- Email: [email protected]
Metallocenophanes bridged by group 13 elements
Accepted Manuscript Title: Metallocenophanes Bridged by Group 13 Elements Author: Hridaynath Bhattacharjee Jens M¨uller PII: DOI: Reference:
S0010-8545(15)00287-8 http://dx.doi.org/doi:10.1016/j.ccr.2015.09.008 CCR 112141
To appear in:
Coordination Chemistry Reviews
Received date: Revised date: Accepted date:
26-6-2015 9-9-2015 15-9-2015
Please cite this article as: H. Bhattacharjee, J. M¨uller, Metallocenophanes Bridged by Group 13 Elements, Coordination Chemistry Reviews (2015), http://dx.doi.org/10.1016/j.ccr.2015.09.008 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Ac
ce
pt
ed
M
an
us
cr
i
Graphical Abstract (for review)
Page 1 of 70
Highlights
The development of strained [1]ferrocenophanes with B, Al, Ga, and In is summarized. The application of [n]metallocenophanes for polymerization is discussed.
Ac ce p
te
d
M
an
us
cr
ip t
Structural aspects of strained and unstrained metallocenophanes are illustrated.
1
Page 2 of 70
Metallocenophanes Bridged by Group 13 Elements
ip t
Hridaynath Bhattacharjee and Jens Müller* Department of Chemistry, University of Saskatchewan, 110 Science Place, Saskatoon, Saskatchewan
cr
S7N 5C9, Canada
us
[email protected]; phone: +1 306 966 2466
an
RECEIVED DATE
ABSTRACT. Metallocenophanes with one or multiple sandwich moieties and at least one group 13
M
element in bridging position are reviewed. The majority of this review summarizes publications
d
concerned with strained sandwich compounds and their application as monomers for metal-containing
te
polymers. Publications up to early 2015 are included.
Ac ce p
Keywords: group 13 elements; ferrocenophanes; ruthenocenophanes; ring-opening polymerization; metallopolymers
Abbreviations: Ar', 2-[(dimethylamino)methyl]phenyl; DSC, differential scanning calorimetry; DLS, dynamic light scattering; Fc, Fe(H4C5)(H5C5); fc, Fe(H4C5)2; FCP, ferrocenophane; GPC, gel permeation chromatography; Mamx, 2,4-di-tert-butyl-6-[(dimethylamino)methyl]phenyl); Me2Ntsi, [(dimethylamino)dimethylsilyl]bis(trimethylsilyl)methyl; Mpysm, [dimethyl(2-pyridyl)silyl]methyl; PFS = poly(ferrocenylsilane); PS, polystyrene; Pytsi, [dimethyl(2pyridyl)silyl]bis(trimethylsilyl)methyl; RCP, ruthenocenophane; ROP, ring-opening polymerization; trisyl, tris(trimethylsilyl)methyl
2
Page 3 of 70
Contents 1. Introduction and scope 2. Aluminum- and gallium-bridged [1]metallocenophanes 2.1. The first generation: trisyl-based ligands
ip t
2.2. Conclusions from the first generation 2.3. The second generation: the Mamx ligand
cr
2.3.1. Metallopolymers
us
2.3.2. DFT calculations 2.4. Conclusions from the second generation
an
2.5. The third generation
4. Boron-bridged [1]metallocenophanes 4.1. Synthesis and mechanistic insights
M
3. Indium-bridged [1]metallocenophanes
Ac ce p
6. [n.n]Metallocenophanes
te
5. [n]Metallocenophanes (n ≥ 2)
d
4.2. Reactivity of bora[1]ferrocenophanes
7. Conclusions
3
Page 4 of 70
1. Introduction and scope [n]Metallocyclophanes are complexes where two metal-coordinated mancude-ring systems are linked by a bridging chain with n atoms in its backbone.1 This is illustrated in Figure 1 for homoleptic (A–C) and heteroleptic compounds (D and E). The largest class is that of [n]metallocenophanes (A), with the
ip t
most examples known for [n]ferrocenophanes (M = Fe). The family of species of type B could be
cr
addressed as [n]metallocenophanes, but are more often called ansa complexes. If one considers the IUPAC definition [1] for metallocenes as sandwich complexes of the type bis(5-
us
cyclopentadienyl)metal, where the d metal is bonded only to the face of two cyclopentadienyl ligands,
an
species of type B are not [n]metallocenophanes as the parent species are not metallocenes. However, in practice, this recommendation is not followed and species such as Cp2MCl2 are usually classified as
M
metallocenes. More importantly, classes A and B are chemically very different. A short bridge in A forces the Cp rings to deviate from their co-planarity in the parent metallocenes Cp2M, which is
d
accompanied by a built-up of intrinsic strain. In contrast, the Cp rings in species B are naturally tilted as
te
the metals in the parent compounds Cp2MX2 (M = Ti, Zr, Hf) are pseudo tetrahedrally coordinated. For
Ac ce p
B the ansa moiety restricts the flexibility of the structure and, commonly, these compounds are used for olefin polymerizations, in particular, in form of their chiral derivatives [2].
The best known class of the heteroleptic sandwich complexes is that of type D, which were coined troticenophane (M = Ti), trovacenophane (M = V), and trochrocenophane (M = Cr) [3-5]. To date, examples of type E are only known for manganese and chromium [6-9]. In addition, other heteroleptic [n]metallocyclophanes are known, which are even rarer; e.g., the CH2SiR2 bridged cobalt complex [Co(4-Me3C4-CH2SiR2-5-C5H4)] equipped with a cyclobutadiene and a Cp moiety [10] and the 1
We define [n]metallocyclophanes similar as IUPAC defined cyclophanes by using the term mancude-ring systems [“rings having (formally) the maximum number of noncumulative double bonds”; IUPAC Gold book (http://goldbook.iupac.org/)]. Therefore, metal-containing “phanes” with metals sandwiched by mancude-ring systems will be referred to as metallocyclophanes. The subgroup where all -coordinated ligands are cyclopentadienyl derivatives will be called metallocenophanes; the subgroup where they are benzene derivatives will be called metalloarenophanes.
4
Page 5 of 70
(tBu2Sn)2 bridged titanium species [Ti{η5-C5H4-(SntBu2)2-η8-C8H7}] equipped with a cyclooctatetraene and a Cp ligand [11].
(ER x)n
X M X
A
(ER x) n
M
B
(ER x) n
ip t
M
C
(ER x)n
(ER x)n
M
M
M
(R xE)
m-1
n
E
F
an
D
us
M
cr
(ER x) n
M
Fig. 1. Metallocyclophanes.
Metallocyclophanes can also exhibit more than one sandwich moiety. Such species are illustrated for
d
metallocenophanes in structure F (Fig. 1), where m expresses the number of sandwich moieties and n is
te
the number of bridging atoms. Using ferrocenophanes (FCPs) as an example, larger ring systems are
Ac ce p
commonly addressed as [nm]ferrocenophanes, whereas for species with just a few repeating units, the meaning of m is usually substituted by repeating n, such as in [n.n]FCPs (m = 2) or [n.n.n]FCPs (m = 3) [12].
Whereas in [n]metallocyclophanes with relatively short bridges the metal-coordinated rings are tilted relative to each other, rings coordinated to the same metal in [nm]metallocyclophanes are coplanar and the structures of the sandwich moieties resembles those of the parent sandwich complex. However, the structures of [n]metallocyclophanes can be significantly different compared to the respective parent sandwich species. This is illustrated for [1]metallocyclophanes in Figure 2, with the common set of angles that is used to measure the degree of distortion. The distortion or intrinsic strain of such a
5
Page 6 of 70
molecule translates into stored energy, a fact that makes these species potential candidates for polymerizations.
ER x
M
us
cr
Fig. 2. Common angles to characterize distortions in [1]metallocyclophanes.
ip t
'
Shortly after the first [n]metallocyclophane was reported by Rinehart et al. in 1957 [A; M = Fe; (ERx)n =
an
(CH2)2C(=O)] [13], the first strained sandwich compound, a [2]FCP with a C2Me4 bridge, was described by the same group [14]. The first strained sandwich compounds with just one atom in bridging positions
M
were silicon-bridged [1]FCPs [ERx = SiPh2, Sifc; fc stands for Fe(H4C5)2] published by Osborne et al. in 1975 [15]. 15 Years later, the first strained bis(benzene) complexes, SiPh2-bridged
d
[1]metalloarenophanes (C) with vanadium and chromium, were published by Elschenbroich et al. [16].
te
In 2004-2005, the research groups of Tamm, Elschenbroich, and Braunschweig, respectively, described
Ac ce p
the first heteroleptic metallocyclophanes of type D (M = Ti [17], V [18], Cr [19]).
It was Manners et al. who finally realized the use of [1]FCPs as monomers for ring-opening polymerization (ROP) and showed that sila[1]ferrocenophanes give poly(ferrocenylsilane)s (PFSs) with high molecular weights [Mw (R = Me) = 5.2 105 Da relative to PS; Scheme 1] [20]. This discovery kick-started tremendous research efforts and, to date, applying the well-established methods of thermal, transition-metal-catalyzed, photolytic, and anionic ROP for FCPs is an elegant way for the preparation of well-defined metallopolymers [21].
6
Page 7 of 70
Scheme 1. Thermal ring-opening polymerization (ROP) of sila[1]ferrocenophanes [20]. 130 °C (R = Me) Fe
SiR 2 230 °C (R = Ph)
1 n
Si R2
Fe
n
PFS
ip t
The most developed strained sandwich compounds are silicon-bridged [1]FCPs (A; Fig. 1), which can undergo living polymerizations [21-23] allowing for the preparation of block copolymers. PFS with
cr
SiMe2-bridging moieties (R = Me; Scheme 1) is partially crystalline, causing block copolymers to form
us
rod- or plate-shaped micelles in block-selective solvents with PFS as the core. These micelles are living and grow to larger micelles of uniform lengths through the addition of further unimers. This
an
crystallization-driven self-assembly (CDSA) of amphiphilic block copolymers has been used recently as bottom-up approach for the fabrication of uniform nanomaterials (see [24-26] for selected recent
M
publications). Compared to silicon-bridged [1]FCPs, other strained sandwich compounds are far less
d
developed and only [1]FCPs bridged by germanium [27] and phosphorus [28-32], and [2]FCPs bridged
te
by carbon [33] could be polymerized with control over molecular weights and molecular weight distributions. Many other strained sandwich compounds could be converted into metallopolymers, but a
Ac ce p
variety of problems brought some of this research to a standstill. For example, many strained monomers are only accessible with bulky bridging moieties; however, for the targeted polymers, often nonencumbered ligands at bridging elements are needed. Another common problem is the lack of solubility of prepared metallopolymers. In order to further advance the area of ROP of strained sandwich compounds, improved monomers are needed in the future.
This review is concerned with metallocenophanes of type A and F (Fig. 1) that contain at least one group 13 element in the backbone of the bridge. Species of type B are not included as these compounds are unstrained and used for a different purpose than those reviewed here. Readers interested in species of type B are referred to the reviews by Braunschweig et al. [34] and Aldridge and Bresner [35]. As 7
Page 8 of 70
species of type D will not be reviewed, readers are directed to the current, promising developments to polymerize these metallocyclophanes [36-38]. In addition, the following list of reviews could be used as a starting point for an overview of the field of metallocyclophanes (see [3-5, 12, 21, 39-43] for selected reviews). With respect to the group 13 containing species discussed in this review, the reviews by
ip t
Braunschweig et al. dealing with boron-bridged metallocyclophanes are particularly relevant [34, 40].
cr
2. Aluminum- and gallium-bridged [1]metallocenophanes
us
Instead of discussing compounds in the group-order of these elements, we start with aluminum- and gallium-bridged [1]metallocenophanes, followed by indium-bridged species, and then close the
an
discussion with boron-bridged compounds. This order might be a surprise to a reader as the first reported group-13-element bridged [n]metallocenophanes were bora[1]ferrocenophanes, published by
M
Braunschweig and Manners et al. in 1997 [44]. After the follow-up article about these highly strained species by the same research groups appeared in 2000, this chemistry came to a standstill [45].
d
Recently, our group revitalized this dormant chemistry by applying ferrocene moieties that were
te
equipped with alkyl groups for the preparation of new bora[1]ferrocenophanes [46, 47]. These alkylated
Ac ce p
ferrocene derivatives were a rational development of our research on aluminum- and gallium-bridged [1]FCPs; hence, we start our discussion with this family of compounds. Without this background information, a discussion of our most recent discoveries in the area of boron-bridged [1]FCPs would lack the appropriate context.
2.1. The first generation: trisyl-based ligands The first strained sandwich compound bridged by a heavier group-13-element (Al–Tl) is the aluminumbridged [1]FCP 1 [48], which was prepared by a common salt-metathesis reaction [3] as illustrated in Scheme 2. Shortly after this finding, the respective gallium-bridged [1]FCP (2) and failed attempts to prepare an indium-bridged [1]FCP were described [49].
8
Page 9 of 70
Scheme 2. Synthesis of the first aluminum-bridged [1]FCP.
Li + Li
Me3 Si Me 3Si
N AlCl2
Al(Pytsi)
Fe
1
cr
2/3 tmeda
Me 2Si
ip t
Fe
In all these cases, the sterically bulky ligand Pytsi had been used (Fig. 3), which is derived from the
us
well-known trisyl ligand, C(SiMe3)3 [50], by replacing one methyl group with a pyridyl moiety 2. The Pytsi ligand had been prepared for the first time by Eaborn and Smith [51] and, subsequently, attached
an
to a variety of elements [51-55]. We had picked the Pytsi ligand as we searched for a ligand that
M
combined intramolecular donor ability with steric bulkiness for the following reasons. First, the speculation that a ligand capable of intramolecular donation was needed was based on experience in
d
using such types of ligands for heavier group 13 elements in the context of GaN deposition by single
te
source precursors for chemical vapour deposition [56-60]. Ligands equipped with one amino moiety for intramolecular coordination give facile access to monosubstituted species of the type RECl2 in one step
Ac ce p
from LiR and ECl3 (E = Al, Ga; R: mono anionic C-N ligand). The intramolecular formation of an N-E dative bond [61] reduces the Lewis acidity of the group 13 element in RECl2 so that the substitution of the second chloride to give R2ECl is significantly slower than the substitution of the first chloride in ECl3. However, the stabilized reagents RECl2 are still reactive and the remaining chloride atoms can be substituted under the right conditions [62, 63]. Second, a sterically bulky ligand was targeted as preliminary results from our laboratory obtained in 2002 from the reaction of Ar'AlCl2 [Ar' = 2(Me2NCH2)C6H4; Fig. 3] [62] with dilithioferrocene revealed that a [1.1]FCP was formed instead of the targeted [1]FCP. Based on this single result, we hoped that sterics would stabilize the targeted strained
2
The bulky ligand C(SiMe3)3 is called trisyl, as an abbreviation for tris(trimethylsilyl)methyl, and is commonly symbolized by Tsi in formulas. Therefore, we used Pytsi and Me2Ntsi for these trisyl derivatives equipped with donor moieties.
9
Page 10 of 70
alumina[1]ferrocenophanes. This thought implied that the [1]FCP was formed as an intermediate but, because of its insufficiently protected aluminum atom, dimerized to the isolated [1.1]FCP. However, there was no direct experimental evidence that a dimerization had occurred. Furthermore, the idea behind using a trisyl-based ligand was due to the fact that the C(SiMe3)3 group had been applied in the
ip t
past to stabilize elements in unusual coordination environments [50].
cr
Table 1 provides an overview of all known aluminum- and gallium-bridged [1]FCPs and
us
[1]ruthenocenophanes ([1]RCPs) and Figure 3 depicts all ligands that had been used. The alumina[1]ferrocenophane 1 [48] and the galla[1]ferrocenophane 2 [49] were isolated in yields of 31
an
and 59%, respectively, and significant amounts of ferrocene were always produced during the synthesis. In an attempt to improve the preparation of these types of aluminum and gallium compounds, the Pytsi
M
ligand was replaced by the related Me2Ntsi ligand (Fig. 3) [64]. This resulted in the isolation of the
Ac ce p
te
and 68%, respectively.
d
alumina[1]ferrocenophane 3 and the respective gallium compound 4 in significantly higher yields of 97
10
Page 11 of 70
Me 2Si
N
NMe2
Me3 Si Me 3Si
Me3 Si Me 3Si
Me 2Ntsi
Pytsi
N Me2Si
N
ip t
Me2Si
Me3 Si Me 3Si
cr
Mpysm
(2-Py)(SiMe 3)2 C
NMe2
us
NMe2
Ar' (R = H) p-tBuAr' (R = tBu) p-Me 3SiAr' (R = SiMe 3)
an
R
M
Mamx
Fig. 3. Intramolecularly coordinating ligands employed for aluminum-, gallium-, and indium-bridged
d
metallocenophanes (Mamx stands for methylaminomethyl-m-xylyl [65]; Mpysm stands for [dimethyl(2-
Ac ce p
te
pyridyl)silyl]methyl [66]).
Table 1. [1]FCPs and [1]RCPs bridged by aluminum or gallium. R' M
ER
R'
structural data a
references
1
XRD:b α = 14.9(3), /' = 39.6/40.5, = 167.9, = 94.7(2)
[48, 67]
Ga(Pytsi)
2
XRD:c α = 15.4(2) {16.4(2)}, /' = 38.3 {37.6}/38.9 {38.5}, = 167.0 {166.3}, = 92.68(13) {92.22(13)}
[49, 68]
Fe / H
Al(Me2Ntsi)
3
XRD:c α = 14.33(14), /' = 39.8/39.4, = 168.2, = 94.51(8)
[64, 67]
Fe / H
Ga(Me2Ntsi)
4
XRD:c α = 15.83(19), /' = 38.3/38.4, = 166.8, = 92.75(11)
[64, 67]
Fe / H
Al[C(SiMe3)2 (2-Py)]
5d
---
[69]
M / R'
ER
Fe / H
Al(Pytsi)
Fe / H
67,
11
Page 12 of 70
Ga[C(SiMe3)2 (2-Py)]
6d
---
[69]
Ru / H
Al(Me2Ntsi)
9
XRD:c α = 20.31(19), /' = 39.9/40.6, = 165.2, = 101.3(2)
[67]
Ru / H
Ga(Me2Ntsi)
10
XRD:c α = 20.91(19), /' = 38.6/38.6, = 163.7, = 98.42(13)
[67]
Fe / H
Al(Mamx)
15
DFT:e α = 13.22, /' = 40.61/40.89, = 169.73, = 95.13
[70]
Fe / H
Ga(Mamx)
16
DFT:e α = 15.80, /' = 38.06/37.78, = 167.30, = 91.59
[71]
Ru / H
Al(Mamx)
17
DFT:e α = 19.37, /' = 41.90/41.46, = 166.44, = 102.80
[70]
Ru / H
Ga(Mamx)
18
DFT:e α = 22.90, /' = 37.49/37.90, = 163.23, = 98.06
[70]
Fe / CHMe2
GaAr'
29
XRD: α = 16.26(9) {16.45(10)}; /' = 40.0(2) {38.9(2)}/39.1(2) {38.4(2)}; = 166.65(3) {166.78(3)}°; = 93.44(10) {92.94(10)}°
[46]
us
cr
ip t
Fe / H
M
an
[a] Angles given in degree; values in braces are from a second independent molecule of the asymmetric unit. [b] The published value of 43.1° for species 1 is wrong. The new values were obtained from the published CIF file with MERCURY [72]. [c] Only the and angles were reported; other angles were obtained from published CIF files with MERCURY [72]. [d] 5 and 6 were not isolated, but could be characterized by 1H NMR spectroscopy in reaction mixtures. [e] Data from optimized geometries on the BP86/TZ2P level of theory.
In the same publication that describes the [1]FCPs 3 and 4, the first aluminum- and gallium-bridged
d
[1]metalloarenophanes were described, namely, [1]chromarenophanes and [1]vanadarenophanes
te
equipped with (Me2Ntsi)Al or (Me2Ntsi)Ga in bridging positions [64]. Subsequently, the first
Ac ce p
[1]molybdarenophanes were prepared by applying the same bridging moieties and synthetic methodologies [73]. As usual, these strained [1]metalloarenophanes were prepared by salt-metathesis reactions using dilithiated sandwich compounds [16, 74-80]. A second motivation to replace the Pytsi moiety with a related one came from our first attempts to prepare [1]chromarenophanes from (LiH5C6)2Crtmeda and (Pytsi)ECl2 (E = Al, Ga). 1H NMR spectroscopy of reaction mixtures gave no indication that the targeted [1]chromarenophanes had been formed, instead, significant amounts of bis(benzene)chromium were present. We speculated that the pyridyl group was the source of protons to produce the parent sandwich compound from the highly basic starting material (LiH5C6)2Crtmeda [81]. As mentioned before, the synthesis of the [1]FCPs 1 and 2 always resulted in FeCp2 in reaction mixtures, which also hinted at the presence of a source of protons [64]. Consequently, a new stabilizing
12
Page 13 of 70
ligand that was similar to Pytsi (Fig.3) was searched for, and resulted, among others [64], in the known ligand Me2Ntsi (Fig. 3) [82]. The use of Me2Ntsi gave not only higher yields for [1]FCPs [64], but also access to the first aluminum- and gallium-bridged [1]metalloarenophanes [64, 73].
ip t
Altering ligands from Pytsi to Me2Ntsi resulted in a change from a five- to a four-membered ring that incorporates aluminum or gallium (Fig. 3 and Scheme 2). At the same time these two ligands were
cr
applied in our group, a similar five- to four-membered ring alteration was tested through applying the
us
(2-Py)(SiMe3)2C ligand. As illustrated in Figure 3, (2-Py)(SiMe3)2C is derived from Pytsi by cutting out the SiMe2 linkage. Salt-metathesis reactions with [(2-Py)(SiMe3)2C]EX2 (EX2 = AlCl2, AlBr2, GaCl2)
an
gave the expected [1]FCPs 5 or 6 (Table 1) as proven by 1H NMR spectroscopy, but their isolation
M
failed and instead of the monomers, low-molecular-weight polymers were obtained.
Moving one step down from iron to ruthenium to prepare [1]RCP is tempting, as the larger Ru atom will
d
result in a significant increase of the tilting of the Cp rings compared to the [1]FCP equipped with the
te
same bridging moiety ERx. An increased α angle should result in an increased strain that, on the other
Ac ce p
hand, should facilitate ROP. To date, only six [1]RCPs are described in the literature [67, 70, 83]. In 2004, Manner et al. described the first [1]RCPs which were prepared with ZrCp'2 (7) and SnMes2 (8) in bridging positions (Fig. 4) [83].
Ru
ZrCp'2 7
Ru
SnMes 2
8
Fig. 4. The first [1]RCPs (Cp' = C5H4tBu; Mes = 2,4,6-Me3C6H2) [83].
As the α angles in tin-bridged [1]FCPs (α = 13.5 – 15.7°) [46, 84-89] are very similar to those in
13
Page 14 of 70
aluminum- or gallium-bridged [1]FCPs (Table 1), it was expected that [1]RCPs could also be prepared with the latter two elements. Consequently, the use of the Me2Ntsi ligand was expanded to [1]RCP and species 9 and 10 were obtained (Table 1) [67].
ip t
The four aluminum- and gallium-bridged [1]FCPs, equipped with the Pytsi (1, 2) or the Me2Ntsi (3, 4) ligand, and the two [1]RCPs 9 and 10 were isolated through crystallization from organic solvents. The
cr
molecular structures for all these six species were solved by single crystal X-ray analysis. The common
us
distortion angles α,, , and (Fig. 2) are listed in Table 1, whereas molecular structures of selected
M
an
examples are shown in Figure 5, 6, and 7.
Si3
C21
Si1 Al1
d
Fe1
C16
te
N1
Ac ce p
Si2
Fig. 5. Molecular structure of 1 with thermal ellipsoids drawn at the 30% probability level. H atoms and half a molecule of FeCp2 are omitted for clarity. The figure was prepared with ORTEP-3 for Windows [90] using the published crystallographic data [48].
14
Page 15 of 70
Si3 C6
Fe1
Ga1 Si1 N1 C1
cr
ip t
Si2
Fig. 6. Molecular structure of 4 with thermal ellipsoids drawn at the 50% probability level. H atoms and
us
half a molecule of benzene are omitted for clarity. The figure was prepared with ORTEP-3 for Windows
M
an
[90] using the published crystallographic data [64].
Si2
Al1 N1
Si1
Si3
Ac ce p
C6
te
Ru1
d
C1
Fig. 7. Molecular structure of 9 with thermal ellipsoids drawn at the 50% probability level. H atoms and half a molecule of n-hexane are omitted for clarity. The figure was prepared with ORTEP-3 for Windows [90] using the published crystallographic data [67].
The aluminum- and gallium-bridged [1]FCPs 1–4 and [1]RCPs 9 and 10 we tested for their propensity to form metallopolymers using thermal-, anionic-, transition-metal-catalyzed-, and photocontrolled ROP methodologies [67]. However, only for one monomer using one ROP method metallopolymers were obtained, while all other attempted ROP were either sluggish or unsuccessful [67]. The 15
Page 16 of 70
galla[1]ferrocenophane 2 could be polymerized with 2 mol% of [Pd(dba)2] (dba = dibenzylideneacetone) in toluene or thf at ambient temperature or 40 °C during 48 h. Characterization of a polymer sample (toluene; 25 °C; 48 h) through Gel Permeation Chromatography (GPC) revealed an Mw of 21.1 kDa (relative to PS) with a broad molecular weight distribution (PDI = 3.0). Attempted
ip t
anionic ROP using a variety of conditions did not result in any reaction of the monomers 3, 4, and 9, or in the case of species 1 and 2 resulted in a substitution of the H atom in para position of the pyridyl
Fe
E
SiMe 3 SiMe2 N
an
Me 3Si
us
spectroscopy, however, their isolation as pure compounds failed [67].
cr
moiety of the Pytsi ligand (Fig. 8). The unexpected [1]FCPs 1Bu and 2Bu were characterized by NMR
M
1 Bu (E = Al) 2 Bu (E = Ga)
te
d
Fig. 8. Substitution products of attempted anionic ROP of [1]FCPs 1 or 2 [67].
Ac ce p
From these mainly unsuccessful polymerization attempts it was concluded that the bulkiness of the employed trisyl-based ligands, Pytsi and Me2Ntsi, prevented polymerizations from occurring. It was speculated that their bulkiness blocked the initiation step or that steric interactions of ligands with the sandwich moieties in the targeted polymers thermodynamically disfavoured their formation [67]. Even though experimental results allowed only to speculate on any details, it was clear that the bulkiness of the ligands attached to the bridging elements needed to be reduced in order to obtain aluminum- or gallium-bridged [1]metallocenophanes that were prone to ROP.
2.2. Conclusions from the first generation As mentioned before, the first attempt in our group to prepare an aluminum-bridged [1]FCP by applying
16
Page 17 of 70
Ac ce p
te
d
M
an
us
cr
ip t
Ar'AlCl2 as the reagent in salt-metathesis reactions gave an [1.1]FCP (11; Scheme 3).
17
Page 18 of 70
Scheme 3. Salt-metathesis reactions resulting in [1.1]FCPs 11-13 [91, 92] and 14 [93].
Me2 N E
A
Fe
Fe
EX2 = AlCl2, GaCl2, InI2
ip t
+ Ar'EX2
E NMe2
Li
us
Fe
cr
11 (E = Al) 12 (E = Ga) 13 (E = In)
Li
Me 2 Si Me2N
B
C(SiMe3 )2 In
X = Cl, I
Fe
Fe In
(Me 3Si)2C
M
+ (Me 2Ntsi)InX2
an
2/3 tmeda
NMe 2 Si Me 2
te
d
14
Ac ce p
As illustrated in Scheme 3A, respective reactions also gave [1.1]FCPs for gallium (12) and indium (13) [92]. As shown in Scheme 3B, employing (Me2Ntsi)InX2 in salt-metathesis reactions also resulted in a [1.1]FCP (14; Scheme 3B) [93]. Note that [1.1]metallocenophanes will be discussed in detail in Chapter 6. The latter result was a surprise, as in respective reactions with (Me2Ntsi)ECl2 (E = Al, Ga) the targeted aluminum- or gallium-bridged [1]FCPs were formed [64]; 1H NMR spectroscopy did not reveal any signs that [1.1]FCPs were also produced (Chapter 2.1). These discoveries revealed that the bulkiness of the ligand attached to the group 13 element has a significant influence on the outcome of salt-metathesis reactions of dilithioferrocene with REX2 (E = Al, Ga, In). The fact that the bulky Me2Ntsi ligand attached to indium gave a [1.1]FCP (14), whereas the same bulky ligand on aluminum or gallium gave [1]FCPs (3, 4; Table 1) can be rationalized based on structural data as illustrated in
18
Page 19 of 70
cr
ip t
Figure 9.
us
Fig. 9. Illustration of the proximity between the Me2Ntsi ligand and the ferrocene moieties (doubleheaded arrows) in the diinda[1.1]ferrocenophane 14 (left) and the galla[1]ferrocenophane 4 (right) [93].
an
HH distances in 14 are 2.15, 2.35, 2.36, and 2.38 Å (SiMe3-H); HH distances in 4 are 2.28 and
M
2.44 Å (SiMe3-H) and 2.35 and 2.53 Å (NMe2-H). Adapted with permission from [93].
d
Copyright © 2008 American Chemical Society.
te
The area of the shortest intramolecular HH distances in the indium species 14 as well as in the gallium
Ac ce p
species 4 were found between the methyl groups of the Me2Ntsi ligand and the α protons of ferrocene moieties [93]. This is indicated with double-headed arrows in Figure 9 and reveals that the available space for a bridging moiety ER in [1.1]FCPs is smaller than in respective [1]FCPs. Indium with a covalent radius of 1.44 Å is significantly larger than aluminum and gallium, which both have nearly the same size (r = 1.25 and 1.26 Å) (all three values for singly bonded, 3-fold-coordinated elements) [94]. Therefore, In-C bonds are significantly longer than respective Al- or Ga-C bonds and, consequently, for indium species, a bulky ligand R of the bridging moiety in a metallocenophane is further away from the sandwich unit. This comparison is only meaningful if the coordination geometry of the element E stays the same for the discussed species, which is the case for the aluminum, gallium, and indium compounds. The structural data as illustrated in Figure 9 lead to the conclusion that aluminum- or gallium-bridged
19
Page 20 of 70
[1.1]FCPs equipped with the Me2Ntsi ligand cannot be formed as the [1.1]FCP structure simply does not offer enough space for the bridging moieties. Instead, [1]FCPs were obtained in salt-metathesis reactions of (Me2Ntsi)AlCl2 or (Me2Ntsi)GaCl2 with dilithioferrocene as the geometry of the strained sandwich compound offers enough space. For indium, a [1]FCP as well as a [1.1]FCP is accessible , as
ip t
both types of compounds offer enough space. The fact that only the [1.1]FCP species was found has been rationalized by assuming that the intrinsic strain in a [1]FCP disfavours its formation and, at the
cr
same time, favours the formation of the unstrained [1.1]FCP [93]. The rationalization why [1.1]FCPs
us
are formed instead of the targeted [1]FCPs when steric hindrance does not play a major role, was also based on the results obtained by employing the one-armed phenyl ligands Ar' or p-tBuAr' (Fig. 3). For
an
aluminum and gallium, switching from bulky trisyl-based ligands (Pytsi, Me2Ntsi; Fig. 3) to the rather flat one-armed phenyl ligand Ar' (Fig. 3) changed the outcome of the salt-metathesis reaction from
M
[1]FCPs to [1.1]FCPs. A similar influence was also observed for bis(benzene)sandwich systems: trisylbased ligands gave [1]metalloarenophanes (E = Al, Ga; M = V, Cr, Mo) [64, 73], while the non-
te
Ac ce p
= Cr, Mo) [95].
d
encumbered ligands Ar' and p-tBuAr' (Fig. 3) gave respective [1.1]metalloarenophanes (E = Al, Ga; M
One can assume that salt-metathesis reactions are non-equilibrium reactions and, therefore, are kinetically controlled. Hence, steric interactions between groups should be discussed for transition states, in particular, for those of the rate determining steps. However, for the above discussed saltmetathesis reactions nothing is known about the kinetics and, therefore, only possible steric interactions for ground-state geometries of products were discussed. We assume that this is a meaningful approach as significant steric interactions in a ground-state geometry of a product between a ligand on the bridging element and the sandwich moiety should also to some extent be present in the transition state leading to this product. Of course, reducing the arguments for a net chemical reaction that is probably composed of different elementary reactions to just arguments of steric interactions in ground-state
20
Page 21 of 70
geometries is clearly an oversimplification. Still, this approach provided our research group with the working hypothesis that the formation of strained [1]FCPs bridged by heavier group 13 elements will only occur if sterically bulky groups block or hinder the formation of the otherwise preferred [1.1]FCPs.
ip t
Inspection of structural data and the following considerations led to the conclusion that the space available for the bridging unit ER decreases from [1]FCPs to poly(ferrocene)s to [1.1]FCPs (Fig. 10).
cr
The parallel alignment of both ferrocenediyl moieties in a [1.1]FCP forces the ligand R to be oriented in
us
such a way that possible steric interactions between R and the sandwich part are maximized. Compared to the relative orientation of the ER group in [1.1]FCPs, that in [1]FCPs is twisted by an angle of ca.
an
90°, which results in a minimization of possible steric interactions. It can be assumed that the relative orientation of the bridging units ER in respective poly(ferrocene)s are somewhere in-between that in
M
[1]FCPs and [1.1]FCPs and, therefore, the available space for ER will be in-between that in [1]FCPs
te
d
and [1.1]FCPs.
E R
Ac ce p
Fe
ER
Fe
n
[1]FCP
PF
Fe
R E Fe E R [1.1]FCP
decreasing space available for ER
Fig. 10. Space restrictions in [1]FCPs, poly(ferrocene)s, and [1.1]FCPs.
On the basis of all considerations discussed above, for the first generation of aluminum- or galliumbridged species one can conclude that strained [1]metallocenophanes were only accessible because the bulky trisyl-based ligands Pytsi or Me2Ntsi did not allow the formation of [1.1]metallocenophanes.
21
Page 22 of 70
However, for steric reasons polymerizations were blocked or sluggish as targeted poly(ferrocene)s offer less space for these bulky ligands compared to the strained monomers.
ip t
2.3. The second generation: the Mamx ligand Our tactics to advance the chemistry of strained aluminum- and gallium-bridged [1]metallocenophanes
cr
was to use steric hindrance cautiously so that [1]metallocenophanes could be accessed that were not
us
overly protected and, therefore, polymerizable. The lessons learned from the first generation of strained sandwich compounds resulted in our working hypothesis that steric hindrance is needed to block or
an
hinder the formation of [1.1]metallocenophanes so that the targeted [1]metallocenophanes are accessible. In Figure 11, the H atoms that exhibit the shortest intramolecular HH distances in the
te
d
M
Ar'Ga-bridged [1.1]FCP 12 are indicated by solid black spheres [46].
Ac ce p
N
Ga
Fe
Fe
Ga
N
Fig. 11. Illustration of the shortest HH distances in the [1.1]FCP 12 (see also Scheme 3). Adapted with permission from [46]. Copyright © 2013 Wiley-VCH.
From this molecular structure it is apparent that a bulky alkyl group adjacent to gallium on the phenyl 22
Page 23 of 70
ring would disfavour or even block the formation of a respective [1.1]FCP. In order to use this blocking tactics, group 13 dihalide species equipped with a modified Ar' ligand were needed (Fig. 12). From the ample theoretical possibilities, we settled on the use of the known Mamx ligand (Fig. 3 and Scheme 4). This bulky aryl ligand had been introduced by Yoshifuji et al. and the needed MamxBr [65, 71] is
ip t
conveniently prepared from 3,5-di-tert-butyltoluene (Scheme 4). Finally, lithium-bromine exchange gives access to the aluminum or gallium dichlorides (Mamx)ECl2 [70, 71] needed for salt-metathesis
us
cr
reactions.
NMe 2
an
EX2 R
M
Fig. 12. Proposed (RAr')EX2 species to avoid formation of [1.1]FCPs.
Ac ce p
te
d
Scheme 4. Mamx-containing aluminum and gallium species [70, 71]. NMe2 Br 1) nBuLi, -78 °C, Et2O MamxBr
2) ECl3 , Et2O
Me 2N
Me2N
E
M
M
E
n Fe Al Ga
Ru
15n 17 n 16 n 18 n
Al Ga
N Me2
Fe
Ru
15 16
17 18
Cl2E (LiC 5H 4) 2M ·tmeda (Mamx)ECl 2
The second tBu group, which is in para position with respect to aluminum or gallium in (Mamx)ECl2, is 23
Page 24 of 70
needed to facilitate the preparation of the targeted aryl bromide by employing a symmetrical rather than an unsymmetrical starting compound (Scheme 4). However, this para-tBu group does not interact sterically with the Cp moieties in metallocenophanes. As shown in Scheme 4, this approach led to [1]FCPs (15, 16) and respective [1]RCPs (17, 18). However, from the four strained species only 18
ip t
could be isolated, while the other three species underwent ROP in reaction mixtures. Still, the isolable gallium-bridged [1]RCP (18) is similarly reactive as species 15, 16, and 17, and polymerized when left
cr
stirring in the reaction mixture [70]. Even though species 15–17 were not isolable, they could be
us
unequivocally identified by 1H NMR spectroscopy from reaction mixtures, in particular, by chemical shifts and signal pattern of Cp protons, which are characteristic for Cs symmetrical aluminum- or
M d
4.4
2.06
4.00
4.6
15
N Me2
4.00 FeCp 2
4.51
Al
Ac ce p
4.72 4.70
Fe
tBu
tBu
te
3.85
4.2
4.0
ppm
2.01
an
gallium-bridged [1]FCPs with donor-stabilizing ligands (see Fig. 13).
Fig. 13. Characteristic shifts and pattern of α- and -protons of 15. 1H NMR spectrum was taken from an aliquot of the reaction mixture after ca. 30 min (C6D6). Adapted with permission from [70]. Copyright © 2012 American Chemical Society.
24
Page 25 of 70
2.3.1. Metallopolymers Molecular weights (Mw) between 8.07 and 106 kDa for polymers 15n–18n (Scheme 4) were estimated by dynamic light scattering (DLS) [70, 71]. GPC was only applied to the polyferrocenylgallane 16n and
ip t
resulted in Mw of 48 kDa (relative to PS; PDI = 3.3) a comparable value as that estimated via DLS (36 kDa) [71]. As the applied characterization methods of GPC and DLS provide only relative values of Mw
cr
the absolute values for all four polymers are unknown [70, 71]. Among the 1H NMR spectra of all four
us
polymers that of 16n stands out as it shows an unusually well resolved signal pattern for the ortho-tBu
Ar
m
Ga Fe
Ar
m
Ga Fe
B
Ar Ga
Fe
M
isotactic
A m
Ga Fe
Ar
r
r
Fe
Ga
Ar Ga
r
Ar
Ga
Fe
Ac ce p
Fe
Ar Ga
Ga Fe
syndiotactic Fe
Ar
d
Ar
C
te
heterotactic
an
group (Fig. 14).
Fig. 14. Illustration of different triads in 16n [diads indicated as r (25aceme) and m (meso)]. The gallium-coordinated NMe2 group at the aryl ligand (Mamx) is omitted for clarity. 1H NMR signal of the ortho-tBu group of 16n exhibiting pentad resolution with an approximate intensity ratio of 1 : 2 : 1 for A : B : C. Reproduced with permission from [71]. Copyright © 2010 American Chemical Society.
The splitting of the signal of the ortho-tBu group into three groups (A, B, and C; Fig. 14) is caused by a triad sensitivity. The approximate intensity ratio of 1 : 2 : 1 (A : B : C) reveals that polymer 16n has a random distribution of stereogenic gallium atoms [71]. This interpretation is further supported by the
25
Page 26 of 70
observed fine structure within the three groups. Whereas the less intense signal groups A and C each appear as three singlets with a 1 : 2 : 1 intensity ratio, the intense group B is split into a set of four singlets of similar intensity. Such an array of 10 singlets is expected for a polymer with a random distribution of pentads, as from 16 possible pentads, only 10 are distinguishable [71]. Why is the ortho-
ip t
tBu group such a sensitive probe for the tacticity of the polymer? This question was addressed when DFT calculations were employed to better understand the reactivity of the monomers 15-18 (see below)
us
cr
[70].
2.3.2. DFT calculations
an
The high reactivity of monomers 15-18 was unexpected. In particular, expected tilt angles α of 15 to 16° for the [1]FCPs 15 and 16 should not impose enough strain that prevents their isolation from reaction
M
mixtures. Furthermore, it was not understood why the only isolable [1]metallocenophane was a [1]RCP (18) and not a less strained [1]FCP. DFT calculations (BP86/TZ2P) [70] were employed to obtain
d
structural as well as thermo-chemical data. In particular, the role of the ortho-tBu group was
te
investigated. This tBu group acts as a sensitive probe for the polymer tacticity and its signal was fully
Ac ce p
resolved into all possible pentads in 16n (Fig. 14). It was speculated that this unusually high sensitivity could be caused by an intimate contact between the ortho-tBu group and ferrocene moieties of the polymer backbone.
R 2 10
M
E 20
3 1
4
6
N Me 2
5
7
Al Ga
R = tBu
R=H
Fe
Ru
Fe
Ru
17 18
15 H
17 H 18 H
15 16
16
H
Fig. 15. Calculated strained sandwich compounds to evaluate the structural effect of the ortho-tBu group of the Mamx ligand.
26
Page 27 of 70
The geometries of the eight strained sandwich compounds (Fig. 15) were optimized to provide structural data. Four of the eight species are identical to the synthesized [1]FCPs and [1]RCPs (15-18), whereas in the second set of four compounds, the ortho-tBu groups on the Mamx ligands were replaced by hydrogen (15H-18H). Comparison of geometric parameters of related pairs (e.g., 15 and 15H) revealed
ip t
the influence of the ortho-tBu group. First, tilt angles α were not very sensitive toward the ortho-tBu group (Fig. 2). If this group was absent, the angle α decreased only by 0.86 and 0.52° for FCPs and by
cr
1.08 and 0.82° for RCPs (Table 1) [70]. However, the ortho-tBu group had a significant effect on the
us
orientation of the aromatic ligand with respect to the sandwich moiety. The M−E−C1 angle decreased between 8.08 (16 to 16H) and 12.73° (17 to 17H) and, at the same time, the tilting of the aromatic ring
an
changed (torsion angle M−E−C1−C2 changed between 5.26 and 6.98°). These effects are shown in Figure 16 for the gallium-bridged [1]RCPs 18 and 18H and clearly illustrate that as the tBu group gets
C10
Ru
Ga N
18
Ac ce p
C20
te
d
M
removed, the aromatic ligand moves toward the freed-up space.
C10
Ru
Ga N
C20
18H
N
Ga
N
Ru
Ga
C1
Ru
C1
C2
C2
Fig. 16. Optimized geometries of the 18 and 18H; hydrogen atoms are omitted for clarity. The doubleheaded arrow illustrates the distance of C1 of 1.003 (18) and 1.518 Å (18H) from the plane Ru−C10−C20 (dotted line). Adapted with permission from [70]. Copyright © 2012 American Chemical Society.
27
Page 28 of 70
To address the question if the significant structural changes imposed by the ortho-tBu group also significantly increases the strain of the [1]metallocenophanes requires quantification of changes in energy. Intrinsic strain of a compound can be best described by the enthalpy of a reaction where the strained species is transformed into unstrained species. Therefore, thermodynamic data of the chemical
cr
ip t
reactions shown in Scheme 5 were calculated.
us
Scheme 5. Chemical reactions applied to calculate strain in [1]FCPs and [1]RCPs [70].
+ MCp2
E
M (1)
an
Me 2N R
(1)
R
E = Al, Ga
19
M = Fe, Ru N Me 2
+ 2 H2
M
Ac ce p
15 - 18 15 H-18 H
R = tBu, H
d
E
te
M
M
M(2)
+
R
H2E
(2) N Me 2 20
The isodesmic reaction 1 should be ideal, as calculated reaction enthalpies can be directly taken as the strain of the starting [1]metallocenophanes. However, the size and floppiness of the bis(metallocenyl) compounds (19) precluded reliable frequency calculations and only ΔE values could be determined. Furthermore, it turned out that all four bis(metallocenyl) compounds equipped with the ortho-tBu group (19; R = tBu) were still strained. The ortho-tBu group pointing toward the metallocenyl unit M(2) forced it to be bent away; in contrast, in species 19 with R = H the metallocenyl unit M(2) was bent in the opposite direction. Overall, reaction 1 could not be used to evaluate the strain of the starting
28
Page 29 of 70
[1]metallocenophanes [70].
In contrast to the first reaction in Scheme 5, the hydrogenation reaction (eq. 2) is not isodesmic, and calculated enthalpies are not equal to the intrinsic strain of respective metallocenophanes. However,
ip t
comparison of the hydrogenation of a tBu containing species (e.g. 15) with that of the respective metallocenophane with R = H (e.g., 15H) resulted in differences in the calculated thermodynamic values
cr
(ΔΔE, ΔΔH, and ΔΔG) that directly provided a measure of the thermodynamic effect of the ortho-tBu
us
group. In all four cases, hydrogenations of species equipped with the tBu group resulted in a larger release of energy; i.e., increases of enthalpies (ΔΔH) between -4.80 and -6.33 kcal/mol were found.
an
That means that the tBu group increased the intrinsic strain in the aluminum- and gallium-bridged [1]FCPs and [1]RCPs on average by 5.5 kcal/mol. Compared to dimethylsila[1]ferrocenophane, the text
M
book example of a strained sandwich compound, with a measured enthalpy of -19 kcal/mol (ΔHROP) [20], the increase of strain caused by the ortho-tBu group of 5.5 kcal/mol is indeed substantial. This was
te
d
the first time that such an effect had been revealed in metallocenophane chemistry [70].
Ac ce p
2.4. Conclusions from the second generation
The application of the Mamx ligand to stabilize aluminum or gallium in bridging position of [1]FCPs or [1]RCPs was a partial success as summarized here: 1) The working hypothesis that the use of a bulky group in ortho-position of the Ar' ligand will block or hinder the formation of [1.1]metallocenophanes to give the targeted [1]metallocenophanes withstood the test.
2) In contrast to the first generation of [1]metallocenophanes, those of the second generation were reactive and gave metallopolymers. The Mamx ligand significantly stabilized the group 13 element so that polymers could be handled under ambient conditions and purified by precipitation into methanol [70]. This stabilizing effect had been used recently for the synthesis of other small molecules and
29
Page 30 of 70
polymers equipped with the (Mamx)Ga moiety [96, 97]. 3) The ortho-tBu group acts as a very sensitive NMR probe and revealed a random tacticity of 16n. DFT calculations shed some light on this surprising result and uncovered that the tBu group pushes against a ferrocenyl moiety in model compounds (19; Scheme 5). Based on this result, one can assume that the
ip t
available space in a polymer is similarly restricted so that the tBu group sterically interacts with ferrocenediyl moieties. Such interactions are probably the reasons for the pentad sensitivity.
cr
4) The average increase in strain in monomers caused by the ortho-tBu group is 5.5 kcal/mol. One
us
would be tempted to say that this increased strain is resulting in the unexpected high reactivity. However, the ortho-tBu group imposes a similar strain on the resulting polymers so that overall this
an
effect cancels out. Therefore, it was concluded that kinetics must govern the reactivity of 15-18. It was speculated that small amounts of dilithioferrocene might initiate the ROP of monomers in reaction
mechanism of ROP still in the dark [70].
M
mixtures, but all attempts to get experimental data to support this speculation failed, leaving the
d
5) Out of the four strained monomers, only species 18 could be isolated, which was achieved by
te
precipitation from the reaction mixture at low temperature. Further purification of 18 through
Ac ce p
crystallization failed. The three [1]metallocenophanes 15-17 could not even be isolated. In order to have a chance to prepare a new polymer for a potential niche application, a control over molecular weights is required. However, the therefore needed living ROP processes require monomers of high purity. As our second generation of monomers failed to fulfill this requirement, we proceeded to the next generation of strained sandwich compounds.
2.5. The third generation For the second generation of strained sandwich compounds our working hypothesis that steric hindrance is needed to block or hinder the formation of [1.1]metallocenophanes so that the targeted [1]metallocenophanes are accessible was put into action by increasing the bulkiness of the ligand on the
30
Page 31 of 70
bridging element. This led to the use of the Mamx ligand as discussed and illustrated in Chapter 2.3 (Fig. 11). For the third generation of strained sandwich compounds we increased the bulkiness of the ferrocene moiety. As evident from the ORTEP plot of the [1.1]FCP shown in Figure 11, bulky alkyl groups in place of the H atoms in positions of the Cp rings should render the resulting [1.1]FCP non-
ip t
accessible. As ferrocene moieties are introduced by dilithioferrocene, the preparation of the third generation required -substituted 1,1’-dilithioferrocene derivatives, which we wanted to prepare from
cr
respective bromides through Li/Br exchange. Hence, we searched for a precursor of type 21 (Fig. 17)
R1
an
methodologies resulting in these type of species are rare.
Br
Br
R4
d
21
M
R2 Fe R 3
us
that was equipped with one, two, three, or four alkyl groups in positions. However, synthetic
Ac ce p
te
Fig. 17. Proposed precursor for the preparation of -substituted dilithioferrocene derivatives.
Scheme 6. Preparation of (Sp,Sp)-1,1'-dibromo-2,2'-di(isopropyl)ferrocene (28) [46, 98].
31
Page 32 of 70
(S)
Ph Ph
O N
B Me 60 mol%
AcCl / AlCl3 Fe
22
OAc
(R)
HNMe2
Fe 2) (CCl2 Br)2, thf, -78 °C
Br
Fe
MeOH/H2 O (R)
(R)
25
NMe2
(R)
NMe2
24
OAc
Ac 2O,
(R)
Br (R)
27
Br
AlMe 3 Fe
CH2 Cl2 , -78 °C
Br
28
OAc
an
Br Fe
M
70 °C
OAc
OH
us
Br
(R)
Ac2 O, pyridine
(R)
1) 4 equiv nBuLi, Et2 O, r.t. 12 h
Fe
26
23
O NMe2
NMe2 (R)
Fe
BH 3·SMe 2 , thf, 0 °C
ip t
CH2Cl2
cr
Fe
OH (R)
O
te
d
We used a well-known approach that was developed by Ugi et al. [99], where it was shown that the – C*H(Me)NMe2 group on one Cp rings is an excellent directing group for diastereoselective lithiation.
Ac ce p
Nearly three decades later, Knochel et al. used the enantioselective CBS reduction [100, 101] (22 to 23; Scheme 6) and extended the known “Ugi amine” chemistry to prepare the “double Ugi amine” 25 among other enantiomerically pure, C2 symmetric ferrocene derivatives [98]. Through a lithiationbromination sequence, amine 25 was diastereoselectively converted into the known species 26 [46, 102]. After conversion of 26 to the acetate 27, the central chirality was removed to leave the planarchiral ferrocene framework in 28 in form of the (Sp,Sp)-isomer behind [46].
The first attempts to lithiate species 28 using the standard conditions of two eq. of nBuLi at -78 °C in thf solutions did not result in a clean lithium-bromine exchange [46]. However, applying the method that Bailey et al. developed for aromatic bromides [103], a quantitative lithium-bromine exchange was 32
Page 33 of 70
accomplished at 0 °C using nBuLi in the solvent mixture of thf and hexanes (1 : 9) (Scheme 7). Even though the solution mainly consists of hexanes, the dilithio species of 28 remains dissolved and does not precipitate. Addition of Ar'GaCl2 to the in situ prepared dilithioferrocene derivative of 28 gave the targeted gallium-bridged [1]FCP 29 (Scheme 7). As discussed before, the respective salt-metathesis
ip t
reaction with dilithioferrocene-tmeda gave the unwanted [1.1]FCP 12 (Scheme 3).
1) 2 equiv nBuLi, 0 °C, thf : hexanes (1 : 9), 30 min
Br
2) 1 equiv Ar'GaCl2
GaAr'
an
Fe
Fe
us
Br
cr
Scheme 7. Synthesis of the gallium-bridged [1]FCP 29 [46].
29
M
28
Checking the reaction mixture by 1H NMR spectroscopy (Fig. 18) revealed that the salt-metathesis
d
reaction resulting in 29 is quite clean as, in addition to the signals caused by this C1 symmetrical
te
product, only weak signals were detected (Fig. 18). The strained [1]FCP was isolated by crystallization
Ac ce p
from the reaction mixture in a yield of 59% [46]. Its molecular structure was solved by single crystal Xray analysis (Fig. 19) and the determined tilt angles for the two independent molecules [16.26(9) and 16.45(10)] and the other common distortion angles , , and are very similar to those of the first generation of gallium-bridged [1]FCPs (see Table 1).
33
Page 34 of 70
Fe Ga N Me2 29
cr
ip t
*
us
Fig. 18. Cp range of the 1H NMR spectrum of 29, taken from an aliquot of the reaction mixture,
an
approximately 5 min after the addition of Ar'GaCl2. The doublet marked by an asterisk is caused by one of the two methylene protons (for the full spectrum see Fig. S7 in reference [46]). Adapted with
C7 C6 C17
Fe1
Ac ce p
Ga1
te
d
M
permission from [46]. Copyright © 2013 Wiley-VCH.
C1
N1
C2
Fig. 19. Molecular structure of 29 with thermal ellipsoids drawn at the 50% probability level. Hydrogen atoms are omitted for clarity. One of two independent molecules is shown. The figure was prepared with ORTEP-3 for Windows [90] using the published crystallographic data [46].
The DSC thermogram of 29 looks as expected for a strained sandwich compound (Fig. 20). However, the endothermic melting peak (onset ca. 187 C) partly overlaps with the exothermic peak (Tmax = 201 34
Page 35 of 70
C); the latter is indicative of ROP. Because of the peak overlap, the HROP could not be measured precisely, but could be estimated to be more negative than -33 kJ mol-1 [46]. This value is very similar to that of similarly tilted tin-bridged [1]FCP tBu2Snfc [ = 14.1(2)] where HROP of -36(±9) kJ mol-1
M
an
us
cr
ip t
was determined [85].
te
[46]. Copyright © 2013 Wiley-VCH.
d
Fig. 20. Differential scanning calorimetry (DSC) thermogram of 29. Reproduced with permission from
Ac ce p
In summary, employing isopropyl groups in -position on Cp rings successfully suppressed the formation of [1.1]FCP and resulted in an isolable and reactive [1]FCP (29). As the enantiomerically pure dibromoferrocene derivative 28 could cleanly be transformed into its dilithio derivative (Scheme 7), we explored its applicability for the synthesis of new [1]FCPs with a variety of bridging elements. To date, elements of group 13 (B [46, 47], Ga [46], In [104]), 14 (Si, Sn [46]), and 15 (P [105]) were successfully incorporated into the bridging position. Some of these results are discussed in Chapters 3 and 4.
3. Indium-bridged [1]metallocenophanes
35
Page 36 of 70
In the first attempt to prepare an inda[1]ferrocenophane the same tactic was followed that led to the first aluminum- and gallium-bridged [1]FCPs (1 and 2; Scheme 2 and Table 1) [48, 49, 68]. However, the 1,1'-disubstituted ferrocene species 30 (Fig. 21) was the only isolable product from the salt metathesis between (Pytsi)InCl2 and (LiC4H4)2Fe2/3tmeda [49]. Changing the stoichiometry of the reaction from
ip t
1 : 1 to 2 : 1 resulted in species 30 in a yield of 57%. At the time of this discovery it could not be clarified why species 30 instead of the targeted [1]FCP was found. As the starting dichloride
cr
(Pytsi)InCl2 formed dimers in the solid state with an In-(-Cl)2-In moiety similar to that found in
us
species 30, the authors speculated that a monomeric indium dihalide might be required for the
an
preparation of an indium-bridged [1]FCP [49].
M
Si23
Si21 Si22
In2
Cl2
d
N21 Fe1
te
Cl1
In1
Ac ce p
N1
Si2
Si1
Si3
Fig. 21. Molecular structure of 1,1'-[(Pytsi)InCl]2fc (30) with thermal ellipsoids drawn at the 50% probability level. Hydrogen atoms and solvent molecules are omitted for clarity. The indium-chlorine bridge is asymmetric [In1-Cl1 = 2.8322(15), In2-Cl2 = 2.8655(15), In1-Cl2 = 2.5087(14), In2-Cl1 = 2.5072(15) Å]. The figure was prepared with ORTEP-3 for Windows [90] using the published crystallographic data [49]. As discussed in Chapter 2.2, employing (Me2Ntsi)InX2 in salt-metathesis reactions resulted in the 36
Page 37 of 70
[1.1]FCP 14 (Scheme 3B) instead of the targeted [1]FCP [93]. This result helped to rationalize the influence of the bulkiness of the bridging moiety on the outcome of the salt-metathesis reaction and led to the working hypothesis that steric hindrance is needed to block or hinder the formation of [1.1]metallocenophanes so that the targeted [1]metallocenophanes are accessible (Chapter 2.3).
ip t
Because indium is significantly larger than aluminum or gallium, a ligand R on indium will result in a smaller steric interaction with a sandwich moiety than the same ligand on aluminum or gallium.
cr
Scheme 8. Synthesis of the first indium-bridged [1]FCPs [104].
us
tBu Me 2N
tBu
In
Fe
In
R' = H
N Me2
+
an
tBu
tBu
+ (Mamx)InCl 2
Fe
M
31 (19%)
+
In tBu
Fe
In
tBu
A
Fe
tBu
Me 2N
n
NMe2
31 2 (23%)
31 n (32%)
d
tBu
Fe Li R'
B + Ar'InCl2
te
Li
R' = iPr
iPr
In
Fe
iPr
+
N Me 2
R' = iPr
N Me 2 iPr 33
Fe
In
tBu
In
Fe
iPr
NMe2 32 2
iPr tBu
+ (Mamx)InCl 2
Fe
iPr
32
C
In
iPr
iPr
Ac ce p
R'
Me2 N
Fe
In (Mamx)
n
33 n (33%)
Scheme 8 shows our recent attempts to prepare inda[1]ferrocenophanes where we combined the 37
Page 38 of 70
chemistry of our second and third generation of aluminum and gallium species [104]. The reaction of (LiC4H4)2Fe2/3tmeda and (Mamx)InCl2 gave a separable mixture of monomer (31), dimer (312), and oligomers (31n) (Scheme 8A). Similarly, reaction of the iPr-substituted dilithioferrocene derivative (generated in situ from dibromide 28) with Ar'InCl2 gave an inseparable mixture from which the
ip t
monomer 32 and its dimer 322 were clearly identified by 1H NMR spectroscopy as the main products (Scheme 8B). Combining the bulky Mamx ligand and iPr groups on ferrocene resulted in a relatively
cr
clean conversion into the inda[1]ferrocenophane 33 (Scheme 8C). However, all attempts of isolation
us
failed as 33 polymerized in the reaction mixture resulting in the poly(ferrocenylindigane) 33n (DLS analysis: Rh = 2.4 0.1 nm; Mw = 24 2 kDa; see [104]). This reactivity is reminiscent of our second
an
generation of the strained sandwich compounds (15-18; Scheme 4) that were equipped with the Mamx ligand, but had no iPr groups on Cp rings. That 33 readily polymerized is still surprising if one
M
considers that it should be significantly less strained than the [1]FCPs (Mamx)Alfc (15) and
d
(Mamx)Gafc (16).
R' Fe
R'
InR
Ac ce p
te
Table 2. Calculated distortion angles in indium-bridged [1]FCPs.
structural data a
references
31
α = 11.44, /' = 37.60/37.62, = 169.99, = 86.24
[104]
Ar'
32 b
α = 10.96, /' = 39.03/38.85, = 170.37, = 87.49
[104]
CHMe2
Mamx
33
α = 11.42, /' = 37.62/37.79, = 170.22, = 86.12
[104]
H
Ar'
34 c
α = 11.18, /' = 38.42/38.27, = 170.12, = 87.08
[104]
R'
R
H
Mamx
CHMe2
[a] Angles given in degree; data from optimized geometries on the BP86/TZ2P level of theory [104]. [b] 32 was not isolated, but could be characterized by 1H NMR spectroscopy in a mixture with 312 (Scheme 8). [c] The hypothetical species 34 (Ar'Infc) was included in the DFT calculations, but had not been synthesized (Scheme 8).
38
Page 39 of 70
Because crystals of the new indium-bridged [1]FCPs could not be obtained, DFT calculations were performed to get structural data. In order to deduce the effects of the tBu and the iPr groups, the hypothetical inda[1]ferrocenophane 34 (Table 2) was included in this study. As shown in Table 2, the distortion angles do not change significantly; e.g., the tilt angles were calculated between 10.96 and
ip t
11.44 for 31-34.
cr
The main difference between the structures of the four compounds 31-34 is due to the relative
us
orientation of the aromatic ligand (Mamx or Ar') with respect to the ferrocene moiety. Similar as it was found for the set of [1]FCPs (15, 16, 15H, 16H) and [1]RCPs (17, 18, 17H, 18H) (Fig. 15 and Fig. 16), the
an
relative orientation changes with the degree of substitution. Overall the structural effects are significant but less pronounced for the indium-bridged [1]FCPs (31-34) compared to the aluminum- and gallium-
M
bridged species (15-18; 15H-18H) [104]. In order to address the question if the iPr and tBu groups influence the strain of the [1]FCPs, the thermochemistry of a hydrogenolysis reaction was investigated
d
(Scheme 9). As discussed before (eq. 2 in Scheme 5), such a hydrogenolysis is a non-isodesmic reaction
te
and only a comparison of two reactions that only differ in the presence or absence of one type of alkyl
Ac ce p
group (iPr or tBu) gives meaningful data that reveals the thermodynamic effect of this particular group. This gave Ho298 between -0.17 to -4.39 kcal mol-1 meaning that in all cases the presence of the particular set of alkyl groups resulted in an increased exothermy of the hydrogenolysis. Species 33 as the most encumbered and structurally distorted species showed with -2.83 (effect of iPr groups) and 4.39 kcal mol-1 (effect of tBu groups) the largest Ho298 values. That means that the alkyl groups significantly increase the intrinsic strain of the inda[1]ferrocenophane 33, which might be causing its unexpectedly high reactivity. Usually, a tilt angle α around 11° is an expression of an amount of strain that is insufficient for ROP. Obviously, compound 33 is an exception.
39
Page 40 of 70
Scheme 9. Hydrogenolysis reaction to evaluate strain in inda[1]ferrocenophanes [104].
Fe
In
+ N Me2
R'
2 H2
Fe
+
H2In
R'
R'
R
H iPr iPr H
tBu H tBu H
R
R N Me 2
us
cr
31 32 33 34
R'
R
ip t
R' R
4. Boron-bridged [1]metallocenophanes
an
In 1997, the first boron-bridged [1]FCPs (36, 37; Table 3) were published by Braunschweig and
M
Manners et al. [44]. This report made some waves in the “phane” community as these species were the first reported [1]FCPs with an element from the second period in bridging position. Second, among the
d
two species prepared, the (Me3Si)2NB-bridged species 37 was structurally characterized revealing a new
te
record for the tilting of the Cp rings. Species 37 with = 32.4(2)° (Fig. 22 and Table 3) replaced the record holder at the time, which was the sulphur-bridged [1]FCP with = 31.05(10)° [106, 107]. To
Ac ce p
date, a total of nine boron-bridged [1]FCPs are known in literature (35-43; Table 3), all of which are equipped with amino groups at boron [44-47].
40
Page 41 of 70
Table 3. Boron-bridged [1]FCPs. R Fe
R1 B N R2
R
R
NR1R2
structural data a
H
NiPr2
35
α = 31.0(2) {31.4(2)}, /' = 35.3(3) {35.6(3)}/34.3(3) {36.1(3)}, = 155.9(2) {156.1(2)}, = 102.0(3) {103.2(3)}
[45]
H
N(SiMe3)tBu
36
---
[44, 45]
H
N(SiMe3)2
37
α = 32.4(2), /' = 33.7(2)/34.0(2), = 155.2(2), = 100.1
CHMe2
NEt2
38 b
---
CHMe2
NiPr2
39
α = 31.9(2), /' = 36(1)/35(1), = 155(2), = 103.0(3)
CHMe2
N(SiMe3)tBu
40 b
---
CHEt2
NEt2
41
---
CHEt2
NiPr2
42
---
CHEt2
N(SiMe3)tBu
43
---
M
an
us
cr
ip t
references
[44, 45]
[47] [46] [47] [47] [47] [47]
Ac ce p
te
d
[a] Angles given in degree; values in braces are from a second independent molecule of the asymmetric unit. [b] Isolated as impure oils [47].
Si1
C10
B1
N1
Fe1
C20
Si2
Fig. 22. Molecular structure of the first boron-bridged [1]FCP (37) with thermal ellipsoids drawn at the 50% probability level. Hydrogen atoms are omitted for clarity. The figure was prepared with ORTEP-3
41
Page 42 of 70
for Windows [90] using the published crystallographic data [44]. After the first report on boron-bridged [1]FCPs appeared in 1997 [44], a follow-up article was published 3 years later [45] describing the synthesis of another [1]FCP (35). The N(SiMe3)2 (37) as well as the N(iPr)2-substituted [1]FCP (35) were structurally characterized revealing highly distorted sandwich
ip t
species (Table 3 and Fig. 22). Nearly 20 years after its publication, species 37 is still the FCP with the most tilted Cp rings [ = 32.4(2)° and = 155.2(2)°]. Recently, our group prepared a series of six
cr
species equipped with three different amino groups based on our third generation of strained FCPs (see
us
Chapter 2.5). These six enantiomerically pure species fall into two sets: one with isopropyl groups (3840; Table 3) and one with 3-pentyl groups (41-43; Table 3) on Cp rings. Out of these six species the
an
molecular structure of only the iPr2NB-bridged compound 39 equipped with iPr groups on ferrocene could be determined in the solid state (Fig. 23). As shown in Table 3, the values for the distortion angles
te
d
M
, , , and , respectively, are very similar for 35, 37, and 39.
C2
Ac ce p
C1
B1
N1
C6
C7
Fig. 23. Molecular structure of the enantiomerically pure [1]FCP 39 with thermal ellipsoids drawn at the 50% probability level. Hydrogen atoms are omitted for clarity. The figure was prepared with ORTEP-3 for Windows [90] using the published crystallographic data [46].
42
Page 43 of 70
4.1. Synthesis and mechanistic insights Similar to all other group-13-bridged [1]FCPs, the boron-bridged compounds 35-43 were prepared by the common salt-metathesis approach using a dilithiated sandwich compound and respective
ip t
amino(dichloro)boranes. Whereas the 35-37 were isolated in yields of 35-44% [45], yields for the chiral [1]FCPs varied between 48% (43) and 65% (42) [47]. The latter yields resulted from optimizing the
cr
reaction conditions for the salt metathesis as follows. Species 39 was the first boron-bridged [1]FCP we
us
made in our group and it was isolated in a yield of only 20% [46]. As shown in Scheme 10, similar as the preparation of the gallium-bridged [1]FCP 29 (Scheme 7), the chiral dilithioferrocene was prepared
an
in situ from the precursor 28 by Li/Br exchange at 0 °C (Scheme 10). After 30 minutes stirring at this temperature, the cold bath was removed and, simultaneously, the addition of a solution of iPr2NBCl2 in
M
hexanes over 10 min was started [46].
Fe Br 28
1
iPr
1) 2 equiv nBuLi, 0 °C, thf : hexanes (1 : 9), 30 min
Ac ce p
Br
te
d
Scheme 10. Preparation of bora[1]ferrocenophane 39.
Fe
B=NiPr2
2) 1 equiv Cl2 BNiPr2
39
B
+
Fe iPr Cl B NiPr2
NiPr 2
Cl 45
H NMR data of the reaction mixture showed the presence of the targeted [1]FCP 39 and the unwanted
bis(boryl)ferrocene 45. The detected ratio for 39 : 45 of 1.0 : 3.5 is comparable to the isolated yield of 20% for 39 [46]. The yield of 39 was improved to 53% by heating the solution of the in situ prepared dilithioferrocene derivative first to 50 °C, before addition of the dichloride [47]. This initiated a series of experiments by combining three different amino(dichloro)boranes and two different dilithioferrocene species in salt-metathesis reactions. Whereas employing iPr2NBCl2 and tBu(Me3Si)NBCl2 to prepare
43
Page 44 of 70
respective [1]FCPs led to the formation of significant amounts of bis(boryl)ferrocenes, the Et2NBbridged [1]FCPs 38 and 41 formed with such high conversions that the small peaks of byproducts in the proton NMR spectra could not reliably be assigned to the expected bis(boryl)ferrocenes. Therefore, the effect of temperature on the product distributions was investigated for the synthesis of the [1]FCPs 39,
ip t
40, 42, and 43 (Table 4).
cr
Table 4. Measured product ratios between [1]FCPs and bis(boryl)ferrocenes (Scheme 11) [47].a
us
reaction temperature NR1R2
product ratio
0 °C r.t.
50 °C
CHMe2
NiPr2
39 : 45
1.0 : 2.4
1.0 : 0.59
CHMe2
NtBu(SiMe3)2
40 : 46
1.0 : 1.4
CHEt2
NiPr2
42 : 48
1.0 : 0.81
CHEt2
NtBu(SiMe3)2
43 : 49
1.0 : 0.52
M
an
CHR2
1.0 : 0.47 1.0 : 0.30 1.0 : 0.12
te
d
[a] Approximate ratios determined by 1H NMR spectroscopy from aliquots of the reaction mixture, taken ca. 20 min after the addition of R1R2NBCl2.
In all four cases, increasing the temperature to 50 °C significantly favoured the formation of the targeted
Ac ce p
[1]FCPs (Table 4). This effect was explained by an obvious reaction mechanism as illustrated in Scheme 11 [47]. Both products stem from a common intermediate (INT), which can either intramolecularly react to form the targeted [1]FCPs or intermolecularly react with additional amino(dichloro)borane. The later reaction channel will be disfavoured by low concentrations of the amino(dichloro)borane and, therefore, slow addition of a dilute solution of this reagent is important. The fact that increased reaction temperature favours the formation of [1]FCPs more strongly reveals that this pathway has the highest activation barrier. This makes sense as some of the strain present in the [1]FCPs must already be built-up in the transition state leading to this species. Based on this insight, the synthesis of the 35 (iPr2NBfc) was reinvestigated using the optimized conditions and the reported yield of 38% [45] was improved to 74% [47]. 44
Page 45 of 70
CHR2 Li(solv) x
k1
+ Cl2 BNR1R 2
R2 HC (solv) x Li
Fe R 2HC (solv) xLi
INT
k3
k2
cr
Fe
CHR2 NR 1R 2 B Cl
R1 Fe
B N
an
R2 HC Cl B
CHR2
us
+ Cl2BNR1R 2 CHR2 NR1 R2 B Fe Cl
ip t
Scheme 11. Proposed reaction mechanism [47].
R2
CHR2
NR 1R 2 44-49
M
38-43
d
In addition to the effect of temperature, data in Table 4 reveal the influence of the different alkyl groups
te
on ferrocene on product ratios (isopropyl compared to 3-pentyl). For all the respective cases, the 3pentyl substituted sandwich species resulted in higher conversions to the targeted [1]FCPs; for example,
Ac ce p
the product ratio at a reaction temperature of 50 °C for the tBu(SiMe3)N-substituted species are 1.0 : 0.47 for iPr and 1.0 : 0.12 for 3-pentyl at ferrocene (Table 4). This effect had been attributed to a steric influence of the CHR2 group (R = Me, Et) on the rate of substitution on R1R2NBCl2. Based on crystallographic data, the most stable conformation of the alkyl groups CHR2 (R = Me, Et) on the ferrocene moiety shows one of the two R groups approximately in the same plane as a Cp ring, whereas the second R group is oriented away from iron and nearly perpendicular to a Cp ring [47]. This is illustrated in Figure 24 with a Newman projection for one CHR2 group at a Cp ring.
45
Page 46 of 70
R R
H Fe
Fig. 24. Newman projection to illustrate the conformation of CHR2 groups on Cp rings.
ip t
The rate constant k2 should be unaffected by the type of alkyl group on the ferrocene moiety, because in
cr
both cases (iPr and 3-pentyl) hydrogen points to the inner area of the sandwich compound where the ring closure happens (Scheme 11). However, the rate constants k1 and k3 will be influenced by the type
us
of alkyl group. For steric reasons one must assume that a borane molecule R1R2NBCl2 will approach a lithiated Cp ring from the least hindered side, which will be opposite to the Fe atom. That means that
an
one R group of the CHR2 substituent will point toward the direction of the incoming borane and, hence,
M
R = Et will reduce k1 and k3 more strongly than R = Me [47]. In short, 3-pentyl groups suppress the formation of bis(boryl)ferrocene derivatives (k3; Scheme 11) more effectively than iPr groups with the
te
d
overall effect that 3-pentyl groups favour the formation of [1]FCPs.
4.2. Reactivity of bora[1]ferrocenophanes
Ac ce p
One of the main motivations to prepare strained sandwich compounds is to use their potential to access new metallopolymers through ROP methodologies. For bora[1]ferrocenophanes one would expect that these highly strained molecules will show a high propensity towards polymerization. Unfortunately, these expectations have not been met yet. As transition-metal-catalyzed ROP works well for silicon- and germanium-bridged [1]FCPs [21, 108-110], Braunschweig and Manners et al. investigated the reactivity of 35-37 towards transition metals. However, employing catalytic amounts of Karstedt’s catalyst did not yield polymers [45]. Even Pt(cod)2, a reagent known to result in just a single insertion of a Pt(cod) moiety into the C-Si bond of Me2Sifc [110], did not result in a reaction with monomer 36 when applied in an equimolar amount [45].
46
Page 47 of 70
However, reactions with carbonyl complexes led to some surprising results. As shown in Scheme 12, insertion of metal carbonyl fragments into Fe-Cp bonds yielded unstrained species. Species 50 was either obtained by using Fe(CO)5 or Fe2(CO)9 in isolated yields of 26 and 64%, respectively [45]. This complex is a boron-bridged analogue of the well-known dimer [Cp(OC)Fe(-CO)]2; a Me2Si-bridged
ip t
analogue of 50 was already published in 1973 [111]. A similar insertion reaction of a metal carbonyl moiety into the Fe-Cp bond had been observed in the reaction with Co2(CO)8 that led to the complex 51
cr
(Scheme 12). Both products were characterized by single crystal X-ray analysis [45]. Similar reactions
us
with carbonyl complexes are not reported for other [1]FCPs, and it is likely that a significant weakening of Fe-Cp bonds, as caused by the high degree of tilting of the Cp rings in bora[1]ferrocenophanes, is
an
needed for such reactions to occur.
B=N iPr
or Fe 2(CO) 9
tBu
Fe
B=N
Co2(CO) 8
iPr B=N iPr
Fe
50
(OC) 2 Co
SiMe3
36
Fe
OC
Ac ce p
35
OC OC
d
Fe
OC
Fe(CO) 5, h
te
iPr
M
Scheme 12. Reactions of bora[1]ferrocenophanes with metal carbonyl complexes.
(OC) 4Co
(CO) 2 Fe
tBu B=N SiMe 3 51
The three [1]FCPs 35-37 were polymerized thermally at 180 °C (37) and 200 °C (35-36). While 36 gave only insoluble material, a minor fraction of the product mixture obtained from monomer 37 was soluble in C6D6. 1H NMR analysis of the dissolved fraction showed signals that were interpreted as being caused by the dimer [(Me3Si)2NBfc]2 [45]. Thermal ROP of 35 resulted in a mixture that was completely soluble in organic solvents. Precipitation from toluene into hexanes gave material (34% 47
Page 48 of 70
yield) that was the targeted polymer as concluded from NMR data. However, this material did not give detectable signals in dynamic light scattering measurements, meaning that the hydrodynamic radius of this polymer was below the detection limit of 2.3 nm, a value that would be around Mn of 9 kDa for PFS
ip t
[45].
From the six chiral [1]FCPs, the two 3-pentyl substituted species 41 and 42 were selected for thermal
cr
ROP as their DSC thermograms looked promising. The obtained glassy solids [TROP = 240 °C (41), 260
us
°C (42)] were dissolved in thf leaving elemental iron particles behind. Repeated precipitations and filtrations resulted in isolation of polymeric materials in 73% (41n) and 70% (42n) yields [47]. GPC
an
analysis gave estimated values of Mw of 10 kDa (relative to PS) for the main elution peak for both polymers. However, CHN analysis did not give the expected values revealing that isolated materials
M
were not the targeted linear polymers in pure form (Scheme 13).
1
R
R
Fe
B NR1 R2
Ac ce p
Fe
R B N R2
te
R
R
d
Scheme 13. Thermal ROP of bora[1]ferrocenophanes.
n
35 -37 (R = H)
41, 42 (R = CHEt2 )
48
Page 49 of 70
ip t cr
us
Fig. 25. DSC thermogram of 42 (heating rate: 10 °C min-1; TROP (onset) = 222 °C; HROP = -63(±5)
an
kJ mol-1). Reproduced with permission from [47]. Copyright © 2014 Wiley-VCH.
M
DSC was used to determine the exothermy of the ROP for 35-37 [45] and the 3-pentyl substituted species 41-43 [47]. Figure 25 displays the DSC thermogram of 42, which shows the expected features;
d
i.e., an endothermic melting peak followed by an exothermic peak indicating the occurrence of ROP.
te
However, for bora[1]ferrocenophanes 35 and 36 overlap of the two peaks prevented the determination
Ac ce p
of HROP [45], whereas for species 43 multiple exothermic peaks were found with the main peak being very broad [47]. These investigations resulted in HROP values of -95 (37), -75(±5) (41), and -63(±5) (42) kJ mol-1. In addition, as we had improved the synthesis for the iPr2NB-bridged [1]FCP 35, its DSC thermogram was remeasured confirming that endothermic and exothermic peaks overlap; however, based on the data a HROP of -73 kJ mol-1 was estimated. These measured heat releases are smaller than what is expected for [1]FCPs with angles greater than 30° [112]. The values should be similar to 130(±20) kJ mol-1 as determined for the sulfur-bridged [1]FCP with an angle of 31.05(10)° [107]. For comparison, the Me2Si-bridged [1]FCP with a tilt angle of 20.8(5)° [113] gave a HROP of ca. -80 kJ mol-1 [20]. This was already pointed out by Braunschweig and Manners et al. and the reduced enthalpy for 37 (HROP = -95 kJ mol-1) was rationalized by the presence of “side group interactions” 49
Page 50 of 70
between the sterically demanding amino groups, which presumably resulted in substantial destabilization of the polymer relative to the monomer [45]. As mentioned above, thermal ROP of 3pentyl decorated [1]FCPs 41 and 42 resulted in iron particles in addition to polymers. Examining the crucibles after DSC measurements of 35, 41, and 42 also showed the presence of elemental iron. It was
ip t
assumed that the thermal extrusion of iron occurs from the starting compound and is a radical process, which is either slightly exothermic or even endothermic. Therefore, the measured exothermy in DSC
cr
thermograms is not representative of the intrinsic strain of [1]FCPs, as a clean ring-opening reaction is
us
not occurring [47].
an
In conclusion one can say that bora[1]ferrocenophanes had not lived up to the expectation with respect to polymer formation. The first set of [1]FCPs (35-37) was plagued with low solubilities of the obtained
M
materials preventing their full characterization. The presence of alkyl groups on the second set of [1]FCPs (41, 42) helped to increase the solubilities of the obtained polymers, but they were of low
d
molecular weight and were not the pure linear polymers. The thermal stability of monomers is so high
te
that the required temperatures for ROP lead to extrusion of iron. One can speculate that amino groups at
Ac ce p
the bridging boron contributes significantly to the stability of the [1]FCPs. Hence, getting access to bora[1]ferrocenophanes that are aryl or alkyl substituted at boron might be key for the preparation of new boron-containing polymers in a controlled way [114, 115].
50
Page 51 of 70
5. [n]Metallocenophanes (n ≥ 2) Many metallocenophanes with more than one atom in bridging position are known; Figure 26 summarizes boron-containing species and Figure 27 summarizes species of the heavier group 13
ip t
elements. With the exception of the boron-bridged species 61, all other compounds are ferrocene
cr
derivatives that cover the range from [2]- to [4]FCPs.
Fe
Fe
E
BR
Fe 54
M
NMe2 R B C
NMe2 B R C Me 2N C B X B C NMe2 C R B Me2N
[PF6]2 N N B Fe
d
Fe
C
te
R' B NMe2
56
Ac ce p
R = Me, Ph, SiMe 3 R' = H, Me,
R
an
53 E = O, S, Se, Te, NR', BNMe 2, CNtBu, CNCy, Pt(PEt3 )2 R = NMe 2, NiPr 2, Br, Me
X
N B Ph NMe 2
BR
52 R = NMe 2, Mes
55
us
BR
BR
Fe
NMe 2 B Ph N
(X = no X, EC 2 H2, 1,4-C 6 H4 ),
O B N N
Fe
57
X = E-C 2H2, 1,4-C 6 H4 R = SiMe3
Pt(PEt 3) 2 R B Fe
R
E
E
E
Fe
BR
Fe
E
B 58
E
R'
E = PPh2 and/or
59 N
R''
N
R/R' = Br, Me, PPh2, OEt, N 2C 3 H3 , NC4 H 8, N2C 5F6 H, N 2C 4 H6 +
E = O, S, NR'
E
B B
Co NMe2
60 E = S, NSiMe3
R = H, F, Cl, Br, Me, Et, nBu, thexyl, OPh, Ph, Mes, NEt 2, NiPr 2, NiPrEt, NPhEt, N(CH 2Ph)Me, Fc
R R B
NMe2
61
OR' B RR
R = iPr, F R' = H, Me
51
Page 52 of 70
Fig. 26. Boron-bridged [n]metallocenophanes (n ≥ 2) (references: 52 [116-118], 53 [119-123], 54 [124],
Fe
SiMe3 O 2 Al
Fe
63
Py N SiMe 3
64
SiMe3 N Fe
62
N
R E
Do N SiMe 3
SiMe 3 N Cl Ga Cl N H SiMe3
cr
E = Al, Ga
In Fe
Cl
N
In Pytsi
30
d
E = Al, Ga R = Me, Et (for Al), Me, Et, Cl (for Ga)
Cl
M
65
SiMe 3
Pytsi
an
SiMe3 N
us
R = H, F, Cl, Br, I, Me, Et, tBu, CH(SiMe3 )2 , Si(SiMe 3) 3 , C R' (R' = nBu, tBu, Ph, SiMe3 ) Do = Py, NEtMe2
Fe R 2E ER 2
ip t
55 [120, 125], 56 [125, 126], 57 [127], 58 [128-132], 59 [133-137], 60 [136, 138], 61 [139, 140]).
te
Fig. 27. Aluminium-, gallium-, and indium-bridged [n]metallocenophanes (n ≥ 2) (references: 30 [49],
Ac ce p
62 [141-145], 63 [146], 64 [145], 65 [143, 145]).
The dibora[2]ferrocenophane 52 (R = NMe2) was reported for the first time in 1997 by Herberhold and Wrackmeyer et al. where a common salt-metathesis was used by reacting freshly prepared dilithioferrocene with the diborane(4) [(Me2N)ClB]2 [116]. Species 52 (R = NMe2) was obtained in a yield of 58% and characterized by NMR spectroscopy and mass spectrometry. The authors noted that the X-ray structural analysis was presented by A. Appel and H. Nöth at a conference; however, it seems that this analysis has not been published.3 Ten years later, species 52 (R = NMe2) was prepared by the flytrap route [3] (61% isolated yield) [117] and its structure analysis showed with = 12.82(16)° the 3
In ref. 2 of publication [122] it was indicated that the structural data for 52 are included in the Diploma Thesis or PhD Thesis of A. Appel (Faculty of Chemistry and Pharmacy, University of Munich, 1993/1996).
52
Page 53 of 70
presence of intrinsic strain. The second known dibora[2]ferrocenophane, 52 (R = Mes), was isolated in 37% and single crystal X-ray analysis revealed a similar tilt angle as for its amino-substituted counterpart [ = 10.5(2)°] [118]. An entire series of [3]FCPs (53) and [4]FCPs (54-56) were obtained by insertion reactions into the B-B bond of the amino-substituted dibora[2]ferrocenophane 52. The same
ip t
chemistry had been explored by using the respective dibora[2]chromarenophane [120, 121, 124]. The diborylation of alkynes was accomplished either stoichiometrically or catalytically, whereas the latter
cr
pathway was performed hetero- or homogeneously (Scheme 14). For stoichiometric pathway the
us
platinum species 53 [E = Pt(PEt3)2; R = NMe2] was prepared and isolated first and then reacted with alkynes. However, addition of catalytic amounts of Pt(PEt3)3 followed by addition of a 10-fold excess of
an
alkyne gave the respective [4]FCPs 55 in high yields (Scheme 14). Similar results were obtained by heterogeneously applying Pt or Pd metals. Obviously, species 53 [E = Pt(PEt3)2; R = NMe2] and
M
respective metal-insertion products are key intermediates in this process. As expected, the platinum compound 53 [E = Pt(PEt3)2; R = NMe2] is less strained compared to its precursor 52 (R = NMe2) as
d
indicated by the small tilt angle of 5.0° (Fig. 28) [120]. On the basis of these first discoveries [120],
te
Braunschweig et al. explored this process further to get access to species 53 (E = CNtBu, CNCy; R =
Ac ce p
NMe2 [121], S and Se [123]), 54 [124], 55 [125], and 56 [125, 126].
Scheme 14. Insertions into the B-B bond of 52 (R = NMe2) [120].
53
Page 54 of 70
R
NMe 2 R B C
R'
Fe
Fe B
Pt sponge or Pd charcoal
NMe2
52
C R' B NMe 2
55 R = Me R' = H, Me
NMe2 B Fe
R'
Pt(PEt3 )2 B NMe2
M
an
us
53
R
cr
[Pt(PEt3) 3]
ip t
B
NMe2
N1 P1 B1 Fe1
d
Pt1
te
B2
Ac ce p
P2
N2
Fig. 28. Molecular structure of the insertion product 53 [E = Pt(PEt3)2; R = NMe2] with thermal ellipsoids drawn at the 50% probability level. Hydrogen atoms are omitted for clarity. The figure was prepared with ORTEP-3 for Windows [90] using the published crystallographic data [120].
R EH
B Fe R B X
X
Fe HE
66
67
X = Cl, Br
E = O, S, NR
54
Page 55 of 70
Fig. 29. Two types of starting compounds for the preparation of FCPs.
The preparation of other species shown in Figure 26 that were not obtained through insertions can be sorted into two groups: one with the ferrocene-containing reagent acting as the electrophile (66) and one
ip t
where it is acting as the nucleophile (67; Fig. 29). The latter reagent was either reacted with a respective element halide reagent directly (in the presence of NEt3 to trap HX) or it was first deprotonated and its
cr
dilithio derivative was applied. An example for the use of a bis(boryl) species (66) is the reaction of
us
fc[BBr(NiPr2)]2 with Li2Se that gave the respective [3]FCP (53 with R = NiPr2 and E = Se; Fig. 26) [119]. Among the compounds depicted in Figure 26 the tribora[3]ferrocenophane 53 (E = BNMe2; R =
an
NMe2) cannot be sorted into any of the two bins as it was synthesized using the common salt-metathesis
M
for FCPs starting from dilithioferrocene and the triborane Me2NB[BCl(NMe2)]2 [122].
The ansa-bridges in species of type 58 are formed by donor-acceptor bonds that many derivatives were
d
prepared by Wagner et al. with a particular emphasis on the dynamic behaviour [128-132]. Depending
te
on the Lewis acidity of the boryl groups and the Lewis basicity of the bridging moieties E only the
Ac ce p
closed form, the open form, or equilibria were found (Fig. 30) [132].
B Fe
E
R
R
B
E
B
R'
Fe
E
R' B
close-58
E = PPh2 and/or
E
N
open- 58 R''
N
Fig. 30. Open and closed form of species 58.
Species 61 stands out as the only non-iron metallocenophane and is an inverse chelate between the 1,1'-
55
Page 56 of 70
bis(boryl)cobaltocenium ion and RO- (Fig. 26).
As shown in Figure 27, the group of aluminum-, gallium-, or indium-bridged [2]metallocenophanes is smaller than the group of boron species (Fig. 26). An analogue of the reactive dibora[2]ferrocenophane
ip t
52 is not known for the heavier group members. Uhl et al. mentioned that all of their attempts to prepare such a [2]FCP bridged by a Ga2 moiety failed, but led to the synthesis of the digalla[1.1]ferrocenophane
cr
72 (Chapter 6) [147]. All the aluminum- and gallium-bridged species were synthesized by Wrackmeyer
us
et al. [141-146] while the only indium-bridged species (30; Fig. 27; see also Chapter 3) had been prepared in our group [49]. The common starting material for compounds 62-65 is 1,1'-
an
bis(trimethylsilylamino)ferrocene (67, E = NSiMe3; Fig. 29) that was either reacted directly with hydrides (H3Al·NEtMe2, H3Al·py; [142, 144, 146]), alkyl species (AlR3, GaR3; R = Me, Et; [143, 145]),
M
or first converted into its dilithio derivative, which was reacted with a variety of aluminum or gallium
te
6. [n.n]Metallocenophanes
d
halides [141, 142, 145].
Ac ce p
The first [1.1]FCPs were carbon-bridged species (Fig. 31; ERx = CH2, CHPh) that had been described by Nesmeyanov and Kritskaya in 1956 [148]. Carbon-bridged [1.1]FCPs were later structurally characterized as syn-isomers (Fig. 31) [149, 150], which is the preferred conformation with the small carbon in bridging position (for an exception see [151]). In the first decades of metallocenophane chemistry, researchers focused on carbon-bridged species, an area that was reviewed in 1986 by Mueller-Westerhoff [12] (see also the review [39]). After the first hetero-atom bridged [1.1]FCP was described by Seyferth et al. in 1982 [Fig. 31; ERx = Sn(nBu)2, SnEt2] [152], this class has grown substantially and [1.1]FCPs with following bridging elements are known: B [153], Al [66, 91, 92, 154], Ga [92, 147, 155-157], In [92, 93, 104], Si [109, 158-165], Sn [85, 87, 88, 152, 166, 167], Pb [168], P [30, 169, 170], As [171], S [172], Zn [173], and Hg [174, 175]. Furthermore, [1.1]RCPs [12, 158] and
56
Page 57 of 70
mixed [1.1]metallocyclophanes [12] with Fe/Ru and Fe/Co, respectively, had been described. In contrast, related arene complexes are very rare and, to the best of our knowledge, the aluminum and gallium-bridged [1.1]chromarenophanes and [1.1]molybdarenophanes are the only examples of this family [95]. From the lists of species above, [1.1]FCPs are the only group-13-bridged
ip t
[1.1]metallocenophanes described in literature. As one can see from Table 5, which provides an overview of all known species, the majority was structurally characterized by single crystal X-ray
cr
analysis and with the exception of the boron compound 68 and the gallium compound 74, all other
us
species exhibit anti conformations in the solid state (Fig. 31).
Fe Fe
Fe ER x
an
R xE E Fe Rx
anti
M
E Rx syn
d
Fig. 31. syn and anti isomers of [1.1]ferrocenophanes
te
Table 5. [1.1]Ferrocenophanes bridged by group 13 elements.
Ac ce p
Rx E
R'
R' Fe
R'
R'
Fe
E Rx
R'
ERx
references
H
BMe2–
68 a
[153]
H
AlAr'
11 a
[66, 91, 92]
H
AlCl(tmeda)
69 a
[154]
H
Al(p-Me3SiAr')
70 a
[66]
H
Al(Mpysm)
71
[66]
H
Ga[CH(SiMe3)2]2
72 a
[147]
H
GaMe(NC5H5) b
73 a
[155, 156]
57
Page 58 of 70
[155, 156]
H
GaMe
75 a
[157]
H
GaAr'
12 a
[66, 92]
H
Ga(p-Me3SiAr')
76 a
[66]
H
Ga(Mpysm)
77
[66]
H
GaAr' / SiMe2 c
78
[176]
H
Ga(p-SiMe3Ar') / SiMe2 c
79 a
[176]
H
InAr'
13 a
[92]
H
In(Me2Ntsi)
14 a
[93]
H
In(Mamx)
312
[104]
CHMe2
InAr'
322
[104]
ip t
74 a
cr
Ga(fc)1/2(NC5H5)
us
H
an
[a] Single crystal X-ray data available. [b] A series of compounds with other donors had been reported in [157]. [c] silagalla[1.1]ferrocenophane
The first reported group-13-bridged [1.1]FCPs were the species 72 and 73 4 and the unusual
M
metallocenophane 74 [155]. Whereas the majority of species shown in Table 5 were prepared by the
te
condensation route (Scheme 15).
d
common salt-metathesis route, the preparation of the gallium compounds 73 and 74 were obtained by a
Ac ce p
Scheme 15. Unusual condensation route to species 73 and 74.
4
Species 72 [147] was published before 73 and 74 [155], however, the latter contribution was submitted two months earlier than that describing 72.
58
Page 59 of 70
N
Me 2 Ga
Me 3Ga
Fe
Me 2 Ga Fe Fe
80
pyridine
Ga Me 2
Ga Me 2
SnCl3
GaMe 2 Fe GaMe 2
n 81
82
N Fe
Fe
Me Ga
- Me3Ga·py
Fe Fe 73
N
M
N
Ga Me
an
Ga 74
toluene,
Fe
+
us
N Ga
cr
N
ip t
SnCl3
d
Heating of the donor-stabilized bis(gallyl)ferrocene 82 resulted in a redistribution of the two ligands Me
te
and fc at gallium, yielding the three Lewis acid-base adducts 73, 74, and Me3Ga·py (Scheme 15). Such
Ac ce p
a redistribution of ligands will require association of species through diborane(6)-type moieties. Such moieties are present in the precursor 81 [156], which was sparingly soluble in non-coordinating solvents. The carousel molecule 74 was isolated as a pure, crystalline compound (Fig. 32) by heating the mixture of 73 and 74 in a high-boiling solvent (e. g., toluene or p-xylene), followed by cooling [155]. In this process, further conversion of 73 into 74 had been observed, a fact that shows that 74 and Me3Ga·py are the end products of the ligand redistribution reaction with the [1.1]FCPs 73 being an intermediate. This type of condensation reaction is reminiscent of that found by Jäkle, Holthausen, and Wagner, where a 1,1'-bis(boryl)ferrocene gave poly(ferrocenylborane)s and diborane(6) [177].
59
Page 60 of 70
N1
Ga1
ip t
Fe1 Fe2
cr
Fe2A
us
Ga1A
an
N1A
M
Fig. 32. Molecular structure of the carousel molecule 74 with thermal ellipsoids drawn at the 50% probability level. Hydrogen atoms and solvent molecules are omitted for clarity. The figure was
te
d
prepared with ORTEP-3 for Windows [90] using the published crystallographic data [155].
Ac ce p
To date, species 68 is the only reported boron-bridged [1.1]FCP [153].5 As shown in Scheme 16, the species had been synthesized by a Lewis acid-base adduct formation between dilithioferrocene and 1,1'bis(dimethylboryl)ferrocene (83) and obtained as the crystalline salt Li(12-crown-4)2[68-Li]. According to the single crystal X-ray analysis, one of the two counter cations is chelated by crown ethers, whereas the other one is trapped inside the cavity of the [1.1]FCP. 7Li NMR data suggests that this unusual structure is preserved in solution as two signals in a 1 : 1 ratio were found. The authors speculated that the lithium ion is “most likely trapped by the electrostatic field originating from the two anionic dimethylborate bridges” [153].
5
Footnote 38 in [45] states that R2NB-bridged [1.1]FCPs have been prepared by Braunschweig and Koblinski; however, these results are still unpublished.
60
Page 61 of 70
Scheme 16. Preparation of the boron-bridged [1.1]FCP 68. Li(12-crown-4) Li Fe
BMe 2 +
BMe2
2 tmeda
B Fe Me2
12-crown-4
Fe
Li
2
Fe Li+
- tmeda
B Me 2 [68-Li] -
cr
ip t
83
B2
us
Li1 Fe1
B1
an
Fe2
M
Fig. 33. Molecular structure of [68-Li]- with thermal ellipsoids drawn at the 50% probability level. Hydrogen atoms and Li(12-crown-4)2]+ are omitted for clarity. The figure was prepared with ORTEP-3
d
for Windows [90] using the published crystallographic data [153].
te
As mentioned above, [1.1]FCPs were either found as syn or as anti isomers (Fig. 31). These isomers are
Ac ce p
organometallic analogues of cyclohexane [178], with the syn isomer being related to the boat isomer and the anti isomer being related to the chair isomer of cyclohexane. Isomers in a syn conformation should be better described as twist-syn isomers (similar as the twist-boat isomer of cyclohexane). An example of such an isomer is the diborata[1.1]ferrocenophane 68 (Fig. 33), where dihedral angles between related Cp rings coordinated to different iron atoms were found as 21.3 and 23.1°, respectively [153]. For comparison, Figure 34 depicts the molecular structure of the gallium compound 72, which serves as a typical example of a [1.1]FCP in an anti conformation (also see Fig. 31).
61
Page 62 of 70
Si1
Ga Si2 Fe'
ip t
Fe
cr
Si2'
Ga'
us
Si1'
Fig. 34. Molecular structure of 72 with thermal ellipsoids drawn at the 50% probability level. Hydrogen
an
atoms are omitted for clarity. The figure was prepared with ORTEP-3 for Windows [90] using the
M
published crystallographic data [147].
d
Bridging elements with small covalent radii prefer the syn conformation, whereas larger bridging
te
elements prefer the anti conformation. In both structures, a bridging atom with a small radius would result in close HH contacts between hydrogen atoms in positon on neighbouring Cp rings; this is
Ac ce p
illustrated for the syn isomer in Figure 35. However, the syn conformation is more flexible than the anti conformation and can form a twisted variant and, hence, allow for a minimization of steric repulsion between -H atoms [12, 178].
Rx E
H H Fe Rx Fe E
H H
Fig. 35. Illustration of steric interactions between -H atoms in syn isomers.
Many of the [1.1]FCPs are highly dynamic in solution and fewer signals than expected are found in 1H 62
Page 63 of 70
and 13C NMR spectra. These facts had been explained by the occurrence of degenerate isomerizations, more specifically, either syn-to-syn or anti-to-anti isomerizations. Considering the relation to cyclohexane that is known for its fast interconversions of conformers, degenerate isomerizations in [1.1]FCPs are not such a surprize. For example, the digalla[1.1]ferrocenophane 72 showed only two
ip t
NMR signals for the 16 Cp protons [147]. With the assumption of free rotations of CH(SiMe3)2 groups at gallium, four 1H NMR signals for Cp protons are expected for such an anti isomer with C2h symmetry
cr
in solution. The fact that only two signals were found shows that the group order of the time-averaged
us
symmetry for 72 must be h = 8 instead of h = 4 (C2h). This can be explained by a degenerate anti-to-anti
an
isomerization, which results in a time-averaged D2h symmetry (h = 8).
Among this group of [1.1]FCPs, only species 78 and 79 (Table 5) have two different elements (Ga and
M
Si) in alternate bridging positions [176]. Both species were prepared by a salt-metathesis approach that also gave mixtures of oligomers among which silagalla[1m]ferrocenophanes (F; Fig. 1; Chapter 1) with
d
up to 12 ferrocenediyl moieties were detected by MALDI-TOF mass spectrometry. To the best of our
te
knowledge, these are the only reported group-13-element bridged [1m]metallocenophanes that have
Ac ce p
more than two sandwich moieties.
7. Conclusions
To date, group-13-element containing [n]metallocenophanes with up to four elements in bridging position are known (n = 1–4). As expected, the majority of these species are ferrocene derivatives, however, a few ruthenocenophanes ([1]RCPs 17 and 18; Scheme 4; Chapter 2.3) and cobaltocenophanes (61; Fig. 26; Chapter 5) had been prepared as well. For metallocenophanes that contain more than one sandwich moiety (F; Fig. 1; Chapter 1) only [1.1]FCPs are reported in the literature. Among this group of species, only 78 and 79 (Table 5; Chapter 6) have two different elements (Ga and Si) in alternate bridging positions. On a personal note, we are fascinated by the possibility to prepare strained species
63
Page 64 of 70
that could become useful monomers for polymerization. As mentioned in the introduction, siliconbridged [1]FCPs have fulfilled this expectation. With respect to metallopolymers, group-13-element bridged metallocenophanes are still in their infancy and the future will reveal if creative synthetic
ip t
chemistry can help them to mature.
Acknowledgments
cr
The authors are grateful to the University of Saskatchewan, the College of Arts and Science (Catalyst
us
Award for H. B.), and the NSERC (Discovery Grant for J. M.) for their support. J. M. would like to thank all his students and collaborators who contributed to the results summarized within this review for
Ac ce p
te
d
M
an
their dedication to phane chemistry.
64
Page 65 of 70
References
Ac ce p
te
d
M
an
us
cr
ip t
[1] A. Salzer, Pure Appl. Chem., 71 (1999) 1557-1585. [2] H.H. Brintzinger, D. Fischer, R. Mülhaupt, B. Rieger, R.M. Waymouth, Angew. Chem., Int. Ed., 34 (1995) 1143-1170. [3] D.E. Herbert, U.F.J. Mayer, I. Manners, Angew. Chem., Int. Ed., 46 (2007) 5060-5081. [4] M. Tamm, Chem. Commun., (2008) 3089-3100. [5] H. Braunschweig, T. Kupfer, Acc. Chem. Res., 43 (2010) 455-465. [6] M. Herberhold, T. Hofmann, S. Weinberger, B. Wrackmeyer, Z. Naturforsch. B, 52 (1997) 10371042. [7] H. Braunschweig, T. Kupfer, K. Radacki, Angew. Chem., Int. Ed., 46 (2007) 1630-1633. [8] H. Braunschweig, M. Friedrich, T. Kupfer, K. Radacki, Chem. Commun., 47 (2011) 3998-4000. [9] H. Braunschweig, F. Breher, S. Capper, K. Duck, M. Fuss, J.O.C. Jimenez-Halla, I. Krummenacher, T. Kupfer, D. Nied, K. Radacki, Chem.-Eur. J., 19 (2013) 270-281. [10] P. Chadha, J.L. Dutton, M.J. Sgro, P.J. Ragogna, Organometallics, 26 (2007) 6063-6065. [11] H. Braunschweig, M. Fuss, T. Kupfer, K. Radacki, J. Am. Chem. Soc., 133 (2011) 5780-5783. [12] U.T. Mueller-Westerhoff, Angew. Chem., Int. Ed., 25 (1986) 702-717. [13] K.L. Rinehart, R.J. Curby, J. Am. Chem. Soc., 79 (1957) 3290-3291. [14] K.L. Rinehart, A.K. Frerichs, P.A. Kittle, L.F. Westman, D.H. Gustafson, R.L. Pruett, J.E. McMahon, J. Am. Chem. Soc., 82 (1960) 4111-4112. [15] A.G. Osborne, R.H. Whiteley, J. Organomet. Chem., 101 (1975) C27-C28. [16] C. Elschenbroich, J. Hurley, B. Metz, W. Massa, G. Baum, Organometallics, 9 (1990) 889-897. [17] M. Tamm, A. Kunst, T. Bannenberg, E. Herdtweck, P. Sirsch, C.J. Elsevier, J.M. Ernsting, Angew. Chem., Int. Ed., 43 (2004) 5530-5534. [18] C. Elschenbroich, F. Paganelli, M. Nowotny, B. Neumüller, O. Burghaus, Z. Anorg. Allg. Chem., 630 (2004) 1599-1606. [19] H. Braunschweig, M. Lutz, K. Radacki, Angew. Chem., Int. Ed., 44 (2005) 5647-5651. [20] D.A. Foucher, B.-Z. Tang, I. Manners, J. Am. Chem. Soc., 114 (1992) 6246-6248. [21] V. Bellas, M. Rehahn, Angew. Chem., Int. Ed., 46 (2007) 5082-5104. [22] R. Rulkens, A.J. Lough, I. Manners, J. Am. Chem. Soc., 116 (1994) 797-798. [23] Y.Z. Ni, R. Rulkens, I. Manners, J. Am. Chem. Soc., 118 (1996) 4102-4114. [24] H. Qiu, Z.M. Hudson, M.A. Winnik, I. Manners, Science, 347 (2015) 1329-1332. [25] Z.M. Hudson, D.J. Lunn, M.A. Winnik, I. Manners, Nat. Commun., 5 (2014) 3372. [26] Z.M. Hudson, C.E. Boott, M.E. Robinson, P.A. Rupar, M.A. Winnik, I. Manners, Nat. Chem., 6 (2014) 893-898. [27] N.S. Ieong, I. Manners, Macromol. Chem. Phys., 210 (2009) 1080-1086. [28] C.H. Honeyman, T.J. Peckham, J.A. Massey, I. Manners, Chem. Commun., (1996) 2589-2590. [29] T.J. Peckham, J.A. Massey, C.H. Honeyman, I. Manners, Macromolecules, 32 (1999) 2830-2837. [30] T. Mizuta, M. Onishi, K. Miyoshi, Organometallics, 19 (2000) 5005-5009. [31] T. Mizuta, Y. Imamura, K. Miyoshi, J. Am. Chem. Soc., 125 (2003) 2068-2069. [32] S.K. Patra, G.R. Whittell, S. Nagiah, C.-L. Ho, W.-Y. Wong, I. Manners, Chem. Eur. J., 16 (2010) 3240-3250. [33] D.E. Herbert, U.F.J. Mayer, J.B. Gilroy, M.J. Lopez-Gómez, A.J. Lough, J.P.H. Charmant, I. Manners, Chem. Eur. J, 15 (2009) 12234-12246. [34] H. Braunschweig, F.M. Breitling, E. Gullo, M. Kraft, J. Organomet. Chem., 680 (2003) 31-42. [35] S. Aldridge, C. Bresner, Coord. Chem. Rev., 244 (2003) 71-92. [36] C.J. Adams, H. Braunschweig, M. Fuss, K. Kraft, T. Kupfer, I. Manners, K. Radacki, G.R. Whittell, Chem.-Eur. J., 17 (2011) 10379-10387. [37] A.C. Tagne Kuate, C.G. Daniliuc, P.G. Jones, M. Tamm, Eur. J. Inorg. Chem., 2012 (2012) 17271733. 65
Page 66 of 70
Ac ce p
te
d
M
an
us
cr
ip t
[38] H. Braunschweig, A. Damme, S. Demeshko, K. Dück, T. Kramer, I. Krummenacher, F. Meyer, K. Radacki, S. Stellwag-Konertz, G.R. Whittell, J. Am. Chem. Soc., 137 (2015) 1492-1500. [39] R.W. Heo, T.R. Lee, J. Organomet. Chem., 578 (1999) 31-42. [40] H. Braunschweig, T. Kupfer, Eur. J. Inorg. Chem., (2012) 1319-1332. [41] F.H. Schacher, P.A. Rupar, I. Manners, Angew. Chem., Int. Ed., 51 (2012) 7898-7921. [42] R.A. Musgrave, A.D. Russell, I. Manners, Organometallics, 32 (2013) 5654−5667. [43] A.D. Russell, R.A. Musgrave, L.K. Stoll, P. Choi, H. Qiu, I. Manners, J. Organomet. Chem., 784 (2015) 24-30. [44] H. Braunschweig, R. Dirk, M. Müller, P. Nguyen, R. Resendes, D.P. Gates, I. Manners, Angew. Chem., Int. Ed., 36 (1997) 2338-2340. [45] A. Berenbaum, H. Braunschweig, R. Dirk, U. Englert, J.C. Green, F. Jäkle, A.J. Lough, I. Manners, J. Am. Chem. Soc., 122 (2000) 5765-5774. [46] S. Sadeh, G. Schatte, J. Müller, Chem. Eur. J., 19 (2013) 13408–13417. [47] S. Sadeh, H. Bhattacharjee, E. Khozeimeh Sarbisheh, J.W. Quail, J. Müller, Chem. Eur. J., 20 (2014) 16320-16330. [48] J.A. Schachner, C.L. Lund, J.W. Quail, J. Müller, Organometallics, 24 (2005) 785-787. [49] J.A. Schachner, C.L. Lund, J.W. Quail, J. Müller, Organometallics, 24 (2005) 4483-4488. [50] C. Eaborn, J.D. Smith, J. Chem. Soc., Dalton Trans., (2001) 1541-1552. [51] S.S. Al-Juaid, C. Eaborn, P.B. Hitchcock, M.S. Hill, J.D. Smith, Organometallics, 19 (2000) 32243231. [52] C. Eaborn, M.S. Hill, P.B. Hitchcock, J.D. Smith, Chem. Commun., (2000) 691-692. [53] S.S. Al-Juaid, A.G. Avent, C. Eaborn, S.M. El-Hamruni, S.A. Hawkes, M.S. Hill, M. Hopman, P.B. Hitchcock, J.D. Smith, J. Organomet. Chem., 631 (2001) 76-86. [54] S.S. Al-Juaid, A.G. Avent, C. Eaborn, M.S. Hill, P.B. Hitchcock, D.J. Patel, J.D. Smith, Organometallics, 20 (2001) 1223-1229. [55] C. Eaborn, M.S. Hill, P.B. Hitchcock, J.D. Smith, J. Chem. Soc., Dalton Trans., (2002) 2467-2472. [56] R.A. Fischer, A. Miehr, H. Sussek, H. Pritzkow, E. Herdtweck, J. Müller, O. Ambacher, T. Metzger, J. Chem. Soc., Chem. Commun., (1996) 2685-2686. [57] J. Müller, R.A. Fischer, H. Sussek, P. Pilgram, R. Wang, H. Pritzkow, E. Herdtweck, Organometallics, 17 (1998) 161-166. [58] A. Devi, W. Rogge, A. Wohlfart, F. Hipler, H.W. Becker, R.A. Fischer, Chem. Vap. Deposition, 6 (2000) 245-252. [59] J. Müller, S. Bendix, J. Chem. Soc., Chem. Commun., (2001) 911-912. [60] A. Devi, R. Schmid, J. Müller, R.A. Fischer, Materials Chemistry of Group-13 Nitrides, in: R.A. Fischer (Ed.) Topics in Organometallic Chemistry, Springer, Heidelberg, 2005, pp. 49-80. [61] A. Haaland, Angew. Chem., Int. Ed., 28 (1989) 992-1007. [62] J. Müller, U. Englert, Chem. Ber., 128 (1995) 493-497. [63] H. Schumann, B.C. Wassermann, S. Schutte, B. Heymer, S. Nickel, T.D. Seuss, S. Wernik, J. Demtschuk, F. Girgsdies, R. Weimann, Z. Anorg. Allg. Chem., 626 (2000) 2081-2095. [64] C.L. Lund, J.A. Schachner, J.W. Quail, J. Müller, Organometallics, 25 (2006) 5817-5823. [65] M. Yoshifuji, K. Kamijo, K. Toyota, Tetrahedron Lett., 35 (1994) 3971-3974. [66] B. Bagh, N.C. Breit, K. Harms, G. Schatte, I.J. Burgess, H. Braunschweig, J. Müller, Inorg. Chem., 51 (2012) 11155−11167. [67] J.A. Schachner, S. Tockner, C.L. Lund, J.W. Quail, M. Rehahn, J. Müller, Organometallics, 26 (2007) 4658-4662. [68] J.A. Schachner, J.W. Quail, J. Müller, Acta Crystallogr., E64 (2008) m517. [69] N.C. Breit, T. Ancelet, J.W. Quail, G. Schatte, J. Müller, Organometallics, 30 (2011) 6150-6158. [70] B. Bagh, G. Schatte, J.C. Green, J. Müller, J. Am. Chem. Soc., 134 (2012) 7924−7936. [71] B. Bagh, J.B. Gilroy, A. Staubitz, J. Müller, J. Am. Chem. Soc., 132 (2010) 1794-1795. [72] Mercury, in, http://www.ccdc.cam.ac.uk/mercury. [73] C.L. Lund, J.A. Schachner, J.W. Quail, J. Müller, J. Am. Chem. Soc., 129 (2007) 9313-9320. 66
Page 67 of 70
Ac ce p
te
d
M
an
us
cr
ip t
[74] C. Elschenbroich, A. Bretschneider-Hurley, J. Hurley, W. Massa, S. Wocadlo, J. Pebler, E. Reijerse, Inorg. Chem., 32 (1993) 5421-5424. [75] C. Elschenbroich, A. Bretschneider-Hurley, J. Hurley, A. Behrendt, W. Massa, S. Wocadlo, E. Reijerse, Inorg. Chem., 34 (1995) 743-745. [76] K.C. Hultzsch, J.M. Nelson, A.J. Lough, I. Manners, Organometallics, 14 (1995) 5496-5502. [77] C. Elschenbroich, E. Schmidt, B. Metz, K. Harms, Organometallics, 14 (1995) 4043-4045. [78] C. Elschenbroich, E. Schmidt, R. Gondrum, B. Metz, O. Burghaus, W. Massa, S. Wocadlo, Organometallics, 16 (1997) 4589-4596. [79] D. Seyferth, Organometallics, 21 (2002) 1520-1530. [80] H. Braunschweig, M. Homberger, C.H. Hu, X.L. Zheng, E. Gullo, G. Clentsmith, M. Lutz, Organometallics, 23 (2004) 1968-1970. [81] C. Lund, J. Müller, unpublished results [82] S.S. Al-Juaid, C. Eaborn, S.M. El-Hamruni, P.B. Hitchcock, J.D. Smith, Organometallics, 18 (1999) 45-52. [83] U. Vogel, A.J. Lough, I. Manners, Angew. Chem., Int. Ed., 43 (2004) 3321-3325. [84] R. Rulkens, A.J. Lough, I. Manners, Angew. Chem., Int. Ed., 35 (1996) 1805-1807. [85] F. Jäkle, R. Rulkens, G. Zech, D.A. Foucher, A.J. Lough, I. Manners, Chem. Eur. J., 4 (1998) 2117-2128. [86] H.K. Sharma, F. Cervantes-Lee, J.S. Mahmoud, K.H. Pannell, Organometallics, 18 (1999) 399403. [87] F. Jäkle, R. Rulkens, G. Zech, J. Massey, I. Manners, J. Am. Chem. Soc., 122 (2000) 4231-4232. [88] T. Baumgartner, F. Jäkle, R. Rulkens, G. Zech, A.J. Lough, I. Manners, J. Am. Chem. Soc., 124 (2002) 10062-10070. [89] A. Berenbaum, F. Jäkle, A.J. Lough, I. Manners, Organometallics, 21 (2002) 2359 -2361. [90] L.J. Farrugia, J. Appl. Crystallogr., 30 (1997) 565. [91] H. Braunschweig, C. Burschka, G.K.B. Clentsmith, T. Kupfer, K. Radacki, Inorg. Chem., 44 (2005) 4906-4908. [92] J.A. Schachner, G.A. Orlowski, J.W. Quail, H.-B. Kraatz, J. Müller, Inorg. Chem., 45 (2006) 454459. [93] J.A. Schachner, C.L. Lund, I.J. Burgess, J.W. Quail, G. Schatte, J. Müller, Organometallics, 27 (2008) 4703-4710. [94] Holleman-Wiberg, Inorganic Chemistry, 1. English ed., Academic Press, San Diego, London, 2001, p 1756. [95] C.L. Lund, J.A. Schachner, I.J. Burgess, J.W. Quail, G. Schatte, J. Müller, Inorg. Chem., 47 (2008) 5992–6000. [96] T. Matsumoto, K. Tanaka, Y. Chujo, J. Am. Chem. Soc., 135 (2013) 4211-4214. [97] T. Matsumoto, Y. Onishi, K. Tanaka, H. Fueno, Y. Chujo, Chem. Commun., 50 (2014) 1574015743. [98] L. Schwink, P. Knochel, Chem. Eur. J., 4 (1998) 950-968. [99] D. Marquarding, H. Klusacek, G. Gokel, P. Hoffmann, I. Ugi, J. Am. Chem. Soc., 92 (1970) 53895393. [100] S. Itsuno, K. Ito, A. Hirao, S. Nakahama, J. Chem. Soc., Chem. Commun., (1983) 469-470. [101] E.J. Corey, R.K. Bakshi, S. Shibata, J. Am. Chem. Soc., 109 (1987) 5551-5553. [102] J. Kang, J.H. Lee, J.S. Choi, Tetrahedron: Asymmetry, 12 (2001) 33-35. [103] W.F. Bailey, M.R. Luderer, K.P. Jordan, J. Org. Chem., 71 (2006) 2825-2828. [104] B. Bagh, S. Sadeh, J.C. Green, J. Müller, Chem. Eur. J., 20 (2014) 2318-2327. [105] E. Khozeimeh Sarbisheh, J.C. Green, J. Müller, Organometallics, 33 (2014) 3508−3513. [106] J.K. Pudelski, D.P. Gates, R. Rulkens, A.J. Lough, I. Manners, Angew. Chem., Int. Ed., 34 (1995) 1506-1508. [107] R. Rulkens, D.P. Gates, D. Balaishis, J.K. Pudelski, D.F. McIntosh, A.J. Lough, I. Manners, J. Am. Chem. Soc., 119 (1997) 10976-10986. 67
Page 68 of 70
Ac ce p
te
d
M
an
us
cr
ip t
[108] N.P. Reddy, H. Yamashita, M. Tanaka, J. Chem. Soc., Chem. Commun., (1995) 2263-2264. [109] Y.Z. Ni, R. Rulkens, J.K. Pudelski, I. Manners, Macromol. Rapid Commun., 16 (1995) 637-641. [110] J.B. Sheridan, K. Temple, A.J. Lough, I. Manners, J. Chem. Soc., Dalton Trans., (1997) 711-713. [111] J. Weaver, P. Woodward, J. Chem. Soc., Dalton Trans., (1973) 1439-1443. [112] J.C. Green, Chem. Soc. Rev., 27 (1998) 263-271. [113] W. Finckh, B.-Z. Tang, D.A. Foucher, D.B. Zamble, R. Ziembinski, A. Lough, I. Manners, Organometallics, 12 (1993) 823-829. [114] F. Jäkle, Chem. Rev., 110 (2010) 3985-4022. [115] F. Cheng, F. Jäkle, Polym. Chem., 2 (2011) 2122-2132. [116] M. Herberhold, U. Dörfler, B. Wrackmeyer, J. Organomet. Chem., 530 (1997) 117-120. [117] H. Braunschweig, F. Seeler, R. Sigritz, J. Organomet. Chem., 692 (2007) 2354-2356. [118] H. Braunschweig, A. Damme, T. Kupfer, Eur. J. Inorg. Chem., 2010 (2010) 4423-4426. [119] M. Herberhold, U. Dörfler, W. Milius, B. Wrackmeyer, J. Organomet. Chem., 492 (1995) 59-63. [120] H. Braunschweig, T. Kupfer, M. Lutz, K. Radacki, F. Seeler, R. Sigritz, Angew. Chem., Int. Ed., 45 (2006) 8048-8051. [121] F. Bauer, H. Braunschweig, K. Schwab, Organometallics, 29 (2010) 934-938. [122] A. Appel, H. Nöth, Z. Anorg. Allg. Chem., 636 (2010) 2329-2342. [123] H. Bera, H. Braunschweig, A. Oechsner, F. Seeler, R. Sigritz, J. Organomet. Chem., 695 (2010) 2609-2613. [124] H. Braunschweig, T. Kupfer, J. Am. Chem. Soc., 130 (2008) 4242-4243. [125] F. Bauer, H. Braunschweig, K. Gruss, T. Kupfer, Organometallics, 30 (2011) 2869-2884. [126] F. Bauer, H. Braunschweig, K. Gruß, C. Lambert, K.K. Pandey, K. Radacki, D. Reitzenstein, Chem. Eur. J., 17 (2011) 5230-5233. [127] L. Ding, K. Ma, F. Fabrizi de Biani, M. Bolte, P. Zanello, M. Wagner, Organometallics, 20 (2001) 1041-1043. [128] F. Jäkle, M. Mattner, T. Priermeier, M. Wagner, J. Organomet. Chem., 502 (1995) 123-130. [129] F. Jäkle, T. Priermeier, M. Wagner, J. Chem. Soc., Chem. Commun., (1995) 1765-1766. [130] F. Jäkle, T. Priermeier, M. Wagner, Organometallics, 15 (1996) 2033-2040. [131] E. Herdtweck, F. Jäkle, G. Opromolla, M. Spiegler, M. Wagner, P. Zanello, Organometallics, 15 (1996) 5524-5535. [132] E. Herdtweck, F. Jäkle, M. Wagner, Organometallics, 16 (1997) 4737-4745. [133] D. Fest, C.D. Habben, J. Organomet. Chem., 390 (1990) 339-344. [134] M. Herberhold, H.D. Brendel, J. Organomet. Chem., 458 (1993) 205-209. [135] B. Wrackmeyer, E.V. Klimkina, H.E. Maisel, W. Milius, M. Herberhold, Inorg. Chim. Acta, 357 (2004) 1703-1710. [136] B. Wrackmeyer, E.V. Klimkina, W. Millus, O.L. Tok, M. Herberhold, Inorg. Chim. Acta, 358 (2005) 1420-1428. [137] T.C. Auch, H. Braunschweig, K. Radacki, R. Sigritz, U. Siemeling, S. Stellwag, Z. Naturforsch. B, 63 (2008) 920-922. [138] M.J.G. Lesley, U. Mock, N.C. Norman, A.G. Orpen, C.R. Rice, J. Starbuck, J. Organomet. Chem., 582 (1999) 116-118. [139] G.E. Herberich, A. Fischer, D. Wiebelhaus, Organometallics, 15 (1996) 3106-3108. [140] G.E. Herberich, U. Englert, A. Fischer, D. Wiebelhaus, Organometallics, 17 (1998) 4769-4775. [141] B. Wrackmeyer, E.V. Klimkina, W. Milius, Inorg. Chem. Commun., 9 (2006) 716-721. [142] B. Wrackmeyer, E.V. Klimkina, W. Milius, Polyhedron, 26 (2007) 3496-3504. [143] B. Wrackmeyer, E.V. Klimkina, W. Milius, Z. Anorg. Allg. Chem., 633 (2007) 1964-1972. [144] B. Wrackmeyer, E.V. Klimkina, W. Milius, Z. Naturforsch. B, 62 (2007) 1259-1266. [145] B. Wrackmeyer, E.V. Klimkina, W. Milius, Z. Anorg. Allg. Chem., 635 (2009) 2532-2540. [146] B. Wrackmeyer, E.V. Klimkina, W. Milius, Eur. J. Inorg. Chem., (2008) 306-311. [147] W. Uhl, I. Hahn, A. Jantschak, T. Spies, J. Organomet. Chem., 637 (2001) 300-303. [148] A.N. Nesmeyanov, I.I. Kritskaya, Bull. Acad. Sci. USSR, Div. Chem. Sci. (Eng. Transl.), (1956) 68
Page 69 of 70
Ac ce p
te
d
M
an
us
cr
ip t
243-244. [149] J.S. McKechnie, B. Bersted, I.C. Paul, W.E. Watts, J. Organomet. Chem., 8 (1967) P29-P31. [150] J.S. McKechnie, C.A. Maier, B. Bersted, I.C. Paul, J. Chem. Soc., Perkin Trans. 2, (1973) 138143. [151] M. Löwendahl, Ö. Davidsson, P. Ahlberg, M. Håkansson, Organometallics, 12 (1993) 2417-2419. [152] D. Seyferth, H.P. Withers, Organometallics, 1 (1982) 1275-1282. [153] M. Scheibitz, R.F. Winter, M. Bolte, H.-W. Lerner, M. Wagner, Angew. Chem., Int. Ed., 42 (2003) 924-927. [154] J.A. Schachner, C.L. Lund, J.W. Quail, J. Müller, Acta Crystallogr., E61 (2005) m682-m684. [155] P. Jutzi, N. Lenze, B. Neumann, H.G. Stammler, Angew. Chem., Int. Ed., 40 (2001) 1423-1427. [156] A. Althoff, P. Jutzi, N. Lenze, B. Neumann, A. Stammler, H.-G. Stammler, Organometallics, 21 (2002) 3018-3022. [157] A. Althoff, P. Jutzi, N. Lenze, B. Neumann, A. Stammler, H.G. Stammler, Organometallics, 22 (2003) 2766-2774. [158] M. Herberhold, T. Bärtl, Z. Naturforsch. B, 50 (1995) 1692-1698. [159] J.W. Park, Y.S. Seo, S.S. Cho, D.M. Whang, K.M. Kim, T.Y. Chang, J. Organomet. Chem., 489 (1995) 23-25. [160] D.L. Zechel, D.A. Foucher, J.K. Pudelski, G.P.A. Yap, A.L. Rheingold, I. Manners, J. Chem. Soc., Dalton Trans., (1995) 1893-1899. [161] N.P. Reddy, N. Choi, S. Shimada, M. Tanaka, Chem. Lett., (1996) 649-650. [162] M.J. MacLachlan, J. Zheng, K. Thieme, A.J. Lough, I. Manners, C. Mordas, R. LeSuer, W.E. Geiger, L.M. Liable-Sands, A.L. Rheingold, Polyhedron, 19 (2000) 275-289. [163] G. Calleja, F. Carré, G. Cerveau, P. Labbé, L. Coche-Guérente, Organometallics, 20 (2001) 42114215. [164] A. Berenbaum, A.J. Lough, I. Manners, Organometallics, 21 (2002) 4415-4424. [165] M. Bao, Y. Hatanaka, S. Shimada, Chem. Lett., 33 (2004) 520-521. [166] A. Clearfield, C.J. Simmons, H.P. Withers Jr., D. Seyferth, Inorg. Chim. Acta, 75 (1983) 139-144. [167] T.Y. Dong, M.Y. Hwang, Y.S. Wen, W.S. Hwang, J. Organomet. Chem., 391 (1990) 377-385. [168] G. Utri, K.E. Schwarzhans, G.M. Allmaier, Z. Naturforsch. B, 45 (1990) 755-762. [169] H. Brunner, J. Klankermayer, M. Zabel, J. Organomet. Chem., 601 (2000) 211-219. [170] T. Mizuta, Y. Imamura, K. Miyoshi, Organometallics, 24 (2005) 990-996. [171] C. Spang, F.T. Edelmann, M. Noltemeyer, H.W. Roesky, Chem. Ber., 122 (1989) 1247-1254. [172] N.S. Ieong, W.Y. Chan, A.J. Lough, M.R. Haddow, I. Manners, Chem.-Eur. J., 14 (2008) 12531263. [173] A.S. Perucha, J. Heilmann-Brohl, M. Bolte, H.W. Lerner, M. Wagner, Organometallics, 27 (2008) 6170-6177. [174] D.A. Lemenovskii, I.F. Urazowski, T.V. Baukova, I.L. Arkhipov, R.A. Stukan, E.G. Perevalova, J. Organomet. Chem., 264 (1984) 283-288. [175] I.G. Kuz'mina, Y.T. Struchkov, D.A. Lemenovsky, I.F. Urazowsky, J. Organomet. Chem., 277 (1984) 147-151. [176] B. Bagh, N.C. Breit, J.B. Gilroy, G. Schatte, J. Müller, Chem. Commun., 48 (2012) 7823-7825. [177] J.B. Heilmann, M. Scheibitz, Y. Qin, A. Sundararaman, F. Jäkle, T. Kretz, M. Bolte, H.W. Lerner, M.C. Holthausen, M. Wagner, Angew. Chem., Int. Ed., 45 (2006) 920-925. [178] J.-M. Löwendahl, M. Håkansson, Organometallics, 14 (1995) 4736-4741.
69
Page 70 of 70
S0010-8545(15)00287-8 http://dx.doi.org/doi:10.1016/j.ccr.2015.09.008 CCR 112141
To appear in:
Coordination Chemistry Reviews
Received date: Revised date: Accepted date:
26-6-2015 9-9-2015 15-9-2015
Please cite this article as: H. Bhattacharjee, J. M¨uller, Metallocenophanes Bridged by Group 13 Elements, Coordination Chemistry Reviews (2015), http://dx.doi.org/10.1016/j.ccr.2015.09.008 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Ac
ce
pt
ed
M
an
us
cr
i
Graphical Abstract (for review)
Page 1 of 70
Highlights
The development of strained [1]ferrocenophanes with B, Al, Ga, and In is summarized. The application of [n]metallocenophanes for polymerization is discussed.
Ac ce p
te
d
M
an
us
cr
ip t
Structural aspects of strained and unstrained metallocenophanes are illustrated.
1
Page 2 of 70
Metallocenophanes Bridged by Group 13 Elements
ip t
Hridaynath Bhattacharjee and Jens Müller* Department of Chemistry, University of Saskatchewan, 110 Science Place, Saskatoon, Saskatchewan
cr
S7N 5C9, Canada
us
[email protected]; phone: +1 306 966 2466
an
RECEIVED DATE
ABSTRACT. Metallocenophanes with one or multiple sandwich moieties and at least one group 13
M
element in bridging position are reviewed. The majority of this review summarizes publications
d
concerned with strained sandwich compounds and their application as monomers for metal-containing
te
polymers. Publications up to early 2015 are included.
Ac ce p
Keywords: group 13 elements; ferrocenophanes; ruthenocenophanes; ring-opening polymerization; metallopolymers
Abbreviations: Ar', 2-[(dimethylamino)methyl]phenyl; DSC, differential scanning calorimetry; DLS, dynamic light scattering; Fc, Fe(H4C5)(H5C5); fc, Fe(H4C5)2; FCP, ferrocenophane; GPC, gel permeation chromatography; Mamx, 2,4-di-tert-butyl-6-[(dimethylamino)methyl]phenyl); Me2Ntsi, [(dimethylamino)dimethylsilyl]bis(trimethylsilyl)methyl; Mpysm, [dimethyl(2-pyridyl)silyl]methyl; PFS = poly(ferrocenylsilane); PS, polystyrene; Pytsi, [dimethyl(2pyridyl)silyl]bis(trimethylsilyl)methyl; RCP, ruthenocenophane; ROP, ring-opening polymerization; trisyl, tris(trimethylsilyl)methyl
2
Page 3 of 70
Contents 1. Introduction and scope 2. Aluminum- and gallium-bridged [1]metallocenophanes 2.1. The first generation: trisyl-based ligands
ip t
2.2. Conclusions from the first generation 2.3. The second generation: the Mamx ligand
cr
2.3.1. Metallopolymers
us
2.3.2. DFT calculations 2.4. Conclusions from the second generation
an
2.5. The third generation
4. Boron-bridged [1]metallocenophanes 4.1. Synthesis and mechanistic insights
M
3. Indium-bridged [1]metallocenophanes
Ac ce p
6. [n.n]Metallocenophanes
te
5. [n]Metallocenophanes (n ≥ 2)
d
4.2. Reactivity of bora[1]ferrocenophanes
7. Conclusions
3
Page 4 of 70
1. Introduction and scope [n]Metallocyclophanes are complexes where two metal-coordinated mancude-ring systems are linked by a bridging chain with n atoms in its backbone.1 This is illustrated in Figure 1 for homoleptic (A–C) and heteroleptic compounds (D and E). The largest class is that of [n]metallocenophanes (A), with the
ip t
most examples known for [n]ferrocenophanes (M = Fe). The family of species of type B could be
cr
addressed as [n]metallocenophanes, but are more often called ansa complexes. If one considers the IUPAC definition [1] for metallocenes as sandwich complexes of the type bis(5-
us
cyclopentadienyl)metal, where the d metal is bonded only to the face of two cyclopentadienyl ligands,
an
species of type B are not [n]metallocenophanes as the parent species are not metallocenes. However, in practice, this recommendation is not followed and species such as Cp2MCl2 are usually classified as
M
metallocenes. More importantly, classes A and B are chemically very different. A short bridge in A forces the Cp rings to deviate from their co-planarity in the parent metallocenes Cp2M, which is
d
accompanied by a built-up of intrinsic strain. In contrast, the Cp rings in species B are naturally tilted as
te
the metals in the parent compounds Cp2MX2 (M = Ti, Zr, Hf) are pseudo tetrahedrally coordinated. For
Ac ce p
B the ansa moiety restricts the flexibility of the structure and, commonly, these compounds are used for olefin polymerizations, in particular, in form of their chiral derivatives [2].
The best known class of the heteroleptic sandwich complexes is that of type D, which were coined troticenophane (M = Ti), trovacenophane (M = V), and trochrocenophane (M = Cr) [3-5]. To date, examples of type E are only known for manganese and chromium [6-9]. In addition, other heteroleptic [n]metallocyclophanes are known, which are even rarer; e.g., the CH2SiR2 bridged cobalt complex [Co(4-Me3C4-CH2SiR2-5-C5H4)] equipped with a cyclobutadiene and a Cp moiety [10] and the 1
We define [n]metallocyclophanes similar as IUPAC defined cyclophanes by using the term mancude-ring systems [“rings having (formally) the maximum number of noncumulative double bonds”; IUPAC Gold book (http://goldbook.iupac.org/)]. Therefore, metal-containing “phanes” with metals sandwiched by mancude-ring systems will be referred to as metallocyclophanes. The subgroup where all -coordinated ligands are cyclopentadienyl derivatives will be called metallocenophanes; the subgroup where they are benzene derivatives will be called metalloarenophanes.
4
Page 5 of 70
(tBu2Sn)2 bridged titanium species [Ti{η5-C5H4-(SntBu2)2-η8-C8H7}] equipped with a cyclooctatetraene and a Cp ligand [11].
(ER x)n
X M X
A
(ER x) n
M
B
(ER x) n
ip t
M
C
(ER x)n
(ER x)n
M
M
M
(R xE)
m-1
n
E
F
an
D
us
M
cr
(ER x) n
M
Fig. 1. Metallocyclophanes.
Metallocyclophanes can also exhibit more than one sandwich moiety. Such species are illustrated for
d
metallocenophanes in structure F (Fig. 1), where m expresses the number of sandwich moieties and n is
te
the number of bridging atoms. Using ferrocenophanes (FCPs) as an example, larger ring systems are
Ac ce p
commonly addressed as [nm]ferrocenophanes, whereas for species with just a few repeating units, the meaning of m is usually substituted by repeating n, such as in [n.n]FCPs (m = 2) or [n.n.n]FCPs (m = 3) [12].
Whereas in [n]metallocyclophanes with relatively short bridges the metal-coordinated rings are tilted relative to each other, rings coordinated to the same metal in [nm]metallocyclophanes are coplanar and the structures of the sandwich moieties resembles those of the parent sandwich complex. However, the structures of [n]metallocyclophanes can be significantly different compared to the respective parent sandwich species. This is illustrated for [1]metallocyclophanes in Figure 2, with the common set of angles that is used to measure the degree of distortion. The distortion or intrinsic strain of such a
5
Page 6 of 70
molecule translates into stored energy, a fact that makes these species potential candidates for polymerizations.
ER x
M
us
cr
Fig. 2. Common angles to characterize distortions in [1]metallocyclophanes.
ip t
'
Shortly after the first [n]metallocyclophane was reported by Rinehart et al. in 1957 [A; M = Fe; (ERx)n =
an
(CH2)2C(=O)] [13], the first strained sandwich compound, a [2]FCP with a C2Me4 bridge, was described by the same group [14]. The first strained sandwich compounds with just one atom in bridging positions
M
were silicon-bridged [1]FCPs [ERx = SiPh2, Sifc; fc stands for Fe(H4C5)2] published by Osborne et al. in 1975 [15]. 15 Years later, the first strained bis(benzene) complexes, SiPh2-bridged
d
[1]metalloarenophanes (C) with vanadium and chromium, were published by Elschenbroich et al. [16].
te
In 2004-2005, the research groups of Tamm, Elschenbroich, and Braunschweig, respectively, described
Ac ce p
the first heteroleptic metallocyclophanes of type D (M = Ti [17], V [18], Cr [19]).
It was Manners et al. who finally realized the use of [1]FCPs as monomers for ring-opening polymerization (ROP) and showed that sila[1]ferrocenophanes give poly(ferrocenylsilane)s (PFSs) with high molecular weights [Mw (R = Me) = 5.2 105 Da relative to PS; Scheme 1] [20]. This discovery kick-started tremendous research efforts and, to date, applying the well-established methods of thermal, transition-metal-catalyzed, photolytic, and anionic ROP for FCPs is an elegant way for the preparation of well-defined metallopolymers [21].
6
Page 7 of 70
Scheme 1. Thermal ring-opening polymerization (ROP) of sila[1]ferrocenophanes [20]. 130 °C (R = Me) Fe
SiR 2 230 °C (R = Ph)
1 n
Si R2
Fe
n
PFS
ip t
The most developed strained sandwich compounds are silicon-bridged [1]FCPs (A; Fig. 1), which can undergo living polymerizations [21-23] allowing for the preparation of block copolymers. PFS with
cr
SiMe2-bridging moieties (R = Me; Scheme 1) is partially crystalline, causing block copolymers to form
us
rod- or plate-shaped micelles in block-selective solvents with PFS as the core. These micelles are living and grow to larger micelles of uniform lengths through the addition of further unimers. This
an
crystallization-driven self-assembly (CDSA) of amphiphilic block copolymers has been used recently as bottom-up approach for the fabrication of uniform nanomaterials (see [24-26] for selected recent
M
publications). Compared to silicon-bridged [1]FCPs, other strained sandwich compounds are far less
d
developed and only [1]FCPs bridged by germanium [27] and phosphorus [28-32], and [2]FCPs bridged
te
by carbon [33] could be polymerized with control over molecular weights and molecular weight distributions. Many other strained sandwich compounds could be converted into metallopolymers, but a
Ac ce p
variety of problems brought some of this research to a standstill. For example, many strained monomers are only accessible with bulky bridging moieties; however, for the targeted polymers, often nonencumbered ligands at bridging elements are needed. Another common problem is the lack of solubility of prepared metallopolymers. In order to further advance the area of ROP of strained sandwich compounds, improved monomers are needed in the future.
This review is concerned with metallocenophanes of type A and F (Fig. 1) that contain at least one group 13 element in the backbone of the bridge. Species of type B are not included as these compounds are unstrained and used for a different purpose than those reviewed here. Readers interested in species of type B are referred to the reviews by Braunschweig et al. [34] and Aldridge and Bresner [35]. As 7
Page 8 of 70
species of type D will not be reviewed, readers are directed to the current, promising developments to polymerize these metallocyclophanes [36-38]. In addition, the following list of reviews could be used as a starting point for an overview of the field of metallocyclophanes (see [3-5, 12, 21, 39-43] for selected reviews). With respect to the group 13 containing species discussed in this review, the reviews by
ip t
Braunschweig et al. dealing with boron-bridged metallocyclophanes are particularly relevant [34, 40].
cr
2. Aluminum- and gallium-bridged [1]metallocenophanes
us
Instead of discussing compounds in the group-order of these elements, we start with aluminum- and gallium-bridged [1]metallocenophanes, followed by indium-bridged species, and then close the
an
discussion with boron-bridged compounds. This order might be a surprise to a reader as the first reported group-13-element bridged [n]metallocenophanes were bora[1]ferrocenophanes, published by
M
Braunschweig and Manners et al. in 1997 [44]. After the follow-up article about these highly strained species by the same research groups appeared in 2000, this chemistry came to a standstill [45].
d
Recently, our group revitalized this dormant chemistry by applying ferrocene moieties that were
te
equipped with alkyl groups for the preparation of new bora[1]ferrocenophanes [46, 47]. These alkylated
Ac ce p
ferrocene derivatives were a rational development of our research on aluminum- and gallium-bridged [1]FCPs; hence, we start our discussion with this family of compounds. Without this background information, a discussion of our most recent discoveries in the area of boron-bridged [1]FCPs would lack the appropriate context.
2.1. The first generation: trisyl-based ligands The first strained sandwich compound bridged by a heavier group-13-element (Al–Tl) is the aluminumbridged [1]FCP 1 [48], which was prepared by a common salt-metathesis reaction [3] as illustrated in Scheme 2. Shortly after this finding, the respective gallium-bridged [1]FCP (2) and failed attempts to prepare an indium-bridged [1]FCP were described [49].
8
Page 9 of 70
Scheme 2. Synthesis of the first aluminum-bridged [1]FCP.
Li + Li
Me3 Si Me 3Si
N AlCl2
Al(Pytsi)
Fe
1
cr
2/3 tmeda
Me 2Si
ip t
Fe
In all these cases, the sterically bulky ligand Pytsi had been used (Fig. 3), which is derived from the
us
well-known trisyl ligand, C(SiMe3)3 [50], by replacing one methyl group with a pyridyl moiety 2. The Pytsi ligand had been prepared for the first time by Eaborn and Smith [51] and, subsequently, attached
an
to a variety of elements [51-55]. We had picked the Pytsi ligand as we searched for a ligand that
M
combined intramolecular donor ability with steric bulkiness for the following reasons. First, the speculation that a ligand capable of intramolecular donation was needed was based on experience in
d
using such types of ligands for heavier group 13 elements in the context of GaN deposition by single
te
source precursors for chemical vapour deposition [56-60]. Ligands equipped with one amino moiety for intramolecular coordination give facile access to monosubstituted species of the type RECl2 in one step
Ac ce p
from LiR and ECl3 (E = Al, Ga; R: mono anionic C-N ligand). The intramolecular formation of an N-E dative bond [61] reduces the Lewis acidity of the group 13 element in RECl2 so that the substitution of the second chloride to give R2ECl is significantly slower than the substitution of the first chloride in ECl3. However, the stabilized reagents RECl2 are still reactive and the remaining chloride atoms can be substituted under the right conditions [62, 63]. Second, a sterically bulky ligand was targeted as preliminary results from our laboratory obtained in 2002 from the reaction of Ar'AlCl2 [Ar' = 2(Me2NCH2)C6H4; Fig. 3] [62] with dilithioferrocene revealed that a [1.1]FCP was formed instead of the targeted [1]FCP. Based on this single result, we hoped that sterics would stabilize the targeted strained
2
The bulky ligand C(SiMe3)3 is called trisyl, as an abbreviation for tris(trimethylsilyl)methyl, and is commonly symbolized by Tsi in formulas. Therefore, we used Pytsi and Me2Ntsi for these trisyl derivatives equipped with donor moieties.
9
Page 10 of 70
alumina[1]ferrocenophanes. This thought implied that the [1]FCP was formed as an intermediate but, because of its insufficiently protected aluminum atom, dimerized to the isolated [1.1]FCP. However, there was no direct experimental evidence that a dimerization had occurred. Furthermore, the idea behind using a trisyl-based ligand was due to the fact that the C(SiMe3)3 group had been applied in the
ip t
past to stabilize elements in unusual coordination environments [50].
cr
Table 1 provides an overview of all known aluminum- and gallium-bridged [1]FCPs and
us
[1]ruthenocenophanes ([1]RCPs) and Figure 3 depicts all ligands that had been used. The alumina[1]ferrocenophane 1 [48] and the galla[1]ferrocenophane 2 [49] were isolated in yields of 31
an
and 59%, respectively, and significant amounts of ferrocene were always produced during the synthesis. In an attempt to improve the preparation of these types of aluminum and gallium compounds, the Pytsi
M
ligand was replaced by the related Me2Ntsi ligand (Fig. 3) [64]. This resulted in the isolation of the
Ac ce p
te
and 68%, respectively.
d
alumina[1]ferrocenophane 3 and the respective gallium compound 4 in significantly higher yields of 97
10
Page 11 of 70
Me 2Si
N
NMe2
Me3 Si Me 3Si
Me3 Si Me 3Si
Me 2Ntsi
Pytsi
N Me2Si
N
ip t
Me2Si
Me3 Si Me 3Si
cr
Mpysm
(2-Py)(SiMe 3)2 C
NMe2
us
NMe2
Ar' (R = H) p-tBuAr' (R = tBu) p-Me 3SiAr' (R = SiMe 3)
an
R
M
Mamx
Fig. 3. Intramolecularly coordinating ligands employed for aluminum-, gallium-, and indium-bridged
d
metallocenophanes (Mamx stands for methylaminomethyl-m-xylyl [65]; Mpysm stands for [dimethyl(2-
Ac ce p
te
pyridyl)silyl]methyl [66]).
Table 1. [1]FCPs and [1]RCPs bridged by aluminum or gallium. R' M
ER
R'
structural data a
references
1
XRD:b α = 14.9(3), /' = 39.6/40.5, = 167.9, = 94.7(2)
[48, 67]
Ga(Pytsi)
2
XRD:c α = 15.4(2) {16.4(2)}, /' = 38.3 {37.6}/38.9 {38.5}, = 167.0 {166.3}, = 92.68(13) {92.22(13)}
[49, 68]
Fe / H
Al(Me2Ntsi)
3
XRD:c α = 14.33(14), /' = 39.8/39.4, = 168.2, = 94.51(8)
[64, 67]
Fe / H
Ga(Me2Ntsi)
4
XRD:c α = 15.83(19), /' = 38.3/38.4, = 166.8, = 92.75(11)
[64, 67]
Fe / H
Al[C(SiMe3)2 (2-Py)]
5d
---
[69]
M / R'
ER
Fe / H
Al(Pytsi)
Fe / H
67,
11
Page 12 of 70
Ga[C(SiMe3)2 (2-Py)]
6d
---
[69]
Ru / H
Al(Me2Ntsi)
9
XRD:c α = 20.31(19), /' = 39.9/40.6, = 165.2, = 101.3(2)
[67]
Ru / H
Ga(Me2Ntsi)
10
XRD:c α = 20.91(19), /' = 38.6/38.6, = 163.7, = 98.42(13)
[67]
Fe / H
Al(Mamx)
15
DFT:e α = 13.22, /' = 40.61/40.89, = 169.73, = 95.13
[70]
Fe / H
Ga(Mamx)
16
DFT:e α = 15.80, /' = 38.06/37.78, = 167.30, = 91.59
[71]
Ru / H
Al(Mamx)
17
DFT:e α = 19.37, /' = 41.90/41.46, = 166.44, = 102.80
[70]
Ru / H
Ga(Mamx)
18
DFT:e α = 22.90, /' = 37.49/37.90, = 163.23, = 98.06
[70]
Fe / CHMe2
GaAr'
29
XRD: α = 16.26(9) {16.45(10)}; /' = 40.0(2) {38.9(2)}/39.1(2) {38.4(2)}; = 166.65(3) {166.78(3)}°; = 93.44(10) {92.94(10)}°
[46]
us
cr
ip t
Fe / H
M
an
[a] Angles given in degree; values in braces are from a second independent molecule of the asymmetric unit. [b] The published value of 43.1° for species 1 is wrong. The new values were obtained from the published CIF file with MERCURY [72]. [c] Only the and angles were reported; other angles were obtained from published CIF files with MERCURY [72]. [d] 5 and 6 were not isolated, but could be characterized by 1H NMR spectroscopy in reaction mixtures. [e] Data from optimized geometries on the BP86/TZ2P level of theory.
In the same publication that describes the [1]FCPs 3 and 4, the first aluminum- and gallium-bridged
d
[1]metalloarenophanes were described, namely, [1]chromarenophanes and [1]vanadarenophanes
te
equipped with (Me2Ntsi)Al or (Me2Ntsi)Ga in bridging positions [64]. Subsequently, the first
Ac ce p
[1]molybdarenophanes were prepared by applying the same bridging moieties and synthetic methodologies [73]. As usual, these strained [1]metalloarenophanes were prepared by salt-metathesis reactions using dilithiated sandwich compounds [16, 74-80]. A second motivation to replace the Pytsi moiety with a related one came from our first attempts to prepare [1]chromarenophanes from (LiH5C6)2Crtmeda and (Pytsi)ECl2 (E = Al, Ga). 1H NMR spectroscopy of reaction mixtures gave no indication that the targeted [1]chromarenophanes had been formed, instead, significant amounts of bis(benzene)chromium were present. We speculated that the pyridyl group was the source of protons to produce the parent sandwich compound from the highly basic starting material (LiH5C6)2Crtmeda [81]. As mentioned before, the synthesis of the [1]FCPs 1 and 2 always resulted in FeCp2 in reaction mixtures, which also hinted at the presence of a source of protons [64]. Consequently, a new stabilizing
12
Page 13 of 70
ligand that was similar to Pytsi (Fig.3) was searched for, and resulted, among others [64], in the known ligand Me2Ntsi (Fig. 3) [82]. The use of Me2Ntsi gave not only higher yields for [1]FCPs [64], but also access to the first aluminum- and gallium-bridged [1]metalloarenophanes [64, 73].
ip t
Altering ligands from Pytsi to Me2Ntsi resulted in a change from a five- to a four-membered ring that incorporates aluminum or gallium (Fig. 3 and Scheme 2). At the same time these two ligands were
cr
applied in our group, a similar five- to four-membered ring alteration was tested through applying the
us
(2-Py)(SiMe3)2C ligand. As illustrated in Figure 3, (2-Py)(SiMe3)2C is derived from Pytsi by cutting out the SiMe2 linkage. Salt-metathesis reactions with [(2-Py)(SiMe3)2C]EX2 (EX2 = AlCl2, AlBr2, GaCl2)
an
gave the expected [1]FCPs 5 or 6 (Table 1) as proven by 1H NMR spectroscopy, but their isolation
M
failed and instead of the monomers, low-molecular-weight polymers were obtained.
Moving one step down from iron to ruthenium to prepare [1]RCP is tempting, as the larger Ru atom will
d
result in a significant increase of the tilting of the Cp rings compared to the [1]FCP equipped with the
te
same bridging moiety ERx. An increased α angle should result in an increased strain that, on the other
Ac ce p
hand, should facilitate ROP. To date, only six [1]RCPs are described in the literature [67, 70, 83]. In 2004, Manner et al. described the first [1]RCPs which were prepared with ZrCp'2 (7) and SnMes2 (8) in bridging positions (Fig. 4) [83].
Ru
ZrCp'2 7
Ru
SnMes 2
8
Fig. 4. The first [1]RCPs (Cp' = C5H4tBu; Mes = 2,4,6-Me3C6H2) [83].
As the α angles in tin-bridged [1]FCPs (α = 13.5 – 15.7°) [46, 84-89] are very similar to those in
13
Page 14 of 70
aluminum- or gallium-bridged [1]FCPs (Table 1), it was expected that [1]RCPs could also be prepared with the latter two elements. Consequently, the use of the Me2Ntsi ligand was expanded to [1]RCP and species 9 and 10 were obtained (Table 1) [67].
ip t
The four aluminum- and gallium-bridged [1]FCPs, equipped with the Pytsi (1, 2) or the Me2Ntsi (3, 4) ligand, and the two [1]RCPs 9 and 10 were isolated through crystallization from organic solvents. The
cr
molecular structures for all these six species were solved by single crystal X-ray analysis. The common
us
distortion angles α,, , and (Fig. 2) are listed in Table 1, whereas molecular structures of selected
M
an
examples are shown in Figure 5, 6, and 7.
Si3
C21
Si1 Al1
d
Fe1
C16
te
N1
Ac ce p
Si2
Fig. 5. Molecular structure of 1 with thermal ellipsoids drawn at the 30% probability level. H atoms and half a molecule of FeCp2 are omitted for clarity. The figure was prepared with ORTEP-3 for Windows [90] using the published crystallographic data [48].
14
Page 15 of 70
Si3 C6
Fe1
Ga1 Si1 N1 C1
cr
ip t
Si2
Fig. 6. Molecular structure of 4 with thermal ellipsoids drawn at the 50% probability level. H atoms and
us
half a molecule of benzene are omitted for clarity. The figure was prepared with ORTEP-3 for Windows
M
an
[90] using the published crystallographic data [64].
Si2
Al1 N1
Si1
Si3
Ac ce p
C6
te
Ru1
d
C1
Fig. 7. Molecular structure of 9 with thermal ellipsoids drawn at the 50% probability level. H atoms and half a molecule of n-hexane are omitted for clarity. The figure was prepared with ORTEP-3 for Windows [90] using the published crystallographic data [67].
The aluminum- and gallium-bridged [1]FCPs 1–4 and [1]RCPs 9 and 10 we tested for their propensity to form metallopolymers using thermal-, anionic-, transition-metal-catalyzed-, and photocontrolled ROP methodologies [67]. However, only for one monomer using one ROP method metallopolymers were obtained, while all other attempted ROP were either sluggish or unsuccessful [67]. The 15
Page 16 of 70
galla[1]ferrocenophane 2 could be polymerized with 2 mol% of [Pd(dba)2] (dba = dibenzylideneacetone) in toluene or thf at ambient temperature or 40 °C during 48 h. Characterization of a polymer sample (toluene; 25 °C; 48 h) through Gel Permeation Chromatography (GPC) revealed an Mw of 21.1 kDa (relative to PS) with a broad molecular weight distribution (PDI = 3.0). Attempted
ip t
anionic ROP using a variety of conditions did not result in any reaction of the monomers 3, 4, and 9, or in the case of species 1 and 2 resulted in a substitution of the H atom in para position of the pyridyl
Fe
E
SiMe 3 SiMe2 N
an
Me 3Si
us
spectroscopy, however, their isolation as pure compounds failed [67].
cr
moiety of the Pytsi ligand (Fig. 8). The unexpected [1]FCPs 1Bu and 2Bu were characterized by NMR
M
1 Bu (E = Al) 2 Bu (E = Ga)
te
d
Fig. 8. Substitution products of attempted anionic ROP of [1]FCPs 1 or 2 [67].
Ac ce p
From these mainly unsuccessful polymerization attempts it was concluded that the bulkiness of the employed trisyl-based ligands, Pytsi and Me2Ntsi, prevented polymerizations from occurring. It was speculated that their bulkiness blocked the initiation step or that steric interactions of ligands with the sandwich moieties in the targeted polymers thermodynamically disfavoured their formation [67]. Even though experimental results allowed only to speculate on any details, it was clear that the bulkiness of the ligands attached to the bridging elements needed to be reduced in order to obtain aluminum- or gallium-bridged [1]metallocenophanes that were prone to ROP.
2.2. Conclusions from the first generation As mentioned before, the first attempt in our group to prepare an aluminum-bridged [1]FCP by applying
16
Page 17 of 70
Ac ce p
te
d
M
an
us
cr
ip t
Ar'AlCl2 as the reagent in salt-metathesis reactions gave an [1.1]FCP (11; Scheme 3).
17
Page 18 of 70
Scheme 3. Salt-metathesis reactions resulting in [1.1]FCPs 11-13 [91, 92] and 14 [93].
Me2 N E
A
Fe
Fe
EX2 = AlCl2, GaCl2, InI2
ip t
+ Ar'EX2
E NMe2
Li
us
Fe
cr
11 (E = Al) 12 (E = Ga) 13 (E = In)
Li
Me 2 Si Me2N
B
C(SiMe3 )2 In
X = Cl, I
Fe
Fe In
(Me 3Si)2C
M
+ (Me 2Ntsi)InX2
an
2/3 tmeda
NMe 2 Si Me 2
te
d
14
Ac ce p
As illustrated in Scheme 3A, respective reactions also gave [1.1]FCPs for gallium (12) and indium (13) [92]. As shown in Scheme 3B, employing (Me2Ntsi)InX2 in salt-metathesis reactions also resulted in a [1.1]FCP (14; Scheme 3B) [93]. Note that [1.1]metallocenophanes will be discussed in detail in Chapter 6. The latter result was a surprise, as in respective reactions with (Me2Ntsi)ECl2 (E = Al, Ga) the targeted aluminum- or gallium-bridged [1]FCPs were formed [64]; 1H NMR spectroscopy did not reveal any signs that [1.1]FCPs were also produced (Chapter 2.1). These discoveries revealed that the bulkiness of the ligand attached to the group 13 element has a significant influence on the outcome of salt-metathesis reactions of dilithioferrocene with REX2 (E = Al, Ga, In). The fact that the bulky Me2Ntsi ligand attached to indium gave a [1.1]FCP (14), whereas the same bulky ligand on aluminum or gallium gave [1]FCPs (3, 4; Table 1) can be rationalized based on structural data as illustrated in
18
Page 19 of 70
cr
ip t
Figure 9.
us
Fig. 9. Illustration of the proximity between the Me2Ntsi ligand and the ferrocene moieties (doubleheaded arrows) in the diinda[1.1]ferrocenophane 14 (left) and the galla[1]ferrocenophane 4 (right) [93].
an
HH distances in 14 are 2.15, 2.35, 2.36, and 2.38 Å (SiMe3-H); HH distances in 4 are 2.28 and
M
2.44 Å (SiMe3-H) and 2.35 and 2.53 Å (NMe2-H). Adapted with permission from [93].
d
Copyright © 2008 American Chemical Society.
te
The area of the shortest intramolecular HH distances in the indium species 14 as well as in the gallium
Ac ce p
species 4 were found between the methyl groups of the Me2Ntsi ligand and the α protons of ferrocene moieties [93]. This is indicated with double-headed arrows in Figure 9 and reveals that the available space for a bridging moiety ER in [1.1]FCPs is smaller than in respective [1]FCPs. Indium with a covalent radius of 1.44 Å is significantly larger than aluminum and gallium, which both have nearly the same size (r = 1.25 and 1.26 Å) (all three values for singly bonded, 3-fold-coordinated elements) [94]. Therefore, In-C bonds are significantly longer than respective Al- or Ga-C bonds and, consequently, for indium species, a bulky ligand R of the bridging moiety in a metallocenophane is further away from the sandwich unit. This comparison is only meaningful if the coordination geometry of the element E stays the same for the discussed species, which is the case for the aluminum, gallium, and indium compounds. The structural data as illustrated in Figure 9 lead to the conclusion that aluminum- or gallium-bridged
19
Page 20 of 70
[1.1]FCPs equipped with the Me2Ntsi ligand cannot be formed as the [1.1]FCP structure simply does not offer enough space for the bridging moieties. Instead, [1]FCPs were obtained in salt-metathesis reactions of (Me2Ntsi)AlCl2 or (Me2Ntsi)GaCl2 with dilithioferrocene as the geometry of the strained sandwich compound offers enough space. For indium, a [1]FCP as well as a [1.1]FCP is accessible , as
ip t
both types of compounds offer enough space. The fact that only the [1.1]FCP species was found has been rationalized by assuming that the intrinsic strain in a [1]FCP disfavours its formation and, at the
cr
same time, favours the formation of the unstrained [1.1]FCP [93]. The rationalization why [1.1]FCPs
us
are formed instead of the targeted [1]FCPs when steric hindrance does not play a major role, was also based on the results obtained by employing the one-armed phenyl ligands Ar' or p-tBuAr' (Fig. 3). For
an
aluminum and gallium, switching from bulky trisyl-based ligands (Pytsi, Me2Ntsi; Fig. 3) to the rather flat one-armed phenyl ligand Ar' (Fig. 3) changed the outcome of the salt-metathesis reaction from
M
[1]FCPs to [1.1]FCPs. A similar influence was also observed for bis(benzene)sandwich systems: trisylbased ligands gave [1]metalloarenophanes (E = Al, Ga; M = V, Cr, Mo) [64, 73], while the non-
te
Ac ce p
= Cr, Mo) [95].
d
encumbered ligands Ar' and p-tBuAr' (Fig. 3) gave respective [1.1]metalloarenophanes (E = Al, Ga; M
One can assume that salt-metathesis reactions are non-equilibrium reactions and, therefore, are kinetically controlled. Hence, steric interactions between groups should be discussed for transition states, in particular, for those of the rate determining steps. However, for the above discussed saltmetathesis reactions nothing is known about the kinetics and, therefore, only possible steric interactions for ground-state geometries of products were discussed. We assume that this is a meaningful approach as significant steric interactions in a ground-state geometry of a product between a ligand on the bridging element and the sandwich moiety should also to some extent be present in the transition state leading to this product. Of course, reducing the arguments for a net chemical reaction that is probably composed of different elementary reactions to just arguments of steric interactions in ground-state
20
Page 21 of 70
geometries is clearly an oversimplification. Still, this approach provided our research group with the working hypothesis that the formation of strained [1]FCPs bridged by heavier group 13 elements will only occur if sterically bulky groups block or hinder the formation of the otherwise preferred [1.1]FCPs.
ip t
Inspection of structural data and the following considerations led to the conclusion that the space available for the bridging unit ER decreases from [1]FCPs to poly(ferrocene)s to [1.1]FCPs (Fig. 10).
cr
The parallel alignment of both ferrocenediyl moieties in a [1.1]FCP forces the ligand R to be oriented in
us
such a way that possible steric interactions between R and the sandwich part are maximized. Compared to the relative orientation of the ER group in [1.1]FCPs, that in [1]FCPs is twisted by an angle of ca.
an
90°, which results in a minimization of possible steric interactions. It can be assumed that the relative orientation of the bridging units ER in respective poly(ferrocene)s are somewhere in-between that in
M
[1]FCPs and [1.1]FCPs and, therefore, the available space for ER will be in-between that in [1]FCPs
te
d
and [1.1]FCPs.
E R
Ac ce p
Fe
ER
Fe
n
[1]FCP
PF
Fe
R E Fe E R [1.1]FCP
decreasing space available for ER
Fig. 10. Space restrictions in [1]FCPs, poly(ferrocene)s, and [1.1]FCPs.
On the basis of all considerations discussed above, for the first generation of aluminum- or galliumbridged species one can conclude that strained [1]metallocenophanes were only accessible because the bulky trisyl-based ligands Pytsi or Me2Ntsi did not allow the formation of [1.1]metallocenophanes.
21
Page 22 of 70
However, for steric reasons polymerizations were blocked or sluggish as targeted poly(ferrocene)s offer less space for these bulky ligands compared to the strained monomers.
ip t
2.3. The second generation: the Mamx ligand Our tactics to advance the chemistry of strained aluminum- and gallium-bridged [1]metallocenophanes
cr
was to use steric hindrance cautiously so that [1]metallocenophanes could be accessed that were not
us
overly protected and, therefore, polymerizable. The lessons learned from the first generation of strained sandwich compounds resulted in our working hypothesis that steric hindrance is needed to block or
an
hinder the formation of [1.1]metallocenophanes so that the targeted [1]metallocenophanes are accessible. In Figure 11, the H atoms that exhibit the shortest intramolecular HH distances in the
te
d
M
Ar'Ga-bridged [1.1]FCP 12 are indicated by solid black spheres [46].
Ac ce p
N
Ga
Fe
Fe
Ga
N
Fig. 11. Illustration of the shortest HH distances in the [1.1]FCP 12 (see also Scheme 3). Adapted with permission from [46]. Copyright © 2013 Wiley-VCH.
From this molecular structure it is apparent that a bulky alkyl group adjacent to gallium on the phenyl 22
Page 23 of 70
ring would disfavour or even block the formation of a respective [1.1]FCP. In order to use this blocking tactics, group 13 dihalide species equipped with a modified Ar' ligand were needed (Fig. 12). From the ample theoretical possibilities, we settled on the use of the known Mamx ligand (Fig. 3 and Scheme 4). This bulky aryl ligand had been introduced by Yoshifuji et al. and the needed MamxBr [65, 71] is
ip t
conveniently prepared from 3,5-di-tert-butyltoluene (Scheme 4). Finally, lithium-bromine exchange gives access to the aluminum or gallium dichlorides (Mamx)ECl2 [70, 71] needed for salt-metathesis
us
cr
reactions.
NMe 2
an
EX2 R
M
Fig. 12. Proposed (RAr')EX2 species to avoid formation of [1.1]FCPs.
Ac ce p
te
d
Scheme 4. Mamx-containing aluminum and gallium species [70, 71]. NMe2 Br 1) nBuLi, -78 °C, Et2O MamxBr
2) ECl3 , Et2O
Me 2N
Me2N
E
M
M
E
n Fe Al Ga
Ru
15n 17 n 16 n 18 n
Al Ga
N Me2
Fe
Ru
15 16
17 18
Cl2E (LiC 5H 4) 2M ·tmeda (Mamx)ECl 2
The second tBu group, which is in para position with respect to aluminum or gallium in (Mamx)ECl2, is 23
Page 24 of 70
needed to facilitate the preparation of the targeted aryl bromide by employing a symmetrical rather than an unsymmetrical starting compound (Scheme 4). However, this para-tBu group does not interact sterically with the Cp moieties in metallocenophanes. As shown in Scheme 4, this approach led to [1]FCPs (15, 16) and respective [1]RCPs (17, 18). However, from the four strained species only 18
ip t
could be isolated, while the other three species underwent ROP in reaction mixtures. Still, the isolable gallium-bridged [1]RCP (18) is similarly reactive as species 15, 16, and 17, and polymerized when left
cr
stirring in the reaction mixture [70]. Even though species 15–17 were not isolable, they could be
us
unequivocally identified by 1H NMR spectroscopy from reaction mixtures, in particular, by chemical shifts and signal pattern of Cp protons, which are characteristic for Cs symmetrical aluminum- or
M d
4.4
2.06
4.00
4.6
15
N Me2
4.00 FeCp 2
4.51
Al
Ac ce p
4.72 4.70
Fe
tBu
tBu
te
3.85
4.2
4.0
ppm
2.01
an
gallium-bridged [1]FCPs with donor-stabilizing ligands (see Fig. 13).
Fig. 13. Characteristic shifts and pattern of α- and -protons of 15. 1H NMR spectrum was taken from an aliquot of the reaction mixture after ca. 30 min (C6D6). Adapted with permission from [70]. Copyright © 2012 American Chemical Society.
24
Page 25 of 70
2.3.1. Metallopolymers Molecular weights (Mw) between 8.07 and 106 kDa for polymers 15n–18n (Scheme 4) were estimated by dynamic light scattering (DLS) [70, 71]. GPC was only applied to the polyferrocenylgallane 16n and
ip t
resulted in Mw of 48 kDa (relative to PS; PDI = 3.3) a comparable value as that estimated via DLS (36 kDa) [71]. As the applied characterization methods of GPC and DLS provide only relative values of Mw
cr
the absolute values for all four polymers are unknown [70, 71]. Among the 1H NMR spectra of all four
us
polymers that of 16n stands out as it shows an unusually well resolved signal pattern for the ortho-tBu
Ar
m
Ga Fe
Ar
m
Ga Fe
B
Ar Ga
Fe
M
isotactic
A m
Ga Fe
Ar
r
r
Fe
Ga
Ar Ga
r
Ar
Ga
Fe
Ac ce p
Fe
Ar Ga
Ga Fe
syndiotactic Fe
Ar
d
Ar
C
te
heterotactic
an
group (Fig. 14).
Fig. 14. Illustration of different triads in 16n [diads indicated as r (25aceme) and m (meso)]. The gallium-coordinated NMe2 group at the aryl ligand (Mamx) is omitted for clarity. 1H NMR signal of the ortho-tBu group of 16n exhibiting pentad resolution with an approximate intensity ratio of 1 : 2 : 1 for A : B : C. Reproduced with permission from [71]. Copyright © 2010 American Chemical Society.
The splitting of the signal of the ortho-tBu group into three groups (A, B, and C; Fig. 14) is caused by a triad sensitivity. The approximate intensity ratio of 1 : 2 : 1 (A : B : C) reveals that polymer 16n has a random distribution of stereogenic gallium atoms [71]. This interpretation is further supported by the
25
Page 26 of 70
observed fine structure within the three groups. Whereas the less intense signal groups A and C each appear as three singlets with a 1 : 2 : 1 intensity ratio, the intense group B is split into a set of four singlets of similar intensity. Such an array of 10 singlets is expected for a polymer with a random distribution of pentads, as from 16 possible pentads, only 10 are distinguishable [71]. Why is the ortho-
ip t
tBu group such a sensitive probe for the tacticity of the polymer? This question was addressed when DFT calculations were employed to better understand the reactivity of the monomers 15-18 (see below)
us
cr
[70].
2.3.2. DFT calculations
an
The high reactivity of monomers 15-18 was unexpected. In particular, expected tilt angles α of 15 to 16° for the [1]FCPs 15 and 16 should not impose enough strain that prevents their isolation from reaction
M
mixtures. Furthermore, it was not understood why the only isolable [1]metallocenophane was a [1]RCP (18) and not a less strained [1]FCP. DFT calculations (BP86/TZ2P) [70] were employed to obtain
d
structural as well as thermo-chemical data. In particular, the role of the ortho-tBu group was
te
investigated. This tBu group acts as a sensitive probe for the polymer tacticity and its signal was fully
Ac ce p
resolved into all possible pentads in 16n (Fig. 14). It was speculated that this unusually high sensitivity could be caused by an intimate contact between the ortho-tBu group and ferrocene moieties of the polymer backbone.
R 2 10
M
E 20
3 1
4
6
N Me 2
5
7
Al Ga
R = tBu
R=H
Fe
Ru
Fe
Ru
17 18
15 H
17 H 18 H
15 16
16
H
Fig. 15. Calculated strained sandwich compounds to evaluate the structural effect of the ortho-tBu group of the Mamx ligand.
26
Page 27 of 70
The geometries of the eight strained sandwich compounds (Fig. 15) were optimized to provide structural data. Four of the eight species are identical to the synthesized [1]FCPs and [1]RCPs (15-18), whereas in the second set of four compounds, the ortho-tBu groups on the Mamx ligands were replaced by hydrogen (15H-18H). Comparison of geometric parameters of related pairs (e.g., 15 and 15H) revealed
ip t
the influence of the ortho-tBu group. First, tilt angles α were not very sensitive toward the ortho-tBu group (Fig. 2). If this group was absent, the angle α decreased only by 0.86 and 0.52° for FCPs and by
cr
1.08 and 0.82° for RCPs (Table 1) [70]. However, the ortho-tBu group had a significant effect on the
us
orientation of the aromatic ligand with respect to the sandwich moiety. The M−E−C1 angle decreased between 8.08 (16 to 16H) and 12.73° (17 to 17H) and, at the same time, the tilting of the aromatic ring
an
changed (torsion angle M−E−C1−C2 changed between 5.26 and 6.98°). These effects are shown in Figure 16 for the gallium-bridged [1]RCPs 18 and 18H and clearly illustrate that as the tBu group gets
C10
Ru
Ga N
18
Ac ce p
C20
te
d
M
removed, the aromatic ligand moves toward the freed-up space.
C10
Ru
Ga N
C20
18H
N
Ga
N
Ru
Ga
C1
Ru
C1
C2
C2
Fig. 16. Optimized geometries of the 18 and 18H; hydrogen atoms are omitted for clarity. The doubleheaded arrow illustrates the distance of C1 of 1.003 (18) and 1.518 Å (18H) from the plane Ru−C10−C20 (dotted line). Adapted with permission from [70]. Copyright © 2012 American Chemical Society.
27
Page 28 of 70
To address the question if the significant structural changes imposed by the ortho-tBu group also significantly increases the strain of the [1]metallocenophanes requires quantification of changes in energy. Intrinsic strain of a compound can be best described by the enthalpy of a reaction where the strained species is transformed into unstrained species. Therefore, thermodynamic data of the chemical
cr
ip t
reactions shown in Scheme 5 were calculated.
us
Scheme 5. Chemical reactions applied to calculate strain in [1]FCPs and [1]RCPs [70].
+ MCp2
E
M (1)
an
Me 2N R
(1)
R
E = Al, Ga
19
M = Fe, Ru N Me 2
+ 2 H2
M
Ac ce p
15 - 18 15 H-18 H
R = tBu, H
d
E
te
M
M
M(2)
+
R
H2E
(2) N Me 2 20
The isodesmic reaction 1 should be ideal, as calculated reaction enthalpies can be directly taken as the strain of the starting [1]metallocenophanes. However, the size and floppiness of the bis(metallocenyl) compounds (19) precluded reliable frequency calculations and only ΔE values could be determined. Furthermore, it turned out that all four bis(metallocenyl) compounds equipped with the ortho-tBu group (19; R = tBu) were still strained. The ortho-tBu group pointing toward the metallocenyl unit M(2) forced it to be bent away; in contrast, in species 19 with R = H the metallocenyl unit M(2) was bent in the opposite direction. Overall, reaction 1 could not be used to evaluate the strain of the starting
28
Page 29 of 70
[1]metallocenophanes [70].
In contrast to the first reaction in Scheme 5, the hydrogenation reaction (eq. 2) is not isodesmic, and calculated enthalpies are not equal to the intrinsic strain of respective metallocenophanes. However,
ip t
comparison of the hydrogenation of a tBu containing species (e.g. 15) with that of the respective metallocenophane with R = H (e.g., 15H) resulted in differences in the calculated thermodynamic values
cr
(ΔΔE, ΔΔH, and ΔΔG) that directly provided a measure of the thermodynamic effect of the ortho-tBu
us
group. In all four cases, hydrogenations of species equipped with the tBu group resulted in a larger release of energy; i.e., increases of enthalpies (ΔΔH) between -4.80 and -6.33 kcal/mol were found.
an
That means that the tBu group increased the intrinsic strain in the aluminum- and gallium-bridged [1]FCPs and [1]RCPs on average by 5.5 kcal/mol. Compared to dimethylsila[1]ferrocenophane, the text
M
book example of a strained sandwich compound, with a measured enthalpy of -19 kcal/mol (ΔHROP) [20], the increase of strain caused by the ortho-tBu group of 5.5 kcal/mol is indeed substantial. This was
te
d
the first time that such an effect had been revealed in metallocenophane chemistry [70].
Ac ce p
2.4. Conclusions from the second generation
The application of the Mamx ligand to stabilize aluminum or gallium in bridging position of [1]FCPs or [1]RCPs was a partial success as summarized here: 1) The working hypothesis that the use of a bulky group in ortho-position of the Ar' ligand will block or hinder the formation of [1.1]metallocenophanes to give the targeted [1]metallocenophanes withstood the test.
2) In contrast to the first generation of [1]metallocenophanes, those of the second generation were reactive and gave metallopolymers. The Mamx ligand significantly stabilized the group 13 element so that polymers could be handled under ambient conditions and purified by precipitation into methanol [70]. This stabilizing effect had been used recently for the synthesis of other small molecules and
29
Page 30 of 70
polymers equipped with the (Mamx)Ga moiety [96, 97]. 3) The ortho-tBu group acts as a very sensitive NMR probe and revealed a random tacticity of 16n. DFT calculations shed some light on this surprising result and uncovered that the tBu group pushes against a ferrocenyl moiety in model compounds (19; Scheme 5). Based on this result, one can assume that the
ip t
available space in a polymer is similarly restricted so that the tBu group sterically interacts with ferrocenediyl moieties. Such interactions are probably the reasons for the pentad sensitivity.
cr
4) The average increase in strain in monomers caused by the ortho-tBu group is 5.5 kcal/mol. One
us
would be tempted to say that this increased strain is resulting in the unexpected high reactivity. However, the ortho-tBu group imposes a similar strain on the resulting polymers so that overall this
an
effect cancels out. Therefore, it was concluded that kinetics must govern the reactivity of 15-18. It was speculated that small amounts of dilithioferrocene might initiate the ROP of monomers in reaction
mechanism of ROP still in the dark [70].
M
mixtures, but all attempts to get experimental data to support this speculation failed, leaving the
d
5) Out of the four strained monomers, only species 18 could be isolated, which was achieved by
te
precipitation from the reaction mixture at low temperature. Further purification of 18 through
Ac ce p
crystallization failed. The three [1]metallocenophanes 15-17 could not even be isolated. In order to have a chance to prepare a new polymer for a potential niche application, a control over molecular weights is required. However, the therefore needed living ROP processes require monomers of high purity. As our second generation of monomers failed to fulfill this requirement, we proceeded to the next generation of strained sandwich compounds.
2.5. The third generation For the second generation of strained sandwich compounds our working hypothesis that steric hindrance is needed to block or hinder the formation of [1.1]metallocenophanes so that the targeted [1]metallocenophanes are accessible was put into action by increasing the bulkiness of the ligand on the
30
Page 31 of 70
bridging element. This led to the use of the Mamx ligand as discussed and illustrated in Chapter 2.3 (Fig. 11). For the third generation of strained sandwich compounds we increased the bulkiness of the ferrocene moiety. As evident from the ORTEP plot of the [1.1]FCP shown in Figure 11, bulky alkyl groups in place of the H atoms in positions of the Cp rings should render the resulting [1.1]FCP non-
ip t
accessible. As ferrocene moieties are introduced by dilithioferrocene, the preparation of the third generation required -substituted 1,1’-dilithioferrocene derivatives, which we wanted to prepare from
cr
respective bromides through Li/Br exchange. Hence, we searched for a precursor of type 21 (Fig. 17)
R1
an
methodologies resulting in these type of species are rare.
Br
Br
R4
d
21
M
R2 Fe R 3
us
that was equipped with one, two, three, or four alkyl groups in positions. However, synthetic
Ac ce p
te
Fig. 17. Proposed precursor for the preparation of -substituted dilithioferrocene derivatives.
Scheme 6. Preparation of (Sp,Sp)-1,1'-dibromo-2,2'-di(isopropyl)ferrocene (28) [46, 98].
31
Page 32 of 70
(S)
Ph Ph
O N
B Me 60 mol%
AcCl / AlCl3 Fe
22
OAc
(R)
HNMe2
Fe 2) (CCl2 Br)2, thf, -78 °C
Br
Fe
MeOH/H2 O (R)
(R)
25
NMe2
(R)
NMe2
24
OAc
Ac 2O,
(R)
Br (R)
27
Br
AlMe 3 Fe
CH2 Cl2 , -78 °C
Br
28
OAc
an
Br Fe
M
70 °C
OAc
OH
us
Br
(R)
Ac2 O, pyridine
(R)
1) 4 equiv nBuLi, Et2 O, r.t. 12 h
Fe
26
23
O NMe2
NMe2 (R)
Fe
BH 3·SMe 2 , thf, 0 °C
ip t
CH2Cl2
cr
Fe
OH (R)
O
te
d
We used a well-known approach that was developed by Ugi et al. [99], where it was shown that the – C*H(Me)NMe2 group on one Cp rings is an excellent directing group for diastereoselective lithiation.
Ac ce p
Nearly three decades later, Knochel et al. used the enantioselective CBS reduction [100, 101] (22 to 23; Scheme 6) and extended the known “Ugi amine” chemistry to prepare the “double Ugi amine” 25 among other enantiomerically pure, C2 symmetric ferrocene derivatives [98]. Through a lithiationbromination sequence, amine 25 was diastereoselectively converted into the known species 26 [46, 102]. After conversion of 26 to the acetate 27, the central chirality was removed to leave the planarchiral ferrocene framework in 28 in form of the (Sp,Sp)-isomer behind [46].
The first attempts to lithiate species 28 using the standard conditions of two eq. of nBuLi at -78 °C in thf solutions did not result in a clean lithium-bromine exchange [46]. However, applying the method that Bailey et al. developed for aromatic bromides [103], a quantitative lithium-bromine exchange was 32
Page 33 of 70
accomplished at 0 °C using nBuLi in the solvent mixture of thf and hexanes (1 : 9) (Scheme 7). Even though the solution mainly consists of hexanes, the dilithio species of 28 remains dissolved and does not precipitate. Addition of Ar'GaCl2 to the in situ prepared dilithioferrocene derivative of 28 gave the targeted gallium-bridged [1]FCP 29 (Scheme 7). As discussed before, the respective salt-metathesis
ip t
reaction with dilithioferrocene-tmeda gave the unwanted [1.1]FCP 12 (Scheme 3).
1) 2 equiv nBuLi, 0 °C, thf : hexanes (1 : 9), 30 min
Br
2) 1 equiv Ar'GaCl2
GaAr'
an
Fe
Fe
us
Br
cr
Scheme 7. Synthesis of the gallium-bridged [1]FCP 29 [46].
29
M
28
Checking the reaction mixture by 1H NMR spectroscopy (Fig. 18) revealed that the salt-metathesis
d
reaction resulting in 29 is quite clean as, in addition to the signals caused by this C1 symmetrical
te
product, only weak signals were detected (Fig. 18). The strained [1]FCP was isolated by crystallization
Ac ce p
from the reaction mixture in a yield of 59% [46]. Its molecular structure was solved by single crystal Xray analysis (Fig. 19) and the determined tilt angles for the two independent molecules [16.26(9) and 16.45(10)] and the other common distortion angles , , and are very similar to those of the first generation of gallium-bridged [1]FCPs (see Table 1).
33
Page 34 of 70
Fe Ga N Me2 29
cr
ip t
*
us
Fig. 18. Cp range of the 1H NMR spectrum of 29, taken from an aliquot of the reaction mixture,
an
approximately 5 min after the addition of Ar'GaCl2. The doublet marked by an asterisk is caused by one of the two methylene protons (for the full spectrum see Fig. S7 in reference [46]). Adapted with
C7 C6 C17
Fe1
Ac ce p
Ga1
te
d
M
permission from [46]. Copyright © 2013 Wiley-VCH.
C1
N1
C2
Fig. 19. Molecular structure of 29 with thermal ellipsoids drawn at the 50% probability level. Hydrogen atoms are omitted for clarity. One of two independent molecules is shown. The figure was prepared with ORTEP-3 for Windows [90] using the published crystallographic data [46].
The DSC thermogram of 29 looks as expected for a strained sandwich compound (Fig. 20). However, the endothermic melting peak (onset ca. 187 C) partly overlaps with the exothermic peak (Tmax = 201 34
Page 35 of 70
C); the latter is indicative of ROP. Because of the peak overlap, the HROP could not be measured precisely, but could be estimated to be more negative than -33 kJ mol-1 [46]. This value is very similar to that of similarly tilted tin-bridged [1]FCP tBu2Snfc [ = 14.1(2)] where HROP of -36(±9) kJ mol-1
M
an
us
cr
ip t
was determined [85].
te
[46]. Copyright © 2013 Wiley-VCH.
d
Fig. 20. Differential scanning calorimetry (DSC) thermogram of 29. Reproduced with permission from
Ac ce p
In summary, employing isopropyl groups in -position on Cp rings successfully suppressed the formation of [1.1]FCP and resulted in an isolable and reactive [1]FCP (29). As the enantiomerically pure dibromoferrocene derivative 28 could cleanly be transformed into its dilithio derivative (Scheme 7), we explored its applicability for the synthesis of new [1]FCPs with a variety of bridging elements. To date, elements of group 13 (B [46, 47], Ga [46], In [104]), 14 (Si, Sn [46]), and 15 (P [105]) were successfully incorporated into the bridging position. Some of these results are discussed in Chapters 3 and 4.
3. Indium-bridged [1]metallocenophanes
35
Page 36 of 70
In the first attempt to prepare an inda[1]ferrocenophane the same tactic was followed that led to the first aluminum- and gallium-bridged [1]FCPs (1 and 2; Scheme 2 and Table 1) [48, 49, 68]. However, the 1,1'-disubstituted ferrocene species 30 (Fig. 21) was the only isolable product from the salt metathesis between (Pytsi)InCl2 and (LiC4H4)2Fe2/3tmeda [49]. Changing the stoichiometry of the reaction from
ip t
1 : 1 to 2 : 1 resulted in species 30 in a yield of 57%. At the time of this discovery it could not be clarified why species 30 instead of the targeted [1]FCP was found. As the starting dichloride
cr
(Pytsi)InCl2 formed dimers in the solid state with an In-(-Cl)2-In moiety similar to that found in
us
species 30, the authors speculated that a monomeric indium dihalide might be required for the
an
preparation of an indium-bridged [1]FCP [49].
M
Si23
Si21 Si22
In2
Cl2
d
N21 Fe1
te
Cl1
In1
Ac ce p
N1
Si2
Si1
Si3
Fig. 21. Molecular structure of 1,1'-[(Pytsi)InCl]2fc (30) with thermal ellipsoids drawn at the 50% probability level. Hydrogen atoms and solvent molecules are omitted for clarity. The indium-chlorine bridge is asymmetric [In1-Cl1 = 2.8322(15), In2-Cl2 = 2.8655(15), In1-Cl2 = 2.5087(14), In2-Cl1 = 2.5072(15) Å]. The figure was prepared with ORTEP-3 for Windows [90] using the published crystallographic data [49]. As discussed in Chapter 2.2, employing (Me2Ntsi)InX2 in salt-metathesis reactions resulted in the 36
Page 37 of 70
[1.1]FCP 14 (Scheme 3B) instead of the targeted [1]FCP [93]. This result helped to rationalize the influence of the bulkiness of the bridging moiety on the outcome of the salt-metathesis reaction and led to the working hypothesis that steric hindrance is needed to block or hinder the formation of [1.1]metallocenophanes so that the targeted [1]metallocenophanes are accessible (Chapter 2.3).
ip t
Because indium is significantly larger than aluminum or gallium, a ligand R on indium will result in a smaller steric interaction with a sandwich moiety than the same ligand on aluminum or gallium.
cr
Scheme 8. Synthesis of the first indium-bridged [1]FCPs [104].
us
tBu Me 2N
tBu
In
Fe
In
R' = H
N Me2
+
an
tBu
tBu
+ (Mamx)InCl 2
Fe
M
31 (19%)
+
In tBu
Fe
In
tBu
A
Fe
tBu
Me 2N
n
NMe2
31 2 (23%)
31 n (32%)
d
tBu
Fe Li R'
B + Ar'InCl2
te
Li
R' = iPr
iPr
In
Fe
iPr
+
N Me 2
R' = iPr
N Me 2 iPr 33
Fe
In
tBu
In
Fe
iPr
NMe2 32 2
iPr tBu
+ (Mamx)InCl 2
Fe
iPr
32
C
In
iPr
iPr
Ac ce p
R'
Me2 N
Fe
In (Mamx)
n
33 n (33%)
Scheme 8 shows our recent attempts to prepare inda[1]ferrocenophanes where we combined the 37
Page 38 of 70
chemistry of our second and third generation of aluminum and gallium species [104]. The reaction of (LiC4H4)2Fe2/3tmeda and (Mamx)InCl2 gave a separable mixture of monomer (31), dimer (312), and oligomers (31n) (Scheme 8A). Similarly, reaction of the iPr-substituted dilithioferrocene derivative (generated in situ from dibromide 28) with Ar'InCl2 gave an inseparable mixture from which the
ip t
monomer 32 and its dimer 322 were clearly identified by 1H NMR spectroscopy as the main products (Scheme 8B). Combining the bulky Mamx ligand and iPr groups on ferrocene resulted in a relatively
cr
clean conversion into the inda[1]ferrocenophane 33 (Scheme 8C). However, all attempts of isolation
us
failed as 33 polymerized in the reaction mixture resulting in the poly(ferrocenylindigane) 33n (DLS analysis: Rh = 2.4 0.1 nm; Mw = 24 2 kDa; see [104]). This reactivity is reminiscent of our second
an
generation of the strained sandwich compounds (15-18; Scheme 4) that were equipped with the Mamx ligand, but had no iPr groups on Cp rings. That 33 readily polymerized is still surprising if one
M
considers that it should be significantly less strained than the [1]FCPs (Mamx)Alfc (15) and
d
(Mamx)Gafc (16).
R' Fe
R'
InR
Ac ce p
te
Table 2. Calculated distortion angles in indium-bridged [1]FCPs.
structural data a
references
31
α = 11.44, /' = 37.60/37.62, = 169.99, = 86.24
[104]
Ar'
32 b
α = 10.96, /' = 39.03/38.85, = 170.37, = 87.49
[104]
CHMe2
Mamx
33
α = 11.42, /' = 37.62/37.79, = 170.22, = 86.12
[104]
H
Ar'
34 c
α = 11.18, /' = 38.42/38.27, = 170.12, = 87.08
[104]
R'
R
H
Mamx
CHMe2
[a] Angles given in degree; data from optimized geometries on the BP86/TZ2P level of theory [104]. [b] 32 was not isolated, but could be characterized by 1H NMR spectroscopy in a mixture with 312 (Scheme 8). [c] The hypothetical species 34 (Ar'Infc) was included in the DFT calculations, but had not been synthesized (Scheme 8).
38
Page 39 of 70
Because crystals of the new indium-bridged [1]FCPs could not be obtained, DFT calculations were performed to get structural data. In order to deduce the effects of the tBu and the iPr groups, the hypothetical inda[1]ferrocenophane 34 (Table 2) was included in this study. As shown in Table 2, the distortion angles do not change significantly; e.g., the tilt angles were calculated between 10.96 and
ip t
11.44 for 31-34.
cr
The main difference between the structures of the four compounds 31-34 is due to the relative
us
orientation of the aromatic ligand (Mamx or Ar') with respect to the ferrocene moiety. Similar as it was found for the set of [1]FCPs (15, 16, 15H, 16H) and [1]RCPs (17, 18, 17H, 18H) (Fig. 15 and Fig. 16), the
an
relative orientation changes with the degree of substitution. Overall the structural effects are significant but less pronounced for the indium-bridged [1]FCPs (31-34) compared to the aluminum- and gallium-
M
bridged species (15-18; 15H-18H) [104]. In order to address the question if the iPr and tBu groups influence the strain of the [1]FCPs, the thermochemistry of a hydrogenolysis reaction was investigated
d
(Scheme 9). As discussed before (eq. 2 in Scheme 5), such a hydrogenolysis is a non-isodesmic reaction
te
and only a comparison of two reactions that only differ in the presence or absence of one type of alkyl
Ac ce p
group (iPr or tBu) gives meaningful data that reveals the thermodynamic effect of this particular group. This gave Ho298 between -0.17 to -4.39 kcal mol-1 meaning that in all cases the presence of the particular set of alkyl groups resulted in an increased exothermy of the hydrogenolysis. Species 33 as the most encumbered and structurally distorted species showed with -2.83 (effect of iPr groups) and 4.39 kcal mol-1 (effect of tBu groups) the largest Ho298 values. That means that the alkyl groups significantly increase the intrinsic strain of the inda[1]ferrocenophane 33, which might be causing its unexpectedly high reactivity. Usually, a tilt angle α around 11° is an expression of an amount of strain that is insufficient for ROP. Obviously, compound 33 is an exception.
39
Page 40 of 70
Scheme 9. Hydrogenolysis reaction to evaluate strain in inda[1]ferrocenophanes [104].
Fe
In
+ N Me2
R'
2 H2
Fe
+
H2In
R'
R'
R
H iPr iPr H
tBu H tBu H
R
R N Me 2
us
cr
31 32 33 34
R'
R
ip t
R' R
4. Boron-bridged [1]metallocenophanes
an
In 1997, the first boron-bridged [1]FCPs (36, 37; Table 3) were published by Braunschweig and
M
Manners et al. [44]. This report made some waves in the “phane” community as these species were the first reported [1]FCPs with an element from the second period in bridging position. Second, among the
d
two species prepared, the (Me3Si)2NB-bridged species 37 was structurally characterized revealing a new
te
record for the tilting of the Cp rings. Species 37 with = 32.4(2)° (Fig. 22 and Table 3) replaced the record holder at the time, which was the sulphur-bridged [1]FCP with = 31.05(10)° [106, 107]. To
Ac ce p
date, a total of nine boron-bridged [1]FCPs are known in literature (35-43; Table 3), all of which are equipped with amino groups at boron [44-47].
40
Page 41 of 70
Table 3. Boron-bridged [1]FCPs. R Fe
R1 B N R2
R
R
NR1R2
structural data a
H
NiPr2
35
α = 31.0(2) {31.4(2)}, /' = 35.3(3) {35.6(3)}/34.3(3) {36.1(3)}, = 155.9(2) {156.1(2)}, = 102.0(3) {103.2(3)}
[45]
H
N(SiMe3)tBu
36
---
[44, 45]
H
N(SiMe3)2
37
α = 32.4(2), /' = 33.7(2)/34.0(2), = 155.2(2), = 100.1
CHMe2
NEt2
38 b
---
CHMe2
NiPr2
39
α = 31.9(2), /' = 36(1)/35(1), = 155(2), = 103.0(3)
CHMe2
N(SiMe3)tBu
40 b
---
CHEt2
NEt2
41
---
CHEt2
NiPr2
42
---
CHEt2
N(SiMe3)tBu
43
---
M
an
us
cr
ip t
references
[44, 45]
[47] [46] [47] [47] [47] [47]
Ac ce p
te
d
[a] Angles given in degree; values in braces are from a second independent molecule of the asymmetric unit. [b] Isolated as impure oils [47].
Si1
C10
B1
N1
Fe1
C20
Si2
Fig. 22. Molecular structure of the first boron-bridged [1]FCP (37) with thermal ellipsoids drawn at the 50% probability level. Hydrogen atoms are omitted for clarity. The figure was prepared with ORTEP-3
41
Page 42 of 70
for Windows [90] using the published crystallographic data [44]. After the first report on boron-bridged [1]FCPs appeared in 1997 [44], a follow-up article was published 3 years later [45] describing the synthesis of another [1]FCP (35). The N(SiMe3)2 (37) as well as the N(iPr)2-substituted [1]FCP (35) were structurally characterized revealing highly distorted sandwich
ip t
species (Table 3 and Fig. 22). Nearly 20 years after its publication, species 37 is still the FCP with the most tilted Cp rings [ = 32.4(2)° and = 155.2(2)°]. Recently, our group prepared a series of six
cr
species equipped with three different amino groups based on our third generation of strained FCPs (see
us
Chapter 2.5). These six enantiomerically pure species fall into two sets: one with isopropyl groups (3840; Table 3) and one with 3-pentyl groups (41-43; Table 3) on Cp rings. Out of these six species the
an
molecular structure of only the iPr2NB-bridged compound 39 equipped with iPr groups on ferrocene could be determined in the solid state (Fig. 23). As shown in Table 3, the values for the distortion angles
te
d
M
, , , and , respectively, are very similar for 35, 37, and 39.
C2
Ac ce p
C1
B1
N1
C6
C7
Fig. 23. Molecular structure of the enantiomerically pure [1]FCP 39 with thermal ellipsoids drawn at the 50% probability level. Hydrogen atoms are omitted for clarity. The figure was prepared with ORTEP-3 for Windows [90] using the published crystallographic data [46].
42
Page 43 of 70
4.1. Synthesis and mechanistic insights Similar to all other group-13-bridged [1]FCPs, the boron-bridged compounds 35-43 were prepared by the common salt-metathesis approach using a dilithiated sandwich compound and respective
ip t
amino(dichloro)boranes. Whereas the 35-37 were isolated in yields of 35-44% [45], yields for the chiral [1]FCPs varied between 48% (43) and 65% (42) [47]. The latter yields resulted from optimizing the
cr
reaction conditions for the salt metathesis as follows. Species 39 was the first boron-bridged [1]FCP we
us
made in our group and it was isolated in a yield of only 20% [46]. As shown in Scheme 10, similar as the preparation of the gallium-bridged [1]FCP 29 (Scheme 7), the chiral dilithioferrocene was prepared
an
in situ from the precursor 28 by Li/Br exchange at 0 °C (Scheme 10). After 30 minutes stirring at this temperature, the cold bath was removed and, simultaneously, the addition of a solution of iPr2NBCl2 in
M
hexanes over 10 min was started [46].
Fe Br 28
1
iPr
1) 2 equiv nBuLi, 0 °C, thf : hexanes (1 : 9), 30 min
Ac ce p
Br
te
d
Scheme 10. Preparation of bora[1]ferrocenophane 39.
Fe
B=NiPr2
2) 1 equiv Cl2 BNiPr2
39
B
+
Fe iPr Cl B NiPr2
NiPr 2
Cl 45
H NMR data of the reaction mixture showed the presence of the targeted [1]FCP 39 and the unwanted
bis(boryl)ferrocene 45. The detected ratio for 39 : 45 of 1.0 : 3.5 is comparable to the isolated yield of 20% for 39 [46]. The yield of 39 was improved to 53% by heating the solution of the in situ prepared dilithioferrocene derivative first to 50 °C, before addition of the dichloride [47]. This initiated a series of experiments by combining three different amino(dichloro)boranes and two different dilithioferrocene species in salt-metathesis reactions. Whereas employing iPr2NBCl2 and tBu(Me3Si)NBCl2 to prepare
43
Page 44 of 70
respective [1]FCPs led to the formation of significant amounts of bis(boryl)ferrocenes, the Et2NBbridged [1]FCPs 38 and 41 formed with such high conversions that the small peaks of byproducts in the proton NMR spectra could not reliably be assigned to the expected bis(boryl)ferrocenes. Therefore, the effect of temperature on the product distributions was investigated for the synthesis of the [1]FCPs 39,
ip t
40, 42, and 43 (Table 4).
cr
Table 4. Measured product ratios between [1]FCPs and bis(boryl)ferrocenes (Scheme 11) [47].a
us
reaction temperature NR1R2
product ratio
0 °C r.t.
50 °C
CHMe2
NiPr2
39 : 45
1.0 : 2.4
1.0 : 0.59
CHMe2
NtBu(SiMe3)2
40 : 46
1.0 : 1.4
CHEt2
NiPr2
42 : 48
1.0 : 0.81
CHEt2
NtBu(SiMe3)2
43 : 49
1.0 : 0.52
M
an
CHR2
1.0 : 0.47 1.0 : 0.30 1.0 : 0.12
te
d
[a] Approximate ratios determined by 1H NMR spectroscopy from aliquots of the reaction mixture, taken ca. 20 min after the addition of R1R2NBCl2.
In all four cases, increasing the temperature to 50 °C significantly favoured the formation of the targeted
Ac ce p
[1]FCPs (Table 4). This effect was explained by an obvious reaction mechanism as illustrated in Scheme 11 [47]. Both products stem from a common intermediate (INT), which can either intramolecularly react to form the targeted [1]FCPs or intermolecularly react with additional amino(dichloro)borane. The later reaction channel will be disfavoured by low concentrations of the amino(dichloro)borane and, therefore, slow addition of a dilute solution of this reagent is important. The fact that increased reaction temperature favours the formation of [1]FCPs more strongly reveals that this pathway has the highest activation barrier. This makes sense as some of the strain present in the [1]FCPs must already be built-up in the transition state leading to this species. Based on this insight, the synthesis of the 35 (iPr2NBfc) was reinvestigated using the optimized conditions and the reported yield of 38% [45] was improved to 74% [47]. 44
Page 45 of 70
CHR2 Li(solv) x
k1
+ Cl2 BNR1R 2
R2 HC (solv) x Li
Fe R 2HC (solv) xLi
INT
k3
k2
cr
Fe
CHR2 NR 1R 2 B Cl
R1 Fe
B N
an
R2 HC Cl B
CHR2
us
+ Cl2BNR1R 2 CHR2 NR1 R2 B Fe Cl
ip t
Scheme 11. Proposed reaction mechanism [47].
R2
CHR2
NR 1R 2 44-49
M
38-43
d
In addition to the effect of temperature, data in Table 4 reveal the influence of the different alkyl groups
te
on ferrocene on product ratios (isopropyl compared to 3-pentyl). For all the respective cases, the 3pentyl substituted sandwich species resulted in higher conversions to the targeted [1]FCPs; for example,
Ac ce p
the product ratio at a reaction temperature of 50 °C for the tBu(SiMe3)N-substituted species are 1.0 : 0.47 for iPr and 1.0 : 0.12 for 3-pentyl at ferrocene (Table 4). This effect had been attributed to a steric influence of the CHR2 group (R = Me, Et) on the rate of substitution on R1R2NBCl2. Based on crystallographic data, the most stable conformation of the alkyl groups CHR2 (R = Me, Et) on the ferrocene moiety shows one of the two R groups approximately in the same plane as a Cp ring, whereas the second R group is oriented away from iron and nearly perpendicular to a Cp ring [47]. This is illustrated in Figure 24 with a Newman projection for one CHR2 group at a Cp ring.
45
Page 46 of 70
R R
H Fe
Fig. 24. Newman projection to illustrate the conformation of CHR2 groups on Cp rings.
ip t
The rate constant k2 should be unaffected by the type of alkyl group on the ferrocene moiety, because in
cr
both cases (iPr and 3-pentyl) hydrogen points to the inner area of the sandwich compound where the ring closure happens (Scheme 11). However, the rate constants k1 and k3 will be influenced by the type
us
of alkyl group. For steric reasons one must assume that a borane molecule R1R2NBCl2 will approach a lithiated Cp ring from the least hindered side, which will be opposite to the Fe atom. That means that
an
one R group of the CHR2 substituent will point toward the direction of the incoming borane and, hence,
M
R = Et will reduce k1 and k3 more strongly than R = Me [47]. In short, 3-pentyl groups suppress the formation of bis(boryl)ferrocene derivatives (k3; Scheme 11) more effectively than iPr groups with the
te
d
overall effect that 3-pentyl groups favour the formation of [1]FCPs.
4.2. Reactivity of bora[1]ferrocenophanes
Ac ce p
One of the main motivations to prepare strained sandwich compounds is to use their potential to access new metallopolymers through ROP methodologies. For bora[1]ferrocenophanes one would expect that these highly strained molecules will show a high propensity towards polymerization. Unfortunately, these expectations have not been met yet. As transition-metal-catalyzed ROP works well for silicon- and germanium-bridged [1]FCPs [21, 108-110], Braunschweig and Manners et al. investigated the reactivity of 35-37 towards transition metals. However, employing catalytic amounts of Karstedt’s catalyst did not yield polymers [45]. Even Pt(cod)2, a reagent known to result in just a single insertion of a Pt(cod) moiety into the C-Si bond of Me2Sifc [110], did not result in a reaction with monomer 36 when applied in an equimolar amount [45].
46
Page 47 of 70
However, reactions with carbonyl complexes led to some surprising results. As shown in Scheme 12, insertion of metal carbonyl fragments into Fe-Cp bonds yielded unstrained species. Species 50 was either obtained by using Fe(CO)5 or Fe2(CO)9 in isolated yields of 26 and 64%, respectively [45]. This complex is a boron-bridged analogue of the well-known dimer [Cp(OC)Fe(-CO)]2; a Me2Si-bridged
ip t
analogue of 50 was already published in 1973 [111]. A similar insertion reaction of a metal carbonyl moiety into the Fe-Cp bond had been observed in the reaction with Co2(CO)8 that led to the complex 51
cr
(Scheme 12). Both products were characterized by single crystal X-ray analysis [45]. Similar reactions
us
with carbonyl complexes are not reported for other [1]FCPs, and it is likely that a significant weakening of Fe-Cp bonds, as caused by the high degree of tilting of the Cp rings in bora[1]ferrocenophanes, is
an
needed for such reactions to occur.
B=N iPr
or Fe 2(CO) 9
tBu
Fe
B=N
Co2(CO) 8
iPr B=N iPr
Fe
50
(OC) 2 Co
SiMe3
36
Fe
OC
Ac ce p
35
OC OC
d
Fe
OC
Fe(CO) 5, h
te
iPr
M
Scheme 12. Reactions of bora[1]ferrocenophanes with metal carbonyl complexes.
(OC) 4Co
(CO) 2 Fe
tBu B=N SiMe 3 51
The three [1]FCPs 35-37 were polymerized thermally at 180 °C (37) and 200 °C (35-36). While 36 gave only insoluble material, a minor fraction of the product mixture obtained from monomer 37 was soluble in C6D6. 1H NMR analysis of the dissolved fraction showed signals that were interpreted as being caused by the dimer [(Me3Si)2NBfc]2 [45]. Thermal ROP of 35 resulted in a mixture that was completely soluble in organic solvents. Precipitation from toluene into hexanes gave material (34% 47
Page 48 of 70
yield) that was the targeted polymer as concluded from NMR data. However, this material did not give detectable signals in dynamic light scattering measurements, meaning that the hydrodynamic radius of this polymer was below the detection limit of 2.3 nm, a value that would be around Mn of 9 kDa for PFS
ip t
[45].
From the six chiral [1]FCPs, the two 3-pentyl substituted species 41 and 42 were selected for thermal
cr
ROP as their DSC thermograms looked promising. The obtained glassy solids [TROP = 240 °C (41), 260
us
°C (42)] were dissolved in thf leaving elemental iron particles behind. Repeated precipitations and filtrations resulted in isolation of polymeric materials in 73% (41n) and 70% (42n) yields [47]. GPC
an
analysis gave estimated values of Mw of 10 kDa (relative to PS) for the main elution peak for both polymers. However, CHN analysis did not give the expected values revealing that isolated materials
M
were not the targeted linear polymers in pure form (Scheme 13).
1
R
R
Fe
B NR1 R2
Ac ce p
Fe
R B N R2
te
R
R
d
Scheme 13. Thermal ROP of bora[1]ferrocenophanes.
n
35 -37 (R = H)
41, 42 (R = CHEt2 )
48
Page 49 of 70
ip t cr
us
Fig. 25. DSC thermogram of 42 (heating rate: 10 °C min-1; TROP (onset) = 222 °C; HROP = -63(±5)
an
kJ mol-1). Reproduced with permission from [47]. Copyright © 2014 Wiley-VCH.
M
DSC was used to determine the exothermy of the ROP for 35-37 [45] and the 3-pentyl substituted species 41-43 [47]. Figure 25 displays the DSC thermogram of 42, which shows the expected features;
d
i.e., an endothermic melting peak followed by an exothermic peak indicating the occurrence of ROP.
te
However, for bora[1]ferrocenophanes 35 and 36 overlap of the two peaks prevented the determination
Ac ce p
of HROP [45], whereas for species 43 multiple exothermic peaks were found with the main peak being very broad [47]. These investigations resulted in HROP values of -95 (37), -75(±5) (41), and -63(±5) (42) kJ mol-1. In addition, as we had improved the synthesis for the iPr2NB-bridged [1]FCP 35, its DSC thermogram was remeasured confirming that endothermic and exothermic peaks overlap; however, based on the data a HROP of -73 kJ mol-1 was estimated. These measured heat releases are smaller than what is expected for [1]FCPs with angles greater than 30° [112]. The values should be similar to 130(±20) kJ mol-1 as determined for the sulfur-bridged [1]FCP with an angle of 31.05(10)° [107]. For comparison, the Me2Si-bridged [1]FCP with a tilt angle of 20.8(5)° [113] gave a HROP of ca. -80 kJ mol-1 [20]. This was already pointed out by Braunschweig and Manners et al. and the reduced enthalpy for 37 (HROP = -95 kJ mol-1) was rationalized by the presence of “side group interactions” 49
Page 50 of 70
between the sterically demanding amino groups, which presumably resulted in substantial destabilization of the polymer relative to the monomer [45]. As mentioned above, thermal ROP of 3pentyl decorated [1]FCPs 41 and 42 resulted in iron particles in addition to polymers. Examining the crucibles after DSC measurements of 35, 41, and 42 also showed the presence of elemental iron. It was
ip t
assumed that the thermal extrusion of iron occurs from the starting compound and is a radical process, which is either slightly exothermic or even endothermic. Therefore, the measured exothermy in DSC
cr
thermograms is not representative of the intrinsic strain of [1]FCPs, as a clean ring-opening reaction is
us
not occurring [47].
an
In conclusion one can say that bora[1]ferrocenophanes had not lived up to the expectation with respect to polymer formation. The first set of [1]FCPs (35-37) was plagued with low solubilities of the obtained
M
materials preventing their full characterization. The presence of alkyl groups on the second set of [1]FCPs (41, 42) helped to increase the solubilities of the obtained polymers, but they were of low
d
molecular weight and were not the pure linear polymers. The thermal stability of monomers is so high
te
that the required temperatures for ROP lead to extrusion of iron. One can speculate that amino groups at
Ac ce p
the bridging boron contributes significantly to the stability of the [1]FCPs. Hence, getting access to bora[1]ferrocenophanes that are aryl or alkyl substituted at boron might be key for the preparation of new boron-containing polymers in a controlled way [114, 115].
50
Page 51 of 70
5. [n]Metallocenophanes (n ≥ 2) Many metallocenophanes with more than one atom in bridging position are known; Figure 26 summarizes boron-containing species and Figure 27 summarizes species of the heavier group 13
ip t
elements. With the exception of the boron-bridged species 61, all other compounds are ferrocene
cr
derivatives that cover the range from [2]- to [4]FCPs.
Fe
Fe
E
BR
Fe 54
M
NMe2 R B C
NMe2 B R C Me 2N C B X B C NMe2 C R B Me2N
[PF6]2 N N B Fe
d
Fe
C
te
R' B NMe2
56
Ac ce p
R = Me, Ph, SiMe 3 R' = H, Me,
R
an
53 E = O, S, Se, Te, NR', BNMe 2, CNtBu, CNCy, Pt(PEt3 )2 R = NMe 2, NiPr 2, Br, Me
X
N B Ph NMe 2
BR
52 R = NMe 2, Mes
55
us
BR
BR
Fe
NMe 2 B Ph N
(X = no X, EC 2 H2, 1,4-C 6 H4 ),
O B N N
Fe
57
X = E-C 2H2, 1,4-C 6 H4 R = SiMe3
Pt(PEt 3) 2 R B Fe
R
E
E
E
Fe
BR
Fe
E
B 58
E
R'
E = PPh2 and/or
59 N
R''
N
R/R' = Br, Me, PPh2, OEt, N 2C 3 H3 , NC4 H 8, N2C 5F6 H, N 2C 4 H6 +
E = O, S, NR'
E
B B
Co NMe2
60 E = S, NSiMe3
R = H, F, Cl, Br, Me, Et, nBu, thexyl, OPh, Ph, Mes, NEt 2, NiPr 2, NiPrEt, NPhEt, N(CH 2Ph)Me, Fc
R R B
NMe2
61
OR' B RR
R = iPr, F R' = H, Me
51
Page 52 of 70
Fig. 26. Boron-bridged [n]metallocenophanes (n ≥ 2) (references: 52 [116-118], 53 [119-123], 54 [124],
Fe
SiMe3 O 2 Al
Fe
63
Py N SiMe 3
64
SiMe3 N Fe
62
N
R E
Do N SiMe 3
SiMe 3 N Cl Ga Cl N H SiMe3
cr
E = Al, Ga
In Fe
Cl
N
In Pytsi
30
d
E = Al, Ga R = Me, Et (for Al), Me, Et, Cl (for Ga)
Cl
M
65
SiMe 3
Pytsi
an
SiMe3 N
us
R = H, F, Cl, Br, I, Me, Et, tBu, CH(SiMe3 )2 , Si(SiMe 3) 3 , C R' (R' = nBu, tBu, Ph, SiMe3 ) Do = Py, NEtMe2
Fe R 2E ER 2
ip t
55 [120, 125], 56 [125, 126], 57 [127], 58 [128-132], 59 [133-137], 60 [136, 138], 61 [139, 140]).
te
Fig. 27. Aluminium-, gallium-, and indium-bridged [n]metallocenophanes (n ≥ 2) (references: 30 [49],
Ac ce p
62 [141-145], 63 [146], 64 [145], 65 [143, 145]).
The dibora[2]ferrocenophane 52 (R = NMe2) was reported for the first time in 1997 by Herberhold and Wrackmeyer et al. where a common salt-metathesis was used by reacting freshly prepared dilithioferrocene with the diborane(4) [(Me2N)ClB]2 [116]. Species 52 (R = NMe2) was obtained in a yield of 58% and characterized by NMR spectroscopy and mass spectrometry. The authors noted that the X-ray structural analysis was presented by A. Appel and H. Nöth at a conference; however, it seems that this analysis has not been published.3 Ten years later, species 52 (R = NMe2) was prepared by the flytrap route [3] (61% isolated yield) [117] and its structure analysis showed with = 12.82(16)° the 3
In ref. 2 of publication [122] it was indicated that the structural data for 52 are included in the Diploma Thesis or PhD Thesis of A. Appel (Faculty of Chemistry and Pharmacy, University of Munich, 1993/1996).
52
Page 53 of 70
presence of intrinsic strain. The second known dibora[2]ferrocenophane, 52 (R = Mes), was isolated in 37% and single crystal X-ray analysis revealed a similar tilt angle as for its amino-substituted counterpart [ = 10.5(2)°] [118]. An entire series of [3]FCPs (53) and [4]FCPs (54-56) were obtained by insertion reactions into the B-B bond of the amino-substituted dibora[2]ferrocenophane 52. The same
ip t
chemistry had been explored by using the respective dibora[2]chromarenophane [120, 121, 124]. The diborylation of alkynes was accomplished either stoichiometrically or catalytically, whereas the latter
cr
pathway was performed hetero- or homogeneously (Scheme 14). For stoichiometric pathway the
us
platinum species 53 [E = Pt(PEt3)2; R = NMe2] was prepared and isolated first and then reacted with alkynes. However, addition of catalytic amounts of Pt(PEt3)3 followed by addition of a 10-fold excess of
an
alkyne gave the respective [4]FCPs 55 in high yields (Scheme 14). Similar results were obtained by heterogeneously applying Pt or Pd metals. Obviously, species 53 [E = Pt(PEt3)2; R = NMe2] and
M
respective metal-insertion products are key intermediates in this process. As expected, the platinum compound 53 [E = Pt(PEt3)2; R = NMe2] is less strained compared to its precursor 52 (R = NMe2) as
d
indicated by the small tilt angle of 5.0° (Fig. 28) [120]. On the basis of these first discoveries [120],
te
Braunschweig et al. explored this process further to get access to species 53 (E = CNtBu, CNCy; R =
Ac ce p
NMe2 [121], S and Se [123]), 54 [124], 55 [125], and 56 [125, 126].
Scheme 14. Insertions into the B-B bond of 52 (R = NMe2) [120].
53
Page 54 of 70
R
NMe 2 R B C
R'
Fe
Fe B
Pt sponge or Pd charcoal
NMe2
52
C R' B NMe 2
55 R = Me R' = H, Me
NMe2 B Fe
R'
Pt(PEt3 )2 B NMe2
M
an
us
53
R
cr
[Pt(PEt3) 3]
ip t
B
NMe2
N1 P1 B1 Fe1
d
Pt1
te
B2
Ac ce p
P2
N2
Fig. 28. Molecular structure of the insertion product 53 [E = Pt(PEt3)2; R = NMe2] with thermal ellipsoids drawn at the 50% probability level. Hydrogen atoms are omitted for clarity. The figure was prepared with ORTEP-3 for Windows [90] using the published crystallographic data [120].
R EH
B Fe R B X
X
Fe HE
66
67
X = Cl, Br
E = O, S, NR
54
Page 55 of 70
Fig. 29. Two types of starting compounds for the preparation of FCPs.
The preparation of other species shown in Figure 26 that were not obtained through insertions can be sorted into two groups: one with the ferrocene-containing reagent acting as the electrophile (66) and one
ip t
where it is acting as the nucleophile (67; Fig. 29). The latter reagent was either reacted with a respective element halide reagent directly (in the presence of NEt3 to trap HX) or it was first deprotonated and its
cr
dilithio derivative was applied. An example for the use of a bis(boryl) species (66) is the reaction of
us
fc[BBr(NiPr2)]2 with Li2Se that gave the respective [3]FCP (53 with R = NiPr2 and E = Se; Fig. 26) [119]. Among the compounds depicted in Figure 26 the tribora[3]ferrocenophane 53 (E = BNMe2; R =
an
NMe2) cannot be sorted into any of the two bins as it was synthesized using the common salt-metathesis
M
for FCPs starting from dilithioferrocene and the triborane Me2NB[BCl(NMe2)]2 [122].
The ansa-bridges in species of type 58 are formed by donor-acceptor bonds that many derivatives were
d
prepared by Wagner et al. with a particular emphasis on the dynamic behaviour [128-132]. Depending
te
on the Lewis acidity of the boryl groups and the Lewis basicity of the bridging moieties E only the
Ac ce p
closed form, the open form, or equilibria were found (Fig. 30) [132].
B Fe
E
R
R
B
E
B
R'
Fe
E
R' B
close-58
E = PPh2 and/or
E
N
open- 58 R''
N
Fig. 30. Open and closed form of species 58.
Species 61 stands out as the only non-iron metallocenophane and is an inverse chelate between the 1,1'-
55
Page 56 of 70
bis(boryl)cobaltocenium ion and RO- (Fig. 26).
As shown in Figure 27, the group of aluminum-, gallium-, or indium-bridged [2]metallocenophanes is smaller than the group of boron species (Fig. 26). An analogue of the reactive dibora[2]ferrocenophane
ip t
52 is not known for the heavier group members. Uhl et al. mentioned that all of their attempts to prepare such a [2]FCP bridged by a Ga2 moiety failed, but led to the synthesis of the digalla[1.1]ferrocenophane
cr
72 (Chapter 6) [147]. All the aluminum- and gallium-bridged species were synthesized by Wrackmeyer
us
et al. [141-146] while the only indium-bridged species (30; Fig. 27; see also Chapter 3) had been prepared in our group [49]. The common starting material for compounds 62-65 is 1,1'-
an
bis(trimethylsilylamino)ferrocene (67, E = NSiMe3; Fig. 29) that was either reacted directly with hydrides (H3Al·NEtMe2, H3Al·py; [142, 144, 146]), alkyl species (AlR3, GaR3; R = Me, Et; [143, 145]),
M
or first converted into its dilithio derivative, which was reacted with a variety of aluminum or gallium
te
6. [n.n]Metallocenophanes
d
halides [141, 142, 145].
Ac ce p
The first [1.1]FCPs were carbon-bridged species (Fig. 31; ERx = CH2, CHPh) that had been described by Nesmeyanov and Kritskaya in 1956 [148]. Carbon-bridged [1.1]FCPs were later structurally characterized as syn-isomers (Fig. 31) [149, 150], which is the preferred conformation with the small carbon in bridging position (for an exception see [151]). In the first decades of metallocenophane chemistry, researchers focused on carbon-bridged species, an area that was reviewed in 1986 by Mueller-Westerhoff [12] (see also the review [39]). After the first hetero-atom bridged [1.1]FCP was described by Seyferth et al. in 1982 [Fig. 31; ERx = Sn(nBu)2, SnEt2] [152], this class has grown substantially and [1.1]FCPs with following bridging elements are known: B [153], Al [66, 91, 92, 154], Ga [92, 147, 155-157], In [92, 93, 104], Si [109, 158-165], Sn [85, 87, 88, 152, 166, 167], Pb [168], P [30, 169, 170], As [171], S [172], Zn [173], and Hg [174, 175]. Furthermore, [1.1]RCPs [12, 158] and
56
Page 57 of 70
mixed [1.1]metallocyclophanes [12] with Fe/Ru and Fe/Co, respectively, had been described. In contrast, related arene complexes are very rare and, to the best of our knowledge, the aluminum and gallium-bridged [1.1]chromarenophanes and [1.1]molybdarenophanes are the only examples of this family [95]. From the lists of species above, [1.1]FCPs are the only group-13-bridged
ip t
[1.1]metallocenophanes described in literature. As one can see from Table 5, which provides an overview of all known species, the majority was structurally characterized by single crystal X-ray
cr
analysis and with the exception of the boron compound 68 and the gallium compound 74, all other
us
species exhibit anti conformations in the solid state (Fig. 31).
Fe Fe
Fe ER x
an
R xE E Fe Rx
anti
M
E Rx syn
d
Fig. 31. syn and anti isomers of [1.1]ferrocenophanes
te
Table 5. [1.1]Ferrocenophanes bridged by group 13 elements.
Ac ce p
Rx E
R'
R' Fe
R'
R'
Fe
E Rx
R'
ERx
references
H
BMe2–
68 a
[153]
H
AlAr'
11 a
[66, 91, 92]
H
AlCl(tmeda)
69 a
[154]
H
Al(p-Me3SiAr')
70 a
[66]
H
Al(Mpysm)
71
[66]
H
Ga[CH(SiMe3)2]2
72 a
[147]
H
GaMe(NC5H5) b
73 a
[155, 156]
57
Page 58 of 70
[155, 156]
H
GaMe
75 a
[157]
H
GaAr'
12 a
[66, 92]
H
Ga(p-Me3SiAr')
76 a
[66]
H
Ga(Mpysm)
77
[66]
H
GaAr' / SiMe2 c
78
[176]
H
Ga(p-SiMe3Ar') / SiMe2 c
79 a
[176]
H
InAr'
13 a
[92]
H
In(Me2Ntsi)
14 a
[93]
H
In(Mamx)
312
[104]
CHMe2
InAr'
322
[104]
ip t
74 a
cr
Ga(fc)1/2(NC5H5)
us
H
an
[a] Single crystal X-ray data available. [b] A series of compounds with other donors had been reported in [157]. [c] silagalla[1.1]ferrocenophane
The first reported group-13-bridged [1.1]FCPs were the species 72 and 73 4 and the unusual
M
metallocenophane 74 [155]. Whereas the majority of species shown in Table 5 were prepared by the
te
condensation route (Scheme 15).
d
common salt-metathesis route, the preparation of the gallium compounds 73 and 74 were obtained by a
Ac ce p
Scheme 15. Unusual condensation route to species 73 and 74.
4
Species 72 [147] was published before 73 and 74 [155], however, the latter contribution was submitted two months earlier than that describing 72.
58
Page 59 of 70
N
Me 2 Ga
Me 3Ga
Fe
Me 2 Ga Fe Fe
80
pyridine
Ga Me 2
Ga Me 2
SnCl3
GaMe 2 Fe GaMe 2
n 81
82
N Fe
Fe
Me Ga
- Me3Ga·py
Fe Fe 73
N
M
N
Ga Me
an
Ga 74
toluene,
Fe
+
us
N Ga
cr
N
ip t
SnCl3
d
Heating of the donor-stabilized bis(gallyl)ferrocene 82 resulted in a redistribution of the two ligands Me
te
and fc at gallium, yielding the three Lewis acid-base adducts 73, 74, and Me3Ga·py (Scheme 15). Such
Ac ce p
a redistribution of ligands will require association of species through diborane(6)-type moieties. Such moieties are present in the precursor 81 [156], which was sparingly soluble in non-coordinating solvents. The carousel molecule 74 was isolated as a pure, crystalline compound (Fig. 32) by heating the mixture of 73 and 74 in a high-boiling solvent (e. g., toluene or p-xylene), followed by cooling [155]. In this process, further conversion of 73 into 74 had been observed, a fact that shows that 74 and Me3Ga·py are the end products of the ligand redistribution reaction with the [1.1]FCPs 73 being an intermediate. This type of condensation reaction is reminiscent of that found by Jäkle, Holthausen, and Wagner, where a 1,1'-bis(boryl)ferrocene gave poly(ferrocenylborane)s and diborane(6) [177].
59
Page 60 of 70
N1
Ga1
ip t
Fe1 Fe2
cr
Fe2A
us
Ga1A
an
N1A
M
Fig. 32. Molecular structure of the carousel molecule 74 with thermal ellipsoids drawn at the 50% probability level. Hydrogen atoms and solvent molecules are omitted for clarity. The figure was
te
d
prepared with ORTEP-3 for Windows [90] using the published crystallographic data [155].
Ac ce p
To date, species 68 is the only reported boron-bridged [1.1]FCP [153].5 As shown in Scheme 16, the species had been synthesized by a Lewis acid-base adduct formation between dilithioferrocene and 1,1'bis(dimethylboryl)ferrocene (83) and obtained as the crystalline salt Li(12-crown-4)2[68-Li]. According to the single crystal X-ray analysis, one of the two counter cations is chelated by crown ethers, whereas the other one is trapped inside the cavity of the [1.1]FCP. 7Li NMR data suggests that this unusual structure is preserved in solution as two signals in a 1 : 1 ratio were found. The authors speculated that the lithium ion is “most likely trapped by the electrostatic field originating from the two anionic dimethylborate bridges” [153].
5
Footnote 38 in [45] states that R2NB-bridged [1.1]FCPs have been prepared by Braunschweig and Koblinski; however, these results are still unpublished.
60
Page 61 of 70
Scheme 16. Preparation of the boron-bridged [1.1]FCP 68. Li(12-crown-4) Li Fe
BMe 2 +
BMe2
2 tmeda
B Fe Me2
12-crown-4
Fe
Li
2
Fe Li+
- tmeda
B Me 2 [68-Li] -
cr
ip t
83
B2
us
Li1 Fe1
B1
an
Fe2
M
Fig. 33. Molecular structure of [68-Li]- with thermal ellipsoids drawn at the 50% probability level. Hydrogen atoms and Li(12-crown-4)2]+ are omitted for clarity. The figure was prepared with ORTEP-3
d
for Windows [90] using the published crystallographic data [153].
te
As mentioned above, [1.1]FCPs were either found as syn or as anti isomers (Fig. 31). These isomers are
Ac ce p
organometallic analogues of cyclohexane [178], with the syn isomer being related to the boat isomer and the anti isomer being related to the chair isomer of cyclohexane. Isomers in a syn conformation should be better described as twist-syn isomers (similar as the twist-boat isomer of cyclohexane). An example of such an isomer is the diborata[1.1]ferrocenophane 68 (Fig. 33), where dihedral angles between related Cp rings coordinated to different iron atoms were found as 21.3 and 23.1°, respectively [153]. For comparison, Figure 34 depicts the molecular structure of the gallium compound 72, which serves as a typical example of a [1.1]FCP in an anti conformation (also see Fig. 31).
61
Page 62 of 70
Si1
Ga Si2 Fe'
ip t
Fe
cr
Si2'
Ga'
us
Si1'
Fig. 34. Molecular structure of 72 with thermal ellipsoids drawn at the 50% probability level. Hydrogen
an
atoms are omitted for clarity. The figure was prepared with ORTEP-3 for Windows [90] using the
M
published crystallographic data [147].
d
Bridging elements with small covalent radii prefer the syn conformation, whereas larger bridging
te
elements prefer the anti conformation. In both structures, a bridging atom with a small radius would result in close HH contacts between hydrogen atoms in positon on neighbouring Cp rings; this is
Ac ce p
illustrated for the syn isomer in Figure 35. However, the syn conformation is more flexible than the anti conformation and can form a twisted variant and, hence, allow for a minimization of steric repulsion between -H atoms [12, 178].
Rx E
H H Fe Rx Fe E
H H
Fig. 35. Illustration of steric interactions between -H atoms in syn isomers.
Many of the [1.1]FCPs are highly dynamic in solution and fewer signals than expected are found in 1H 62
Page 63 of 70
and 13C NMR spectra. These facts had been explained by the occurrence of degenerate isomerizations, more specifically, either syn-to-syn or anti-to-anti isomerizations. Considering the relation to cyclohexane that is known for its fast interconversions of conformers, degenerate isomerizations in [1.1]FCPs are not such a surprize. For example, the digalla[1.1]ferrocenophane 72 showed only two
ip t
NMR signals for the 16 Cp protons [147]. With the assumption of free rotations of CH(SiMe3)2 groups at gallium, four 1H NMR signals for Cp protons are expected for such an anti isomer with C2h symmetry
cr
in solution. The fact that only two signals were found shows that the group order of the time-averaged
us
symmetry for 72 must be h = 8 instead of h = 4 (C2h). This can be explained by a degenerate anti-to-anti
an
isomerization, which results in a time-averaged D2h symmetry (h = 8).
Among this group of [1.1]FCPs, only species 78 and 79 (Table 5) have two different elements (Ga and
M
Si) in alternate bridging positions [176]. Both species were prepared by a salt-metathesis approach that also gave mixtures of oligomers among which silagalla[1m]ferrocenophanes (F; Fig. 1; Chapter 1) with
d
up to 12 ferrocenediyl moieties were detected by MALDI-TOF mass spectrometry. To the best of our
te
knowledge, these are the only reported group-13-element bridged [1m]metallocenophanes that have
Ac ce p
more than two sandwich moieties.
7. Conclusions
To date, group-13-element containing [n]metallocenophanes with up to four elements in bridging position are known (n = 1–4). As expected, the majority of these species are ferrocene derivatives, however, a few ruthenocenophanes ([1]RCPs 17 and 18; Scheme 4; Chapter 2.3) and cobaltocenophanes (61; Fig. 26; Chapter 5) had been prepared as well. For metallocenophanes that contain more than one sandwich moiety (F; Fig. 1; Chapter 1) only [1.1]FCPs are reported in the literature. Among this group of species, only 78 and 79 (Table 5; Chapter 6) have two different elements (Ga and Si) in alternate bridging positions. On a personal note, we are fascinated by the possibility to prepare strained species
63
Page 64 of 70
that could become useful monomers for polymerization. As mentioned in the introduction, siliconbridged [1]FCPs have fulfilled this expectation. With respect to metallopolymers, group-13-element bridged metallocenophanes are still in their infancy and the future will reveal if creative synthetic
ip t
chemistry can help them to mature.
Acknowledgments
cr
The authors are grateful to the University of Saskatchewan, the College of Arts and Science (Catalyst
us
Award for H. B.), and the NSERC (Discovery Grant for J. M.) for their support. J. M. would like to thank all his students and collaborators who contributed to the results summarized within this review for
Ac ce p
te
d
M
an
their dedication to phane chemistry.
64
Page 65 of 70
References
Ac ce p
te
d
M
an
us
cr
ip t
[1] A. Salzer, Pure Appl. Chem., 71 (1999) 1557-1585. [2] H.H. Brintzinger, D. Fischer, R. Mülhaupt, B. Rieger, R.M. Waymouth, Angew. Chem., Int. Ed., 34 (1995) 1143-1170. [3] D.E. Herbert, U.F.J. Mayer, I. Manners, Angew. Chem., Int. Ed., 46 (2007) 5060-5081. [4] M. Tamm, Chem. Commun., (2008) 3089-3100. [5] H. Braunschweig, T. Kupfer, Acc. Chem. Res., 43 (2010) 455-465. [6] M. Herberhold, T. Hofmann, S. Weinberger, B. Wrackmeyer, Z. Naturforsch. B, 52 (1997) 10371042. [7] H. Braunschweig, T. Kupfer, K. Radacki, Angew. Chem., Int. Ed., 46 (2007) 1630-1633. [8] H. Braunschweig, M. Friedrich, T. Kupfer, K. Radacki, Chem. Commun., 47 (2011) 3998-4000. [9] H. Braunschweig, F. Breher, S. Capper, K. Duck, M. Fuss, J.O.C. Jimenez-Halla, I. Krummenacher, T. Kupfer, D. Nied, K. Radacki, Chem.-Eur. J., 19 (2013) 270-281. [10] P. Chadha, J.L. Dutton, M.J. Sgro, P.J. Ragogna, Organometallics, 26 (2007) 6063-6065. [11] H. Braunschweig, M. Fuss, T. Kupfer, K. Radacki, J. Am. Chem. Soc., 133 (2011) 5780-5783. [12] U.T. Mueller-Westerhoff, Angew. Chem., Int. Ed., 25 (1986) 702-717. [13] K.L. Rinehart, R.J. Curby, J. Am. Chem. Soc., 79 (1957) 3290-3291. [14] K.L. Rinehart, A.K. Frerichs, P.A. Kittle, L.F. Westman, D.H. Gustafson, R.L. Pruett, J.E. McMahon, J. Am. Chem. Soc., 82 (1960) 4111-4112. [15] A.G. Osborne, R.H. Whiteley, J. Organomet. Chem., 101 (1975) C27-C28. [16] C. Elschenbroich, J. Hurley, B. Metz, W. Massa, G. Baum, Organometallics, 9 (1990) 889-897. [17] M. Tamm, A. Kunst, T. Bannenberg, E. Herdtweck, P. Sirsch, C.J. Elsevier, J.M. Ernsting, Angew. Chem., Int. Ed., 43 (2004) 5530-5534. [18] C. Elschenbroich, F. Paganelli, M. Nowotny, B. Neumüller, O. Burghaus, Z. Anorg. Allg. Chem., 630 (2004) 1599-1606. [19] H. Braunschweig, M. Lutz, K. Radacki, Angew. Chem., Int. Ed., 44 (2005) 5647-5651. [20] D.A. Foucher, B.-Z. Tang, I. Manners, J. Am. Chem. Soc., 114 (1992) 6246-6248. [21] V. Bellas, M. Rehahn, Angew. Chem., Int. Ed., 46 (2007) 5082-5104. [22] R. Rulkens, A.J. Lough, I. Manners, J. Am. Chem. Soc., 116 (1994) 797-798. [23] Y.Z. Ni, R. Rulkens, I. Manners, J. Am. Chem. Soc., 118 (1996) 4102-4114. [24] H. Qiu, Z.M. Hudson, M.A. Winnik, I. Manners, Science, 347 (2015) 1329-1332. [25] Z.M. Hudson, D.J. Lunn, M.A. Winnik, I. Manners, Nat. Commun., 5 (2014) 3372. [26] Z.M. Hudson, C.E. Boott, M.E. Robinson, P.A. Rupar, M.A. Winnik, I. Manners, Nat. Chem., 6 (2014) 893-898. [27] N.S. Ieong, I. Manners, Macromol. Chem. Phys., 210 (2009) 1080-1086. [28] C.H. Honeyman, T.J. Peckham, J.A. Massey, I. Manners, Chem. Commun., (1996) 2589-2590. [29] T.J. Peckham, J.A. Massey, C.H. Honeyman, I. Manners, Macromolecules, 32 (1999) 2830-2837. [30] T. Mizuta, M. Onishi, K. Miyoshi, Organometallics, 19 (2000) 5005-5009. [31] T. Mizuta, Y. Imamura, K. Miyoshi, J. Am. Chem. Soc., 125 (2003) 2068-2069. [32] S.K. Patra, G.R. Whittell, S. Nagiah, C.-L. Ho, W.-Y. Wong, I. Manners, Chem. Eur. J., 16 (2010) 3240-3250. [33] D.E. Herbert, U.F.J. Mayer, J.B. Gilroy, M.J. Lopez-Gómez, A.J. Lough, J.P.H. Charmant, I. Manners, Chem. Eur. J, 15 (2009) 12234-12246. [34] H. Braunschweig, F.M. Breitling, E. Gullo, M. Kraft, J. Organomet. Chem., 680 (2003) 31-42. [35] S. Aldridge, C. Bresner, Coord. Chem. Rev., 244 (2003) 71-92. [36] C.J. Adams, H. Braunschweig, M. Fuss, K. Kraft, T. Kupfer, I. Manners, K. Radacki, G.R. Whittell, Chem.-Eur. J., 17 (2011) 10379-10387. [37] A.C. Tagne Kuate, C.G. Daniliuc, P.G. Jones, M. Tamm, Eur. J. Inorg. Chem., 2012 (2012) 17271733. 65
Page 66 of 70
Ac ce p
te
d
M
an
us
cr
ip t
[38] H. Braunschweig, A. Damme, S. Demeshko, K. Dück, T. Kramer, I. Krummenacher, F. Meyer, K. Radacki, S. Stellwag-Konertz, G.R. Whittell, J. Am. Chem. Soc., 137 (2015) 1492-1500. [39] R.W. Heo, T.R. Lee, J. Organomet. Chem., 578 (1999) 31-42. [40] H. Braunschweig, T. Kupfer, Eur. J. Inorg. Chem., (2012) 1319-1332. [41] F.H. Schacher, P.A. Rupar, I. Manners, Angew. Chem., Int. Ed., 51 (2012) 7898-7921. [42] R.A. Musgrave, A.D. Russell, I. Manners, Organometallics, 32 (2013) 5654−5667. [43] A.D. Russell, R.A. Musgrave, L.K. Stoll, P. Choi, H. Qiu, I. Manners, J. Organomet. Chem., 784 (2015) 24-30. [44] H. Braunschweig, R. Dirk, M. Müller, P. Nguyen, R. Resendes, D.P. Gates, I. Manners, Angew. Chem., Int. Ed., 36 (1997) 2338-2340. [45] A. Berenbaum, H. Braunschweig, R. Dirk, U. Englert, J.C. Green, F. Jäkle, A.J. Lough, I. Manners, J. Am. Chem. Soc., 122 (2000) 5765-5774. [46] S. Sadeh, G. Schatte, J. Müller, Chem. Eur. J., 19 (2013) 13408–13417. [47] S. Sadeh, H. Bhattacharjee, E. Khozeimeh Sarbisheh, J.W. Quail, J. Müller, Chem. Eur. J., 20 (2014) 16320-16330. [48] J.A. Schachner, C.L. Lund, J.W. Quail, J. Müller, Organometallics, 24 (2005) 785-787. [49] J.A. Schachner, C.L. Lund, J.W. Quail, J. Müller, Organometallics, 24 (2005) 4483-4488. [50] C. Eaborn, J.D. Smith, J. Chem. Soc., Dalton Trans., (2001) 1541-1552. [51] S.S. Al-Juaid, C. Eaborn, P.B. Hitchcock, M.S. Hill, J.D. Smith, Organometallics, 19 (2000) 32243231. [52] C. Eaborn, M.S. Hill, P.B. Hitchcock, J.D. Smith, Chem. Commun., (2000) 691-692. [53] S.S. Al-Juaid, A.G. Avent, C. Eaborn, S.M. El-Hamruni, S.A. Hawkes, M.S. Hill, M. Hopman, P.B. Hitchcock, J.D. Smith, J. Organomet. Chem., 631 (2001) 76-86. [54] S.S. Al-Juaid, A.G. Avent, C. Eaborn, M.S. Hill, P.B. Hitchcock, D.J. Patel, J.D. Smith, Organometallics, 20 (2001) 1223-1229. [55] C. Eaborn, M.S. Hill, P.B. Hitchcock, J.D. Smith, J. Chem. Soc., Dalton Trans., (2002) 2467-2472. [56] R.A. Fischer, A. Miehr, H. Sussek, H. Pritzkow, E. Herdtweck, J. Müller, O. Ambacher, T. Metzger, J. Chem. Soc., Chem. Commun., (1996) 2685-2686. [57] J. Müller, R.A. Fischer, H. Sussek, P. Pilgram, R. Wang, H. Pritzkow, E. Herdtweck, Organometallics, 17 (1998) 161-166. [58] A. Devi, W. Rogge, A. Wohlfart, F. Hipler, H.W. Becker, R.A. Fischer, Chem. Vap. Deposition, 6 (2000) 245-252. [59] J. Müller, S. Bendix, J. Chem. Soc., Chem. Commun., (2001) 911-912. [60] A. Devi, R. Schmid, J. Müller, R.A. Fischer, Materials Chemistry of Group-13 Nitrides, in: R.A. Fischer (Ed.) Topics in Organometallic Chemistry, Springer, Heidelberg, 2005, pp. 49-80. [61] A. Haaland, Angew. Chem., Int. Ed., 28 (1989) 992-1007. [62] J. Müller, U. Englert, Chem. Ber., 128 (1995) 493-497. [63] H. Schumann, B.C. Wassermann, S. Schutte, B. Heymer, S. Nickel, T.D. Seuss, S. Wernik, J. Demtschuk, F. Girgsdies, R. Weimann, Z. Anorg. Allg. Chem., 626 (2000) 2081-2095. [64] C.L. Lund, J.A. Schachner, J.W. Quail, J. Müller, Organometallics, 25 (2006) 5817-5823. [65] M. Yoshifuji, K. Kamijo, K. Toyota, Tetrahedron Lett., 35 (1994) 3971-3974. [66] B. Bagh, N.C. Breit, K. Harms, G. Schatte, I.J. Burgess, H. Braunschweig, J. Müller, Inorg. Chem., 51 (2012) 11155−11167. [67] J.A. Schachner, S. Tockner, C.L. Lund, J.W. Quail, M. Rehahn, J. Müller, Organometallics, 26 (2007) 4658-4662. [68] J.A. Schachner, J.W. Quail, J. Müller, Acta Crystallogr., E64 (2008) m517. [69] N.C. Breit, T. Ancelet, J.W. Quail, G. Schatte, J. Müller, Organometallics, 30 (2011) 6150-6158. [70] B. Bagh, G. Schatte, J.C. Green, J. Müller, J. Am. Chem. Soc., 134 (2012) 7924−7936. [71] B. Bagh, J.B. Gilroy, A. Staubitz, J. Müller, J. Am. Chem. Soc., 132 (2010) 1794-1795. [72] Mercury, in, http://www.ccdc.cam.ac.uk/mercury. [73] C.L. Lund, J.A. Schachner, J.W. Quail, J. Müller, J. Am. Chem. Soc., 129 (2007) 9313-9320. 66
Page 67 of 70
Ac ce p
te
d
M
an
us
cr
ip t
[74] C. Elschenbroich, A. Bretschneider-Hurley, J. Hurley, W. Massa, S. Wocadlo, J. Pebler, E. Reijerse, Inorg. Chem., 32 (1993) 5421-5424. [75] C. Elschenbroich, A. Bretschneider-Hurley, J. Hurley, A. Behrendt, W. Massa, S. Wocadlo, E. Reijerse, Inorg. Chem., 34 (1995) 743-745. [76] K.C. Hultzsch, J.M. Nelson, A.J. Lough, I. Manners, Organometallics, 14 (1995) 5496-5502. [77] C. Elschenbroich, E. Schmidt, B. Metz, K. Harms, Organometallics, 14 (1995) 4043-4045. [78] C. Elschenbroich, E. Schmidt, R. Gondrum, B. Metz, O. Burghaus, W. Massa, S. Wocadlo, Organometallics, 16 (1997) 4589-4596. [79] D. Seyferth, Organometallics, 21 (2002) 1520-1530. [80] H. Braunschweig, M. Homberger, C.H. Hu, X.L. Zheng, E. Gullo, G. Clentsmith, M. Lutz, Organometallics, 23 (2004) 1968-1970. [81] C. Lund, J. Müller, unpublished results [82] S.S. Al-Juaid, C. Eaborn, S.M. El-Hamruni, P.B. Hitchcock, J.D. Smith, Organometallics, 18 (1999) 45-52. [83] U. Vogel, A.J. Lough, I. Manners, Angew. Chem., Int. Ed., 43 (2004) 3321-3325. [84] R. Rulkens, A.J. Lough, I. Manners, Angew. Chem., Int. Ed., 35 (1996) 1805-1807. [85] F. Jäkle, R. Rulkens, G. Zech, D.A. Foucher, A.J. Lough, I. Manners, Chem. Eur. J., 4 (1998) 2117-2128. [86] H.K. Sharma, F. Cervantes-Lee, J.S. Mahmoud, K.H. Pannell, Organometallics, 18 (1999) 399403. [87] F. Jäkle, R. Rulkens, G. Zech, J. Massey, I. Manners, J. Am. Chem. Soc., 122 (2000) 4231-4232. [88] T. Baumgartner, F. Jäkle, R. Rulkens, G. Zech, A.J. Lough, I. Manners, J. Am. Chem. Soc., 124 (2002) 10062-10070. [89] A. Berenbaum, F. Jäkle, A.J. Lough, I. Manners, Organometallics, 21 (2002) 2359 -2361. [90] L.J. Farrugia, J. Appl. Crystallogr., 30 (1997) 565. [91] H. Braunschweig, C. Burschka, G.K.B. Clentsmith, T. Kupfer, K. Radacki, Inorg. Chem., 44 (2005) 4906-4908. [92] J.A. Schachner, G.A. Orlowski, J.W. Quail, H.-B. Kraatz, J. Müller, Inorg. Chem., 45 (2006) 454459. [93] J.A. Schachner, C.L. Lund, I.J. Burgess, J.W. Quail, G. Schatte, J. Müller, Organometallics, 27 (2008) 4703-4710. [94] Holleman-Wiberg, Inorganic Chemistry, 1. English ed., Academic Press, San Diego, London, 2001, p 1756. [95] C.L. Lund, J.A. Schachner, I.J. Burgess, J.W. Quail, G. Schatte, J. Müller, Inorg. Chem., 47 (2008) 5992–6000. [96] T. Matsumoto, K. Tanaka, Y. Chujo, J. Am. Chem. Soc., 135 (2013) 4211-4214. [97] T. Matsumoto, Y. Onishi, K. Tanaka, H. Fueno, Y. Chujo, Chem. Commun., 50 (2014) 1574015743. [98] L. Schwink, P. Knochel, Chem. Eur. J., 4 (1998) 950-968. [99] D. Marquarding, H. Klusacek, G. Gokel, P. Hoffmann, I. Ugi, J. Am. Chem. Soc., 92 (1970) 53895393. [100] S. Itsuno, K. Ito, A. Hirao, S. Nakahama, J. Chem. Soc., Chem. Commun., (1983) 469-470. [101] E.J. Corey, R.K. Bakshi, S. Shibata, J. Am. Chem. Soc., 109 (1987) 5551-5553. [102] J. Kang, J.H. Lee, J.S. Choi, Tetrahedron: Asymmetry, 12 (2001) 33-35. [103] W.F. Bailey, M.R. Luderer, K.P. Jordan, J. Org. Chem., 71 (2006) 2825-2828. [104] B. Bagh, S. Sadeh, J.C. Green, J. Müller, Chem. Eur. J., 20 (2014) 2318-2327. [105] E. Khozeimeh Sarbisheh, J.C. Green, J. Müller, Organometallics, 33 (2014) 3508−3513. [106] J.K. Pudelski, D.P. Gates, R. Rulkens, A.J. Lough, I. Manners, Angew. Chem., Int. Ed., 34 (1995) 1506-1508. [107] R. Rulkens, D.P. Gates, D. Balaishis, J.K. Pudelski, D.F. McIntosh, A.J. Lough, I. Manners, J. Am. Chem. Soc., 119 (1997) 10976-10986. 67
Page 68 of 70
Ac ce p
te
d
M
an
us
cr
ip t
[108] N.P. Reddy, H. Yamashita, M. Tanaka, J. Chem. Soc., Chem. Commun., (1995) 2263-2264. [109] Y.Z. Ni, R. Rulkens, J.K. Pudelski, I. Manners, Macromol. Rapid Commun., 16 (1995) 637-641. [110] J.B. Sheridan, K. Temple, A.J. Lough, I. Manners, J. Chem. Soc., Dalton Trans., (1997) 711-713. [111] J. Weaver, P. Woodward, J. Chem. Soc., Dalton Trans., (1973) 1439-1443. [112] J.C. Green, Chem. Soc. Rev., 27 (1998) 263-271. [113] W. Finckh, B.-Z. Tang, D.A. Foucher, D.B. Zamble, R. Ziembinski, A. Lough, I. Manners, Organometallics, 12 (1993) 823-829. [114] F. Jäkle, Chem. Rev., 110 (2010) 3985-4022. [115] F. Cheng, F. Jäkle, Polym. Chem., 2 (2011) 2122-2132. [116] M. Herberhold, U. Dörfler, B. Wrackmeyer, J. Organomet. Chem., 530 (1997) 117-120. [117] H. Braunschweig, F. Seeler, R. Sigritz, J. Organomet. Chem., 692 (2007) 2354-2356. [118] H. Braunschweig, A. Damme, T. Kupfer, Eur. J. Inorg. Chem., 2010 (2010) 4423-4426. [119] M. Herberhold, U. Dörfler, W. Milius, B. Wrackmeyer, J. Organomet. Chem., 492 (1995) 59-63. [120] H. Braunschweig, T. Kupfer, M. Lutz, K. Radacki, F. Seeler, R. Sigritz, Angew. Chem., Int. Ed., 45 (2006) 8048-8051. [121] F. Bauer, H. Braunschweig, K. Schwab, Organometallics, 29 (2010) 934-938. [122] A. Appel, H. Nöth, Z. Anorg. Allg. Chem., 636 (2010) 2329-2342. [123] H. Bera, H. Braunschweig, A. Oechsner, F. Seeler, R. Sigritz, J. Organomet. Chem., 695 (2010) 2609-2613. [124] H. Braunschweig, T. Kupfer, J. Am. Chem. Soc., 130 (2008) 4242-4243. [125] F. Bauer, H. Braunschweig, K. Gruss, T. Kupfer, Organometallics, 30 (2011) 2869-2884. [126] F. Bauer, H. Braunschweig, K. Gruß, C. Lambert, K.K. Pandey, K. Radacki, D. Reitzenstein, Chem. Eur. J., 17 (2011) 5230-5233. [127] L. Ding, K. Ma, F. Fabrizi de Biani, M. Bolte, P. Zanello, M. Wagner, Organometallics, 20 (2001) 1041-1043. [128] F. Jäkle, M. Mattner, T. Priermeier, M. Wagner, J. Organomet. Chem., 502 (1995) 123-130. [129] F. Jäkle, T. Priermeier, M. Wagner, J. Chem. Soc., Chem. Commun., (1995) 1765-1766. [130] F. Jäkle, T. Priermeier, M. Wagner, Organometallics, 15 (1996) 2033-2040. [131] E. Herdtweck, F. Jäkle, G. Opromolla, M. Spiegler, M. Wagner, P. Zanello, Organometallics, 15 (1996) 5524-5535. [132] E. Herdtweck, F. Jäkle, M. Wagner, Organometallics, 16 (1997) 4737-4745. [133] D. Fest, C.D. Habben, J. Organomet. Chem., 390 (1990) 339-344. [134] M. Herberhold, H.D. Brendel, J. Organomet. Chem., 458 (1993) 205-209. [135] B. Wrackmeyer, E.V. Klimkina, H.E. Maisel, W. Milius, M. Herberhold, Inorg. Chim. Acta, 357 (2004) 1703-1710. [136] B. Wrackmeyer, E.V. Klimkina, W. Millus, O.L. Tok, M. Herberhold, Inorg. Chim. Acta, 358 (2005) 1420-1428. [137] T.C. Auch, H. Braunschweig, K. Radacki, R. Sigritz, U. Siemeling, S. Stellwag, Z. Naturforsch. B, 63 (2008) 920-922. [138] M.J.G. Lesley, U. Mock, N.C. Norman, A.G. Orpen, C.R. Rice, J. Starbuck, J. Organomet. Chem., 582 (1999) 116-118. [139] G.E. Herberich, A. Fischer, D. Wiebelhaus, Organometallics, 15 (1996) 3106-3108. [140] G.E. Herberich, U. Englert, A. Fischer, D. Wiebelhaus, Organometallics, 17 (1998) 4769-4775. [141] B. Wrackmeyer, E.V. Klimkina, W. Milius, Inorg. Chem. Commun., 9 (2006) 716-721. [142] B. Wrackmeyer, E.V. Klimkina, W. Milius, Polyhedron, 26 (2007) 3496-3504. [143] B. Wrackmeyer, E.V. Klimkina, W. Milius, Z. Anorg. Allg. Chem., 633 (2007) 1964-1972. [144] B. Wrackmeyer, E.V. Klimkina, W. Milius, Z. Naturforsch. B, 62 (2007) 1259-1266. [145] B. Wrackmeyer, E.V. Klimkina, W. Milius, Z. Anorg. Allg. Chem., 635 (2009) 2532-2540. [146] B. Wrackmeyer, E.V. Klimkina, W. Milius, Eur. J. Inorg. Chem., (2008) 306-311. [147] W. Uhl, I. Hahn, A. Jantschak, T. Spies, J. Organomet. Chem., 637 (2001) 300-303. [148] A.N. Nesmeyanov, I.I. Kritskaya, Bull. Acad. Sci. USSR, Div. Chem. Sci. (Eng. Transl.), (1956) 68
Page 69 of 70
Ac ce p
te
d
M
an
us
cr
ip t
243-244. [149] J.S. McKechnie, B. Bersted, I.C. Paul, W.E. Watts, J. Organomet. Chem., 8 (1967) P29-P31. [150] J.S. McKechnie, C.A. Maier, B. Bersted, I.C. Paul, J. Chem. Soc., Perkin Trans. 2, (1973) 138143. [151] M. Löwendahl, Ö. Davidsson, P. Ahlberg, M. Håkansson, Organometallics, 12 (1993) 2417-2419. [152] D. Seyferth, H.P. Withers, Organometallics, 1 (1982) 1275-1282. [153] M. Scheibitz, R.F. Winter, M. Bolte, H.-W. Lerner, M. Wagner, Angew. Chem., Int. Ed., 42 (2003) 924-927. [154] J.A. Schachner, C.L. Lund, J.W. Quail, J. Müller, Acta Crystallogr., E61 (2005) m682-m684. [155] P. Jutzi, N. Lenze, B. Neumann, H.G. Stammler, Angew. Chem., Int. Ed., 40 (2001) 1423-1427. [156] A. Althoff, P. Jutzi, N. Lenze, B. Neumann, A. Stammler, H.-G. Stammler, Organometallics, 21 (2002) 3018-3022. [157] A. Althoff, P. Jutzi, N. Lenze, B. Neumann, A. Stammler, H.G. Stammler, Organometallics, 22 (2003) 2766-2774. [158] M. Herberhold, T. Bärtl, Z. Naturforsch. B, 50 (1995) 1692-1698. [159] J.W. Park, Y.S. Seo, S.S. Cho, D.M. Whang, K.M. Kim, T.Y. Chang, J. Organomet. Chem., 489 (1995) 23-25. [160] D.L. Zechel, D.A. Foucher, J.K. Pudelski, G.P.A. Yap, A.L. Rheingold, I. Manners, J. Chem. Soc., Dalton Trans., (1995) 1893-1899. [161] N.P. Reddy, N. Choi, S. Shimada, M. Tanaka, Chem. Lett., (1996) 649-650. [162] M.J. MacLachlan, J. Zheng, K. Thieme, A.J. Lough, I. Manners, C. Mordas, R. LeSuer, W.E. Geiger, L.M. Liable-Sands, A.L. Rheingold, Polyhedron, 19 (2000) 275-289. [163] G. Calleja, F. Carré, G. Cerveau, P. Labbé, L. Coche-Guérente, Organometallics, 20 (2001) 42114215. [164] A. Berenbaum, A.J. Lough, I. Manners, Organometallics, 21 (2002) 4415-4424. [165] M. Bao, Y. Hatanaka, S. Shimada, Chem. Lett., 33 (2004) 520-521. [166] A. Clearfield, C.J. Simmons, H.P. Withers Jr., D. Seyferth, Inorg. Chim. Acta, 75 (1983) 139-144. [167] T.Y. Dong, M.Y. Hwang, Y.S. Wen, W.S. Hwang, J. Organomet. Chem., 391 (1990) 377-385. [168] G. Utri, K.E. Schwarzhans, G.M. Allmaier, Z. Naturforsch. B, 45 (1990) 755-762. [169] H. Brunner, J. Klankermayer, M. Zabel, J. Organomet. Chem., 601 (2000) 211-219. [170] T. Mizuta, Y. Imamura, K. Miyoshi, Organometallics, 24 (2005) 990-996. [171] C. Spang, F.T. Edelmann, M. Noltemeyer, H.W. Roesky, Chem. Ber., 122 (1989) 1247-1254. [172] N.S. Ieong, W.Y. Chan, A.J. Lough, M.R. Haddow, I. Manners, Chem.-Eur. J., 14 (2008) 12531263. [173] A.S. Perucha, J. Heilmann-Brohl, M. Bolte, H.W. Lerner, M. Wagner, Organometallics, 27 (2008) 6170-6177. [174] D.A. Lemenovskii, I.F. Urazowski, T.V. Baukova, I.L. Arkhipov, R.A. Stukan, E.G. Perevalova, J. Organomet. Chem., 264 (1984) 283-288. [175] I.G. Kuz'mina, Y.T. Struchkov, D.A. Lemenovsky, I.F. Urazowsky, J. Organomet. Chem., 277 (1984) 147-151. [176] B. Bagh, N.C. Breit, J.B. Gilroy, G. Schatte, J. Müller, Chem. Commun., 48 (2012) 7823-7825. [177] J.B. Heilmann, M. Scheibitz, Y. Qin, A. Sundararaman, F. Jäkle, T. Kretz, M. Bolte, H.W. Lerner, M.C. Holthausen, M. Wagner, Angew. Chem., Int. Ed., 45 (2006) 920-925. [178] J.-M. Löwendahl, M. Håkansson, Organometallics, 14 (1995) 4736-4741.
69
Page 70 of 70