Spring-loaded Iron(II) complexes as magnetogenic probes reporting on a chemical analyte in water

Article abstract of DOI:10.1002/ejic.201403183

A paramagnetic quality may be established in an aqueous sample if it contains a molecular probe that responds to the addition of a chemical analyte by turning from a diamagnetic state into a paramagnetic one (off-on). We explore here a stable, low-spin, binary iron(II) complex that stores so much potential energy that its transformation by the target analyte leads to its fragmentation into a high-spin complex despite the imposing strength of the chelate effect. The underlying ligand is a mixed aminal, the components of which can be freely varied to optimize response kinetics with the initial probe stability preserved. With decreasing leaving-group character of the azole component, the probe stability improves, and the response kinetics diminish. An optimal arrangement can be found with a pyrazole paired with an electron-deficient aromatic carbaldehyde component (nitro substitution). The drastic electronic reversion associated with the reduction of the nitro group to an amino group is the principal reason for the observation of an initially stable probe that suffers swift fragmentation when reacted with a reductive analyte. A new low-spin iron(II) triazacyclononane-based complex, selected from four different candidates, acts as a molecular probe that turns a diamagnetic aqueous sample into a paramagnetic one as a response to a chemical analyte. The four complexes are prepared by a highly convergent synthetic protocol and give precious insights into structure-activity relationship (SAR) tendencies.

Full text of DOI:10.1002/ejic.201403183

Job/Unit: I43183
/KAP1
Date: 10-02-15 14:33:15
Pages: 8
FULL PAPER
DOI:10.1002/ejic.201403183
Spring-Loaded Iron(II) Complexes as Magnetogenic
Probes Reporting on a Chemical Analyte in Water
Corentin Gondrand,[a][‡] Fayçal Touti,[a][‡] Estelle Godart,[a]
Yevgen Berezhanskyy,[a] Erwann Jeanneau,[b] Philippe Maurin,[a] and
Jens Hasserodt*[a]
Keywords: Cage compounds / Iron / Magnetic properties / Macrocyclic ligands / N ligands
A paramagnetic quality may be established in an aqueous
sample if it contains a molecular probe that responds to the
addition of a chemical analyte by turning from a diamagnetic
state into a paramagnetic one (off–on). We explore here a
stable, low-spin, binary iron(II) complex that stores so much
potential energy that its transformation by the target analyte
leads to its fragmentation into a high-spin complex despite
the imposing strength of the chelate effect. The underlying
ligand is a mixed aminal, the components of which can be
freely varied to optimize response kinetics with the initial
probe stability preserved. With decreasing leaving-group
character of the azole component, the probe stability im-
proves, and the response kinetics diminish. An optimal ar-
rangement can be found with a pyrazole paired with an elec-
tron-deficient aromatic carbaldehyde component (nitro sub-
stitution). The drastic electronic reversion associated with the
reduction of the nitro group to an amino group is the princi-
pal reason for the observation of an initially stable probe that
suffers swift fragmentation when reacted with a reductive
analyte.
possible for several sample environments, including bio-
logical ones, that do not contain any other paramagnetic
components. As eligible detection techniques, electron para-
magnetic (EPR) spectroscopy may be cited as a direct
method, and NMR spectroscopy may be cited as an indi-
rect one (see Figure 4, b and c).
Introduction
Proving the presence of a particular chemical analyte in
a sample is the raison d’être for the discipline of analytical
chemistry but also plays an essential role in the rapidly ex-
panding field of chemical imaging.[1] Today, several physical
detection modes serve this purpose, including electromag-
netic waves (absorbance, fluorescence, phosphorescence,
and interferometry), radioactivity [autoradiography, single-
photon emission computed tomography (SPECT)], or elec-
tric currents (conductimetry). Although it has been rela-
tively neglected, a magnetic mode of detection offers several
advantages over classical modes. For example, the signal is
not constantly emitted, as is the case for radioactive com-
pounds, but is emitted only when an external magnetic field
is applied. Also, the signal and the molecule from which
it originates do not experience any fatigue, as is found for
fluorescent or radioactive molecules. The signal emitted by
the magnetic molecule is not attenuated during its passage
through the sample, and finally, highly specific detection is
The design of molecular probes that actively respond to
a chemical stimulus by changing the magnitude of their
emitted signal is attractive because they offer better signal-
to-background ratios and, thus, improved sensitivity. To the
best of our knowledge, the examples of coordination com-
pounds that respond to a change in their chemical composi-
tion by changing their magnetic properties do so only
through gradual, reversible reactions.[2–4] We introduced a
new probe concept that is based on low-spin iron(II) binary
complexes[5–7] that undergo an irreversible reaction with the
target analyte for high analyte specificity, optimal magni-
tude of signal change, and best adaptation to aqueous me-
dia. Our choice of the particular constitution of the com-
plex has been motivated by the highly competitive nature
of water as the solvent, the need to translate a chemical
event into a readable signal, and the requirement to over-
come the imposing strength of the chelate effect. We re-
cently communicated a molecular probe that irreversibly
establishes a paramagnetic quality (“on”) in an initially dia-
magnetic aqueous sample (“off”) upon the addition of a
chemical analyte.[8] It did so through fragmentation into
three components once transformed by the target analyte.
The fragmentation implied the cleavage of a five-membered
chelate ring. This required the identification of a molecular
[a] Laboratoire de Chimie, UMR CNRS UCBL 5182,
Université de Lyon–ENS de Lyon,
46, allée d’Italie, 69364 Lyon CEDEX 07, France
E-mail: jens.hasserodt@ens-lyon.fr
http://www.ens-lyon.fr/CHIMIE/
[b] Centre de Diffractométrie Henri Longchambon, Université de
Lyon–UCBL,
Site CLEA, Bâtiment ISA, 3ème étage
5, rue de la Doua, 69100 Villeurbanne, France
[‡] These authors contributed equally.
Supporting information for this article is available on the
WWW under http://dx.doi.org/10.1002/ejic.201403183.
Eur. J. Inorg. Chem. 0000, 0–0
1
© 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Job/Unit: I43183
/KAP1
Date: 10-02-15 14:33:15
Pages: 8
www.eurjic.org
FULL PAPER
moiety that stores much potential energy (spring-loading) cyclic base triazacyclononane (tacn) bearing two picolyl
without compromising the initial probe stability. Once pendent arms (dptacn).[3,7] The condensation reaction is
chemically transformed by the target analyte, the resulting convergent and does not lead to any other aminals aside
moiety should lead to swift fragmentation. Although the from the targeted mixed aminal. The condensation was ex-
probe responded at neutral pH, near room temperature, tensively explored by Katritzky et al. in the 1990s[9] but
and fairly rapidly, we were interested in whether its stability, never before applied to a secondary amine of such complex-
response kinetics, or both could be improved. Here, we re- ity, let alone a macrocyclic one.
port on the expansion of this molecular system and explore
The ligands that led to the previously reported 6, 7, and
the generality of its synthetic access and fragmentation 8 (Scheme 2) were obtained in benzene under reflux with
mechanism. Thus, we varied two of the three components the help of a Dean–Stark apparatus to remove water and
of the efficient condensation reaction that leads to the drive the reaction to completion. Unfortunately, when these
underlying hexadentate ligand of the probe. Five new conditions were applied to the present precursors, they sys-
iron(II) complexes were prepared, four of which were iso- tematically led to the degradation of the precursors. How-
lated, and two of which were obtained as crystals that led to ever, the present five ligands formed readily at room tem-
X-ray structures. They exhibit electronic spectra, solution- perature in dichloromethane in the presence of activated
phase magnetic moments, and bond angles and lengths sim- molecular sieves. This success under milder conditions sug-
ilar to those of formerly reported representatives of this sys- gests that the heteroaromatic carbaldehydes based on thio-
tem. One of the new species, which displays a heteroarom- phene (1–2) and furan (3–5) form thermodynamically more
atic analyte-reactive trigger moiety, shows faster response favorable mixed aminals than those from benzaldehyde de-
kinetics than those of the previously reported parent sys- rivatives (6–8). As the aminals were prone to rapid hydroly-
tem. The present results make apparent the need for a com- sis, as previously reported for 6–8, their isolation was pre-
promise between probe stability in the absence of analyte cluded. The quality of dptacn remains of utmost impor-
and tendency to fragment in its presence. In this regard, a tance to the success of the condensation reaction. Indeed,
general trend can be derived for the variation of the azole poor conversion could often not be avoided even if the
portion.
macrocycle was stored under an argon atmosphere. We sus-
pect that the secondary amine moiety reacts with carbon
dioxide and becomes deactivated.[10] Therefore, to ensure
the success of the reaction, the sample should be recent,
preferably prepared freshly.
Results and Discussion
The five target complexes were obtained through a con-
venient two-step, one-pot reaction (Scheme 1). This con-
firms the high versatility of this synthetic pathway, as sug-
gested by previous studies.[8]
Scheme 1. Structure and synthesis of iron(II) complexes 1–5 (a crys-
b
talline material unless stated otherwise; powder).
The first step is the preparation of the hexadentate li-
gand, a mixed N,N-acetal (aminal), from three precursors.
Two of them, a heteroaromatic aldehyde and an azole, are
obtained from commercial sources; the third is the macro- Scheme 2. Previously reported iron(II) complexes.[12–14,8]
Eur. J. Inorg. Chem. 0000, 0–0
2
© 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Job/Unit: I43183
/KAP1
Date: 10-02-15 14:33:15
Pages: 8
www.eurjic.org
FULL PAPER
The second step consists of the trapping of the mixed Table 1. Selected bond lengths [Å] and angles [°] for 2.
aminals by in situ complexation with an iron(II) ion. The
Fe–N(1)
Fe–N(2)
Fe–N(3)
Fe–N(4)
Fe–N(5)
Fe–N(6)
1.976
1.984
1.985
1.987
2.015
1.925
N(1)–Fe–N(2)
N(1)–Fe–N(3)
N(1)–Fe–N(4)
N(1)–Fe–N(5)
N(1)–Fe–N(6)
N(2)–Fe–N(3)
N(2)–Fe–N(4)
N(2)–Fe–N(5)
N(2)–Fe–N(6)
N(3)–Fe–N(4)
N(3)–Fe–N(5)
N(3)–Fe–N(6)
93.03
83.93
169.53
97.72
94.15
96.69
83.17
169.05
95.02
86.82
86.57
168.22
solution of the iron salt, which is added in a titration-like
manner, has to be strictly anhydrous.[11] All of the com-
plexes, except 4, are fairly stable in dichloromethane, aceto-
nitrile, and water. Once the (swift) complexation is com-
plete, the resulting complexes can be obtained in high pu-
rity by chromatography on a reversed-phase column (C18
cartridge) with a water/MeCN mobile phase, except for 4,
which decomposes under these conditions. The purification
of the present complexes is somewhat more difficult than
that of previously reported complexes 6–8 owing to the
higher polarity of the furanylene and thiophenylene units
compared to that of a phenylene moiety. Complexes 1, 2, 3, Table 2. Selected bond lengths [Å] and angles [°] for 5.
and 5 could finally be obtained in isolated yields of 10 to
14%, as proven by mass spectrometry (Supporting Infor-
mation) and elemental analysis. Complexes 1, 2, and 5 crys-
tallized from concentrated MeCN solutions of the complex
superposed by a layer of ether.
Fe–N(1)
Fe–N(2)
Fe–N(3)
Fe–N(4)
Fe–N(5)
Fe–N(6)
1.954
1.970
1.990
1.990
2.001
1.922
N(1)–Fe–N(2)
N(1)–Fe–N(3)
N(1)–Fe–N(4)
N(1)–Fe–N(5)
N(1)–Fe–N(6)
N(2)–Fe–N(3)
N(2)–Fe–N(4)
N(2)–Fe–N(5)
N(2)–Fe–N(6)
N(3)–Fe–N(4)
N(3)–Fe–N(5)
N(3)–Fe–N(6)
92.82
84.86
169.98
96.31
93.60
97.38
83.98
170.40
93.59
86.73
86.50
168.92
The crystals for 2 and 5 were suitable for X-ray diffrac-
tion analysis. The solved structures (Figures 1 and 2,
Table S1 in the Supporting Information) confirm the mono-
nuclear, binary nature of the complexes. The bond angles
(listed in Tables 1 and 2) are in congruence with a quasi-
octahedral coordination geometry. The two structures high-
light the spatial separation of the trigger unit that is suscep-
tible to the target analyte (nitro group) on the one side and
the signal-generating coordination complex (the magne-
togenic unit) on the other side.[7]
The iron–nitrogen bond lengths (Table 3) are all slightly
below 2.0 Å, as for the analogous structures of 6–8 and the
parent diamagnetic complex 9 (FeTPTACN) reported pre-
viously. By comparison, the paramagnetic molecule 10 exhi-
bits longer bond lengths of ca. 2.2 Å.[14] This proves the
firmly low-spin nature of the present complexes 1, 2, 3, and
5 in the solid state. In addition, the triazole ring in 2 is
bound to the aminal moiety through its N(2) nitrogen
atom, as was the case for triazole-bearing 7. However, 5
displays a nitrogen atom connectivity that may form to
avoid steric clash between the indazole unit and the subse-
quent pendent picolyl group; this bonding fashion was also
observed in the analogous benzotriazole-containing com-
plex 8.[15–18]
Figure 1. Thermal ellipsoid representation of the complex unit in
2 showing the quasi-octahedral geometry (ellipsoids drawn at 50%
probability; the hydrogen atoms of the macrocycle and the two
picolyl units and solvent molecules have been omitted for clarity).
Table 3. Comparison of average Fe–N bond lengths [Å] for 2, 5,
and 6–9 in the solid state.
Average Fe–N bond length Reference
2
1.979
1.971
1.979
1.983
1.977
1.990
2.2
this work
5
this work
[8]
6
[8]
7
[8]
8
[13]
[14]
9
10
The electronic spectra of all isolated complexes have been
acquired for crystalline samples when possible (see Table 4
and the Supporting Information). The spectra are charac-
terized by the presence of a π–π* band (column A in
Table 4) at λ ≈ 250 nm (presence of picolyl groups) and a
metal-to-ligand charge-transfer band (MLCT) at λ ≈
400 nm. These wavelengths and their associated molar ex-
Figure 2. Thermal ellipsoid representation of the complex unit in
5 showing the quasi-octahedral geometry (ellipsoids drawn at 50%
probability; the hydrogen atoms of the macrocycle and the two
picolyl units and solvent molecules have been omitted for clarity).
Eur. J. Inorg. Chem. 0000, 0–0
3
© 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Job/Unit: I43183
/KAP1
Date: 10-02-15 14:33:15
Pages: 8
www.eurjic.org
FULL PAPER
tinction coefficients (ε) suggest a fully low-spin state for rings reliably leads to completely low-spin iron(II) com-
these iron(II) complexes, as illustrated by the similarity with plexes (off status)[6] in a solvent as competitive as water.[7]
the spectrum of the model compound 9.[13,19] Complexes 1,
Three prerequisites for good probe performance should
3, and 6, which contain a pyrazole moiety, exhibit a split- be put forward: (1) robustness (stability in the absence of
ting of the MLCT that might be explained by the dissym- the analyte), (2) efficient chemical conversion by reaction
metric substitution/coordination motif; by comparison, the with the analyte, and (3) swift fragmentation of the initial
other complexes, which contain a triazole unit, show a sin- product as a result of analyte conversion. These qualities
gle, shoulderless band. Complexes 5 and 8, which contain were verified by UV/Vis spectroscopy and mass spectral
indazole and benzotriazole units, respectively, give rise to monitoring.
bands of a more complex shape. All of the newly prepared
We first tested the stability of the complexes. The robust-
complexes also exhibit an additional π–π* band (labeled B ness of this line of iron(II) complexes is quite dependent on
in Table 4) at λ ≈ 300 nm owing to the presence of a nitro- the composition of the aqueous medium, as we observed
thiophenyl or nitrofuranyl unit.[20] This band is absent in for our previously reported probe 11 (Scheme 2). Complex
the spectra of 6, 7, and 8 because they comprise a phenylene 11 is stable in blood serum[8] but slowly degrades in phos-
unit that habitually absorbs below 200 nm.
phate-buffered saline (PBS, pH 7.4; UV spectroscopic data
in the Supporting Information). When samples of 1, 2, 3,
and 5 in 50 mm PBS (pH 7.4) at 37 °C were monitored by
UV spectroscopy (Figure 3) over 2 h, the spectra of 1, 3,
and 5 showed practically no evolution over time (2 h),
where those for 2 indicate degradation. Mass spectral moni-
toring at the same time points confirms this behavior (Sup-
porting Information). As a general trend, maximum sta-
bility in the phenylene series can only be obtained with the
strongest electron-withdrawing group (nitro) at the para
position (6–8), whereas a less definite electronic effect
(acylated amine), as found in 11, weakens the construction.
When moving from the phenylene series to heteroaromatic
spacers in the present study, this trend appears to occur
slightly earlier and renders even a nitro-substituted thio-
phene derivative (2) unstable. Therefore, only 1, 3, and 5
were retained for subsequent analysis.
Table 4. Major UV/Vis spectroscopy data [λmax, nm (ε, m–1 cm–1)]
of 1–3 and 5–9 in aqueous solution.
π-π* A
π-π* B
MLCT
Reference
1
2
3
5
6
7
8
9
254 (12207) 313 (12662) 393 (8419), 412 (8258) this work
254 (8148) 307 (6637)
254 (7968) 304 (9352)
389 (7470)
394 (5575), 412 (5558) this work
this work
252 (13830) 302 (11279) 402 (11658)
this work
[8]
255 (15859)
250 (14231)
255 (21935)
248 (3184)
–
–
–
–
–
397 (6657), 416 (6697)
391 (7559)
436 (12417)
432 (11970)
329 (2870)
[8]
[8]
[6,19]
[6]
10 237 (3179)
In view of our principal interest in aqueous media as the
sample environment (and the magnetogenic response to the
presence of a chemical analyte therein), we determined the
magnetic status of our target complexes in solution. The 1H
NMR spectra of 1, 2, 3, and 5 in D2O (Supporting Infor-
mation) are all in congruence with their diamagnetic nature,
as all peaks lie in the usual 10 ppm window. The method
introduced by Evans (see Table 5 and the Supporting Infor-
mation) confirms this hypothesis and furnishes values for
the effective magnetic moment (μeff) of ca. 0.8 μB at
293 K.[21,22] This corresponds to the result for model low-
spin complex 9 (FeTPTACN) and indicates the preservation
of the low-spin status in solution.[6] By contrast, 10
(Scheme 2), the ultimate product of our desired activation
process, exhibits a moment of 5.5 μB, as determined pre-
viously (Table 5).[6] Thus, it can be concluded that our
choice of the macrocycle tacn as the base system in con-
junction with the presence of three pendent arms with an
imine-type nitrogen atom and six five-membered chelate
Figure 3. Monitoring by UV/Vis spectroscopy of the stability of 1–
3 and 5 in PBS (50 mm, pH 7.4) at 37 °C. Blue line: freshly dis-
solved sample. Green line: after 2 h incubation.
Table 5. Magnetic moments μeff [μB] of 1–3, 5, and 9 in aqueous
solution at 293 K.
The efficiency of probe conversion was now assessed.
The chemical analyte used here, molecular hydrogen in the
presence of a solid-state catalyst (Pd-C 10%), is known to
reduce aromatic nitro groups very quickly. We initiated
probe conversion over 10 min of catalytic hydrogenation,
removed the catalyst by filtration, and analyzed the crude
solution immediately. Although 1, 3 and 5 all showed clean
reduction to the corresponding amino derivatives (as dem-
μeff
Reference
1
0.74
0.89
0.75
0.66
0.9
this work
this work
this work
2
3
5
this work
[6]
9
[6]
10
5.5
Eur. J. Inorg. Chem. 0000, 0–0
4
© 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Job/Unit: I43183
/KAP1
Date: 10-02-15 14:33:15
Pages: 8
www.eurjic.org
FULL PAPER
onstrated in Figures S20–S22), the UV/Vis (Figure 4, a) and sulting from the fragmentation of 3 is paramagnetic. In-
MS spectra (Figures S23 and S24) for the amines resulting deed, the application of the Evans NMR method to the
from 1 and 5 did not evolve over 2 h. On the other hand, initial sample and the totally converted sample of 3 fur-
the mass spectral signal corresponding to the amine inter- nished magnetic moments of 0.87 and 2.79 μB, respectively.
mediate of 3 disappears over 1 h (Figure 4, a and Support- This demonstrates that 3 is an effective magnetogenic probe,
ing Information, Figure S25). The monitoring of the same whereas 1, 2, and 5 are not. By analogy with the previously
process by UV/Vis spectroscopy (see Supporting Infor- reported probe 7, the amine derivative of 3 loses a pyrazole
mation) yielded a half-life of 8.6 min, which reflects slightly unit to afford a characteristic 5-methylenefuran-2-(5H)-
faster response kinetics than previously observed for probe imine intermediate (Figure S25), which is hydrolyzed to 10
7 (25 min) under the same conditions. To determine the fate (Scheme 3).[26] The final value of 2.79 μB can be explained
of the organic ligand upon complete disappearance of the by the accumulation of one full equivalent of free pyrazole
amine derivative of 3, the sample was brought to pH 13 and in the sample, which partially coordinates to the iron center
extracted with dichloromethane. The 1H NMR spectrum of at the newly established free coordination site and, thus,
the extract solely exhibits signals (see Figure S28) corre- effectively renders a limited proportion of complex 10 low-
sponding to the presence of the dptacn base system and spin. This phenomenon can be expected to be much less
pyrazole; the signals of the furan unit are absent, a fact that pronounced in a real-world setting, as concentrations
is in congruence with the claims by other authors that it should remain significantly below the 6 mm that were used
suffers degradation.[23–25]
for the present measurements.
Scheme 3. Mechanism of magnetogenic response by 3 in analogy
with 7 (50 mm PBS, pH 7.4).[8]
As a general trend, the increased stability of the initial
probe and the lack of susceptibility to fragmentation after
conversion when moving from triazole to pyrazole deriva-
tives is confirmed. This may be explained by the signifi-
cantly higher pKa value of pyrazole (pKa = 14) compared
to that of triazole (pKa = 10) and, thus, its much poorer
leaving-group character. The move from a nitrated phenyl-
ene unit in our former report to a nitrated thiophenylene
or furanylene moiety in the present work causes the overall
Figure 4. Monitoring of the fragmentation of the intermediates re- probe stability to diminish. This may explain why the sta-
sulting from the activation of 1, 3, and 5. (a) Monitoring by UV/
bility and fragmentation speed of 3 (combination of a fur-
Vis spectroscopy of a 0.05 mm PBS solution (0.02 mm for 5; pH
anylene and a pyrazole unit) and that of previously reported
7.4); blue line: immediately after activation; green line: after 2 h
7 (combination of a phenylene and a triazole unit) are sim-
ilar. Two factors may also explain the higher tendency of
heteroaryl complexes to fragment compared to that of the
incubation at 37 °C; (b, c) T1 monitoring (left, 4 mm) and 1H NMR
spectra (right, 6 mm) of 3 in PBS before (b) and after (c) activation.
The longitudinal relaxation time (T1) of the water formerly reported phenylene ones: (a) the heterocycles exer-
hydrogen nuclei of a sample of 3 before and after activation cise less steric hindrance owing to their cycle size (five-
was monitored with an NMR spectrometer (Figure 4, b). membered) and the absence of an α substituent on one of
The initial value was 2.3 s, which is somewhat below the the atoms flanking the atom connected to the aminal
value for pure water; this can be explained by the powder carbon atom; this should allow the adoption of larger dihe-
quality of 3 and the presence of traces of impurities, which dral angles and, thus, the attainment of lower-lying transi-
can have a significant impact on relaxation times, as we tion states of elimination (see Scheme 3); (b) the heteroaryl
have experienced in other contexts. After the complete dis- spacers benefit from higher electron density than that of a
appearance of the amine resulting from the reduction of 3, phenylene moiety, and this may lower the energy content of
the relaxation time of 290 ms suggests that the species re- the intermediate resulting from azole elimination.
Eur. J. Inorg. Chem. 0000, 0–0
5
© 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Job/Unit: I43183
/KAP1
Date: 10-02-15 14:33:15
Pages: 8
www.eurjic.org
FULL PAPER
lected, and the solvents were removed in vacuo to yield a red resin.
The resin was recrystallized by the liquid/liquid diffusion of diethyl
ether into a saturated MeCN solution at 4 °C to yield red-brown
Conclusions
We have explored five binary, low-spin iron(II) complexes
of significant structural variation for their ability to frag- microcrystals (0.13 g, 14%). C26H33B2F8FeN8O3.5S (775.1): calcd.
C 40.2, H 4.2, N 14.4; found C 40.4, H 3.8, N 14.1. MS (ESI): m/z
ment in aqueous media into a high-spin complex if their
aromatic nitro group is reduced to an amino group by cata-
lytic hydrogenation. Four of the complexes gave rise to
(%) = 287.1 (100) [M]2+
.
Complex 2: The same procedure was applied with 2H-1,2,3-triazole
complexes that can be isolated. Although two of the pyr- and 5-nitrothiophene-2-carbaldehyde to yield red-brown crystals
(0.12 g, 12.5%) suitable for X-ray diffraction analysis.
C25H34B2F8FeN9O4.5S (794.1): calcd. C 37.8, H 4.3, N 15.8; found
azole derivatives are inert toward fragmentation, the one
containing the nitrofuranylene trigger moiety fragments
readily once reduced. It does so at a rate somewhat faster
C 37.7, H 4.0, N 15.5. MS (ESI): m/z (%) = 287.5 (100) [M]2+
.
than that of the previously reported reference compound. Complex 3: The same procedure was applied with 1H-pyrazole and
5-nitrofuran-2-carbaldehyde to yield a red-brown microcrystalline
powder (90 mg, 10%). Attempts to crystallize the powder from
MeCN solutions were unsuccessful; the powder was used for subse-
quent studies. C26.5H32.5B2Cl1.5F8FeN8O4 (809.7): calcd. C 39.3, H
4.0, N 13.8; found C 39.1, H 3.7, N 13.5. MS (ESI): m/z (%) =
This work has demonstrated that the unusual synthetic ac-
cess to this system can be further generalized and extended
to heteroaryl carbaldehydes. The spring-loaded trigger unit
(the mixed aminal) makes the system rare in its capacity
to fragment and, thus, overcome the strong chelate effect
necessary for the observation of the initially stable probe
structures. However, true aqueous robustness for this line
of iron(II) complexes can only be found with a nitro substit-
uent as the trigger group; the drastic inductive reversion
associated with the reduction of the nitro group to an
amino group is the principal reason for the observation of
an initially stable probe that suffers swift fragmentation
when reacted with a reductive analyte. Although catalytic
hydrogenation served as a model analyte, other analytes
such as dithionite[27,28] or nitroreductase/NADH[26,29–32]
convert similar probes just as well. The latter analyte is
habitually the target in the identification of hypoxia as a
biomarker of cancer metastasis.
279.1 100 [M]2+
.
Complex 5: The same procedure was applied with 1,2-benzopyr-
azole and 5-nitrofuran-2-carbaldehyde to yield red-brown crystals
(0.10 g, 10%) suitable for X-ray diffraction analysis.
C34H38B2F8FeN10O3 (864.2): calcd. C 47.3, H 4.4, N 16.2; found
C 47.2, H 4.2, N 16.4. MS (ESI): m/z (%) = 304.1 (100) [M]2+
.
X-ray Crystallography: Crystals of 2 and 5 were selected and
mounted on a Gemini kappa-geometry diffractometer (Agilent
Technologies UK, Ltd.) equipped with an Atlas CCD detector and
a Mo-Kα graphite-monochromated radiation source (λ = 0.7107 Å).
The intensities were collected at low temperature with the CrysAlis-
Pro software to a maximum 2θ value of 59.2°.[33] Reflection in-
dexing, unit-cell parameter refinement, Lorentz polarization cor-
rection, peak integration, and background determination were per-
formed. An analytical absorption correction was applied by using
the modeled faces of the crystal.[34] The crystal parameters and
details of the data collection are summarized in the Supporting
Information. The structures were solved by direct methods with
SIR97.[35] and the least-square refinement on F2 was achieved with
the CRYSTALS software.[36] All non-hydrogen atoms were refined
anisotropically. The hydrogen atoms were all located in a difference
map, but those attached to carbon atoms were repositioned geo-
metrically. The H atoms were initially refined with soft restraints
on the bond lengths and angles to regularize their geometry (C···H
in the range 0.93–0.98 and N···H in the range 0.86–0.89 Å) and
Uiso(H) values (1.2–1.5 times the Ueq values of the parent atom),
after which the positions were refined with riding constraints.
Experimental Section
General: Dry dichloromethane (DCM) was obtained by drying
commercially available DCM for 24 h with thermally activated
molecular sieves (3 Å, sieves activated by 24 h heating at 315 °C).
Routine chemicals were supplied by Sigma–Aldrich Co., Alfa
Aesar, Acros Organics, and Tokyo Chemical Industries and used
without further purification. All 1H NMR spectra were acquired
at 297 K with a Bruker AVANCE 300 spectrometer (300 and
75 MHz for 1H and 13C, respectively). Chemical shifts (δ) are re-
ported in ppm and referenced to residual solvent signals. Unit
masses were measured by direct injection of samples into the mass
analyzer of an Agilent Technologies 1260 Infinity LC–MS system
running in electrospray ionization (ESI) mode. UV/Vis spectra were
recorded with a UV-670 UV/Vis spectrophotometer (JASCO Inc.).
CCDC-989146 (for 2) and -989147 (for 5) contain the supplemen-
tary crystallographic data for this paper. These data can be ob-
tained free of charge from The Cambridge Crystallographic Data
Centre via www.ccdc.cam.ac.uk/data_request/cif.
Complex 1: In a 50 mL flask, activated 3 Å molecular sieves (2.5 g),
1H-pyrazole (131 mg, 1.93 mmol, 1.5 equiv.), and 5-nitrothio-
phene-2-carbaldehyde (303 mg, 1.93 mmol, 1.5 equiv.) were added
Magnetic Moment Determination: The solution-phase magnetic
moments of 1–3 and 5 were determined by the Evans method with
the Schubert refinement.[21,22] For each complex, a pure sample was
dissolved in H2O/D2O (90:10) containing an additional 2%
tBuOH. The frequency shift of the tBuOH NMR signal between
this solution and that of a reference was measured in situ by using
a coaxial NMR tube; the corresponding magnetic moments of the
complexes were calculated from the frequency shifts. The formulas,
calculation parameters, and additional information are summa-
rized in Figure S13 and Table S2.
to
a solution of 1,4-bis(pyridin-2-ylmethyl)-1,4,7-triazacyclo-
nonane[6] (400 mg, 1.28 mmol) in dry dichloromethane (15 mL).
This mixture was stirred under an inert atmosphere for 15 h
at room temperature. An anhydrous acetonitrile solution of
[Fe(BF4)2·(CH3CN)6][11] was then added in a titration-like manner.
The resulting solution was filtered, and the filtrate was concen-
trated in vacuo. The residue was dissolved in a minimal volume
of water/MeCN (95:5 v/v); the resulting solution was subjected to
reversed-phase analytical chromatography with a preconditioned
C18 cartridge (Agilent; 20 mL of immobile phase). The fractions
were analyzed by mass spectrometry, the purest ones were col-
T1 Measurements: The longitudinal relaxation times (T1) of the
protons of the water molecules in aqueous solutions of 1–3 and 5
were determined by NMR spectroscopy. For each complex, a pure
Eur. J. Inorg. Chem. 0000, 0–0
6
© 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Job/Unit: I43183
/KAP1
Date: 10-02-15 14:33:15
Pages: 8
www.eurjic.org
FULL PAPER
[9] A. R. Katritzky, X. F. Lan, J. Z. Yang, O. V. Denisko, Chem.
Rev. 1998, 98, 409–548.
[10] H. Lepaumier, D. Picq, P.-L. Carrette, Ind. Eng. Chem. Res.
2009, 48, 9061–9067.
[11] D. Coucouvanis, in: Inorganic Syntheses, Wiley, New York,
2002, 33.
[12] K. Wieghardt, E. Schoffmann, B. Nuber, J. Weiss, Inorg. Chem.
1986, 25, 4877–4883.
[13] L. Christiansen, D. N. Hendrickson, H. Toftlund, S. R. Wilson,
C. L. Xie, Inorg. Chem. 1986, 25, 2813–2818.
[14] L. Spiccia, G. D. Fallon, M. J. Grannas, P. J. Nichols, E. R. T.
Tiekink, Inorg. Chim. Acta 1998, 279, 192–199.
[15] M. C. Lopez, R. M. Claramunt, P. Ballesteros, J. Org. Chem.
1992, 57, 5240–5243.
[16] P. Ballesteros, R. M. Claramunt, M. C. Lopez, J. Elguero, G.
Gomez-Alarcon, Chem. Pharm. Bull. 1988, 36, 2036–2041.
[17] P. Ballesteros, J. Elguero, R. M. Claramunt, Tetrahedron 1985,
41, 5955–5963.
[18] M. C. Gallego-Preciado, P. Ballesteros, R. M. Claramunt, M.
Cano, J. V. Heras, E. Pinilla, A. Monge, J. Organomet. Chem.
1993, 450, 237–244.
[19] S. J. Dorazio, P. B. Tsitovich, S. A. Gardina, J. R. Morrow, J.
Inorg. Biochem. 2012, 117, 212–219.
[20] F. S. Boig, G. W. Costa, I. Osvar, J. Org. Chem. 1953, 18, 775–
778.
[21] D. F. Evans, J. Chem. Soc. 1959, 2003–2005.
[22] E. M. Schubert, J. Chem. Educ. 1992, 69, 62–62.
[23] I. Parveen, D. P. Naughton, W. J. Whish, M. D. Threadgill,
Bioorg. Med. Chem. Lett. 1999, 9, 2031–2036.
[24] S. A. Everett, M. A. Naylor, K. B. Patel, M. R. Stratford, P.
Wardman, Bioorg. Med. Chem. Lett. 1999, 9, 1267–1272.
[25] S. Ferrer, D. P. Naughton, M. D. Threadgill, Tetrahedron 2003,
59, 3437–3444.
sample was dissolved in the indicated solvent containing an ad-
ditional drop of D2O. The H NMR spectra of the solution were
1
acquired by using an inversion–recovery sequence with increasing
recovery times τ. The area of the NMR signal of the water protons,
α, is proportional to the total magnetization of the water protons
in the solution and was extracted for each spectrum and plotted
against τ. The T1 values were calculated by fitting this curve with
the exponential function α(τ) = α0(1–2e–τ/T ).
1
Activation Experiments:[8] Pure samples of 1–3 or 5 were dissolved
in PBS (50 mm, pH 7.4) to give 4 mm solutions. Pd/C (10% w/w,
0.1 equiv.) was added, and hydrogen gas was bubbled through the
suspension with a needle for 5–10 min until the total reduction of
the complex was observed by mass spectrometry. The solution was
filtered and subsequently incubated at 37 °C. The activation was
monitored by MS, UV/Vis spectroscopy, T1 measurements, and
Evans NMR spectroscopy measurements.
Supporting Information (see footnote on the first page of this arti-
cle): MS spectra; X-ray structural analyses; UV/Vis spectra; NMR
spectra; solution-phase magnetic moment measurements; UV mon-
itoring of the robustness of 11 in PBS; monitoring of robustness,
conversion, and fragmentation by MS; and NMR and UV/Vis
monitoring of fragmentation.
Acknowledgments
F. T. and C. G. thank the French Ministère de la Recherche for
PhD fellowships. A postdoctoral fellowship to Y. B. by the Bulluk-
ian Foundation, Lyon, France is gratefully acknowledged. Erwann
Jeanneau is thanked for performing the X-ray structural analyses.
[26] Z. Li, X. Li, X. Gao, Y. Zhang, W. Shi, H. Ma, Anal. Chem.
2013, 85, 3926–3932.
[27] X. P. Qian, O. Hindsgaul, Chem. Commun. 1997, 1059–1060.
[28] F. Touti, P. Maurin, L. Canaple, O. Beuf, J. Hasserodt, Inorg.
Chem. 2012, 51, 31–33.
[29] J. D. Chapman, A. P. Reuvers, J. Borsa, A. Petkau, D. R.
McCalla, Cancer Res. 1972, 32, 2616–2624.
[30] J. M. Berry, C. Y. Watson, W. J. D. Whish, M. D. Threadgill, J.
Chem. Soc. Perkin Trans. 1 1997, 1147–1156.
[31] N. P. Mahmud, S. W. Garrett, M. D. Threadgill, Anti-Cancer
Drug Des. 1998, 13, 655–662.
[1] Committee on Revealing Chemistry through Advanced Chemi-
cal Imaging, National Research Council of the national Acade-
mies, Visualizing Chemistry: The Progress and Promise of Ad-
vanced Chemical Imaging, National Academies Press, Washing-
ton DC, 2006.
[2] M. P. Shores, C. M. Klug, S. R. Fiedler, in: Spin-Crossover Ma-
terials (Ed.: M. A. Halcrow), Wiley, Oxford, UK, 2013, p. 281.
[3] B. Weber, in: Spin-Crossover Materials (Ed.: M. A. Halcrow),
Wiley, Oxford, UK, 2013, p. 55.
[32] R. F. Borch, J. Liu, J. P. Schmidt, J. T. Marakovits, C. Joswig,
J. J. Gipp, R. T. Mulcahy, J. Med. Chem. 2000, 43, 2258–2265.
[4] P. B. Tsitovich, J. A. Spernyak, J. R. Morrow, Angew. Chem.
Int. Ed. 2013, 52, 13997–14000; Angew. Chem. 2013, 125, [33] CrysAlisPro, version 1.171.34.49, Agilent Technologies, 2011.
14247.
[34] R. C. Clark, J. S. Reid, Acta Crystallogr., Sect. A 1995, 51, 887–
897.
[5] J. Hasserodt, WO 2005/094,903, 2005.
[6] V. Stavila, M. Allali, L. Canaple, Y. Stortz, C. Franc, P. Mau-
rin, O. Beuf, O. Dufay, J. Samarut, M. Janier, J. Hasserodt,
New J. Chem. 2008, 32, 428–435.
[7] J. Hasserodt, J. L. Kolanowski, F. Touti, Angew. Chem. Int. Ed.
2014, 53, 60–73; Angew. Chem. 2014, 126, 60.
[8] F. Touti, P. Maurin, J. Hasserodt, Angew. Chem. Int. Ed. 2013,
52, 4654–4658; Angew. Chem. 2013, 125, 4752.
[35] A. Altomare, M. C. Burla, M. Camalli, G. L. Cascarano, C.
Giacovazzo, A. Guagliardi, A. Moliterni, G. Polidori, R.
Spagna, J. Appl. Crystallogr. 1999, 32, 115–119.
[36] P. W. Betteridge, J. R. Carruthers, R. I. Cooper, K. Prout, D. J.
Watkin, J. Appl. Crystallogr. 2003, 36, 1487–1487.
Received: December 12, 2014
Published Online: ᭿
Eur. J. Inorg. Chem. 0000, 0–0
7
© 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Job/Unit: I43183
/KAP1
Date: 10-02-15 14:33:15
Pages: 8
www.eurjic.org
FULL PAPER
Magnetogenic Probes
C. Gondrand, F. Touti, E. Godart,
Y. Berezhanskyy, E. Jeanneau, P. Maurin,
J. Hasserodt* .................................... 1–8
Spring-Loaded Iron(II) Complexes as
Magnetogenic Probes Reporting on
Chemical Analyte in Water
a
Keywords: Cage compounds / Iron / Mag-
netic properties / Macrocyclic ligands / N
ligands
A new low-spin iron(II) triazacyclononane-
based complex, selected from four different
candidates, acts as a molecular probe that
turns a diamagnetic aqueous sample into a
paramagnetic one as a response to a chemi-
cal analyte. The four complexes are pre-
pared by a highly convergent synthetic
protocol and give precious insights into
structure–activity relationship (SAR) tend-
encies.
Eur. J. Inorg. Chem. 0000, 0–0
8
© 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim