Stereospecific templated synthesis of a triruthenium butadiyne-linked cyclic porphyrin trimer

Article abstract of DOI:10.1039/a605847g

A bis(ethynyl)-substituted porphyrin monomer of Ru(CO) has been prepared and oligomerised by Glaser-Hay coupling to give a cyclic butadiyne-linked trimer. Use of tri(4-pyridyl)triazine as template during the coupling ensured that only a single stereoisomer of the trimer was formed: all three CO groups are forced to be outside the cavity, leaving three potential ruthenium-binding sites facing into the cavity.

Full text of DOI:10.1039/a605847g

View Article Online / Journal Homepage / Table of Contents for this issue
Stereospecific templated synthesis of a triruthenium butadiyne-linked
cyclic porphyrin trimer
Valérie Marvaud, Anton Vidal-Ferran, Simon J. Webb and Jeremy K. M. Sanders*,†
Cambridge Centre for Molecular Recognition, University Chemical Laboratory,
Lensfield Road, Cambridge CB2 1EW, UK
A bis(ethynyl)-substituted porphyrin monomer of Ru(CO) has been prepared and oligomerised by Glaser–Hay
coupling to give a cyclic butadiyne-linked trimer. Use of tri(4-pyridyl)triazine as template during the coupling
ensured that only a single stereoisomer of the trimer was formed: all three CO groups are forced to be outside the
cavity, leaving three potential ruthenium-binding sites facing into the cavity.
R
R
One of the key aims of supramolecular chemistry is to create
enzyme mimics capable of recognition and catalysis. The first
major milestone of our research in this area 1 was the controlled
synthesis2 of a butadiyne-linked cyclic Zn3 trimer and related
dimers and tetramers using templated Glaser–Hay coupling of
preformed porphyrin (por) monomers [equation (1)]. This
NH
N
NH
N
X
X
R
R
I
R = CH2CH2CO2Me
X = Br or I
᎐
᎐
᎐
᎐
᎐
por᎐C᎐C᎐H + H᎐C᎐C᎐por → por᎐C᎐C᎐C᎐C᎐por (1)
᎐
᎐
᎐
trimer has proved to be a versatile object for molecular recogni-
tion, binding a range of amine ligands,3 stereoselectively
accelerating an exo-Diels–Alder reaction,4 and also catalysing
acyl transfers.5 However, it was an important part of our strat-
egy that we should create a series of receptors using the same
diarylporphyrin monomer I as a building block while exploring
different metal recognition properties and a range of cavity
shapes and sizes.6,7
CO
R
R
N
N
Ru
N
N
X
X
R
R
MeOH
1
2
3
X = I
X = C≡C–SiMe3
X = C≡CH
In this paper we expand our repertoire of recognition build-
ing blocks to include the diamagnetic d6 ruthenium carbonyl
porphyrin 1, and describe the stereospecific templated synthesis
via 2 and 3 of the triruthenium complex 4. Ruthenium porphy-
rins have several potentially attractive properties for molecular
recognition: (a) the lifetimes of amine–Ru(CO)–porphyrin
complexes are much longer than their zinc analogues, so that
metal complexation and decomplexation are in slow exchange
on the NMR chemical shift time-scale;8 this makes the spectra
much easier to interpret than in the zinc case and allows direct
measurement of the co-operativity of binding of ruthenium
porphyrins to bidentate ligands;9 (b) the ruthenium centre also
binds soft sulfur and phosphorus ligands, greatly expanding
the range of substrates that can be recognised;10 (c) the exterior
face of the cavity can in principle be selectively blocked with
carbonyl groups (this would have the effect that ligands could
only bind inside the cavity, thus simplifying binding and kinetic
analyses); the CO group is also spectroscopically useful, being
sensitive to the trans ligand; (d) in a mixed-metal porphyrin
CO
R
R
N
N
N
Ru
N
R
R
N
N
N
R
R
N
N
N
N
N
Hβ
N
N
R
Ru
R
R
R
Ru
oligomer, a Ru᎐O porphyrin could be used to oxidise selectively
Hα
᎐
N
N
N
N
OC
a
substrate bound to another porphyrin moiety in the
CO
molecule.11
Unfortunately the tri(4-pyridyl)triazine template is so strong-
ly held in complex 4 that it cannot be cleanly removed; how-
ever, the monomer building block 1 and this template approach
have more recently been combined with our stepwise oligopor-
phyrin synthesis1,6 to give a mixed-metal Zn2Ru trimer where
the template can be removed to leave a catalytically competent
cavity.12
R
R
4
Results and Discussion
Our standard route2 to metallated cyclic oligomers involves
construction of the cavity via Glaser coupling of preformed
† E-Mail: jkms@cam.ac.uk
J. Chem. Soc., Dalton Trans., 1997, Pages 985–990
985

View Article Online
zinc porphyrin building blocks in the presence of a suitable
template; for a trimer, tri(4-pyridyl)triazine is our usual tem-
plate. The template and zinc are then removed with dilute acid
to afford the free-base trimer which may be remetallated to
generate binding sites. All attempts to insert ruthenium into this
free-base trimer failed to give satisfactory products in this final
step, apparently due to partial hydrogenation of the acetylenic
links. Similar results were obtained when ruthenium insertion
into an acetylenic monomer was attempted. We therefore pro-
tected the triple bonds of the free-base trimer by complexation
with six Co2(CO)6 groups13 and subsequently succeeded in the
ruthenium metallation, but all attempts at deprotection using
ammonium cerium nitrate,13 iron() nitrate13 or tetrabutylam-
monium fluoride14 were unsuccessful, the porphyrins them-
selves being destroyed.
CO
Ru
X
X
N
X
CO
Ru
X
A new synthetic strategy was therefore required, the principle
of which was to introduce the ruthenium into the porphyrin
monomer before the acetylenic linkers; the last step would then
be the formation of the cyclic trimer. Initially the bromopor-
phyrin I (X = Br) was synthesized by standard procedures2,15
but this proved insufficiently reactive for the Pd-catalysed coup-
ling that was required for attachment of the acetylene linkers,
so the iodo derivative I (X = I) was used for all subsequent
work on ruthenium porphyrins. Insertion of ruthenium into I
was accomplished using a modification of the published 16 pro-
cedure to give complex 1. Palladium-catalysed coupling of 1
N
H2
X
X
CO
Ru
᎐
with Me Si᎐C᎐CH then yielded the protected acetylenic
N
᎐
3
Hα
Hβ
monomer 2; the Me3Si protecting groups were removed using
tetrabutylammonium fluoride in refluxing chloroform to give 3.
Oxidative coupling of compound 3 in the absence of tem-
plate gave, as expected, a multiplicity of products: these could
be identified by mass and NMR spectroscopy (with and with-
out added complementary ligands) as dimers, trimers and
tetramers with CO randomly placed inside or outside the cavity,
but it was not possible to separate cleanly any individual prod-
ucts. However, using 0.35 equivalent template (per monomer
unit) gave in 32% isolated yield the ruthenium carbonyl trimer
complex 4 with the three carbon monoxide groups outside the
cavity and template inside;‡ this was fully characterised as
described below. As expected, increasing the proportion of
template decreased the yield of desired trimer, and led to pro-
duction of some material with internal CO groups. It is not
clear whether the templating involves trapping of linear inter-
mediates as observed in zinc porphyrin analogues18 or pre-
assembly of three monomer units onto a single template before
coupling takes place, which is the mechanism employed in more
strongly binding situations.19 Removal of the template from the
final complex 4 proved very difficult, the porphyrin being sensi-
tive to strong acid and oxidation, and the ruthenium–carbonyl
bond sensitive to photochemical excitation. This difficulty in
removing the template by protonation, by comparison with the
ease of its removal from porphyrin trimers containing only a
single Ru᎐N bond, is compelling evidence for the presence of
multiple strong binding sites and therefore for binding within
the cavity.
Hγ
Scheme 1 Three exchanging atropisomers of a ruthenium porphyrin;
cis–trans isomerisation is faster than complexation–decomplexation
1
Fig. 1 The 250 MHz H NMR spectrum of compound 4 in CDCl3
solution
cally in the free base or zinc forms, but in the ruthenium series
with bound pyridine the three different environments (Scheme
1) give rise to multiple signals for most of the porphyrin reson-
ances; the shift differences between atropisomers do not exceed
0.09 ppm (H2), and for most signals are comparable with the
width of the multiplet. Signals due to bound pyridine also
always appear as complex ill resolved multiplets due to this
overlap of signals from different atropisomers. The upfield
shifts experienced by the protons of bound pyridine are sub-
stantially larger than in the zinc cases: for Hα ∆δ is 7.7 rather
than 6.0 ppm, and for Hβ it is around 2.3 rather than 1.9 ppm.
There is no evidence from other resonances that this is due to a
larger ring current in the ruthenium case, so it is almost
certainly due to the fact that the Ru is in the plane of the
Spectroscopic characterisation
For the products 1–3, addition of small amounts of pyrid-
ine to CDCl3 solutions helped solubility and improved NMR
spectral resolution; it appears that the ruthenium porphyrin
always requires a sixth ligand, and in the absence of added
ligand either porphyrin–porphyrin aggregation or weak inter-
molecular ligation of the sidechain ester occurs. The monomers
1–3 exist as a mixture of slowly exchanging cis- and trans-
atropisomers; these are not readily distinguished spectroscopi-
‡ Similarly, Glaser coupling in the presence of 0.5 equivalent of 4,4Ј-
bipyridyl gave the corresponding ruthenium carbonyl dimer stereo-
specifically. This compound and its stereoisomers are described
elsewhere.17
porphyrin (or even slightly displaced towards the CO group 20
pulling the ligand into the porphyrin ring current.
)
986
J. Chem. Soc., Dalton Trans., 1997, Pages 985–990

View Article Online
For cyclic oligomers the influence of the porphyrin ring cur-
rent allowed us to identify the site of the ligand, either inside or
outside the cavity. As an example, the H NMR spectrum of
Molecular modelling was carried out on a Silicon Graphics
Indigo workstation using CERIUS2 Version 2.0 (BIOSYM/
Molecular Simulations) with the UNIVERSAL 1.01 force
field;21 atom and bond parameters were taken directly from the
program database. Electrostatic charge calculations were per-
formed before minimisation using the charge-equilibration
method.22 Structures were minimised using the conjugate-
gradient algorithm with a root mean square force convergence
value of 0.01 kcal molϪ1 ÅϪ1 (cal = 4.184 J). Dynamics calcula-
tions were performed at 400 K with 500 steps at 0.001 ps
intervals.
1
compound 4 is shown in Fig. 1. The ring-current induced shifts
of the pyridine α- and β-protons of the template in the complex
(at δ 1.41 and 5.57, respectively) show that the three pyridyl
groups bind inside the cavity. The internal aromatic protons of
the host (H2) experience an even larger downfield shift, to δ
9.03, than in the zinc analogue,3 presumably due to the closer
proximity of the guest ring current. The carbonyl IR stretch is
sensitive to the trans ligand, shifting from 1931 cmϪ1 for metha-
nol adducts to 1944 cmϪ1 for pyridine adducts.
Syntheses
Molecular modelling
Model building and simple molecular mechanics calculations2,3
on the Zn3 trimer had shown that the cavity is too large to bind
the template without some distortion of the host or guest, but
serious modelling of these complex systems has not been previ-
ously pursued. In the absence of suitable crystals for X-ray
analysis, and with the advent of force fields that are appropriate
for metal ions in porphyrins,21,22 we undertook molecular
modelling of complex 4.
Fig. 2 summarises the modelling results: the template is
indeed too small to fit perfectly within the undistorted cavity,
but the trimer framework is flexible enough to respond to the
geometrical demands of the ligand. The porphyrin units flex
inwards [Fig. 2(a)] to optimise contact with the ligand: the pre-
dicted Ru᎐N (pyridyl) distance of 2.14 Å is very similar to
the 2.193 Å found experimentally in [Ru(tpp)(CO)(py)]20
(H2tpp = 5,10,15,20-meso-tetraphenylporphyrin, py = pyridine),
while the inward flexing of the porphyrin is essentially the same
as that found in the crystal structure of a related Zn4 tetramer–
tetrapyridylporphyrin complex.23 Furthermore, rotation about
the porphyrin–aryl linkage brings the whole porphyrin unit
closer to the ligand and imparts a helical twist to the trimer
framework [Fig. 2(b)]; distortion of such a linkage away from
perpendicular has been observed experimentally in the crystal
structure of a Zn3 trimer–pyridine complex.24
5,15-Bis(3-iodophenyl)-2,8,12,18-tetra(2-methoxycarbonyl-
ethyl)-3,7,13,17-tetramethyl porphyrin I. Palladium on carbon
(10%, 350 mg) was added to a solution of 5,5Ј-dibenzyl-
oxycarbonyl-3,3Ј-di(2-methoxycarbonylethyl)-4,4Ј-dimethyl-
2,2Ј-dipyrrolylmethane (6.15 g, 10 mmol)2 in tetrahydrofuran
(200 cm3) containing 1% triethylamine and the mixture was
stirred under a hydrogen atmosphere for 3 h. The catalyst was
removed (Celite) and the filtrate evaporated. Trifluoroacetic
acid (50 cm3, argon saturated) was added under argon at 0 ЊC
and the solution was stirred for 20 min followed by 20 min at
room temperature (periodically the reaction vessel was evacu-
ated and then saturated with argon to remove CO2). At this
stage, the solution was orange-brown. It was cooled to Ϫ30 ЊC
(liquid N2 in PriOH) and 3-iodobenzaldehyde (2.32 g, 0.01
mol)25 in methanol (50 cm3, argon saturated) was added by a
cannula and stirred for 2 h. During this period the temperature
was allowed to rise from Ϫ30 to Ϫ20 ЊC. 2,3-Dichloro-5,6-
dicyano-1,4-benzoquinone (2.35 g, 10 mmol) was then added
and the mixture stirred for 30 min before careful addition of
triethylamine (50 cm3). The resulting mixture was poured into
chloroform (250 cm3) and the organic layer washed with water
(2 × 500 cm3). The solvent was evaporated, and the solid passed
through a silica column [CHCl3–NEt3 (99:1) as eluent]. The red
porphyrin fractions were gathered, the solvents removed and
the product purified by at least two recrystallisations from
dichloromethane–methanol. The resulting red solid was filtered
off and dried in vacuo to yield compound I (2.55 g, 46% yield).
λmax/nm (CH2Cl2) (ε/dm3 molϪ1 cmϪ1) = 410 (27400), 507.5
(2000), 541 (550), 575.5 (760) and 627.5 (110). NMR (250 MHz,
CDCl3): 1H, δ Ϫ2.51 (s, 2 H, NH), 2.54 (s, 12 H, CH3 of pyrrole),
3.16 (t, J = 7.8, 8 H, CH2CH2CO2Me), 3.66 (s, 12 H, CO2CH3),
4.36 (t, J = 7.6 Hz, 8 H, CH2CH2CO2Me), 7.46–8.45 (m, 8 H,
CH of phenyl) and 10.29 (s, 2 H, H, meso); 13C, δ 15.01 (CH3 of
pyrrole), 21.94 (CH2CH2CO2Me), 36.93 (CH2CH2CO2Me),
51.77 (CO2CH3), 93.74 (CI), 97.11 (meso-CH), 116.78 (meso-C
of aryl), 129.39, 131.12, 137.70, 141.61 (CH of phenyl), 137.18,
141.26, 141.60, 144.06, 144.90 (pyrrole ring + C1 of phenyl) and
Conclusion
The templating approach used in this study provides a route to
ruthenium porphyrin oligomers with binding sites exclusively
facing into the cavity; such molecules have potential for cata-
lysis and for molecular electronics and electrochemistry. The
combined strength of three ruthenium–pyridine binding sites
has prevented satisfactory removal of the template in this par-
ticular case, even though the ligand is too small for strain-free
complexation. The way forward is clearly to reduce the number
of ruthenium centres as it is known that templates are readily
removed from zinc oligomers.
173.42 (C᎐O). Positive-ion FAB mass spectrum: m/z 1115.9
᎐
(M+) (C52H52I2N4O8 requires 1114.2).
Experimental
All solvents were distilled before use. All other reagents were
obtained commercially as reagent-grade chemicals and used
without further purification. For large-scale experiments (>10
mg) porphyrins were recrystallised by layered addition of
methanol to a concentrated solution in chloroform or dichlo-
romethane. Column chromatography was performed with 60
mesh silica gel or alumina activated II-III. Thin-layer prepara-
tive chromatography was performed with silica gel, type 60 on
20 × 20 cm glass plates (2.5 mm layer thickness).
The UV/VIS spectra were recorded with a Perkin-Elmer or
a Uvikon 810 spectrophotometer, IR spectra of chloroform
solutions with a Perkin-Elmer 1710 spectrometer in Fourier-
transform mode, 1H and 13C NMR spectra on Bruker WM-250
or AM-400 MHz spectrometers and fast atom bombardment
(FAB) mass spectra on a Kratos MS-50 instrument using a
m-nitrobenzyl alcohol matrix.
Carbonyl[5,15-bis(3-iodophenyl)-2,8,12,18-tetra(2-methoxy-
carbonylethyl)-3,7,13,17-tetramethylporphyrinato](methanol)-
ruthenium 1. A mixture of compound I (540 mg, 4.84 × 10Ϫ4
mol), [Ru3(CO)12] (650 mg, 1.01 × 10Ϫ3 mol) and decalin (20
cm3) was degassed, saturated with argon and heated overnight
at 130 ЊC under argon. The progress of the reaction was moni-
tored by UV/VIS spectroscopy. Upon completion of the reac-
tion half of the decalin was removed under vacuum, and the
red-orange product was filtered off, washed with cold metha-
nol, dissolved in the minimum of chloroform–methanol (20:1)
and then chromatographed on silica with chloroform elution.
The initial fractions contained an excess of [Ru3(CO)12] which
was collected and recycled, after which the ruthenium carbonyl
complex 1 was collected as a red solution. The solvent was
evaporated and the product recrystallised in a mixture of
chloroform–methanol. Yield 466 mg (75%). λmax/nm
J. Chem. Soc., Dalton Trans., 1997, Pages 985–990
987

View Article Online
(a)
(b)
Fig. 2 Two views of complex 4 generated by molecular modelling using CERIUS2 with the UNIVERSAL 1.01 force field: (a) from above, showing
how the porphyrin units are bent inwards to allow formation of Ru᎐N bonds; (b) from the side showing how partial rotation of the porphyrin–aryl
bond imparts a helical twist to the host and again brings the metal centres closer to the ligand
988
J. Chem. Soc., Dalton Trans., 1997, Pages 985–990

View Article Online
(CHCl3 + MeOH) = 305, 401.4, 523 and 554.2. ν˜max/cmϪ1
(CHCl3) = 1931 [CO, Ru(CO)]. H NMR (250 MHz, CDCl3 +
and the progress of the reaction was monitored by TLC
1
(chloroform). The product was washed with water (3 × 300
cm3),§ the solvent was evaporated and the product purified by
preparative TLC on silica with chloroform as eluent. The major
fraction corresponded to the expected ruthenium trimer isomer
(10.6 mg, 32%). λmax/nm (CHCl3) = 334.7, 401.7, 525.3 and
C5H5N) (mixture of conformers): δ 1.18–1.24 (m, 2 H, Hα of py
bound to Ru), 2.40 (3 s, 12 H, CH3 of pyrrole), 3.14 (m, 8 H,
CH2CH2CO2Me), 3.60 (3 s, 12 H, CO2CH3), 4.19 (br s, 8 H,
CH2CH2CO2Me), 5.07–5.17 (m, 2 H, Hβ of py), 6.01 (m, 1 H,
Hγ of py), 7.45 (t, 2 H, aryl, J = 8), 7.96 (d, 2 H, aryl, J = 8), 8.02
(d, 2 H, aryl, J = 8 Hz), 8.36 + 8.45 (2 s, 2 H, aryl), 9.80 (s, 2 H,
meso). Positive-ion FAB mass spectrum: m/z 1243 (M+) and
1215 (M Ϫ CO) (C53H50I2N4O9Ru requires 1242.1).
1
554.6. ν max/cmϪ1 (CHCl3) = 1944 [CO, Ru(CO)]. H NMR (250
MHz, CDCl3) (one isomer): δ 1.41 (d, 6 H, J = 6, Hα of
template), 2.39 (s, 36 H, CH3 of pyrrole), 2.99 (m, 24 H,
CH2CH2CO2Me), 3.39 (s, 36 H, CO2CH3), 4.02 (m, 24 H,
CH2CH2CO2Me), 5.57 (d, 6 H, J = 6, Hβ of template), 7.30 (d,
6 H, J = 8, aryl), 7.53 (t, 6 H, J = 8, aryl), 7.85 (d, 6 H, J = 8 Hz,
aryl), 9.03 (s, 6 H, CH of phenyl) and 9.38 (s, 6 H, meso).
Positive-ion FAB mass spectrum: m/z 3421 (M+) and 3336.8
(M Ϫ 3 CO) (C189H162N18O27Ru3 requires 3420.9).
Carbonyl(methanol){2,8,12,18-tetra(2-methoxycarbonyl-
ethyl)-3,7,13,17-tetramethyl-5,15-bis[3-(trimethylsilylethynyl)-
phenyl]porphyrinato}ruthenium 2. Complex
1
(450 mg,
3.5 × 10Ϫ4 mol) was dissolved in freshly distilled tetrahydro-
furan (25 cm3); triethylamine (25 cm3, freshly distilled), hexa-
kis(acetato)tripalladium() (12 mg, Johnson Matthey) and
triphenylphosphine (25 mg) were added and the mixture was
carefully degassed and saturated with argon. Trimethylsilyl-
acetylene (3 cm3) was added via a syringe and the reaction
mixture was heated overnight at 80 ЊC under an argon atmos-
phere. After evaporation of the solvent, the product was dis-
solved in the minimum of chloroform–methanol (20:1) and
chromatographed on silica gel with chloroform elution. The
product was recrystallised from chloroform–methanol to yield
Acknowledgements
We thank Dr. Harry Anderson for preliminary results and
stimulating discussions, Dr. Nick Bampos for advice and NMR
spectra, the European Community Human Captial and Mobil-
ity Programme (A. V. F. and V. M.), Cambridge Common-
wealth Trust, Ethyl Corporation and the New Zealand Vice
Chancellor’s Committee (S. J. W.) for their generous financial
support, and the EPSRC Mass Spectrometry Service for FAB
mass spectra.
the red-orange complex
2
(335 mg, 78%). λmax/nm
(CHCl3) = 310 (br), 401.3, 522.8 and 553.9. ν max/cmϪ1
᎐
§ The trimer solution should not be dried with potassium or sodium
carbonate as it appears to form an uncharacterised adduct from which
it could not be reisolated.
(CHCl ) = 1931 [CO, Ru(CO)], 2153 (C᎐CSi) and 846 (Si᎐Me).
᎐
3
1H NMR (250 MHz, CDCl3 + C5H5N): δ 0.25, 0.27 (2 s, 18 H,
SiMe3), 1.22 (m, 2 H, Hα of py bound to Ru), 2.36–2.40 (3 s, 12
H, CH3 of pyrrole), 3.11 (m, 8 H, CH2CO2Me), 3.59 (2 s, 12 H,
CO2CH3), 4.18 (br s, 8 H, CH2CH2CO2Me), 5.13 (m, 2 H, Hβ
of py), 6.0 (m, 1 H, Hγ of py), 7.69 (m, 2 H, aryl), 7.88 (m, 2 H,
aryl), 8.0–8.2 (m, 4 H, aryl) and 9.80 (br s, 2 H, meso). Positive-
ion FAB mass spectrum: m/z 1182.5 (M+) and 1154 (M Ϫ CO)
(C63H68N4O9RuSi2 requires 1182.4).
References
1 J. K. M. Sanders, Comprehensive Supramolecular Chemistry, eds.
J. L. Atwood, J. E. D. Davies, D. D. Macnicol and F. Vögtle,
Elsevier, Amsterdam, 1996, vol. 9, pp. 131–164.
2 H. L. Anderson and J. K. M. Sanders, J. Chem. Soc., Perkin Trans.
1, 1995, 2223; S. Anderson, H. L. Anderson and J. K. M. Sanders,
J. Chem. Soc., Perkin Trans. 1, 1995, 2247.
3 H. L. Anderson, S. Anderson and J. K. M. Sanders, J. Chem. Soc.,
Perkin Trans. 1, 1995, 2231.
4 C. J. Walter and J. K. M. Sanders, Angew. Chem., 1995, 107, 223;
Angew. Chem., Int. Ed. Engl., 1995, 34, 217.
5 L. G. Mackay, R. S. Wylie and J. K. M. Sanders, J. Am. Chem. Soc.,
1994, 116, 3141.
6 A. Vidal-Ferran, C. M. Müller and J. K. M. Sanders, J. Chem. Soc.,
Chem. Commun., 1994, 2657.
7 L. G. Mackay, H. L. Anderson and J. K. M. Sanders, J. Chem. Soc.,
Perkin Trans. 1, 1995, 2269; H. L. Anderson, C. J. Walter, A. Vidal-
Ferran, R. A. Hay, P. A. Lowden and J. K. M. Sanders, J. Chem.
Soc., Perkin Trans. 1, 1995, 2275.
8 S. S. Eaton and G. R. Eaton, Inorg. Chem., 1977, 16, 72.
9 H. L. Anderson, C. A. Hunter, M. N. Meah and J. K. M. Sanders,
J. Am. Chem. Soc., 1990, 112, 5780.
10 H. Ogoshi, H. Sugimoto and Z. Yoshida, Bull. Chem. Soc. Jpn.,
1978, 51, 2369; T. Boschi, G. Bontempelli and G.-A. Mazzocchin,
Inorg. Chim. Acta, 1979, 37, 155; P. Le Maux, H. Bahri and
G. Simonneaux, J. Chem. Soc., Chem. Commun., 1991, 1350.
11 J. T. Groves and R. Quinn, J. Am. Chem. Soc., 1985, 107, 5790.
12 S. J. Webb and J. K. M. Sanders, unpublished work.
13 D. Seyferth, M. O. Nestle and A. T. Wehmen, J. Am. Chem. Soc.,
1975, 97, 7417.
14 S. L. Schreiber, T. Sammakia and W. E. Crowe, J. Am. Chem. Soc.,
1986, 108, 3128.
15 M. J. Gunter and L. N. Mander, J. Org. Chem., 1981, 46, 4792.
16 D. P. Rillema, J. K. Nagle, L. F. Barringer and T. J. Meyer, J. Am.
Chem. Soc., 1981, 103, 56.
17 V. Marvaud and J. K. M. Sanders, unpublished work.
18 S. Anderson, H. L. Anderson and J. K. M. Sanders, J. Chem. Soc.,
Perkin Trans. 1, 1995, 2255.
19 D. W. J. McCallien and J. K. M. Sanders, J. Am. Chem. Soc., 1995,
117, 6611.
20 R. G. Little and J. A. Ibers, J. Am. Chem. Soc., 1973, 95, 8583.
21 A. K. Rappé, C. J. Casewit, K. S. Colwell, W. A. Goddard and
W. M. Skiff, J. Am. Chem. Soc., 1992, 114, 10024.
22 W. A. Goddard and A. K. Rappé, J. Phys. Chem., 1991, 95,
3358.
Carbonyl[5,15-bis(3-ethynylphenyl)-2,8,12,18-tetra(2-
methoxycarbonylethyl)-3,7,13,17-tetramethylporphyrinato]-
(methanol)ruthenium 3. The protected monomer 2 (256 mg,
2.1 × 10Ϫ4 mol) was dissolved in chloroform (50 cm3) contain-
ing tetrahydrofuran (0.5 cm3, freshly distilled) and the mixture
heated to reflux under dry air. Tetrabutylammonium fluoride
(1 cm3, of a 1.1 mol dmϪ3 solution in tetrahydrofuran) was
added and the reaction mixture was stirred for 10 min. Two
spatulas full of CaCl2 were added to the cooled mixture to
remove the excess of fluoride and the product was washed with
water (3 × 300 cm3). The solvent was evaporated, the product
dissolved in the minimum of chloroform–methanol (20:1)
and chromatographed on silica gel with chloroform elution.
It was recrystallised from a CHCl3–MeOH mixture to yield a
red-orange product (175 mg, 78%). λmax/nm (CHCl3) = 311
(br), 401.6, 523.7 and 554.4. ν max/cmϪ1 (CHCl3) = 1935.7
1
[CO, Ru(CO)]. H NMR (250 MHz, CDCl3 + C5H5N): δ 1.23
(m, 2 H, Hα of py bound to Ru), 2.39 (2 s, 12 H, CH3 of
pyrrole), 3.13 (m, 8 H, CH2CH2CO2Me), 3.59 (s, 12 H,
CO2CH3), 4.19 (br s, 8 H, CH2CH2CO2Me), 5.13 (m, 2 H, Hβ of
py), 6.0 (m, 1 H, Hγ of py), 7.70 (m, 2 H, aryl), 7.94 (m, 2 H,
aryl), 8.09 (m, 2 H, aryl), 8.13 (s, 2 H, aryl), 8.23 (s, 2 H, aryl)
and 9.81 (s, 2 H, meso). Positive-ion FAB mass spectrum: m/z
1039.4 (M+) and 1011.3 (M Ϫ CO) (C57H52N4O9Ru requires
1038.3).
Trimer 4. Deprotected porphyrin monomer 3 (30 mg,
2.8 × 10Ϫ5 mol) and tri(4-pyridyl)triazine (3 mg, 9.9 × 10Ϫ6 mol,
0.35 equivalent) were dissolved in dichloromethane (200 cm3,
freshly distilled over CaH2). After 5 min, when a bright clear
solution was obtained, CuCl (0.207 g) was added followed
by N,N,NЈ,NЈ-tetramethylethane-1,2-diamine (0.3 cm3). The
mixture was stirred under dry air for 2 h at room temperature,
J. Chem. Soc., Dalton Trans., 1997, Pages 985–990
989

View Article Online
23 S. Anderson, H. L. Anderson, A. Bashall, M. McPartlin and
J. K. M. Sanders, Angew. Chem., 1995, 107, 1196; Angew. Chem.,
Int. Ed. Engl., 1995, 34, 1096.
25 I. R. L. Barker and W. A. Waters, J. Chem. Soc., 1952, 150.
24 H. L. Anderson, A. Bashall, K. Henrick, M. McPartlin and J. K. M.
Sanders, Angew. Chem., 1994, 106, 445; Angew. Chem., Int. Ed.
Engl., 1994, 33, 429.
Received 22nd August 1996; Paper 6/05847G
990
J. Chem. Soc., Dalton Trans., 1997, Pages 985–990