Synthesis of new pyrrole–pyridine-based ligands using an in situ Suzuki coupling method

The compounds 6-(pyrrol-2-yl)-2,2‘-bipyridine, 2-(pyrrol-2-yl)-1,10-phenanthroline and 2-(2-(N-methylbenz[d,e]imidazole)-6-(pyrrol-2-yl)-pyridine were synthesized by using an in situ generated boronic acid for the Suzuki coupling. Crystals of the products could be grown and exhibited interesting structures by X-ray analysis, one of them showing a chain-like network with the adjacent molecules linked to each other via intermolecular N–H…N hydrogen bonds.


Introduction
Tridentate ligands have recently received attention in the area of rare-earth-metal complex chemistry [1][2][3][4][5]. Many rare-earthmetal cations, such as europium(III), have the tendency to be nine-coordinating species. Besides their triple positive charge, this coordination sphere must be saturated to achieve high quantum yields when the target is to optimize its luminescent properties [6]. In the "classic" complexes, i.e., the β-diketonates [6,7], this is often achieved by combining three "saturating ligands" that consist of negatively charged β-diketonates with a "neutral ligand", which is often bidentate or tridentate (Scheme 1, left). The coordination sphere and the positive charge of the rare-earth metal cation are thus saturated and neu-tralized, respectively. In a recent development, the Bünzli group has succeeded in synthesizing homoleptic complexes, in which only one type of ligand binds to the central europium(III) via two neutral and one negatively charged atom (Scheme 1, middle) [2][3][4]. As three ligands bind to the rare-earth cation, the resulting complex is neutral, and its coordination number is nine and thereby saturated.
Our interest in this type of complex is driven by its potential analytical application for luminescence degradation measurements on photocatalytic surfaces [8]. Currently, β-diketonate complexes are in use for this analysis, but homoleptic Scheme 1: β-diketonate complexes (left), homoleptic complexes (middle) and planned homoleptic complexes of europium(III) (right).

Scheme 2:
Pyrrole-pyridine-based structures synthesized in this study. complexes may be advantageous in our opinion, because they only consist of one type of ligand, which is then subjected to photocatalytically produced radicals. Therefore, we tried to broaden the scope by synthesizing a new class of ligands for homoleptic complexes, which should bind to europium(III) via a pyrrolate anion (Scheme 1, right). The decision to choose a negatively charged heterocycle as a binding unit was based on the idea to enlarge the π-system of the ligand, thereby making it possible to absorb longer wavelengths of light (λ > 350 nm). The target structures are shown in Scheme 2.

Results and Discussion
Our first retrosynthetic approach included a Suzuki coupling of the alpha-substituted boronic acid of Boc-protected pyrrole 7 with the heteroaryl bromides 8-10, as shown in Scheme 3.
Compound 7 is described in literature [10], but it could not be purified by column chromatography and therefore was not isolated as a pure product. In addition, reports on the stability of this boronic acid show that it is not suitable for long-term storage [10]. We therefore applied a modification of the Suzuki coupling that was also used to prepare [2.2]paracyclophane-Scheme 4: Synthesis of the heteroaryl bromides used in the coupling reaction. derivatives [11]. This comprised the in situ reaction of the freshly prepared boronic acid/ester with the heteroaryl bromides 8-10. These starting compounds could be prepared by using literature procedures, as shown in Scheme 4. Substance 8 was synthesized by following the standard literature procedure [12]. Compound 9 was synthesized by principally the same reaction path [12,13]. 2-Bromopyridine-6-benzimidazole 10 was published in a patent [14], and its preparation was adapted from [4]. The following in situ Suzuki coupling uses Boc-protected pyrrole [15], which was subjected to freshly prepared LDA at −78 °C and quenched with trimethylborate. This gave rise to the intermediate 21/22 (Scheme 5).
Afterwards the heteroaryl bromides 8-10 were added along with the catalyst and base in aqueous media and the reaction mixture was heated under reflux for several hours. The reaction times were not minimized. In principal the ester can react with the boronic acid, starting with the addition of the aqueous base (K 2 CO 3 ). Yet, structure 21 is already activated for the coupling step, since a negative charge is located on the boron atom. In Scheme 6 the coupling reactions are shown; detailed information about the reaction conditions is given in Table 1.
Reaction 1 of aryl bromide 8 to product 4 was repeated under the same conditions and with the same amounts of starting material and catalyst. The workup procedure was changed, which affected the yield only to a very small extent. In the original publication [11] the aryl bromide coupling partner was used in excess. As 19 is by far easier to prepare in large quantities, we altered the protocol and used it in excess, which produced even a slightly better yield than the other way round (Table 1). Finally, deprotection of 4 with hydrochloric acid after the coupling gave 1 in 93% yield. Compound 2 was directly isolated as Boc-deprotected product, but in a much lower yield. We believe the lower yield is caused by the workup procedure, because TLC indicated nearly full consumption of 19. During column purification with aluminium oxide or silica (both were tried) we noticed smearing of the blue fluorescent product under UV irradiation. Even deactivated aluminium oxide provides  enough acidity to split the Boc-moiety, which probably causes the generated amine to partly remain on the column. In the case of 6, using an excess of 19 lead to better yields, but it has to be mentioned that the workup procedure was also adjusted, thus it cannot be directly compared. Further increasing the equivalents of 19 and the base did not significantly enhance the resulting yield ( Table 1). The total reaction time was not systematically minimized. After 20 h under reflux 6 was isolated in 87% yield; a reaction time of 70 h delivered 85% of 6. Compound 3 was synthesized in 70% yield by deprotection of 6. In this case  another method using sodium methoxide was applied, because reaction with hydrochloric acid did not take place and the unprotected starting material was recovered. We noticed no definite influence on the product yield when using different scales (2 mmol to 12 mmol).

X-Ray analysis
The molecular structure of compound 1 is shown in Figure 1.  for 2,2'-pyrrolylpyridines revealed 20 hits, all with cis geometry, discounting rigid fused-ring systems and one sterically hindered di-tert-butyl system; the corresponding absolute torsion angles ranged from 0 to 25°, mean value 7.5°. The molecular packing of 1 ( Figure 2) is determined by a classical hydrogen bond N17-H17 … N1 involving the peripheral rings, which connects the molecules via the a-glide plane to form chains parallel to the a-axis.
The structure of the methanol solvate of compound 2 is shown in Figure 3. The 1,10-phenanthroline ring system is planar (mean deviation 0.01 Å), and the pyrrole ring subtends an interplanar angle of 9° to it. The methanol molecule fits neatly into the "bay" region of the parent molecule, forming classical hydrogen bonds N19-H19 … O99 and O99-H99 … N1. The molecules associate into pairs by ring stacking across inversion centres (Figure 4). The interplanar distance is ca. 3.35 Å.
The structure of the ethanol solvate of compound 3 is shown in Figure 5. The central ring is almost coplanar with the pyrrole ring (interplanar angle 9°), but subtends an angle of 38° with the benzimidazole system. Ethanol forms one classical hydrogen bond within the asymmetric unit, acting as a donor, but it also acts as a hydrogen bond acceptor via the a-glide plane. The overall effect is to form helical chains of alternating residues of 3 and ethanol, parallel to the a-axis ( Figure 6).
The crystallographic data of compounds 1 to 3 are summarized in Table 2.

Conclusion
The new target structures 1-3 were successfully synthesized in good to acceptable yields by applying an in situ variation of the Suzuki coupling as the main reaction step. The crystal structures of these compounds could be obtained by X-ray analysis. They exhibit interesting, but very different structural features: 1 forms chains that consist of molecules directly interconnected by N-H … N hydrogen bonds, whereas 2 retains the solvent molecule methanol in the "bay" region of the molecule to form stacked dimers. Structure 3 forms a chain-like structure, but retains the solvent molecule ethanol, which connects the mole- cules via hydrogen bonds. The idea was to synthesize europium(III) complexes containing the new ligands. Since the pyrrole amine is a very weak acid (pK a = 23.0) [17] it can only be deprotonated by hard bases, such as n-butyllithium or sodium hydride [18]. Therefore we chose to work in water-free conditions with the europium precursors EuCl 3 (no crystallization water) or Eu[N(SiMe 3 ) 2 ] 3 , which are commercially available. To date we did not succeed in synthesizing the target complexes depicted on the right-hand side of Scheme 1. This   class of compounds may also be of interest for other areas of chemistry. The pyrrole-pyridine structural motif is featured in current studies, owing to its complexation properties towards first-row transition metals (Fe, Co, Ni, Cu, Zn) [18] and ruthenium [19]. The intramolecular proton transfer of these species is also of interest for vibrational spectroscopy measurements [20]. Synthesis and characterization of the heteroaryl bromides 8 and 9 starting from 2,2'-bipyridine (11) and 1,10-phenanthroline (14), as well as the synthesis of compound 10 and 19 are shown in Supporting Information File 1. Absorption, excitation and emission spectra of compounds 1-3 are included as well.