Comparison of Carbazole and Fluorene Donating Effects on the Two-Photon Absorption and Nitric Oxide Photorelease Capabilities of a Ruthenium–Nitrosyl Complex

Article abstract of DOI:10.1002/ejic.201700895

A ruthenium–nitrosyl derivative of formula [Ru^II(CzT)(bipy)(NO)](PF₆)₃ [CzT = 4′-(N-ethylcarbazol-3-yl)-2,2′:6′,2′′-terpyridine, bipy = 2,2′-bipyridine] has been synthesized and fully characterized, and compared with the previously reported [Ru^II(FT)(bipy)(NO)](PF₆)₃ complex [FT = 4′-(9,9-dihexyl-9H-fluoren-2-yl)-2,2′:6′,2′′-terpyridine]. Additionally, the X-ray crystal structure of [Ru^II(CzT)(bipy)(NO₂)](PF₆), the precursor of [Ru^II(CzT)(bipy)(NO)](PF₆)₃, is reported. The presence of a tertiary amine in the carbazole unit leads to redshifted charge-transfer transitions towards the electron-withdrawing Ru–NO fragment and hence enhanced two-photon absorption (TPA) properties. In contrast, the quantum yield of the NO· photorelease process is lower for the carbazole-containing complex. The issue of optimization of the TPA versus NO·-release capabilities is addressed.

Full text of DOI:10.1002/ejic.201700895

A Journal of
Accepted Article
Title: Comparison of carbazole and fluorene donating effects on the
two-photon absorption and nitric oxide photo-release capabilities
in a ruthenium-nitrosyl complex
Authors: Alejandro Enriquez-Cabrera, Pascal Georges Lacroix,
Isabelle Sasaki, Sonia Malet-Ladeira, Norberto Farfan,
Rodrigo Barba-Barba, Gabriel Ramos-Ortiz, and Isabelle
Malfant
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: Eur. J. Inorg. Chem. 10.1002/ejic.201700895
Link to VoR: http://dx.doi.org/10.1002/ejic.201700895

European Journal of Inorganic Chemistry
10.1002/ejic.201700895
Comparison of carbazole and fluorene donating effects on the two-photon
absorption and nitric oxide photo-release capabilities in a ruthenium-
nitrosyl complex
Alejandro Enriquez-Cabrera,[a1,b] Pascal G. Lacroix*,[a] Isabelle Sasaki,[a] Sonia Malet-
[
a]
[b]
[c]
[c]
Ladeira, Norberto Farfan, Rodrigo M. Barba-Barba, Gabriel Ramos-Ortiz, and Isabelle
Malfant.*[a]
Keywords: ruthenium complexes, nitric oxide, drug delivery, density functional calculations,
photochemistry, electronic structure, charge transfer.
II
Abstract: A ruthenium nitrosyl derivative of formula [Ru (CzT)(bipy)(NO)](PF
6
3
) , where
CzT is the 4’-(N-ethyl-carbazol-3-yl)-2,2’:6’,2’’-terpyridine, and bipy the 2,2’-bipyridine has
been synthesized and fully characterized, and compared to the previously reported
II
[
Ru (FT)(bipy)(NO)](PF
6
)
3
complex (FT = 4’-(9,9-dihexyl-9H-fluorene-2-yl)-2,2’:6’,2’’-
II
terpyridine). Additionally, the X-Ray crystal structure of [Ru (CzT)(bipy)(NO
2
)](PF ),
6
II
precursor of [Ru (CzT)(bipy)(NO)](PF
6
3
) , is reported. The presence of a tertiary amine on the
carbazole unit leads to red-shifted charge transfer transitions towards the withdrawing Ru-NO
fragment and hence enhanced two-photon absorption (TPA) properties. By contrast, the
quantum yield of the NO· photo-release process appears lowered in the carbazole derivative.
The issue of the optimization of the TPA vs NO· release capabilities is addressed.
Introduction
Nitric oxide (NO·) is a highly reacting radical naturally produced by nitric oxide
synthase (NOS) which plays a key role in various physiological processes such as blood
[a]
CNRS, Laboratoire de Chimie de Coordination (LCC)
05 route de Narbonne, BP44099, Toulouse Cedex 4, 31077 France
2
E-mail : pascal.lacroix@lcc-toulouse.fr
isabelle.malfant@lcc-toulouse.fr
[
b]
c]
Facultad de Quimica, Departamento de Quimica Orgánica
Universidad Nacional Autónoma de México
0
4510 México D.F., México
[
Centro de Investigaciones en Óptica (CIO) A.P. 1-948, 37000 León, Gto. (México)
1
This article is protected by copyright. All rights reserved.

European Journal of Inorganic Chemistry
10.1002/ejic.201700895
regulation, immune response, neurotransmission, respiration and cytotoxic activity in tumor
cells by apoptosis.[1,2] However, at concentration different than its physiological level, NO·
can induce various cardiovascular, neurological, pulmonary diseases, atherosclerosis and
cancer.[3-5] In this context, there has been an increasing interest for the investigation of
exogenous donors, capable to release NO· locally and quantitatively. Among them, ruthenium
nitrosyl complexes have been recognized as the most promising candidates, in relation to their
generally low toxicity, good stability and capability of releasing NO· under light irradiation,
exclusively,[6] taking advantage of the noninvasive and highly controllable characteristics of
light. Additionally, while most NO· donors require to be irradiated in the 300-500 nm
wavelengths domain, the use of the two-photon absorption (TPA) technique,[7-11] in which the
molecules absorb simultaneously two photons instead of one, allows working at double
wavelength, in the 600-1200 nm therapeutic window of relative transparency of the biological
media.[12] Owing to these appealing perspectives, the issue of designing ruthenium nitrosyl
complexes exhibiting sizeable TPA capabilities is naturally addressed.
We have recently reported on ruthenium nitrosyl complexes built up from 4’-(2-
fluorenyl)-2,2’:6’,2’’-terpyridine ligands, capable of releasing NO· with good quantum yields,
[13]
under irradiation at= 405 nm.
In these compounds, an electron-rich fluorenyl unit has
been introduced on the ligand in relation to the widely reported capabilities of fluorene to
enhance the TPA properties of molecules.[14-17] Our best result in this approach is provided by
II
3+
3+
the [Ru (FT)(bipy)(NO)] complex ([1] in Scheme 1), in which bipy is the 2,2’-bipyridine
and FT is the 4’-(9,9-dihexyl-9H-fluorene-2-yl)-2,2’:6’,2’’-terpyridine in which two
additional hexyl chains are introduced in order to increase the solubility. The TPA capability
of [1]3+ expressed as its molecular cross-section (TPA) was found to be equal to 108  18
GM, a value probably far from being optimized. Indeed, it has been shown that increasing the
“push-pull” character of the -conjugated skeleton usually leads to an increase of TPA in
dipolar molecules.[7-11] Therefore, the idea of replacing the fluorenyl unit by a more donating
fragment arises naturally.
II
3+
In the present contribution, we report on the new [Ru (CzT)(bipy)(NO)] complex
3
+
3+
(
[2] in Scheme 1), a chromophore related to [1] in which the slightly electron-donor-9,9-
dihexyl-9H-fluorene unit is replaced by a 9-ethylcarbazole, in which a tertiary amine is
conjugated within the three-cyclic structure with the expected outcome of an enhanced push-
pull effect towards the withdrawing Ru(NO) unit and hence, a better TPA response. After the
report of the synthesis and characterization of [2](PF
6
)
3
, its spectroscopic and photochemical
2
This article is protected by copyright. All rights reserved.

European Journal of Inorganic Chemistry
10.1002/ejic.201700895
properties are compared to those of the parent [1](PF
6
)
3
derivative. The observed differences
are analyzed computationally within the framework of the time dependent density functional
3
+
3+
3+
3+
theory (TD-DFT), which will also be applied to [1’] and [2’] , isomers of [1] and [2] , in
order to fully evaluate the capabilities of fluorene and carbazole to behave as electron donors
towards a withdrawing fragment present at positions 2- and 3- on the phenyl ring of the donor.
Finally, the capabilities of carbazole-containing [Ru (terpy)(bipy)(NO)]3+ towards both
efficient NO· release and TPA properties are critically evaluated.
II
Scheme 1. Molecular formula of the four ruthenium nitrosyl complexes under investigation.
Results and Discussion
Synthesis and Characterization
We have previously reported on various synthetic approaches towards ruthenium
nitrosyl complexes built up from terpyridine ligands, with the aim of avoiding (i) the
II
]2+ homoleptic species[13b] and (ii) the tedious
formation of undesirable [Ru (terpyridine)
2
purification step required to separate the trans/cis isomers when two chlorides are present as
ligands in the coordination sphere of the ruthenium center[13a]. The general route reported on
the present investigation for the synthesis of [2](PF
6
3
) is shown in Scheme 2, which implies
first the synthesis of the 4’-(N-ethyl-carbazol-3-yl)-2,2’:6’,2’’-terpyridine ligand (E).
3
This article is protected by copyright. All rights reserved.

European Journal of Inorganic Chemistry
10.1002/ejic.201700895
6 3
Scheme 2. Synthetic route towards ligand E (top) and [2](PF ) ruthenium nitrosyl complex (bottom).
E was obtained in a similar manner as for the fluorene derivatives[13] previously
reported, and using a modification of the previously reported method.[18] It implied first
brominating the carbazole using DMF as a solvent, after which an alkylation was performed
using a novel methodology[19] involving water, sodium hydroxide and a phase transfer catalyst
(tetrabutyl ammonium iodide) at 30°C with vigorous stirring for 24 hrs resulting in the pure
product B after recrystallization. This approach avoids the use of organic solvents like DMF
that are normally used for the alkylation of carbazole, and the use of the harsh reagent
NaH.[20] The preparation of N-ethyl-carbazol-3-carboxaldehyde (C) and consequent
condensation with 2-acetylpyridine in basic medium results in the product (E)-3-(9-ethyl-9H-
carbazol-3-yl)-1-(pyridin-2-yl)prop-2-en-1-one (D); the ligand E was obtained using a
Kröhnke condensation between D and the Kröhnke salt 1-(2-oxo-2-pyridin-2-yl-
ethyl)pyridinium iodide. Note that, if hexyl chains were to be used instead of ethyl chains, the
products B and C would be oils requiring a purification by column chromatography.
The procedures towards the different complexes were the same as for the previously
reported fluorene-based complexes.[13a] First reacting the ligand E with 1 eq. of RuCl
in
3
ethanol resulted in complex F, which was reacted with 2,2’-bipyridine in the presence of an
excess of LiCl and triethylamine as a reducing agent to give complex G. By using sodium
nitrite, the chloride ligand was replaced by a nitro ligand (H), and finally using 12M HCl to
4
This article is protected by copyright. All rights reserved.

European Journal of Inorganic Chemistry
10.1002/ejic.201700895
afford the metathesis of the nitro ligand to the nitrosyl obtaining the desired complex
[
2](PF ) .
6
3
During the course of the reaction a qualitative exploration of the color of the product
either in solution or in solid state) suggested that the reaction may be completed, since F is
is purple again.
(
brown and G is purple, H is red and the final nitrosyl complex [2’](PF
)
6 3
This evolution is readily confirmed by the UV-Vis spectra which show that G (chlorido
ligand) exhibits an absorption maximum located at 517 nm, H (nitro ligand) at 490 nm and
the final [2’](PF
6
)
3
at 517nm, making it possible to follow the different reaction steps by UV-
1
Vis spectroscopy. Additionally, H-NMR spectroscopy is a valuable tool for determining the
ligand attached to the ruthenium center, due to the shift of the H-6 of the bipyridine that goes
from 10.30 ppm to 9.95 ppm and 9.35 ppm, for the chloride, nitro, and nitrosyl ligands,
respectively, as illustrated in Figure 1. This feature is related to the fact that H-6 is spatially
-
-
+
close to the ligands (Cl , NO
2
, NO ) in the ruthenium complexes, so any substitution leads to
a significant chemical shift due to a different electronic environment. The nature of the ligand
-1
(
NO
2
/NO) is confirmed by IR spectroscopy with the presence of bands at 1331 cm and 1295
-1
-1
cm (nitro) or at 1937 cm (nitrosyl).
1
H-NMR (400MHz, CD CN)
3
H
H
H
H
H
H
Figure 1
1H-NMR spectra showing the sensitivity in the chemical shift of H-6 (in red color) on going
from G (top), to H (middle), and [2]3+ (bottom).
Structural studies
5
This article is protected by copyright. All rights reserved.

European Journal of Inorganic Chemistry
10.1002/ejic.201700895
To date, we have reported on three X-ray crystal structures of ruthenium nitrosyl
complexes built up from fluorenyl-terpyridine ligands.[13] In any case, the crystals suitable for
the X-ray diffraction measurements were grown by slow diffusion of diethylether into a
concentrated solution of the desired complex in acetonitrile. This approach failed here, in the
case of carbazole-based species, which appears much more difficult to crystallize despite
many attempts carried out in various solvent conditions. Nevertheless, suitable crystals of
complex H which contains the Ru(NO ) core instead of Ru(NO), were finally obtained from a
2
diffusion of diethylether into a concentrated solution of the complex made of a 1:1 mixture of
acetone/DMF.
II
[
Ru (bipy)(E)(NO
2
)](PF
6
) (H) crystallizes in the monoclinic P 2 /n space group. The
1
main crystal data are gathered in Table 1. The overall structure results from a tridimensional
array of cationic complexes linked to each other in short contacts through (PF
)- anionic
6
bridges. The asymmetric unit cell, shown in Figure 2, is built up from one
II
[
Ru (bipy)(E)(NO
2
)](PF ) entity with one molecule of DMF. Importantly, a disorder appears
6
-
at the NO
2
ligand, which was successfully treated by introducing both a nitro group and a
chlorine atom in a ratio 90:10 after refinement, with a Ru-Cl distance of 2.32(2) Å. This
distance falls in the 2.317(3) – 2.394(1) Å range of Ru-Cl bond lengths observed in our
II
(NO)](PF
) complexes,[21] and therefore appears
previous report on [Ru (R-terpyridine)(Cl)
2
6
chemically satisfactory. The unexpected presence of chlorine in the crystal structure is likely
related to the fact that H is synthesized from complex G, which contains the RuCl core
instead of Ru(NO
traces of the product obtained at the previous synthetic step. Additionally, the ruthenium-
nitrogen distance within the Ru(NO ) core is equal to 2.009(8) Å, a rather short value but
compatible with the references found in the literature. Indeed, 18 entries are found from the
Cambridge crystallographic data base, containing the “Ru-terpyridine-NO ” core. Among
them the shortest Ru-NO
distances are 2.013(2) Å, [22] 2.029(2) Å.[23,24], and 2.029(3) Å.[25]
2
). The sample used for the crystal growth could therefore have contained
2
2
2
II
+
The molecular structure of the [Ru (bipy)(E)(NO )] cation appears less distorted than
2
that of the previously reported [Ru (4’-fluorenylterpyridyne)(bipy)(NO)]3+ complex (non-
II
3+
alkylated version of [1] ), in particular at the terpyridine level where the longest deviation
from the mean plane (18 atoms) was 0.143 Å, a value reduced to 0.090 Å in the present case.
Additionally, the angle between the averaged mean planes of terpyridine and carbazole (11
atoms) is found equal to 5.55°, which indicates a complete  delocalization. The coordination
6
This article is protected by copyright. All rights reserved.

European Journal of Inorganic Chemistry
10.1002/ejic.201700895
sphere around the ruthenium(II) atom is pseudo octahedral, with selected bond lengths and
angles listed in Table 2.
Table 1.Crystal data and structure refinement parameters for H.
_
________________________________________________________
Chemical formula
Formula weight
Crystal system
Space group
T (K)
C
39
H
30Cl0.1
N
6.9
6
O1.8Ru, F P, C
3
7
H NO
946.76
monoclinic
P2 /n
1
100(2)
Wavelength (Mo K) (Å)
a (Å)
0.71073
11.6045(15)
14.4506(19)
23.486(3)
90
b (Å)
c (Å)


(°)
(°)
93.979(4)
90
(°)
V (Å3)
3928.9(9)
4
Z
-3
D
calc. (g cm )
1.601
-1
Absorption coefficient (mm )
Reflections collected
uniques
0.527
68343
8075
R
R
int
0.0969
0.0434
0.0877
0.1010
1.046
a
1
wR b
[I>2(I)]
for all]
GOF (F2)
2
[
_
________________________________________________________
a
푅 = ∑|ꢀ퐹 ꢀ − ꢀ퐹 ꢀ|⁄∑ꢀ퐹 ꢀ
1
0
퐶
0
b
2
0
2
퐶
2
2
0
2
1/2
푤푅 = [∑[푤(퐹 − 퐹 ) ]⁄∑[(푤퐹 ) ]]
2
7
This article is protected by copyright. All rights reserved.

European Journal of Inorganic Chemistry
10.1002/ejic.201700895
Figure 2
Table 2
Asymmetric unit cell for complex H, with relevant atom labeling scheme. Displacement
ellipsoids are drawn at the 50 % probability level. The hydrogen atoms are omitted for clarity.
Selected bond lengths (Å) and angles (°) in complex H compared to those of the non-alkylated
3
+ a
[
1] .
H
[1]3+
H
[1]3+
_
_________________________________________________________________________
Ru-N(1)
2.009(8)
2.064(3)
2.087(3)
1.762(2)
N(1)-Ru-N(4)
89.9(3)
87.4(2)
90.6(2)
93.13(10)
98.14(9)
94.91(10)
Ru-N(2)bpy
Ru-N(3)bpy
2.0871(19) N(1)-Ru-N(5)
2.0931(19) N(1)-Ru-N(6)
2.081(2)
Ru-N(4)terpy 2.069(3)
Ru-N(5)terpy 1.965(3)
Ru-N(6)terpy 2.074(3)
1.9814(18) N(4)-Ru-N(5)
79.10(11)
79.63(12)
79.79(8)
79.35(8)
2.085(2)
N(5)-Ru-N(6)
_
_________________________________________________________________________
data from reference [13a]
a
3+
The averaged data indicates a slight contraction of 0.014 Å in H, compared to [1] ,
for the 5 Ru-N bond lengths involving the terpyridine and bipyridine ligands. The presence of
a longer Ru-N distance (2.009(8) Å) for the nitrate ligand in H versus 1.763(2) Å to the
3+
nitrosyl ligand in [1] , may qualitatively account for this difference. While the position of the
ruthenium atom in the vicinity of the terpyridine (N(4), N(5) and N(6)) is roughly the same
8
This article is protected by copyright. All rights reserved.

European Journal of Inorganic Chemistry
10.1002/ejic.201700895
for both derivatives in distances and angles, a sizeable difference is observed in the position
of N(1), with a Ru-N(1) bond nearly perpendicular to the terpyridine-ruthenium plane in H, in
3
+
striking contrast to [1] , in which the Ru-NO bond is bent in the direction of the bipyridine
with N(I)-Ru-N(4), Ru-N(5), and Ru-N(6) angles significantly larger than 90°. By contrast,
the coordination sphere around the ruthenium atom of H exhibits a stronger octahedral
character.
Although no experimental crystal structure is available for ruthenium nitrosyl
complexes with carbazole, it may be interesting to target the donating capabilities of the
different ligands on the basis of structural insights obtained by DFT computations along the
3
+
3+
3+
3+
[
1] , [2] , [1’] , [2’] series. The relevant data are gathered in Table 3. In particular, the
torsion angle between the averaged planes of the terpyridine and those of the donating (either
fluorene or carbazole) substituents appears to be a parameter of interest to evaluate the charge
transfer capabilities allowed by -overlaps between the organic fragments. It is known that
the transfer integral between the subunits, and hence the overlap integral vary as the cosine
of the torsion angle,[26] thus leading to enhanced push-pull effects towards the withdrawing
3+
nitrosyl at lower angle values. It appears clearly from the table that the structure of [2] is the
most favorable one in terms of charge transfer effect and hence, TPA capabilities.
3+
3+
3+
Table 3
Selected geometrical parameters in the DFT computed structures of the [1] , [2] , [1’] ,
3
+
[
2’] : bond lengths (Å) and torsion angle (°) between the planes of the terpyridine and the
donating (fluorene or carbazole) substituent.
C-C bond
torsion angle
_
_________________________________________
3
3
+
+
[
[
[
1]
2]
1.467
1.461
1.472
1.467
31.28
28.30
33.19
30.77
3
3
+
1’]
2’]
+
[
_
_________________________________________
Additionally, the frequency of the (NO) stretching vibration is also a good indicator
of the amount of electron density transferred to the strongly withdrawing nitrosyl ligand
through the well-known Ru → *NO back-bonding.[6b,27] Indeed, the electron density is
invariably transferred to an antibonding orbital, with the outcome of a lowering of the N-O
bond order and hence, a lowering of its infrared stretching frequency. In the present case,
II
9
This article is protected by copyright. All rights reserved.

European Journal of Inorganic Chemistry
10.1002/ejic.201700895
-1

(NO) is equal to 1937 cm in [2](PF
6
)
3
, which appears slightly switched to lower energy
-
1
versus the previously reported value of 1942 cm for [1](PF
keep in mind that great care must be taken in the analysis of the energy shifts observed for
(NO) in the solid state. For instance, it has been reported that ruthenium nitrosyl complexes
with terpyridine could exhibit different (NO) frequencies, depending on the nature of the
counter ions (e.g. with PF -[28] and Cl-[29]), due to different solid state environments.
Nevertheless, the DFT computations performed on isolated cationic complexes confirm the
6
3
) . However, it is important to

6
-1
tendency (Table 4). The lowering of 5 cm observed in [2](PF
1](PF compound, suggests that a carbazole unit behaves as a better electron donor than a
fluorene, as anticipated from the effect of its electron-rich nitrogen atom. Along this line, [2]3+
6
)
3
, with respect to the parent
[
)
6 3
3+
3+
3+
3+
appears to be the best candidate along the [1] , [2] ,[1’] ,[2’] series of complexes.
Altogether, computed structural data and infrared stretching frequencies correlate fairly well,
and suggest that the push-pull charge transfer capabilities varies in the following order: [2]3+
3+
3+
3+
>
[2’]  [1] > [1’] .
Table
4
6 3 6 3
Experimental Infrared (NO) stretching frequency for [1](PF ) and [2](PF ) , compared to the
3+
3+
3+
3+
DFT computed values obtained for [1] , [2] ,[1’] ,[2’] .
compound
DFT computation
experimental (NO)
_
________________________________________________
1]3+
2]3+
2045 cm-1
2039 cm-1
1942 cm in [1](PF
-1
)
[
[
6
3
3
-1
1937 cm in [2](PF
6
)
[
1’]3+
2’]3+
2048 cm-1
2045 cm-1
[
_
_______________________________________________
Spectroscopic Properties
The UV-visible spectra of [1](PF
6
)
3
and [2](PF
6
3
) in acetonitrile are compared in
Figure 3. In both cases, the complexation leads to the appearance of low-lying transitions
which are absent in the related terpyridine ligands. This observation suggests that he Ru(NO)
core is strongly involved in these transitions, which are all potentially active in NO· release
1
0
This article is protected by copyright. All rights reserved.

European Journal of Inorganic Chemistry
10.1002/ejic.201700895
processes. At first glance, the effect of the substitution of fluorene by carbazole is that of a
global red shift. More precisely, [2](PF is dominated by a low-lying intense band centered
at 517 nm ( = 14 600 L.mol .cm ), which has to be compared with that of [1](PF
)
6 3
-1
-1
)
6 3
-1
-1
observed at 453 nm ( = 16 700 L.mol .cm ). The sizeable shift of 64 nm is indicative of an
enhancement of the push-pull effect, readily ascribed to the better donating capability of the
carbazole, as discussed in the previous section.
Figure 3
UV-visible spectra of [1](PF
6 3 6 3
) (blue) and [2](PF ) (red) in acetonitrile (top), with that of their
related terpyridine ligand (bottom).
In order to get more theoretical insight on the origin and full extent of these
differences, the TD-DFT computed spectra are shown in Figure 4. The global red shift on
3
+
3+
going from [1] to [2] is confirmed by the calculation. In particular, the wavelengths for the
low-energy dominant transitions are compared to the experiment in Table 5. The slight
discrepancy observed in the high energy transition of [2](PF
6
3
) (380 nm / 436 nm) remains
-1
below 3 500 cm , which is still acceptable in the case of molecules containing heavy atoms,
in which the transitions have a strong charge transfer character.[30] The fair agreement between
computation and UV-visible data makes the analysis relevant at the orbital level.
1
1
This article is protected by copyright. All rights reserved.

European Journal of Inorganic Chemistry
10.1002/ejic.201700895
3
+
3+
3+
3+
Figure 4
Table 5
TD-DFT computed electronic spectra for [1] , [2] ,[1’] , and [2’] .
-
1
-1
UV-visible spectra (max in nm and in L.mol .cm ) compared to TD-DFT data (max and
oscillator strengths f) for [1](PF and [2](PF
6
)
3
6 3
) .
TD-DFT
UV-visible

max
f

max

_
______________________________________________________
[
1](PF
6
)
3
432 0.336
63 0.398
480 0.297
80 0.522
453 16 700
367 18 900
517 14 600
436 11 300
3
[
2](PF
6
)
3
3
_
______________________________________________________
1
2
This article is protected by copyright. All rights reserved.

European Journal of Inorganic Chemistry
10.1002/ejic.201700895
3
+
3+
3+
The description of the transitions involved in the electronic properties of [1] , [2] ,[1’] ,
and [2’]3+ are shown in Table 6. It clearly appears that the analysis of the properties at the
molecular level can be restricted to a limited set of orbitals: HOMO and LUMO for the
fluorenyl [1], and HOMO-1 HOMO and LUMO for the carbazole [2] derivatives. These
orbitals are shown in Figure 5. The donating character of the fluorene and carbazole
substituents and the withdrawing character of the ruthenium-nitrosyl units clearly appear from
the Figure together with the higher energy of the occupied levels of the carbazole, in strong
indication of (i) its enhanced donating character and (ii) the resulting push-pull effect
3
+
expected in [2] .
1
3
This article is protected by copyright. All rights reserved.

European Journal of Inorganic Chemistry
10.1002/ejic.201700895
Table 6
Relevant TD-DFT data for the main transitions of ruthenium complexes: absorption maxima (max in nm), oscilator strengh (f), change in dipole
moment ( in D) in the low-lying transition, composition of the configuration interaction (CI) and dominant character of the charge transfer.
Compound
transition

max
f

composition of CI expansion*
character
_________________________________________________________________________________________________________________
[
[
1]3+
1 → 2
→ 7
432
363
0.336
0.398
30.9 167→168 (80%)
fluorene → Ru(NO)
1
15.0 167→170 (34%) + 164→169 (18%)
fluorene-terpy → terpy-Ru(NO)
1’]3+
1 → 3
411
339
336
0.091
0.200
0.190
36.3 167→168 (79%)
164→169 (68%) + 165→168 (16%)
167→170 (74%)
fluorene → Ru(NO)
1
1
→ 10
→ 11
fluorene-terpy → terpy-Ru(NO)
fluorene → terpy-Ru(NO)
[
[
2]3+
1 → 2
→ 8
480
380
0.297
0.522
27.5 167→168 (86%)
7.70 167→170 (70%)
carbazole → Ru(NO)
1
carbazole → terpy-Ru(NO)
2’]3+
1 → 3
434
362
346
0.337
0.379
0.245
27.9 166→168 (81%)
carbazole → Ru(NO)
1
1
→ 10
→ 12
166→170 (35%) + 164→169 (17%)
166→170 (43%) + 164→169 (23%)
carbazole-terpy → terpy-Ru(NO)
carbazole-terpy → terpy-Ru(NO)
_
*
________________________________________________________________________________________________________________
Excitations contributing more than 15% to the transitions. For the four compounds, orbital 167 is the HOMO and orbital 168 is the LUMO.
1
4
This article is protected by copyright. All rights reserved.

European Journal of Inorganic Chemistry
10.1002/ejic.201700895
3+
3+
3+
3+
Figure 5
Main orbitals and relative energies computed by DFT for [1] , [2] ,[1’] , and [2’] .
1
5
This article is protected by copyright. All rights reserved.

European Journal of Inorganic Chemistry
10.1002/ejic.201700895
Cyclic voltammetry experiments provide additional features for enhanced ligand
donating capabilities in [2](PF 3. The redox potentials measured for [1](PF 3, and [2](PF
6
)
6
)
)
6 3
are presented in Table 7. The presence of two reduction waves in the voltamograms (Figure
S1) is consistent with previous reports on Ru(terpy)(NO) derivatives,[31-33] and is ascribed to
+
-
RuNO → RuNO· and RuNO· → RuNO reduction processes with a potential difference of
E = 700 mV for each compounds leading to comproportionation constants (Kc) of 0.7 
012 (RT lnKc = E/0.059). Slightly lower values of reduction potentials are observed in
2](PF , with an offset of 20 mV. This difference indicates the presence of a less reducible

1
[
)
6 3
3+
nitrosyl fragment on [2] . This feature indicates an enhanced electron density on the nitrosyl
and therefore an enhanced donating effect induced by the carbazole compared to that of the
fluorenyl ligand. The weak lowering value of 20 mV is related to the fact that the redox
features are centered on the Ru(NO) core with similar environment in both complexes. By
contrast, the differences in the orbitals is more pronounced at the ligand centered occupied
levels (fluorene vs carbazole) leading to a more sizeable shift in the absorption spectra (Figure
3
).
Table 7 Electrochemical data in acetonitrile (E1/2 vs SCE in Volts) for [1](PF
)3, and [2](PF ) .
6 6 3
+
NO·/NO-
Compounds
NO /NO·
_
_____________________________________________________
[
[
1](PF
2](PF
6
)
)
3
0.45 (reversible)
0.43 (reversible)
-0.25 (irreversible)
-0.27 (irreversible)
6
3
_
_____________________________________________________
TPA properties
To quantify the TPA properties we employed the Z-scan technique (due to absence of
significant fluorescence in the ruthenium complexes the use of other techniques such as two-
photon excited fluorescence are precluded). Within the present experimental conditions, an
incident wavelength of 800 nm leads to a TPA absorption at 400 nm, which does not
correspond the absorption maxima of the ruthenium complexes (453 nm and 517 nm, for
[
1](PF
6
)
3
and [2](PF ) , respectively). Nevertheless, and apart from the fact that two-photon
6
3
electronic spectra may be significantly different than one-photon spectra, it is important to
1
6
This article is protected by copyright. All rights reserved.

European Journal of Inorganic Chemistry
10.1002/ejic.201700895
keep in mind that the 800 nm wavelength is the most widely used wavelength in biophotonics
applications for practical reasons, and therefore was the one selected here.
At the molecular level, the quantification of the TPA properties is expressed by a
cross-section (TPA), expressed in Göppert-Mayer units (1 GM = 10-50 cm s photons-1
4
-1
molecules ). With Z-scan we measured the nonlinear absorption coefficient ( ) in acetonitrile
solutions of [2]3+ at the concentration of 10 mol.L . Samples of high concentrations are
necessary in Z-scan to assure traces of transmission detected at the far field with good signal
to noise ratio. Figure 6 presents typical traces of normalized transmission (T(z)) (see
-2
-1
Experimental Section) obtained from Z-scan experiments in samples of [2](PF
a 1 mm-thick quartz cuvette. The presented traces correspond to pulse energies from 180nJ to
60nJ and a pulsewidth of 350 fs . It is observed that the Z-scan traces exhibit good symmetry
6
3
) contained in
4
around z = 0 position and the normalized transmission decreased as the intensity of pulses
was increased. The intensity dependence of transmission is a distinctive feature of a nonlinear
absorption process. Good fitting to the experimental data using the mathematical formalism
for Z-scan[34] corroborated TPA as the origin of the nonlinear absorption shown in Figure 6.
Notice that multistep two-photon excitation due to absorption of excited states is avoided
since the employed infrared excitation (800 nm) is far from the one-photon resonance (515
nm) and the tail of lineal absorption (see Figure 3). We also point out that despite the high
intensities no artifacts on the Z-scan traces were detected, i.e. thermo-optical effects, which
usually introduce strong distortion. The Z-traces were repeated several times and the mean

value for each pulse energy is displayed in the inset of Figure 6. According to these results
the coefficient showed a negligible dependence on the energy of excitation for the range of
energies employed in the Z-scan experiments. Thus, the average for [2](PF and the
corresponding TPA from a total of 25 plots resulted in the values shown in Table 8. This
Table also shows the and TPA values for the fluorene-based parent compound [1](PF
The TPA value obtained for [2](PF is equal to 159.2 GM which indicates an
increase of about 1.5 times that of the fluorene-based parent compound [1](PF . This
enhancement in the TPA response is consistent with the red shifting of 64 nm of the low-lying
6 3
)
6
) .
3
)
6 3
)
6 3
3+
3+
absorption band after the substitution of fluorene in [1] by carbazole in [2] , pointing out
the potential effect of the donating capabilities of the tertiary amine present at the carbazole
moiety.
1
7
This article is protected by copyright. All rights reserved.

European Journal of Inorganic Chemistry
10.1002/ejic.201700895
Figure 6
Table 8
6 3
Normalized transmission in Z-scan experiments for [2](PF ) with different energies of
excitation at the wavelength of 800 nm. Inset: average values of the nonlinear absorption
coefficient  as a function of energies.
6 3
Nonlinear absorption coefficients (), and TPA cross-sections (TPA), for [1](PF ) and
[
6 3
2](PF ) at incident laser wavelength of 800 nm.
-
11
Compound 10 cm/W
TPA (GM)
ref
_
_______________________________________________________
[
[
2](PF
6
6
)
3
3
3.85  0.55
2.63  0.43
159.2  22 this work
108.0  18 [Ref 13a]
1](PF
)
_
_______________________________________________________
NO release in [1](PF
6
)
3
and [2](PF
6 3
)
As previously reported in several Ruthenium nitrosyl complexes with fluorene
terpyridine ligand,[13] the resulting release of nitric oxide is followed by the formation of a
solvent bound ruthenium(III) photoproduct, according to the following equation:
ℎ휈
II
+
III
(
L)Ru -NO + solvent →ꢁ (L)Ru -solvent + NO·
(1)
1
8
This article is protected by copyright. All rights reserved.

European Journal of Inorganic Chemistry
10.1002/ejic.201700895
The efficient photo-induced NO· release in acetonitrile solution by [2](PF
6
3
) complex
under standard (one-photon) excitation at 436 nm can be farther demonstrated by using EPR
spectroscopy since spin trapping combined with EPR spectroscopy is considered as one of the
best methods for the direct detection of NO· radicals.[35] Namely, we used Iron(II)-N-methyl-
II
D-glucamine dithiocarbamato [Fe (MGD)
2
] to trap NO· due to its high probability of adduct
formation and to the high stability of its spin adduct. Figure 7 shows the characteristic triplet
-3
-1
signal with a hyperfine splitting constant value of a
N
= 1.2 10 cm and a g-factor of g =
II
-NO] adduct.[36] The weak
2
.039. This is consistent with the literature report for [Fe (MGD)
2
signal observed in the control spectrum (top of Figure 7) is due to a trace of NO∙ and related
to the fact that the manipulation is never strictly conducted in the dark.
30000
20000
10000
0
-
-
-
10000
20000
30000
3
0000
0000
0000
0
2
1
-
-
-
10000
20000
30000
3
380
3430
3480
Magnetic field (G)
Figure 7
2 6 3
Triplet EPR signals from NO· trapped by [Fe(MGD) ] for [2](PF ) upon one photon
excitation at room temperature and at  > 400 nm (Hg lamp) (bottom), and control signal
before irradiation (top).
The changes in the electronic absorption spectrum of [2](PF
6
3
) exposed to 436 nm
light in acetonitrile are shown in Figure 8. The presence of isosbestic points at 338, 406, 426
II
+
and 496 nm indicates a clean conversion of the Ru (NO ) complexes to related photolysed
species. No back-reaction is observed when the light is turned off. In the photolysed species,
new bands located at 320, 370 and 472 nm arise. It is interesting to observe that the shape of
II
the final product (red line in Figure 8) corresponds to a Ru complex (lack of band at 600 nm
1
9
This article is protected by copyright. All rights reserved.

European Journal of Inorganic Chemistry
10.1002/ejic.201700895
characteristic of RuIII species).[13c] This suggests that the photoproduct may be reduced
II
immediately after the photoreaction, leading to a Ru complex as the final observable
derivative. Nevertheless, the isosbestic points evidenced in Figure 8 indicate that this second
step process is probably extremely fast. Similar features are observed in [1](PF
6
3
) (Figure S2).
The observation of final Ru species instead of the expected RuIII species appears to be a
rather general feature in ruthenium nitrosyl complexes containing five -acceptor pyridines in
the metal coordination sphere.[13a]
II
Before irradiation
5
4
3
2
1
0000
0000
0000
0000
0000
0
After 20h irradiation
3
00
400
500
600
700
λ nm
Figure 8
Evolution in the absorption spectra of [2](PF
36 nm.
6 3
) in acetonitrile under irradiation at =
4
The quantum yield of NO· release (NO) observed for [2](PF
irradiation is equal to 0.01. This value appears significantly reduced compared to the 0.03
value measured for [1](PF . The electronic transitions in the carbazole complex turn out to
6
)
3
under 436 nm light
)
6 3
be less efficient in NO· release (lower NO) than those of the fluorene-terpyridine complex,
although they are associated to a more sizable ligand → Ru(NO) charge transfer. This
observation raises the issue of the mechanism involved in the NO· release process and that of
the NO vs TPA optimization. This will be addressed in the last section.
Additional considerations on NO and TPA
In the search for efficient NO· donors inducted by two-photon excitation, it is
important to keep in mind that the figure of merit is not the quantum yield of photo-release
(
NO) or the molecular cross section (TPA), but rather the NO  TPA product, so called
“uncaging cross-section” of the photo-induced process. Therefore, the search for the best
2
0
This article is protected by copyright. All rights reserved.

European Journal of Inorganic Chemistry
10.1002/ejic.201700895
candidates requires both properties to be considered simultaneously. So far, very few
estimations of photorelease (NO) and TPA (TPA) capabilities are available in
Ru(terpy)(bipy)(NO) complexes (Table 9). It is important to keep in mind that NO is
estimated at one-photon, while TPA refers to the two-photon process. Nevertheless, NO is the
number of NO∙ generated divided by the number of excited states promoted. Clearly, the
efficiency of the photoreaction [Ru(NO)]* → [Ru] + (NO) does not imply to consider what
route took place towards the excited state ([Ru(NO)]*. Along this line, a one-photon and a
two-photon excitations lead to the same quantum yield.
A first tendency seems to appear in Table 9, which suggests that enhancing the TPA
values would lead to lowering NO, along this series of complexes, in contrast to our initial
intuition that the efficiency of the NO· release could reflect the strength of the charge transfer
behavior towards the nitrosyl.
Table 9
Parameters of interest for the two-photon induced release of NO· (TPA in GM and

NO
)
in
ruthenium-terpyridine-nitrosyl complexes. The values in parentheses are the wavelength of the
incident irradiation (in nm).
Compounds

NO

TPA
NO  TPA
_
_______________________________________________________________
a
a
[
1](PF
6
)
3
0.06 (405) 108.0 (800)
.03 (436)
0.01 (436)
6.5
3.2
1.5
0
[
2](PF
6
)
3
159.2 (800)
_
_______________________________________________________________
a
ref 13a
In fact, the mechanism of the NO· release process is still controversial. Indeed, several
II
+
reviews have discussed the dissociative mechanism of M NO as resulting from photo-
+
induced intramolecular electron transfers to NO and concomitant oxidation of the metal
III
III
[37,6b]
center to M . Then, the resulting M NO· species undergoes a “facile” NO· release.
Therefore, the presence of an electron-rich ligand should encourage this process.
Nevertheless, the final breaking of the Ru-NO bond remains unexplained within this picture,
except by emphasizing the strong antibonding character of the d-* overlap between the
ruthenium and nitrogen atoms, which leads to a reduction of the bond order. On the other
II
+
hand, we have recently observed that, in a series of [Ru (R-terpy)(Cl)
2
(NO)] complexes,
2
1
This article is protected by copyright. All rights reserved.

European Journal of Inorganic Chemistry
10.1002/ejic.201700895
there is no clear correlation between the donating/withdrawing character of R and the NO
values.[21] Furthermore, a recent theoretical investigation has suggested that the NO· release
could require more than one photon, with the contribution of an intermediate triplet state.[38]
Altogether, these features tend to suggest that enhancing the donating capabilities of the
ligands may not be the only fruitful strategy in this approach towards efficient NO· donors,
which could also encompass molecules of greater complexity (e.g. polymetallic Ru(NO)
complexes), than those envisioned in the initial stage of these investigations. We are currently
exploring these possibilities.
Conclusion
In a research effort aimed at designing ruthenium nitrosyl complexes with two-photon
induced NO· release capabilities, we have observed that replacing a fluorenyl unit by a
carbazole on a terpyridine ligand leads to a sizeable enhancement of the intramolecular push-
pull properties from the substituted terpyridine to the withdrawing nitrosyl ligand, which can
be evidenced by spectroscopy and electrochemistry. This electronic features lead to the
outcomes of higher TPA cross-section (TPA) and lower quantum yield (NO). The first
experimental data available indicate that the two-photon encaging cross-section of the overall
two-photon induced NO· release (NO  TPA) is surprisingly less favorable when the donating
capabilities of the ligands are more pronounced, within this series of
II
3+
[
Ru (terpy)(bipy)(NO)] complexes. Finding a definitive rationale to this issue is not
straightforward. Electronic and structural features are likely to play a role. Further
investigations are currently under progress.
Experimental section
Material and Equipment
2
-Acetylpyridine (Alfa-Aesar), carbazole (Sigma-Aldrich), iodoethane (Sigma-
Aldrich), 2,2'-bipyridine (TCI), triethylamine (Sigma-Aldrich), N-bromosuccinimide (Sigma-
Aldrich), lithium chloride (ACROS), sodium nitrite (Alfa-Aesar), lithium chloride (ACROS),
NH
4
PF
6
(Alfa-Aesar) and RuCl
3
.3H O (STREM) nBuLi 1.6 M (hexanes) (Sigma-Aldrich)
2
2
2
This article is protected by copyright. All rights reserved.

European Journal of Inorganic Chemistry
10.1002/ejic.201700895
were used without further purification. The solvents were analytical grade and used without
further purification. Elemental analyses were performed at LCC with a Perkin Elmer 2400
1
serie II Instrument. H NMR spectra were obtained at 298K in CDCl
3
, CD CN as internal
3
reference and were recorded on a Bruker Avance 300 or a Bruker Avance 400. Infrared
spectra were recorded on a Perkin Elmer 1725 Spectrometer.
Synthesis
3
-Bromocarbazole (A). Carbazole (2 g, 11.96 mMol) was dissolved in DMF (25 mL),
cooled to 0° C and covered with aluminum to protect it from the light. NBS (2.128 g, 11.96
mMol) was dissolved in DMF (25 mL) and was added dropwise to the solution containing the
carbazole, and the resulting mixture was allowed to reach room temperature. The reaction was
left 3 hrs and then was poured into 100 mL of a mixture of ice-water. The solid was filtrated
with vacuum, then it was recrystallized from ethanol resulting in 2.20 g (8.97 mMol) 75% of a
grey-pink solid (m.p. 199-200 °C).
3
-Bromo-9-ethylcarbazole (B). Compound A (3 g, 12.19 mMol) and tetrabutyl
ammonium iodide (300 mg) were put in a flask, 25 mL of a solution of 50% sodium
hydroxide in water were added, followed by ethyl iodide (2.92 mL, 36.5 mMol). The mixture
was stirred vigorously at 30 °C for 24 hrs under nitrogen atmosphere. The aqueous phase was
extracted three times with ethyl acetate (30 mL) and then dried with sodium sulfate and
evaporated at reduced pressure. The solid obtained was then recrystallized from ethanol
1
leading to 2.58 g (9.41 mMol) 77% of a white solid (m.p. 79-80 °C). H NMR (300 MHz,
CDCl
3
)  8.25 (dd, J = 2.0, 0.6 Hz, 1H), 8.10 (ddd, J = 7.8, 1.2, 0.6 Hz, 1H), 7.59 (dd, J = 8.7,
.0 Hz, 1H), 7.53 (dd, J = 7.0, 1.2 Hz, 1H), 7.46 (dt, J = 8.3, 1.2 Hz, 1H), 7.38 – 7.28 (m, 1H),
.33 – 7.23 (m, 1H), 4.40 (q, J = 7.2 Hz, 2H), 1.47 (t, J = 7.2 Hz, 3H).
2
7
9
-Ethylcarbazol-3-carboxaldehyde (C). Compound B (2 g, 7.29 mMol) was placed in
a flask, purged with nitrogen, then diluted with dry THF (20 mL) and cooled down to -74° C.
nBuLi 11 M (1.3 mL, 14.6 mMol) was added dropwise resulting in a yellow solution, which
was kept at -74° C for 1 hr and then dry DMF (2.8 mL, 36.5 mMol) was added in one step.
The reaction mixture was kept at -74° C for 1 hr, then HCl(conc.) (3 mL, ~36 mMol) was added
in one step and the reaction was allowed to reach room temperature and left for 1 hr at room
temperature. The reaction was diluted with water (20 mL) and extracted three times with ethyl
2
3
This article is protected by copyright. All rights reserved.

European Journal of Inorganic Chemistry
10.1002/ejic.201700895
acetate (30 mL). The organic phase was dried with sodium sulfate and then evaporated under
reduced pressure. It was purified using silicagel column with hexane-ethyl acetate (9:1) as
eluent, yielding 1.23 g (5.54 mMol) 76% of a transparent oil. 1H NMR (300 MHz, CDCl
) 
3
1
8
8
0.14 (s, 1H), 8.66 (dd, J = 1.6, 0.6 Hz, 1H), 8.21 (ddd, J = 8.2, 1.2, 0.8 Hz, 1H), 8.06 (dd, J =
.6, 1.6 Hz, 1H), 7.59 (ddd, J = 8.2, 7.0, 1.2 Hz, 1H), 7.53 (d, J = 8.6 Hz, 1H), 7.51 (dt, J =
.2, 1.2 Hz, 1H), 7.38 (ddd, J = 8.2, 7.0, 1.2 Hz, 1H), 4.46 (q, J = 7.2 Hz, 2H), 1.52 (t, J = 7.2
Hz, 3H).
(E)-3-(9-ethyl-9H-carbazol-3-yl)-1-(pyridin-2-yl)prop-2-en-1-one (D). Compound C
(2.56 g, 11.46 mMol) was suspended in ethanol 96% (100 mL) and then 2-acetylpyridine (1.3
mL, 11.46 mMol) was added. A solution containing sodium hydroxide (900 mg, 22.5 mMol)
in 25 mL of water was added dropwise to the reaction mixture. It was stirred overnight at
room temperature which led to the appearance of a yellow solid. It was filtrated under vacuum
and then it washed with water and cold ethanol, yielding to 3.4 g (10.4 mMol) 90% of the
1
desired product. H NMR (400 MHz, CDCl
3
)  8.78 (ddd, J = 4.8, 1.8, 0.9 Hz, 1H), 8.46 (d, J
=
1.8 Hz, 1H), 8.35 (d, J = 15.9 Hz, 1H), 8.27 – 8.23 (m, 1H), 8.20 (d, J = 15.9 Hz, 1H), 8.14
(d, J = 7.7 Hz, 1H), 7.89 – 7.83 (m, 2H), 7.53 – 7.44 (m, 2H), 7.39 (t, J = 8.3 Hz, 2H), 7.33 –
7
.24 (m, 1H), 4.33 (q, J = 7.2 Hz, 2H), 1.43 (t, J = 7.2 Hz, 3H).
4
’-(N-ethyl-carbazol-3-yl)-2,2’:6’,2’’-terpyridine (E). A mixture of compound D (1 g,
3
mMol), 1-(2-oxo-2-pyridin-2-yl-ethyl)pyridinium iodide (1.0 g, 3 mMol), and ammonium
acetate (2.36 g, 30 mMol) was suspended in ethanol (50 mL), and refluxed for 20 hrs. After
cooling down to room temperature, the resulting green solid was filtrated under vacuum. It
was purified by column chromatography using silica and a mixture of hexane : acetone (8:2)
1
as an eluent yielding 392 mg (9.19 mMol) 30% of the resulting ligand as a yellow solid. H
NMR (400 MHz, CDCl
H), 8.67 (d, J = 1.6 Hz, 1H), 8.23 (d, J = 7.7 Hz, 1H), 8.05 (dd, J = 8.5, 1.8 Hz, 1H), 7.88
td, J = 7.7, 1.8 Hz, 2H), 7.53 – 7.49 (m, 1H), 7.49 – 7.47 (m, 1H), 7.43 (d, J = 8.1 Hz, 1H),
3
)  8.87 (s, 2H), 8.77 (dd, J = 4.7, 0.8 Hz, 2H), 8.70 (d, J = 8.0 Hz,
2
(
7
7
.35 (ddd, J = 7.4, 4.8, 1.0 Hz, 2H), 7.32 – 7.26 (m, 1H), 4.40 (q, J = 7.2 Hz, 2H), 1.47 (t, J =
.2 Hz, 3H).
III
[
Ru Cl
3
(E)] (F). The ligand E (150 mg, 0.35 mMol) was suspended with RuCl x
3
H
2
O (92 mg, 0.35 mMol) in ethanol (25 mL). The reaction was protected from the light, and
2
4
This article is protected by copyright. All rights reserved.

European Journal of Inorganic Chemistry
10.1002/ejic.201700895
refluxed for 3 hrs; it was left to reach room temperature and then cooled with ice. The
resulting solid was filtrated under vacuum and rinsed with ethanol and diethyl ether, yielding
2
12 mg (0.33 mMol, 95%) of a dark brown powder.
II
[
Ru (bipy)(Cl)(E)](Cl) (G). A mixture of complex F (100 mg, 0.158 mMol), 2,2’-
bipyridine (2.25 mg, 0.158 mMol), LiCl (40 mg, 0.946 mMol) and trimethylamine (0.1 mL)
was added to a flask and suspended in ethanol 75% (20 mL). The reaction was protected from
the light, and refluxed for 4 hrs. Then, it was left to reach room temperature and then cooled
with ice. The resulting solid was filtrated under vacuum and rinsed with water, ethanol and
1
diethyl ether, yielding 100 mg (0.134 mMol, 85%) of a purple solid. H NMR (300 MHz,
CD
3
CN)  10.28 (d, J = 5.6 Hz, 1H), 9.20 (s, 1H), 9.03 (s, 2H), 8.67 – 8.53 (m, 3H), 8.42 (d, J
=
7.8 Hz, 1H), 8.37 – 8.24 (m, 3H), 7.98 (ddd, J = 7.4, 5.7, 1.3 Hz, 1H), 7.88 – 7.76 (m, 3H),
7
7
7
.74 – 7.62 (m, 4H), 7.60 – 7.53 (m, 1H), 7.45 (d, J = 5.8 Hz, 1H), 7.32 (t, J = 7.3 Hz, 1H),
.28 – 7.17 (m, 1H), 6.99 (ddd, J = 7.4, 5.8, 1.3 Hz, 2H), 4.55 (q, J = 7.2 Hz, 2H), 1.50 (t, J =
.2 Hz, 3H).
II
[
Ru (bipy)(E)(NO
2
)](PF ) (H). A mixture of complex G (155 mg, 0.205 mMol) and
6
NaNO (142 mg, 2.05 mMol) was suspended in ethanol 75% (20 mL). The reaction was
2
protected from the light, and refluxed for 4 hrs. Then, it was left to reach room temperature
and then cooled with ice. NH PF (100 mg, 0.61 mMol) dissolved in 3 mL of water was
4
6
added to the reaction, after which a red precipitate appeared. It was filtrated under vacuum
and then rinsed with water, ethanol and diethyl ether, yielding 122 mg (0.679 mMol, 70%) of
-
1
-1
1
the desired complex H, as a red solid. FTIR-ATR: 1331 cm and 1295 cm (-NO
NMR (400 MHz, CD CN)  9.95 (dd, J = 5.5, 1.3 Hz, 1H), 8.95 (d, J = 1.9 Hz, 1H), 8.84 (s,
H), 8.61 (d, J = 8.2 Hz, 1H), 8.49 (d, J = 7.9 Hz, 2H), 8.38 (d, J = 8.0 Hz, 1H), 8.30 (dd, J =
2
). H
3
2
7
1
.8, 1.4 Hz, 2H), 8.27 – 8.21 (m, 1H), 7.99 (ddd, J = 7.4, 5.6, 1.3 Hz, 1H), 7.92 (td, J = 7.9,
.6 Hz, 2H), 7.82 (d, J = 8.7 Hz, 2H), 7.78 (d, J = 4.9 Hz, 2H), 7.68 (d, J = 8.3 Hz, 1H), 7.61
(ddd, J = 8.3, 7.0, 1.2 Hz, 1H), 7.47 (d, J = 5.7 Hz, 1H), 7.36 (ddd, J = 8.0, 7.0, 1.0 Hz, 1H),
7
2
.31 (ddd, J = 7.7, 5.5, 1.3 Hz, 2H), 7.08 (ddd, J = 7.3, 5.6, 1.2 Hz, 1H), 4.58 (q, J = 7.2 Hz,
H), 1.52 (t, J = 7.2 Hz, 3H).
II
[
Ru (bipy)(E)(NO)](PF
6
)
3
([2’](PF
6
3
) ). A mixture of ethanol (15 mL) and HCl(conc.) (4
mL, ~50 mMol) was added to complex H (34 mg, 38.86 Mol). The reaction was protected
2
5
This article is protected by copyright. All rights reserved.

European Journal of Inorganic Chemistry
10.1002/ejic.201700895
from the light, and heated at 60° C for 2 hrs. After cooling down with ice, NH
4
PF (100 mg,
6
0
.61 mMol)) dissolved in water (3 mL) was added to the reaction mixture, which led to the
appearance of a purple precipitate. It was filtrated under vacuum and then rinsed with water,
ethanol and diethyl ether, yielding 31 mg (26.9 Mol, 70%) of a purple solid. FTIR-ATR:.
-1
-1
-1
1
1
937 cm (-NO). UV-Vis: max = 517 nm ( = 14 600 L.mol .cm ). H NMR (400 MHz,
CD CN)  9.35 (d, J = 5.5 Hz, 1H), 9.23 (s, 2H), 9.21 (d, J = 2.0 Hz, 1H), 8.95 (dd, J = 8.2,
3
0
8
8
.9 Hz, 2H), 8.84 (d, J = 8.2 Hz, 1H), 8.73 (td, J = 8.0, 1.4 Hz, 1H), 8.65 (d, J = 8.3 Hz, 1H),
.53 (td, J = 7.9, 1.5 Hz, 2H), 8.45 (dd, J = 8.8, 2.1 Hz, 1H), 8.42 (d, J = 8.2 Hz, 1H), 8.35 –
.26 (m, 2H), 8.07 – 8.03 (m, 2H), 7.93 (d, J = 8.8 Hz, 1H), 7.77 – 7.72 (m, 3H), 7.70 – 7.65
(
m, 1H), 7.54 (ddd, J = 7.5, 6.0, 1.3 Hz, 1H), 7.49 – 7.44 (m, 1H), 7.43 – 7.41 (m, 1H), 4.60
q, J = 7.2 Hz, 2H), 1.53 (t, J = 7.2 Hz, 3H). Anal. Calcd. for C39 OP Ru: C 40.78, H
.63, N 8.54 found: C 40.17, H 2.45, N 8.25.
(
H
30
F
18
N
7
3
2
Crystallographic data
Data were collected at low temperature (100(2) K) on a Bruker Kappa Apex II diffractometer
equipped with a 30 W air-cooled microfocus, using MoKα radiation (λ = 0.71073 Å), and an
Oxford Cryosystems Cryostream cooler device. Phi- and omega- scans were used for data
collection. The structure was solved by intrinsic phasing method (SHELXT)[39]
All non-
.
hydrogen atoms were refined anisotropically by means of least-squares procedures on F² with
the aid of the program SHELXL[40]. All the hydrogen atoms were refined isotropically at
calculated positions using a riding model. The crystal structure with only NO
2
coordinated to
the Ru atom was not fully acceptable. Anisotropic displacements parameters and the residual
electron density suggested a disorder between the NO
after refinement.
2
group and a Cl atom in a ratio 90:10
CCDC 1564051 (for H) contains the supplementary crystallographic data for this paper.
These data can be obtained free of charge from the Cambridge Crystallographic data center.
Computational studies
3
+
3+
3+
3+
The ruthenium complexes [1] , [2] , [1’] , and [2’] were optimized in gas phase
using the Gaussian-09 program package[41] within the framework of the Density Functional
2
6
This article is protected by copyright. All rights reserved.

European Journal of Inorganic Chemistry
10.1002/ejic.201700895
Theory (DFT). The double- basis set 6-31G* was used for all atoms except the heavy
ruthenium atom, for which the LANL2DZ basis set was applied to account for relativistic
effects.[42] The two hexyl chains present in [1] , and [1’]3+ were replaced by two methyl
groups to help the convergence process. To be consistent with our previous report,[13c] and in
agreement with a previous investigation of ruthenium-nitrosyl by Mascharak,[43] we have
selected the hybrid functional B3PW91 for the optimization. B3PW91 has been shown to
outperform other hybrid functionals (e.g. B3LYP) and pure functionals (e.g. PW91) in
numerous cases of ruthenium complexes, especially when back bonding ligands (like NO) are
present.[43,44] The vibrational analyses were performed at the same level to verify that the
stationary points correspond to minima on the potential energy surfaces. The UV-visible
electronic spectra were then computed at the CAM-B3LYP/6-31G* level, which was selected
for its good efficiency to reproduce experimental transition energies. Solvent effects were
included by using the polarizable continuum model (PCM) implemented in Gaussian09 for
3+
[
45]
acetonitrile ( = 35.688). Molecular orbitals were plotted with GABEDIT 2.4.8.
The
optimized geometries are provided in supporting information.
Electrochemistry
Electrochemical experiments were performed at room temperature in a homemade
airtight three-electrode cell connected to a vacuum/argon line. The reference electrode was a
saturated calomel electrode (SCE) separated from the solution by a bridge compartment. The
2
counter electrode was a platinum wire of ca. 1 cm apparent surface. The working electrode
was a Pt microdisk (radius = 0.25 mm). The supporting electrolyte (nBu
4
N)(PF ) (Fluka, 99%
6
electrochemical grade) was used as received and simply degassed under argon. Acetonitrile
was freshly purified prior to use. The solutions used during the electrochemical studies were
-3
-1
-1
-1
typically 10 mol.L and 10 mol.L in supporting electrolyte. Before each measurement, the
solutions were degassed by bubbling argon through them, and the working electrode was
polished with a polishing machine (Presi P230).
Z-scan measurements
The Z-scan technique[46] was used to measure the nonlinear absorption coefficient of the
samples at 800 nm using short laser pulses of 350 fs at 1 KHz of repetition rate. Molecules
under study were dissolved in acetonitrile at the concentration of 1×10-2 mol.L . Z-scan
-1
2
7
This article is protected by copyright. All rights reserved.

European Journal of Inorganic Chemistry
10.1002/ejic.201700895
traces for each solution were measured at different energies (180, 250, 360 and 460 nJ per
pulse) The sample were measured at least four times for each energy. To verify the validity of
our measurements, the same Z-scan apparatus was first used to measure in the close aperture
approach the nonlinear refractive index n
2
of the standard CS
2
. The results led to n values in
2
the range 1 - 2 ×10-15 cm /W, which is in very good agreement with the values accepted for
2
this standard reference. Then, the laser dye rhodamine B (RB) (dissolved in methanol at the
-2
-1
concentration of 1×10 mol.L ) was measured for comparison of TPA, along with the sample
at test in the open aperture approach. In this case, the value of TPA of RB was found equal to
1
0412 (for a pulse of 350 fs) also in very good agreement to the value accepted in the
literature (reference value 120 GM).[47] The nonlinear absorption coefficient  of each sample
was obtained after fitting the normalized transmission T(z) to Z-scan formalism. A total of 25
plots of Z-scans were analyzed. The TPA cross-section (TPA) is obtained from the following
expression:
ℏ
푁
휎푇푃퐴
=

Were N is the molecular density and  is the optical frequency.
Photochemistry
Kinetic studies on the photolysis reactions were carried out with a diode array Hewlett
-5
-1
Packart 8454A spectrophotometer. Solutions of 3 mL of [1](PF
6
)
3
(5.35.10 mol.L ) and
-5
-1
[
6 3
2](PF )
(7.35.10 mol.L ) in non-deoxygenated acetonitrile were used. The optical fiber
was fixed laterally from the cuvette. Absorption spectra were taken after each minute, in fast
scan mode, during a period of 7 hours for both complexes, which allows reaching apparent
stable absorption conditions. The UV-visible spectra were recorded under irradiation realized
with a Muller reactor device equipped with a cooling water filter and a mercury arc lamp
equipped with appropriate interference filter to isolate the desired irradiation wavelength
(max = 436 nm, intensity 8 mW). The light intensity was determined by using ferrioxalate
actinometer. The sample solutions were placed in a quartz cuvette of 1 cm path -length stirred
continuously. The temperature was maintained at 27°C during the whole experiment.
Quantum yield measurements: Light intensities were determined before each photolysis
experiments by chemical actinometry procedure. The actinometers used were potassium
2
8
This article is protected by copyright. All rights reserved.

European Journal of Inorganic Chemistry
10.1002/ejic.201700895
-6
-1 -1
-7
-1 -1
ferrioxalate to irr=436 nm (I
0
= 1.01.10 mol.L .s for [1](PF
) was determined by the program Sa3.3 written by D.
Lavabre and V. Pimienta.[50] It allows the resolution of the differential equation d d[ At ]
6
)
3
and I
0
= 8.13.10 mol.L .s
for [2](PF
6
)
3
). The quantum yield (

=
A
λ
A
−
Φ I = −Φ ꢂbs I F where I is the intensity of the light absorbed by the precursor; F,
퐴 a 퐴 A 0 a
ꢋ
ꢄ1ꢅ10ꢆꢇꢈꢉToꢊꢌ
ꢏ ; ꢂbs , the absorbance of [1] or [2]3+ before
λ
3+
the photokinetic factor ꢃF =
ꢋ
Aꢍꢎ
Toꢊ
A
λ
irradiation; ꢂbs , the total absorbance; I
0
, the incident intensity measured at 436 nm. The
ꢐꢑt
λ
equation was fitted with the experimental data ꢂbsꢐꢑt = 푓(푡) and 2 parameters 

and 

(

is the molar extinction coefficient measured at the end of the reaction) at two wavelengths
(
irr=436 nm, obs=470 nm for [1]3+ and 550 nm for [2] ).obs was chosen because it
3+
corresponds to a large difference between molar extinction coefficient at the initial and final
time of the photochemical reaction. Simulation and optimization procedures were performed
by using numerical integration and a non-linear minimization algorithm for the fitting of the
model to the experimental data.[48,49]
[
1]3+
ꢅꢔ
ꢅ1 4ꢖ6
ꢅ1
ꢅ1 470
ꢅ1
ꢅ1
[
ꢂ] = 5.35 ꢒꢓ mꢕl. L ,ε
= ꢒ55ꢒꢗ mꢕl . L. cm ,ε
퐴
= ꢒ5ꢘꢓꢗ mꢕl . L. cm
,
,
0
퐴
4
ꢖ6
ꢅ1
ꢅ1
470
ꢅ1
ꢅ1
ε
= ꢒꢗ3ꢗ9 mꢕl . L. cm , ε퐵 = ꢘꢘꢓꢒ9 mꢕl . L. cm
퐵
[
2]3+
ꢅꢔ
ꢅ1 4ꢖ6
ꢅ1
ꢅ1 ꢔꢔ0
ꢅ1
ꢅ1
[
ꢂ] = ꢙ.35 ꢒꢓ mꢕl. L ,ε
= ꢒꢒꢘ9ꢘ mꢕl . L. cm ,ε
퐴
= ꢒꢒ93ꢒ mꢕl . L. cm
0
퐴
4
ꢖ6
ꢅ1
ꢅ1
ꢔꢔ0
ꢅ1
ꢅ1
ε
= ꢒꢘꢒꢓꢚ mꢕl . L. cm , ε퐵 = 8ꢒ8ꢙ mꢕl . L. cm
퐵
Supporting Information: Crystal data, DFT geometries, voltamograms for [1](PF
6
3
) and
[
2](PF (S1), and changes in absorption spectra for [1](PF (S2).
6
)
3
6 3
)
Acknowledgements
The work has been performed within the framework of the French-Mexican international
laboratory (LIA-LCMMC). The authors thank CNRS (France) and CONACyT (México) for
financial supports, Marine Tassé for recording the photo release properties, and Dr. Alix
2
9
This article is protected by copyright. All rights reserved.