Synthesis of functionalized hexadentate iminopyridine FeII complexes - Toward anion-dependent spin switching in polar media

Article abstract of DOI:10.1002/ejic.201201086

We report the syntheses and characterizations of low-spin FeII complexes of hexadentate ligands poised for aniontriggered spin-state switching in polar solutions: [Fe(L^5-OH)]- (BF₄)₂ (1) and [Fe(L^5-ONHtBu)](BF₄)₂ (3), in which L5-OH and L5-ONHtBu are tripodal iminopyridine ligands that contain methanolic or tert-butylamide functional groups, respectively, bound meta to the pyridyl N donor atom. Solid-state evidence for strong hydrogen bonding between Cl- anions and all three amide functional groups in [Fe(L^5-ONHtBu)]²⁺ is provided by the crystal structure of {[Fe(L^5-ONHtBu)]∪Cl} _2- [FeCl₄] (2). In ambient-temperature acetonitrile solutions of 1 and 3, chloride ion titrations produce marked changes in the 1H NMR spectra, including large downfield shifts for the amide NH and hydroxy OH resonances, which indicates strong anion binding events. Interestingly, for amide-containing 3, we observe small changes in magnetic susceptibility as (nBu₄N)Cl is added, which suggests that spin-state control by anion-cation interactions may be accessible for related compounds with weaker ligand fields.

Full text of DOI:10.1002/ejic.201201086

FULL PAPER
DOI:10.1002/ejic.201201086
Synthesis of Functionalized Hexadentate Iminopyridine
FeII Complexes – Toward Anion-Dependent Spin Switching
in Polar Media
CLUSTER
ISSUE
Ashley M. McDaniel,[a] Christina M. Klug,[a] and
Matthew P. Shores*[a]
Keywords: Spin crossover / Iron / Anions / Host–guest systems / Sensors
We report the syntheses and characterizations of low-spin
FeII complexes of hexadentate ligands poised for anion-
triggered spin-state switching in polar solutions: [Fe(L5-OH)]-
(BF4)2 (1) and [Fe(L5-ONHtBu)](BF4)2 (3), in which L5-OH and
L5-ONHtBu are tripodal iminopyridine ligands that contain
methanolic or tert-butylamide functional groups, respec-
tively, bound meta to the pyridyl N donor atom. Solid-state
evidence for strong hydrogen bonding between Cl– anions
and all three amide functional groups in [Fe(L5-ONHtBu)]2+ is
provided by the crystal structure of {[Fe(L5-ONHtBu)]ʚCl}2-
[FeCl4] (2). In ambient-temperature acetonitrile solutions of
1 and 3, chloride ion titrations produce marked changes in
the H NMR spectra, including large downfield shifts for the
amide NH and hydroxy OH resonances, which indicates
strong anion binding events. Interestingly, for amide-con-
taining 3, we observe small changes in magnetic suscep-
tibility as (nBu4N)Cl is added, which suggests that spin-state
control by anion–cation interactions may be accessible for re-
lated compounds with weaker ligand fields.
1
state switching schemes, it appears that anion-sensing and
spin-crossover complexes are a natural fit, because an elec-
trostatic cation–anion interaction offers a strong chemical
trigger for SC property perturbation, and most SC com-
plexes are cationic.
Introduction
We and others are interested in understanding the factors
that control spin-state switching in solution so as to develop
new applications for spin-crossover (SC) based materials.[1]
As small energy differences separate high-spin (HS) and
low-spin (LS) states in a typical FeII SC system, noncova-
lent interactions between a complex and its environment
significantly impact the observed spin states. Under the
right conditions – a suitably balanced ligand field, complex
solution stability, and sufficiently perturbing host–guest in-
teractions – spin-state switching may act as a reporter in a
chemical sensing scheme.
Anions provide an attractive substrate for sensing as they
play important roles in biology and the environment.[2] In
recognition of their impacts on living systems, interest in
effective anion sensing has spurred the design of molecular
sensors.[2a,3] The incorporation of metal atoms into sensing
architectures has become increasingly widespread; they are
used to influence complex geometry, add additional re-
porting functionality (e.g., optical, electrochemical, or mag-
netic differences in the anion-free and -associated states), or
to be the site of anion binding itself.[2a,4] Specific to spin-
Recently, we have demonstrated that homo- and hetero-
leptic FeII complexes featuring the bidentate ditopic ligand
H2bip demonstrate solution-phase anion-dependent spin
switching.[5] However, to minimize ligand exchange in solu-
tion, these studies must be performed in dichloromethane,
a solvent of low polarity. The extension of this proof-of-
concept work into more polar media (such as alcohols,
acetonitrile, or water) would be more relevant to the aque-
ous conditions of biological media or environmental sam-
ples. To do this, ligand lability must be minimized. Hexa-
dentate ligands based on the Schiff-base condensation of
tris(2-aminoethyl)amine (tren) with pyridinecarbaldehydes
combine with FeII to produce SC species;[6] a variety of FeII
complexes with related tripodal ligands also demonstrate
tunable SC behavior. However, none of these complexes
have been shown to act as anion receptors,[7] which is un-
surprising given the absence of anion-binding functional
groups.
Competent anion binding is optimized when multiple hy-
drogen-bonding groups are oriented toward a single bind-
ing site or pocket and when hydrogen bonding is coupled
with electrostatic interactions.[8] Poised NH and OH groups
can encourage direct anion–ligand interactions in tripodal
complexes (Scheme 1). Metal-based anion receptors can ef-
[a] Department of Chemistry, Colorado State University,
Fort Collins, CO 80523-1872, USA
Fax: +1-970-491-1801
E-mail: matthew.shores@colostate.edu
Homepage: http://franklin.chem.colostate.edu/shores/
Supporting information for this article is available on the
WWW under http://dx.doi.org/10.1002/ejic.201201086.
Eur. J. Inorg. Chem. 2013, 943–950
943
© 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
fectively combine these two binding modes, as shown in the FeII center would provide positive charge to attract anions
recent work by Fabbrizzi and co-workers on a (tris)imid- and also organize the hydrogen-bond donor groups into a
azolium cage organized around an FeII ion, which is cap- pocket where anions could bind. The ligand field provided
able of binding spherical and linear anions.[9] A second ex- by the diimine ligand arms would be near the SC regi-
ample uses CuII ions to preorganize a tripodal ligand to me.[11a,12] Distinct from the more commonly observed be-
sense phosphates.[10]
havior of HS (6-position) complexes, in which successful
anion binding increases ligand-field strength to cause
HSǞLS switching,[5,13] here anion binding to LS 5-position
complexes could cause geometric perturbation of the FeII
center by pushing the ligand arms out, which would lead
to lengthening of the Fe–N bonds and a transition to the
HS state. Herein, we report on the anion-binding behavior
of FeII complexes of L5-OH and L5-ONHtBu, in which we find
competent cation–chloride interactions in polar solution
and potential evidence for spin-state switching as a result
of anion binding.
Scheme 1. Tripodal ditopic ligand family targeting anion-triggered
spin-state switching.
Results and Discussion
To observe FeII anion-dependent spin switching in solu-
tion, several criteria need to be met. The complex must
competently bind anions, and the host–guest association
Syntheses
The preparation of L5-OH is relatively straightforward
must sufficiently perturb the ligand field to prompt spin- and necessitates only one or two steps from precursors
state switching. The spin-switching event must take place in found in the literature, as shown in Scheme 2.
a relatively narrow temperature range between 0 and 100 °C
The 5-hydroxy-substituted aldehyde starting material re-
for aqueous solutions and in slightly larger windows for acts cleanly with tren to afford the desired tripodal ligand;
polar organic solvents such as methanol or acetonitrile. it does not precipitate from solution and is isolated as an
These last two criteria are highly dependent on the ligand oil. In contrast, the methyl ester functionalized ligand[14]
field.
and the amide-based ligand L5-ONHtBu are isolated as free
In this work, we targeted ligands based on pyridines flowing powders.
functionalized at the 5-position as shown in Scheme 1; our
The synthesis of L5-ONHtBu is more challenging. We
recent efforts on the HS and SC 6-position analogues have aimed to produce an amide-functionalized aldehyde precur-
been reported elsewhere.[11] Chelation of the ligand to an sor by converting a 5-position methyl ester to an amide by
Scheme 2. Synthetic pathway for L5-OH
.
Scheme 3. Synthetic pathway for ligand L5-ONHtBu. The steps in the top row of the synthetic scheme are modified from ref.[17] The steps
in the middle row are adapted from ref.[14]
Eur. J. Inorg. Chem. 2013, 943–950
944
© 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
modification of a published synthetic route.[15] This was un- is consistent with the behavior of the 5-position methyl ester
successful, perhaps because of increased steric bulk of the complex, which is also LS at room temperature even though
tert-butylamine group relative to the diaminopropane an electron-withdrawing substituent is present and is ex-
group. A successful alternative synthetic path (Scheme 3) pected to weaken the ligand field.[14]
was adapted from a set of published procedures and starts
from pyridine-2,5-dicarboxylic acid. Protection at the 2-po-
sition as a methyl ester[16] is followed by generation of an
Solid-State Structural Characterization
acyl chloride from an unprotected acid at the 5-position.[17]
To explore anion–cation interactions in the solid state,
The acyl chloride for L5-ONHtBu is produced by using thio-
crystals of 1·CH3CN and 2·3CH3CN·(CH3CH2)2O were
nyl chloride and reacts readily with tert-butylamine to gen-
structurally characterized at 100 and 120 K, respectively.
erate the desired amide functionalization.[17] The amide
An Fe-containing cation for each complex is shown in Fig-
groups are unaffected by the reduction and oxidation steps
ure 1; full renderings are shown in Figures S1 and S2 (Sup-
needed to convert the 2-position ester to the aldehyde in a
porting Information). The Fe–N bond lengths for the 5-
way analogous to the steps described for the synthesis of
position functionalized 1·CH3CN and 2·3CH3CN·
L5-OH, and final condensation to produce L5-ONHtBu pro-
ceeds smoothly.
1
Both of the new ligands were characterized by H NMR
spectroscopy and positive-ion ESI-MS. The major ESI-MS
peak for each of the isolated ligands corresponds to the
1
sodium adduct of the ligand. The H NMR spectrum for
each ligand clearly shows loss of the aldehyde resonance
downfield of δ = 10 ppm and a new resonance in the δ =
8.0–8.4 ppm region owing to imine formation. In each li-
gand, the three arms tethered to the apical N atom of tren
are equivalent on the 1H NMR timescale, and the reso-
nances are split in ways consistent with the expected three-
fold symmetry. The ligands were used in subsequent steps
without further purification.
Metalation of the ligands by FeII is straightforward, and
the resulting compounds are shown in Scheme 4. Solutions
of FeII salts react rapidly with solutions of the free ligand
to form the desired complexes. The salts [Fe(L5-OH)]X [X =
–
2–
–
(BF4 )2 (1), SO4 (S1), (BPh4 )2 (S2)] and [Fe(L5-ONHtBu)]X
–
–
{X = (Cl–)[(FeCl4)2–]0.5 (2), (BF4 )2 (3), (BPh4 )1.5(Cl–)0.5
(S3)} are purple and are all stable in air. Interestingly,
attempts to directly prepare the chloride salt of
[Fe(L5-ONtBu)]2+ result instead in the formation of a salt that
contains both chloride and [FeCl4]2– anions. Evans-method
room-temperature 1H NMR spectra indicate that all the
FeII centers in these salts (except for [FeCl4]2–) are LS. This
Figure 1. One complex cation and hydrogen-bonded anion from
the asymmetric unit cell in the structures of 1·CH3CN (top) and
2·3CH3CN·(CH3CH2)2O (bottom), shown at 40% probability. First
coordination sphere and atoms participating in hydrogen-bonding
interactions are labeled. Dark red, blue, red, gray, green, and white
atoms correspond to Fe, N, O, C, Cl, H atoms, respectively. Hydro-
gen atoms bonded to carbon atoms and minor disordered compo-
nents have been omitted for clarity.
Scheme 4. Tripodal FeII iminopyridine complexes with anion-bind-
ing substituents.
Eur. J. Inorg. Chem. 2013, 943–950
945
© 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
(CH3CH2)2O are typical of low-spin FeII and range from ity holding the chloride ion. Additionally, there are close
1.948(3) to 1.996(3) Å and 1.848(3) to 1.991(3) Å, respec- contacts between the imine group and the Cl atoms of the
tively.[12] The apical nitrogen atom of the podand ligand lies [FeCl4]2– anion.
3.460(4) Å from the Fe center in 1·CH3CN and has nearly
planar geometry with C–N–C angles of 119.3(4) to
120.9(4)°. Similarly, the apical nitrogen atom in
Solution Anion-Binding Studies
2·3CH3CN·(CH3CH2)2O is also nearly planar and lies
Although cation–anion interactions have been observed
3.489(3) Å from the Fe center with C–N–C angles of in the solid state for FeII complexes with 5- or 6-position-
118.9(3)–120.6(3)°. These values contrast with the 6-posi- functionalized tripodal ligands and salts of [Fe(L6-OH)]2+
tion-functionalized [Fe(L6-OH)](BPh4)2, which remains high show temperature-dependent changes in solid-state interac-
spin at 100 K.[11a] There, the Fe–N bonds are longer than tions and spin state,[11a] neither anion binding in solution
those found in low-spin FeII complexes (ca. 1.9 Å) and nor anion-dependent spin-state behavior has been observed
range from 2.159(3) to 2.277(3) Å. The apical N atom is in solution for tripodal ligand FeII complexes. Intrigued by
also clearly pyramidalized with more acute C–N–C angles the evidence of strong chloride binding in the solid-state
that range from 116.90(17) to 117.28(17)° and a shorter N– structure containing 2, we explored the solution behavior
Fe distance of 3.242(2) Å. Key Fe–N distances and the of these tripodal complexes toward anion interaction.
1
angles around the apical N atom for each of the structures
are tabulated in Table S1 (Supporting Information).
The degree of interaction can be probed by H NMR
spectroscopy. In diamagnetic compounds, strong interac-
The complex cations in each of the structures participate tions between anions and the OH or NH protons result in
in hydrogen-bonding interactions in the solid state downfield shifts of the proton resonances owing to electro-
(Table 1). For 1·CH3CN, in which no free halide ions are static deshielding. The addition of excess Cl– ions to RuII-
present, all but one of the hydroxy groups participate in based amide anion receptors in CD3CN produces down-
hydrogen-bonding interactions with hydroxy groups on ad- field shifts of nearly 2 ppm for protons involved in hydrogen
jacent cations. The O–O distances compare closely to an bonding.[20] Similar shifts are observed for hydrogen-bond-
average reported O–O distance of 2.76(10) Å for OH···OH ing protons in LS FeII-based imidazolium anion recep-
hydrogen-bonding interactions.[18] In the structure of tors.[9] The LS nature of complexes such as 1 and 3 should
2·3CH3CN·(CH3CH2)2O, which contains a halide ion, there allow for the unambiguous detection of anion binding. If
are multiple direct hydrogen bonds between the anions and spin-state switching accompanies anion binding, magnetic
the amide functional groups of the ligands. The N–Cl dis- susceptibility increases can be detected by the Evans
tances in 2·3CH3CN·(CH3CH2)2O are very close to the re- method.
1
ported mean N–Cl distance of 3.181(6) Å for sp2 NH···Cl
The H NMR spectrum of the Fe(BF4)2 salt of L5-OH, 1,
interactions.[19] These interactions are comparable to those in CD3CN does not display paramagnetic peak shifting or
found in the previously reported [Fe(L6-OH)]Br2 struc- broadening, which indicates that the complex remains LS
ture,[11a] in which the mean O–Br distance for an OH···Br in solution. All of the peaks in the H NMR spectrum can
1
interaction is 3.254(8) Å.[19]
be assigned, including that of the OH proton at δ =
3.38 ppm. The position of this peak is not shifted greatly
from the chemical shift for the OH proton in the free ligand
(δ = 3.50 ppm in CDCl3). The similarity of the peak posi-
tions suggests that the tetrafluoroborate ions do not inter-
act strongly with the hydroxy groups in solution. However,
upon addition of excess (nBu4N)Cl to a solution of 1, the
1H NMR spectrum changes [Figures 2 and S4 (Supporting
Information)]. Small downfield shifts of 0.05 and 0.22 ppm
are observed for the imine proton (H3) and the proton at
the 6-position (H6), respectively. More dramatically, the OH
proton (H8) shifts 1.6 ppm downfield to δ = 5.0 ppm. The
magnitude of this shift indicates that there is strong interac-
tion in solution between the OH proton and the added Cl–
ions. In addition to the large δ shift, the signal splittings of
the OH proton and the adjacent methylene protons (H7)
Table 1. Hydrogen-bonding interactions in the crystal structures of
1·CH3CN and 2·3CH3CN·(CH3CH2)2O.
Complex
Interacting atoms (DH–A)[a] Distance [Å]
1·CH3CN
O1–O4
O3–O6
O4–O5
2.744(5)
2.882(5)
2.755(5)
2·3CH3CN·(CH3CH2)2O
N4–Cl1
N7–Cl1
N10–Cl1
N14–Cl2
N17–Cl2
N20–Cl2
3.169(4)
3.186(3)
3.194(4)
3.199(4)
3.220(4)
3.201(3)
[a] Complete labeling for donor and acceptor atoms is presented in
Figure S3 (Supporting Information).
There are several close contacts between the BF4– anions change: they now each appear as a doublet of doublets and
and the hydroxy groups on the complex cation in 1·CH3CN have integrations of three protons, which indicates that the
[Table S2 (Supporting Information)], but none of these in- methylene protons are no longer equivalent. A likely expla-
teractions have distances short enough to be considered hy- nation for this is that the interacting Cl– ion effectively locks
drogen bonds. In 2·3CH3CN·(CH3CH2)2O, many of the the methanol group into a single conformation, which ori-
close contacts occur between the 6-position hydrogen atom ents one hydrogen atom on the methylene group toward
on the pyridine moiety and the bound Cl– ion; however, the metal center and the other away from it, resulting in
these hydrogen atoms are not pointing directly into the cav- magnetically inequivalent positions.
Eur. J. Inorg. Chem. 2013, 943–950
946
© 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
Figure 2. 300 MHz 1H NMR spectra of [Fe(L5-OH)](BF4)2 (1) in
CD3CN at 24.9 °C (top) and after the addition of 2 equiv. of
(nBu4N)Cl (bottom). The signal marked with * is due to diethyl
ether.
Figure 3. Chemical-shift changes for the aromatic protons in
[Fe(L5-ONHtBu)](BF4)2 (3) during titration of (nBu4N)Cl. The 1H
NMR spectra were collected in CD3CN at 400 MHz and 23.5 °C.
Red: amide H7, green: imine H3, blue: 3-position H4, purple: 4-
position H5, black: 6-position H6; H atoms labelled as in Fig-
ure S10 (Supporting Information). Lines are provided as a visual
guide.
Complexes of L5-ONHtBu show similar anion dependence
in their 1H NMR spectra. For both {[Fe(L5-ONHtBu)]ʚ-
Cl}2[FeCl4] (2) and [Fe(L5-ONHtBu)](BF4)2 (3), all proton
resonances can be assigned, including those of the NH pro-
tons, which lie at δ = 9.12 and 6.79 ppm, respectively. Alter-
ation of the counteranion does not obviously influence the
spin state of the [Fe(L5-ONHtBu)]2+ complex cation at room
temperature, but the large difference in NH chemical shift brizzi’s tris(imidazolium) FeII system has also been studied
implies anion binding is operative in solutions containing through 1H NMR titrations, and association constants have
Cl– ions.
been calculated [6300 m–1 for Cl– in CD3CN/D2O (4:1)].[9]
To further study the solution interaction between chlor- However, direct comparison to our iminopyridine systems
ide ions and [Fe(L5-ONHtBu)]2+, (nBu4N)Cl was used to ti- is difficult, because anion binding is highly sensitive to sol-
trate 1.27 equiv. of Cl– ions to a solution of 3, and the pro- vent polarity and competing hydrogen-bonding interac-
1
ton shifts were monitored by H NMR spectroscopy [Fig- tions. Titrations with a very similar tris(imidazolium) FeII
ures 3 and S5 (Supporting Information)]. From the titration complex have been done in CD3CN: these give imidazolium
data, downfield shifts are seen for the amide proton, as ex- C–H peak shifts of δ = 2.67 ppm after addition of 1 equiv.
pected from comparison of the separately collected spectra of (nBu4N)Cl and yield an association constant for Cl– of
of 2 and 3. The amide peak ultimately shifts +2.2 ppm, 1.4ϫ106 m–1.[21] It is likely that [Fe(L5-ONHtBu)]2+ would not
which indicates very strong cation–anion interactions. The bind anions as strongly as these systems as it is a dicationic
proton at the 6-position also shows modest downfield shifts species, whereas the tris(imidazolium) FeII complex cations
upon chloride titration, which is consistent with its proxim- have charges of 5+.
ity to the bound anion. The other aromatic protons on the
One intriguing observation from the chloride titration
pyridine and the imine proton all show small upfield shifts, study of 3 is that the capillary and bulk solvent TMS peaks
as shown in Figure 3. Similar upfield shifts for protons are split upon addition of (nBu4N)Cl [Figures S6 and S7
pointing away from the anion-containing cavity were ob- (Supporting Information)]. It is tempting to conclude that
served for the tris(imidazolium) FeII complex studied by such behavior is indicative of an anion-triggered increase in
Fabbrizzi and co-workers.[9]
magnetic susceptibility. We note that slight increases in χMT
Focusing on the large downfield shifts, it is useful to for 3 by itself when heated above room temperature in the
compare this result to the heteroleptic RuII complex re- solid state (Experimental Section) suggest that the com-
ported by Beer and co-workers (also studied in CD3CN), pound is close to the spin-crossover region, and anion–cat-
in which a bipyridine ligand decorated with two tert-butyl ion interactions could further perturb the HS/LS ratio. Sev-
amide groups is poised for selective anion chelation.[20] For eral possible origins of this behavior are discussed in detail
this RuII complex, the amide NH signal undergoes a in the Supporting Information. Briefly, the shifts are small,
+1.62 ppm shift upon addition of 1 equiv. of Cl– ions, and do not appear to increase at higher temperatures in acetoni-
quantitative analysis of the titration data gives a stability trile solution [Figure S8a (Supporting Information)], and
constant of 6700 m–1. In the FeII systems described here, the apparent χMT changes depend on the concentration of
we expect comparable association constants for the L5-OH the species, which indicates interference from experimental
system and even stronger association for the L5-ONHtBu sys- conditions [e.g., inhomogeneous fields for the capillary
tem. This improved anion binding is likely because of the standard; see also Figure S8 (Supporting Information)].
tripodal pocket created by FeII complexation, in contrast to However, real susceptibility increases cannot be ruled out
the bidentate chelation available in the RuII system. Fab- for [Fe(L5-ONHtBu)]2+ when combined with chloride; this
Eur. J. Inorg. Chem. 2013, 943–950
947
© 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
the oxidizing agent, because it allows for easier isolation of the
product. 2,5-Pyridinedimethanol (0.486 g, 3.49 mmol) was dis-
solved in CHCl3 (100 mL). Activated MnO2 (6.07 g, 69.9 mmol)
was added, and the suspension was stirred at room temperature.
After 45 min, the mixture was filtered through Celite, the solids
were washed with CHCl3 (3ϫ 100 mL), and the filtrate was col-
lected. The solvent was removed by rotary evaporation to give a
yellow solid, which was purified by column chromatography
(EtOAc on silica gel) to give 0.165 g (34%) of light yellow powder.
1H NMR (300 MHz, CDCl3): δ = 4.86 (s, 2 H, CH2OH), 7.90 (d,
1 H, 3-H), 7.96 (d, 1 H, 4-H), 8.75 (s, 1 H, 6-H), 10.05 (s, 1 H,
CHO) ppm. These values qualitatively match previously reported
chemical shifts.[25]
suggests that similar complexes with slightly weaker ligand
fields may show a strong spin-state response to added
anions.
Conclusions
Two new tripodal iminopyridine ligands and their FeII
coordination complexes have been synthesized. Ligands
with functionalization at the 5-pyridine position form exclu-
sively LS complexes at room temperature; they are air-
stable in the solid state and in solution. The complexes in-
teract strongly with anions in the solid state and in solution.
Structural characterization of 2 clearly shows a chloride
anion bound by the three amide protons in the pocket cre-
ated by FeII coordination to the ligand. The addition of
chloride ions to 1 and 3 produces large downfield shifts for
the hydroxy and amide protons, respectively, in the NMR
spectra, which indicates strong cation–anion interactions in
solution.
{6,6Ј,6ЈЈ-[(1E,1ЈE,1ЈЈE)-{[Nitrilotris(ethane-2,1-diyl)]tris(azanylyl-
idene)}tris(methanylylidene)]tris(pyridine-6,3-diyl)}trimethanol
(L5-OH): In air, 5-(hydroxymethyl)picolinaldehyde (0.203 g,
1.46 mmol) was dissolved in methanol (6 mL). A methanolic solu-
tion (5 mL) of tris(2-aminoethyl)amine (tren, 0.071 g, 0.49 mmol)
was added, and the resulting yellow solution was heated at reflux
for 14 h. The solvent was removed by rotary evaporation to leave
1
0.237 g (97%) of an orange oil. H NMR (300 MHz, CDCl3): δ =
2.94 (t, 6 H, 1-H), 3.50 (t, 3 H, 8-H), 3.70 (t, 6 H, 2-H), 4.73 (s, 6
H, 7-H), 7.62 (dd, 3 H, 5-H), 7.82 (d, 3 H, 4-H), 8.08 (s, 3 H, 3-
H), 8.54 (d, 3 H, 6-H6) ppm. ESI-MS (MeOH): m/z = 526.27 [M +
Na]+. The compound was used in subsequent metalation reactions
without further purification.
The observation of apparent magnetic susceptibility
changes in solutions of [Fe(L5-ONHtBu)]2+ titrated with
chloride ions suggests that anion-triggered spin-state
switching may emerge in FeII complexes with a slightly
weaker ligand field. For tripodal iminopyridine ligands, the
available results[6,11a,14] demonstrate that functionalization
at all three 6- and 5-positions generate exclusively HS and
LS FeII complexes, respectively. Moving toward the spin-
crossover regime may be achieved by employing mixed-arm
tripodal ligands: Drago and co-workers have demonstrated
HS–LS tunability by adjusting steric bulk at the 6-posi-
tion;[6] the incorporation of anion-binding moieties has not
[Fe(L5-OH)](BF4)2 (1): Solid Fe(BF4)2·6H2O (50.3 mg, 0.149 mmol)
was added to a stirring methanolic solution (3 mL) of L5-OH
(73.4 mg, 0.146 mmol), which resulted in an immediate color
change to violet-plum. The mixture was stirred for 2 h, and a dark
plum precipitate formed. The precipitate was collected by filtration
to give 94.5 mg (88%) of plum-colored solid. 1H NMR (300 MHz,
CD3CN): δ = 3.08 (m, 6 H, 1-H), 3.39 (m, 3 H, 8-H), 3.53 (m, 6
H, 2-H), 4.50 (d, 6 H, 7-H), 6.96 (s, 3 H, 3-H), 8.06 (d, 3 H, 5-H),
been explored. Alternatively, changing trigonal twisting[22] 8.19 (d, 3 H, 4-H), 9.09 (d, 3 H, 6-H) ppm. ESI-MS (CH3CN):
m/z = 279.67 [Fe(L5-OH)]2+. C27H33B2F8FeN7O3 (733.05): calcd. C
by modification or substitution of the tren backbone may
44.24, H 4.54, N 13.38; found C 43.70, H 5.02, N 13.36, and C
afford spin-crossover complexes without impacting an al-
43.56, H 4.80, N 13.25. Crystals suitable for X-ray diffraction were
ready demonstrated robust anion binding pocket. Both of
grown by diffusion of diethyl ether into a concentrated solution of
these strategies are being pursued in our laboratory to tar-
get anion-triggered spin-state switching in polar media.
the complex in acetonitrile.
6,6Ј,6ЈЈ-[(1E,1ЈE,1ЈЈE)-{[Nitrilotris(ethane-2,1-diyl)]tris(azanylyl-
idene)}tris(methanylylidene)]tris[N-(tert-butyl)nicotinamide]
(L5-ONHtBu): A solution of tren (0.131 g, 0.896 mmol) in acetonitrile
(2 mL) was added to a solution of N-(tert-butyl)-6-formylnicotin-
Experimental Section
Preparation of Compounds: Unless otherwise noted, compound ma-
nipulations were performed either inside dinitrogen-filled
glovebox (MBRAUN Labmaster 130) or by Schlenk techniques on
amide (0.600 g, 2.91 mmol) in acetonitrile (8 mL). Upon addition,
the reaction mixture darkened to a dark yellow color. The reaction
mixture was stirred at room temperature for 16 h, and then the
a
an inert-gas (N2) manifold. The preparations of diethyl pyridine- solvent was removed in vacuo to afford 0.520 g (85%) of yellow
1
2,5-dicarboxylate,[23] 2,5-pyridinedimethanol,[24] and 6-(meth-
oxycarbonyl)nicotinic acid[16] have been described previously. Lit-
erature procedures for generating amides from acyl chlorides[17]
were modified to use 6-(methoxycarbonyl)nicotinic acid as the
starting material and tert-butylamine as the reagent to give methyl
5-(tert-butylcarbamoyl)picolinate. N-(tert-Butyl)-6-formylnicotin-
amide was synthesized analogously to methyl 6-formylnicotin-
ate.[14] Pentane was distilled from sodium metal and subjected to
three freeze–pump–thaw cycles. Other solvents were sparged with
dinitrogen, passed through molecular sieves, and degassed prior to
use. All other reagents were obtained from commercial sources and
were used without further purification.
powdered product. H NMR (300 MHz, CD3CN): δ = 1.45 (s, 27
H, 8-H), 2.92 (t, 6 H, 1-H), 3.72 (t, 6 H, 2-H), 6.86 (s, 3 H, 7-H),
7.85 (d, 3 H, 4-H), 8.00 (dd, 3 H, 5-H), 8.22 (s, 3 H, 3-H), 8.78 (d,
3 H, 6-H) ppm. ESI-MS (CH3OH): m/z = 733.53 [M + Na]+. The
compound was used in subsequent metalation reactions without
further purification.
{[Fe(L5-ONHtBu)]ʚCl}2[FeCl4] (2): A solution of L5-ONHtBu (0.070 g,
0.098 mmol) in acetonitrile (3 mL) was added to a stirred mixture
of FeCl2 (0.019 g, 0.15 mmol) in acetonitrile (1 mL). The resulting
dark purple solution was stirred at room temperature for 2 h, and
then the solvent was removed in vacuo to give a purple residue.
This solid was triturated with diethyl ether (10 mL) to give a dark
purple powder. The powder was collected by filtration and washed
with diethyl ether (3ϫ 3 mL) to give 0.064 g (72%) of product. 1H
5-(Hydroxymethyl)picolinaldehyde: A synthesis using SeO2 as the
oxidant has been reported in a patent;[25] here, MnO2 was used as
Eur. J. Inorg. Chem. 2013, 943–950
948
© 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
NMR (400 MHz, CD3CN): δ = 1.23 (27 H, 8-H), 3.08 (6 H, 1-H),
3.42 (3 H, 2-H), 3.56 (3 H, 2-H), 7.56 (3 H, 6-H), 8.22 (3 H, 4-H),
8.34 (3 H, 5-H), 8.98 (3 H, 7-H), 9.07 (3 H, 3-H) ppm.
ESI-MS (CH3CN): m/z = 801.33 [Fe(L5-ONHtBu)Cl]+, 383.27 [Fe-
(L5-ONHtBu)]2+. C78H108Cl6Fe3N20O6 (1802.10): calcd. C 51.99, H
6.04, N 15.55; found C 51.87, H 6.16, N 15.33. Crystals suitable
for X-ray diffraction were grown by diffusion of diethyl ether into
a concentrated solution of the complex in acetonitrile.
[Fe(L5-ONHtBu)](BF4)2 (3): A solution of Fe(BF4)2·6H2O (0.024 g,
0.076 mmol) in acetonitrile (3 mL) was added to a solution of
L5-ONHtBu (0.054 g, 0.076 mmol) in acetonitrile (2 mL). The re-
sulting dark purple solution was stirred at room temperature for
2 h, and then the solvent was removed in vacuo to give a purple
residue. This solid was triturated with diethyl ether (10 mL) to give
a dark purple powder. The powder was collected by filtration,
washed with diethyl ether (3ϫ 3 mL), and dried in vacuo to give
0.061 g (85%) of product. 1H NMR (400 MHz, CD3CN): δ = 1.31
(27 H, 8-H), 3.09 (6 H, 1-H), 3.56 (3 H, 2-H), 3.71 (3 H, 2-H), 6.79
(3 H, 7-H), 7.37 (3 H, 6-H), 8.34 (3 H, 4-H), 8.45 (3 H, 5-H), 9.27
(3 H, 3-H) ppm. ESI-MS (CH3 C N) : m/z = 853.0 [Fe-
(L5-ONHtBu)(BF4)]+, 383.27 [Fe(L5-ONHtBu)]2+. Magnetic suscep-
tibility (SQUID, 295 K): 0.62 cm3 Kmol–1 (μeff = 1.80 μB); (395 K):
0.95 cm3 Kmol–1 (μeff = 2.22 μB). C39H56B2F8FeN10O4 (3·H2O)
(958.38): calcd. C 48.88, H 5.89, N 14.61; found C 48.94, H 5.74,
N 14.35.
ization effects by using SAINT, and semiempirical absorption cor-
rections were applied by using SADABS.[26] The structure was
solved by direct methods and refined against F2 with the
SHELXTL 6.14 software package.[27] Unless otherwise noted, ther-
mal parameters for all non-hydrogen atoms were refined anisotrop-
ically. Hydrogen atoms were added at the ideal positions and were
refined by using a riding model where the thermal parameters were
set at 1.2 times those of the attached carbon atom (1.5 for methyl
protons). In the structure of 1·CH3CN, positional disorder in one
of the BF4 anions was treated by splitting atoms F15 and F16 over
two positions (refined to a 56:44 ratio); all thermal parameters were
treated anisotropically. Solvate molecule disorder in the structure
of 2·3CH3CN·(CH3CH2)2O was modeled in the following manner.
Two acetonitrile molecules were modeled at full occupancy. The
occupancy of the third acetonitrile molecule was originally tied to
a free variable that refined to 70–80% occupancy; this was rounded
to full occupancy to give a chemically more reasonable value. Posi-
tive residual electron density near the terminal C atom (C82) may
be due to further disorder. SAME, DELU and SIMU restraints
were used on each of the acetonitrile molecules with the acetonitrile
containing N21, C79, and C80 as the model for SAME. Mean-
while, the diethyl ether molecule is disordered over two positions
at two sites. Because of the large amount of disorder in this solvent
molecule, H atoms were not added to the model. Both disordered
solvent molecules were treated isotropically. CCDC-901283 (for
1·CH3CN) and CCDC-901284 [for 2·3CH3CN·(CH3CH2)2O] con-
tain the supplementary crystallographic data for this paper. These
data can be obtained free of charge from The Cambridge Crystallo-
graphic Data Centre via www.ccdc.cam.ac.uk/data_request/cif.
Structure Determinations: Crystals suitable for X-ray analyses were
coated with Paratone-N oil and supported on a Cryoloop before
being mounted on a Bruker Kappa Apex II CCD diffractometer
under a stream of dinitrogen. Data collection was performed at 100
or 120 K with Mo-Kα radiation and a graphite monochromator,
targeting complete coverage and fourfold redundancy. Initial lattice
parameters were determined from at least 500 reflections harvested
from 36 frames; these parameters were later refined against all data.
Crystallographic data and metric parameters are presented in
Table 2. Data were integrated and corrected for Lorentz and polar-
Other Physical Methods: Absorption spectra were obtained with a
Hewlett–Packard 8453 spectrophotometer in glass cuvettes with
1 cm path lengths in air and near-IR spectra were recorded with a
Cary 500 spectrophotometer; all experiments were performed at
room temperature. Mass spectrometric measurements were per-
formed in either the positive-ion or negative-ion mode with a Fin-
Table 2. Crystallographic data[a] for 1 and 2.
1·CH3CN
2·3CH3CN·(CH3CH2)2O
Empirical formula
Formula mass
Color, habit
Crystal size [mm]
Crystal system
Space group
Z
a [Å]
b [Å]
c [Å]
α [°]
C29H36B2F8FeN8O3
774.13
purple parallelepiped
0.25ϫ0.19ϫ0.15
orthorhombic
Pna21
C88H117Cl6Fe3N23O7
1989.30
purple plate
0.39ϫ0.18ϫ0.14
triclinic
¯
P1
2
8
15.4403(3)
12.2839(2)
34.8855(6)
90
90
90
6616.6(2)
1.554
100(2)
11.9443(11)
21.0685(18)
21.1533(18)
73.208(4)
89.076(4)
84.615(4)
5073.4(8)
1.302
β [°]
γ [°]
V [Å3]
dcalcd. [g/cm3]
T [K]
120(2)
F(000)
3184
11691
89312
947
0.547
1.056
4.42 (10.28)
2084
21768
72359
1117
0.643
1.023
6.00 (14.33)
No. of unique reflections
No. of observed reflections
No. of parameters
μ(Mo-Kα) [mm–1]
GOF
R1 (wR2) [%][b]
2
2 2
[a] Obtained with graphite-monochromated Mo-Kα (λ = 0.71073 Å) radiation. [b] R1 = Σ||Fo| – |Fc||/Σ|Fo|, wR2 = {Σ[w(Fo – Fc ) ]/
Σ[w(Fo2)2]}1/2 for Fo Ͼ 4σ(Fo).
Eur. J. Inorg. Chem. 2013, 943–950
949
© 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
[9]
V. Amendola, M. Boiocchi, B. Colasson, L. Fabbrizzi, M.-J. R.
Douton, F. Ugozzoli, Angew. Chem. 2006, 118, 7074; Angew.
Chem. Int. Ed. 2006, 45, 6920–6924.
S. L. Tobey, B. D. Jones, E. V. Anslyn, J. Am. Chem. Soc. 2003,
125, 4026–4027.
a) C. M. Klug, A. M. McDaniel, S. R. Fiedler, K. A. Schulte,
B. S. Newell, M. P. Shores, Dalton Trans. 2012, 41, 12577–
12585; b) A. M. McDaniel, Ph. D. Dissertation, Colorado
State University, Fort Collins, CO, 2012.
H. A. Goodwin in Spin Crossover in Transition Metal Com-
pounds I, vol. 233 (Eds.: P. Gütlich, H. A. Goodwin), Springer
Berlin, Heidelberg, 2004, pp. 59–90.
Z. Ni, S. R. Fiedler, M. P. Shores, Dalton Trans. 2011, 40, 944–
950.
A. M. McDaniel, A. K. Rappé, M. P. Shores, Inorg. Chem.
2012, 51, 12493–12502.
H.-S. Chong, S. V. Torti, R. Ma, F. M. Torti, M. W. Brechbiel,
J. Med. Chem. 2004, 47, 5230–5234.
K. Isagawa, M. Kawai, Y. Fushizaki, Nippon Kagaku Zasshi
1967, 88, 553–556.
a) B. J.-S. Wang, J. Chem. Educ. 1974, 51, 631; b) D. L. Pavia,
G. M. Lampman, G. S. Kriz, R. G. Engel, Introduction to Or-
ganic Laboratory Techniques: A Small Scale Approach, 1st ed.,
Saunders College Publishing, Fort Worth, 1998.
S. Wallwork, Acta Crystallogr. 1962, 15, 758–759.
T. Steiner, Acta Crystallogr., Sect. B: Struct. Sci. 1998, 54, 456–
463.
nigan LCQ Duo mass spectrometer, equipped with an analytical
electrospray ion source and a quadrupole ion trap mass analyzer.
1H NMR spectra were measured with Varian INOVA 300 MHz
or 400 MHz instruments. Magnetic susceptibilities in solution were
determined by the Evans method.[28] Magnetic susceptibility mea-
surements in the solid state were performed with a Quantum De-
sign model MPMS-XL superconducting quantum interference de-
vice (SQUID) magnetometer in the temperature range 295–395 K
under a field of 1000 Oe. One square inch of aluminum foil and six
inches of copper wire were used as the sample holder. The data
were corrected by subtracting the measured susceptibility of an
empty sample holder. Diamagnetic corrections were applied by
using Pascal’s constants.[29]
[10]
[11]
[12]
[13]
[14]
[15]
[16]
[17]
Supporting Information (see footnote on the first page of this arti-
cle): Preparations of compounds S1–S3; key bond/internuclear dis-
1
tances and angles; H NMR spectra including chloride titrations.
Acknowledgments
This work was supported by the National Science Foundation
(CHE-1058889).
[18]
[19]
[1] M. P. Shores, C. M. Klug, S. R. Fiedler in Spin-Crossover Ma-
terials – Properties and Applications (Ed.: M. A. Halcrow),
John Wiley & Sons, Chichester, 2012.
[2] a) R. Vilar, Eur. J. Inorg. Chem. 2008, 357–367; b) A. Bianchi,
K. Bowman-James, E. García-España, Supramolecular Chemis-
try of Anions, Wiley-VCH, 1997; c) J. L. Sessler, P. A. Gale,
W.-S. Cho, Anion Receptor Chemistry, The Royal Society of
Chemistry, Cambridge, UK, 2006.
[20] P. D. Beer, S. W. Dent, T. J. Wear, J. Chem. Soc., Dalton Trans.
1996, 2341–2346.
[21] K. Sato, Y. Sadamitsu, S. Arai, T. Yamagishi, Tetrahedron Lett.
2007, 48, 1493–1496.
[22] a) D. L. Reger, J. D. Elgin, M. D. Smith, F. Grandjean, L.
Rebbouh, G. J. Long, Polyhedron 2006, 25, 2616–2622; b) D. L.
Reger, J. R. Gardinier, J. D. Elgin, M. D. Smith, D. Hautot,
G. J. Long, F. Grandjean, Inorg. Chem. 2006, 45, 8862–8875.
[3] a) P. D. Beer, P. A. Gale, Angew. Chem. 2001, 113, 502; Angew.
Chem. Int. Ed. 2001, 40, 486–516; b) P. A. Gale, S. E. Garcia-
Garrido, J. Garric, Chem. Soc. Rev. 2008, 37, 151–190; c) J. W. [23] P. Bisel, J. P. Gies, G. Schlewer, C. G. Wermuth, Bioorg. Med.
Steed, Chem. Soc. Rev. 2009, 38, 506–519.
Chem. Lett. 1996, 6, 3025–3028.
[4] a) P. D. Beer, S. R. Bayly in Anion Sensing, vol. 255 (Ed.: I.
Stibor), Springer Berlin, Heidelberg, 2005, pp. 373–378; b)
C. R. Rice, Coord. Chem. Rev. 2006, 250, 3190–3199.
[5] a) Z. Ni, M. P. Shores, J. Am. Chem. Soc. 2008, 131, 32–33; b)
Z. Ni, A. M. McDaniel, M. P. Shores, Chem. Sci. 2010, 1, 615–
621; c) Z. Ni, M. P. Shores, Inorg. Chem. 2010, 49, 10727–
10735.
[6] M. A. Hoselton, L. J. Wilson, R. S. Drago, J. Am. Chem. Soc.
1975, 97, 1722–1729.
[7] a) H. Z. Lazar, T. Forestier, S. A. Barrett, C. A. Kilner, J.-F.
Létard, M. A. Halcrow, Dalton Trans. 2007, 4276–4285; b)
M. A. Halcrow, Polyhedron 2007, 26, 3523–3576; c) M. Sered-
yuk, A. B. Gaspar, J. Kusz, P. Gütlich, Z. Anorg. Allg. Chem.
2011, 637, 965–976.
[24] M. I. Dawson, R. Chan, P. D. Hobbs, W. Chao, L. J. Schiff, J.
Med. Chem. 1983, 26, 1282–1293.
[25] M. I. Dawson, R. L. S. Chan, P. D. Hobbs, SRI International,
USA. 1985, 9 pages.
[26] APEX 2, Bruker Analytical X-ray Systems, Inc., Madison, WI,
2008.
[27] G. M. Sheldrick, SHELXTL, Version 6.14, Bruker Analytical
X-ray Systems, Inc., Madison, WI, 1999.
[28] a) D. F. Evans, J. Chem. Soc. 1959, 2003–2005; b) D. H. Live,
S. I. Chan, Anal. Chem. 1970, 42, 791–792; c) D. Ostfeld, I. A.
Cohen, J. Chem. Educ. 1972, 49, 829; d) E. M. Schubert, J.
Chem. Educ. 1992, 69, 62; e) D. H. Grant, J. Chem. Educ. 1995,
72, 39.
[29] G. A. Bain, J. F. Berry, J. Chem. Educ. 2008, 85, 532–536.
[8] P. D. Beer, P. A. Gale, D. K. Smith, Supramolecular Chemistry,
Oxford University Press, Oxford, New York, 1999.
Received: September 15, 2012
Published Online: January 16, 2013
Eur. J. Inorg. Chem. 2013, 943–950
950
© 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim