The mononuclear complex Ru(danapy)Cl2 shows a very
broad absorption at 322 cmϪ1, which could comprise and
include several stretching modes, and a weak absorption at 254
cmϪ1, which might be a Ru᎐N (pyridine) mode.
1
indicating acetonitrile co-ordination. The H NMR spectrum
of Ru2(dcnapy)Cl4 in CD3CN shows a fast conversion of the
symmetric complex into the asymmetric complex (50% conver-
sion within 2 min). The conversions of the symmetric into the
asymmetric compounds are significantly faster in acetonitrile
than in water, showing the stronger co-ordination properties of
acetonitrile. The observed bathochromic shift instead of the
usual hypsochromic shift for both compounds suggests that the
acetonitrile co-ordination destablilizes the ruthenium() state
compared to the original complex.
The spectrum of Ru(danapy)Cl2 in acetonitrile shows an
additional band at 547 nm compared to the spectrum obtained
in water. The original MLCT band of the complex at 385 nm is
not affected by dissolution in acetonitrile. The visible spectrum
of the dinuclear complex with the danapy ligand, Ru2(danapy)-
Cl6, does not seem to be affected upon dissolution of the
complex in acetonitrile. In fact, only a small red shift for the
intraligand band at 297 nm is observed.
Cyclic voltammetry
The reduction potentials, RuIII–RuII, of ruthenium complexes
are known to depend on the presence of back-bonding ligands
in the co-ordination sphere with the potential increasing as
the number of such ligands is increased.22 Qualitatively, this
change of potential can be attributed to a stabilization of the
ruthenium() t2g level by increased back bonding, whereas π
bonding between ruthenium() and pyridine-type ligands is
thought to be insignificant.23 On varying the ligand environ-
ment surrounding the RuII the potential of the couple changes,
thus yielding information about the relative π interaction
between the metal and ligand.
The π-accepting properties of the naphthyridine ligand may
be manipulated by adding substituents to the naphthyridine
rings. It was expected that the π-accepting properties of these
ligands would increase in the order dmnapy < napy < dcnapy
< danapy.24
Unfortunately, it was impossible to obtain cyclic voltam-
mograms of Ru2(napy)2Cl4 and Ru2(ppyz)2(dmso)2Cl2 in
acetonitrile due to their poor solubility. Attempts to measure
the electrochemical response in other solvents like water, dmso
and mixed solvents did not give satisfactory results. The
ruthenium complexes with dmnapy, dcnapy and the mono-
nuclear complex with danapy all gave an irreversible oxidation
wave at ϩ0.98 V vs. Ag–AgCl in acetonitrile–tetrabutyl-
ammonium hexafluorophosphate solution which can be attrib-
uted to chloride oxidation.25 The complex Ru2(danapy)Cl6 did
not show any electrochemical response in acetonitrile or water.
The dcnapy complex showed an additional oxidation wave at
1.26 V vs. Ag–AgCl with a peak separation of 60–70 mV,
depending on the scan rate, indicating a (quasi) reversible one-
electron transfer (Fig. 2). The complex Ru2(dmnapy)Cl4 did not
seem to give this additional response in acetonitrile at first
sight; however, after prolonged standing at room temperature
of the solution an additional wave was observed at 1.28 V vs.
Ag–AgCl with a peak separation of 60–70 mV. As discussed
above, NMR spectroscopy had shown that in acetonitrile the
symmetrical dcnapy complex was immediately converted into
the asymmetrical complex, whereas the symmetric dmnapy
complex only slowly converts into the asymmetric compound.
The observed electrochemical responses around 1.25 V there-
Infrared spectroscopy
Fourier-transform infrared spectra (4000–200 cmϪ1) were
obtained for all complexes and ligands. In SUP 57294 only the
characteristic bands which change substantially upon complex-
ation are listed. Absorptions due to napy from 1600 to 650 cmϪ1
have been previously assigned, while those below 650 cmϪ1 have
been attributed to ligand deformations.16 As reported before,
the change in the position of the skeletal modes of free napy
(1558, 1228, 1105 and 760 cmϪ1) with respect to the complexes
indicates co-ordination.9,10 This was the case for all novel com-
plexes and no evidence of non-co-ordinated ligands was found
in any of the spectra of the complexes. In the spectrum of
Ru2(ppyz)2(dmso)2Cl4 the frequencies at 1019 and 953 cmϪ1
originate from co-ordinated dmso, as the infrared spectrum of
the crude complex of ruthenium with ppyz before recrystalliz-
ing from dmso does not show these frequencies. Bonding of
dmso through the sulfur usually causes17 an increase in νSO
to about 1100 cmϪ1 (from 1055 cmϪ1 for free dmso), whereas a
shift to the lower range of 1000–900 cmϪ1 is indicative of co-
ordination by oxygen.18 Therefore, the dmso in this complex
appears to be bound through the oxygen atom. For all com-
plexes distinct absorbances in the far-infrared region were also
observed compared to the unco-ordinated ligands. However,
the assignment of these frequencies to metal–ligand inter-
actions (Ru᎐Cl, Ru᎐N) by comparison of the spectra of the
unco-ordinated ligands and the ruthenium complexes is dif-
ficult. This is due to the fact that complex formation may acti-
vate ligand vibrations which are inactive in the free state, and
that Ru᎐Cl and Ru᎐N vibrations occur in the same infrared
region.
Staniewicz et al.19 have reported a low-frequency infrared
study on [Ru(napy)4][PF6]2 and could assign four distinct bands
(315, 293, 255 and 227 cmϪ1) to Ru᎐N stretching modes, by
comparing the infrared spectrum of free napy and that of
4564
J. Chem. Soc., Dalton Trans., 1997, Pages 4561–4570

View Article Online
Fig. 2 Cyclic voltammogram of Ru2(dcnapy)Cl4 (1.5 × 10Ϫ3 ) in
MeCN–NBu4ClO4 solution (platinum working electrode, Ag–AgCl
reference electrode, scan rate 100 mV sϪ1
4+
N
N
Cl
H2O
H2O
OH2
OH2
Ru
Ru
O
H
N
N
[Ru2(napy)2(H2O)4Cl(OH)][ClO4]4•3H2O
N
N
N
dmso
Cl
Cl
Cl
dmso
Fig. 3 Spectra showing the pH dependence of the signals of complex 2
in D2O
Ru
Ru
Cl
N
N
N
HClO4. They suggested that this irreversibility arises because of
the precipitation of the dimer onto the electrode surface.
Owing to the difficult electrochemical analysis of the new
complexes it was not possible to draw conclusions about the
effect of the different ligand substituents on the electrochemical
properties of the ruthenium complexes. All six compounds and
their proposed structures are summarized in Scheme 2.
[Ru2(ppyz)2(dmso)2Cl4]•3H2O
R
MeCN
or
H2O
N
N
H2O
R
R
N
N
R
Oxidation catalysis: general observations
Cl
Cl
Cl
Cl
or MeCN
Cl
Cl
Ru
Ru
Ru
Ru
Cl
Cl
Owing to the complicated electrochemical results it was not
possible to study the catalytic oxidation capabilities of the
complexes towards alcohols and sugars by cyclic voltammetry
or by controlled-potential electrolysis as studied before by Gerli
and Reedijk5c for the dinuclear ruthenium bpy complex. There-
fore, the oxidation experiments were performed with chemical
co-oxidants like NaBrO3 for alcohols and O2/aldehyde for
the epoxidation of trans-stilbene. Most attention will be de-
voted to the reactions catalysed by the complex [Ru2(napy)2-
(H2O)4Cl(OH)]4ϩ 2, both because of its resemblance to the
ruthenium–bpy dimer and for its more detailed structural
characterization compared to those of the other new complexes.
Proton NMR spectra of catalyst 2 were taken at different pH.
As can be seen in Fig. 3, the spectrum is best resolved at acidic
pH. At more neutral pH the signals become broad and change
to an apparent paramagnetic spectrum at pH 8.9. Only at
extremely basic solution (pH 13.1), the signals become sharper
again, albeit with the appearance of some additional species.
When the pH of this solution is reversed from 13.1 to 2.2 the
original spectrum, measured at pH 2.2, does not return, indicat-
ing that the conversion is irreversible. The transformation of
the species obtained in acidic condition (pH 2.2) to species
causing broader NMR signals at neutral conditions (pH 6.7) is
fully reversible though. These experiments suggest that at acidic
and neutral pH the complex is stable, which is contrary to the
situation at basic pH.
R = OMe, [Ru2(dmnapy)Cl4]
R = Cl, [Ru2(dcnapy)Cl4]
N
N
N
N
N
N
N
N
N
N
N
N
Cl
Cl
Cl
Cl
Cl
Ru
Cl
Ru
Cl
Ru
OH2
Cl
OH2
[Ru(danapy)Cl2]•2H2O
[Ru2(danapy)Cl6]•4H2O
water molecules may also co-ordinate
Scheme 2
fore seem to originate from the secondary asymmetric com-
plexes.
In aqueous solution an electrochemical response for the com-
plex [Ru2(napy)2(H2O)4Cl(OH)][ClO4]4 could be obtained.
However, this response appeared to be strongly influenced by
adsorption phenomena on the electrode. Variation in solvent
(phosphate, triflate, acetate buffer, acetonitrile), or electrodes
(platinum, glassy carbon) did not solve this problem. Meyer
and co-workers3a addressed the same phenomena discussing the
electrochemical irreversible response of the bpy dimer in 0.1 
A 31P NMR spectrum of complex 2 in phosphate buffer dis-
plays two signals, at δ 4.23 and Ϫ11.25, respectively, shifted
from the signal of non-co-ordinated phosphate, indicating two
J. Chem. Soc., Dalton Trans., 1997, Pages 4561–4570
4565

View Article Online
types of co-ordination of phosphate to the ruthenium core.27
A
1H NMR spectrum of 2 in acetate buffer revealed one methyl
signal at δ 2.61 that could be assigned to co-ordinated acetate,
with an integral showing that two acetate groups are co-
ordinated to one dinuclear complex. These results show that
both acetate and phosphate can co-ordinate to complex 2.
However, these results suggest that two acetate molecules each
replace two water molecules and bridge between the two
ruthenium ions but that phosphate co-ordinates in two different
modes to the ruthenium core, indicating at least other co-
ordination modes than simple bridging.
Table 2 Oxidation of cyclohexanol by NaBrO3, catalysed by several
ruthenium complexes at 60 ЊC for 15 h. Ratio catalyst:substrate:
NaBrO3 is 1:1000:4000; 100% selectivity for cyclohexanone
Conversion (%)
(turnovers)
Catalyst
2
3
4
5
6
7
[Ru2(napy)2(H2O)4Cl(OH)][ClO4]4ؒ3H2O
41
(410)
73
(730)
70
(700)
89
(890)
55
(550)
80
Ru2(ppyz)2(dmso)2Cl4(H2O)3
Ru2(dmnapy)Cl4
Upon addition of a ten-fold excess of an alcohol to complex
2 in phosphate or acetate buffer no changes in the NMR spec-
trum were observed, showing the stability of 2 in the presence
of an alcoholic group and no association between the alcohol
and the complex, at least when the complex is present in the
Ru2(dcnapy)Cl4
Ru(danapy)Cl2(H2O)2
Ru2(danapy)Cl6(H2O)4
RuIII form. Upon addition of an excess of NaBrO3 the three
(800)
2
sharp signals of H2,7, H4,5 and H3,6 disappear and six broad
signals appear in the aromatic region at δ 9.04, 8.87, 8.51, 8.13,
7.48 and 7.21. After addition of an excess of an alcohol (for
instance n-butanol) and stirring at room temperature for 1 d the
three (somewhat broad) signals reappear at δ 9.44, 9.07 and
8.18, albeit shifted downfield compared to those of the original
complex (δ 9.18, 8.62 and 7.83, respectively). This experiment
suggests that the original symmetric complex is oxidized by
NaBrO3 to yield an asymmetric high-valent complex, which is
subsequently reduced by an alcohol to yield again a symmetric
complex.
Table 3 Oxidation of n-butanol by NaBrO3, catalysed by several
ruthenium complexes at room temperature for 15 min. Ratio catalyst:
substrate:NaBrO3 is 1:1000:2000
Conversion (%)
(turnovers)
Catalyst
Products (%)
2
9
Butyraldehyde (1.5)
Butyric acid (7.5)
Butyric acid (35)
(90)
35
(350)
94
(940)
72
(720)
90
3
4
5
6
Butyraldehyde (2.5)
Butyric acid (91.5)
Butyraldehyde (12.5)
Butyric acid (58.5)
Butyraldehyde (2)
Butyric acid (57)
Since the above-described experiments do not give quanti-
tative information about the high-valent oxidation state of
the catalyst, a spectrometric titration was done with CeIV as
oxidant. The absorption spectrum of complex 2 in aqueous
solution is characterized by a visible band at 393 nm which
is characteristic of metal-to-ligand charge-transfer dπ → π*
transitions. The MLCT band shifts to higher energy by 30 nm
in 0.1  CF3SO3H solution. A spectrophotometric titration of 2
with CeIV in 0.1  CF3SO3H shows a two-electron oxidation of
(900)
Propanoic acid (31)
Butyraldehyde (1.5)
Butyric acid (44)
7
89
(890)
Propanoic acid (43.5)
[RuIIIRuIII(napy)2(H2O)4Cl(OH)]4ϩ to the RuIV analogue, with
2
Catalytic oxidation of alcohols
an isosbestic point at 390 nm. A plot of the absorbance of 2 at
384 nm (also observed at other wavelengths) vs. the Ce:Ru
mole ratio has an apparent end-point at Ce:Ru of 2 0.2:1.
Although the decrease is quite linear (correlation coefficient =
0.987), some of the deviation of linearity could be caused by
disproportionation reactions of RuIV2, similar to those of the
dinuclear ruthenium–bpy complex. Addition of an excess of
Complexes 2 through 7 were tested for their catalytic activity in
the oxidation of a primary and a secondary alcohol (Table 2).
Cyclohexanol was chosen as an aliphatic secondary alcohol
and NaBrO3 as the terminal oxidant. All catalysts showed a
catalytic activity for this substrate, yielding only cyclohexanone.
The mononuclear danapy complex appears to be less active
than the dinuclear danapy catalyst.
NaBrO3 to the RuIII species gives a similar absorbance spec-
2
trum [as of the high-valent complex obtained with cerium,
λ (ε × 102) = 254 (344), 300 (288) and 374 nm (109 Ϫ1 cmϪ1)],
however at a slower rate. Addition of an excess of alcohol
to this solution causes the return of the original low-valent
species [λ (ε × 10Ϫ2) = 258 (243), 305 (109) and 393 nm (96 Ϫ1
cmϪ1)].
As a primary alcohol, n-butanol was selected. All complexes
show a high or very high activity towards n-butanol. The
dmnapy complex 4 is especially active with 940 turnovers in 15
min at room temperature. Complexes 2 and 3 show a lower
catalytic activity towards this substrate than do the other com-
plexes (Table 3). A difference in selectivity between the danapy
complexes on one side and the other complexes on the other
is observed. The danapy complexes show, besides alcohol
oxidation activity, also carbon–carbon bond cleavage, while the
other catalysts only yield butyraldehyde and butyric acid. In
contrast to what is found for cyclohexanol (see above) the
mononuclear danapy complex is as active as the dinuclear
danapy complex.
A detailed comparison of the results obtained in oxidations
of these substrates with various primary oxidants in combi-
nation with ruthenium catalysts with literature data is difficult.
A major problem is that reaction conditions (solvent, reaction
time, reaction temperature, amount and type of co-oxidant) are
not comparable. In Table 4 the number of turnovers per hour
of the dcnapy complex towards cyclohexanol is compared to
literature results.29,30 From this comparison one can conclude
that the dcnapy complex seems quite reactive in the sense that it
shows more turnovers per hour than the other complexes. It
should be realized, however, that this high activity is seen at a
To study the mechanistic features of this reaction in some
more detail, cyclobutanol was treated in the catalytic system
(co-oxidant NaBrO3). Over a wide range of pH (3–11) both
cyclobutanone and acyclic products were found as reaction
products at all pH, which indicates that the catalytic system can
operate via both a one- and two-electron transfer in the oxid-
ation step.28 The absorption results and the cyclobutanol
oxidation result suggest that the RuIII complex is oxidized to
2
RuIV2 by NaBrO3 after which this high-valent species can react
either by two subsequent one-electron steps or by a two-
electron step back to RuIII2, thereby oxidizing alcoholic
substrates.
Although catalytic characteristics of the other complexes
have also been studied, a detailed discussion is less relevant
because of limited information about their exact structures in
water. In fact the poor solubility in water and organic solvents
allowed only limited oxidation studies with Ru2(ppyz)2(dmso)2-
Cl4 and Ru2(napy)2Cl4.
4566
J. Chem. Soc., Dalton Trans., 1997, Pages 4561–4570

View Article Online
Table 4 Oxidation of secondary alcohols, catalysed by different ruthenium complexes
Turnovers
Catalyst
T/ЊC
Ref.
per hour
Co-oxidant
Solvent
[RuO2(bpy){IO3(OH)3}]ؒ1.5H2Oa
r.t.
r.t.
r.t.
r.t.
30(a)
30(b)
30(c)
30(d)
30(d)
30(e)
f
8
19
18
20
6
20
59
NBu4IO4
NBu4IO4
mmoc
mmo
mmo
e
CH2Cl2
CH2Cl2
CH2Cl2
MeCN
MeCN
Water
[RuO2(HIO6)2]6Ϫ a
[PPh4][RuO2(O2CMe)Cl2]b
[Ru2O6(py)4]ؒ3.5H2Oa
[HNC5H4But-4][RuO2Cl3(NC5H4But-3)]a r.t.
cis-[Ru(dcbpy)2(OH2)2]2ϩ d
[Ru2(dcnapy)Cl4]a
r.t.
60 ЊC
NaBrO3
Water
a
b
c
d
r.t. = Room temperature. Cyclohexanol as substrate. Cyclooctanol as substrate. N-Methylmorpholine N-oxide. dcbpy = 6,6Ј-Dichloro-2,2Ј-
bipyridine; cyclobutanol as substrate. e Electrooxidation, glassy carbon electrode. This study.
f
Table 5 Epoxidation of trans-stilbene by O2/butyraldehyde, catalysed
by several ruthenium complexes at 40 ЊC for 20 h
higher temperature and another co-oxidant than for the other
complexes and in water instead of acetonitrile or methylene
chloride.
Conversion (%)
(turnovers)
Selectivity for
epoxide (%)
Catalyst
Epoxidation of trans-stilbene
2
42
(420)
9
(90)
18
(180)
95
(950)
91
(910)
38
81
98
73
87
78
58
The catalytic epoxidation of olefins is both an important indus-
trial technology and a useful synthetic method,31 and several
catalytic epoxidation systems are known using RuCl3 with bi-
pyridyl or substituted phenanthrolines.32 The most effective
catalyst to date is that reported by the group of Griffith,30a
[RuO2(bpy){IO3(OH)3}]ؒ1.5H2O, which is able to epoxidize
trans-stilbene with 100% selectivity to trans-stilbene oxide with
249 turnovers in 15 h at 2 ЊC, with NaIO4 as a co-oxidant. From
an environmental and economical point of view dioxygen or
hydrogen peroxide would be better co-oxidants to use. However,
hydrogen peroxide is known to be easily decomposed on
ruthenium and has therefore to be used in a large (100-fold)
excess to yield reasonable amounts of product.33 The use of
molecular oxygen in the liquid-phase synthesis of organic com-
pounds is limited because its triplet ground state precludes reac-
tion with singlet organic compounds. Recently,34 metal (includ-
ing ruthenium)-catalysed epoxidations have been described
with the use of a combination of an aldehyde and dioxygen.
Mechanistically these systems are rather complex as several
species present in the reaction mixture might effect the epoxid-
ation of the olefin.35 The mechanism of epoxidation is generally
believed to proceed by the autoxidation of butyraldehyde to
peracids and alkyl hydroperoxides, which can then be used as
oxygen-transfer agents in the epoxidation of alkenes. Besides
epoxidation a notorious side reaction is oxidative cleavage of
the double bond, yielding two aldehyde molecules. Ruthenium
dioxide without amine ligands present is known selectively
to catalyse oxidative cleavage of carbon–carbon bonds of
terminal and α/β unsaturated carbonyl compounds using
molecular oxygen/aldehyde to give the corresponding carbonyl
compounds.36 It is known that the selectivity of the epoxidation
reaction can be improved by the addition of amine ligands to
the reaction mixture, or by the use of ruthenium amine
complexes.35d
All novel catalysts are tested for their activity in the epoxid-
ation reaction of trans-stilbene. With NaBrO3 or NaIO4 as
co-oxidants only small amounts of epoxide are formed with
benzaldehyde being a major side product. When dioxygen is
used in a free-radical autoxidation better selectivities for the
epoxide are obtained (see Table 5). All complexes proved to be
active for high turnovers (from 90 for the ppyz complex to 950
for the dcnapy complex) in 20 h under mild conditions. It can-
not be excluded that the complexes act as radical initiators
for an autoxidation reaction (epoxidation via RCO3). However
the fact that the different complexes show not only different
catalytic rates but also different selectivities (from 58 to 98%)
suggests more involvement of the ruthenium complexes in the
mechanism than being just a radical initiator.
3
4
5
6
7
(380)
complex is the poorest catalyst and the dcnapy complex (which
is easily converted into a mononuclear complex in aqueous
solution as discussed before) is the most reactive catalyst for
epoxidation with the mononuclear danapy complex as second
best. The dinuclear napy, danapy and dmnapy complexes (the
latter is known to keep its structure in solution as discussed) all
give a lower yield of epoxide.
In conclusion the mononuclear danapy and dcnapy com-
plexes are quite active catalysts for the epoxidation reaction
with a reactivity comparable to the data reported for [RuO2-
(bpy){IO3(OH)3}]ؒ1.5H2O. Their selectivity for the epoxidation
reaction compared to carbon–carbon cleavage is quite good,
although lower than reported30a for [RuO2(bpy){IO3(OH)3}]ؒ
1.5H2O.
Conclusion
Although the complexes described in this paper were originally
synthesized with the aim to obtain a series of dinuclear com-
plexes only different in their electrochemical characteristics,
several types of complexes were isolated. Seven new complexes
which were characterized by NMR, Fourier-transform IR, UV/
VIS spectroscopy and subjected to electrochemical study.
According to their properties they have been divided into three
subgroups.
Group I consists of ruthenium dinuclear complexes having
two ligands co-ordinated to two ruthenium atoms, 3 and 1
and the aquated derivative 2. Group II consists of ruthenium
dinuclear complexes having only one ligand co-ordinated to
two ruthenium atoms, 4 and 5. Proton NMR spectra for both
complexes in aqueous solution and in acetonitrile reveal the
conversion of a symmetrical form, indicating a conversion into
dinucleating behaviour of the ligand, into an asymmetrical
form, suggesting mononucleating behaviour of the ligand.
Group III consists of a mono- and a di-nuclear complex with
the ligand danapy, 6 and 7 (see Scheme 2 for the suggested
structural formulae of the new complexes according to the
obtained analytical data). The electrochemical analysis did not
allow conclusions about the effect of the different ligand sub-
As can be seen in Table 5, the dinuclear complexes appear to
be less reactive than the mononuclear complexes. The ppyz
J. Chem. Soc., Dalton Trans., 1997, Pages 4561–4570
4567

View Article Online
stituents on the electrochemical properties of the ruthenium
complexes.
argon for 4 h, cooled and filtered to remove sodium chloride.
The methanol was removed under reduced pressure. The off-
white solid was dissolved in the minimum volume of methylene
chloride and passed through a plug of neutral alumina. The
solvent was removed and the product recrystallized from the
The novel naphthyridine complexes have been tested for their
catalytic reactivity in the oxidation of cyclohexanol, n-butanol
and the epoxidation of trans-stilbene. All complexes prove to be
quite reactive towards aliphatic primary and secondary alcohol
oxidation. For the primary alcohol oxidation, the danapy com-
plexes show an additional carbon–carbon bond cleavage
reactivity compared to the other complexes. For catalyst 2
NMR and spectrophotometric data suggest that the symmetric
RuIII2 core is oxidized by a two-electron step to an asymmetric
RuIV2 oxo complex, which in turn oxidizes the substrate, thereby
1
minimum volume of methanol, yield 3.6 g, 72.4%. H NMR
[(CD3)2SO vs. SiMe4]: δ 8.21 (d, J = 9, 2 H), 6.94 (d, J = 9 Hz,
2 H) and 4.01 (s, 3 H). M.p. 73 ЊC.
2,7-Di(phenylazo)-1,8-naphthyridine (danapy). A suspension
was prepared of finely ground 2,7-diamino-1,8-naphthyridine
(0.4 g, 2.5 mmol) in an aqueous sodium hydroxide–benzene
mixture (10 g of NaOH in 30 cm3 of water and 15 cm3 of
benzene). To this was added nitrosobenzene (1.6 g, 15.6 mmol)
as a solid in one portion. The green suspension was heated to
50 ЊC for 90 min, cooled and diluted with water (80 cm3).
Extraction with methylene chloride (4 × 100 cm3) yielded a
brown solution, which was dried over MgSO4 and the solvent
removed. The dark solid was chromatographed on neutral
alumina with methylene chloride. A bright green band eluted
quickly from the column and was discarded. A second orange
band was recovered, and the solvent removed to yield 2,7-
di(phenylazo)-1,8-naphthyridine as an orange powder (260 mg,
22%). This material was sufficiently pure to be used without
being reduced again to a symmetric RuIII complex. From
2
the results of the epoxidation reaction of trans-stilbene it can
be concluded that dinuclear complexes (which retain their
dinuclear structure in aqueous solution) are less active than the
mononuclear complexes (or dinuclear complexes that become
mononuclear in aqueous solution). The catalysts have a selectiv-
ity for the epoxide between 73 (complex 7) and 99% (complex
3), using environmental friendly dioxygen/butyraldehyde as
co-oxidant/reductant.
Experimental
Reagents and substrates
1
further purification. H NMR (CDCl3 vs. SiMe4): δ 8.45 (d,
J = 9, 2 H), 8.16 (m, 4 H), 8.09 (d, J = 9, 2 H), 7.60 (dd, J = 3, 4
The compound RuCl3(H2O)x (x is approximately 3) was used
as obtained on a loan scheme from Johnson Matthey. 1,8-
Naphthyridine was prepared as reported by Paudler and
Kress.37 Other chemicals were purchased (analytical grade)
and used without further purification or were prepared as
described below. For analytical measurements Millipore water
was used.
Hz, 4 H) and 7.52 (m, 2 H).
Complex synthesis
Ru2(napy)2Cl4. 1,8-Naphthyridine (225 mg), RuCl3(H2O)3
(480 mg) and LiCl (700 mg) were dissolved in methanol (60
cm3) and refluxed for 5 h. After cooling a precipitate was filtered
off. After the solid was extracted five times with boiling water, it
was dried under reduced pressure yielding a yellow compound
Ligand synthesis
1
1. Yield 937 mg (91%). H NMR [(CD3)2SO vs. SiMe4]: δ 9.88
(br s, 2 H), 8.51 (br s, 2 H) and 7.80 (br s, 2 H).
2,7-Dihydroxy-1,8-naphthyridine. A suspension was prepared
of finely ground 2-amino-7-hydroxy-1,8-naphthyridine (20.0 g,
124 mmol) in concentrated sulfuric acid (200 cm3) in a round-
bottomed flask (500 cm3) containing a stir bar. The flask was
placed in an ice-bath and finely ground sodium nitrite (10.6 g,
124 mmol) added. The solution was stirred for 15 min and
poured over crushed ice. It was made neutral by the addition of
saturated sodium carbonate solution, causing the precipitation
of a yellow-brown solid. The solid was filtered off, washed well
with water and air dried. The yield of the product was 15.2 g
(76%) and though contaminated by about 5% of the starting
[Ru2(napy)2(H2O)4Cl(OH)][ClO4]4. The compounds Ru2-
(napy)2Cl4 (100 mg) and AgClO4 (120 mg) were dissolved in
water–acetone (1:3 v/v, 100 cm3) and refluxed for 2 h. After
cooling, the solution was filtered and reduced in volume to 10
cm3, under reduced pressure. This solution was allowed to stand
at room temperature for at least 1 week under air until green
crystals of complex 2 precipitated. Yield 83 mg (51%). 1H
NMR [D2O vs. dds (4,4-dimethyl-4-silapentane sulfonic acid)]:
δ 8.92 (d, J = 5, 2 H), 8.56 (d, J = 9, 2 H) and 7.81 (t, J = 7 Hz,
2 H).
1
material was sufficiently pure to be used in the next step. H
NMR [(CD3)2SO vs. SiMe4]: δ 7.78 (d, J = 9, 2 H), 6.28 (d, J = 9
Hz, 2 H) and 3.21 (br s, 2 H).
Ru2(ppyz)2(dmso)2Cl4. The compounds RuCl3(H2O)3 (126
mg), pyrido[2,3-b]pyrazine (68 mg) and LiCl (200 mg) were dis-
solved in methanol (20 cm3) and refluxed for 7 h. After cooling,
the precipitate was filtered off and redissolved in the minimum
volume of dmso (3 cm3). This solution was allowed to stand at
room temperature for some days until red crystals of complex 3
precipitated. Yield 40 mg (11%). 1H NMR [(CD3)2SO vs.
SiMe4]: δ 9.52 (m, 1 H), 9.46 (dd, J = 9, 1 H), 9.11 (dd, J = 9, 1
H), 8.56 (t, J = 10 Hz, 1 H) and 8.05 (m, 1 H).
2,7-Dichloro-1,8-naphthyridine (dcnapy). A flask was charged
with finely ground dry 2,7-dihydroxy-1,8-naphthyridine (8.0 g,
50 mmol). The reaction vessel was cooled in ice and phosphorus
pentachloride (24.4 g, 100 mmol) added, followed by phos-
phorus trichloride oxide (16.2 g, 1.5 mmol). After the initial
exothermic reaction had subsided the cooling bath was
removed and the reaction mixture refluxed for 3 h. The reaction
was cooled and ice (≈300 g) was added. The product was filtered
off, washed well with water and air dried to yield 2,7-dichloro-
1,8-naphthyridine (7.3 g, 73.8%) as a solid. This material was
found to be pure according to 1H NMR spectroscopy
[(CD3)2SO vs. SiMe4]: δ 8.15 (d, J = 9, 2 H) and 7.61 (d, J = 9
Hz, 2 H).
Ru2(dmnapy)Cl4. The compounds dmnapy (62 mg) and
RuCl3(H2O)3 (85 mg) were dissolved in ethyl acetate (10 cm3)
and refluxed for 2 h. The brownish red precipitate of complex 4
was filtered off, washed with several portions of ethyl acetate
and cold methanol and dried under reduced pressure. Yield 84
1
mg (49%). H NMR (D2O vs. dds): δ 8.50 (d, J = 9, 2 H), 7.25
(d, J = 9 Hz, 2 H) and 4.23 (s, 6 H).
2,7-Dimethoxy-1,8-naphthyridine (dmnapy). A solution of
sodium methoxide in methanol was prepared by dissolving
sodium metal (2.32 g, 100 mmol) in dry methanol (300 cm3). To
this was added 2,7-dichloro-1,8-naphthyridine (5.0 g, 25 mmol).
The suspension was heated at reflux under an atmosphere of
Ru2(dcnapy)Cl4. This complex was synthesized by the same
procedure as Ru2(dmnapy)Cl4 5. Yield 57%. 1H NMR (D2O vs.
dds): δ 8.49 (d, J = 9, 2 H) and 7.76 (d, J = 9 Hz, 2 H).
4568
J. Chem. Soc., Dalton Trans., 1997, Pages 4561–4570

View Article Online
Ru(danapy)Cl2 and Ru2(danapy)Cl6. The compounds danapy
(110 mg) and RuCl3(H2O)3 (177 mg) were dissolved in methanol
(20 cm3) and refluxed for 2 h. After cooling a red crystalline
precipitate was filtered off and dried under reduced pressure. To
the filtrate, reduced in volume under reduced pressure to 5 cm3,
was added diethyl ether (200 cm3). After standing overnight at
4 ЊC a green microcrystalline product precipitated, which was
filtered off and dried under reduced pressure. Yield red precipi-
without the ruthenium catalyst yielded at most 10% conversion
under the same reaction conditions.
Acknowledgements
This work has been sponsored by the Netherlands Ministry
of Agriculture, Nature Management and Fishery, through its
special programme on Carbohydrate Oxidation, supervised
by the Programme Commission Carbohydrates. One of us
(T. X. N.) thanks the Netherlands Organisation for Scientific
Research for support. The authors are indebted to the EU for a
grant as Host Institute in the EU Programme Human Capital
and Mobility (1994–1997). We thank Johnson Matthey
(Reading, UK) for their generous loan of RuCl3ؒ3H2O.
1
tate, Ru(danapy)Cl2 6, 61 mg (34%). H NMR (CD3OD vs.
SiMe4): δ 8.72 (d, J = 8, 1 H), 8.54 (d, J = 9, 1 H), 8.46 (d, J = 8,
1 H), 8.37 (d, J = 9, 1 H), 8.25 (m, 2 H), 8.08 (br d, J = 7, 2 H),
7.70 (m, 4 H) and 7.55 (t, J = 7 Hz, 2 H). Yield green precipitate,
1
Ru2(danapy)Cl6 7, 22 mg (8.9%). H NMR (CD3OD): δ 8.37
(t, J = 10, 2 H), 8.23 (d, J = 9, 1 H), 8.09 (d, J = 9, 2 H), 7.94
(m, 1 H), 7.90 (m, 1 H), 7.66 (d, J = 8, 2 H), 7.56 (m, 3 H) and
7.44 (d, J = 9 Hz, 1 H).
References
1 K. B. Wiberg, Oxidation in Organic Chemistry part A, Academic
Press, New York, 1965; D. G. Lee, M. van den Engh and
W. S. Trakanowski (Editors), Oxidation in Organic Chemistry part B,
Academic Press, New York, 1973.
2 See, for example, W. P. Griffith, Chem. Soc. Rev., 1991, 21, 179;
L. Roecker and T. J. Meyer, J. Am. Chem. Soc., 1987, 109, 746;
M. S. Thompson and T. J. Meyer, J. Am. Chem. Soc., 1982, 104,
4106; L. Roecker and T. J. Meyer, J. Am. Chem. Soc., 1986, 108,
4067; C. M. Che, W. T. Tang, K. Y. Lee, K. Y. Wong and T. C. Lau,
J. Chem. Soc., Dalton Trans., 1992, 1551.
3 (a) J. A. Gilbert, D. S. Eggleston, W. R. Murphy, D. A. Geselowitz,
S. W. Gersten, D. J. Hodgson and T. J. Meyer, J. Am. Chem. Soc.,
1985, 107, 3855; (b) K. Honda and A. J. Frank, J. Chem. Soc., Chem.
Commun., 1984, 1635; (c) C. D. Ellis, J. A. Gilbert, W. R. Murphy
and T. J. Meyer, J. Am. Chem. Soc., 1983, 105, 4843.
4 J. K. Hurst, J. Zhou and Y. Lei, Inorg. Chem., 1992, 31, 1010;
J.-P. Collin and J. P. Sauvage, Inorg. Chem., 1986, 25, 135.
5 (a) P. Comte, M. K. Nazeeruddin, F. P. Rotzinger, A. J. Frank and
M. Gratzel, J. Mol. Catal., 1989, 52, 63; (b) D. Geselowitz and
T. J. Meyer, Inorg. Chem., 1990, 29, 3894; (c) A. Gerli and J. Reedijk,
J. Mol. Catal., 1994, 89, 101; (d) W. Kutner, J. A. Gilbert,
A. Tomaszewsky, T. J. Meyer and R. W. Murray, J. Electroanal.
Chem., Interfacial Electrochem., 1986, 205, 185; (e) L. Roecker,
W. Kutner, J. A. Gilbert, M. Simmons, R. W. Murray and T. J.
Meyer, Inorg. Chem., 1985, 24, 3784.
6 W. J. Vining and T. J. Meyer, Inorg. Chem., 1986, 25, 2023.
7 S. J. Raven and T. J. Meyer, Inorg. Chem., 1988, 27, 4478.
8 W. Kutner, J. A. Gilbert, A. Tomaszewsky, T. J. Meyer and R. W.
Murray, J. Electroanal. Chem., Interfacial Electrochem., 1986, 205,
185.
9 A. Tiripicchio, M. T. Camellini, R. Uson, L. A. Oro, M. A. Ciriano
and F. Viguri, J. Chem. Soc., Dalton Trans., 1984, 125; R. W.
Brookes and R. L. Martin, Inorg. Chem., 1975, 14, 528; L. Sacconi,
C. Mealli and D. Gatteschi, Inorg. Chem., 1974, 13, 1985;
D. Gatteschi, C. Mealli and L. Sacconi, Inorg. Chem., 1976, 15,
2774; K. Emerson, A. Emad, R. W. Brookes and R. L. Martin,
Inorg. Chem., 1973, 12, 978; C. Mealli and F. Zanobini, J. Chem.
Soc., Chem. Commun., 1982, 97; T. Tsuda, S. Ohba, M. Takahashi
and M. Ito, Acta Crystallogr., Sect. C, 1989, 45, 887.
10 W. R. Tikkanen, C. Kruger, K. D. Bombden, W. L. Jolly, W. C.
Kaska and P. C. Ford, Inorg. Chem., 1984, 23, 3633; J. M. Epstein,
J. C. Dewan, D. L. Kepert and A. H. White, J. Chem. Soc., Dalton
Trans., 1974, 1949; A. Clearfield and P. Singh, Chem. Commun.,
1970, 389; E. L. Enwall and K. Emerson, Acta Crystallogr., Sect. B,
1979, 35, 2562; P. Singh, A. Clearfield and I. Bernda, J. Coord.
Chem., 1971, 1, 29; G. W. Bushnell, K. R. Dixon and M. A. Kahn,
Can. J. Chem., 1978, 56, 450; J. C. Dewan, D. L. Kepert and
A. H. White, J. Chem. Soc., Dalton Trans., 1975, 490; C. Mealli and
L. Sacconi, Acta Crystallogr., Sect. B, 1977, 33, 710; M. A.
Cavanaugh, V. M. Capo, C. J. Alexander and M. L. Good, Inorg.
Chem., 1976, 15, 2615; J. R. Wagner and D. G. Hendricker, J. Inorg.
Nucl. Chem., 1975, 37, 1375; R. L. Bodner and D. G. Hendricker,
Inorg. Chem., 1973, 12, 33.
Analytical methods
1
The apparatus for H NMR, cyclic voltammetry and UV/VIS
spectroscopy has been described before.26 Elemental analyses
were performed by the microanalytical laboratory of the Uni-
versity of Groningen. Ruthenium complexes are notorious for
causing problems in complete combustion analyses. Also in this
case we needed a slightly higher error window (maximum error
of 0.8% versus a normal maximum error of 0.5%) to explain
our data. Fourier-transform infrared spectra (4000–200 cmϪ1
)
were recorded on a Perkin-Elmer spectrometer (Paragon 1000)
in ultrapure caesium iodide (Johnson Matthey) or polyethylene.
Spectrophotometric redox titrations were carried out by adding
aliquots of a 1.8 × 10Ϫ3  solution of CeIV in 0.1  CF3SO3H to
aliquots (0.5 cm3) of a 3.1 × 10Ϫ4  solution of complex 1 in 0.1
 CF3SO3H. The volume was adjusted to 3 cm3 by addition of
0.1  CF3SO3H, and the changes were monitored in the range
200–600 nm. The Ce:Ru mole ratio was varied from 0 to 6:1.
Cerium() solutions in 0.1  CF3SO3H were prepared from
[NH4]2[Ce(NO3)6].
Catalytic procedure for the oxidation of alcohols
All catalytic reactions were conducted under dioxygen. Specific
blank experiments did not show any activity towards the sub-
strate unless stated otherwise. All reaction substrates were
analysed by gas chromatography and found to be satisfactory
compared to commercial samples. Specific procedures have
been described for n-butanol and cyclohexanol oxidation.26
A
general example is as follows: substrate (100 mg) was dissolved
in water (5 cm3) after which 4 molar equivalents of NaBrO3 and
0.001 molar equivalent of ruthenium complex were added. The
reaction mixture was stirred at the desired temperature and for
the desired time after which it was extracted with ether and
methylene chloride (in order to extract the substrate and prod-
ucts which did not contain an acid group), acidified and again
extracted (in order to extract the products which contain an
acid group). The organic layers were dried over MgSO4 and
analysed by gas chromatography. To increase solubility of
Ru2(ppyz)2(dmso)2Cl4, 10% v/v dmso was added to solutions in
which this complex was used as a catalyst.
Catalytic procedure for the epoxidation of trans-stilbene
The conditions for the oxidation of trans-stilbene were as
follows. In the case of NaBrO3 or NaIO4 as a co-oxidant,
trans-stilbene (0.15 mmol), catalyst (0.15 × 10Ϫ3 mmol) and
NaBrO3 or NaIO4 (0.3 mmol) were dissolved in a mixture of
water (2 cm3) and 1,2-dichloroethane (2 cm3). This mixture was
stirred for 24 h at 40 ЊC. In the case of dioxygen as a co-oxidant,
trans-stilbene (0.15 mmol), isobutyraldehyde (54 µl) and
catalyst (0.15 × 10Ϫ3 mmol) were dissolved in a vigorously
stirred mixture of water (2 cm3) and 1,2-dichloroethane (2 cm3).
This mixture was stirred for 24 h at 40 ЊC. Blank experiments
11 R. J. Staniewicz and G. G. Hendricker, J. Am. Chem. Soc., 1977, 99,
6581.
12 A. E. M. Boelrijk, M. M. van Velzen, T. X. Neenan, J. Reedijk,
H. Kooijman and A. L. Spek, J. Chem. Soc., Chem. Commun., 1995,
2465.
13 G. R. Newkome, S. J. Garbis, V. K. Majestic, F. R. Fronczek and
G. Chiari, J. Org. Chem., 1981, 46, 833.
14 J.-P. Collin and M.-T. Youinou, Inorg. Chim. Acta, 1992, 201, 29.
15 H. Schmidbaur and K. C. Dash, J. Am. Chem. Soc., 1973, 95, 4855.
J. Chem. Soc., Dalton Trans., 1997, Pages 4561–4570
4569

View Article Online
16 W. L. F. Armarego, G. B. Barlin and E. Spinner, Spectrochim. Acta,
1966, 22, 117.
17 R. S. McMillan, A. Mercer, B. R. James and J. Trotter, J. Chem.
Soc., Dalton Trans., 1975, 1006.
18 I. P. Evans, A. Spenser and G. Wilkinson, J. Chem. Soc., Dalton
Trans., 1973, 205.
19 R. J. Staniewicz, D. G. Hendricker and P. R. Griffiths, Inorg. Nucl.
Chem. Lett., 1977, 13, 467.
30 (a) A. J. Bailey, W. P. Griffith, A. J. P. White and D. J. Williams,
J. Chem. Soc., Chem. Commun., 1994, 1833; (b) A. El-Hendawy,
W. P. Griffith, B. Piggott and D. J. Williams, J. Chem. Soc., Dalton
Trans., 1988, 1983; (c) W. P. Griffith and J. M. Jolliffe, in Dioxygen
Activation and Homogeneous Catalysis Oxidation, ed. L. I. Simandi,
Elsevier, Amsterdam, 1991, p. 395; (d) A. C. Dengel, A. M.
El-Hendawy, W. P. Griffith, C. A. O’Mahoney and D. J. Williams,
J. Chem. Soc., Dalton Trans., 1990, 737; (e) K. Y. Wong, C. M. Che
and F. C. Anson, J. Electroanal. Chem., Interfacial Electrochem.,
1991, 319, 207.
20 E. A. Seddon and K. R. Seddon, The Chemistry of Ruthenium,
Elsevier, Amsterdam, 1984.
21 R. A. Krause and K. Krause, Inorg. Chem., 1980, 19, 2600.
22 G. M. Brown, T. R. Weaver, F. R. Keene and T. J. Meyer, Inorg.
Chem., 1976, 15, 190.
23 A. Juris, V. Balzani, F. Barigelletti, S. Campagna, P. Belser and
A. von Zelewski, Coord. Chem. Rev., 1988, 84, 85.
31 K. A. Jørgensen, Chem. Rev., 1989, 89, 431; R. A. Sheldon, in
Aspects of Homogeneous Catalysis, ed. R. Ugo, Reidel, Dordrecht,
1981, vol. 4, p. 1.
32 C. Eskenazi, G. Balavoine, F. Meunier and H. Rivière, J. Chem. Soc.,
Chem. Commun., 1985, 111; G. Balavoine, C. Eskenazi, F. Meunier
and H. Rivière, Tetrahedron Lett., 1984, 25, 3187.
33 J. M. Fisher, A. Fulford and P. S. Bennet, J. Mol. Catal., 1992, 77,
229.
34 T. Mukaiyama, in The Activation of Dioxygen and Homogeneous
Catalytic Oxidation, eds. D. H. R. Barton, A. E. Martell and D. T.
Sawyer, Plenum, New York, 1993, p. 133; S. I. Murahashi, Y. Oda
and T. Naota, J. Am. Chem. Soc., 1992, 144, 7913.
35 (a) R. A. Sheldon and J. K. Kochi, Metal Catalyzed Oxidations
of Organic Compounds, Academic Press, New York, 1981; (b)
T. Mukaiyama, T. Yamada, T. Nagata and K. Imagawa, Chem.
Lett., 1993, 327; (c) T. Yamada, K. Imagawa, T. Nagata and
T. Mukaiyama, Chem. Lett., 1992, 2231; (d) G. A. Barf and
R. A. Sheldon, J. Mol. Catal., 1995, 102, 23 and refs. therein.
36 K. Kaneda, S. Haruna, T. Imanaka and K. Kawamoto, J. Chem.
Soc., Chem. Commun., 1990, 1467.
24 K. J. Takeuchi, G. J. Samuels, S. W. Gerstein, J. A. Gilbert and
T. J. Meyer, Inorg. Chem., 1983, 22, 1407; C. M. Che and W. H.
Leung, J. Chem. Soc., Chem. Commun., 1987, 1376; S. Goswami,
A. R. Chakravarty and A. Chakravorty, Inorg. Chem., 1983, 22, 602.
25 Handbook of Chemistry and Physics, ed. D. R. Lide, CRC Press,
New York, 75th edn., 1991.
26 (a) A. E. M. Boelrijk and J. Reedijk, J. Mol. Catal., 1994, 89, 63;
(b) A. E. M. Boelrijk, Recl. Trav. Chim. Pays-Bas, 1994, 113, 411;
(c) A. E. M. Boelrijk, A. M. J. Jorna and J. Reedijk, J. Mol. Catal.,
1995, 103, 73; (d) A. M. J. Jorna, A. E. M. Boelrijk, H. J. Hoorn
and J. Reedijk, Reactive and Functional Polymers, 1996, 29, 101; (e)
A. E. M. Boelrijk, H. Dorst and J. Reedijk, Recl. Trav. Chim.
Pays-Bas, 1996, 115, 536.
27 D. G. Gorenstein, in Phosphorus-31 NMR, Principles and
Applications, ed. D. G. Gorenstein, Academic Press, Orlando, FL,
1984, ch. 1.
28 J. Rocˇek and A. E. Radkowski, J. Am. Chem. Soc., 1973, 95, 7123.
29 W. P. Griffith, Chem. Soc. Rev., 1992, 21, 179 and refs. therein.
37 W. W. Paudler and T. J. Kress, J. Org. Chem., 1967, 32, 832.
Received 1st May 1997; Paper 7/02983G
4570
J. Chem. Soc., Dalton Trans., 1997, Pages 4561–4570