Influence of the alkyl chain length on the self-assembly of amphiphilic iron complexes: An analysis of x-ray structures

Article abstract of DOI:10.1002/chem.201304653

Several new amphiphilic iron complexes were synthesised and characterised by single crystal X-ray structure analysis. The Schiff-base-like equatorial ligands contain long alkyl chains in their outer periphery with chain lengths of 8, 12, 16 and 22 carbon atoms. As axial ligands methanol, pyridine, 4-aminopyridine, 4-(dimethylamino)pyridine and 1,2-bis(4-pyridyl)ethane were used. X-ray structure analysis of the products reveals different coordination numbers, depending on the combination of equatorial and axial ligand. The driving force for this is the self-assembly to lipid-layer-like arrangements. This can be controlled through the chain lengths and the dimension of the axial ligands in a crystal-engineering-like approach. For this an empirical rule is introduced concerning the crystallisation behaviour of the complexes. The efficacy of this rule is confirmed with the crystallisation of an octahedral complex with two docosyl (C22) chains in the outer periphery. The rule is also applied to other ligand systems. Looking ahead: Based on the X-ray structures of 15 amphiphilic iron complexes the self-assembly parameter, sap, is introduced, providing the ability to predict when lipid-layer-like structures are observed (see figure). This was used for the crystallisation of a complex with two C22 alkyl chains and can also be applied to other systems.

Full text of DOI:10.1002/chem.201304653

DOI: 10.1002/chem.201304653
Full Paper
&
Self-Assembly
Influence of the Alkyl Chain Length on the Self-Assembly of
Amphiphilic Iron Complexes: An Analysis of X-ray Structures
Stephan Schlamp, Peter Thoma, and Birgit Weber*[a]
Dedicated to Prof. Dr. I.-P. Lorenz on the occasion of his 70th birthday
Abstract: Several new amphiphilic iron complexes were syn-
thesised and characterised by single crystal X-ray structure
analysis. The Schiff-base-like equatorial ligands contain long
alkyl chains in their outer periphery with chain lengths of 8,
12, 16 and 22 carbon atoms. As axial ligands methanol, pyri-
dine, 4-aminopyridine, 4-(dimethylamino)pyridine and 1,2-
bis(4-pyridyl)ethane were used. X-ray structure analysis of
the products reveals different coordination numbers, de-
pending on the combination of equatorial and axial ligand.
The driving force for this is the self-assembly to lipid-layer-
like arrangements. This can be controlled through the chain
lengths and the dimension of the axial ligands in a crystal-
engineering-like approach. For this an empirical rule is intro-
duced concerning the crystallisation behaviour of the com-
plexes. The efficacy of this rule is confirmed with the crystal-
lisation of an octahedral complex with two docosyl (C22)
chains in the outer periphery. The rule is also applied to
other ligand systems.
Introduction
Designing multifunctional materials is of great importance as
they allow exploring many new potential applications. In this
frame, adding additional properties to spin crossover (SCO)
compounds would lead to materials that cannot only be
switched from the low-spin (LS) to the high-spin (HS) state
through thermal excitation, or induced by light, pressure or
host inclusion,[1–5] but exhibit further functionalities. One possi-
bility to introduce additional properties to such materials is the
synthesis of amphiphilic spin crossover complexes. This could
result in liquid crystalline behaviour of the material, where syn-
ergetic effects between phase transition and spin transition
(ST) are possible.[6–12] Another interesting aspect is the self-as-
sembly of such complexes in solution that might lead to the
formation of micelles or inverse micelles and can influence the
ST properties in solution.[13,14] Further the self-assembly of am-
phiphilic complexes on frontiers can be used for the formation
of Langmuir–Blodgett (LB) films.[15] Thus, such systems are not
only of interest as bulk material but also for the nanostructur-
ing of SCO materials. For a purposeful synthesis of such sys-
tems it is important to understand the self-assembly in solu-
tion as this will influence the behaviour on frontiers and the
crystal packing in the bulk material.
Scheme 1. General structure of the equatorial (top) and axial (bottom) li-
gands discussed in this work and their abbreviations.
synthesise octahedral complexes. Therefore, we designed
N2O2-coordinating Schiff-base-like equatorial ligands with long
alkyl chains in the outer periphery (Scheme 1). In combination
with various axial ligands, octahedral iron(II) complexes with
thermal spin crossover activity can be realised,[16–18] and the in-
fluence of the hydrophobic substituents can be investigated.
In a previous work we showed, that the amphiphilic complexes
of this family can self-assemble to lipid-layer-like arrange-
ments,[16,17] with one of those complexes exhibiting a highly
cooperative spin transition with a wide hysteresis loop.[17] The
X-ray structure of this complex was solved before and after the
spin transition, revealing that the volume change of the unit
cell is very small (DV/Vꢀ2.9%),[17] and in the region of thermal
For SCO complexes with iron(II), the most widely used metal
ion in spin crossover research,[3] it is commonly necessary to
[a] S. Schlamp, Dr. P. Thoma, Prof. Dr. B. Weber
Inorganic Chemistry II, Universitꢀt Bayreuth
Universitꢀtsstraße 30, NW 1, 95440 Bayreuth (Germany)
E-mail: weber@uni-bayreuth.de
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/chem.201304653.
Chem. Eur. J. 2014, 20, 1 – 13
1
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Full Paper
contraction. Despite this an up to 47 K wide thermal hysteresis
loop is observed that can be explained with a hydrogen bond
network between the polar head groups of the amphipilic
complexes and significant changes in the structure of the com-
plex upon spin transition.[1,5,17] In contrast to this gradual spin
transitions are observed for complexes with relatively short
alkyl chain substituents.[16,18] The X-ray structure of one of
those complexes in the HS and the LS state reveals, that the
volume change upon spin transition is higher compared to the
example with the long alkyl chain. However, with DV/Vꢀ4.0%
it is still small for a spin crossover complex.[18] For this example
with a very gradual spin transition no lipid-layer-like structure
was observed.
like ligands H2L(a-d)(Cn+1) were firstly converted with iron(II)
acetate in methanol to give the penta- or hexa-coordinated
[Fe(La-d)(Cn+1)(MeOH)1–2] complexes with methanol as axial
ligand. Further treatment with aromatic N-donor ligands, like
pyridine or imidazole that replace the methanol in the axial
position, leads to a [N4O2] coordination sphere that shifts the
iron(II) metal centre into the right energy region for the obser-
vation of thermal spin crossover.[4,19]
In the one hand, monomeric spin crossover complexes can
be synthesised using axial ligands like pyridine (py), 4-amino-
pyridine (apy) or (4-dimethylamino)pyridine (dmap). These
neutral compounds have been proved to be good candidates
for SCO, but are also quite air sensitive and the thermal stabili-
ty is not too high.[4,19] On the other hand, coordination poly-
mers can be realised using axial ligands like 4,4’-bipyridine
(bipy), 1,2-bis(4-pyridyl)ethane (bpea) or 1,2-bis(4-pyridyl)-
ethene (bpee). These show an increased thermal stability as
well as a dramatically increased stability against oxidation in
comparison to monomeric SCO complexes and give higher
yields due to low solubility.[4,20] Additionally, such 1D polymeric
SCO complexes are attractive for the formation of nanoparti-
cles.[21]
There appears to be a connection between the appearance
of lipid-layer-like (lamellar) structures and the cooperativity of
the spin transition. Additionally, lamellar arrangements are in-
teresting for further nanostructuring (e.g., LB film formation) or
with regard to potential liquid crystalline properties. Thus, the
question arises of whether there are factors which we can con-
trol in a crystal-engineering-like approach to obtain the desired
lamellar structural motif. A second question to be answered is
how to control the formation of octahedral complexes. First in-
vestigations on a system with short octyl chains demonstrated
that often unwanted penta-coordinated products are ob-
tained.[18]
For the synthesis of the final complexes similar reaction con-
ditions were used as for the well investigated Schiff-base-like
iron(II) complexes with no alkyl chains in the outer periphery.[4]
The combination of four different axial ligands with a total of
sixteen different equatorial ligands (the lengths of the two
alkyl chains are varied from 8, 12, and 16 up to 22 carbon
atoms and four different substituents combinations for R1 and
R2 are used) should allow the synthesis of a large pool of new
octahedral complexes. It turned out, that the desired octahe-
dral complexes were not obtained as easily as for the unmodi-
fied system.[18] The difference is especially pronounced for the
complexes with bridging axial ligands, which are readily
formed for the unsubstituted Schiff-base-like ligands.[4] In con-
It is difficult to investigate the self-assembly of the highly air
sensitive iron(II) complexes directly in solution. Therefore, we
decided to have a look at the structures of the solid products
that can be considered as the result of the self-assembly in so-
lution. Thus, the X-ray structures of 15 complexes of the am-
phiphilic Schiff-base-like ligands used in our group were com-
pared. The results lead to a rule which allows us to predict,
under which conditions the purposeful synthesis of octahedral
complexes with lamellar structures is possible. It will become
clear why of the alkyl chain lengths used for the ligand (C8,
C12, C16 and C22), the longest
chains were necessary for the
synthesis of an octahedral com-
Table 1. Overview of the complexes discussed in this work.
plex with two bulky dmap (N,N-
dimethylaminopyridine) as axial
ligands. Application of this rule
to other systems shows that it is
not limited to the complexes in-
vestigated in our group. In
Scheme 1 the general structure
of the ligands together with the
used abbreviations are given.
Complex
Abbreviation
Leq
Lax
Oxidation
state
Desired
product
Ref.
[FeLa(C8)(dmap)]
1
La(C8)
dmap
dmap
dmap
–
dmap
dmap
dmap
apy
MeOH
MeOH
bpea
bpea
apy
iron(II)
iron(II)
iron(II)
iron(III)
iron(II)
iron(II)
iron(II)
iron(II)
iron(II)
iron(II)
iron(II)
iron(II)
iron(II)
iron(II)
iron(II)
iron(II)
iron(II)
no
no
no
no
yes
yes
yes
no
yes[a]
yes[a]
no
no
no
no
yes
yes
yes
this work
[7]
[7]
this work
[7]
[7]
[7]
this work
[5]
[FeLc(C8)(dmap)]
[FeLc(C8)(dmap)]
[(FeLc(C8))2O]
2a
2b
3
Lc(C8)
Lc(C8)
Lc(C8)
[FeLc(C8)(dmap)2]ꢁMeOH LS
[FeLc(C8)(dmap)2]ꢁMeOH HS
[FeLc(C8)(dmap)2]ꢁ2MeOH
[FeLc(C8)(apy)]
4aLS
4aHS
4b
5
Lc(C8)
Lc(C8)
Lc(C8)
Lc(C8)
[FeLa(C12)(MeOH)2]
[FeLc(C12)(MeOH)2]
[FeLa(C12)2(bpea)]
[{FeLb(C12)(MeOH)}2(bpea)]
[FeLa(C12)(apy)]
[FeLb(C12)(dmap)(MeOH)]
[FeLa(C16)(py)2]ꢁ0.25H2O LS
[FeLa(C16)(py)2]ꢁ0.25H2O HS
[FeLd(C22)(dmap)2]ꢁ0.5EtOH
6
7
8
9
La(C12)
Lc(C12)
La(C12)
Lb(C12)
La(C12)
Lb(C12)
La(C16)
La(C16)
Ld(C22)
[5]
Results and Discussion
this work
this work
this work
this work
[6]
Synthesis of the complexes
10
11
12LS
12HS
13
dmap
py
py
An overview of the complexes
discussed in this work is given in
Table 1.
[6]
this work
dmap
To obtain octahedral iron(II)
complexes, the free Schiff-base-
[a] Starting material.
&
&
Chem. Eur. J. 2014, 20, 1 – 13
www.chemeurj.org
2
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Paper
Scheme 2. Exemplary reaction pathway with the desired (top) and obtained reaction product 9 (bottom).
trast to this, dimeric structures are obtained frequently, as illus-
trated in Scheme 2. Here, we focus on the synthetic ap-
proaches in which single crystals in high enough quality for an
X-ray structure analysis of the product were obtained.
whereas the other is methanol (from the solvent). For the
parent Schiff-base-like complexes only one example for such
an octahedral complex is known.[22]
Chain length of eight carbon atoms
X-ray structure analysis
Crystals suitable for X-ray structure analysis were obtained
quite frequently for the complexes with the C8 alkyl chains.
However, in most cases the quality of the diffraction data was
low. Thus, of the seven different structures (compounds 1, 2a,
In the following the X-ray structures of eight new amphiphilic
complexes are compared with those of seven previously pub-
lished complexes of this ligand system. The focus is set on the
analysis of the crystal packing,
especially the arrangement of
the alkyl chains and the van der
Table 2. Selected bond lengths [ꢂ] and angles [8] within the inner coordination sphere and spin state of the
complexes discussed in this work.
Waals (vdW) interactions be-
tween them. For complexes 1,
2a, 3, 4b, 5 and 10, the quality
of the diffraction data was low,
so only the relative conforma-
tion of the complex and the ori-
entation in the crystal is dis-
cussed. For the other examples
intermolecular interactions are
also considered. In the Support-
ing Information (Table S1), the
crystallographic data of the
newly presented complexes are
summarised. Selected bond
lengths and angles within the
first coordination sphere are
compared in Table 2. The iron
centre of the complexes is either
penta-coordinated or hexa-coor-
dinated. For the latter case often
mixed axial ligands were ob-
tained, of which one is the de-
FeꢁNeq [ꢂ]
FeꢁOeq [ꢂ]
FeꢁNax [ꢂ]
O-Fe-O [8]
Spin
state
1
2.1/2.0
2.1/2.0
1.9/2.0
2.0/2.0
2.1
2.1
2.127(4)
FeꢁOax: 1.8/1.8
2.013(2)/2.018(2)
2.216(3)/2.217(3)
2.0/2.0
101
100
HS
HS
HS
HS
LS
HS
LS
HS
HS
HS
HS
2a[7]
2b[7]
3[a]
2.073(4)/2.085(4)
2.0/2.0; 2.0/2.1
1.909(2)/1.915(2)
2.065(3)/2.080(3)
1.9/1.9
2.002(3)/1.980(3)
1.9/1.9; 1.9/2.0
1.948(1)/1.957(1)
2.002(3)/2.012(3)
2.0/2.0
101.62(14)
91/94
91.68(6)
107.16(11)
92
4aLS[7]
4aHS[7]
4b[7]
5
2.0/2.1
2.0/2.0
2.1
97
6[5]
2.093(2)/2.095(2)
2.088(2)/2.094(2)
2.077(3)/2.075(3);
2.075(3)/2.077(3)
2.107(4)/2.101(5);
2.089(5)/2.102(4)
2.1/2.0; 2.1/2.0
2.089(2)/2.095(2)
1.897(2)/1.907(2)
2.059(3)/2.086(2)
2.009(1)/2.030(1)
2.007(2)/2.021(2)
1.997(3)/1.978(3);
1.995(3)/1.977(3)
2.026(3)/2.040(3);
2.028(3)/2.026(3)
2.0/2.0; 2.0/2.0
2.002(2)/2.026(2)
1.935(2)/1.947(2)
2.001(2)/1.999(2)
FeꢁOax: 2.203(1)/2.230(2)
FeꢁOax: 2.201(2)/2.233(2)
2.130(3); 2.132(5)
108.93(5)
109.37(8)
103.58(10);
102.92(12)
112.29(16);
111.60(16)
107; 105
108.87(9)
88.80(7)
106.10(9)
7[5]
8[b]
9[c]
2.227(4); 2.242(4);
FeꢁOax: 2.228(4)/2.227(4)
2.1; 2.1
2.181(3); FeꢁOax: 2.267(3)
2.021(2)/2.014(2)
2.284(3)/2.280(4);[f]
2.284(3)/2.288(7)[g]
2.186(4)/2.224(4);
2.201(4)/2.235(4)
HS
10[d]
11
12LS[6]
12HS[6]
HS
HS
LS
HS
13[e]
2.113(4)/2.120(4);
2.090(4)/2.134(4)
2.030(3)/2.075(3);
2.016(3)/2.062(3)
111.66(12);
112.26(13)
HS
[a] Fe1; Fe3. [b] Fe1; Fe2. [c] Fe1; Fe2. [d] Fe1; Fe2. [e] Fe1; Fe2. [f, g] Disorder.
sired
N-heterocyclic
ligand
Chem. Eur. J. 2014, 20, 1 – 13
www.chemeurj.org
3
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Full Paper
2b, 3, 4a, 4b and 5), only two were of good enough quality
to be discussed fully (2b and 4a), whereas for the others only
the molecular setup and the relative orientation of the alkyl
chains is discussed. The different products for the combination
of Lc with dmap were presented recently.[18] Two different sam-
ples were obtained for both, the octahedral complex 4, and
the penta-coordinated complex 2.[18] In all cases significant dif-
ferences were observed for the relative orientation of the alkyl
chains. For the penta-coordinated samples indications for vdW
interactions were observed between the alkyl chains but not
for the octahedral samples.[18] Here, the motif of the crystal
structures of the penta-coordinated complexes 1 and 5 and of
the iron(III)-m-oxido complex 3 are presented. The complexes
¯
crystallise in the space group P1 (1) and (3) and P21/c (5). The
bond lengths around the iron centre are within the expected
range for complexes of this type (FeꢁOeq/FeꢁNeq: 1.9–2.1 ꢂ,
FeꢁNax: 2.1 ꢂ).[9] The distances to the bridging oxygen atom in
the m-oxido complex are clearly shorter (ca. 1.8 ꢂ). This is also
the case for the O-Fe-O angle (91–948) in comparison to the N-
substituted complexes (97–1018; Table 2).
The molecular setup of the complexes 1, 3 and 5 is dis-
played in Figure 1. For the complexes with the C8 alkyl chains
a wide variety of relative orientations of the alkyl chains is ob-
served. They are either parallel to each other but bent down
with regard to the equatorial Schiff-base-like ligand (Figure 1,
top; 1 and 2b), parallel to each other and in plane with the
equatorial ligand (Figure 1, middle) or spread out and parallel
with regard to the equatorial ligand (Figure 1, bottom; 5 and
2a).
The packing of 1 is very similar to that of 2b (Figure 2, top).
The almost planar planes of the equatorial ligand including the
iron are arranged above each other, the alkyl chains are bent
in the direction of the dmap of the neighbouring molecule.
Like 2b, the dmap ligands in 1 are arranged parallel to each
other and also almost parallel to the C8 alkyl chains, and the
chains are also parallel to each other and form a lipid-layer-like
arrangement. In the packing of 2b vdW interactions between
the alkyl chains and p–p interactions between the planes of
the equatorial ligand are observed.
Figure 1. Molecular setups of 1 (top), 3 (middle) and 5 (bottom). Hydrogen
atoms are omitted for clarity. Ellipsoids for the crystal structures are drawn
at the 50% probability level.
The arrangement of the molecules of 5 in the crystal
(Figure 2, bottom) is very similar to the second modification of
the penta-coordinated complex 2a. As the quality of both
structures is low, no discussion about the potential vdW inter-
actions between the alkyl chains is possible.
chains they are weak and do not dominate the crystal packing.
They are easily surpassed by other, energetically more favour-
able conditions, such as hydrogen bonds or p–p interactions.
Chain length of twelve carbon atoms
In the m-oxido sample of [FeLc(C8)] (3; Figure 2, middle), the
only iron(III) complex in this study, two inequivalent molecules
each with two iron centres (Fe-O-Fe angles: ca. 1548) are
stacked above each other, shifted sideways in a small angle.
The chains point in the same direction. Next to these two mol-
ecules, another pair is arranged, the alkyl chains of which point
in the opposite direction, so the molecules form a zipper-type
structure in which the polar head-groups are arranged next to
each other as the nonpolar tail groups are.
Crystals suitable for X-ray structure analysis were obtained for
six complexes with the chain length of 12 carbon atoms. The
two precursor complexes with two methanol as axial ligands, 6
and 7, crystallised readily out of the reaction mixture.[16] The
crystal structures of two dinuclear complexes with bpea (8, 9),
one complex with apy (10; motif) and one hexa-coordinated
complex with one dmap and one methanol as axial ligands
(11) were investigated and are discussed in the following. Crys-
tallographic data and selected bond lengths and angles are
summarised in Table S1 and Table 2, respectively. In all cases
lipid-layer-like structures are observed. However, we did not
succeed with the isolation of crystalline material with the de-
For complexes with relatively short C8 alkyl chains the crys-
tallisation mode is quite unpredictable. The strength of the
vdW interactions increases with the dimension of the hydro-
phobic part of the molecule. For the complexes with C8 alkyl
&
&
Chem. Eur. J. 2014, 20, 1 – 13
www.chemeurj.org
4
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Paper
Figure 3. Molecular setups of 11 (top), and 10 (bottom). Hydrogen atoms
are omitted for clarity. Ellipsoids are drawn at the 50% probability level.
two inequivalent iron centres whereas 11 crystallises in the
monoclinic space group P21/c with only one molecule in the
asymmetric unit and four molecules in the unit cell. Figure 3
displays the molecular setup of the two mononuclear
complexes and Figure 4 the packing of the molecules in the
crystal.
The bond lengths to the equatorial coordinating N and
O atoms are with about 2.1 and 2.0 ꢂ very similar for both
compounds and in the typical range for complexes of this
ligand type, as are the O-Fe-O angles of about 105–1098 that
clearly indicate the high spin state in both complexes. In 11,
the bond length to the O atom of the axially coordinating
methanol is almost 0.1 ꢂ longer than to the N atom of the ax-
ially attached dmap. One hydrogen bond between the O atom
of the methanol and the O atom of one of the keto groups of
the equatorial ligand of the neighboured complex molecule is
observed (O7–H51···O3, Table 3). One non-classical hydrogen
bond between C24–H24C···O1 connects the methyl group of
a dmap pointing at the OꢁFeꢁO site with the atom O1 of the
complex shifted above. For both complexes the lipid-layer-like
arrangement is supported by hydrogen bonds between the
polar head-groups.
Figure 2. Packing in the crystal of 1 along [0 0 1] (top), 3 along [1 0 0]
(middle) and 5 along [0 1 0] (bottom). Hydrogen atoms are omitted.
sired octahedral [N4O2] coordination sphere, which is interest-
ing with regard to magnetic properties.
The dinuclear complexes 8 and 9 both crystallised out of the
reaction mixtures of the corresponding iron–methanol com-
plexes with the bidentade bridging bpea in an about 30 equiv
excess after one day at room temperature (Scheme 2). For the
The mononuclear complex 10 crystallises in the triclinic
¯
space group P1 with four formula units in the unit cell and
Chem. Eur. J. 2014, 20, 1 – 13
www.chemeurj.org
5
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Full Paper
Figure 6 their packing in the crystal. The average bond lengths
and angles within the first coordination sphere do not differ
much from the other complexes and are listed in Table 2. Like
11, in 9 two hydrogen bonds are formed between the
O atoms of the coordinating methanols and the O atoms of
the keto groups of the neighboured complexes. (O17–
H17···O26 and O27–H27···O15, see Table 3 and Figure 6). For 8,
no classical hydrogen bonds are possible. Surprisingly, a very
short contact between atom Fe2 and H15B of atom C15 of
one of the ethyl groups in the equatorial ligand surrounding
the Fe1 metal centre is found that is shorter by 0.3 ꢂ than the
sum of the van der Waals radii (2.89 ꢂ). The angle N13 (of the
bpea)-Fe2-H15B is 168.18 and the angle to the carbon atom
C15 (N13-Fe2-C15) is 177.87(13)8 (length Fe2ꢁC15: 3.524(6) ꢂ).
Also, for Fe1 a contact shorter by 0.2 ꢂ than the sum of the
van der Waals radii is observed to C118 (Fe1ꢁC118: 3.491(4) ꢂ,
aN3-Fe1-C118: 177.89(10)8) and H11H (Fe1ꢁH11H: 2.90 ꢂ,
aN3-Fe1-H11H: 1688) of one of the ethyl groups in the equa-
torial ligand surrounding Fe2. An intermolecular network is
formed in which the carbon atoms of the CH3 of the ethylester
groups are connected with the iron atoms in a nearly ideal
1808 angle.
All complexes with C12 alkyl chains, the four molecules dis-
cussed in this manuscript and the two methanol complexes al-
ready described,[16] crystallise in a lipid-layer-like structure. The
alkyl chains form layers with average layer–layer distances of
about 4.8 (6), 4.8 (7), 4.1 (8), 4.4 (9), 4.2 (10) and 4.2 ꢂ (11).
Stabilising van der Waals interactions between the alkyl
chains (the London dispersion forces) can be considered, when
the distances between the atoms are in the region of the sum
of the van der Waals radii plus 0.3–0.4 ꢂ. In the case of 8, 10
and 11, such stabilising interactions are dominant. Only few
destabilising shorter contacts are observed and the average H–
H distance is about 2.7 ꢂ. In the case of 9, the ratio of stabilis-
ing and destabilising interactions is around 1:1. It can be as-
sumed that additional strong interactions (hydrogen bonds) in-
fluence the crystal packing. In the case of the methanol com-
plexes 6 and 7 the layer–layer distance is relatively long. In
agreement with this, fewer stabilising contacts between the
alkyl chains are observed.
Figure 4. Packing in the crystal of 10 (top) along [1 0 0] and 11 (bottom)
along [0 0 1]. Hydrogen atoms are omitted for clarity. Hydrogen bonds in 11
are drawn in dashed lines.
Table 3. Short contacts and hydrogen bonds and non-classical hydrogen
bonds [ꢂ] of the obtained crystal structures.
D–H···A
D–H
H···A
D···A
D–H···A
11
9
O7–H51···O3[a]
0.79(4)
0.98
0.82
0.82
0.98
0.98
0.98
0.98
0.95
0.98
1.95(4)
2.57
1.87
1.90
2.90
2.89
2.57
2.58
2.36
2.38
2.730(4)
3.395(5)
2.686(4)
2.689(4)
3.491(4)
3.524(6)
3.323(7)
3.381(11)
3.266(6)
3.202(7)
170(4)
142
175
163
120
123
134
139
160
141
C24–H24···O1[b]
O17–H17···O26[c]
O27–H27···O15[d]
C118–H11H···Fe1[e]
C15–H15B···Fe2[f]
C137–H13H···O108[g]
C30–H30B···N106[h]
C32–H32···O105[h]
C45–H45B···O8[i]
8
13
[a] x,ꢁ1/2ꢁy,ꢁ1/2+z; [b] x,y,1+z; [c] 1ꢁx,1/2+y,1/2ꢁz; [d] ꢁx,ꢁ1/2+
y,1/2ꢁz; [e] x,y,z; [f] ꢁ1+x,y,z; [g] 1+x,y,z; [h] 2ꢁx,1ꢁy,1ꢁz; [i] ꢁ1+x,y,z.
Chain length of sixteen carbon atoms
The structure of an octahedral spin crossover complex with
two alkyl chains of 16 carbon atoms was published recently in
the HS and the LS state (12LS and 12HS).[17] As for the com-
plexes with C12 alkyl chains, a lipid-layer-like structure is ob-
served in the crystal packing. The layer–layer distance between
the alkyl chains is with an average of 4.3 ꢂ (both spin states) in
the same region as observed for most of the complexes with
Cs12 alkyl chains. In agreement with this, numerous stabilising
vdW contacts can be observed.
corresponding parent compounds with no alkyl chains in the
outer periphery those reaction conditions, even with a tenfold
excess of axial ligand, always led to the corresponding coordi-
nation polymer.[20,23]
¯
While 8 crystallises in the space group P1 with two mole-
cules in the unit cell, 9 crystallises in P21/c with four molecules
in the unit cell. Despite the comparably high excess of axial
ligand, both complexes do not precipitate as coordination
polymers but form dinuclear units with two iron centres per
bridging ligand. For 9 the iron centre is saturated with metha-
nol at the sixth coordination site whereas in 8 it is penta-coor-
dinated. Figure 5 displays the molecular setup of 8 and 9 and
Chain length of 22 carbon atoms
We succeeded with the crystallisation of an octahedral iron(II)
complex with a Schiff-base-like ligand with two C22 alkyl
&
&
Chem. Eur. J. 2014, 20, 1 – 13
www.chemeurj.org
6
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Paper
chains and two dmap as axial li-
gands. Single crystals of high
enough quality for X-ray struc-
ture analysis were obtained and
the molecular setup could be
elucidated.
Compound
13
([Fe(Ld)(C22)(dmap)2]ꢁ0.5EtOH)
crystallises in the triclinic space
¯
group P1 with four molecules in
the unit cell and two inequiva-
lent iron centres bearing the de-
sired octahedral [N4O2] coordina-
tion sphere. Selected bond
lengths and angles are given in
Table 2, and crystallographic
data in Table S1 in the Support-
ing Information. Similar to the
hexa-coordinated complex 9, the
FeꢁNeq and FeꢁOeq bond lengths
are with about 2.1 ꢂ, respective-
ly, 2.0 ꢂ in the expected magni-
tude. The O-Fe-O angle of about
1128 clearly indicates that the
complex is in the HS state at
173 K. The molecular setup of 13
is displayed in Figure 7 (top).
The coordination sphere around
the iron(II) is a distorted octahe-
dron. The Schiff-base-like ligand
can be described as saddle
shaped. The chelate six rings
with the delocalised p system
are bent in the direction of the
phenylene substituents. The
angles with the [N2O2] plane are
168/238 (O1C1C2C3N1-plane and
N2C10C11C12O2-plane) for Fe1
and 258/158 for Fe2. The iron
centre is slightly shifted out of
the [N2O2] plane by 0.08 ꢂ.
Figure 5. Molecular setups of 8 (top) and 9 (bottom). Hydrogen atoms are omitted for clarity. Ellipsoids are drawn
at the 50% probability level.
The bond length to the
N atom of the axial dmap ligand
is slightly elongated at the more
shielded site of the ligand. The
planes of the dmap rings
assume a staggered conforma-
tion with an angle of 748 for Fe1
and 528 for Fe2. The C22 alkyl
chains are not arranged in plane
with the equatorial ligand but
point out of the [N2O2] plane
with an angle of 228 at Fe1 and
288 at Fe2, bent about 458 side-
ward. This is different to the
complexes with C12 and C16
alkyl chains.
Figure 6. Packing in the crystal of 8 along [0 1 0] (top) and 9 along [1 1 0] (bottom). Hydrogen atoms are omitted
for clarity. Hydrogen bonds in 9 are drawn in dashed lines.
Chem. Eur. J. 2014, 20, 1 – 13
www.chemeurj.org
7
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Full Paper
(dmap)2] is a low-spin complex
in the entire temperature range
(see Supporting Information, Fig-
ure S1). The difference to the
spin state of 13 (HS at 173 K)
can be explained with the miss-
ing additional ethanol molecule
in the crystal packing.
Prediction of the molecular
setup of amphiphilic complexes
and their arrangement in the
crystal
The results from X-ray structure
analysis show clearly, that with
an alkyl chain length of C12 or
higher, lipid-layer-like arrange-
ments are the preferred structur-
al motif. However, often not the
desired reaction product is ob-
tained. This implies that the for-
mation of lipid-layer-like ar-
rangements prevents the forma-
tion of octahedral complexes.
Thus, the question arises, what
parameters control the forma-
tion of lipid-layer-like arrange-
ments. In 1976, Israelachvili et al.
predicted how amphiphiles self-
assemble in solution, depending
on the geometry of the mole-
cules.[24] For this, he introduced
the critical packing parameter
[cpp, Eq. (1)]:[24]
cpp ¼ V Aꢁ1 Lꢁ1
ð1Þ
It correlates the volume of the
hydrophobic part (V), in relation
to the area of the head group
(A) and the length of the hydro-
Figure 7. Molecular setup (top) and packing in the crystal (bottom) of 13, packing along [1 0 0]. Hydrogen atoms
are omitted for clarity. Ellipsoids for the crystal structures are drawn at the 50% probability level. Non-classical hy-
drogen bonds are drawn in dashed lines.
phobic part (L). If cppꢀ1, in
In the crystal packing, the molecules are clearly arranged in
the lipid-layer-like structure with a layer–layer distance of
about 4.00 ꢂ (Figure 7, bottom). In comparison with the layer
distances of the other complexes bearing this structure motif,
this is short. Several non-classical hydrogen bonds are ob-
served between the polar head-groups (C137–H13H···O108,
C30–H30B···N106, C32–H32···O105 and C45–H45B···O8; see
Table 3) indicating a dense packing. Most of the contacts be-
tween the alkyl chains are in the region of the sum of the van
der Waals radii plus 0.3–0.4 ꢂ. This indicates strong stabilising
vdW interactions between the alkyl chains of this complex.
Unfortunately, the amount of crystals of 13 was not enough
for magnetic measurements on the single crystals. A separately
prepared powder sample with the composition [Fe(Ld)(C22)-
a polar solvent the molecules bearing two alkyl chains will ar-
range in lipid layers. If the cpp is smaller than one, flexible mi-
cellar structures are realised, and if it is bigger than 1, in non-
polar solvents inverse micelles are formed. It should be point-
ed out, that for all of our reactions a polar solvent was used.
On the basis of this, we tried to establish a rule for our com-
plexes regarding how they will arrange in solution and thus in
the solid. In order to obtain this rule, the broadness and the
height of the polar head and the length of the whole molecule
were measured (exemplarily shown for the complex 12LS in
Figure 8). The results are listed in Table 4.
First attempts to predict the crystal packing with the help of
the critical packing parameter (cpp) of Israelachvili et al. were
not successful. It then became obvious, that the following
&
&
Chem. Eur. J. 2014, 20, 1 – 13
www.chemeurj.org
8
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Paper
with this, other packing patterns are observed. Examples for
this are the hexa-coordinated complexes 4a/4b.
The values would match exactly, for example, for a penta-co-
ordinated complex [FeLc(C8)(MeOH)]. Indeed, the synthesised
fine crystalline powder of the methanol precursor has this
composition. Additionally it is extremely air sensitive—a further
indication for penta-coordination. Unfortunately, no crystals of
high enough quality could be isolated. For the penta-coordi-
nated complexes with one N-heterocyclic axial ligand (2a/2b,
1 and 5) the ratio is closer to 1 compared to the octahedral
systems. Thus, lipid-layer-like structures, for example, with bent
alkyl chains as for 1 and 2b, are observed. For the m-oxido
complex 3 the ratio is smaller than 1 and again another struc-
tural motif is observed.
For the complexes with alkyl chain lengths of 12 carbon
atoms the situation is different. Penta- and hexa-coordinated
products are obtained. Due to the longer hydrophobic tails
the importance of the vdW interactions as structure determin-
ing element increases and lipid-layer-like structures are ob-
served for every complex. The alkyl chains are arranged along
the [N2O2] plane without significant differences in the angles.
However, reactions with dmap, apy or the bridging bpea did
not give the desired hexa-coordinated products. Only spin
crossover inactive methanol complexes 6 and 7,[16] mixed de-
rivatives like 11, dinuclear complexes as 8 and 9 or penta-coor-
dinated compounds (10) crystallised. Application of Equa-
tion (2) explains this behaviour. As can be seen in Table 4, for
all of the crystallised complexes with C12 alkyl chains the
values for (H+B)/L are almost exactly around 1. The two
penta-coordinated complexes with slightly smaller values
reveal both short contacts to other atoms at the empty coordi-
nation site. Adding bigger axial ligands (for example, two
dmap and substituent b: Bꢀ13.0 ꢂ, Hꢀ16.30 ꢂ, Lꢀ23.89 ꢂ;
sapꢀ1.23) would cause a loss of the ability of the complex to
crystallise in the lipid-layer-like structure. This is the reason
why 11, for example, crystallises in this unusual modification
bearing one methanol instead of a second dmap. It also ex-
plains the formation of dinuclear complexes instead of coordi-
nation polymers, despite the large excess of axial ligand.
For complexes with C16 alkyl chains the situation is very
similar. Unfortunately, only the crystal structure of one complex
12 (HS and LS) could be obtained so far.[17] The axial pyridine li-
gands fit exactly in combination with substituent a and the
C16 alkyl chains ((H+B)/L for the complex in HS: 1.00) to give
a lipid-layer-like structure in the crystal.
Figure 8. Dimensions of the compounds taken in account for the calculation
of the sap. L denotes the length, H the height and B the broadness of the
complex.
Table 4. Height (H), broadness (B) and length (L) of the complexes dis-
cussed in this work.
H [ꢂ]
B [ꢂ]
L [ꢂꢁ1
]
(H+B)/L
1
8.10
8.11
8.12
17.13
14.68
14.24
14.24
14.40
14.28
14.11
14.31
16.86
14.91
15.47
13.26
15.27
12.83
16.91
17.44
16.36
16.19
16.02
15.90
19.70
19.15
18.32
18.12
17.50
23.79
23.98
23.56
24.25
23.90
23.86
29.44
29.30
32.20
1.56
1.42
1.41
0.81
1.58
1.67
1.66
1.20
1.05
0.96
0.94
1.00
0.92
1.05
0.96
1.00
1.01
2a
2b
3
4aLS
4aHS
4b
5
6
7
8
9
10
11
12LS
12HS
13
1.72
15.90
16.29
16.00
6.74
8.07
8.07
6.71
10.87
6.78
12.22
11.48
11.97
16.26
straight forward relation [the self-assembly parameter, sap,
Eq. (2)] can be used. When the sum of the broadness (B) and
the height (H) of the polar head-group, divided through the
entire length of the molecule (L) is around 1, a lipid-layer-like
packing in the crystal can be expected:
To add more bulky ligands, like dmap, following Equa-
tion (2), the chain length had to be elongated. Consequently,
the corresponding ligand and the methanol precursor with
alkyl chain lengths of 22 carbon atoms was synthesised and
converted with dmap. The complex 13 is hexa-coordinated
and exhibits a lipid-layer-like arrangement of the molecules in
the crystal. The structural situation for the C22 compound is
more complicated. The calculated length of 13 is 40.6 ꢂ
(Figure 9). This is too long for the dmap ligands, even if the
high steric demand of the equatorial ligand is considered (cal-
culated sap=0.81). This can be the reason, why the nonpolar
chains are bent quite strongly sideward. As shown in Figure 9,
sap ¼ ðH þ BÞ=L
ð2Þ
The complexes with alkyl chains of eight carbon atoms have
an entire length, L, of about 20.4 ꢂ, when the chains are ideally
arranged. Adding, for example, two dmap ligands in the axial
positions would cause an inappropriate relation of (H+B)/L
(substituent c: Bꢀ14.4 ꢂ, Hꢀ16.30 ꢂ; sapꢀ1.50). In agreement
Chem. Eur. J. 2014, 20, 1 – 13
www.chemeurj.org
9
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Full Paper
a better result is obtained if only the length of the alkyl chain
is considered (1.08). This illustrates the influence of the differ-
ent numbers of alkyl chains on the sap.
Conclusion
In this work, the crystal structures of seven published and
eight new amphiphilic iron complexes with Schiff-base-like li-
gands and nonpolar alkyl chain groups with lengths of 8, 12,
16 and 22 carbon atoms are presented and discussed. With
regard to the spin transition properties of the central metal
atom, a [N4O2] coordination sphere and self-assembly to lipid-
layer-like structures is desired. This can be achieved, if the
newly introduced sap (self-assembly parameter) is considered.
The sap allows us to predict, which general requirements have
to be fulfilled for the observation of octahedral complexes
with lipid-layer-like structures in the solid. For the synthesis of
octahedral complexes with rather bulky axial ligands as the
dmap, long alkyl chains are necessary. The synthesis of
a ligand with C22 alkyl chains allowed the crystallisation of
a corresponding octahedral iron(II) complex with dmap. This is,
to the best of our knowledge, the complex with the longest
alkyl chains in which crystallisation succeeded.
Figure 9. Measurement of the length taken in account for the calculation of
(H+B)/L for complex 13. L denotes the length of the complex. The plane on
the left is spanned by the ultimate four hydrogen atoms of the complex.
a plane was spanned at the outer H atoms of the two phenyl-
ene rings of the equatorial ligand. Measuring the distance be-
tween the plane and the latest H atoms of the alkyl chains
leads to an average length of 32.2 ꢂ, which denotes the actual
length of the complex in the crystal, and a sap of 1.01. Conse-
quently, it should be possible to add more bulky axial ligands
like phenylpyridine without losing the lipid-layer-like
arrangement.
This empirically derived rule to design octahedral SCO com-
plexes can also be successfully applied to explain the crystalli-
sation behaviour of other ligand systems. Depending on the
system, slight adjustments may be necessary. Additionally, we
gain a further insight into the self-assembly of such complexes
in solution. This is of great importance for the nanostructuring
of SCO systems (e.g., micelles or inverse micelles) or the crea-
tion of thin layers (e.g., Langmuir–Blodgett films).
The next question to be answered is, whether this concept
can be transferred to other amphiphilic complexes.[12,25] In the
field of spin crossover research, amphiphilic systems are so far
reported for iron(II),[8,26] cobalt(II),[9] iron(III)[11,13,14,27] and
manganese(III).[10]
Experimental Section
A few X-ray structures are available for complexes with alkyl
chains ꢂC8. In contrast to our system, in all cases counter ions
are involved. For manganese(III) complexes with the bapen
ligand and C6 alkyl chains different orientations of the alkyl
chains in the crystal packing are observed.[10] For a similar
iron(III) complex with the sal2trien ligand and C8 alkyl chains
already a layered structure is observed with interactions be-
tween the polar heads (and the counter ions) and interactions
between the nonpolar chains.[27] A similar situation is observed
for an iron(II) complex of the C6tren ligand with three C6 alkyl
chains.[7d]
Synthesis
The synthesis of the iron complexes and syntheses with the di-
amino precursor were carried out under an argon atmosphere
using Schlenk-tube techniques. The solvents therefore were puri-
fied as described in the literature,[30] and distilled under an atmos-
phere of argon. The synthesis of 1,2-dioctyloxybenzene, 1,2-dido-
decyloxybenzene, 1,2-dihexadecyloxybenzene, 1,2-didocosyloxy-
benzene, 4,5-dioctyloxy-1,2-dinitrobenzene, 4,5-didodecyloxy-1,2-
dinitrobenzene, 4,5-dihexadecyloxy-1,2-dinitrobenzene, 4,5-didoco-
syloxy-1,2-dinitrobenzene, 1,2-diamino-4,5-dioctyloxybenzene, 1,2-
diamino-4,5-didodecyloxybenzene, 1,2-diamino-4,5-dihexadecyloxy-
benzene, 1,2-diamino-4,5-didocosyloxybenzene,[16,17] ethoxymeth-
The cobalt(II) complexes investigated by Hayami and co-
workers differ from our system as the two alkyl chains point in
opposite directions and again counter ions are involved.[6,28]
Layered structures are observed that differ, however, quite
strongly from lipid-layer-like arrangements.
yleneethylacetoacetate,[31]
methoxymethylenemethylacetoace-
tate,[32] ethoxymethyleneacetylacetone,[31] ethoxymethylenepheny-
lacetoacetate,[33] H2La-c, [Fe(La-c)(Cn+1)(MeOH)1–2],[16,17] and iron(II)
acetate[34] are described in the literature. Pyridine (py), 4-aminopyri-
dine (apy; Alfa Aesar, 98%), 4-(dimethylamino)pyridine (dmap;
Merck, ꢃ99%) and 1,2-bis(4-pyridyl)ethane (bpea; Aldrich, 99%)
were purchased and used as received. For the complexes 5, 9 and
13 the amount of crystals was not enough to perform elemental
analysis.
An iron(III) complex of a Schiff base ligand with n-dodecyl
chains showed a very similar behaviour to our complexes with
C12 alkyl chains. X-ray structure analysis revealed self-assembly
to lipid-layer-like structures for this neutral complex in the
solid state.[29] Between the polar head-groups a network of hy-
drogen bonds is observed. For this system, the calculated
value of (H+B)/L is with 0.87 significantly smaller than 1 and
[Fe(La)(C8)(dmap)] (1): [Fe(La)(C8)(MeOH)2] (0.40 g, 0.52 mmol) and
dmap (2.1 g, 17.19 mmol, 33.1 equiv) were dissolved in methanol
(17 mL) and heated to reflux for 90 min. After storing at ꢁ308C
&
&
Chem. Eur. J. 2014, 20, 1 – 13
www.chemeurj.org
10
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Paper
black crystalline precipitate was filtered and washed two times
with methanol (2.5 mL). In the residual solution black needle-like
crystals were formed at 78C within few days. Yield: 0.02 g (5%;
820.84 gmolꢁ1). Elemental analysis calcd (%) for C43H64FeN4O8: C
62.92, H 7.86, N 6.83; found: C 62.97, H 8.09, N 7.32.
for C76H122FeN2O10: C 71.33, H 9.61, N 2.19; found: C 71.45, H 10.05,
N 2.48.
[Fe(Ld)(C22)(dmap)2]ꢀ0.5EtOH (13): [Fe(Ld)(C22)(MeOH)2] (0.25 g,
0.20 mmol) and dmap (1.22 g, 10.02 mmol, 50.0 equiv) were dis-
solved in a mixture of toluene (8 mL) and ethanol (2 mL) and
heated to reflux for 90 min. After cooling to ꢁ308C ethanol (5 mL)
were added and the green solution again stored at ꢁ308C for
11 d. The brown precipitate was filtered and dried in vacuum.
Yield: 0.13 g (45%; 1482.92 gmolꢁ1). In the residual solution black
needle-like crystals were formed.
[(Fe(Lc)(C8))2O] (3): [Fe(Lc)(C8)(MeOH)1–2] (0.3 g, 0.41 mmol) was
dissolved in pyridine (6 mL) and heated to reflux for 15 min. After
cooling to RT, H2O (4 mL) was added and the mixture was shortly
heated again until it boiled. After cooling to RT, fine crystalline
brown precipitate was filtrated. Yield: 0.27 g (79%; 828.81 gmolꢁ1).
Elemental analysis calcd (%) for C44H60FeN4O8: C 63.76, H 7.30, N
6.76; found: C 63.66, H 7.21, N 7.27. Needle-like crystals of m-oxido
complex were formed in the remaining solution standing in air.
[Fe(Ld)(C22)(dmap)2] (powder): [Fe(Ld)(C22)(MeOH)2] (0.17 g,
0.13 mmol) and dmap (0.81 g, 6.65 mmol, 50 equiv) were heated
to reflux in ethanol (30 mL) for 90 min. After 3 d the brown precipi-
tate was filtered, washed with ethanol (3 mL) and dried in vacuum.
Yield: 0.19 g (98%; 1459.89 gmolꢁ1). Elemental analysis calcd (%)
for C88H134FeN6O8: C 72.40, H 9.25, N 5.76; found: C 72.20, H 9.45, N
5.97.
[Fe(Lc)(C8)(apy)] (5): [Fe(Lc)(C8)(MeOH)1–2] (0.29 g, 0.39 mmol) and
apy (1.11 g, 11.79 mmol, 30.2 equiv) were dissolved in methanol
(10 mL) and heated to reflux for 1 h. After storing the black solu-
tion for 4 d at 68C, black block-like crystals were isolated. Yield:
0.08 g (25%; 858.84 gmolꢁ1).
X-ray diffraction
[{Fe(La)(C12)}2(bpea)] (8): [Fe(La)(C12)(MeOH)2] (0.34 g, 0.39 mmol)
and bpea (2.15 g, 11.67 mmol, 29.9 equiv) were dissolved in metha-
nol (10 mL) and heated to reflux for 75 min. After 1 d at RT, green-
ish-black needles were filtrated and washed with methanol (3 mL).
Yield: 0.06 g (16%; 1806.00 gmolꢁ1). Elemental analysis calcd (%)
for C100H152Fe2N6O16: C 66.50, H 8.48, N 4.65; found: C 66.64, H 8.70,
N 4.72.
The intensity data of 1, 2, 3, 9, 10 and 11 were collected with
a Stoe IPDS II diffractometer, the intensity data of 5 with a Bruker
Nonius Kappa CCD diffractometer and the intensity data of 13
with a Bruker D8 Venture diffractometer using graphite-monochro-
mated MoKa radiation. The data were corrected for Lorentz and po-
larisation effects. The structures were solved by direct methods
(SIR97: 1, 10, 11;[35] SIR2004: 3, 5;[36] and SHELXS-97: 8, 9, 13[37]) and
[{Fe(Lb)(C12)(MeOH)}2(bpea)] (9): [Fe(Lb)(C12)] (0.27 g, 0.36 mmol)
and bpea (1.83 g, 9.93 mmol, 27.5 equiv) were suspended in meth-
anol (10 mL) and heated to reflux for 1 h. After 1 d at RT, black pla-
telet-like crystals were filtrated. Yield: 0.02 g (3%; 1749.98 gmolꢁ1).
2
refined by full-matrix least-square techniques against F0 (SHELXL-
97, SHELXH-97 for 3, 10 and 13).[37] The hydrogen atoms were in-
cluded at calculated positions with fixed displacement parameters,
allowed to ride on their parent atoms. If not denoted differently,
for methyl and hydroxyl groups the torsion angles were allowed to
be refined according to the electron density. In the structure of 11,
the proton of the methanol hydroxyl group was directly located
according its electron density. If not denoted differently in the cif-
file, all non-hydrogen atoms were refined anisotropically. Due to
bad quality of the data of 1, 3, 5 and 10 only the general molecu-
lar setup could be investigated. For 3 and 5, twin refinement was
conducted based on twin law:
[Fe(La)(C12)(apy)] (10): [Fe(La)(C12)(MeOH)2] (0.22 g, 0.37 mmol)
and apy (1.03 g, 10.94 mmol, 29.6 equiv) were dissolved in metha-
nol (10 mL) and heated to reflux for 1 h. After cooling to RT, black
platelet-like crystals were isolated. Yield: 0.13 g (39%;
904.99 gmolꢁ1). Elemental analysis calcd (%) for C49H76FeN4O8: C
65.03, H 8.46, N 6.19; found: C 65.34, H 8.55, N 6.33.
[Fe(Lb)(C12)(dmap)(MeOH)] (11): [Fe(Lb)(C12)] (0.26 g, 0.35 mmol)
and dmap (1.17 g, 9.58 mmol, 27.4 equiv) were dissolved in metha-
nol (10 mL) and heated to reflux for 1 h. After 14 d at 68C black
needle-like crystals were filtrated and washed with methanol
(3 mL). Yield: 0.01 g (3%; 905.04 gmolꢁ1). Elemental analysis calcd
(%) for C50H80FeN4O7: C 66.35, H 8.91, N 6.19; found: C 66.64, H
9.09, N 6.15.
H2Ld(C22):
1,2-Diamino-4,5-didocosyloxybenzene
(1.2 g,
1.58 mmol) was suspended in ethanol (200 mL) saturated with
argon and ethoxymethylenephenylacetoacetate (0.94 g, 3.80 mmol,
2.4 equiv) was added dropwise. The yellow suspension was heated
to reflux for 90 min and then stored at ꢁ308C, overnight. The pre-
cipitate was filtered and washed with ethanol (20 mL). The orange
product was recrystallised from ethanol (30 mL). Yield: 1.42 g
found by PLATON.[38] ORTEP-III[39] was used for the structure repre-
sentation, Schakal-99[40] and Mercury[41] for the representation of
the molecule packing. CCDC-952444 (1), 952445 (3), 952442 (5),
952449 (8), 952446 (9), 952447 (10), 952448 (11) and 952443 (13)
contain the supplementary crystallographic data for this paper.
These data can be obtained free of charge from The Cambridge
Crystallographic Data Centre via www.ccdc.cam.ac.uk/data_re-
quest/cif.
(77%; 1161.72 gmolꢁ1). H NMR (296 K, CDCl3): d=0.87 (t, 6H, J=
8 Hz, CH3), 1.20–1.64 (m, 80H, CH3(CH2)19, CH3(Et)), 1.75–1.90 (m,
4H, CH2CH2O), 3.95–4.11 (m, 8H, CH2O), 6.82 (s, 2H, NHCqCH), 7.30–
7.75 (m, 10H, Phenyl-CH), 8.30 (dd, 2H, 3J=12.9 Hz, C=CH),
1
3
3
12.14 ppm (dd, 2H, J=12.9 Hz, NH); elemental analysis calcd (%)
for C74H116N2O8: C 76.51, H 10.06, N 2.41; found C 76.76, H 9.96, N
2.71.
[Fe(Ld)(C22)(MeOH)2]: H2Ld(C22) (1.14 g, 0.98 mmol) and iron(II)
acetate (0.36 g, 2.06 mmol, 2.1 equiv) were mixed in methanol
(160 mL) and heated to reflux for 5 h. The brown precipitate was
filtered, washed with methanol (2ꢁ15 mL) and dried in vacuum.
Yield: 0.93 g (74%; 1279.63 gmolꢁ1). Elemental analysis calcd (%)
Chem. Eur. J. 2014, 20, 1 – 13
www.chemeurj.org
11
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Full Paper
[20] W. Bauer, W. Scherer, S. Altmannshofer, B. Weber, Eur. J. Inorg. Chem.
2011, 2803–2818.
Acknowledgements
[21] a) J. Larionova, L. Salmon, Y. Guari, A. Tokarev, K. Molvinger, G. Molnꢇr,
A. Bousseksou, Angew. Chem. 2008, 120, 8360–8364; Angew. Chem. Int.
Ed. 2008, 47, 8236–8240; b) F. Volatron, L. Catala, E. Riviꢈre, A. Gloter, O.
Stꢄphan, T. Mallah, Inorg. Chem. 2008, 47, 6584–6586; c) E. Coronado,
J. R. Galꢇn-Mascarꢉs, M. Monrabal-Capilla, J. Garcꢅa-Martꢅnez, P. Pardo-
IbꢇÇez, Adv. Mater. 2007, 19, 1359–1361; d) I. Boldog, A. B. Gaspar, V.
Martꢅnez, P. Pardo-IbaÇez, V. Ksenofontov, A. Bhattacharjee, P. Gꢃtlich,
J. A. Real, Angew. Chem. 2008, 120, 6533–6537; Angew. Chem. Int. Ed.
2008, 47, 6433–6437; e) J. R. Galꢇn-Mascarꢉs, E. Coronado, A. Forment-
Aliaga, M. Monrabal-Capilla, E. Pinilla-Cienfuegos, M. Ceolin, Inorg.
Chem. 2010, 49, 5706–5714; f) S. M. Neville, C. Etrillard, S. Asthana, J.-F.
Lꢄtard, Eur. J. Inorg. Chem. 2010, 282–288; g) T. Forestier, A. Kaiba, S.
Pechev, D. Denux, P. Guionneau, C. Etrillard, N. Daro, E. Freysz, J.-F.
Lꢄtard, Chem. Eur. J. 2009, 15, 6122–6130.
Financial supports from the German Science Foundation
(SFB840; project A10) and the University of Bayreuth are ac-
knowledged. We thank S. Albrecht and P. Mayer (University of
Munich) for the collection of the single crystal X-ray data of 5
and 13.
Keywords: crystal engineering · head-tail compounds · iron ·
N,O ligands · self-assembly
[1] M. A. Halcrow, in Spin-Crossover Materials (Ed.: M. A. Halcrow), Wiley,
2013, pp. 147–169.
[22] W. Bauer, B. Weber, Acta Crystallogr. Sect. C 2008, 64, m237.
[23] a) W. Bauer, M. M. Dꢊrtu, Y. Garcia, B. Weber, CrystEngComm 2012, 14,
1223; b) T. M. Pfaffeneder, S. Thallmair, W. Bauer, B. Weber, New J. Chem.
2011, 35, 691–700.
[24] J. N. Israelachvili, D. J. Mitchell, B. W. Ninham, J. Chem. Soc. Faraday
Trans. 2 1976, 72, 1525–1568.
[2] a) M. A. Halcrow, Coord. Chem. Rev. 2009, 253, 2493–2514; b) A. B. Kou-
driavtsev, W. Linert, J. Struct. Chem. 2010, 51, 335–365; c) K. S. Murray,
Aust. J. Chem. 2009, 62, 1081–1101; d) S. Brooker, J. A. Kitchen, Dalton
Trans. 2009, 7331–7340; e) A. Gaspar, M. Seredyuk, P. Gꢃtlich, J. Mol.
Struct. 2009, 924–926, 9–19; f) J.-F. Lꢄtard, J. Mater. Chem. 2006, 16,
2550–2559; g) Spin Crossover in Transition Metal Compounds I–III (Eds.:
P. Gꢃtlich, H. Goodwin), Springer, Heidelberg, 2004; h) J. A. Kitchen, S.
Brooker, Coord. Chem. Rev. 2008, 252, 2072–2092; i) O. Sato, J. Tao, Y.-Z.
Zhang, Angew. Chem. 2007, 119, 2200–2236; Angew. Chem. Int. Ed.
2007, 46, 2152–2187.
[25] K. S. Murray, in Spin-Crossover Materials (Ed.: M. A. Halcrow), Wiley,
2013, pp. 1–54.
[26] a) S. Hayami, N. Motokawa, A. Shuto, R. Moriyama, N. Masuhara, K.
Inoue, Y. Maeda, Polyhedron 2007, 26, 2375–2380; b) M. Seredyuk, A. B.
Gaspar, V. Ksenofontov, Y. Galyametdinov, M. Verdaguer, F. Villain, P. Gꢃt-
lich, Inorg. Chem. 2010, 49, 10022–10031; c) M. Seredyuk, A. B. Gaspar,
V. Ksenofontov, Y. Galyametdinov, M. Verdaguer, F. Villain, P. Gꢃtlich,
Inorg. Chem. 2008, 47, 10232–10245.
[27] C. Gandolfi, C. Moitzi, P. Schurtenberger, G. G. Morgan, M. Albrecht, J.
Am. Chem. Soc. 2008, 130, 14434–14435.
[28] S. Hayami, Y. Shigeyoshi, M. Akita, K. Inoue, K. Kato, K. Osaka, M. Takata,
R. Kawajiri, T. Mitani, Y. Maeda, Angew. Chem. 2005, 117, 4977–4981;
Angew. Chem. Int. Ed. 2005, 44, 4899–4903.
[29] A. M. Ako, M. S. Alam, M. Rahman, J. P. Hill, N. M. Sanchez-Ballester, K.
Ariga, G. Buth, C. E. Anson, A. K. Powell, Chem. Eur. J. 2012, 18, 16419–
16425.
[30] H. G. O. Becker, Organikum, Organisch-Chemisches Grundpraktikum,
Johann Ambrosius Barth, Berlin, 1993.
[3] P. Gꢃtlich, A. Hauser, H. Spiering, Angew. Chem. 1994, 106, 2109–2141;
Angew. Chem. Int. Ed. Engl. 1994, 33, 2024–2054.
[4] B. Weber, Coord. Chem. Rev. 2009, 253, 2432–2449.
[5] M. A. Halcrow, Chem. Soc. Rev. 2011, 40, 4119–4142.
[6] Y. Garcꢅa, B.-L. Su, Y. Komatsu, K. Kato, Y. Yamamoto, H. Kamihata, Y. H.
Lee, A. Fuyuhiro, S. Kawata, S. Hayami, Eur. J. Inorg. Chem. 2012, 2769–
2775.
[7] a) S. Hayami, R. Moriyama, A. Shuto, Y. Maeda, K. Ohta, K. Inoue, Inorg.
Chem. 2007, 46, 7692–7694; b) M. Seredyuk, A. B. Gaspar, V. Ksenofon-
tov, Y. Galyametdinov, J. Kusz, P. Gꢃtlich, Adv. Funct. Mater. 2008, 18,
2089–2101; c) Y. Bodenthin, U. Pietsch, H. Mçhwald, D. G. Kurth, J. Am.
Chem. Soc. 2005, 127, 3110–3114; d) M. Seredyuk, A. B. Gaspar, V. Kse-
nofontov, Y. Galyametdinov, J. Kusz, P. Gꢃtlich, J. Am. Chem. Soc. 2008,
130, 1431–1439.
[31] L. Claisen, Justus Liebigs Ann. Chem. 1897, 297, 1–98.
[32] W. Bauer, T. Ossiander, B. Weber, Z. Naturforsch. B 2010, 65, 323–328.
[33] E.-G. Jꢆger, E. Hꢆussler, M. Rudolph, A. Schneider, Z. Anorg. Allg. Chem.
1985, 525, 67–85.
[8] A. B. Gaspar, M. Seredyuk, P. Gꢃtlich, Coord. Chem. Rev. 2009, 253,
2399–2413.
[9] S. Hayami, Y. Komatsu, T. Shimizu, H. Kamihata, Y. H. Lee, Coord. Chem.
Rev. 2011, 255, 1981–1990.
[34] B. Weber, R. Betz, W. Bauer, S. Schlamp, Z. Anorg. Allg. Chem. 2011, 637,
102–107.
[10] C. Gandolfi, T. Cotting, P. N. Martinho, O. Sereda, A. Neels, G. G. Morgan,
M. Albrecht, Dalton Trans. 2011, 40, 1855–1865.
[35] A. Altomare, M. C. Burla, M. Camalli, G. L. Cascarano, C. Giacovazzo, A.
Guagliardi, A. G. G. Moliterni, G. Polidori, R. Spagna, J. Appl. Crystallogr.
1999, 32, 115–119.
[36] M. C. Burla, R. Caliandro, M. Camalli, B. Carrozzini, G. L. Cascarano, L.
de Caro, C. Giacovazzo, G. Polidori, R. Spagna, J. Appl. Crystallogr. 2005,
38, 381–388.
[11] P. N. Martinho, C. J. Harding, H. Mꢃller-Bunz, M. Albrecht, G. G. Morgan,
Eur. J. Inorg. Chem. 2010, 675–679.
[12] S. Hayami, in Spin-Crossover Materials (Ed.: M. A. Halcrow), Wiley, 2013,
pp. 321–345.
[13] C. Gandolfi, G. G. Morgan, M. Albrecht, Dalton Trans. 2012, 41, 3726–
3730.
[37] G. Sheldrick, Acta Crystallogr. Sect. A 2008, 64, 112–122.
[38] A. L. Spek, PLATON: A Multipurpose Crystallographic Tool, Utrecht Uni-
versity, Utrecht, The Netherlands, 2008.
[39] a) L. Farrugia, J. Appl. Crystallogr. 1997, 30, 565; b) C. K. Johnson, M. N.
Burnett, ORTEP-III, Oak-Ridge National Laboratory, Oak-Ridge, TN, 1996.
[40] E. Keller, Schakal-99, University of Freiburg, Freiburg, Germany, 1999.
[41] C. F. Macrae, P. R. Edgington, P. McCabe, E. Pidcock, G. P. Shilds, R. Taylor,
M. Towler, J. van de Streek, J. Appl. Crystallogr. 2006, 39, 453–457.
[14] P. N. Martinho, Y. Ortin, B. Gildea, C. Gandolfi, G. McKerr, B. O’Hagan, M.
Albrecht, G. G. Morgan, Dalton Trans. 2012, 41, 7461–7463.
[15] a) P. N. Martinho, C. Rajnak, M. Ruben, in Spin-Crossover Materials (Ed.:
M. A. Halcrow), Wiley, 2013, pp. 375–404; b) J. A. Kitchen, N. G. White,
C. Gandolfi, M. Albrecht, G. N. L. Jameson, J. L. Tallon, S. Brooker, Chem.
Commun. 2010, 46, 6464; c) D. R. Talham, Chem. Rev. 2004, 104, 5479–
5502.
[16] S. Schlamp, P. Thoma, B. Weber, Eur. J. Inorg. Chem. 2012, 2759–2768.
[17] S. Schlamp, B. Weber, A. D. Naik, Y. Garcia, Chem. Commun. 2011, 47,
7152–7154.
Received: November 27, 2013
Revised: February 4, 2014
[18] S. Schlamp, K. Dankhoff, B. Weber, New J. Chem. 2014; DOI: 10.1039/
c3nj00991b.
[19] B. Weber, E.-G. Jꢆger, Eur. J. Inorg. Chem. 2009, 465–477.
Published online on && &&, 0000
&
&
Chem. Eur. J. 2014, 20, 1 – 13
www.chemeurj.org
12
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Paper
FULL PAPER
Looking ahead: Based on the X-ray
structures of 15 amphiphilic iron com-
plexes the self-assembly parameter, sap,
is introduced that allows to predict
when lipid-layer-like structures are ob-
served (see figure). This was used for
the crystallisation of a complex with
two C22 alkyl chains and can also be
applied to other systems.
&
Self-Assembly
S. Schlamp, P. Thoma, B. Weber*
&& – &&
Influence of the Alkyl Chain Length on
the Self-Assembly of Amphiphilic Iron
Complexes: An Analysis of X-ray
Structures
Chem. Eur. J. 2014, 20, 1 – 13
www.chemeurj.org
13
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ