Ruthenium Cumulenylidene Complexes Bearing Heteroscorpionate Ligands Rutheniumkumulenylidenkomplexe mit Heteroskorpionatliganden Der Naturwissenschaftlichen Fakultät der Friedrich-Alexander-Universität Erlangen-Nürnberg zur Erlangung des Doktorgrades Dr. rer. nat. vorgelegt von Frank Strinitz aus Nürnberg
252
Embed
Ruthenium Cumulenylidene Complexes Bearing ... · PDF fileHeteroscorpionate Ligands Rutheniumkumulenylidenkomplexe mit Heteroskorpionatliganden ... „Allenylidene Complexes Based
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Ruthenium Cumulenylidene Complexes Bearing
Heteroscorpionate Ligands
Rutheniumkumulenylidenkomplexe mit Heteroskorpionatliganden
8! Appendix .......................................................................................................................................201!8.1! Details of the Structure Determinations ................................................................................. 202!8.2! Cyclic Voltammetry .................................................................................................................. 211!8.3! Myoglobin Assay of CORMs .................................................................................................... 214!8.4! List of Abbreviations and Symbols .......................................................................................... 216!8.5! List of Compounds .................................................................................................................... 219!
Scheme 11. Synthesis of the first allenylidene complex by E. O. FISCHER et al.[4]
This route allows the formation of mono-heteroatom substituted allenylidene complexes but a
different approach is required for di-heteroatom substitution patterns. H. FISCHER et al.
established a route starting from [(CO)5M(THF)] (M = Cr, W), which yields upon addition of
deprotonated 3,3,3-tris(dimethylamino)prop-1-yne the intermediary metalate (Scheme 12).[83]
The abstraction of one amino group with BF3 etherate directly afforded the corresponding
neutral allenylidene complex. This route has also been employed on the synthesis of the
cumulogous pentatetraenylidene complex.[83]
Scheme 12. Synthesis of di-heteroatom substituted allenylidene complexes via alkynyl metalates.[83]
The current popularity of ruthenium allenylidene complexes can also be attributed to the
discovery of J. SELEGUE in 1982, that the direct conversion of [RuCpCl(PMe3)2] with 1,1-
diphenylprop-2-yn-1-ol led to a cationic allenylidene complex (Scheme 13).[84] This method is
based on the spontaneous dehydration of propargyl alcohols after η2-coordination of the
alkyne to 16 VE complexes that form intermediary hydroxyl vinylidene complexes.[84]
(CO)5M COC2H5
HCN(CH3)2
Ph (CO)5M C C CN(CH3)2
Ph1. EX3/CH2Cl22. THF- C2H5OH
M = Cr, EX3 = BF3M = W, EX3 = Al(C2H5)3
C C C(NMe2)3[(CO)5M(THF)] C CM(CO)5 C(NMe2)3
[(CO)5M] C C CNMe2
NMe2BF3 OEt2.
+
M = W, Cr
17
!State of Knowledge
!! !
Scheme 13. Synthesis of the first allenylidene complex based on a propargyl alcohol by J. SELEGUE and the mechanism of the dehydration.[84]
This method has proven to be suitable for a wide variety of ruthenium systems although
certain limitations are known. For a successful transformation i) the metal precursor needs to
form a coordinatively unsaturated 16 VE complex that allows η2-coordination of the alkyne.
The next limitation ii) is oftentimes the reluctance of the 3-hydroxyvinylidene to undergo
dehydration (Scheme 14-a). This behavior is especially distinctive when electron-rich metal
fragments are used. Depending on the used propargyl alcohol also iii) the competitive
formation of alkenylvinylidene complexes can occur, if the Cδ is carrying protons and thus
allows an 3,4-elimination of water with respect to the ruthenium center.[85-88] In some cases the
high stability of 3-hydroxyvinylidenes requires the treatment with acidic Al2O3 to complete
the conversion towards the allenylidene complex.[89-90] Calculations on the half-sandwich
system [RuCp(PH3)2]+ have emphasized the importance of protic solvents (e.g. MeOH) for the
transformation of the propargyl alcohol to the allenylidene complex (Scheme 14-b).[91]
RuMe3P PMe3
+ C CH CPh
OHPh
RuMe3P
Me3P CC
CPh
Ph
RuMe3P PMe3
RuMe3P
Me3P CCH
Cl
- Cl-
C CH CPh
OHPh
- H2O
- H2O, Cl-
Ph PhOH
18
!State of Knowledge
!! !
Scheme 14. a) Synthesis of allenylidene and alkenylvinylidene complexes via 3-hydroxyvinylidene intermediate; b) Mechanism of methanol catalyzed allenylidene formation.
The different ruthenium allenylidene complexes can be divided into three different groups,
namely neutral 16 VE complexes, cationic 18 VE complexes and neutral 18 VE complexes.
2.4.1 Neutral 16 Valence Electron Complexes
Starting from [RuCl2(PPh3)3] and [RuCl2(PPh3)4] A. HILL and P. NOLAN independently
explored the reaction with the propargyl alcohol 1,1-diphenylprop-2-yn-1-ol which led to the
complex [RuCl2(═C═C═CPh2)(PPh3)2] or upon addition of PCy3 to the complex to
[RuCl2(═C═C═CPh2)(PCy3)2],[92-95] which is closely related to Grubbs first generation catalyst
[RuCl2(═CHPh)(PCy3)2]. Nevertheless, the synthesis of [RuCl2(═C═C═CPh2)(PPh3)2] has
shown to be strongly temperature dependent as high temperatures allow the selective
formation of this 16 VE complex. Lower temperatures and the addition of NaPF6 led to the
additional formation of two 18 VE complexes that can be explained as dimeric forms of the
parent 16 VE complex. To compensate the lack in electron density either strongly donating
solvents are reported in the crystal structures of the complexes ([RuCl2(═C═C═CPh2)-
C CH COH
R2
R1
[M] C CH
COH
R1 R2
R2 = CHR3R4
- H2O[M] C C
H
C CR3
R4R1
[M] C C CR2
R1
[Ru] C CC O
HOH
Me
HHH
C C[Ru] CO
HH
HHO
Me
H[Ru] C C C
HH
OH
HOMe
H
- H2O
19
!State of Knowledge
!! !
(PPh3)2(solvent)] (solvent = H2O, MeOH, EtOH) or the formation of a symmetric or
asymmetric unit can be observed (Scheme 15).[96]
Scheme 15. Synthesis of monomeric, neutral dimeric and cationic dimeric diphenyl allenylidene complexes based on [RuCl2(PPh3)3].[96]
For the dimeric species two AB quartet patterns can be observed in the 31P NMR spectrum,
which can be assigned to the two different dimers (Scheme 15). Based on spectroscopic
analysis, one of the dimers contains two bridging and two terminal chlorido ligands, leading
to a neutral 18 VE complex. The second dimer is a monocationic complex that contains three
bridging chlorido ligands. The positive charge is compensated by a chloride anion that is
released during dimerization but can also be exchanged by addition of NaPF6 allowing to shift
the equilibrium towards the cationic dimer.
The monomeric allenylidene complex [RuCl2(═C═C═CPh2)(PPh3)2] can also be used as a
precursor for the selective formation of 18 VE complexes. Addition of dppe (1,2-
bis(diphenylphosphino)ethane) and KPF6 leads to the cationic complex [RuCl(═C═C═CPh2)-
(dppe)2]PF6 while the addition of carbon monoxide leads to the formation of
[RuCl2(═C═C═CPh2)(CO)(PPh3)2].[93]
The chemistry of the tricyclohexylphosphine-based complex [RuCl2(═C═C═CPh2)(PCy3)2] is
even more versatile as it allows a series of conversions. Reaction with the potassium salt of
[RuCl2(PPh3)3]HC C C(Ph)2OH
toluene, Δ, 4 hRu
Cl
ClPh3P
PPh3C C C
Ph
Ph
HC C C(Ph)2OHTHF25 °C2h
Ru C C CPh2Cl
ClCl
PPh3
Ph3P
RuCCPh2CPPh3
Ph3P
Cl
ClRuRu
Cl
Cl
CC
CPh2Ph3P
PPh3PPh3
Ph3P
CC
Ph2C
X
for X = ClTHF
Δ
NaPF6
X = Cl, PF6
20
!State of Knowledge
!! !
the hydridotris(1-pyrazolyl)borate ligand leads to the formation of the neutral 18 VE complex
[Ru(HB(pz)3)Cl(═C═C═CPh2)(PPh3)].[93] Addition of the dimeric [Ru2Cl4(η6-MeC6H4iPr)2]
leads to the formation of the dimer [Ru2Cl4(═C═C═CPh2)(PCy3)(η6-MeC6H4iPr)] which in
this case shows two bridging chlorido ligands with two terminal chlorido ligands, thus
resulting in a neutral complex (Scheme 16).[92] In comparison the conversion of the
allenylidene complex with the N-heterocyclic carbene ligand 1,3-bis((2,6-di-isopropyl-
phenyl)imidazol-2-ylidene) (IPr) leads to the 16 VE complex [RuCl2(═C═C═CPh2)-
(PCy3)(IPr)], which undergoes rearrangement to the indenylidene complex [RuCl2(3-
phenylindenylid-1-ene)(PCy3)(IPr)].[94]
Several of the aforementioned complexes have been tested as precatalysts in metathesis
reactions and this topic will be discussed later within this work (2.4.4, page 28).
Scheme 16. Synthesis of [Ru2Cl4(═C═C═CPh2)(PCy3)(η-MeC6H4iPr)], [RuCl2(═C═C═CPh2)(PCy3)(IPr)] and
Following this initial work, based on the dppm ligand, the complexes based on the dppe
system led to even more stable systems as indicated by the in situ generation of
[RuCl(dppe)2]PF6 from cis-[RuCl2(dppe)2] upon addition of NaPF6.[100-101] This system has
proven to be quite versatile and allows on the one hand systems similar to the dppm based
allenylidene complexes bearing a chlorido ligand trans to the allenylidene unit.[102] On the
other hand, highly unsaturated alkynyl allenylidene ruthenium complexes are achievable
(Scheme 18),[102] which are especially interesting due to their potential to create a C–C bond,
RuPh2P
Ph2P Cl
Cl
PPh2
PPh2
Ru
Ph2P PPh2
PPh2Ph2P
Cl C C CR
HPF6
C CH CHR(OH)NaPF6
R = Ph, p-PhCl, p-PhF, p-PhOMe
22
!State of Knowledge
!! !
by carbon-rich ligand coupling.[103-105] In addition, their dual alkynyl donor and allenylidene
acceptor functionalities raise interest in building linear conjugated organometallics.[102]
Scheme 18. Synthesis of highly unsaturated alkynyl allenylidene complexes via deprotonation of allenylidene complexes based on trans-[RuCl(═C═CHR)dppe2]PF6.[102]
It is noteworthy that these mixed alkynyl allenylidene complexes are stable towards the
addition of methanol at Cα and Cγ in contrast to other systems.[106-107] The steric hindrance and
the strong electron donating capabilities of the dppe ligands reduce the reactivity allowing
only stronger nucleophiles like sodium methoxide to react selectively with the Cγ carbon atom
resulting in ruthenium diacetylide complexes.[102, 108-109] Starting from the mixed alkynyl
allenylidene complex trans-[Ph2C═C═C═(dppe)2Ru—C≡C―CHPh]PF6, the oxidation with
Ce(IV) ammonium nitrate allows the isolation of the first real bis(allenylidene) metal
complex (Scheme 19). The previously reported bis(allenylidene) complex trans-[(dppm)2Ru-
(═C═C═C(OMe)(CH═CPh2))2]2+ shows due to the presence of a donor group on the
unsaturated chain an elevated bis(alkynyl) character and can better be described as trans-
(dppm)2Ru[—C≡C—C(═OMe)(C═CPh2)]22+.[110-112] Especially interesting is the behavior
upon one electron reduction due to the formation of a stable radical complex with an electron
pair delocalized identically on both sides of the alkynyl allenylidene complex. This
observation shows the possibility of these carbon-rich systems to mediate conductivity
Scheme 19. Reduction of trans-[Ph2C═C═C═(dppe)2Ru—C≡C―CHPh]PF6 with ammonium cerium(IV) nitrate to trans-[Ph2C═C═C═(dppe)2Ru═C═C═CPh2][B(C6F5)4]2.[110]
While carbon-rich ruthenium allenylidene complexes are no rarity, systems based on
polyaromatic ligands are rare. One interesting spacer is the dipropargyl alcohol 10,10′-
diethynyl-10H,10′H-[9,9′]bianthracenylidene-10,10′-diol that features the interesting
nonplanar bianthracenylidene moiety. Bis(allenylidene) ruthenium(II) complexes based on
this system can be obtained by reaction with two equivalents [RuCl(dppe)2]OTf. The addition
of one equivalent affords the expected monoallenylidene derivative and the electrochemical
and spectroelectrochemical properties were measured in detail (Scheme 20).[113] Both
techniques highlighted the presence of electronic communication between the metal centers
through the multiconjugated organic chains and allow allenylidene-centered reversible
reductions, which are not interrupted on passing from mononuclear allenylidene complexes to
dinuclear bisallenylidene complexes. This synthesis also allows the stepwise formation of
heterobimetallic allenylidene complexes via stepwise reaction of the propargyl alcohol with
selected ruthenium and rhenium precursors.
Ru
Ph2P PPh2
PPh2Ph2P
C C C
PF6
CPh
PhCC
Ph
PhH Ru
Ph2P PPh2
PPh2Ph2P
C C C
X2
CPh
PhCC
Ph
Ph1. CeIV, CH2Cl22. KB(C6F5)4X = B(C6F5)4
24
!State of Knowledge
!! !
Scheme 20. Synthesis of monoallenylidene and homobimetallic bisallenylidene complexes based on the bianthracenylidene linker.[113]
The dppe ligand system also allows the formation of further dinuclear allenylidene complexes
based on homocoupling reactions of ruthenium diyne complexes and heterocoupling reactions
between diyne complexes with allenylidene complexes.[102, 108, 114-116] Even larger trinuclear
complexes are accessible from aromatic spacers bearing three propargyl alcohols as
substitutes.[117]
HO
OH
CH
CH
Ru
Ph2P PPh2
PPh2Ph2P
C C CClOH
CH
Ru
Ph2P PPh2
PPh2Ph2P
C C CCl C C C Ru
Ph2P PPh2
PPh2Ph2P
Cl
(OTf)2
OTf
+ 1 eq. [RuCl(dppe)2]OTf
+ 2 eq. [RuCl(dppe)2]OTf
25
!State of Knowledge
!! !
The group of κ3 coordinated cationic ruthenium allenylidene complexes consists mainly of
N,N,N and S,S,S ligands, which are either facial or meridional coordinating. Cationic 18 VE
ruthenium allenylidene complexes are known for systems based on 1,4,7-trithia-
2,6-bis(oxazolyn-2"-yl)pyridines[122] and bis(pyrazol-1-yl)pyridines[123] but are mostly limited
to the diphenyl allenylidene complexes.
The ligand system with the next higher coordination, the κ4-N,N,N,N-makrocycle 1,5,9,13-
tetramethyl-1,5,9,13-tetraazacyclohexadecane (16-TMC), employed by C.-M. CHE et al.
selectively forms the trans positioned allenylidene complexes bearing heteroatom donor units
in the allenylidene residues.[124-125] The use of the dipyridyl allenylidene unit as a “molecular
clip” allows the coordination of zinc or ruthenium ions to form either homo- or hetero-
bimetallic allenylidene complexes (Scheme 21). DFT and TD-DFT calculations and
experimental data showed delocalization along the [Ru═C═C═C(2-py)2Ru] moiety in the
MLCT giving rise to NIR absorptions. This behavior highlights the potential application of
allenylidene ligands as molecular bridges to allow electronic communication between remote
functional groups.
Scheme 21. Synthesis of a heterobimetallic ruthenium zinc allenylidene complex and a homobimetallic ruthenium allenylidene complex based on 16-TMC.[125]
Important for this group of heteroatom based bimetallic complexes is the trans-arrangement
of the chlorido and allenylidene ligand due to the competitive behavior of both towards the
NN
N N
Me Me
Me
Ru
Me
Cl C CN
N
NN
N N
Me Me
Me
Ru
Me
ClN
NZn
Cl
Cl
NN
N N
Me Me
Me
Ru
Me
Cl C C CN
NRu(acac)2
C C C
C
+ ZnCl2
+ cis-[Ru(acac)2-(CH3CN)2]
26
!State of Knowledge
!! !
binding site. The irreversible formation of the allenylidene unit outcompetes the reversible κ1
coordination of the N-donor. Similar results were obtained with the trans-[RuCl(dppe)2]
system.[126-128]
The group of cationic half-sandwich complexes does not only include the classical η5-
cyclopentadienyl, η5-indenyl and η6-arene ruthenium complexes but also tethered type ligands
in which an ancillary κ1-coordinating donor atom is introduced leading to η5: κ1(L)- or
η6: κ1(L)-coordination. Most commonly η5 half-sandwich complexes bearing two phosphine
ligands and one allenylidene moiety are reported following a similar procedure to the classical
approach by J. SELEGUE (Scheme 13) leading to a general structure [Ru(η5-Ring)-
(═C═C═CR1R2)(L1)(L2)][X] with X– = BF4–, BPh4
–, PF6– and L1, L2 = PPh3, PMe3, PiPr3,
dppe.[89-90, 129-138]
Recently, several remarkable examples by E. NAKAMURA et al. based on a ruthenium(II)
fullerene-cyclopentadienyl complex bearing allenylidene ligands have been isolated (Figure
7).[139]
Figure 7. Ruthenium allenylidene complexes bearing a fullerene-cyclopentadienyl ligand.[139]
The focus of these complexes lies on the interaction of the physical and chemical properties
of the allenylidene and fullerene moieties. On the one hand the bulkiness of the C60Me5 ligand
leads to regio- and stereoselectivity for nucleophilic additions. On the other hand the intense
absorptions of the aforementioned complexes in the visible and NIR region are of particular
interest for their potential use in photophysical applications.[139-144]
N,N-diallyltosylamide into N-tosyldihydropyrrole. Three general trends could be observed:
i) The activity of the catalyst increases with the electron richness off the phosphine
ligand in the series PPh3 < PiPr3 < PCy3 the.
ii) The employed counter anion has a drastic impact on the reactivity which increases
with the sequence BF4– < BPh4
– ≈ PF6– < TfO–.
iii) While several substituted diarylallenylidene complexes have been tested the most
potent was the simple diphenylallenylidene which showed values comparable to
[RuCl2(═CHPh)(PCy3)2].[164-166]
Scheme 24. Conversion of N,N-diallyltosylamide into N-tosyldihydropyrrole catalyzed by the precatalyst [RuCl(═C═C═CR2)(η6-p-cymene)(PR3)][X]. [164-166]
A common rearrangement for diphenylallenylidene complexes is the formation of an
indenylidene system that leads to a stronger repulsive interaction between the indenylidene
group and the arene ligand resulting in a dissociation of the later. The generated vacant sites
are required for substrate binding and the utility of isolated indenylidene complexes has been
reported in detail by P. DIXNEUF et al.[158, 167] Due to the rigid structure of the fluorenyl unit no
rearrangement into carbene complexes is possible thus, leading to a different mechanism that
has not yet been fully understood.[163, 166]
2.5 Heteroscorpionate Chemistry
The aforementioned bdmpza ligand has been first reported by A. OTERO et al. in 1999 and can
be synthesized starting from bis(3,5-dimethylpyrazol-1-yl)methane.[43] Deprotonation with
n-butyllithium followed by reaction with carbon dioxide leads to the lithium compound
[Li(H2O)(bdmpza)4].[43] A more versatile one-pot synthesis has been introduced by N.
BURZLAFF et al. starting from either 3,5-dimethylpyrazole or unsubstituted pyrazole.[42] The
reaction with dichloroacetic or dibromoacetic acid under basic conditions in the presence of a
phase transfer catalyst allows after acidic workup the direct isolation of the protonated ligands
Hbdmpza or Hbpza (bis(pyrazol-1-yl)acetic acid). In comparison to bpza the methyl
NTs
NTs[RuCl(=C=C=CR2)-
(η6-p-cymene)(PR3)][X]- C2H4
30
!State of Knowledge
!! !
substituents of the bdmpza ligand increase the solubility in common organic solvents and
furthermore the steric hindrance. This leads in comparison to the air-stable complex
[Ru(bpza)Cl(PPh3)2] to an oxygen sensitive complex [Ru(bdmpza)Cl(PPh3)2] (Figure 8). The
methyl substituents and the corresponding steric hindrance causes a smaller angle between the
nitrogen donor atoms and the ruthenium center as observed in the crystal structure and thus
leads to a labilization of the phosphine ligands.[168] This is remarkable as the closely related Tp
complex [Ru(Tp)Cl(PPh3)2] and the Cp complex [Ru(Cp)Cl(PPh3)2] are air stable compounds
which indicates that the steric hindrance of the bdmpza ligand strongly increases the tendency
to release one phosphine ligand.[169-171]
Figure 8. Ruthenium based triphenylphosphine complexes bearing the Tp scorpionate ligand ([Ru(Tp)Cl(PPh3)2], left) and the heteroscorpionate ligands bdmpza ([Ru(bdmpza)Cl(PPh3)2], middle) and the ligand bpza
([Ru(bpza)Cl(PPh3)2], right).[168-169]
The class of pyrazole based N,N,O heteroscorpionates contains mainly acetic acid based
systems, although more flexible ligands based on propionic acid have been introduced by E.
DÍEZ-BARRA et al. as sodium salt of 3,3-bis(pyrazol-1-yl)propionate (Na[bpzp]) and 3,3-
bis(3,5-dimethylpyrazol-1-yl)propionate (Na[bdmpzp]).[172] The free acids and coordination
properties have been explored by N. BURZLAFF et al.[173] Synthesis of these elongated N,N,O
ligands involves a double Michael Addition of the pyrazole precursor to methyl propiolate.
Depending on the aqueous workup this either affords the free acid or the corresponding
sodium salt (Scheme 25). The coordination behavior of these two ligands has been explored
with manganese carbonyl complexes as the CO vibrations can be monitored by IR
spectroscopy and thus allows probing the electron donating and accepting properties of the
ligands. Addition of [MnBr(CO)5] to the in-situ formed potassium salt (K[bdmpzp]) leads to
the formation of the corresponding manganese(I) carbonyl complex fac-
[Mn(bdmpzp)(CO)3].[173] Closely related are the complexes fac-[Re(bdmpza)O3] and fac-
[Tc(bpza)O3] which attracted attention for their possible application regarding radio-
NN N
N
Me
Me
Me
MeRu
OO
Cl PPh3Ph3P
NN N
N
Ru
OO
Cl PPh3Ph3P
N
NN
N N
N
Ru
BH
PPh3Ph3P Cl
31
!State of Knowledge
!! !
pharmaceutical purposes.[174-175] In comparison to the previously reported complexes fac-
[Mn(bdmpza)(CO)3] and fac-[Re(bdmpza)(CO)3] the elongated propionate based complex
show only slight deviations in their chemical properties as for example the carbonyl ligand
trans to the carboxylate anchor shows a longer distance to the manganese center and the
coordination geometry around the manganese center is closer to perfect octahedral geometry
due to the decreased strain.[173] For further work on the chemistry of N,N,N, N,N,S, N,N,O and
N,N,Cp scorpionate ligands several reviews are available.[44, 176-180]
Scheme 25. Synthesis of the propionic acid based ligands Hbpzp and Hbdmpzp via methyl propiolate; formation of the manganese (I) and rhenium (I) carbonyl complexes [Mn(bdmpzp)(CO)3] and [Re(bdmpzp)(CO)3].[172-173]
The aforementioned system [Ru(bdmpza)Cl(PPh3)2] has proven to show rich follow-up
chemistry due to the labile triphenylphosphine and chlorido ligands. Moreover, the complex
has been used as a model for the active site of 2-oxoglutarate dependent iron enzymes, which
are often difficult to investigate due to their paramagnetic ferrous high-spin constitution. In
comparison, the ruthenium based system with its low-spin state allows NMR characterization
of the complexes. Conversion of [Ru(bdmpza)Cl(PPh3)2] with acetate or benzoate allows the
formation of the κ2 coordinated neutral ruthenium complexes [Ru(bdmpza)(O2CMe)(PPh3)]
and [Ru(bdmpza)(O2CPh)(PPh3)] (Scheme 26).[181] In similar fashion the reaction of
[Ru(bdmpza)Cl(PPh3)2] with thallium 2-oxocarboxylates Tl[O2CC(O)R] (R = Ph,
CH2CH2CO2H) produces κ2O1,O2-2-oxocarboxylato complexes which can also be synthesized
via the intermediary acetato or benzoato complexes due to the higher acidity of the
oxocarboxylic acid. This is especially relevant for the catalytic cycle of the 2-oxoglutarate
dependent enzymes, which has been postulated to show the exchange of a carboxylato ligand
by a 2-oxocarboxylato ligand as a regenerative step.[182-183] The hemilabile behavior of the
κ1O1,O1´ ligands allows the isolation of the water and acetonitrile adducts [Ru(bdmpza)-
(O2CMe)(H2O)(PPh3)] and [Ru(bdmpza)(O2CPh)(MeCN)(PPh3)] which are promising
candidates for further reactions.[181]
NN N
N
Me
Me
Me
MeMO
CO COCO
NNH
R
R
1. NaH2.3. H+/H2O
CO2CH3 ONN N
N
R
R
R
R
COOH1. KOtBu2. [MnBr(CO5)]
R = H, Me M = Mn, Re
32
!State of Knowledge
!! !
Scheme 26. Formation of carboxylato and 2-oxocarboxylato ruthenium(II) complexes.[181]
The reaction of [Ru(bdmpza)Cl(PPh3)2] with pyridine, acetonitrile, carbon monoxide and
sulfur dioxide has also been reported and has shown that the N,N,O ligand bdmpza leads to a
preferred arrangement around the ruthenium center with the remaining triphenylphosphine
ligand positioned trans to a pyrazole unit. The chlorido ligand is positioned trans to the
acetate anchor in the aforementioned cases and leaves the newly introduced ligand in trans
position to the remaining pyrazole unit. This behavior has been observed for the complexes
allenylidene complexes. The reaction of the bisphosphine ruthenium complex with a series of
terminal alkynes led via 1,2-H shift of the intermediary η2 coordinated alkyne to the
vinylidene complexes with the vinylidene ligand positioned trans to the pyrazole or
carboxylate unit. The relation of the two formed isomers is attributed to the sterical hindrance
of the respective alkyne complex during the η2 coordination preferring the arrangement trans
to the pyrazole moiety for larger substituents. Up to now, the phenyl, tolyl, propyl and butyl
substituted vinylidene complexes of the general formula [Ru(bdmpza)Cl(═C═CHR)(PPh3)]
have been reported (Scheme 28).[61, 186]
Scheme 28. Synthesis of bdmpza based ruthenium vinylidene complexes with the cumulenylidene ligand positioned either trans to a pyrazole or the carboxylate unit (R = phenyl, tolyl, propyl, butyl).[186]
Due to the sensibility of the vinylidene complexes towards oxidation, a separation of the two
occurring structural isomers cannot be achieved. Hence, the reactivity of the ruthenium
precursor with propargyl alcohols was explored leading to the diphenyl and ditolyl substituted
allenylidene complexes of the general formula [Ru(bdmpza)Cl(═C═C═CR2)(PPh3)]
(R = phenyl, tolyl) which are stable towards air and humidity (Scheme 29). Separation via
column chromatography allows the isolation of both structural isomers, which show different
chemical and physical properties. Studies on the reactivity of these allenylidene complexes in
metathesis reactions were disappointing as no catalytic activity could be observed, which is
attributed to the stability of the 18 VE complex that does not undergo ligand dissociation to
the reactive 16 VE species.
NN N
N
Me
Me
Me
MeRu
OO
Cl PPh3Ph3P
- PPh3
NN N
N
Me
Me
Me
MeRu
OO
C ClPh3P
NN N
N
Me
Me
Me
MeRu
OO
CClPh3PC
H RC H
R
+C CH R+
34
!State of Knowledge
!! !
Scheme 29. Synthesis of bdmpza based ruthenium allenylidene complexes with the cumulenylidene ligand positioned either trans to a pyrazole or the carboxylate unit (R = phenyl, tolyl).[186]
As mentioned in chapter 2.4.3 for the Tp based allenylidene complexes, the addition of
nucleophiles usually occur either in α or γ position. For weak nucleophiles, like ammonia, the
addition to the γ position on the complex [Cr(═C═C═C(OMe)NMe2)(CO)5] has been
reported.[187] The rearrangement of the Cγ ammonia adduct to the α-aminocarbene is known
for the complex [Re(═C═C═CPh2)(CO)2(triphos)][OTf] (triphos = 1,1,1-tris(diphenyl-
phosphinomethyl)ethane).[188] In the case of the reaction of [Ru(bdmpza)Cl-
(═C═C═C(tolyl)2)(PPh3)] with methylamine, the product can be obtained as aminocarbene
complex indicating the higher reactivity of the α carbon atom which is in good agreement
with DFT calculations (Scheme 30).[189]
Scheme 30. Addition of methylamine to the complex [Ru(bdmpza)Cl(═C═C═C(tolyl)2)(PPh3)] yielding the corresponding aminocarbene complex.[189]
NN N
N
Me
Me
Me
MeRu
OO
Cl PPh3Ph3P- PPh3
NN N
N
Me
Me
Me
MeRu
OO
C ClPh3P
NN N
N
Me
Me
Me
MeRu
OO
CClPh3PC
C
+C CH CR2OH+
CCR
R
R R
NN N
N
Me
Me
Me
MeRu
OO
CClPh3P CC
Tol
Tol
+ NH2Me
NN N
N
Me
Me
Me
MeRu
OO
CClPh3PNHMe
HCC Tol
Tol
35
3 OBJECTIVE AND AIMS
36
!Objective and aims
!! !
During the last years a rich variety of heteroscorpionate complexes based on transition metals
has been reported by the BURZLAFF group and other working groups and extensively reviewed
by A. OTERO.[44, 173, 176, 180, 184-185, 189-190] In the field of manganese(I) carbonyl complexes bearing
heteroscorpionate ligands I. HEGELMANN and L. PETERS of the BURZLAFF group synthesized
several complexes of the general formula [Mn(L1-3)(CO)3] with the heteroscorpionate ligands
bis(3,5-dimethylpyrazol-1-yl)acetate (L1 = bdmpza), bis(pyrazolyl)acetate (L2 = bpza) and
3,3-bis(3,5-dimethylpyrazol-1-yl)propionate (L3 = bdmpzp).[42, 173] Given the interest in
medicinal applications of heteroscorpionate complexes, the carbon monoxide release
properties of these and closely related manganese(I) carbonyl complexes should be
investigated as first project within this thesis. Closely related is a topic previously
investigated by S. TAMPIER and G. TÜRKOGLU concerning ruthenium carbonyl complexes
with heteroscorpionate ligands. A single crystal X-ray structure determination of the dinuclear
complex [Ru(bdmpza)(CO)(μ2-CO)]2 was obtained by serendipity and in consequence the
rational synthesis of this compound should be explored within this work.
Especially the complex [Ru(bdmpza)Cl(PPh3)2] has shown a versatile organometallic
chemistry including ligand exchange reactions leading to the 2-oxocarboxylato complex
[Ru(bdmpza)(O2C(CO)Me)(PPh3)].[185] As these can be classified as bioinorganic model
complexes for iron enzymes, the topic of aminophenol ligands and their non-innocent
behavior should be explored starting from previous results of M. KECK.
Moreover, a major part of this thesis should be the synthesis and characterization of carbon-
rich cumulenylidene complexes, i.e. ruthenium allenylidene and vinylidene complexes. The
starting point was previous work by H. KOPF on the synthesis of neutral bdmpza based
ruthenium complexes with the general formula [Ru(bdmpza)Cl(L)(PPh3)].[61, 186, 189] Thus
within this work several new carbon-rich propargyl alcohols should be synthesized and the
reaction to the corresponding ruthenium allenylidene complex should be performed. The
properties of the resulting allenylidene complexes should be characterized with a focus on
cyclic voltammetry and absorption spectroscopy. As a side project the formation of water-
soluble carbon-rich allenylidene complexes should be explored for possible applications in
non-covalent functionalization of carbon allotropes.
37
4 RESULTS AND DISCUSSION
38
!Results and Discussion
!! !
4.1 Manganese Based Photo-CORMs
Recently, P. KURZ and coworker compared the CO-release properties of [Mn(bdmpza)(CO)3]
(1) (bdmpza = bis(3,5-dimethylpyrazol-1-yl)acetate), that was published by the BURZLAFF
group some years ago, to that of a related tris(pyrazol-1-yl)methane (tpm) complex.[42, 191]
Given the interest in medicinal applications of heteroscorpionate complexes, it was decided to
explore the carbon monoxide release properties of such heteroscorpionate manganese(I)
carbonyl complexes further. During the last decade, the BURZLAFF group has reported on the
synthesis of various manganese(I) complexes based on heteroscorpionate ligands. Up to date,
the complexes [Mn(bdmpza)(CO)3] (1), [Mn(bpza)(CO)3] (2) (bpza = bis(pyrazol-1-
yl)acetate), and [Mn(bdmpzp)(CO)3] (3) (bdmpzp = 3,3-bis(3,5-dimethylpyrazol-1-yl)-
propionate) have been described.[42, 173] Due to the rising interest in CORMs, especially with
Alfama´s lead compound fac-[Mo(CO)3(histidinate)]Na (ALF-186),[192] which also features a
κ3 coordinated N,N,O motif, we decided to synthesize the missing link, a manganese(I)
complex bearing a 3,3-bis(pyrazol-1-yl)propionic acid (Hbpzp) and to study the complexes
towards their CO-release properties (Scheme 31).
Scheme 31. Synthesis of heteroscorpionate complexes [Mn(bdmpza)(CO)3] (1), [Mn(bpza)(CO)3] (2), [Mn(bdmpzp)(CO)3] (3) and [Mn(bpzp)(CO)3] (4).[42, 173]
The ligands and complexes were synthesized according to reported procedures.[42, 173]
Deprotonation of the free acids (e.g. Hbpzp) with potassium tert-butylate lead to the
potassium carboxylates (e.g. K[bpzp]). Reaction of [MnBr(CO)5] with these carboxylates
resulted in the formation of tricarbonyl complexes [Mn(heteroscorpionate)(CO)3] (1-4). The
new complex [Mn(bpzp)(CO)3] (4) could be filtered off after spontaneous precipitation from
the reaction mixture. Complex 4 is stable towards oxygen as a solid but decomposes in
solution within some hours, as described previously for complexes 1-3.[42, 173] In comparison to
NN N
N
R
R
R
RMn
OO
OC COCO
NN N
N
R
R
R
RMnOC COCO
OO
or
R = Me (1), H (2) R = Me (3), H (4)
[MnBr(CO)5] potassium carboxylateof the scorpionate+
39
!Results and Discussion
!! !
the dimethylpyrazole based complex [Mn(bdmpzp)(CO)3] (3), its solubility in most solvents
is lower. The two pyrazolyl groups of 4 give rise to only one set of signals in the 1H NMR
spectrum in accordance with the expected Cs structure. Due to poor solubility, no carbonyl
signals could be detected in the 13C NMR spectrum. Nevertheless, the IR spectrum shows
three carbonyl signals (A’, A’’ and A’) at 2026, 1932 and 1915 cm–1, as expected for a facial
coordinated tricarbonyl complex with Cs symmetry.
Furthermore, the complex [MnBr(CO)3(Hpz)2] (5) was included in this study representing a
simplified analogue of complexes 2 and 4 and that - from a synthetic point of view - is a lot
simpler to prepare (Scheme 32). Moreover, it was tried to synthesize the analogous
imidazole based manganese(I) complex but the reaction of one equivalent [MnBr(CO)5] with
two equivalents of imidazole in CH2Cl2 over 5 h only led to the formation of the cationic
Scheme 32. Synthesis of manganese based complexes [MnBr(CO)3(HPz)2] (5) and [Mn(CO)3(HIm)3]Br (6).[193]
The high symmetry (C3v) of complex 6 is emphasized by the 1H NMR and 13C NMR spectra.
Both spectra show only one set of signals for all relevant positions including the amine
protons at 12.97 ppm in the 1H NMR spectrum and the carbonyl ligands at 220.3 ppm in the 13C NMR spectrum. ESI-MS experiments showed that the cation can be detected at
m/z = 343.05 which is the characteristic [M]+ signal for the complex after loss of the counter
ion.
The 13C NMR and IR data of the carbonyl ligands in complexes 1-6 is listed in Table 2. The
characteristic facial coordinating motif of the heteroscorpionate N,N,O ligands leads to the
formation of complexes with Cs symmetry as has been shown previously.[42-43, 173] This reduced
symmetry compared to the C3v symmetrical tpm M(CO)3 based complexes gives rise to three
IR absorption bands (A’, A’’ and A’) for the CO vibrations. Due to its higher C3v symmetry,
complex 6 exhibits only two characteristic CO vibrations in the IR spectrum.
MnOC
OC N
Br
N
CO
MnOC
N N
CO
N
CO
HN
NH
NH
NH
NH
[MnBr(CO)5]2 HPz 3 HIm
5 6
Br
40
!Results and Discussion
!! !
Ligand Complex δ (CO) [ppm] ṽ (CO)(KBr) [cm–1]
bdmpza 1 219.8 2038, 1954, 1931
bpza 2 219.3 2039, 1956, 1917
bdmpzp 3 222.0 2030, 1925, 1899
bpzp 4 / 2028, 1947, 1929
2 × Hpz, 1 × Br– 5 222.8 2035, 1940, 1918
3 × HIm 6 220.3 2024, 1907
Table 2. Overview of relevant 13C NMR and IR spectroscopic data of compounds 1-6.
CO release properties
In order to study the properties of the new compounds as photoactivatable CO releasing
molecules (Photo-CORMs), the manganese complexes 1-6 were investigated using the
UV/Vis based myoglobin assay. Complex 1 has been reported previously to show Photo-
CORM activity,[191] but due to a slightly different assay, it was decided to reassess the values
to facilitate comparison with compounds 2-6. Prior to measuring the release upon UV
excitation, the stability of each compound was tested in the dark. Complexes 1-3 were stable
and did not show any decomposition during 6 h. Nevertheless, complexes 4-6 showed
CO release upon dissolution in aqueous media in the dark. Similar behavior has previously
been reported for the precursor [Mn(CO)5]Cl, which forms the aqua complex
[Mn(CO)3(H2O)3]+.[194] Complex 4 exhibited slow release within a timeframe of 24 h yielding
1.42 ± 0.04 eq. of carbon monoxide with an average half life of 217 min (Chapter 8.3, Figure
64). For bis(imidazol-2-yl)propionate (bip) based complexes, a betain-like structure has been
reported with a κ2 coordination of the bis(imidazo-2-yl)methane moiety and a dissociated
propionate anchor.[195] A similar dissociation of the carboxylate donor might explain the low
stability of 4 in solution, but up to now it was not possible to isolate such an intermediate. The
pyrazole-based compound 5 shows faster decomposition in the dark, releasing 1.14 ± 0.09 eq.
CO with t1/2 = 124 min (Chapter 8.3, Figure 65). In contrast, the cationic imidazole complex 6
shows the fastest CO release in the dark with a t1/2 = 73 min and the release of 2 eq. CO
41
!Results and Discussion
!! !
(1.94 ± 0.39) (Chapter 8.3, Figure 66). Nevertheless, it was decided to take a look at the effect
of photoactivation of 5 and 6 with UV light to see if complete CO release can be achieved.
Figure 9. Fitted average CO release of 6 (squares) and 5 (dots) measured via myoglobin assay.
For complex 5, a strong acceleration of the CO release was achieved via UV excitation
(365 nm), lowering the release time to t1/2 = (6.40 ± 0.14) min and yielding an increased
amount of equivalents CO (2.17 ± 0.09 eq.) (Figure 9). In comparison, 6 shows an even
higher activity under these conditions, releasing about 2.5 equivalents of carbon monoxide in
20 min (t1/2 = 5.56 min, 2.60 ± 0.35 eq.). Due to the different coordination sphere of the
manganese(I) center with two pyrazole or three imidazole ligands, a direct comparison is
difficult. However, removal of the ligand backbone from either tris(imidazol-2-yl)phosphanes
or bis(pyrazol-1-yl)acetate leads to manganese tricarbonyl complexes that are not stable under
the conditions of the myoglobin assay.[196]
In the next step, the Photo-CORM properties of the compounds, which showed no CO release
in the dark were evaluated. For 1, a t1/2 = 6.73 min with a release of 2.38 ± 0.11 eq. of carbon
monoxide was observed upon photoactivation at 365 nm, which is in good accordance with
the literature value of 2.5 eq. (Figure 10).[191] Compound 2 lacks the methyl substituents as in
complex 1 and shows slower CO release. With 2.15 ± 0.11 eq. and t1/2 = 11.35 min, the
complex seems to be more stable than the methyl substituted one, which is in agreement with
previous observations regarding these ligands.[168] In a similar context, the ruthenium based
system [Ru(bpza)Cl(PPh3)2] is quite stable towards oxygen, whereas the bulkier methyl-
0 5 10 15 20 25 300.0
0.5
1.0
1.5
2.0
2.5
3.0
eq. (CO)
t [min]
42
!Results and Discussion
!! !
substituted [Ru(bdmpza)Cl(PPh3)2] is highly oxidation sensitive and decomposes within hours
in the presence of oxygen.[168]
Figure 10. Fitted average CO release of 1 (squares), 2 (dots) and 3 (triangles) measured via myoglobin assay.
Extending the acetate anchor to a propionate leads to complex 3, which has the fastest
CO release kinetics of the compounds presented within this work. Full release of 2 eq. CO has
been observed after less than 15 min with t1/2 = 3.77 min (2.06 ± 0.09 eq.). This might be due
to the propionate based ligand that allows more dynamic behavior in the resulting complexes.
The less rigid coordination seems to facilitate the dissociation of carbonyl ligands compared
to the acetate based ones. The difference in CO-release rate between 3 and 4 seems to be
strongly influenced by the methyl substituents. Obviously, in this case, the methyl groups
stabilize the complex, hinder the mobility of the carboxylate anchor and prevent solvent-
controlled dissociation of CO molecules for 3. The proton-substituted complex 4 is however,
not stable in solution and solvent-controlled dissociation of the carbonyl ligands occurs in the
dark.
In Table 3, the properties of several literature known Photo-CORMs are collected.
Comparison with the heteroscorpionate complexes shows that on average, only two of the
three CO ligands per complex are released. The half life of the complexes presented in this
work is shorter than most of the compounds described so far in the literature, but varies
significantly with the ligand system.
0 5 10 15 20 25 300.0
0.5
1.0
1.5
2.0
2.5
eq. (CO)
t [min]
43
!Results and Discussion
!! !
Complex t1/2 [min] eq. (CO) Reference
1 [Mn(bdmpza)(CO)3] 7 2.38 [a]
2 [Mn(bpza)(CO)3] 11 2.15 [a]
3 [Mn(bdmpzp)(CO)3] 4 2.06 [a]
[Mn(tpm)(CO)3]PF6 20 1.96 [36]
fac-[Mn(his)(CO)3] 93 1.26 [197]
[Mn2(CO)10] (CORM-1) not determined 0.68 [32]
Table 3. Photoinduced CO release upon UV excitation for manganese(I) carbonyl complexes, with half-lifes and number of CO molecules released per complex determined with the myoglobin assay; [a] reported in this work.
ethylenediamine).[215-218] Furthermore, the complex [Ru(Me3tacn)(OH2)(O2CCF3)](O2CCF3)2
was shown to be an effective catalyst for homogeneous oxidation of alkenes by tert-butyl
hydroperoxide (TBHP) as an oxidant.[219] Thus, in previous experiments some of the
ruthenium bdmpza complexes mentioned above, such as [Ru(bdmpza)Cl(PPh3)2] and
[Ru(bdmpza)(OAc)(PPh3)], were tested for their catalytic activity in similar alkene
45
!Results and Discussion
!! !
epoxidations.[220] Unfortunately, rather poor catalytic activity with only 2−3 turnovers was
observed, due to large quantities of O═PPh3 byproduct, which inhibit the catalytic
epoxidation. Thus, it was decided to focus on phosphine-free complexes for further studies. S.
TAMPIER and G. TÜRKOGLU synthesized [Ru(bdmpza)Cl(CO)2] (9) and [Ru(2,2-bdmpzp)Cl-
(CO)2] (10) starting from [RuCl2(CO)2]n[221] that has been reported by J. VOS et al. as quite
useful and easily accessible precursor for phosphine free ruthenium chemistry.[222-223] Addition
of the respective potassium salt of the heteroscorpionate ligand leads to formation of the
mononuclear ruthenium complex (Scheme 33).
Scheme 33. Synthesis of dicarbonyl complexes [Ru(bdmpza)Cl(CO)2] (9) and [Ru(2,2-bdmpzp)Cl(CO)2] (10).
In addition the formation of the dinuclear byproduct [Ru(bdmpza)(CO)(μ2-CO)]2 was
observed in the FAB+ mass spectrum as indicated by the molecular ion peak (m/z 810, 4%). Thus, it was not surprising that, in attempts to crystallize complex 9, crystals of this
byproduct [Ru(bdmpza)(CO)(μ2-CO)]2 (12) suitable for an single-crystal X-ray structure
determination were isolated by S. TAMPIER.[222] The structure determination revealed its
molecular structure as a dinuclear μ2-CO complex as reported in his dissertation.[222]
There are several procedures described in the literature for the synthesis of the analogous
cyclopentadienyl compound [Ru(η5-C5H5)(CO)(μ2-CO)]2. Thus, attempts were undertaken to
rationally synthesize this compound. E. O. FISCHER et al. synthesized [Ru(η5-C5H5)(CO)(μ2-
CO)]2 via reaction of the ruthenium(II) precursor [Ru(CO)2I2][224] with an excess of sodium
cyclopentadienide, Na[C5H5].[225] Attempts to adopt this procedure by using potassium bis(3,5-
dimethylpyrazol-1-yl)acetate instead of Na[C5H5] failed because of the insolubility of
K[bdmpza] in the aliphatic solvent. In further attempts an oxidative addition of bis(3,5-
dimethylpyrazol-1-yl)acetic acid to [Ru3(CO)12] was tested. This should result in the hydrido
complex [Ru(bdmpza)H(CO)2] (11), which might then be oxidized by oxygen and dimerize to
[Ru(bdmpza)(CO)(μ2-CO)]2 (12) as reported for [Ru(η5-C5H5)(CO)(μ2-CO)]2.[226-227] Indeed,
NN N
N
Ru
OO
OC ClCO
Me
Me
Me
Me
R
NN N
N
Me
Me
Me
Me
CO2HR
R = H (7)R = Me (8)
R = H (9)R = Me (10)
1. KOtBu2. [RuCl2(CO)2]n
46
!Results and Discussion
!! !
the 1H NMR spectrum of the resulting product indicated formation of two hydrido complexes
by two singlet signals at −13.32 and −10.10 ppm (in CDCl3) (Scheme 34). These signals have
been assigned to two structural isomers, the symmetrical hydrido complex [Ru(bdmpza)-
H(CO)2] (11A) and the unsymmetrical hydrido complex [Ru(bdmpza)H(CO)2] (11B). In the
Cs symmetric isomer 11A the hydrido ligand resides trans to the carboxylate and only one set
of signals is observed for the two 3,5-dimethylpyrazole donors in the 1H NMR spectrum,
whereas the C1 symmetric, chiral, but racemic complex [Ru(bdmpza)H(CO)2] (11B) shows
two sets of signals in the 1H NMR spectrum, instead.
Scheme 34. Synthesis of complexes [Ru(bdmpza)H(CO)2] (11A,B) and [Ru(bdmpza)(CO)(μ2-CO)]2 (12).
The solubility of the hydrido complex [Ru(bdmpza)H(CO)2] (11A,B) is rather poor in most
solvents apart from CHCl3 and CH2Cl2. Unfortunately, the complex decomposes quickly in
CDCl3 by formation of the chlorido complex, a reactivity that was reported for other hydrido
complexes such as [Ru(η5-C5H5)H(CO)(PPh3)].[225] In CD2Cl2 the stability of the complex is
slightly better. Thus, only 1H NMR data could be obtained so far. Nevertheless, ESI-MS data
and elemental analysis prove the formation of 11A,B. Surprisingly, so far it was not possible
to isolate 12 from solutions of the hydrido complex [Ru(bdmpza)H(CO)2] (11A,B), that had
been exposed to air. Obviously, the hydrido complex 11A,B seems to be quite unreactive
regarding oxygen. Even heating under reflux in nonpolar solvents such as n-heptane and
NN N
N
Ru
OO
OC HCO
Me
Me
Me
Me
Hbdmpza[Ru3(CO)12]
HOAc
[Ru(OAc)(CO)2]n
toluene, Δ
THF, Δ- HOAc
Hbdmpza
NN N
N
Ru
OO
OC COH
Me
Me
Me
Me
+
O
NN
NNOO
NN
N
O
Ru
OC
RuCO
CO
OC
Me
Me
Me
Me
Me
Me
Me
Me N
11A 11B
12
47
!Results and Discussion
!! !
applying aerobic conditions did not yield complex 12 but mostly unreacted 11A,B. Thus,
another attempt was undertaken by reacting the acetate polymer catena-[Ru(OAc)(CO)2]n
with Hbdmpza. The polymer catena-[Ru(OAc)(CO)2] is readily available but is also easily
accessible by reacting [Ru3(CO)12] with acetic acid.[228] It has been successfully applied in the
syntheses of various dinuclear ruthenium(I) complexes before.[229-230] Reaction in THF at
reflux for 24 h replaced the acetate of catena-[Ru(OAc)(CO)2]n by bis(3,5-dimethylpyrazol-1-
yl)acetic acid and resulted in the target complex [Ru(bdmpza)(CO)(μ2-CO)]2 (12) in a yield of
30%. The constitution of the molecule is confirmed by elemental analysis as well as by ESI-
MS data in acetonitrile, which show a 100% peak at m/z 405.02 (100) assigned to a
[Ru(bdmpza)(CO)2]+ fragment and a small (4%) molecular ion peak at m/z 810.05. Due to the
low solubility of 12 in all common deuterated solvents, only 1H NMR data could be obtained
so far. As expected for the C2h-symmetric molecule depicted in Scheme 34, only one set of
signals is observed, with the methyl singlet signals at 2.35 (Me3) and 2.62 ppm (Me5). The
pyrazole CH proton is found at 6.04 ppm and the methine proton at 6.31 ppm. In theory at
least three isomeric forms of complex 12 might be possible: (I) terminal trans-CO/μ2-CO
bridged, (II) terminal cis-CO/μ2-CO bridged, (III) nonbridged. Apparently, according to the
NMR data only one of these possible isomeric forms seems to be present in solution. This is
in contrast to [Ru(η5-C5H5)(CO)(μ2-CO)]2, where an equilibrium of various isomeric forms
was reported.[231-234] The bdmpza ligand exhibits its typical IR vibrations at 1673 cm–1 (as-
CO2–) and 1559 cm–1
(C═N) as expected for κ3 coordination. The IR spectrum in solution
(CHCl3) is almost identical with that obtained in a KBr matrix. IR vibrations (CHCl3) at
1978 cm–1 (terminal CO) and 1761 cm–1 (μ2-CO) agree well with those reported for μ2-CO
isomers of [Ru(η5-C5H5)(CO)(μ2-CO)]2 (ν(CO) (CHCl3) 2009 cm–1 (terminal CO) and
Thus, owing to the observed very strong μ2-CO vibration one μ2-CO isomer seems to
dominate in the solid state as well as in solution. Nevertheless, a very weak shoulder around
2010 cm–1 and a weak signal at 1950 cm–1 might indicate traces of a nonbridged species.
Similar results could be obtained with the sterically less demanding bpza ligand, as the
bridged complex could be synthesized.[235] However, the complex [Ru(bpza)Cl(CO)2] (13)
was still missing. 13 can be easily synthesized in analogy to [Ru(bdmpza)Cl(CO)2] (9) by
adding the potassium salt of the bpza ligand and heating under reflux (Scheme 35).
48
!Results and Discussion
!! !
Scheme 35. Synthesis of dicarbonyl complex [Ru(bpza)Cl(CO)2] (13).
The complex [Ru(bdmpza)Cl(CO)2] (9) shows high solubility in CH2Cl2 and CHCl3, in
comparison [Ru(bpza)Cl(CO)2] (13) remains completely insoluble and can only be
characterized as DMSO or methanol solution. 13 shows in the 1H NMR spectrum six protons
for the two pyrazolyl units indicating that the chlorido ligand is positioned trans to a pyrazole
moiety and thus an asymmetric bpza ligand without a mirror plane is isolated. The 13C NMR
spectrum confirms the presence of two asymmetric carbonyl ligands at 194.8 and 193.9 ppm
that are positioned trans to the second pyrazole moiety and the acetate anchor. The complex
shows two intense absorptions at 2081 and 2014 cm–1 in the IR absorption spectrum in a
similar region to 9 (2074 cm–1, 2005 cm–1).[236] Two additional very weak absorptions at 1768
and 1760 cm–1 indicate that traces of the dimeric complex could be present however, attempts
of removal via recrystallization from DMSO did not lead to disappearance. The complex
could further be characterized via ESI-MS experiments showing the presence of the sodium
adduct of 13 as 100% peak at m/z 406.91 (100) assigned to a [Ru(bpza)Cl(CO)2 + Na]+
cluster. Crystals suitable for a single crystal X-ray structure determination were obtained by
dissolving 13 in boiling methanol and afterwards vapor diffusion of Et2O into the methanolic
solution at room temperature. [Ru(bpza)Cl(CO)2] (13) crystallizes as two independent
molecules in the space group C2/c and shows co-crystallization of methanol and Et2O. The
complex shows the arrangement deduced from the NMR data and shows the chlorido ligand
trans to a pyrazole moiety leading to two non-equivalent carbonyl ligands trans to the
remaining pyrazole unit and trans to the carboxylate anchor (Figure 11).
NN N
N
Ru
OO
OC ClCO
NN N
N
CO2H 1. KOtBu2. [RuCl2(CO)2]n
13
THF
49
!Results and Discussion
!! !
Figure 11. Preliminary molecular structure of [Ru(bpza)Cl(CO)2] (13). Hydrogen atoms and solvent molecules have been omitted for clarity.
To elucidate the spectroscopic properties and the binding situation in [Ru(bdmpza)(CO)(μ2-
CO)]2 (12) further, DFT calculations were performed by E. HÜBNER starting from the X-ray
structure determination data. The resulting geometry of the DFT calculations was almost
identical with the geometry of the X-ray structure determination. The spin density of the two
electrons forming the Ru−Ru bond is mainly located at the metal centers and the bridging
carbonyl ligands (Figure 12). Surprisingly, the spin density plot does not resemble the contour
plots of two dz2 orbitals but the contour plots of dxy, dxz or dyz orbitals. This implies that the
Ru−Ru bond is better described as a π bond than as a σ bond. In order to verify the IR signals
of 12, DFT calculations were performed. It is well-known for the chosen B3LYP/6-31G*
DFT functional and basis set, that calculated vibrational frequencies are typically
overestimated in comparison to experimental data. These errors arise from the neglect of
anharmonicity effects, incomplete incorporation of electron correlation, and the use of finite
basis sets in the theoretical treatment.[237] In order to achieve a correlation with observed
spectra, a scaling factor of approximately 0.96 has to be applied.[237] Depending on the
examined vibration, this factor differs slightly even in the same molecule and is usually
greater for lower energies.[238]
O4C41
N21 N11
N22 N12
RuC31
ClO3
O1
O2
50
!Results and Discussion
!! !
Figure 12. Spin density plot regarding the electrons forming the Ru−Ru bond.
We were especially interested in the two carbonyl vibrations, which were predicted (unscaled)
at 2078 cm−1 (terminal CO) and at 1851 cm−1 (μ2-CO). This leads to expected vibrations at
1995 and 1777 cm−1. Both values agree well with the experimental data. In further agreement
with the experimental data, the trans geometry of the bridged isomer of 12 was found to be
the lowest in energy. The energy difference between the bridged and nonbridged species
(Figure 13) was found to be rather small, with ΔE = 22 kJ/mol in comparison to an energy
difference of ΔE = 45 kJ/mol between the cis and trans geometries. The low energy
difference toward the unbridged isomer implies a rather high possibility of finding the
nonbridged isomer in solution, which may agree with the data of the IR spectra discussed
above. The strong asymmetric IR vibrations of the nonbridged CO were predicted (unscaled)
at 2075 and 2047 cm−1, which should result in vibrations around 1992 and 1965 cm−1.
polymer catena-[Ru(OAc)(CO)2]n with Hbdmpza. The poly-mer catena-[Ru(OAc)(CO)2]n is readily available but is alsoeasily accessible by reacting [Ru3(CO)12] with acetic acid.54 Ithas been successfully applied in the syntheses of variousdinuclear ruthenium(I) complexes before.42,49 Reaction in THFat reflux for 24 h replaced the acetate of catena-[Ru(OAc)-(CO)2]n by bis(3,5-dimethylpyrazol-1-yl)acetic acid andresulted in the target complex [Ru(bdmpza)(CO)(μ2-CO)]2(6) in a yield of 30%. The constitution of the molecule isconfirmed by elemental analysis as well as by ESI MS data inacetonitrile, which show a 100% peak at m/z 405.02 (100)assigned to a [Ru(bdmpza)(CO)2]
+ fragment and a small (4%)molecular ion peak at m/z 810.05. Due to the low solubility of6 in all common deuterated solvents, only 1H NMR data couldbe obtained so far. As expected for the C2h-symmetric moleculedepicted in Figure 3, only one set of signals is observed, withthe methyl singlet signals observed at 2.35 (Me3) and 2.62(Me5) ppm. The pyrazole CH proton is found at 6.04 ppm andthe methine proton at 6.31 ppm. In theory at least threeisomeric forms of complex 6 might be possible: (I) terminaltrans-CO/μ2-CO bridged, (II) terminal cis-CO/μ2-CO bridged,(III) nonbridged. Apparently, according to the NMR data onlyone of these possible isomeric forms seems to be present insolution. This is in contrast to the case for [Ru(η5-C5H5)(CO)(μ2-CO)]2, where an equilibrium of variousisomeric forms was reported.50,55 The bdmpza ligand exhibitsits typical IR vibrations at 1673 cm−1 (as-CO2
−) and 1559 cm−1
(CN) as expected for κ3 coordination. The IR spectrum insolution (CHCl3 solvent) is almost identical with that obtainedin a KBr matrix. IR vibrations (CHCl3) at 1978 cm
−1 (terminalCO) and 1761 cm−1 (μ2-CO) agree well with those reportedfor μ2-CO isomers of [Ru(η5-C5H5)(CO)(μ2-CO)]2 (ν(CO)(CHCl3 solvent) 2009 cm−1 (terminal CO) and 1768 cm−1
55b Thus, owing to the observed verystrong μ2-CO vibration one μ2-CO isomer seems to dominatein the solid state as well as in solution. Nevertheless, a veryweak shoulder around 2010 cm−1 and a weak signal at 1950cm−1 might indicate traces of a nonbridged species. Due to thesteric hindrance of the bdmpza ligands and in accord with DFTcalculations (see below), the μ2-CO isomer cis-[Ru(bdmpza)-(CO)(μ2-CO)]2 (isomer II) with cis geometry of the terminalCO ligands seems to be thermodynamically disfavored. Thus,in accordance with the solid-state structure (Figure 3) trans-[Ru(bdmpza)(CO)(μ2-CO)]2 (6) (isomer I) is the mainisomeric form. To elucidate the spectroscopic properties andthe binding situation in [Ru(bdmpza)(CO)(μ2-CO)]2 (6)further, DFT calculations were performed for 6 starting fromthe X-ray structure determination data. The resulting geometryof the DFT calculations was almost identical with the geometryof the X-ray structure determination. The spin density of thetwo electrons forming the Ru−Ru bond is mainly located at themetal centers and the bridging carbonyl ligands (Figure 4).Surprisingly, the spin density plot does not resemble thecontour plots of two dz2 orbitals but the contour plots of dxy,dxz, or dyz orbitals. This implies that the Ru−Ru bond is betterdescribed as a π bond than as a σ bond. In order to verify the IRsignals of 6, DFT calculations on 6 were performed. It is well-known for the chosen B3LYP/6-31G* DFT functional andbasis set that calculated vibrational frequencies are typicallyoverestimated in comparison to experimental data. These errorsarise from the neglect of anharmonicity effects, incompleteincorporation of electron correlation, and the use of finite basis
sets in the theoretical treatment.56 In order to achieve acorrelation with observed spectra, a scaling factor ofapproximately 0.96 has to be applied.56 Depending on theexamined vibration, this factor differs slightly even in the samemolecule and is usually greater for lower energies.57 We wereespecially interested in the two carbonyl vibrations, which werepredicted (unscaled) at 2078 cm−1 (terminal CO) and at 1851cm−1 (μ2-CO). This leads to expected vibrations at 1995 and1777 cm−1. Both values agree well with the experimental data.In further agreement with the experimental data, the transgeometry of the bridged isomer of 6 was found to be the lowestin energy. The energy difference between the bridged andnonbridged (Figure 5) species was found to be rather small,
with ΔE = 22 kJ/mol in comparison to an energy difference ofΔE = 45 kJ/mol between the cis and trans geometries. The lowenergy difference toward the unbridged isomer implies a ratherhigh possibility of finding the nonbridged isomer in solution,which may agree with the data of the IR spectra discussedabove. The strong asymmetric IR vibrations of the nonbridgedCO were predicted (unscaled) at 2075 and 2047 cm−1, whichshould result in vibrations around 1992 and 1965 cm−1.Experiments regarding the oxidation of cyclohexene
mediated by complexes 3 and 4 were carried out inunstabilized, HPLC grade CH2Cl2 under a dinitrogen
Figure 4. Spin density plot regarding the electrons forming the Ru−Ru bond.
Figure 5. Calculated geometry of a nonbridged isomer of 6.
Figure 13. Calculated geometry of a nonbridged isomer of 12.
Preliminary studies by G. TÜRKOGLU revealed the promising potential of
[Ru(bdmpza)Cl(CO)2] (9) and [Ru(2,2-bdmpzp)Cl(CO)2] (10) as epoxidation catalysts, with a
TON for complex 9 up to 20.[223] Even higher TON values might be accessible by
optimization of the reaction conditions or by applying soluble iodosylbenzene derivatives as
oxidant.
Currently the complex [Ru(bpza)Cl(CO)2] (13) is tested in cooperation with the group of S.
MÉNAGE as center for an artificial oxygenase. Therefore, complex 13 was incorporated into an
enzyme pocket and the isolated hybrid was employed in the oxidation of styrene yielding
styrene glycol, which is not accessible by the complex or protein itself. Currently, the single
crystal X-ray structure determination of the hybrid is solved and refined and the parameters of
the catalysis are tuned for a greener reaction.
polymer catena-[Ru(OAc)(CO)2]n with Hbdmpza. The poly-mer catena-[Ru(OAc)(CO)2]n is readily available but is alsoeasily accessible by reacting [Ru3(CO)12] with acetic acid.54 Ithas been successfully applied in the syntheses of variousdinuclear ruthenium(I) complexes before.42,49 Reaction in THFat reflux for 24 h replaced the acetate of catena-[Ru(OAc)-(CO)2]n by bis(3,5-dimethylpyrazol-1-yl)acetic acid andresulted in the target complex [Ru(bdmpza)(CO)(μ2-CO)]2(6) in a yield of 30%. The constitution of the molecule isconfirmed by elemental analysis as well as by ESI MS data inacetonitrile, which show a 100% peak at m/z 405.02 (100)assigned to a [Ru(bdmpza)(CO)2]
+ fragment and a small (4%)molecular ion peak at m/z 810.05. Due to the low solubility of6 in all common deuterated solvents, only 1H NMR data couldbe obtained so far. As expected for the C2h-symmetric moleculedepicted in Figure 3, only one set of signals is observed, withthe methyl singlet signals observed at 2.35 (Me3) and 2.62(Me5) ppm. The pyrazole CH proton is found at 6.04 ppm andthe methine proton at 6.31 ppm. In theory at least threeisomeric forms of complex 6 might be possible: (I) terminaltrans-CO/μ2-CO bridged, (II) terminal cis-CO/μ2-CO bridged,(III) nonbridged. Apparently, according to the NMR data onlyone of these possible isomeric forms seems to be present insolution. This is in contrast to the case for [Ru(η5-C5H5)(CO)(μ2-CO)]2, where an equilibrium of variousisomeric forms was reported.50,55 The bdmpza ligand exhibitsits typical IR vibrations at 1673 cm−1 (as-CO2
−) and 1559 cm−1
(CN) as expected for κ3 coordination. The IR spectrum insolution (CHCl3 solvent) is almost identical with that obtainedin a KBr matrix. IR vibrations (CHCl3) at 1978 cm
−1 (terminalCO) and 1761 cm−1 (μ2-CO) agree well with those reportedfor μ2-CO isomers of [Ru(η5-C5H5)(CO)(μ2-CO)]2 (ν(CO)(CHCl3 solvent) 2009 cm−1 (terminal CO) and 1768 cm−1
55b Thus, owing to the observed verystrong μ2-CO vibration one μ2-CO isomer seems to dominatein the solid state as well as in solution. Nevertheless, a veryweak shoulder around 2010 cm−1 and a weak signal at 1950cm−1 might indicate traces of a nonbridged species. Due to thesteric hindrance of the bdmpza ligands and in accord with DFTcalculations (see below), the μ2-CO isomer cis-[Ru(bdmpza)-(CO)(μ2-CO)]2 (isomer II) with cis geometry of the terminalCO ligands seems to be thermodynamically disfavored. Thus,in accordance with the solid-state structure (Figure 3) trans-[Ru(bdmpza)(CO)(μ2-CO)]2 (6) (isomer I) is the mainisomeric form. To elucidate the spectroscopic properties andthe binding situation in [Ru(bdmpza)(CO)(μ2-CO)]2 (6)further, DFT calculations were performed for 6 starting fromthe X-ray structure determination data. The resulting geometryof the DFT calculations was almost identical with the geometryof the X-ray structure determination. The spin density of thetwo electrons forming the Ru−Ru bond is mainly located at themetal centers and the bridging carbonyl ligands (Figure 4).Surprisingly, the spin density plot does not resemble thecontour plots of two dz2 orbitals but the contour plots of dxy,dxz, or dyz orbitals. This implies that the Ru−Ru bond is betterdescribed as a π bond than as a σ bond. In order to verify the IRsignals of 6, DFT calculations on 6 were performed. It is well-known for the chosen B3LYP/6-31G* DFT functional andbasis set that calculated vibrational frequencies are typicallyoverestimated in comparison to experimental data. These errorsarise from the neglect of anharmonicity effects, incompleteincorporation of electron correlation, and the use of finite basis
sets in the theoretical treatment.56 In order to achieve acorrelation with observed spectra, a scaling factor ofapproximately 0.96 has to be applied.56 Depending on theexamined vibration, this factor differs slightly even in the samemolecule and is usually greater for lower energies.57 We wereespecially interested in the two carbonyl vibrations, which werepredicted (unscaled) at 2078 cm−1 (terminal CO) and at 1851cm−1 (μ2-CO). This leads to expected vibrations at 1995 and1777 cm−1. Both values agree well with the experimental data.In further agreement with the experimental data, the transgeometry of the bridged isomer of 6 was found to be the lowestin energy. The energy difference between the bridged andnonbridged (Figure 5) species was found to be rather small,
with ΔE = 22 kJ/mol in comparison to an energy difference ofΔE = 45 kJ/mol between the cis and trans geometries. The lowenergy difference toward the unbridged isomer implies a ratherhigh possibility of finding the nonbridged isomer in solution,which may agree with the data of the IR spectra discussedabove. The strong asymmetric IR vibrations of the nonbridgedCO were predicted (unscaled) at 2075 and 2047 cm−1, whichshould result in vibrations around 1992 and 1965 cm−1.Experiments regarding the oxidation of cyclohexene
mediated by complexes 3 and 4 were carried out inunstabilized, HPLC grade CH2Cl2 under a dinitrogen
Figure 4. Spin density plot regarding the electrons forming the Ru−Ru bond.
Figure 5. Calculated geometry of a nonbridged isomer of 6.
4.3 Ruthenium Heteroscorpionate Complexes with Aminophenol Based
Ligands
The oxidative ring cleavage of substituted aromatic compounds such as catechols and
o-aminophenols is most commonly performed by mononuclear non-heme iron
dioxygenases.[239-241] Some play important roles in human metabolism, for example in
tryptophan degradation. 3-Hydroxyanthranilate (HAA) is O2-mediated cleaved by the HAA-
3,4-dioxygenase (HAD) and reacts to quinolinate (Scheme 36).[242-243]
Scheme 36. Catalyzed reaction from 3-hydroxyanthranilate (HAA) to quinolinate.[242, 244]
In 2012 A. FIEDLER et al. reported the first synthetic intermediate of this enzyme in form of
the Fe2+–ISQ (ISQ = iminobenzosemiquinonate) complex.[244] Reaction of the Tp based iron
complex [(Ph2Tp)Fe(OBz)] with the sterically demanding aminophenol ligand 2-amino-4,6-di-
tert-butylphenol (tBuAPH2; 2-amino-4,6-di-tert-butylphenolate = tBuAPH–) yields the κ2
coordinated complex [(Ph2Tp)Fe(2+)(tBuAPH)] which mimics the enzyme-substrate complex.[244]
Reaction of this complex with 2,4,6-tri-tert-butylphenoxy radical (TBBP
•) leads to an iron(II)
complex bound to an ISQ radical.[244] The resulting Fe2+–SQ (SQ = semiquinone) complex is
often invoked as intermediate for the mechanism of catechol dioxygenases although all other
relevant models feature [Fe3+–catecholate]+ units.[245-247] Further one-electron oxidation using
Ag[SbF6] allows isolation of a cationic complex that shows an oxidation state that can be
attributed to a [Fe3+–ISQ–]+ or [Fe2+–IBQ]+ complex (IBQ = iminobenzoquinonate).[244]
Recently T. PAINE et al. reported the first functional model for 2-aminophenol dioxygenases,
namely APD (2-aminophenol-1,6-dioxygenase) and HAD.[248] The non-heme complex [(6-
Me3-TPA)Fe2+(4-tBuHAP)](ClO4) (6-Me3-TPA = tris(6-methyl-2-pyridylmethyl)amine, tBuHAP = 2-amino-4-tert-butylphenol), which is readily available from a one-pot synthesis
shows reactivity with dioxygen and formation of 4-tert-butyl-2-picolinic acid through C–C
bond cleavage of 2-amino-4-tert-butylphenol.[248]
N
NH2OH-O2C
CO2H-O2C
NH2
COOHCHO
-O2CO2
HAD
non-enzymatic
- H2O
53
!Results and Discussion
!! !
Due to the high sensibility of Fe2+ based heteroscorpionate complexes bearing aminophenol
ligands it was decided to start from the commonly used precursor [Ru(bdmpza)Cl(PPh3)2]
(14). M. KECK synthesized during his master thesis a dark blue complex bearing the tBuAPH2
ligand. Due to the lack of a single crystal X-ray structure determination it was supposed from
analytical data that the complex should be of the general formula [Ru(bdmpza)-
(tBuISQ)(PPh3)]Cl (Scheme 37). Strong antiferromagnetic spin-spin coupling might lead in this
case to diamagnetic coupling in NMR spectroscopy allowing the observation of the imino
proton at 14.19 ppm in the 1H NMR spectrum. It was discussed that two theoretical binding
modes for the tBuAPH– could occur with the imino and hydroxo functionality positioned trans
to the carboxylate anchor and one pyrazole unit and vice versa.
Scheme 37. Synthesis of [Ru(bdmpza)(tBuISQ)(PPh3)]Cl (15A, 15B) by M. KECK and its supposed structures a) and b).
The high solubility of complex 15 in polar and nonpolar solvents led to difficulties in
obtaining crystals suitable for a single crystal X-ray structure determination. Nevertheless,
dissolving complex 15 in a hot mixture of CH2Cl2 and n-hexane, layered with pure n-hexane
led to the formation of crystals.
The result of a single crystal X-ray structure determination shows that the predicted binding
mode is not in agreement with the observed structure (Figure 14). Instead of a κ2 coordinated tBuISQ ligand a κ1 coordination of a possibly neutral tBuIBQ or monoanionic tBuISQ occurs,
which is unprecedented in literature. The APH2 based ligand coordinates with the imino
moiety trans to a pyrazole unit of the bdmpza ligand. A PPh3 and a chlorido ligand occupy the
two remaining coordination sites and the absence of counter ions indicates the formation of a
neutral complex.
NN N
N
Ru
OO
Ph3P PPh3Cl
Me
Me
Me
Me
KtBuAPH
NN N
N
Ru
OO
PPh3
Me
Me
Me
Me
NHO
tBu
tBu
NN N
N
Ru
OO
PPh3
Me
Me
Me
Me
OHN
tBu
tBu
a) b)
- Cl-
54
!Results and Discussion
!! !
Figure 14. Molecular structure of [Ru(bdmpza)Cl(tBuISQ/tBuIBQ)(PPh3)] (15). Thermal ellipsoids are drawn at the 50% probability level. Most hydrogen atoms have been omitted for clarity. Selected bond lengths (Å) and angles (deg): Ru–N(11) = 2.1495(17), Ru–N(21) = 2.098(2), Ru–O(1) = 2.1037(14), Ru–P(1) = 2.3318(6), Ru–Cl(1) = 2.3956(6), Ru–N(61) = 1.962(2), N(61)–C(61) = 1.310(3); N(11)–Ru–N(21) = 85.29(7), O(1)–Ru–N(21) = 85.72(7), O(1)–Ru–N(11) = 85.57(6), O(1)–Ru–P(1) = 85.76(4), P(1)–Ru–Cl(1) = 99.45(2), P(1)–Ru–N(61) = 87.43(5), O(1)–Ru–N(61) = 97.17(7), N(11)–Ru–P(1) = 170.81(5), N(21)–Ru–Cl(1) = 92.35(5), Cl(1)–Ru–N(61) = 84.39(6), Ru–N(61)–C(61) = 138.72(17).
For “non-innocent” APH2 ligands several oxidation states are known which in conclusion
allow the interaction with the metal center in form of redox chemistry.[249] Starting from the
deprotonated tBuAP2– the first one electron oxidation leads to the aforementioned anionic tBuISQ radical, which in the next oxidation step forms the tBuIBQ compound (Scheme 38).[249]
Scheme 38. Forms of the tert-butyl substituted aminophenol ligand seperated by one-electron redox steps.[249]
Cl1
N21N11
N12 N22
Ru
H61
C1
O62
N61
P1
C62
C61
O1
C66
C63
O2
C65
C64
ab
c
tBu
tBu
O
NH
tBu
tBu
O
NH
tBu
tBu
O-
NH-
- e-
+ e-- e-
+ e-
tBuAP2- tBuISQ tBuIBQ
55
!Results and Discussion
!! !
Bond length [Å] 15 16 [Ru(acac)2(ISQ)] [RuCl(terpy)(ISQ)]+
Ru–N(61) 1.962(2) 1.962(5) 1.906(3) 1.942(8)
N(61)–C(61) 1.310(3) 1.293(8) 1.340(4) 1.312(12)
C(61)–C(62) 1.498(3) 1.485(9) 1.439(4) 1.433(13)
C(62)–O(62) 1.231(2) 1.221(9) 1.291(4) 1.270(11)
C(62)–C(63) 1.470(3) 1.441(11) 1.424(5) 1.416(13)
C(63)–C(64) 1.350(3) 1.280(11) 1.363(6) 1.322(13)
C(64)–C(65) 1.449(3) 1.425(11) 1.409(7) 1.446(15)
C(65)–C(66) 1.347(3) 1.351(10) 1.345(6) 1.377(14)
C(66)–C(61) 1.432(3) 1.406(9) 1.411(5) 1.433(13)
Table 4. Selected bond lengths of the ruthenium iminoquinone complexes 15, 16, [Ru(acac)2(ISQ)][250] and [RuCl(terpy)(ISQ)]ClO4.[251]
Bond length [Å] 15 16 [Ru(bdmpza)Cl(PPh3)2] [Ru(bdmpza)Cl2(PPh3)]
Table 5. Selected bond lengths of the ruthenium iminoquinone complexes 15 and 16 and closely related heteroscorpionate complexes [Ru(bdmpza)Cl(PPh3)2] (14) and [Ru(bdmpza)Cl2(PPh3)].[168]
In comparison with literature known [Ru3+–ISQ] complexes the bond lengths of the ISQ
ligands are in good agreement, although the alternation between located double and single
bonds is more pronounced for complex 15 (Table 4).[250-251] In addition the oxygen–carbon
bond is shortened due to the lack of interaction with the ruthenium center. To further
understand the oxidation state of the metal center, a look at the heteroscorpionate ligand is
useful. The closely related compounds [Ru(bdmpza)Cl(PPh3)2] (14) and [Ru(bdmpza)-
Cl2(PPh3)] show ruthenium(II) and ruthenium(III) centers, respectively, which in comparison
to complex 15, highlight the proposed ruthenium(III) structure (Table 5).[168] Especially the
bond Ru–N(11) which is positioned trans to a PPh3 ligand and thus not directly influenced by
ligand exchange shows similar values for 15 and the ruthenium(III) complex. This
56
!Results and Discussion
!! !
emphasizes the assumption of a [Ru3+–ISQ] compound, although they do not allow a final
decision between an ISQ or IBQ ligand. Questionable remains the reaction pathway as the
metal and the ligand both undergo one-electron oxidation in absence of an oxidant. Possibly
contamination with oxygen might play a key role and lead to the low reported yields.
Hence it was decided to synthesize the analogous complex based on unsubstituted
2-aminophenol (APH2) similar to [Ru(acac)2(ISQ)][250] and [RuCl(terpy)(ISQ)]ClO4.[251]
Deprotonation of APH2 with potassium tert-butylate followed by addition of [Ru(bdmpza)Cl-
(PPh3)2] (14) at room temperature led to the formation of a dark solution (Scheme 39). After
two chromatography steps the complex was obtained as a dark blue solid in extremely low
yields.
Scheme 39. Synthesis of ISQ or IBQ complex [Ru(bdmpza)Cl(ISQ/IBQ)(PPh3)] (16).
In accordance with the expected structure ESI-MS experiments showed the presence of the
molecular ion (m/z 753.12 (5%) M+) and the closely related sodium adduct (m/z 742.15
(100%) [M – Cl + Na]+). The 13C NMR and 1H NMR spectrum show the pattern of an
asymmetric [Ru(bdmpza)] fragment with four independent signals for the methyl substituents.
The imino proton results in a signal at 14.77 ppm in the 1H NMR spectrum. The carbonyl
moiety of the ISQ/IBQ ligand gives rise to a signal at 171.3 ppm in the 13C NMR spectrum
indicating a similar binding motif as 15 with an uncoordinated keto moiety. The lower
solubility of 16 in nonpolar solvents in comparison to 15 allows crystallization from CH2Cl2
solution layered with n-hexane.
The result of a single crystal X-ray structure determination shows that the previously
observed κ1 binding motif also occurs for the sterically less demanding ISQ/IBQ ligand. The
arrangement around the ruthenium center is similar to 15 with the imino moiety of the
ISQ/IBQ ligand positioned trans to one pyrazole donor forcing the PPh3 ligand in trans
NN N
N
Me
Me
Me
MeRu
OO
ClPh3PNH
O
HO NH2
THF
16 (IBQ)
1. KOtBu2. 14
NN N
N
Me
Me
Me
MeRu
OO
ClPh3PNH
O
16 (ISQ)
or
57
!Results and Discussion
!! !
position of the second pyrazole donor and the remaining chlorido ligand trans to the
carboxylate anchor (Figure 15).
Figure 15. Molecular structure of [Ru(bdmpza)Cl(ISQ/IBQ)(PPh3)] (16). Thermal ellipsoids are drawn at the 50% probability level. Most hydrogen atoms have been omitted for clarity. Selected bond lengths (Å) and angles (deg): Ru–N(21) = 2.102(5), Ru–N(11) = 2.133(5), Ru–O(1) = 2.088(4), Ru–P(1) = 2.3300(18), Ru–Cl(1) = 2.3835(17), Ru–N(61) = 1.962(5), N(61)–C(61) = 1.293(8); N(11)–Ru–N(21) = 84.9(2), O(1)–Ru–N(21) = 86.4(2), O(1)–Ru–N(11) = 85.96(18), O(1)–Ru–P(1) = 85.26(12), P(1)–Ru–Cl(1) = 99.56(6), P(1)–Ru–N(61) = 88.53(16), O(1)–Ru–N(61) = 97.9(2), N(11)–Ru–P(1) = 170.61(14), N(21)–Ru–Cl(1) = 91.18(18), Cl(1)–Ru–N(61) = 84.06(16), Ru–N(61)–C(61) = 137.9(5).
The bond lengths in the ISQ/IBQ ligand of complex 16 are listed in Table 4 and are in good
agreement with the closely related complex 15 and especially the two unsubstituted ISQ/IBQ
complexes by G. LAHIRI et al.[250-251] However, especially the double bond character of C(63)–
C(64) is more pronounced with a bond length of 1.280(11) Å for 16 compared to 1.350(3) Å
for [Ru(bdmpza)Cl(tBuISQ/tBuIBQ)(PPh3)] (15). The comparison with the ruthenium(II) and
ruthenium(III) complexes [Ru(bdmpza)Cl(PPh3)2] (14) and [Ru(bdmpza)Cl2(PPh3)] (Table 5)
indicates that the complex can best be described as [Ru3+–ISQ] due to the shortened
ruthenium nitrogen bonds between the bdmpza ligand and the ruthenium center indicating a
ruthenium(III) center (2.133(5) Å and 2.102(5) Å for 16 in comparison to 2.184(3) Å and
2.109(3) Å for [Ru(bdmpza)Cl2(PPh3)]), although a [Ru2+–IBQ] system can not be ruled out.
Cl
O62
H61
N11
C62N61
N21Ru
N12
C61
N22
C63
P1C64
C66C65O1
O2
a b
c
58
!Results and Discussion
!! !
4.4 Carbon-rich Ruthenium Allenylidene Complexes
Parts of this chapter have been published:
Strinitz, F.; Waterloo, A.; Tucher, J.; Hübner, E.; Tykwinski, R. R.; Burzlaff, N., Eur. J.
Inorg. Chem. 2013, 5181-5186.
Starting from the previous results of the BURZLAFF group concerning the formation of
carbene, vinylidene and allenylidene complexes and their reaction behavior it was decided to
investigate the possible substitution patterns of the propargyl alcohols employed.[61, 189]
As mentioned in chapter 2.5 only two structural isomers can be observed for
[Ru(bdmpza)Cl(PPh3)2] (14) based cumulenylidene complexes. For easy nomenclature the
isomer with the cumulenylidene moiety trans to the pyrazole unit is indicated with “A” and
the isomer with the cumulenylidene moiety trans to the carboxylate anchor with “B”. In a
similar way the bdmpza ligand is numbered for NMR and single crystal X-ray structure
determination purposes. The numbering is dependent on the position of the PPh3 ligand and
follows the depicted scheme (Scheme 40).
Scheme 40. Numbering scheme used for cumulenylidene complexes.
Typically the straight forward synthesis of ruthenium allenylidene complexes following
Selegue´s route starts from substituted propargyl alcohols leading to the dissociation of water
from the intermediary hydroxyvinylidene complexes. Depending on the used metal fragment,
the dissociation of water requires the addition of catalytical amounts of acid, which allows the
isolation of the labile vinylidene species. Nevertheless, for the facially coordinating bdmpza
ligand, it has been observed that the intermediary vinylidene complexes can be detected via 1H NMR due to the characteristic vinylidene proton but the complex cannot be isolated and
reacts directly to the allenylidene complex.[61] P. DIXNEUF et al. have previously shown that a
reversible addition of sodium methoxide to the cationic allenylidene complex trans-[(dppm)2-
ClRu(═C═C═CPh2]PF6 yields the corresponding neutral alkynyl complex trans-[(dppm)2Cl-
Ru(–C≡C–CPh2(OMe)].[252] In an attempt to synthesize an isolable vinylidene complex based
on [Ru(bdmpza)Cl(PPh3)2] (14) it was decided to start from the 3,5-di-tert-butyl substituted
methoxy ether of the conventional used diphenyl propargyl alcohol. This compound has
recently been shown to be an effective building block for forming stabilized organic
cumulenes with up to ten carbon atoms.[253] Furthermore, the presence of the ether group
might enhance the formation of the vinylidene complex due to the reduced leaving potential
of the methanol unit in comparison to the free hydroxyl group.
The synthesis of the intended neutral vinylidene complex started from [Ru(bdmpza)Cl(PPh3)2]
(14) and excess propargyl alcohol 1,1-bis-(1,3-di-tert-butylphenyl)-1-methoxy-2-propyne in
THF (Scheme 41). Initially, no apparent color change was observed. After 3 d, however, a
strong purple color was visible and the formation of the allenylidene complex
[Ru(bdmpza)Cl(═C═C═C(PhtBu2)2)(PPh3)] (19A, 19B) was completed by heating for 4 h
under reflux. Due to the facial coordinating motif of the bdmpza ligand, the formation of two
structural isomers was observed, as has been reported previously.[61] The relatively high
stability (no degradation over days was observed) allowed the separation via column
chromatography under aerobic conditions affording a purple (19A) and a red isomer (19B). 13C NMR spectra revealed for 19B characteristic signals for a ruthenium allenylidene complex
at 314.7 ppm (d, 2JCP = 18.3 Hz, Cα), 234.6 ppm (Cβ) and 152.4 ppm (Cγ) for the allenylidene
unit, as well as a singlet in the 31P NMR spectrum at 34.5 ppm.
60
!Results and Discussion
!! !
Scheme 41. Synthesis of ruthenium vinylidene intermediate [Ru(bdmpza)Cl(═C═CH(COMe(PhtBu2)(PPh3)] (17), carbonyl complex [Ru(bdmpza)Cl(CO)(PPh3)] (18A, 18B) and allenylidene complex [Ru(bdmpza)Cl(═C═C═C(PhtBu2)2)(PPh3)] (19A, 19B).
NN N
N
Me
Me
Me
MeRu
OO
C ClPh3P
NN N
N
Me
Me
Me
MeRu
OO
CClPh3PC
CC
CtBu
tBu tBu
tBu
tBu
tBu
tButBu
19A 19B
NN N
N
Me
Me
Me
MeRu
OO
ClPh3P PPh3
OMe
H
tBu
tBu tBu
tBu+
NN N
N
Me
Me
Me
MeRu
OO
CClPh3P C H
C
tBu
tBu
tBu
tBuOMe
NN N
N
Me
Me
Me
MeRu
OO
ClPh3P C
+
NN N
N
Me
Me
Me
MeRu
OO
ClPh3P C
17
18A
18B
14
O
O
61
!Results and Discussion
!! !
For compound 19A the 13C NMR spectrum revealed strong contamination with the carbonyl
complex [Ru(bdmpza)Cl(CO)(PPh3)] (18A, carbonyl trans to pyrazole), which can be formed
by oxygen induced bond cleavage of the vinylidene intermediate as has been shown
previously for [Ru(bdmpza)Cl(═C═CHPh)(PPh3)].[61] Unfortunately, all attempts to avoid
formation of 18A during synthesis or separation of 19A via column chromatography provided
only impure product. Nevertheless, the assignment of the two structural isomers A and B to
the respective symmetry and positions was accomplished based on comparisons to previously
reported two-dimensional NMR experiments (ROESY) and APT 13C NMR measurements.[61]
Namely, type B complexes show cross-peaks between the methyl substituents in the 3- and 3´
positions with the aryl protons of the allenylidene moiety in the ROESY spectrum, which
indicate an arrangement trans to the carboxylate anchor.[61]
Cyclic voltammetric analyses were performed on the precursor [Ru(bdmpza)Cl(PPh3)2] (1)
and on the resulting allenylidene complex 19B. The exhibited electrochemical properties are
summarized in Chapter 8.2. The reversibility of the redox processes shows strong dependence
on the solvent used, as voltammograms recorded in acetonitrile lead to irreversible oxidations
and reductions indicating side reactions of the allenylidene complexes with acetonitrile. The
voltammograms recorded in dichloromethane with nBu4NPF6 (0.1 M) as electrolyte and
referenced to the ferrocene/ferrocenium couple as internal standard at a scan rate of 100 mV/s
feature exclusively reversible and quasi-reversible processes. For the used precursor
[Ru(bdmpza)Cl(PPh3)2] (14) one reversible oxidation at 394 mV can be observed, which is
attributed to the Ru(II)/Ru(III) couple. For the ferrocene/ferrocenium couple literature reports
a peak separation of 78 mV (83 mV in our setup) in dichloromethane,[254] which is a good
indication that the peak separation of 73 mV and the peak current ratio ipa/ipc = 0.80 for 14
confirm a reversible one-electron oxidation. For the first ruthenium allenylidene complex
[Ru(bdmpza)Cl(═C═C═C(PhtBu2)2)(PPh3)] (19B) two quasi-reversible redox processes can
be observed. The oxidation of the ruthenium center happens at a lower potential of 265 mV
compared to 14, indicating a possible electron releasing effect of the used allenylidene ligand
in comparison to the PPh3 ligand. The reduced peak current ratio indicates however, that the
reversibility is lowered in comparison to 14. A second redox process at –1631 mV can be
attributed to the quasi-reversible reduction (ipa/ipc = 0.67) of the allenylidene moiety as
reported previously for the systems [Cl(dppe)2Ru(═C═C═CPh2)]PF6 (–1.03 V) and [Cl(16-
TMC)Ru(═C═C═CPh2)]PF6 (–1.27 V).[124, 255] In comparison it is obvious that the two redox
62
!Results and Discussion
!! !
couples of the neutral bdmpza allenylidene complex 19B are more cathodic in comparison to
the aforementioned cationic allenylidene complexes (Δ = 0.60 V/0.36 V).
The intense color of the allenylidene complexes can best be characterized via UV/Vis
absorption spectroscopy. Hence the comparison with the previously reported bdmpza based
allenylidene complexes [Ru(bdmpza)Cl(═C═C═CPh2)(PPh3)] and [Ru(bdmpza)Cl-
(═C═C═C(tol)2)(PPh3)] is a good starting point. For type B isomers absorption maxima of
495 and 507 nm have been reported for solutions in CH2Cl2.[61] The closely related complex
[Ru(bdmpza)Cl(═C═C═C(PhtBu2)2)(PPh3)] (19B) shows a maximum at 506 nm with a molar
extinction coefficient of approximately 13000 L mol–1cm–1 (Figure 16). This transition was
assigned to a metal-to-ligand charge-transfer (MLCT) for the diphenyl and ditolyl substituted
allenylidene complexes in literature.[61] However, newer calculations performed for complexes
within this work indicate, that this transition corresponds to a metal-perturbed π-π*
transition.[256] A further transition that can be observed in the NIR region at 1024 nm with an
extremely low extinction coefficient indicating a forbidden transition, which might belong to
the MLCT in which the HOMO–1, HOMO and LUMO have been involved for the
pentacenequinone based allenylidene complexes (Figure 17).[256] Due to lack of a single
crystal X-ray structure determination no TD-DFT calculations were performed on this
complex and no definite answer can be given on this topic.
63
!Results and Discussion
!! !
Figure 16. Absorption spectrum of 19B in CH2Cl2.
Figure 17. Magnification of the relevant parts of the NIR region displaying forbidden MLCTs for 19B.
400 600 800 1000 1200 1400 1600
0
5000
10000
15000
20000ε([L(m
ol.1(cm
.1]
W ave leng th([nm]
800 1000 1200 1400 16000
50
100
150
200
ε([L(m
ol.1(cm
.1]
W ave leng th([nm]
64
!Results and Discussion
!! !
Figure 18. Molecular structure of [Ru(bdmpza)Cl(CO)(PPh3)] (18B). Thermal ellipsoids are drawn at the 50% probability level. Hydrogen atoms have been omitted for clarity. Selected bond lengths (Å) and angles (deg): Ru–N(11) = 2.156(3), Ru–N(21) = 2.082(3), Ru–O(1) = 2.119(2), Ru–P(1) = 2.3360(10), Ru–Cl(1) = 2.3880(11), Ru–C(3) = 1.923(5), C(3)–O(3) = 1.009(5); N(11)–Ru–N(21) = 82.68(11), O(1)–Ru–N(11) = 85.60(10), O(1)–Ru–N(21) = 87.15(10), O(1)–Ru–P(1) = 86.27(7), P(1)–Ru–Cl(1) = 90.05(4), P(1)–Ru–C(3) = 95.27(11), O(1)–Ru–C(3) = 177.34(13), N(21)–Ru–P(1) = 97.08(8), N(11)–Ru–Cl(1) = 89.87(8), Cl(1)–Ru–C(3) = 87.26(11), Ru–C(3)–O(3) = 176.0(4).
Attempts to obtain crystals of 19B suitable for X-ray diffraction by layering a solution in
dichloromethane with n-hexane leads, within several weeks, to bond cleavage of the
allenylidene unit with conservation of the relative geometry, providing complex 18B as
illustrated in Figure 18. In comparison to the previously reported carbonyl complex 18A, the
carbonyl ligand in 18B is in trans position to the carboxylate. Thus, the chlorido and
triphenylphosphine ligands are in trans position to the pyrazole units. This observation is
unexpected since the direct carbonylation of [Ru(bdmpza)Cl(PPh3)2] (14) and decomposition
of the resulting vinylidene complexes leads exclusively to the carbonyl complex 18A with the
carbonyl trans to a pyrazole.[61] The ruthenium(II) center is facially coordinated by the
bdmpza ligand resulting in a slightly distorted octahedral geometry caused by the rather rigid
and strained coordination geometry of the heteroscorpionate ligand. The Ru–C(3)
(1.923(5) Å) bond is slightly elongated and C(3)–O(3) (1.009(5) Å) contracted in comparison
to the other structural isomer 18A (Ru–C(3) = 1.831(5) Å, C(3)–O(3) = 1.151(6) Å) as a
O2
O1
C2
Cl1P1
C1
Ru1
N12
N11
N22
N21
C3
O3
ab
c
65
!Results and Discussion
!! !
result from the trans-orientation of the carboxylate group. This is in contrast to the pyrazolyl
donor, which is a σ and π donor as well as a π acceptor and shows no trans influence, as
previously discussed and supported by DFT calculations for the dissociation energies of
N,N,O ligands.[173] This observation suggests that the steric demand of the four tert-butyl
groups reduces the stability in comparison to the analogous unsubstituted diphenyl
allenylidene complex, i.e., substituted with only two phenyl rings.
66
!Results and Discussion
!! !
4.4.2 Fluorene Based Allenylidene Complexes
Closely related to the diphenyl allenylidene complexes are systems bearing a fluorene group
on Cγ. Fluorene based allenylidene complexes have previously shown to inhibit the
rearrangement of the allenylidene moiety into the corresponding indenylidene complex.[167]
NMR spectroscopic experiments have shown that [(η6-p-cymene)RuCl(═C═C═(FN))(PCy3)]-
OTf (FN = fluorenyl) reacts upon addition of HOTf to the alkenylcarbyne, but no further
transformation could be observed.[167] Furthermore polyfluorenes are organic electro-
luminescent materials that have been applied to devices in photonics and optoelectronics.[257-
259] Following the route described above, addition of excess amounts of 9-ethynylfluoren-9-ol
to [Ru(bdmpza)Cl(PPh3)2] (14) led to the formation of a deep purple solution (Scheme 42).
The increased stability of the obtained structural isomers allowed separation via column
chromatography under aerobic conditions yielding a purple (20A, allenylidene trans to
pyrazole) and a red isomer (20B, allenylidene trans to carboxylate).
Scheme 42. Synthesis of bdmpza based ruthenium allenylidene complexes [Ru(bdmpza)Cl(═C═C═(FN))(PPh3)] (FN = fluorenyl) (20A, 20B).
For 20A the characteristic signals for the allenylidene chain are observed in the 13C NMR
spectrum at 300.6 (d, 2JC,P = 27.6 Hz, Cα), 236.4 (d, 3JC,P = 4.6 Hz, Cβ) and 141.0 ppm (Cγ)
with doublets for Cα and Cβ caused by coupling with the phosphorus atom of the
triphenylphosphine ligand. Furthermore, the IR spectrum shows an intense band at 1910 cm–1
in the IR spectrum corresponding to the cumulenylidene ligand. The 31P NMR spectrum
NN N
N
Me
Me
Me
MeRu
OO
C ClPh3P
NN N
N
Me
Me
Me
MeRu
OO
CClPh3PC
CC
C
20A 20B
NN N
N
Me
Me
Me
MeRu
OO
ClPh3P PPh3
+
OH
H+ THF
14
67
!Results and Discussion
!! !
consists of one singlet at 34.6 ppm and ESI-MS experiments showed the presence of the
protonated monocationic species (m/z 825.16 (100%) MH+). Similar spectroscopic values are
obtained for the second isomer 20B with signals at 314.4 (d, 2JC,P = 19.3 Hz, Cα), 256.2 (Cβ)
and 141.6 ppm (Cγ) in the 13C NMR spectrum (P–C coupling observable for Cα), while the 31P NMR spectrum shows one singlet at 30.9 ppm, which is shifted upfield in comparison to
20A. The IR spectrum shows the cumulenic stretch at 1903 cm–1 a slightly lower value than
that for 20A and the ESI-MS experiments reveal the main observable signal that is consistent
with the protonated monocationic species (m/z 825.16 (100%) MH+). The assignment of the
geometries was in accordance to the previous synthesized bdmpza based allenylidene
complexes and could be verified by single crystal X-ray structure determinations of both
isomers. Crystals were obtained from solutions in CH2Cl2 layered with n-hexane. 20A and
20B represent the first single-crystal X-ray structure determinations of fluorene based
allenylidene complexes.[107, 166-167, 252, 260-261] Complex 20A crystallizes as an racemic mixture
(space group Pbca) as [Ru(bdmpza)Cl(═C═C═(FN))(PPh3)] × H2O with one water molecule
bound via a hydrogen bond to the carbonyl moiety of the carboxylate unit (Figure 19). The
molecular structure exhibits a slightly distorted octahedral geometry at the Ru(II) center with
the allenylidene positioned trans to a pyrazole donor, the triphenylphosphine trans to the
second pyrazole donor and the chlorido ligand trans to the carboxylate anchor. In comparison
to the diphenyl allenylidene complex [Ru(bdmpza)Cl(═C═C═CPh2)(PPh3)], the bdmpza
ligand of 20A shows only slight deviations.[61] The Ru–C(61) bond is with 1.865(3) Å similar
to the analogous [Ru(bdmpza)Cl(═C═C═CPh2)(PPh3)] (1.886(5) Å)[61] and the octahedral Tp
but considerably longer than in pentacoordinated 16 VE ruthenium allenylidene complexes
like [RuCl2(═C═C═CPh2)(PCy3)2] (1.794(11) Å).[95] The allenylidene chain deviates slightly
from the linear geometry (∠Ru–C(61)–C(62) = 175.1(3)°, ∠C(61)–C(62)–C(63) = 172.7(3)°)
with the fluorenyl moiety remaining in plane with the C═C═C moiety.
68
!Results and Discussion
!! !
Figure 19. Molecular structure of [Ru(bdmpza)Cl(═C═C═(FN))(PPh3)] (20A). Thermal ellipsoids are drawn at the 50% probability level. Hydrogen atoms and one water molecule have been omitted for clarity. Selected bond lengths (Å) and angles (deg): Ru–N(11) = 2.172(2), Ru–N(21) = 2.231(2), Ru–O(1) = 2.094(2), Ru–P(1) = 2.3121(9), Ru–Cl(1) = 2.3552(9), C(63)–C(67) = 1.472(4), C(63)–C(64) = 1.470(4), C(64)–C(65) = 1.406(4), C(65)–C(66) = 1.466(5), C(66)–C(67) = 1.407(4), Ru–C(61) = 1.865(3), C(61)–C(62) = 1.247(4), C(62)–C(63) = 1.352(4); N(11)–Ru–N(21) = 83.05(9), O(1)–Ru–N(11) = 86.32(9), O(1)–Ru–N(21) = 82.43(9), O(1)–Ru–P(1) = 92.72(6), P(1)–Ru–Cl(1) = 92.09(3), P(1)–Ru–C(61) = 85.66(9), O(1)–Ru–C(61) = 92.39(11), N(21)–Ru–P(1) = 99.15(7), N(11)–Ru–Cl(1) = 89.19(7), Cl(1)–Ru–C(61) = 96.26(9), Ru–C(61)–C(62) = 175.1(3), C(61)–C(62)–C(63) = 172.7(3).
As has been described for the structural related butatriene 4-(9H-fluoren-9-ylidene)-2-
methylbuta-2,3-dienal (COH(CH3)C═C═C═(FN)),[262] the bond lengths of the five-membered
ring of the fluorenyl unit of 20A show less bond length alternation than the parent
fluorenone,[263] which indicates a strong delocalization of the electron density from the
allenylidene moiety to the fluorenyl unit. Also noticeable are strong solid-state π-π stacking
interactions between two neighboring fluorenyl units (Figure 20), with an interplanar distance
of 3.45 Å as calculated from the least-squares plane generated from the carbon atoms of one
fluorenyl moiety to the plane of its neighbor.
Cl1
N21N11
N22N12
Ru1
C1
P1C61C62
O1
C2
C64C63
O2
C65 C67C66
abc
69
!Results and Discussion
!! !
Bond length [Å] 20A 20B Fluorene COH(CH3)C═C═C═(FN)
C63–C64 1.470(4) 1.469(7) 1.486 1.469
C64–C65 1.406(4) 1.408(7) 1.390 1.390
C65–C66 1.466(5) 1.463(8) 1.475 1.471
C66–C67 1.407(4) 1.407(7) 1.390 1.404
C67–C63 1.472(4) 1.466(7) 1.486 1.465
Table 6. Selected bond lengths of the five membered rings of the ruthenium allenylidene complexes 20A and 20B, fluorene and the structural related butatrien COH(CH3)C═C═C═(FN).[262-263]
Figure 20. π–π stacking interactions between two molecules of 20A a) top view and b) side view.
a) b)
70
!Results and Discussion
!! !
Figure 21. Molecular structure of [Ru(bdmpza)Cl(═C═C═(FN))(PPh3)] (20B). Thermal ellipsoids are drawn at the 50% probability level. Hydrogen atoms and two molecules CH2Cl2 have been omitted for clarity. Selected bond lengths (Å) and angles (deg): Ru–N(11) = 2.146(4), Ru–N(21) = 2.081(4), Ru–O(1) = 2.144(3), Ru–P(1) = 2.3552(12), Ru–Cl(1) = 2.4077(11), C(63)–C(67) = 1.466(7), C(63)–C(64) = 1.469(7), C(64)–C(65) = 1.408(7), C(65)–C(66) = 1.463(8), C(66)–C(67) = 1.407(7), Ru–C(61) = 1.855(5), C(61)–C(62) = 1.247(7), C(62)–C(63) = 1.363(7); N(11)–Ru–N(21) = 83.54(15), O(1)–Ru–N(11) = 85.55(13), O(1)–Ru–N(21) = 86.49(13), O(1)–Ru–P(1) = 85.75(9), P(1)–Ru–Cl(1) = 88.10(4), P(1)–Ru–C(61) = 95.27(15), O(1)–Ru–C(61) = 178.55(16), N(21)–Ru–P(1) = 98.00(11), N(11)–Ru–Cl(1) = 90.02(11), Cl(1)–Ru–C(61) = 89.85(14), Ru–C(61)–C(62) = 175.7(4), C(61)–C(62)–C(63) = 177.2(5).
The second structural isomer 20B shows solid-state characteristics similar to that of 20A,
including a distorted octahedral geometry (Figure 21). Compound 20B crystallizes in space
group P–1 as a racemic mixture. The chlorido and PPh3 ligand are now positioned trans to
pyrazole donors, placing the allenylidene unit trans to the carboxylate anchor. Therefore, the
respective bond lengths differ slightly in comparison to isomer 20A. For example, there is
shortening of the Ru–C(61) bond to 1.855(5) Å and a similar contraction of the Ru–N(11)
bond, which can be explained by the reduced trans influence in this structural isomer because
of the π accepting pyrazole and allenylidene ligand are no longer positioned trans to each
other. Additionally, the allenylidene chain is slightly less distorted from linearity than in 20A
(∠Ru–C(61)–C(62) = 175.7(4)°, ∠C(61)–C(62)–C(63) = 177.2(5)°). This can be explained by
the different packing motif in the solid state. The change in the relative positions around the
ruthenium center also leads to reduced distances between the fluorenyl moiety and one phenyl
ring of the triphenylphosphine ligand of the complex. This close proximity of the phenyl rings
N21
C65
C64
N22
P1C63
C66
C62C61
C67
C1
Ru1
O1
C2
O2 N12
N11
Cl1
ab
c
71
!Results and Discussion
!! !
appears to hinder the π-π stacking interaction between two neighboring fluorenyl units in the
solid state, which also seems to result in the smaller angles in ∠Ru–Cα–Cβ and ∠Cα–Cβ–Cγ of
20A.
The fluorene based allenylidene complexes [Ru(bdmpza)Cl(═C═C═(FN))(PPh3)] (20A, 20B)
show a reversible oxidation at 389 mV/371 mV indicating a positive shift in peak potential in
comparison to 19B (Chapter 8.2). Interesting is the appearance of two redox processes at
negative voltages. While the process at –1273 mV (20A, 20B) is again attributed to the
reduction of the allenylidene moiety and appears more anodic in comparison to 19B, a second
quasi-reversible/reversible process appears at –1932 mV (20A) and –1937 mV (20B), that we
assign to the reduction of the fluorenyl moiety. Although the position of the allenylidene
moiety trans to the pyrazole or carboxylate moiety strongly influences the physical and
chemical properties of the complex no obvious differences in electrochemical properties
could be observed for these two structural isomers.
The UV/Vis absorption spectra of [Ru(bdmpza)Cl(═C═C═(FN))(PPh3)] (20A, 20B) recorded
in CH2Cl2 show several intense absorptions (Figure 22). The strong absorptions below 400 nm
can be attributed to ligand centered π–π* transitions involving the bdmpza and PPh3 ligand.
As mentioned previously the strong absorption at 546 nm (20A) or 517 nm (20B) with molar
extinction coefficients around 15000 L mol–1 cm–1 correspond to a metal-perturbed π-π*
transition of the allenylidene moiety. Again weak transitions can be observed in the NIR
region for both complexes with a signal at 1053 nm with a shoulder at 919 nm for 20B
(Figure 23). For 20A, two distinct signals at 1201 nm and 944 nm can be observed. These
transitions can be assigned to HOMO → LUMO and HOMO–2 → LUMO excitations, which
are MLCT transitions (Table 7).
Observed values Calculated values
Compound Wavelength [nm]
Absorption coefficient [M–1 cm–1]
Wavelength [nm]
Transition dipole moment [debye]
Transition
20A 1053
919
226
175
906
813
0.12
0.54
HOMO ! LUMO
HOMO–2 ! LUMO
Table 7. Calculated and measured transitions for 20A in the NIR region.
72
!Results and Discussion
!! !
Figure 22. Absorption spectrum of 20A (black) and 20B (grey) in CH2Cl2.
Figure 23. Magnification of the relevant parts of the NIR region displaying forbidden MLCTs for 20A (black) and 20B (grey); signal caused by CH2Cl2 is indicated by *.
400 600 800 1000 1200 1400 1600
0
5000
10000
15000
20000
ε([L(m
ol.1(cm
.1]
W ave leng th([nm]
800 1000 1200 1400 16000
100
200
ε'[L'm
ol-1'cm
-1]
W ave leng th'[nm]
*
73
!Results and Discussion
!! !
To further understand the absorption spectra TD-DFT calculations (time-dependent DFT)
have been performed by E. HÜBNER for complex 20A that explain the absorptions at the edge
of the NIR region. Especially, the HOMO → LUMO and HOMO–2 → LUMO transitions
with low transition dipole moments of 0.12 and 0.54 debye seem to correspond to forbidden
MLCT transitions. The calculated geometries emphasize that the LUMO is delocalized over
the ruthenium center as well as the entire allenylidene moiety and the fluorenyl unit, whereas
the HOMO and HOMO–2 are mainly located on the ruthenium center and Cα and Cβ (Figure
24). Furthermore, the calculated absorptions are in good agreement with the measured values
(Table 7).
LUMO HOMO
HOMO–1 HOMO–2
Figure 24. Orbital diagrams of the LUMO (–2.9 eV), HOMO (–5.2 eV), HOMO–1 (–5.3 eV) and HOMO–2 (–5.4 eV) of [Ru(bdmpza)Cl(═C═C═(FN))(PPh3)] (20A).
74
!Results and Discussion
!! !
4.4.3 Anthraquinone Based Allenylidene Complexes
Although heteroatom substituted allenylidene complexes based on 4,5-diazafluorene[126-127]
and cyclopentadithiophene[128] have been reported, little is known about allenylidene
complexes with larger polyaromatic substituents. To date, only few complexes are discussed
in literature, such as, for example, trans, trans-[(dppe)2RuCl(C═C═C(bianth)C═C═C)ClRu-
(dppe)2](OTf)2 that is based on the extended conjugated system [9,9′]bianthracenylidene-
10,10′-dione are discussed in literature.[113] Notably in this system, the close proximity of the
protons of two anthrone units of the bianthrone moiety results significant strain and a non-
planar organic spacer. For further studies on the π-π stacking interactions between
polyaromatic allenylidene units and electron transfer properties between the ruthenium(II)
center and the organic substituents it was decided to look into anthraquinone based
allenylidene complexes.
The use of anthraquinone derivatives is manifold ranging from the anthraquinone oxidation
process for hydrogen peroxide production to dye precursors.[264-265] A recent highlight was the
construction of a metal-free organic-inorganic aqueous flow battery by B. HUSKINSON and M.
MARSHAK et al.[266] 9,10-Anthraquinone-2,7-disulphonic acid (AQDS) undergoes rapid and
reversible two-electron two-proton reduction in sulfuric acid. Combination of the couple
quinone/hydroquinone with the couple Br2/Br– with glassy carbon electrodes allows the
formation of promising flow batteries for electrical energy storage at greatly reduced cost.[266]
O. WENGER et al. recently published a bpy (bpy = 2,2´-bipyridine) based ruthenium complex
bearing an anthraquinone moiety in its periphery.[267] The thermodynamics and kinetics of the
intramolecular electron transfer between the [Ru(bpy)3]2+ core and the anthraquinone unit
linked via one up to three xylene linkers in the complex [Ru(bpy)2(bpy–xyn–AQ)]2+ (xy = p-
xylene, n = 0-3) was investigated. It was shown that electron transfer between the ruthenium
core and the anthraquinone unit can be triggered by photoexcitation leading to the charge
separated state. The solvent influence on the electron transfer indicates a finite proton density
transfer rather than a full PCET (proton coupled electron transfer).[267] This redox behavior
shows similarities to the electron transfer cascade in photosynthetic reaction centers of
bacteria.[268]
For the formation of the first anthraquinone based allenylidene complex the anthraquinone
(AQ) based 10-ethynyl-10-hydroxyanthracen-9-one (24) is promising. The required precursor
75
!Results and Discussion
!! !
24 is known to the literature, however, an appealing high yield synthesis is missing.[269-271] The
classic approach to 24 begins with the formation of a lithium acetylide, via reaction of
gaseous acetylene with lithium in liquid ammonia followed, by the addition of anthraquinone
leading to the monosubstituted propargyl alcohol 24. This synthesis can be mimicked by the
addition of the commercially available suspension of sodium acetylide in xylenes to
anthraquinone. These procedures, however, offer low yields of 24, and they are also
unattractive because of difficult purification due to the low solubility of 24. In analogy to the
pentacenequinone based synthesis of the monopropargyl alcohol the addition of trimethylsilyl
(TMS) acetylene followed by the desilylation allows the high yield synthesis of ketone 24
(Scheme 43).[256, 272] In the first step of the reaction, a substoichiometric amount of n-BuLi is
added to trimethylsilylacetylene in dry THF. In the following, step the lithium acetylide was
added dropwise to an excess amount of anthraquinone (21) in THF to avoid the formation of
the bis-adduct 9,10-bis((trimethylsilyl)ethynyl)-9,10-dihydroanthracene-9,10-diol (22). After
aqueous workup, the unreacted anthraquinone can be removed via column chromatography on
silica with CH2Cl2 as eluent yielding the ketone 23. The desymmetrization of anthraquinone
via acetylide addition leads to the appearance of four aromatic signals in the 1H NMR
spectrum of 23 at 8.17, 8.08, 7.71 and 7.51 ppm, with second-order coupling patterns
characteristic of an ortho-substituted arene. Furthermore, singlets for the alcohol and TMS
groups are observed at 3.16 and 0.16 ppm, respectively. The 13C NMR spectrum shows the
moiety at 183.1 ppm and the three characteristic signals for a propargyl alcohol groups at
106.7, 91.5 and 66.4 ppm. The removal of the TMS group from 23 to give 24 leads to no
change in the coupling pattern of the aryl protons in the 1H NMR spectrum, while the
appearance of an additional signal corresponding to the alkyne proton at 2.71 ppm can be
observed concurrent with the loss of the singlet of the TMS group. The solubility of 24 is
significantly decreased, however, and a 13C NMR spectrum can only be recorded in DMSO-d6
and shows the terminal alkyne carbon appearing at 76.1 ppm and the keto moiety at
182.4 ppm.
76
!Results and Discussion
!! !
Scheme 43. Synthesis of the TMS protected ketone 23, the deprotected propargyl alcohol 24 and the undesired bisadduct 22.
Similar to the method described for the fluorenyl based systems, the preparation of the
corresponding ruthenium allenylidene complex (25A, 25B) was carried out by using an
excess amount of propargyl alcohol 24 (Scheme 44). The formation of the intense purple
color and the appearance of a peak at 1880 cm–1 in the IR spectrum, characteristic for
allenylidene complexes, confirmed the successful conversion of the anthraquinone based.
Separation of the two structural isomers was achieved following the procedure described
above.
O
O O
O
OH
HO
OH
OH
TMS
TMS
H
TMS
1. n-BuLi / TMS-acetylene (< 1.0 eq.)2. H2O
THF
KOHMeOH, H2O
21 23
24
22
77
!Results and Discussion
!! !
Scheme 44. Synthesis of [Ru(bdmpza)Cl(═C═C═(AO))(PPh3)] (AO = anthrone) (25A, 25B).
For the first isomer 25A the allenylidene carbon atoms Cα (292.1 ppm, d, 2JC,P = 26.8 Hz), Cβ
(251.0 ppm, d, 3JC,P = 5.0 Hz) and Cγ (141.4 ppm, d, 4JC,P = 3.0 Hz) appear as doublets in the 13C NMR spectrum including long-range 4JC,P coupling between the triphenylphosphine ligand
and Cγ. A singlet is found in the 31P NMR spectrum at 30.1 ppm resulting from the
triphenylphosphine ligand, which also supports the suggested structure. ESI-MS experiments
again show the major observable signal resulting from the protonated monocationic species
(m/z 863.14 (100%) MH+). For the structural isomer 25B with the allenylidene unit positioned
trans to the carboxylate, the 13C NMR spectrum shows a downfield shift for Cα (309.6 ppm, d, 2JC,P = 19.8 Hz) and Cβ (277.0 ppm) relative to 25A. The third allenylidene carbon Cγ appears
almost unchanged at 140.5 ppm, and a signal at 29.1 ppm in the 31P NMR spectrum confirms
the triphenylphosphine ligand. The change in coordination geometry gives rise to an increase
of 16 cm–1 (1896 cm–1) in the cumulene vibration compared to 25A. For the other ruthenium
allenylidene complexes reported within this work, no clear trend for the allenylidene
absorptions in the IR spectrum could be observed regarding A and B type isomers. However,
in this case the difference might be explainable by a reduced linearity of the allenylidene
moiety as described in the following part. ESI-MS experiments showed that the change in
coordination in this complex strongly influence the stability of 25B although the
monocationic complex can be observed the intensity is low (m/z 863.14 (4%) MH+). The
major signals result from the dissociation of the chlorido ligand followed by decarboxylation
NN N
N
Me
Me
Me
MeRu
OO
C ClPh3P
NN N
N
Me
Me
Me
MeRu
OO
CClPh3PC
CC
C
25A25B
NN N
N
Me
Me
Me
MeRu
OO
ClPh3P PPh3
++ THF
OH
O
H
O
O
14
78
!Results and Discussion
!! !
of the bdmpza ligand (m/z 783.18 (88%) [M – Cl – CO2]+) and can be followed by addition of
one solvent molecule (m/z 824.21 (100%) [M – Cl – CO2 + MeCN]+).
Figure 25. Molecular structure of [Ru(bdmpza)Cl(═C═C═(AO))(PPh3)] (25A). Thermal ellipsoids are drawn at the 50% probability level. Hydrogen atoms and solvent molecules have been omitted for clarity. Selected bond lengths (Å) and angles (deg): Ru–N(11) = 2.138(3), Ru–N(21) = 2.193(3), Ru–O(1) = 2.077(2), Ru–P(1) = 2.3325(8), Ru–Cl(1) = 2.3761(8), Ru–C(31) = 1.868(3), C(31)–C(32) = 1.243(5), C(32)–C(33) = 1.362(5); N(11)–Ru–N(21) = 83.07(10), O(1)–Ru–N(11) = 87.20(9), O(1)–Ru–N(21) = 84.01(10), O(1)–Ru–P(1) = 83.87(6), P(1)–Ru–Cl(1) = 100.47(3), P(1)–Ru–C(31) = 86.76(10), O(1)–Ru–C(31) = 97.22(12), N(11)–Ru–P(1) = 170.16(8), N(21)–Ru–Cl(1) = 85.76(8), Cl(1)–Ru–C(31) = 92.65(10), Ru–C(31)–C(32) = 177.0(3), C(31)–C(32)–C(33) = 175.2(4).
The assignment of the relative geometry could be verified for both complexes from X-ray
crystal structure analysis that were performed on crystals obtained from solutions in CH2Cl2
layered with n-hexane (Figure 25). Complex 25A (allenylidene trans to pyrazole) crystallizes
as racemic mixture [Ru(bdmpza)Cl(═C═C═(AO))(PPh3)] × CH2Cl2 in space group P–1 with
one solvent molecule disordered and parted over three positions. The distorted octahedral
geometry is affected by the strained bdmpza ligand that shows values comparable to the
fluorenyl allenylidene complex 20A discussed above. The change from the central 5-
membered ring in the fluorenyl moiety to the 6-membered ring in the anthraquinone based
Cl1
N21 N11
N22
Ru1
N12
C31P1C32
C1
C33
O1
C2
C36
O3
O2
ab
c
79
!Results and Discussion
!! !
system in complex 25A leads to increased steric repulsion between the anthrone moiety and
the triphenylphosphine ligand. This results in a smaller bond angle ∠O(1)–Ru–P(1) (83.9°)
compared to 92.7° in 20A and considerable greater angle ∠P(1)–Ru–Cl(1) (100.5°) in
comparison to 92.1° (20A). The allenylidene unit itself shows rather unremarkable values of
d(Ru–C(31)) = 1.868(3) Å, ∠Ru–C(31)–C(32) = 177.0(3)°, and ∠C(31)–C(32)–
C(33) = 175.2(4)°.
Figure 26. π–π stacking interactions between two molecules of 25A a) top view and b) side view.
Similar to complex 20A, π-π stacking interactions between two anthrone units are observed
with approximately two thirds of the anthrone area affected resulting in a mean interplanar
distance of 3.37 Å (Figure 26), as calculated between the least-squares plane generated from
the carbon atoms of one anthrone moiety to the plane of its neighbor. For comparison, it is
noted that anthraquinone shows a similar slipped stack arrangement in the solid state, with a
mean interplanar distance of 3.48 Å.[273]
For the second structural isomer 25B (allenylidene trans to carboxylate) crystals of
[Ru(bdmpza)Cl(═C═C═(AO))(PPh3)] (Figure 27) in space group P–1 were obtained. Similar
to the structures discussed above the positioning of the π donor and acceptor pyrazole trans to
the chlorido ligand and the allenylidene trans to the carboxylate anchor leads to a loss of the
trans-influence for both ligands. This change in the coordination sphere results in a shortened
Ru–N(21) bond with 2.0740(18) Å and a similar Ru–C(61) bond with 1.862(2) Å. The bond
angles of the allenylidene chain are with ∠Ru–C(61)–C(62) = 174.4(2)° and ∠C(31)–C(32)–
C(33) = 169.8(3)° slightly reduced compared to 25A in contrast to the fluorenyl system
(20A/20B).
a) b)
80
!Results and Discussion
!! !
Figure 27. Molecular structure of [Ru(bdmpza)Cl(═C═C═(AO))(PPh3)] (25B). Thermal ellipsoids are drawn at the 50% probability level. Hydrogen atoms have been omitted for clarity. Selected bond lengths (Å) and angles (deg): Ru–N(11) = 2.1355(19), Ru–N(21) = 2.0740(18), Ru–O(1) = 2.1367(15), Ru–P(1) = 2.3546(6), Ru–Cl(1) = 2.4100(6), Ru–C(61) = 1.862(2), C(61)–C(62) = 1.232(3), C(62)–C(63) = 1.354(3); N(11)–Ru–N(21) = 83.87(7), O(1)–Ru–N(11) = 85.11(7), O(1)–Ru–N(21) = 86.94(7), O(1)–Ru–P(1) = 85.94(5), P(1)–Ru–Cl(1) = 87.68(2), P(1)–Ru–C(61) = 95.87(7), O(1)–Ru–C(61) = 178.13(8), N(11)–Ru–P(1) = 170.75(5), N(21)–Ru–Cl(1) = 174.12(5), Cl(1)–Ru–C(61) = 87.84(7), Ru–C(61)–C(62) = 174.4(2), C(61)–C(62)–C(63) = 169.8(3).
The explanation for this bent allenylidene unit can be deduced from the solid state packing
motif of two neighboring complexes. The space filling model clarifies that only a small
overlap of two anthrone units is observed due to the presence of one phenyl ring (dark grey)
of the triphenylphosphine ligand on top of the anthrone moiety, which thus blocks the π-π
interactions as observed for 25A and forcing the neighboring allenylidene chain into a slightly
bent structure in the solid state (Figure 28).
Cl1
N11
N12
O3
O2
O1
C67
C2
Ru1
C63
C61C62
C1
P1
N22
N21
a
b
c
81
!Results and Discussion
!! !
Figure 28. Space-filling model of complex 25B; phenyl ring highlighted in dark grey, anthrone moiety highlighted in light grey.
For the anthraquinone based allenylidene complexes [Ru(bdmpza)Cl(═C═C═(AO))(PPh3)]
(25A, 25B) the influence of the anthrone unit is clearly visible and the difference between
both structural isomers is obvious (Chapter 8.2). Three reversible redox processes can be
observed and can be attributed to the Ru(II)/Ru(III) couple (466 mV/641 mV), the
allenylidene moiety (–1013 mV/–870 mV) and the anthrone moiety (–1479 mV/–1315 mV).
Apparently, the facile reduction of the anthrone unit leads to a positive shift for all three redox
processes and this effect is especially prominent in the B-type isomer with the allenylidene
moiety trans to the pyrazole unit.
The UV/Vis absorption spectra of [Ru(bdmpza)Cl(═C═C═(AO))(PPh3)] (25A, 25B)
recorded in CH2Cl2 show several intense absorptions (Figure 29). The strong absorptions
below 400 nm can again be attributed to ligand centered π–π* transitions involving the
bdmpza and PPh3 ligand. As mentioned previously the strong absorption at 578 nm (25A) or
550 nm (25B) with molar extinction coefficients around 14000 L mol–1 cm–1 correspond to a
metal-perturbed π-π* transition of the allenylidene moiety and are bathochromic shifted in
comparison to 20A/B. Again weak transitions can be observed in the NIR region for both
complexes with two signals at 1331 and 939 nm for 25B (Figure 30). For 25A one broad
signal at 1131 nm can be detected. These transitions can be assigned to HOMO → LUMO
and HOMO–1 → LUMO excitations, which are MLCT transitions (Table 8).
82
!Results and Discussion
!! !
Figure 29. Absorption spectrum of 25A (black) and 25B (grey) in CH2Cl2.
Figure 30. Magnification of the relevant parts of the NIR region displaying forbidden MLCTs for 25A (black) and 25B (grey); signal caused by CH2Cl2 is indicated by *.
400 600 800 1000 1200 1400 1600
0
5000
10000
15000
20000
25000
30000
ε)[L)m
ol/1)cm
/1]
W ave leng th)[nm]
800 1000 1200 1400 16000
50
100
150
200
ε([L(m
ol.1(cm
.1]
W ave leng th([nm]
*
83
!Results and Discussion
!! !
Observed values Calculated values
Compound Wavelength [nm]
Absorption coefficient [M–1 cm–1]
Wavelength [nm]
Transition dipole moment [debye]
Transition
25A 1131 184 997
879
0.85
0.25
HOMO ! LUMO
HOMO–1 ! LUMO
25B
1331
939
155
204
1204
905
0.68
0.24
HOMO ! LUMO
HOMO–1 ! LUMO
Table 8. Calculated and measured transitions for 25A and 25B in the NIR region.
For further understanding the absorption spectra TD-DFT calculations have again been
performed by E. HÜBNER for complexes 25A and 25B, which explain the absorptions at the
edge of the NIR region. Especially, the HOMO → LUMO and HOMO–1 → LUMO
transitions with low transition dipole moments between 0.24 and 0.85 debye seem to
correspond to forbidden MLCT transitions for both structural isomers (Table 8). The
calculated geometries emphasize that the LUMO is delocalized over the ruthenium center as
well as the entire allenylidene moiety and the anthrone unit for both complexes (Figure 31
and Figure 32). The B type isomer seems to show a stronger involvement of the anthrone unit
to the calculated orbital in comparison to the A type complex. The HOMO and HOMO–1 are
mainly located on the ruthenium center, Cα and Cβ for both isomers. The comparison of the
calculated absorptions with the measured values (Table 8) confirm the assumption of the
MLCT between the calculated orbitals.
84
!Results and Discussion
!! !
LUMO HOMO
HOMO–1 HOMO–2
Figure 31. Orbital diagrams of the LUMO (–3.2 eV), HOMO (–5.3 eV), HOMO–1 (–5.5 eV) and HOMO–2 (–5.7 eV) of [Ru(bdmpza)Cl(═C═C═(AO))(PPh3)] (25A).
85
!Results and Discussion
!! !
LUMO HOMO
HOMO–1 HOMO–2
Figure 32. Orbital diagrams of the LUMO (–3.2 eV), HOMO (–5.2 eV), HOMO–1 (–5.5 eV) and HOMO–2 (–5.8 eV) of [Ru(bdmpza)Cl(═C═C═(AO))(PPh3)] (25B).
86
!Results and Discussion
!! !
4.4.4 Pentacenequinone Based Allenylidene Complexes
Functionalized acenes have proven to be good candidates for small-molecule semiconductor
applications and have been widely explored over the past decade.[274-276] In comparison to
field-effect transistors (FETs) based on single-crystal, polycrystalline or amorphous silicon,
those based on acene molecules allow the realization of large-area, mechanically flexible, and
low-cost devices.[277] Likewise, the direct precursors to pentacenes, namely pentacene-
quinones, are potentially useful organic semiconductors in their own right.[278] To date, the
functionalization of the framework of pentacenes and pentacenequinones has focused on
organic substituents and is accomplished mainly by appending alkyl, aryl, and alkyne residues
to the framework to influence the HOMO–LUMO gap and packing motif in the crystalline
state. [274-276, 279] The organometallic chemistry of pentacenes has been little explored,[280] and
organometallic derivatives of pentacenequinone were unknown.
As a starting point for the pentacenone component, the known 13-hydroxy-13-
[(triisopropylsilyl)ethynyl]pentacen-6-one (27)[272] was converted in cooperation with the
group of R. TYKWINSKI into the propargyl alcohol 13-ethynyl-13-hydroxypentacen-6-one (28)
by desilylation with TBAF (tetra-n-butylammonium fluoride) in THF (Scheme 45).[272]
Scheme 45. Synthesis of the TIPS (triisopropylsilyl) protected alcohol 27 and deprotection of the propargyl alcohol 28 (reaction conditions: I) 1. n-BuLi, TIPS-acetylene, THF, 2. H2O; II) TBAF, THF).
A similar approach as for the previously described complexes was used to obtain the
corresponding heteroscorpionate allenylidene bdmpza complexes 29A and 29B. To the
precursor [Ru(bdmpza)Cl(PPh3)2] (14), 1.5 equiv. of 13-ethynyl-13-hydroxypentacen-6-one
O O OHO
TIPS
OHO
H
I) II)
27 2826
87
!Results and Discussion
!! !
(28) was added in THF. This led to the formation of a deep-blue solution after 4 d of stirring
at room temperature (Scheme 46).
Scheme 46. Synthesis of [Ru(bdmpza)Cl(═C═C═(PCO))(PPh3)] (PCO = pentacenone) (29A, 29B).
As a result of the facial κ3 coordination of the bdmpza ligand the two expected structural
isomers were formed, namely 29A and 29B. The neutral 18 VE allenylidene complexes 29A and 29B are rather stable towards oxygen and water. Therefore, the complexes were purified
by column chromatography under aerobic conditions, although a clean separation of the two
isomers was hindered in comparison to all previously describe bdmpza based allenylidene
complexes.[61] However, it was possible to isolate trace amounts of the major isomer 29A
from a mixture of 29A and 29B by column chromatography on silica gel
(CH2Cl2/acetone/n-hexane, 1:1:1, v/v/v). Surprisingly, all attempts to obtain pure 29B by
chromatographic separation resulted in samples that contained traces of isomer 29A.
Nevertheless, crystals suitable for an X-ray structure determination of 29B were obtained by
slow diffusion of n-hexane into a solution of the complex 29B in CH2Cl2 (Figure 33).
NN N
N
Me
Me
Me
MeRu
OO
C ClPh3P
NN N
N
Me
Me
Me
MeRu
OO
CClPh3PC
CC
C
29A 29B
NN N
N
Me
Me
Me
MeRu
OO
ClPh3P PPh3
+
+THF
28
O
O
OH
O
H
88
!Results and Discussion
!! !
Figure 33. Molecular structure of [Ru(bdmpza)Cl(═C═C═(PCO))(PPh3)] (29B). Thermal ellipsoids are drawn at the 50% probability level. Hydrogen atoms and two molecules dichloromethane have been omitted for clarity. Selected bond lengths (Å) and angles (deg): Ru–N(11) = 2.0881(16), Ru–N(21) = 2.1394(17), Ru–O(1) = 2.1458(14), Ru–P(1) = 2.3435(5), Ru–Cl(1) = 2.3884(5), Ru–C(101) = 1.861(2), C(101)–C(102) = 1.263(3), C(102)–C(103) = 1.363(3); N(11)–Ru–N(21) = 83.59(7), O(1)–Ru–N(11) = 85.01(6), O(1)–Ru–N(21) = 86.76(6), O(1)–Ru–P(1) = 86.23(4), P(1)–Ru–Cl(1) = 89.036(18), P(1)–Ru–C(101) = 97.39(6), O(1)–Ru–C(101) = 173.68(7), N(11)–Ru–P(1) = 99.60(5), N(21)–Ru–Cl(1) = 87.14(5), Cl(1)–Ru–C(101) = 95.20(6), Ru–C(101)–C(102) = 167.78(17), C(101)–C(102)–C(103) = 163.2(2).
The ruthenium complex 29B exhibits an octahedral geometry that is slightly distorted due to
the facial coordinating N,N,O ligand with the allenylidene unit positioned trans to the
carboxylate and the PPh3 as well as the chlorido ligand trans to the pyrazole donors. In
comparison with other ruthenium allenylidene complexes, the Ru–C3 chain is extremely bent
with angles ∠Ru–C(101)–C(102) = 167.78(17)° and ∠C(101)–C(102)–C(103) = 163.2(2)°
(Table 9). These distorted angles might be caused by crystal packing effects and are
unprecedented for mononuclear ruthenium allenylidene complexes.[281] The distortion of the
sp carbon chain can be explained by strong π–π stacking interactions in the solid state that
Cl1 N21N22
C103
C102
C106
C101
O3
Ru1
C1O1
C2
O2
P1
N12N11
a
bc
89
!Results and Discussion
!! !
force the allenylidene unit into this bent structure. In the crystal structure, two pentacenone
units are stacked with less than half a phenyl ring slippage and a mean interplanar distance of
3.63 Å (Figure 34).
Figure 34. π–π stacking interactions between two molecules of 29B from a) side view and b) top view.
This is in good agreement with the distance reported for 27 (3.60 Å), but is longer than in
pentacenequinone (ca. 3.4 Å).[282-283] These favorable π-stacking interactions could lead to
aggregation in solution, which would explain the difficult separation of the isomers by
column chromatography. In addition, the pentacenone units are arranged parallel throughout
the crystal lattice with a diagonal distance of around 3.9/4.0 Å between two neighboring
pentacenone units, giving a structure resembling a staircase. TD-DFT calculations by E.
HÜBNER on a single molecule and neglecting π interactions led to an almost linear
90
!Results and Discussion
!! !
allenylidene moiety (Table 9). This supports the assumption that the considerable deviations
in the solid state are a result of strong π–π stacking interactions. Cyclic voltammetric (CV)
analysis of complex 29B revealed several ligand based and one ruthenium based redox
transitions (see Chapter 8.2). The Ru(II)/Ru(III) couple is totally irreversible and appears at
+0.92 V. A weak reversible oxidation is observed at +0.12 V and derives most likely from the
pentacenone unit. Two reversible peaks at –0.48 and –0.88 V indicate facile reduction of the
pentacenone and allenylidene moieties. The latter peak at –0.88 V agrees well with values of
other ruthenium allenylidene complexes reported in the literature.[125] Further reversible
reductions are observed at –1.40 and –1.83 V. In summary, these data seem to indicate that
29B has potential electron-acceptor properties.
The UV/Vis spectrum of 29B recorded in CH2Cl2 is depicted in Figure 35 and a magnification
of the relevant parts of the NIR region is given in Figure 36. The strong absorption bands
(55000 L mol–1 cm–1 for 29B) below 400 nm have been assigned to ligand-centered (LC) π–
π* transitions involving the PPh3 and bdmpza ligands. An additional metal-perturbed π–π*
transition can be observed at 605 nm (ε ≈ 14000 L mol–1 cm–1), and the broad transitions at
lower wavelengths (830–1400 nm) can be attributed to HOMO–1→LUMO and
HOMO→LUMO excitations, which are MLCT transitions.
Figure 35. Absorption spectrum of 29B in CH2Cl2.
400 600 800 1000 1200 14000
10000
20000
30000
40000
50000
60000
ε)[L)m
ol/1)cm
/1]
W ave leng th)[nm]
91
!Results and Discussion
!! !
Figure 36. Magnification of the relevant parts of the NIR region displaying forbidden MLCTs for 29B.
To further understand the UV/Vis data, DFT calculations were performed on 29B by E.
HÜBNER. TD-DFT calculations of the excited states revealed two absorption bands located at
the edge of the NIR region (830–1400 nm, 370 L mol–1 cm–1). The two bands were assigned to
metal-to-ligand charge-transfer transitions. The absorption band at lower wavelength for the
first excited state was calculated to be at 1114 nm (i.e., 1.11 eV) and correlates mainly to a
HOMO→LUMO transition. The second absorption band was calculated to be at 857 nm (i.e.,
1.45 eV) and correlates to a HOMO–1→LUMO transition. Both occupied orbitals are mainly
located on the ruthenium center in the d orbitals with some of the electron density extending
towards the adjacent chlorido ligand as well as the oxygen atom of the carboxylate group of
the bdmpza ligand. The lowest unoccupied orbital is delocalized mainly throughout the
pentacenone ligand (Figure 37). The high degree of delocalization might explain the long-
wave absorption bands because, as a consequence, the LUMO is expected to have a rather
low energy, leading to a small energy difference between the occupied and unoccupied
orbitals. For both transitions, the dipole moment was calculated to be rather small (0.65 and
0.17 debye), which indicates forbidden transitions. The HOMO→LUMO gap, calculated as
the difference between the calculated orbital energies of the ground state (DFT) of 29B, is
2.0 eV, and this correlates reasonably well with the long-wavelength absorption bands found
experimentally. The longest wavelength absorption bands obtained by the more accurate TD-
900 1000 1100 1200 1300 14000
50
100
150
200
250
300
350
400
ε([L(m
ol.1(cm
.1]
W ave leng th([nm]
*
92
!Results and Discussion
!! !
DFT (1.1 to 1.4 eV), however, match the experimental CV data better, revealing a HOMO–
LUMO gap of around 1.4 eV, which agrees as well as the UV/Vis data (0.9 to 1.5 eV).
LUMO HOMO
HOMO–1
Figure 37. Orbital diagrams of the HOMO–1 (–5.4 eV), HOMO (–5.1 eV), and LUMO (–3.06 eV) of [Ru(bdmpza)Cl(═C═C═(PCO))(PPh3)] (29B).
93
!Results and Discussion
!! !
Experimental Calculated
Distances [Å] 29A 29B 29A 29B
Ru–Cα 1.859(3) 1.861(2) 1.868 1.875
Cα–Cβ 1.244(5) 1.263(3) 1.263 1.265
Cβ–Cγ 1.365(5) 1.363(3) 1.359 1.359
Angles [°]
Ru–Cα–Cβ 172.8(3) 167.78(17) 174.41 174.51
Cα–Cβ–Cγ 179.0(4) 163.2(2) 176.40 175.75
Table 9. Selected distances and angles for complexes 29A and 29B determined from the X-ray crystal structure and theoretical calculations (LACVP*/B3LYP).
The same TD-DFT calculations were performed for complex 29A. For the three calculated
occupied orbitals HOMO–2, HOMO–1 and HOMO the orbitals are again mainly located on
the ruthenium center in the d orbitals with some of the electron density extending towards the
adjacent chlorido ligand as well as the oxygen atom of the carboxylate group of the bdmpza
ligand (Figure 38).
The LUMO is delocalized mainly throughout the pentacenone ligand. However, the change in
geometry from the allenylidene unit positioned trans to the carboxylate unit (29B) to trans to
the pyrazole unit (29A) seems to reduce the delocalization of the HOMO within the
pentacenone unit (Figure 38).
94
!Results and Discussion
!! !
LUMO HOMO
HOMO–1 HOMO–2
Figure 38. Orbital diagrams of the LUMO (–3.0 eV), HOMO (–5.2 eV), HOMO–1 (–5.3 eV) and HOMO–2 (–5.5 eV) of [Ru(bdmpza)Cl(═C═C═(PCO))(PPh3)] (29A).
In comparison the second structural isomer (29A) shows the expected behavior for a bdmpza
based allenylidene complex in crystalline state. The arrangement with the allenylidene unit
positioned trans to a pyrazole unit forces the chlorido ligand trans to the carboxylate anchor.
The allenylidene chain exhibits common angles with ∠Ru–C(61)–C(62) = 172.8(3)° and
∠C(61)–C(62)–C(63) = 179.0(4)°. This indicates that in the solid state less repulsion occurs,
compared to complex 29B, thus, allowing the allenylidene moiety to maintain its preferred
geometry.
95
!Results and Discussion
!! !
Figure 39. Molecular structure of [Ru(bdmpza)Cl(═C═C═(PCO))(PPh3)] (29A). Thermal ellipsoids are drawn at the 50% probability level. Hydrogen atoms and two molecules dichloromethane have been omitted for clarity. Selected bond lengths (Å) and angles (deg): Ru–N(11) = 2.142(3), Ru–N(21) = 2.196(3), Ru–O(1) = 2.079(2), Ru–P(1) = 2.3325(10), Ru–Cl(1) = 2.3770(9), Ru–C(61) = 1.859(3), C(61)–C(62) = 1.244(5), C(62)–C(63) = 1.365(5); N(11)–Ru–N(21) = 83.03(10), O(1)–Ru–N(11) = 87.15(10), O(1)–Ru–N(21) = 83.38(10), O(1)–Ru–P(1) = 83.79(7), P(1)–Ru–Cl(1) = 99.43(3), P(1)–Ru–C(61) = 89.24(11), O(1)–Ru–C(61) = 95.41(13), N(11)–Ru–P(1) = 170.45(8), N(21)–Ru–Cl(1) = 88.72(7), Cl(1)–Ru–C(61) = 92.13(11), Ru–C(61)–C(62) = 172.8(3), C(61)–C(62)–C(63) = 179.0(4).
The packing motif in the solid state resembles for 29A more or less the packing motif of the
anthraquinone based complex (25A) although due to the extended ring system only partial
overlap of four phenyl rings can be observed (Figure 40). This can be attributed to the fifth
phenyl ring being slightly forced out of the planar system due to repulsive interactions with
the PPh3 ligand. For a possible application of pentacene or pentacenequinone based
compounds in, for example, field effect transistors, efficient overlap of the planar π-systems is
crucial in order to allow for charge transport along this axis.[274]
Cl1
O3C74
C63
N11
C62C61
Ru1
N21
N12N22
P1
C1
O1
C2O2
a
bc
96
!Results and Discussion
!! !
Figure 40. π–π stacking interactions between two molecules of 29A from a) top view and b) side view.
The arrangement observed for 29A and 29B in the crystalline state renders the complexes
promising candidates for metal-tuned FETs or “organic” metal-semiconductor field-effect
transistors (OMESFETs), whereas the electron-accepting ability and low-energy absorption
characteristics might be tuned for use in solar cells. Both aspects present an appealing starting
point for new kinds of functionalized organic semiconductors.
a) b)
97
!Results and Discussion
!! !
4.4.5 Vinylidene Complex Bearing a Malonodinitrile Substituted Pentacenequinone
A common way to modulate the electron-accepting properties of quinones is the formation of
the corresponding tetracyano-p-quinodimethane (In the case of unsubstituted quinone:
7,7,8,8-tetracyano-p-quinodimethane (TCNQ)) by reacting the quinone with malonodinitrile
in the presence of titanium(IV) chloride.[284-285] Especially TCAQ (11,11,12,12-tetracyano-
9,10-anthraquinodimethane) has played an important role in the area of organic electron
acceptors and its properties have been extensively reviewed.[286] The major drawback of most
TCNQ based polyaromatic systems is the stabilization of the resulting radical anion resulting
from the reduction of the electron acceptor. M. HANACK et al. synthesized a series of
symmetrical acene based TCNQ derivatives like TCPQ (15,15,16,16-tetracyano-6,13-
pentacenequinodimethane) that however, were poorer electron acceptors as indicated by the
more negative reduction potentials observed in CV analysis.[285] Unsymmetrical acenes
bearing acetylene moieties on one side and malonodinitrile units one the opposing side are
currently a project investigated by A. WATERLOO in the group of R. TYKWINSKI. Starting from
the previously used 13-ethynyl-13-hydroxypentacen-6-one (28), the reaction with
malonodinitrile in the presence of TiCl4 leads to the isolation of 2-(13-(dicyanomethyl)-13-
ethynylpentacen-6(13H)-ylidene)malononitrile (30) as the reaction with the hydroxyl moiety
could not be avoided. In cooperation, it was decided to investigate the reaction of this
acetylene derivative with the [Ru(bdmpza)Cl(PPh3)2] (14) precursor to explore the effects of
the electron withdrawing groups. Reacting [Ru(bdmpza)Cl(PPh3)2] (14) with 30 in THF led
within minutes to a dark blue solution very similar to 29A/29B, nevertheless the solution
obtained is highly sensitive towards oxygen as it turns from blue to brownish green within
hours (Scheme 47). From this lack of stability it was expected that the formed complex 31
might be the corresponding vinylidene complex (For the sake of simplicity the pentacene-
quinone derivative of the vinylidene complex 31 is referred to as PCN).
98
!Results and Discussion
!! !
Scheme 47. Synthesis of [Ru(bdmpza)Cl(═C═CH(PCN))(PPh3)] (PCN = pentacenone based tetracyano derivative) (31).
Due to the poor stability under aerobic conditions, the isolation of complex 31 via column
chromatography was performed under nitrogen atmosphere with CH2Cl2/acetone (1:1, v/v,
silica). It was concluded that compound 31 has the vinylidene moiety positioned trans to a
pyrazole unit as no second structural isomer could be observed. In the 1H NMR spectrum the
bdmpza ligand can be assigned by its characteristic four methyl signals at 2.51, 2.15, 1.94 and
1.91 ppm. One of the pyrazole protons in 4 position is shifted upfield to 4.98 ppm while the
second one shows a more common value with 5.93 ppm. Furthermore, the vinylidene proton
shows an extremely low value for an aromatic vinylidene complex of 3.82 ppm, which might
result from the strong electron withdrawing groups of the substituent.[61] The aliphatic proton
at 3.47 ppm and several aromatic protons between 8.68 and 7.65 ppm characterize the PCN
moiety. Moreover, only one set of signals of the PCN moiety can be observed indicating free
rotation around the Cβ–Cγ bond. The 13C NMR spectrum confirms the aforementioned
assumption that the vinylidene ligand is positioned trans to a pyrazole moiety as a long range
P–C coupling can be observed for one 4 position pyrazole carbon atom at 106.6 ppm
(4JC,P = 2.9 Hz).[186] Additionally, the presence of Cα gives rise to a doublet downfield shifted
to 365.1 ppm (2JC,P = 39.1 Hz) clearly providing evidence for a vinylidene complex.
Furthermore, a set of four signals of the cyano substituents at 119.1, 113.4, 110.6 and
110.1 ppm confirm the obtained complex 31. In addition the pentacenequinone backbone
NN N
N
Me
Me
Me
MeRu
OO
CClPh3P CH
CNNC
CNCN
31
H
NC CN
CNCN
NN N
N
Me
Me
Me
MeRu
OO
ClPh3P PPh3+ THF
30
99
!Results and Discussion
!! !
leads to several aromatic signals between 134.6 and 126.9 ppm. The 31P NMR spectrum
supports the assumption that only one structural isomer has been formed as one strongly
downfield shifted singlet at 44.6 ppm can be observed. The IR spectrum confirms the
presence of two weak nitrile vibrations at 2198 and 2126 cm–1 arising from the two different
sets of nitrile moieties. Finally ESI-MS experiments allowed the detection of the protonated
complex 31 (m/z 1077.21 (15%) MH+) and its sodium adduct (m/z 1099.20 (100%)
[M + Na]+).
The intense blue color of complex 31 seems to indicate an allenylidene complex, thus UV/Vis
absorption spectroscopy was performed to compare the spectrum to the closely related
ruthenium allenylidene complexes 29A and 29B based on pentacenequinone. The signals
below 450 nm resemble the typical pattern for ruthenium bdmpza based complexes. However,
a very broad absorption with maxima at 578 and 692 nm and high molar extinction
coefficients between 6000 and 8000 L mol–1 cm–1 can be observed (Figure 41), although these
extinction coefficients are high, in comparison to the allenylidene complexes reported
previously, the values are moderate. A possible explanation would be that the electron
withdrawing nitrile substituents lead to a strong bathochromic shift and the absorptions
observed are similar to the allenylidene complexes metal-perturbed π–π* transitions, but in
this case of the vinylidene unit. A further feature of 31 is an intense absorption in the NIR
region between 1000 and 1087 nm with an molar extinction coefficient around 800 L
mol–1 cm–1 that could not yet be attributed to a certain transition due to the lack of a single
crystal X-ray structure determination and thereafter TD-DFT calculations (Figure 42).
Nevertheless, a HOMO–LUMO transition seems logical in analogy to values reported in this
work for the allenylidene complexes.
100
!Results and Discussion
!! !
Figure 41. Absorption spectrum of 31 in CH2Cl2.
Figure 42. Magnification of the relevant parts of the NIR region displaying forbidden MLCTs for 31; signal caused by CH2Cl2 is indicated by *.
400 600 800 1000 1200 1400 16000
5000
10000
15000
20000
25000
30000
35000ε)[L)m
ol/1)cm
/1]
W ave leng th)[nm]
800 1000 1200 1400 16000
200
400
600
800
1000
ε'[L'm
ol-1'cm
-1]
W ave leng th'[nm]
*
101
!Results and Discussion
!! !
4.4.6 Benzotetraphenone Based Allenylidene Complexes
In the next step the focus was more on the packing motif in the solid state. For theoretical
device formation a 3D electronic communication in the solid state is required. This can be
achieved either by intramolecular transport within a polymer or intermolecular via close
distance interactions by π-π stacking interactions.[272, 287] Based on the previous results, it was
assumed that the allenylidene substituent is too close to the PPh3 and the bdmpza ligand to
show extended intermolecular stacking interactions. Hence it was decided to look into related
carbon-rich compounds that show substitution patterns extending the size of the polyaromatic
system in the opposing direction to the allenylidene chain. For comparable size, 7H-
benzo[no]tetraphen-7-one (10,11-BzBT, 34) consisting of five six-membered rings was
chosen. Due to the so called 1,7 interaction the molecule is greatly distorted from a planar
geometry, even in comparison to the closely related 7H-benzo[hi]chrysene-7-one (8,9-BzBT,
32) and 13H-dibenz[a,de]anthracen-13-one (5,6-BzBT, 33) (Figure 43).[288]
Figure 43. 7H-benzo[hi]chrysene-7-one (8,9-BzBT, 32), 13H-dibenz[a,de]anthracen-13-one (5,6-BzBT, 33) and structures of 7H-benzo[no]tetraphen-7-one (10,11-BzBT, 34) (steric repulsion indicated with *).[288]
Starting from compound 34 is most promising as two of the phenyl rings are facing away
from the keto moiety and in consequence from the future allenylidene unit. Thus it was
decided to investigate the follow up chemistry of 10,11-BzBT as up to now only the nitration
of the aromatic backbone, the dimerization of two units and the reduction of the keto moiety
is know to literature.[289-291]
O O O
32 33 34* *
* *
102
!Results and Discussion
!! !
Scheme 48. Synthesis of the TMS-ethynyl alcohol 35 and the deprotected propargyl alcohol 36 (reaction conditions: I) 1. n-BuLi, TMS-acetylene, THF, 2. H2O; II) KOH, MeOH, H2O).
Starting from 7H-benzo[no]tetraphen-7-one (34) (Scheme 48), the addition of excess amounts
of lithiated TMS-acetylene leads to the quantitative formation of the corresponding propargyl
alcohol indicated by the characteristic singlets of the alcohol proton at 2.59 ppm and the TMS
group at 0.23 ppm in the 1H NMR spectrum. In the 13C NMR spectrum of compound 35, the
two relevant alkyne signals appear at 107.6 and 93.2 ppm, while that of the tetrahedral carbon
is observed at 69.9 ppm. The product gives a signal in negative mode of ESI-MS analysis that
can be attributed to a chloride adduct of 35 (m/z 413.11 (18%) [M + Cl]–). Deprotection of the
alkyne is achieved in methanol with potassium hydroxide. 7-Ethynyl-7H-benzo[no]tetraphen-
7-ol (36) shows the additional signal of alkyne proton at 2.91 ppm in the 1H NMR spectrum,
while the singlet of the TMS group is lost. In the 13C NMR spectrum, the loss of the methyl
resonance from the TMS group and the shift of the terminal alkyne carbon to 76.4 ppm
support the successful transformation from 35 to 36.
O
34 35 36
OH OH
TMS H
I) II)
103
!Results and Discussion
!! !
Scheme 49. Synthesis of [Ru(bdmpza)Cl(═C═C═(BT))(PPh3)] (37A, 37B).
The preparation of the corresponding benzotetraphene (BT) based ruthenium allenylidene
complexes 37A/37B was carried out by combinig equimolar amounts of propargyl alcohol 36
and [Ru(bdmpza)Cl(PPh3)2] (14) (Scheme 49). The reaction mixture turns to deep blue,
similar to the reaction to give pentacenequinone based allenylidene complexes 29A/29B
indicating in both cases a strong influence of the size of the aromatic group on the color of the
allenylidene complex. The separation of the two structural isomers can be achieved by
column chromatography as described for the anthraquinone based allenylidene complexes
25A/25B. For the major isomer 37A, Cα shows a characteristic signal in the 13C NMR
spectrum at 273.6 ppm with a coupling constant of 2JC,P = 19.2 Hz. For Cβ (221.1 ppm, d, 3JC,P = 3.5 Hz) and Cγ (139.8 ppm, d, 4JC,P = 1.7 Hz) the signals are also shifted upfield in
comparison to the fluorenone, anthraquinone and pentacenequinone based systems. The
allenylidene stretch appears in the IR spectrum at 1903 cm–1. A singlet in the 31P NMR
spectrum of 37A is found at 35.2 ppm for the PPh3 ligand, supporting assignment of the
allenylidene trans to a pyrazole moiety. ESI-MS experiments confirm the formation of the
complex through detection of the molecular ion (m/z 934.18 (100%) [M]+) and similar to 24B
the decarboxylation followed by chloride dissociation can be observed (m/z 855.22 (23%)
[M – CO2 – Cl]+). For isomer 37B similar upfield shifted signals for Cα (289.5 ppm, d, 2JC,P = 18.4 Hz), Cβ (237.2 ppm) and Cγ (138.0) can be observed. The IR spectrum reveals an
allenylidene stretch in a similar region at 1907 cm–1. The PPh3 ligand shows a singlet in
NN N
N
Me
Me
Me
MeRu
OO
C ClPh3P
NN N
N
Me
Me
Me
MeRu
OO
CClPh3PC
CC
C
37A
37B
NN N
N
Me
Me
Me
MeRu
OO
ClPh3P PPh3
+
+ THF
36
14
OH
H
104
!Results and Discussion
!! !
31P NMR spectrum of 37B confirming the assignment that the allenylidene is positioned trans
to the carboxylate anchor. These values are overall in good agreement with the previously
observed NMR chemical shifts for type B isomers in comparison to type A isomers.[61]
ESI-MS experiments again show the appearance of a signal that is characteristic for the
ionized complex (m/z 934.18 (100%) [M]+). Layering a solution of 37A in CH2Cl2 with
n-hexane gave crystals of complex 37A suitable for a single crystal X-ray structure
determination. The compound crystallizes as racemic mixture in the space group P–1 with
two disordered molecules CH2Cl2 in the asymmetric unit. A graphical presentation of the
compound is illustrated in Figure 44. As mentioned previously for type A isomers the typical
strained coordination of the bdmpza ligand is observed and the allenylidene unit is
coordinated trans to a pyrazole group, which leaves the PPh3 ligand trans to the second
pyrazole and the chlorido ligand trans to the carboxylate anchor.
105
!Results and Discussion
!! !
Figure 44. Molecular structure of [Ru(bdmpza)Cl(═C═C═(BT))(PPh3)] (37A). Thermal ellipsoids are drawn at the 50% probability level. Hydrogen atoms and two molecules dichloromethane have been omitted for clarity. Selected bond lengths (Å) and angles (deg): Ru–N(11) = 2.159(5), Ru–N(21) = 2.199(5), Ru–O(1) = 2.091(4), Ru–P(1) = 2.2936(17), Ru–Cl(1) = 2.3803(16), Ru–C(61) = 1.878(6), C(61)–C(62) = 1.239(8), C(62)–C(63) = 1.374(8); N(11)–Ru–N(21) = 85.5(2), O(1)–Ru–N(11) = 86.85(19), O(1)–Ru–N(21) = 85.14(19), O(1)–Ru–P(1) = 86.09(13), P(1)–Ru–Cl(1) = 95.91(6), P(1)–Ru–C(61) = 89.08(17), O(1)–Ru–C(61) = 95.5(2), N(11)–Ru–P(1) = 172.60(14), N(21)–Ru–Cl(1) = 89.08(15), Cl(1)–Ru–C(61) = 90.02(17), Ru–C(61)–C(62) = 174.1(5), C(61)–C(62)–C(63) = 176.0(6).
The main feature of the structure of 37A is the nearly linear allenylidene moiety with ∠Ru–
C(61)–C(C62) = 174.1(5) and ∠C(61)–C(62)–C(63) = 176.0°(6). The benzotetraphene group
is non-planar, due to hydrogen-hydrogen repulsion that forces the phenalene and naphthalene
portions out of planarity. For rational description of the distortion between the two units the
twisting angle around the pseudo bond C(72)–C(74) is defined. The torsion angle ∠C(75)–
C(74)–C(72)–C(71) = 32.58° is similar to the angle observed for the parent ketone (33.4°).[288]
Also the distance C(71)–C(75) = 3.006 Å for 37A is close to the precursor which shows a
distance of 2.993 Å.[288] The combination of reduced planarity and of an extended π-system
facing away from the ruthenium center allows the formation of a stepwise arrangement within
the crystal lattice. While the pentacenequinone based systems did not show any extended
stacking motifs, several π–π stacking interactions can be observed for complex 37A (Figure
45). The mean distance between two neighboring phenalene units accounts to 3.39 Å, which
Cl1
N21N11
N22N12
RuC61
C62P1
C63
C71C72
O1
C73
C74
O2
C75
ab
c
106
!Results and Discussion
!! !
indicates attractive interactions. For comparison the interplanar distance in the parent 10,11-
BzBT (34) has been reported with 3.582 Å for the face to face stacked ketone.[288] For the
naphthalene unit three short contact interactions between 3.25 and 3.62 Å to the neighboring
phenalene unit can be observed. This slip-stacked arrangement leads to a staircase type
arrangement in the crystal that might allow charge transport between the polyaromatic
systems.
Figure 45. π–π stacking interactions between two molecules of 37A from a) side view and b) top view.
For the second structural isomer 37B only a single crystal X-ray structure determination of
poor quality could be obtained due to a strongly disordered crystal. Nevertheless, qualitative
interpretation of the structure reveals that due to steric hindrance of the remaining ligands no
π–π stacking interactions can be observed for this isomer.
The voltammograms of the benzotetraphene based allenylidene complexes
[Ru(bdmpza)Cl(═C═C═(BT))(PPh3)] (37A, 37B) show mainly similarities to the fluorene
based allenylidene complexes 20A and 20B, although the potentials again show dependence
on the structural isomer (Chapter 8.2). The reversible redox process involving the Ru(II)
center appears at 87 mV (37A) or 118 mV (37B) indicating a more cathodic process due
electron releasing properties of the benzotetraphene unit. The first redox process at negative
voltage can best be described as quasi-reversible for the reduction involving the allenylidene
moiety (–1228 mV (37A); –1168 mV (37B)) due to the lower peak current ratios. The second
process for the reduction of the benzotetraphene unit itself (–1914 mV (37A); –1859 mV
(37B)) is fully reversible.
a) b)
107
!Results and Discussion
!! !
The UV/Vis spectra of the benzotetraphene based systems 37A/B recorded in CH2Cl2 share
several common features for ruthenium allenylidene complexes as previously reported (Figure
46).[61, 125, 127, 256] The strong absorptions at wavelengths less than ~450 nm have been assigned
to ligand centered (LC) π–π* transitions involving the PPh3 and bdmpza ligands. An
additional metal perturbed π–π* transition can be observed at 654 nm for 37B with a molar
extinction coefficient around 15000 L mol–1 cm–1. In comparison to the complex of type B a
further increase in absorption energy can be observed for the complex bearing the
allenylidene unit positioned trans to the pyrazole moiety (708 nm for 37A) with a comparable
extinction coefficient. Again, absorption bands are observed in the NIR region (Figure 47),
thus E. HÜBNER performed TD-DFT calculations of the excited states on the basis of crystal
structures of 37A. The results of the calculations of the excited states revealed two absorption
bands at 1058 nm (1.17 eV) and 905 nm (1.37 eV) that can be assigned to MLCTs which are
in good agreement with the experimental values (1097 nm, 989 nm) (Table 10). The first
absorption correlates mainly to the HOMO → LUMO transition with the second one
assignable to the HOMO–1 → LUMO transition. The HOMO and HOMO–1 orbital may be
described as ruthenium d orbitals with a small contribution of the chlorido ligand and the
carboxylate anchor of the bdmpza ligand (Figure 48). For the HOMO orbital electron density
can be observed to extend towards the allenylidene chain. For the LUMO a high degree of
delocalization along the allenylidene unit into the phenalene moiety of the benzotetraphene
unit can be observed. Overall, this results in a rather low energy LUMO leading to a small
energy difference between the occupied and unoccupied orbitals and thus long-wave
absorptions can be observed. The two small transition dipole moments that are calculated
(1.54 and 0.51 debye) indicate forbidden transitions, which correlate well with the low
absorption coefficients that were observed experimentally (764 and 708 L mol–1 cm–1).
Furthermore, the HOMO–LUMO gap calculated as the difference between the calculated
orbital energies of the ground state (DFT) of 37A correlates less with the experimental CV
data (calcd.: 2.0 eV, observed 1.36 eV).
108
!Results and Discussion
!! !
Figure 46. Absorption spectrum of 37A (black) and 37B (grey) in CH2Cl2.
Figure 47. Magnification of the relevant parts of the NIR region displaying forbidden MLCTs for of 37A (black) and 37B (grey); signal caused by CH2Cl2 is indicated by *.
400 600 800 1000 1200 1400 16000
5000
10000
15000
20000
25000
30000ε)[L)m
ol/1)cm
/1]
W ave leng th)[nm]
800 1000 1200 1400 16000
200
400
600
800
*
ε([L(m
ol.1(cm
.1]
W ave leng th([nm]
109
!Results and Discussion
!! !
LUMO HOMO
HOMO–1 HOMO–2
Figure 48. Orbital diagrams of the LUMO (–2.9 eV), HOMO (–4.9 eV), HOMO–1 (–5.1 eV) and HOMO–2 (–5.3 eV) of [Ru(bdmpza)Cl(═C═C═(BT))(PPh3)] (37A).
Observed values Calculated values
Compound Wavelength [nm]
Absorption coefficient [M–1 cm–1]
Wavelength [nm]
Transition dipole moment [debye]
Transition
37A 1097
989
764
708
1058
905
1.54
0.51
HOMO ! LUMO
HOMO–1 ! LUMO
Table 10. Calculated and measured transitions for 37A in the NIR region.
110
!Results and Discussion
!! !
4.4.7 Larger Quinoidal Polyaromatic Compounds
Based on the previous results even larger graphene like systems were considered as suitable
precursors in which an extension of the substituent along the direction of the allenylidene
chain would lead to 3D stacking patterns. Derived from anthraquinone and pentacenequinone,
several dimerized and fused systems are viable (Figure 49), although no synthetic route for
peripentacenequinone has been described, so far.[292]
Figure 49. Structures of the acenequinones (bisanthracenquinone/bianthrone (38), bispentacenequinone (40)) and fused acenequinones (bisanthenequinone (39), „fused bispentacenequinone“ (41), peripentacenequinone
(42)).[292]
O
O
O
O
O
O
O
OO
O
38 39
40 41
42
111
!Results and Discussion
!! !
From the structures shown in Figure 49, bisanthenequinone is a promising candidate due to its
planarity and its considerable larger π system compared to the previously used systems. The
bisanthenequinone structure is the lead structure for many naturally occurring pigments like
hypericin controlling biophysical processes like phototoxicity and antiseptic actions.[293-294]
The first synthesis of 39 was described by R. SCHOLL in 1910 via oxidation of meso-
benzdianthrone with potassium dichromate in concentrated sulfuric acid.[295] Given the recent
interest in carbon-rich chemistry several groups have used bisanthenequinone based systems.
H. BOCK et al. used this as building block to create soluble and liquid-crystalline ovalenes,[296]
which are of great interest for devices like field effect transistors and solar cells.[297-302] Closely
related are dicyano ovalene diimides reported by J. WU et al. that allow solution processing of
OFET devices, which show high electron mobility under aerobic conditions.[303] Further
compounds by J. WU et al. focused on the reactivity of the meso positions and allowed the
formation of NIR dyes that undergo several reversible redox processes.[304] Similar results
were obtained for the parent bisanthene core that without meso-substituents rapidly
decomposes.[305]
Based on these promising results it was decided to focus on the synthesis of an allenylidene
complex bearing a bisanthenequinone unit. Starting from commercially available bianthrone
(38) required as first step a photocyclization reaction previously described by S. ARABEI and
coworker.[306] Illumination was performed with a mercury vapor lamp and a solution of 38 in
benzene, which reacted via intermediary helianthrone (43) in one step cleanly to
bisanthenequinone (39, Scheme 50). Experiments showed that the cyclization can either be
performed under an argon atmosphere or without inert gas, although the yields are higher in
the presence of oxygen. The presence of oxygen leads to an orange product in comparison to
a brownish product after inert gas conditions, although the spectroscopic datas are equivalent.
112
!Results and Discussion
!! !
Scheme 50. Photocyclization reaction for the synthesis of bisanthenequinone (39).[306]
The poor solubility of bisanthenequinone (39) leads to precipitation from benzene and after
removal of unconverted bianthrone the IR spectrum shows a shift for the two characteristic
quinone signals. For 38 these two signals appear at 1667 and 1594 cm–1 and are shifted for 39
to lower wavenumbers at 1659 and 1582 cm–1. For further characterization, it was decided to
measure a 1H NMR spectrum in deuterated concentrated sulfuric acid as no other solvent
allows dissolution of 39. The signals confirm the high symmetry of the bisanthenequinone
unit as only three signals can be observed. Two well resolved doublets at 8.63 and 8.22 ppm
and a further signal that appears as triplet at 7.26 ppm are in agreement with the reported
structure.
In the following step, the conversion of the quinone moiety into a propargyl alcohol by
nucleophilic addition of acetylene precursors was attempted. As promising candidates sodium
acetylide, ethynylmagnesium bromide and lithiated trimethylsilyl acetylide were added in
equimolar amounts to 39 to form the monoaddition product (Scheme 51) in analogy to the
anthraquinone and pentacenequinone systems.
O
O
O
O38 39
O
O43
hν hν
113
!Results and Discussion
!! !
Scheme 51. Attempted syntheses of a bisanthenone propargyl alcohol (44).
Nevertheless, all described attempts failed to produce a propargyl alcohol. Due to the poor
solubility of the resulting compounds, a positive identification of the obtained compounds is
questionable. In comparison to the literature known bisanthenequinone derivatives that bear
bulky or electron withdrawing groups on the reactive zigzag edges 44 lacks stabilization.[307]
This might cause decomposition of 44 after formation. Another explanation might be a report
by J. WU et al. indicating that lithiated species can be added to the bisanthenequinone
backbone and not directly to the quinone moiety allowing the formation of a mixture of
substances.[304] Based on these results, two routes seem promising. On the one hand, the
symmetrical functionalization, known in literature to obtain 7,14-bis(triisopropylsilylethynyl)-
phenanthro[1,10,9,8-opqra]perylene and after deprotection the formation of bimetallic
allenylidene complexes.[305] On the other hand functionalization of the armchair area of the
bisanthenequinone unit, which shows diene character, might be an opportunity.[303] Diels-
Alder reaction with 1,4-naphthaquinone can introduce quinoidal systems that in a next step
could be converted into a suitable precursor with strongly enhanced solubility in comparison
to the unsubstituted bisanthenequinone system (Scheme 52).[308]
O
O39
O44
OH
R
1.2. H2O
R R´
R = H, R´ = NaR = H, R´ = MgBrR = TMS, R´ = Li
114
!Results and Discussion
!! !
Scheme 52. Functionalization of the bay region of soluble bisanthene derivatives by J. WU et al.[308]
tButBu
tButBu
tButBu
tButBu
O
O
O
O
+ benzeneΔ
115
!Results and Discussion
!! !
4.4.8 Carbon-Rich Allenylidene Complexes Based on [RuCl2(PPh3)3]
It has been reported previously that [RuCl2(PPh3)3] is a versatile starting compound for quite
stable 16 VE ruthenium allenylidene complexes like [RuCl2(═C═C═CPh2)(PPh3)2].[93] This
complexes can be further stabilized as 18 VE complexes via coordination of solvents to form
systems such as [RuCl2(═C═C═CPh2)(PPh3)2(solvent)] (solvent = H2O, MeOH, EtOH).[96] A
controlled dimerization to less reactive 18 VE bimetallic allenylidene complexes [(Ph3P)4(μ-
Cl)3Ru2(═C═C═CAr2)2]PF6 has been reported.[309] Therefore, the corresponding fluorenone
based allenylidene complex [RuCl2(═C═C═(FN))(PPh3)2] (45) was targeted by applying the
conditions for the analogous diphenyl allenylidene complex. Heating [RuCl2(PPh3)3] with
9-ethynylfluoren-9-ol in THF for 2 h under reflux led to the formation of a deep red solution
(Scheme 53).
Scheme 53. Synthesis of [RuCl2(═C═C═(FN))(PPh3)2] (45).
The progress of the reaction could be monitored via IR spectroscopy by the disappearance of
the alkyne peak and appearance of the resulting allenylidene peak at 1922 cm–1. After
recrystallization from CH2Cl2/n-hexane, the compound was further characterized via its
characteristic 13C NMR spectrum showing the Cα at 313.8 and Cβ at 239.1 ppm. Both signals
are shifted downfield in comparison to the analogous diphenyl allenylidene complex. The
high symmetry of complex 45 leads to only one singlet for two triphenylphosphine ligands in
the 31P NMR spectrum at 29.1 ppm. ESI-MS experiments verified the proposed structure as a
monocationic species (m/z 849.12 (100%), [M – Cl]+) that fits complex 45 after loss of one
chlorido ligand. The compound is easily obtained and long storage under anaerobic conditions
is possible. Solutions of 45 in chlorinated solvents like CHCl3 and CH2Cl2 tend to form
dimeric complexes as indicated by the formation of four doublets in 31P NMR spectrum. It is
literature known that, if no strongly coordinating solvents like alcohols or pyridine are present
the stabilization of the 16 VE complex can occur via dimerization processes.[93] For the
Ru
Ph3P Cl
PPh3Cl
C C C[RuCl2(PPh3)3]OH
+ Δ
THF
45
H
116
!Results and Discussion
!! !
[RuCl2(PPh3)3] based complexes two different structures are possible with one cationic (45A)
and one neutral 18 VE complex (45B) (Scheme 54). In accordance to the previously reported
dimers the two doublets at 47.8 (d, 2JP,P = 35.6 Hz) and 47.1 ppm (d, 2JP,P = 37.3 Hz) can be
assigned to the neutral allenylidene complex that consists of two µ2 bridged chlorido ligands
and two terminal chlorido ligands with the allenylidene units positioned trans to each other.
In consequence the two remaining doublets at 37.3 (d, 2JP,P = 26.7 Hz) and 34.8 ppm (d, 2JP,P = 26.7 Hz) are caused by the cationic form that contains a Ru2Cl3-cluster as a central
feature with two ruthenium(II) centers bridged by three chlorido ligands resulting in a
monocationic complex. The positive charge is compensated by a chloride counter anion that
has been displaced from the monomeric ruthenium allenylidene complex.
4.5 Ruthenium Heteroscorpionate Cumulenylidene Complexes as
Molecular Slides
Carbon nanotubes (CNTs)[310-312] are a key material in the nanotechnological progress[313] due
to their physical properties (electrical, optical, mechanical, etc.) and their nanometer scale
size.[314] The potential applications are ranging from electronics,[315] sensing[316] and energy
conversion[317-321] to biological functions.[322] The first discovered CNTs were multiwalled
carbon nanotubes (MWCNTs) consisting of several layers of tubes.[323] However, the leading
nanotubular structures in terms of possible applications are singlewalled carbon nanotubes
(SWCNTs), which can be described as small graphene like sheets that have been rolled up to
cylinders.[314] However, strong intermolecular bundling in SWCNTs leads to difficulties in
exfoliation of single strands and dispersing them afterwards in solution, especially in aqueous
media. The common ways are the covalent functionalization, i.e. functionalization of the open
edges or sidewall and the noncovalent interactions of aromatic molecules or macromolecules
with the sidewall. The advantage of noncovalent functionalization is the preservation of the
pristine sp2 hybrid state and the inherent electron transport properties.[324]
A wide variety of metallophthalocyanines and their CNT donor-acceptor systems are known
to literature and have been extensively reviewed.[314] Only few metal complex based systems
focusing on classical complex chemistry are known.[325-326] A first example was reported from
S. WONG and coworker, who described the addition of Wilkinson´s catalyst to oxidized
CNTs.[327] On the one hand, this increased the solubility in a variety of organic solvents and
exfoliation of larger nanotube bundles. On the other hand, the CNT worked as reusable
catalyst support for homogenous hydrogenation of cyclohexene, demonstrating the conserved
activity of the catalyst.[327] To name an example for the phthalocyanine based systems a
pyrene conjugate by D. GULDI et al. can be named.[328] The pyrene anchor allows the
noncovalent functionalization of CNTs and spectroscopic and photoelectrochemical
techniques were used to characterize the resulting adducts. Integration into photoactive
electrodes allowed the detection of stable and reproducible photocurrents.[328] A further
interesting approach is the reversible solubilization performed by A. IKEDA et al. based on a
[Cu(bpy-R2)2]2+ derivative bearing two cholesteryl groups.[329] The square planar copper(II)
complex allows π-π stacking interactions with the CNTs and leads to solutions of the
aggregated compound. Upon reduction with ascorbic acid, the coordination geometry of the
120
!Results and Discussion
!! !
copper(I) center changes to tetrahedral and leads to precipitation of the CNTs. Reoxidation
with oxygen leads to the former geometry and again a stable solution is formed. This behavior
was tested in several cycles and allows the purification of metallic and semiconducting CNTs
via modulation of the redox state of the copper center.[329]
Recently, A. MARTI et al. reported a series of cationic ruthenium(II) bpy complexes that are
able to solubilize CNTs in aqueous media via noncovalent interactions.[330] The complexes
[Ru(bpy)2L1]2+, [Ru(bpy)2L2]2+ and [Ru(bpy)2L3Ru(bpy)2]4+ allow π-π stacking interactions
with the CNTs and allow high individualization in water (Figure 50).
Figure 50. Ligands L1 (dppz = dipyrido[3,2-a:2´.3´-c]-phenazine), L2 (dppn = benzo[i]dipyrido-[3,2-a:2´.3´-c]phenazine) and L3 (tpphz = tetrapyrido[3,2-a:2´.3´-c:3´´,2´-h:2´´´,3´´´-j]phenazine) used by A. MARTI et al.[330]
4.5.1 Polyaromatic Ruthenium Vinylidene Complexes
Given the interest of the BURZLAFF group in carbon-rich cumulenylidene ruthenium
complexes it was decided to synthesize a series of complexes that could be suitable for
noncovalent functionalization of carbon nanotubes. Due to the simplicity of synthesis the first
focus was on ruthenium vinylidene complexes bearing the bdmpza ligand. From previous
work it was known that for acetylene compounds with small substituents a mixture of A and
B type isomers is formed (see Scheme 28). Due to the sensibility of vinylidene complexes
towards oxygen the separation was avoided, as column chromatography was not favorable.
Thus two important questions are, if larger polyaromatic substituents stabilize the vinylidene
complexes and if the steric demand leads to selective formation of A type isomers.
With the phenylacetylene based vinylidene complex already known to literature the next
larger 2-ethynyl-6-methoxynaphthalene was picked as suitable precursor. Reaction of excess
amounts of the ethynyl substituted naphthalene with [Ru(bdmpza)Cl(PPh3)2] (14) in THF at
N
N N
N N
N N
N N
N N
N
N
N
L1 L2 L3
121
!Results and Discussion
!! !
room temperature afforded an orange solution. Reducing of the solvent and cooling in the
freezer led to crystallization of the complex [Ru(bdmpza)Cl(═C═CH(6-methoxy-
naphthalene))(PPh3)] (48) (Scheme 58).
Scheme 58. Synthesis of the vinylidene complex [Ru(bdmpza)Cl(═C═CH(6-methoxynaphthalene))(PPh3)] (48).
The 1H and 13C NMR spectrum exhibit only one set of signals characteristic for an A type
isomer indicating the influence of the steric demand of the substituent on the coordination
pattern. The 1H NMR spectrum of 48 shows the characteristic Hß, gained by the 1,2-H shift, at
5.11 ppm with a 4JH,P coupling constant of 4.7 Hz due to the PPh3 ligand. Furthermore, the
four methyl groups of the bdmpza ligands appear at 1.87, 2.39, 2.46 and 2.53 ppm and the
methoxy group at 3.88 ppm. The complete absence of further signals in the aliphatic region
proves the clean formation of one isomer. In the 13C NMR spectrum, the Cα signal is found at
363.2 ppm as a doublet with the coupling constant 2JC,P = 24.8 Hz and the Cß signal appears at
109.0 ppm as a doublet with the coupling constant 3JC,P = 3.0 Hz. The singlet at 37.3 ppm in
the 31P NMR spectrum confirms the assignment to an A type isomer in comparison to the
literature value of 37.5 ppm for the A type isomer of the complex [Ru(bdmpza)Cl-
(═C═CHPh)(PPh3)] (B type: 32.3 ppm).[61] The absence of vibrations between 2200 and
1800 cm–1 in the IR spectrum that would indicate alkyne moieties and the detection of the
potassium adduct of complex 48 (m/z 867.12 (100%) [M + K]+) in ESI-MS experiments
confirm the structure.
Based on this promising result, it was decided to synthesize a pyrene (Pyr) based vinylidene
complex for possible applications as noncovalent linker to CNTs. Pyrene derivatives linked
via a variety of spacer groups are widely spread in literature for CNT exfoliation and further
applications.[331-342] In organometallic chemistry one example of pyrene substituted complexes
NN N
N
Me
Me
Me
MeRu
OO
CClPh3P CH
NN N
N
Me
Me
Me
MeRu
OO
ClPh3P PPh3
H+
48
THF OMe
OMe
122
!Results and Discussion
!! !
is reported by R. WINTER et al., demonstrating the non-innocent behavior of a ruthenium
vinylpyrenyl complex [(Pyr)CH═CH)Ru(CO)Cl(PPh3)3] allowing the large π ligand to
heavily participate in electron-transfer processes.[343] The first example of a pyrene based
vinylidene complex [Ru(κ1-OAc)(κ2-OAc)(═C═CH(Pyr))(PPh3)2] has been reported by J.
LYNAM et al. in 2012 in a study that emphasized the similar donor/acceptor properties of
vinylidene and isocyanide ligands.[344]
The synthesis of the corresponding bdmpza based ruthenium complex was achieved similar to
48. [Ru(bdmpza)Cl(PPh3)2] (14) and 2.2 equivalents of 1-ethynylpyrene were stirred in THF
at room temperature and after reducing the solvent and storing in the freezer the crystalline
complex [Ru(bdmpza)Cl(═C═CH(Pyr))(PPh3)] (49) was obtained (Scheme 59).
Scheme 59. Synthesis of the vinylidene complex [Ru(bdmpza)Cl(═C═CH(Pyr))(PPh3)] (49).
Again only one set of signals characteristic for an A type isomer could be observed. The 1H NMR spectrum indicates that the larger pyrene substituent leads to a deshielding of Hß and
thus to an increased chemical shift with 5.78 ppm and a coupling constant of 4JH,P = 4.7 Hz.
The larger polyaromatic substituent shows less influence on the bdmpza ligand as the four
methyl substituents of 49 appear with 1.88, 2.39, 2.49 and 2.55 ppm in a similar region
compared to 48. In the 13C NMR spectrum, Cα and Cß give rise to doublets at 359.1
(2JC,P = 23.1 Hz) at 111.8 ppm (3JC,P = 2.6 Hz). The deshielding influence apparently only
influences Cß significantly as the Cα position appears further upfield shifted. In addition all 16
carbon atoms of the pyrene unit could be detected and the C–H multiplicities were assigned
NN N
N
Me
Me
Me
MeRu
OO
CClPh3P CH
NN N
N
Me
Me
Me
MeRu
OO
ClPh3P PPh3
H+
49
THF
123
!Results and Discussion
!! !
via APT 13C NMR experiments. Further prove for the molecular structure is provided by the
singlet at 37.1 ppm in the 31P NMR spectrum and the detection of the molecular ion in ESI-
MS experiments (m/z 872.16 (100%) M+).
Finally, the geometry of complex 49 could be unambiguously characterized by a single
crystal X-ray structure determination. Crystals suitable for analysis were obtained from a
concentrated solution in CH2Cl2 layered with n-pentane stored in a Schlenk flask under argon
atmosphere with a septum allowing slow evaporation. The complex crystallizes as racemic
mixture in space group P–1 with one molecule CH2Cl2 and a strongly disordered molecule n-
pentane in the asymmetric unit (Figure 51).
Figure 51. Molecular structure of [Ru(bdmpza)Cl(═C═CH(Pyr))(PPh3)] (49). Thermal ellipsoids are drawn at the 50% probability level. Hydrogen atoms (except the vinylidene proton) and solvent molecules have been omitted for clarity. Selected bond lengths (Å) and angles (deg): Ru–N(11) = 2.162(5), Ru–N(21) = 2.216(10), Ru–O(1) = 2.091(4), Ru–P = 2.322(4), Ru–Cl = 2.388(4), Ru–C(71) = 1.781(9), C(71)–C(72) = 1.315(10), C(72)–C(73) = 1.482(8); N(11)–Ru–N(21) = 81.3(2), O(1)–Ru–N(11) = 86.8(2), O(1)–Ru–N(21) = 83.7(2), O(1)–Ru–P = 87.05(17), P–Ru–Cl = 96.16(14), P–Ru–C(71) = 87.5(2), O(1)–Ru–C(71) = 97.5(3), N(11)–Ru–P = 173.61(13), N(21)–Ru–Cl = 86.2(2), Cl–Ru–C(71) = 92.3(2), Ru–C(71)–C(72) = 173.7(5), C(71)–C(72)–C(73) = 123.8(6).
Complex 49 shows the typical strained octahedral coordination of the bdmpza ligand with the
vinylidene ligand positioned trans to one pyrazole moiety. The angles and distances for the
O2O1
H72P
N22N12
C72
C71
Ru
N11N21
C73
Cl
a
b
c
124
!Results and Discussion
!! !
bdmpza ligand are in good agreement with the previously reported bdmpza based vinylidene
complex [Ru(bdmpza)Cl(═C═CH(Tol))(PPh3)] (Tol = tolyl).[61] The vinylidene moiety
exhibits an angle ∠Ru–C(71)–C(72) = 173.7(5)° with bond lengths dRu–C(71) = 1.781(9) Å
and dC(71)–C(72) = 1.315(10) Å that are similar to the tolyl complex and the pyrene based
Table 11. Overview of characteristic bond lengths and angles of ruthenium vinylidene complexes [Ru(bdmpza)Cl(═C═CH(Pyr))(PPh3)] (49), [Ru(bdmpza)Cl(═C═CH(Tol))(PPh3)] and [Ru(κ1-OAc)(κ2-OAc)-(═C═CH(Pyr))(PPh3)2].[61, 344]
The torsion angle ∠C(61)–C(62)–C(63)–C(64) is a good indicator for possible conjugation
between the metal center and the pyrenyl unit. Complex 49 shows an angle of approximately
36°. This is quite large in comparison to the tolyl based complex [Ru(bdmpza)Cl-
(═C═CH(Tol))(PPh3)], which has a torsion angle of –18°. The rotation around the C(62)–
C(63) axis should not be hindered and in solution a planar arrangement might be achievable.
125
!Results and Discussion
!! !
Figure 52. Absorption spectrum of 49 in CH2Cl2.
Figure 53. Magnification of the relevant parts of the NIR region displaying forbidden MLCTs for 49; signal caused by CH2Cl2 is indicated by *.
400 600 800 1000 1200 1400 1600
0
5000
10000
15000
20000
25000
30000ε)[L)m
ol/1)cm
/1]
W ave leng th)[nm]
800 1000 1200 1400 16000
50
100
150
200
250
ε([L(m
ol.1(cm
.1]
W ave leng th([nm]
126
!Results and Discussion
!! !
The absorption spectrum of complex 49 measured in CH2Cl2 shows the expected pattern for a
ruthenium vinylidene complex based on bdmpza (Figure 52). Two intense absorptions at 406
and 389 nm with high molar extinction coefficients (~21000 L mol–1 cm–1) can be attributed to
a metal-perturbed π–π* transition of the vinylidene moiety and ligand centered π–π*
transitions of the bdmpza and PPh3 ligand. In addition, a further weak absorption can be
observed in the NIR region at 921 nm (~200 L mol–1 cm–1) as reported for the vinylidene
Complexes 48 and 49 are remarkably stable in the presence of oxygen. This is noteworthy
since for possible applications like exfoliation and non-covalent functionalization of CNTs,
higher stability is required to simplify procedures.
127
!Results and Discussion
!! !
4.5.2 Pyrene Based Allenylidene Complexes
As described in the previous chapters, ruthenium allenylidene complexes have proven to be
good candidates for remarkably stable organometallic compounds. A common disadvantage is
however, that the Cγ position needs to be stabilized with aryl substituents as protons or alkyl
substituents might lead to decomposition or rearrangement into vinylvinylidene
complexes.[157] Hence, pyrenophenone (50) was picked as conveniently available compound,
as it can be easily obtained by a Friedel-Crafts Acylation of pyrene with benzoyl chloride
(Scheme 60).[345-346] This compound has recently been employed as starting material for
several ethenes showing aggregation-enhanced excimer emission and electroluminescence.[346]
Furthermore, it is reported that 50 can undergo a Scholl Reaction creating an additional five-
or six-membered ring (Scheme 60, compound 52), depending on the publication.[291, 345, 347]
Therefore, it was decided to a) synthesize from pyrenophenone (50) the corresponding
propargyl alcohol 51 and convert the alcohol into the bdmpza based allenylidene complex and
b) explore the Scholl reaction and if possible synthesize a second propargyl alcohol 53 with
extended π system.
128
!Results and Discussion
!! !
Scheme 60. Synthetic overview of propargyl alcohols 53 and 51 that should be available starting from pyrenophenone (50).[291, 345, 347]
Pyrenophenone (50) was synthesized according to literature via a Friedel-Crafts Acylation of
pyrene.[346] Since the synthesis of the compounds 52 and 53 was not as straight forward as
expected these compounds will be discussed in detail in chapter 4.6.
Crystals of pyrenophenone (50) suitable for a single crystal X-ray structure determination
could be obtained from slow evaporation of a solution of 50 in a mixture of CH2Cl2 and n-
hexane. The ketone crystallizes in space group P–1 and confirms the previously reported
connectivity of the pyrenyl residue (Figure 54). The keto moiety shows a bond length dC(1)–
O(1) = 1.2227(15) Å and the three surrounding angles are close to 120°, which is in good
agreement with free rotation around the single bonds.
O
OH
H
OO
or
or
OH OH
H
H
50
51
52
53
129
!Results and Discussion
!! !
Figure 54. Molecular structure of pyrenophenone (50). Thermal ellipsoids are drawn at the 50% probability level. Hydrogen atoms have been omitted for clarity. Selected bond lengths (Å) and angles (deg): C(1)–O(1) = 1.2227(15), C(1)–C(2) = 1.4989(17), C(1)–C(21) = 1.4964(18), C(2)–C(3) = 1.3960(18), C(3)–C(4) = 1.3836(19), C(21)–C(26) = 1.3943(19), C(25)–C(26) = 1.393(2); O(1)–C(1)–C(2) = 119.81(11), O(1)–C(1)–C(21) = 120.15(11), C(2)–C(1)–C(21) = 120.04(11).
a)
b)
Figure 55. π–π stacking interactions between two molecules of 50 from a) top view and b) side view.
In the solid state, strong π–π stacking interactions between two neighboring pyrene units can
be observed (Figure 55). The pyrene units overlap with approximately 75% of their surface
O1C22
C1
C23C21
C2
C3
C15
C4
C14
C5
C16
C24
C13
C6
C26
C17
C12
C7
C8C25
C11
C9
C10
130
!Results and Discussion
!! !
area and an average distance of 3.46 Å can be observed. The plane of the phenyl moiety is
66.2° rotated against the plane formed by the pyrene and blocks further stacking interactions.
Reacting pyrenophenone (50) with ethynylmagnesium bromide in THF leads to the formation
of the brownish propargyl alcohol 1-phenyl-1-(pyren-1-yl)prop-2-yn-1-ol (51), which can be
isolated after aqueous workup (Scheme 61).
Scheme 61. Synthesis of propargyl alcohol 51 starting from pyrenophenone (50).
The addition of the acetylene unit leads to the characteristic acetylene proton at 2.98 ppm and
the alcohol proton at 3.86 ppm in the 1H NMR spectrum. In the 13C NMR spectrum the
corresponding signals for the acetylene unit appear at 86.7 and 76.8 ppm and the sp3 carbon
atom results in a signal at 74.7 ppm. Further proof for the structure are ESI-MS experiments
that show the sodium adduct (m/z 355.11 (30%) [M + Na]+) and the IR spectrum that shows a
characteristic alkyne absorption at 2114 cm–1.
In the next step the corresponding pyrenophenone based allenylidene complex was
synthesized by addition of propargyl alcohol 51 to the complex [Ru(bdmpza)Cl(PPh3)2] (14)
in THF and stirring for 4 d at room temperature until a deep red color could be observed
(Scheme 62). The separation of the resulting allenylidene complexes
[Ru(bdmpza)Cl(═C═C═C(PyrPh))(PPh3)] (54A/54B) was achieved via column
chromatography with CH2Cl2/acetone = 1:1 yielding a purple isomer 54A (allenylidene trans
to pyrazole) and a red isomer 54B (allenylidene trans to carboxylate). Complex 54A shows
the Cα carbon atom as doublet at 304.6 ppm (2JC,P = 26.3 Hz) in the 13C NMR spectrum and
the Cβ gives rise to a singlet at 231.2 ppm in the expected region. The PPh3 ligand leads as
expected to a singlet in the 31P NMR spectrum at 32.2 ppm. Further proof for the allenylidene
moiety is the appearance of a characteristic absorption in the IR spectrum at 1918 cm–1 and
O
OH
H
5051
H MgBr+
1. THF2. H2O
131
!Results and Discussion
!! !
the observation of a weak signal during ESI-MS experiments (m/z 960.19 (0.2%) [M]+). The
second structural isomer 54B shows downfield shifted values for Cα and Cβ with a doublet at
315.6 (2JC,P = 19.8 Hz) and a singlet at 241.9 ppm in the 13C NMR spectrum. Further evidence
for the allenylidene complex is the singlet in the 31P NMR spectrum at 32.2 ppm. However,
this signal has the identical value as structural isomer 54A and is usually the easiest indicator
for the geometry and allows in this case no statement. In the IR spectrum the typical
absorption appears at 1916 cm–1 and ESI-MS experiments allowed the observation of the
molecular ion (m/z 960.19 (14%) [M]+).
Scheme 62. Synthesis of [Ru(bdmpza)Cl(═C═C═C(PyrPh))(PPh3)] (54A, 54B).
Layering a solution of 54B in CH2Cl2 with n-hexane gave crystals suitable for a single crystal
X-ray structure determination. The compound crystallizes as racemic mixture in the space
group P–1. A molecular presentation of the compound is illustrated in Figure 56. As
mentioned previously for type B isomers the typical strained coordination of the bdmpza can
be observed and the allenylidene unit is coordinated trans to a pyrazole leaving the PPh3
ligand trans to the second pyrazole and the chlorido ligand trans to the carboxylate anchor.
Furthermore, this structure determination proves that the assignment to the A and B type
isomers is correct. The characteristic allenylidene angles are with ∠Ru–C(61)–
C(62) = 172.1(2) and ∠C(61)–C(62)–C(63) = 165.7°(3) strongly bent and show values similar
to the pentacenequinone based allenylidene complex 29B. In the case of complex 54B the π–
NN N
N
Me
Me
Me
MeRu
OO
C ClPh3P
NN N
N
Me
Me
Me
MeRu
OO
CClPh3PC
CC
C
54A
54B
NN N
N
Me
Me
Me
MeRu
OO
ClPh3P PPh3
+14THF
51
OH
H
132
!Results and Discussion
!! !
π stacking interactions between the neighboring pyrenyl moieties is reduced as the pyrenyl
units are showing less than half a pyrene overlap. However, π–π stacking interactions
between the pyrenyl moiety and one pyrazolyl unit can be observed, which seems to be
responsible for a larger distortion of the bdmpza ligand, the reduced linearity of the
allenylidene unit and the reduced planarity of the pyrenyl moiety.
Figure 56. Molecular structure of [Ru(bdmpza)Cl(═C═C═C(PyrPh))(PPh3)] (54B). Thermal ellipsoids are drawn at the 50% probability level. Hydrogen atoms have been omitted for clarity. Selected bond lengths (Å) and angles (deg): Ru–N(11) = 2.1458(19), Ru–N(21) = 2.086(2), Ru–O(1) = 2.1477(16), Ru–P(1) = 2.3521(6), Ru–Cl(1) = 2.3999(7), Ru–C(61) = 1.850(2), C(61)–C(62) = 1.250(4), C(62)–C(63) = 1.366(4); N(11)–Ru–N(21) = 84.67(8), O(1)–Ru–N(11) = 84.85(7), O(1)–Ru–N(21) = 87.18(7), O(1)–Ru–P(1) = 86.81(5), P(1)–Ru–Cl(1) = 87.93(2), P(1)–Ru–C(61) = 95.68(7), O(1)–Ru–C(61) = 176.51(9), N(11)–Ru–P(1) = 171.03(6), N(21)–Ru–Cl(1) = 173.29(5), Cl(1)–Ru–C(61) = 91.96(8), Ru–C(61)–C(62) = 172.1(2), C(61)–C(62)–C(63) = 165.7(3).
The reduced planarity is apparent for C(72), the carbon atom that connects the pyrenyl unit to
Cγ, as it deviates from the plane calculated for all pyrenyl carbon atoms by 0.16 Å. The mean
distance between the pyrazolyl moiety and the plane calculated for the pyrenyl moiety is
3.48 Å, indicating the strong interactions possible for the pyrene unit. The arrangement
Cl1
O2
N11
N12
O1
Ru
C61P1
C62
N22
N21
C63
C72
133
!Results and Discussion
!! !
suggests that the pyrenyl moiety should allow further π–π stacking interactions with carbon
allotropes in solution. However, the solubility is limited to polar solvents like acetone,
CH2Cl2, CHCl3 and mixtures containing aforementioned solvents and nonpolar solvents like
n-hexane and n-pentane as apparently the pyrenyl unit reduces the polarity of the complex in
comparison to the quinone based allenylidene complexes.
Cyclic voltammetric analyses were again performed on the pyrenophenone based allenylidene
complexes 54A and 54B (Chapter 8.2). Both compounds show a behavior similar to the
benzotetraphene based allenylidene complexes [Ru(bdmpza)Cl(═C═C═C(BT))(PPh3)] (37A,
37B). The oxidation process involving the ruthenium(II) center appears at 252 mV (54A) or
320 mV (54B). For 54A this process is reversible, however, for 54B this process is
irreversible and followed by a second oxidation process at higher potential that shows no
backward peak. The reduction processes are all best described as irreversible as the forward
scan shows signals similar to the previously reported allenylidene complexes within this
work. The backward scan however, shows extremely weak current intensities with larger peak
separations.
The UV/Vis spectra of the pyrenophenone based complexes 54A and 54B recorded in CH2Cl2
share several common features with the other ruthenium allenylidene complexes bearing the
bdmpza ligand (Figure 57). The strong absorptions at wavelengths less than ~300 nm have
been assigned to ligand-centered (LC) π–π* transitions involving the PPh3 and bdmpza
ligands. However, the intense absorption around 330 nm (54A: 336 nm, 29000 L mol–1 cm–1;
54B: 335 nm, 27000 L mol–1 cm–1) seems to correspond to the parent pyrenophenone moiety
50. An additional metal-perturbed π–π* transition can be observed at 543 nm for 54A with a
molar extinction coefficient around 24000 L mol–1 cm–1. This extinction coefficient is
noticeably larger in comparison to the allenylidene complexes discussed so far, although the
absorption maximum is close to the diphenyl based allenylidene complex
[Ru(bdmpza)Cl(═C═C═CPh2)(PPh3)] (519 nm, 17000 L mol–1 cm–1).[61] In comparison to the
complex of type A, complex 54B shows with an absorption at 523 nm a decrease in
absorption energy and the extinction coefficient (19000 L mol–1 cm–1) is in the common range
if compared to 37A. Again absorption bands are observed in the NIR region around 1050 nm
(45A) and 900 nm (45B) that can be attributed to HOMO–LUMO transitions and can best be
described as MLCTs (Figure 58).
134
!Results and Discussion
!! !
Figure 57. Absorption spectrum of 54A (black) and 54B (grey) in CH2Cl2; signal caused by switching lamp is indicated by *.
Figure 58. Magnification of the relevant parts of the NIR region displaying forbidden MLCTs for of 54A (black) and 54B (grey); signal caused by CH2Cl2 is indicated by *.
300 400 500 600 700
0
10000
20000
30000
40000)ε)[L
)mol
/1)cm
/1]
W ave leng th)[nm]
*
800 900 1000 1100 1200 13000
20
40
60
80
100
120
140
160
180
200
*
**ε*[L
*mol
01*cm
01]
W ave leng th*[nm]
*
135
!Results and Discussion
!! !
4.5.3 Carbon-Rich Ruthenium Allenylidene Complexes Bearing the PTA Ligand
For possible applications of the carbon-rich allenylidene complexes, e.g. in the exfoliation of
carbon allotropes, solubility in aqueous media or at least in alcohols is required. Hence, it was
decided to adapt the ligand sphere of the ruthenium precursor to allow the formation of water-
soluble allenylidene complexes.
Two types of water-soluble phosphines are commonly used in organometallic chemistry. On
the one hand, sulfonated derivatives of the classical triphenylphosphine ligand like the
triphenylphosphine-3,3′,3′′-trisulfonic acid trisodium salt (TPPTS) and on the other hand the
adamantane derived 1,3,5-triaza-7-phosphaadamantane (PTA). For ruthenium allenylidene
complexes, especially the PTA ligand has proven successful, as the cyclopentadienyl and
hydridotrispyrazolyl based complexes [Ru(Cp)(═C═C═CPh2)(PTA)(PPh3)](CF3SO3)[348] and
[Ru(Tp)(═C═C═CPh2)(PTA)(PPh3)]PF6[261] have been reported. Precursors for these
complexes are the chlorido complexes [Ru(Cp)Cl(PTA)(PPh3)][349] and [Ru(Tp)Cl-
(PTA)(PPh3)].[350] Reaction of the latter with propargyl alcohols led to the formation of
cationic allenylidene complexes due to chloride abstraction. As for the bdmpza based
ruthenium triphenylphosphine complex [Ru(bdmpza)Cl(PPh3)2] (14) usually one phosphine
ligand can be replaced easily, it was decided to synthesize the PTA analogues
[Ru(bdmpza)Cl(PTA)(PPh3)] (55) and [Ru(bdmpza)Cl(PTA)2] (56), each (Scheme 63).
136
!Results and Discussion
!! !
Scheme 63. Synthesis of [Ru(bdmpza)Cl(PTA)(PPh3)] (55) and [Ru(bdmpza)Cl(PTA)2] (56).
The reaction of [Ru(bdmpza)Cl(PPh3)2] (14) with one equivalent PTA in THF under reflux
allows the almost quantitative exchange of one PPh3 ligand. Complex 55 shows in the IR
spectrum the characteristic carboxylate absorption at 1661 cm–1 in CH2Cl2. The 1H NMR
spectrum shows two multiplets around 7.66 and 7.26 ppm that can be assigned to the PPh3
ligand and the PTA ligand leads to four quartets showing an AB pattern in the aliphatic region
at 4.37, 4.23, 3.89 and 3.78 ppm. The bdmpza ligand exhibits two asymmetric protons in the
4- and 4´ position indicating the expected asymmetric structure. The 13C NMR spectrum
repeats the conclusions drawn from the proton NMR spectrum as the characteristic doublets
from the PTA ligand at 73.1 (3JC,P = 5.8 Hz) and 52.4 ppm (1JC,P = 14.8 Hz) can be observed as
well as four asymmetric methyl substituents at 16.7, 14.0, 11.5 and 11.5 ppm. Further
evidence is the coupling pattern in the 31P NMR spectrum showing one doublet of the PPh3
ligand at 41.2 ppm (2JP,P = 43.5 Hz) and one doublet of the PTA ligand at –27.5 ppm
(2JP,P = 43.5 Hz) caused by the two different phosphine ligands. Additional ESI-MS
experiments show the presence of the molecular ion as major observable compound (m/z
803.17 (100%) M+). Complex 55 is nicely soluble in polar solvents like chlorinated solvents,
THF and alcohols, but unfortunately insoluble in water.
NN N
N
Me
Me
Me
MeRu
OO
ClPh3P PPh3
NN N
N
Me
Me
Me
MeRu
OO
Cl
NN
N
P
NN
N
P
NN N
N
Me
Me
Me
MeRu
OO
ClPh3P
NN
N
P+ 1.0 eq. PTA
+ 2.2 eq. PTA
- PPh3
- 2 PPh3
55
56
137
!Results and Discussion
!! !
Figure 59. Molecular structure of [Ru(bdmpza)Cl(PTA)(PPh3)] (55). Thermal ellipsoids are drawn at the 50% probability level. Hydrogen atoms have been omitted for clarity. Selected bond lengths (Å) and angles (deg): Ru–N(11) = 2.184(3), Ru–N(21) = 2.184(3), Ru–O(1) = 2.112(2), Ru–P(1) = 2.2621(10), Ru–P(2) = 2.3027(10), Ru–Cl(1) = 2.4048(9); N(11)–Ru–N(21) = 79.22(11), O(1)–Ru–N(11) = 87.42(10), O(1)–Ru–N(21) = 86.97(10), O(1)–Ru–P(1) = 86.90(7), P(1)–Ru–Cl(1) = 93.08(4), P(1)–Ru–P(2) = 91.86(4), N(11)–Ru–P(1) = 170.19(8), N(21)–Ru–Cl(1) = 88.84(8).
Crystals of 55 suitable for single crystal X-ray structure determination could be obtained from
a concentrated solution in CH2Cl2 layered with n-hexane (Figure 59). The complex
crystallizes as enantiomeric mixture in space group Pbca with one co-crystalized CH2Cl2 per
asymmetric unit. As expected the PPh3 and PTA ligand are similar to the parent ruthenium
complex [Ru(bdmpza)Cl(PPh3)2] (14) positioned trans to the pyrazolyl moieties, thus the
chlorido ligand is positioned trans to the carboxylate anchor. The smaller Tolman cone angle
and the resulting reduced steric demand of the PTA ligand in comparison to the PPh3 ligand
leads to a Ru–P(1) bond length of 2. 2621(10) Å and a longer Ru–P(2) bond length of
2.3027(10) Å. The reduced repulsion between the two phosphine ligands is also responsible
for a smaller angle ∠O(1)–Ru–P(1) = 86.90(7)° in comparison to the larger PPh3 ligand with
the angle ∠O(1)–Ru–P(2) = 93.20(7)°. The angle between both phosphine ligands is in
consequence reduced from 94.12(6)° for [Ru(bdmpza)Cl(PPh3)2] (14) to 91.86(4)° for
Cl1
N11 N21
N22N12
Ru
P2
P1
O1
O2
a
b
c
138
!Results and Discussion
!! !
complex 55.[61] However, the rigid structure of the bdmpza ligand leads to an almost
symmetrical coordination of the bdmpza ligand with identical ruthenium–nitrogen bond
lengths of 2.184(3) Å for Ru–N(11) and Ru–N(12).
In analogy the synthesis of [Ru(bdmpza)Cl(PTA)2] (56) was attempted starting from
[Ru(bdmpza)Cl(PPh3)2] (14) with two equivalents PTA either in THF or toluene under reflux.
The reaction formed a mixture of 56 and [Ru(bdmpza)Cl(PTA)(PPh3)] (55). Employing 2.2
equivalents of PTA and longer reaction times in THF under reflux led to a full conversion to
56. However, it was not possible to purify complex 56 convincingly as the excess amounts of
PTA could not be removed. Complex 56 shows high solubility in water as expected from the
similar cyclopentadienyl ruthenium complex [Ru(Cp)Cl(PTA)2].[351] The 1H NMR spectrum
shows a triplet at 4.17 and a multiplet around 4.01 ppm that can be assigned to the two
coordinated PTA ligands. The signals of the bdmpza ligand seem to indicate that the reduced
steric demand of the PTA ligand might allow a dynamic exchange between the two PTA
ligands and the chlorido ligand as for the protons in position 4 of the pyrazole moiety only
one signal at 5.79 ppm can be observed but the four methyl substituents lead to one singlet
consisting of two methyl substituents at 1.80 ppm and two further singlets at 2.12 and
2.21 ppm representing one methyl group each. The 13C NMR spectrum shows a symmetrical
complex as only one set of signals can be observed as the Me3 and Me3´ substituents give one
signal at 13.6 ppm and in consequence the Me5 and Me5´ substituents lead to one signal at
11.6 ppm. In the 31P NMR spectrum a triplet of the PTA ligand at –53.8 ppm can be observed
that cannot be conclusively explained by the suspected structure. Preliminary experiments
showed that the reaction of 56 with 9-ethynyl-9-fluorenol does not allow the formation of
allenylidene complexes as the formation of a complicated mixture of compounds can be
observed. The hot reaction mixture shows an intense red color that disappears upon cooling to
room temperature, which is in good agreement with previous reports. According to these,
PTA, if present in solution, can add to Cα of allenylidene ligands hereby forming an α-
phosphonioallenyl species.[352] Hence it was decided to focus on [Ru(bdmpza)Cl(PTA)(PPh3)]
(55) as precursor as the reaction with propargyl alcohols should displace the PPh3 ligand that
has previously not shown any addition reactions to the allenylidene moiety.
Reacting 55 with twofold excess of 9-ethynyl-9-fluorenol in THF at room temperature for two
days did not lead to any formation of an allenylidene complex in contrast to the
[Ru(bdmpza)Cl(PPh3)2] (14) based fluorenyl substituted complex [Ru(bdmpza)Cl-
(═C═C═(FN))(PPh3)] (7A/7B). However, heating to reflux for 16 h allowed the formation of
139
!Results and Discussion
!! !
a deep red allenylidene complex [Ru(bdmpza)Cl(═C═C═(FN))(PTA)] (57A/57B) that upon
heating for further 24 h under nitrogen atmosphere decomposes (Scheme 64). The
implementation of the polar PTA ligand led to a strongly decreased Rf value for the
CH2Cl2/acetone eluent mixture for column chromatography, which is usually employed in the
purification of bdmpza based allenylidene complexes. Therefore, a solvent mixture of
acetone/water (95:5 v/v) was used to separate the allenylidene complexes from the crude
product, yielding two fractions. The first purple complex was assigned to the structure 57A
and the second red complex was assigned to 57B. Due to the extremely low yields, i.e. 11%
for 57A and below 1% for 57B, only complex 57A could be characterized satisfyingly.
Complex 57A shows the typical IR absorption at 1923 cm–1 in CHCl3 indicating the
successful formation of an allenylidene complex. In the 1H NMR spectrum the asymmetric
bdmpza ligand can be observed, as four methyl substituents are present at 2.89, 2.58, 2.51 and
2.24 ppm. The PTA ligand is characterized by one singlet at 4.52 ppm and a doublet at
4.24 ppm (2JH,H = 6.8 Hz). In the 13C NMR spectrum the allenylidene moiety is
unambiguously assigned to the doublet at 294.2 ppm (2JC,P = 26.0 Hz, Cα) and a singlet at
230.3 ppm (Cβ). The remaining aromatic protons can be assigned to the fluorenyl moiety and
the bdmpza ligand confirms the asymmetric pattern with four different chemical shifts of
16.1, 13.0, 11.5 and 11.3 ppm. The PTA ligand shows the expected two doublets at 73.8
[M + H + MeCN – PPh3]+; Elemental analysis calcd (%) for
C54H42ClN4O3PRu × 1.25 CH2Cl2: C 62.10, H 4.20, N 5.24; found C 62.11, H 4.19, N 5.23.
NN N
N
Me
Me
Me
MeRu
OO
C ClPh3P
NN N
N
Me
Me
Me
MeRu
OO
CClPh3PC
C
C
A B
C
O
O
178
!Experimental Section
!! !
29B: Yield of 29A/29B: 46 mg (0.048 mmol, 43%).
The most prominent signals of compound 29B could be extracted from the combined spectra.
Due to the strong contamination with 29A the signals for Cα and Cβ could not be observed. 1H NMR (300 MHz, CD2Cl2): δ = 8.81 (s, 2H, PCO–H), 8.59 (s, 2H, PCO–H), 6.85 (s, 1H,
empirical formula C44H51ClN5O3PRu C36H35ClN5O3PRu × 0.25 CH2Cl2 formula weight [g mol–1] 865.39 774.41 crystal color / habit green block blue block crystal system triclinic triclinic space group, Z P–1 (No. 2), 2 P–1 (No. 2), 2 a [Å] 10.9262(8) 9.777(3) b [Å] 12.5322(12) 12.733(3) c [Å] 17.0543(16) 14.860(4) α [°] 107.131(7) 91.91(2) β [°] 93.042(8) 102.05(3) γ [°] 111.744(7) 108.86(2) V [Å3] 2037.9(3) 1701.8(8) θ [°] 6.2-26.5 6.25-26.5 h min, max – 13 to 13 – 12 to 12 k min, max – 15 to 15 – 15 to 15 l min, max – 21 to 21 – 18 to 17 F(000) 900 793 μ(Mo-Kα) [mm–1] 0.536 0.67 crystal size [mm] 0.254 × 0.225 × 0.169 0.165 × 0.099 × 0.079 Dcalcd [g cm–3] 1.41 1.511 T [K] 150(2) 150(2) reflections collected 17293 15195 indep. reflections 8277 6951 obs. reflections (>2σI) 6961 4138 parameter 504 450 wt. Parameter a 0.0206 0.0681 wt. Parameter b 1.2936 0.0000 R1, wR2 (obsd.) 0.0316, 0.0661 0.0738, 0.1428 R1, wR2 (overall) 0.0435, 0.071 0.1389, 0.1703 Diff. Peak / hole [e/Å3] 0.358 / –0.502 1.29 / –0.748 Goodness-of-fit on F2 1.041 1.014
Table 13. Structure determination details of complexes 15 and 16.
204
!Appendix
!! !
[Ru(bdmpza)Cl(═C═C═(FN))-(PPh3)] (20A)
[Ru(bdmpza)Cl(═C═C═(FN))-(PPh3)] (20B)
empirical formula C45H38ClN4O2PRu × H2O C45H38ClN4O2PRu × 2 CH2Cl2 formula weight [g mol–1] 852.30 1004.14 crystal color / habit brown block red block crystal system orthorombic triclinic space group, Z Pbca (No. 61), 8 P–1 (No. 2), 2 a [Å] 15.534(3) 12.0270(10) b [Å] 19.993(4) 12.4223(6) c [Å] 24.814(5) 14.9179(13) α [°] 90 98.766(5) β [°] 90 95.117(8) γ [°] 90 97.919(6) V [Å3] 7707(3) 2167.9(3) θ [°] 6.21-26.5 6.21-26.5 h min, max – 19 to 19 – 13 to 15 k min, max – 24 to 24 – 15 to 15 l min, max – 30 to 31 – 18 to 18 F(000) 3504.0 1024.0 μ(Mo-Kα) [mm–1] 0.565 0.752 crystal size [mm] 0.320 × 0.141 × 0.094 0.250 × 0.170 × 0.102 Dcalcd [g cm–3] 1.469 1.538 T [K] 150(2) 150(2) reflections collected 62130 25952 indep. reflections 7851 8825 obs. reflections (>2σI) 6066 7812 parameter 506 554 wt. Parameter a 0.0116 0.0901 wt. Parameter b 17.0595 10.8236 R1, wR2 (obsd.) 0.0389, 0.0762 0.0654, 0.1739 R1, wR2 (overall) 0.0622, 0.0855 0.0737, 0.181 Diff. Peak / hole [e/Å3] 0.819 / –0.523 1.513 / –1.81 Goodness-of-fit on F2 1.172 1.061
Table 14. Structure determination details of complexes 20A and 20B.
205
!Appendix
!! !
[Ru(bdmpza)Cl(═C═C═(AO))-(PPh3)] (25A)
[Ru(bdmpza)Cl(═C═C═(AO))-(PPh3)] (25B)
empirical formula C46H38ClN4O3PRu × CH2Cl2 C46H38ClN4O3PRu formula weight [g mol–1] 947.22 862.29 crystal color / habit brown block red block crystal system triclinic triclinic space group, Z P–1 (No. 2), 2 P–1 (No. 2), 2 a [Å] 12.7183(13) 12.1411(7) b [Å] 12.9619(4) 12.4107(11) c [Å] 14.5615(11) 14.8427(12) α [°] 86.985(5) 93.010(8) β [°] 69.949(6) 98.019(5) γ [°] 77.142(6) 98.291(5) V [Å3] 2197.7(3) 2185.4(3) θ [°] 6.22-26.5 6.2-26.5 h min, max – 15 to 15 – 15 to 15 k min, max – 16 to 16 – 15 to 15 l min, max – 18 to 18 – 18 to 18 F(000) 968 884 μ(Mo-Kα) [mm–1] 0.621 0.499 crystal size [mm] 0.307 × 0.168 × 0.112 0.305 × 0.282 × 0.254 Dcalcd [g cm–3] 1.431 1.310 T [K] 153(2) 150(2) reflections collected 54811 31254 indep. reflections 8972 8938 obs. reflections (>2σI) 8522 8005 parameter 585 509 wt. Parameter a 0.0188 0.0510 wt. Parameter b 6.5351 1.7391 R1, wR2 (obsd.) 0.0450, 0.1097 0.0315, 0.0932 R1, wR2 (overall) 0.0475, 0.1111 0.0362, 0.096 Diff. Peak / hole [e/Å3] 1.378 / –0.753 0.612 / –0.557 Goodness-of-fit on F2 1.233 1.089
Table 15. Structure determination details of complexes 25A and 25B.
206
!Appendix
!! !
[Ru(bdmpza)Cl(═C═C═(PCO))-(PPh3)] (29A)
[Ru(bdmpza)Cl(═C═C═(PCO))-(PPh3)] (29B)
empirical formula C54H42ClN4O3PRu × 2 CH2Cl2 C54H42ClN4O3PRu × 2 CH2Cl2 formula weight [g mol–1] 1132.26 1132.26 crystal color / habit violet plate yellow block crystal system triclinic triclinic space group, Z P–1 (No. 2), 2 P–1 (No. 2), 2 a [Å] 12.4473(11) 12.3917(5) b [Å] 13.2515(10) 13.4181(7) c [Å] 15.4170(9) 15.9204(3) α [°] 82.283(6) 90.222(3) β [°] 83.333(6) 93.896(4) γ [°] 85.279(6) 109.245(17) V [Å3] 2497.1(3) 2492.46(17) θ [°] 6.2-26.5 6.2-26.5 h min, max – 15 to 15 – 14 to 15 k min, max – 16 to 16 – 16 to 16 l min, max – 19 to 18 – 19 to 16 F(000) 1156.0 1156 μ(Mo-Kα) [mm–1] 0.664 0.665 crystal size [mm] 0.272 × 0.154 × 0.113 0.223 × 0.226 × 0.389 Dcalcd [g cm–3] 1.506 1.509 T [K] 150(2) 153(2) reflections collected 46376 24746 indep. reflections 10207 10166 obs. reflections (>2σI) 7515 9079 parameter 662 662 wt. Parameter a 0.0243 0.0859 wt. Parameter b 3.7984 0.5338 R1, wR2 (obsd.) 0.0506, 0.091 0.029, 0.0742 R1, wR2 (overall) 0.0829, 0.1015 0.0351, 0.078 Diff. Peak / hole [e/Å3] 0.749 / –0.677 1.735 / –0.789 Goodness-of-fit on F2 1.037 1.054
Table 16. Structure determination details of complexes 29A and 29B.
207
!Appendix
!! !
[Ru(bdmpza)Cl(CO)(PPh3)] (18B)
[Ru(bdmpza)Cl(═C═C═(BT))-(PPh3)] (37A)
empirical formula C31H30ClN4O3PRu C53H42ClN4O2PRu × 2 CH2Cl2 formula weight [g mol–1] 674.08 1104.25 crystal color / habit red block blue plate crystal system monoclinic triclinic space group, Z P 21/n (No. 14), 4 P–1 (No. 2), 2 a [Å] 16.4731(10) 11.095(2) b [Å] 10.7331(8) 11.579(2) c [Å] 18.2521(13) 20.263(4) α [°] 90 103.341(18) β [°] 113.487(4) 92.260(13) γ [°] 90 94.627(16) V [Å3] 2959.7(4) 2520.1(8)
θ [°] 6.2-26.5 6.21-26.5 h min, max – 12 to 20 – 13 to 13 k min, max – 9 to 13 – 14 to 14 l min, max – 22 to 20 – 25 to 25 F(000) 1376.0 1128.0 μ(Mo-Kα) [mm–1] 0.713 0.654 crystal size [mm] 0.165 × 0.135 × 0.120 0.216 × 0.143 × 0.139 Dcalcd [g cm–3] 1.513 1.455 T [K] 150(2) 150(2) reflections collected 19549 27332 indep. reflections 6050 10282 obs. reflections (>2σI) 4479 6684 parameter 374 644 wt. Parameter a 0.0368 0.0663 wt. Parameter b 3.7134 7.1940 R1, wR2 (obsd.) 0.0443, 0.0909 0.0792, 0.1699 R1, wR2 (overall) 0.073, 0.1019 0.1302, 0.1928 Diff. Peak / hole [e/Å3] 0.76 / –0.521 1.942 / –1.012 Goodness-of-fit on F2 1.021 1.047
Table 17. Structure determination details of complexes 18B and 37A.
208
!Appendix
!! !
[Ru(bdmpza)Cl(═C═CH(Pyr))-(PPh3)] (49)
[Ru(bdmpza)Cl-(═C═C═C(PhPyr))(PPh3)] (54B)
empirical formula C48H40ClN4O2PRu × C5H12 C55H44ClN4O2PRu formula weight [g mol–1] 1029.4 960.43 crystal color / habit red prism black block crystal system triclinic triclinic space group, Z P–1 (No. 2), 2 P–1 (No. 2), 2 a [Å] 11.995(6) 11.8575(4) b [Å] 15.01(3) 12.6866(5) c [Å] 15.12(6) 14.7270(4) α [°] 108.26(16) 89.741(3) β [°] 103.18(11) 84.460(2) γ [°] 94.47(7) 81.454(3) V [Å3] 2484(11) 2180.46(13) θ [°] 6.21-26.5 2.82-26.73 h min, max – 15 to 15 – 14 to 14 k min, max – 18 to 18 – 15 to 16 l min, max – 18 to 18 – 18 to 18 F(000) 1064 988 μ(Mo-Kα) [mm–1] 0.554 0.507 crystal size [mm] 0.249 × 0.211 × 0.202 0.2259 × 0.1757 × 0.1611 Dcalcd [g cm–3] 1.376 1.463 T [K] 150(2) 180(2) reflections collected 28077 19715 indep. reflections 10098 9197 obs. reflections (>2σI) 6629 8051 parameter 608 581 wt. Parameter a 0.0961 0.0169 wt. Parameter b 3.2770 2.4714 R1, wR2 (obsd.) 0.0746, 0.1735 0.0349, 0.0759 R1, wR2 (overall) 0.1257, 0.2024 0.0428, 0.0796 Diff. Peak / hole [e/Å3] 1.363 / –1.283 0.402 / –0.344 Goodness-of-fit on F2 1.027 1.102
Table 18. Structure determination details of complexes 49 and 54B.
209
!Appendix
!! !
[Ru(bdmpza)Cl(PTA)(PPh3)] (55) Pyrenophenone (50)
empirical formula C36H42ClN7O2P2Ru × CH2Cl2 C23H14O formula weight [g mol–1] 888.15 306.34 crystal color / habit yellow block yellow block crystal system orthorombic triclinic space group, Z Pbca (No. 61), 8 P–1 (No. 2), 2 a [Å] 20.263(3) 8.6960(5) b [Å] 9.0183(6) 8.9100(3) c [Å] 41.404(5) 11.1780(4) α [°] 90 69.723(3) β [°] 90 80.622(4) γ [°] 90 69.626(4) V [Å3] 7566.0(14) 760.71(6) θ [°] 6.21-27.5 2.97-27.49 h min, max – 26 to 25 – 11 to 11 k min, max – 11 to 11 – 11 to 11 l min, max – 53 to 53 – 14 to 14 F(000) 3648 320 μ(Mo-Kα) [mm–1] 0.756 0.08 crystal size [mm] 0.25 × 0.13 × 0.08 0.26 × 0.20 × 0.11 Dcalcd [g cm–3] 1.559 1.337 T [K] 150(2) 150(2) reflections collected 90917 21718 indep. reflections 8575 3482 obs. reflections (>2σI) 6344 2743 parameter 473 217 wt. Parameter a 0.0400 0.0836 wt. Parameter b 20.5593 0.1325 R1, wR2 (obsd.) 0.0504, 0.1036 0.0454, 0.1301 R1, wR2 (overall) 0.0797, 0.1146 0.061, 0.1434 Diff. Peak / hole [e/Å3] 1.007 / –1.318 0.358 / –0.293 Goodness-of-fit on F2 1.061 1.099
Table 19. Structure determination details of compounds 55 and 50.
210
!Appendix
!! !
6,6a-dihydro-11H-indeno[2,1-a]pyren-11-one (63)
11H-indeno[2,1-a]pyren-11-one (64)
empirical formula C23H14O C23H12O formula weight [g mol–1] 306.34 304.33 crystal color / habit red block yellow block crystal system orthorombic orthorombic space group, Z Pna21 (No. 33), 4 P212121 (No. 19), 4 a [Å] 14.4440(18) 4.5831(8) b [Å] 7.484(4) 15.039(3) c [Å] 13.509(3) 20.610(4) α [°] 90 90 β [°] 90 90 γ [°] 90 90 V [Å3] 1460.3(9) 1420.6(4) θ [°] 3.2-27.52 1.98-27.19 h min, max – 18 to 18 – 5 to 2 k min, max – 9 to 9 – 19 to 19 l min, max – 17 to 13 – 26 to 26 F(000) 640 632 μ(Mo-Kα) [mm–1] 0.084 0.086 crystal size [mm] 0.22 × 0.16 × 0.12 0.24 × 0.18 × 0.14 Dcalcd [g cm–3] 1.393 1.423 T [K] 150(2) 150(2) reflections collected 16840 11834 indep. reflections 3115 3150 obs. reflections (>2σI) 2450 2385 parameter 217 217 wt. Parameter a 0.0636 0.074 wt. Parameter b 0.6588 0.0000 R1, wR2 (obsd.) 0.00546, 0.1231 0.0498, 0.1224 R1, wR2 (overall) 0.00774, 0.1362 0.0787, 0.1446 Diff. Peak / hole [e/Å3] 0.519 / –0.303 1.513 / –1.81 Goodness-of-fit on F2 1.069 1.106
Table 20. Structure determination details of compounds 63 and 64.
211
!Appendix
!! !
8.2 Cyclic Voltammetry
Figure 62. Cyclic voltammogram for complexes 14, 19B, 20A, 20B, 25A, 25B, 37A, 37B, 54A and 54B in CH2Cl2 with nBu4NPF6 (0.1 M) as supporting electrolyte at a scan rate of 100 mV/s (vs. Fc/Fc+) (* indicates signal corresponding to Fc/Fc+).
212
!Appendix
!! !
Figure 63. Cyclic voltammogram for complex 29B in MeCN with nBu4NPF6 (0.1 M) as supporting electrolyte at a scan rate of 200 mV/s (vs. Fc/Fc+).
213
!Appendix
!! !
Table 21. Cyclic voltammograms were recorded at 20 °C in dichloromethane, with nBu4NPF6 (0.1 M) as electrolyte, potentials are given relatively to ferrocenium/ferrocene as internal standard, at a scan rate of 100 mV/s.
Compound
E1/2 (m
V)
ΔEp a (m
V)
ipa /ipc
E1/2 (m
V)
ΔEp a (m
V)
ipa /ipc
E1/2 (m
V)
ΔEp a (m
V)
ipa /ipc
reduction processes
Ru(II)/ Ru(III)
14 - - - - - -
394
73
0.80
19B
–1631
92
0.67
- - -
265
82
0.71
20A
–1932
85
0.45
–1273
64
0.76
389
73
0.94
20B
–1937
74
0.96
–1273
64
0.64
371
73
0.91
25A
–1479
74
0.99
–1013
73
0.94
466
64
0.98
25B
–1315
91
0.88
–870
82
0.94
641
83
0.95
37A
–1914
82
0.83
–1228
83
0.71
87
73
0.86
37B
–1859
64
0.97
–1168
73
0.61
188
72
0.97
54A
–2178
105
0.23
–1548
64
0.17
252
73
0.85
54B
- - -
–1182
64
0.07
320
64
0.55
214
!Appendix
!! !
8.3 Myoglobin Assay of CORMs
Figure 64. CO release of [Mn(bpzp)(CO)3] 4 in the dark measured via myoglobin assay.
Figure 65. CO release of [MnBr(HPz)2(CO)3] 5 in the dark measured via myoglobin assay.
0 200 400 600 8000.0
0.2
0.4
0.6
0.8
1.0
eq. (
CO
)
t [min]
0 50 100 150 200 250 300 350 400 4500.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
eq. (
CO
)
t [min]
215
!Appendix
!! !
Figure 66. CO release of [Mn(HIm)3(CO)3]Br 6 in the dark measured via myoglobin assay.