Highly flexible synthesis of indenylethylamines as ligand precursors for titanium complexes

Various indenylethylamines are synthesized for the first time by reductive amination of 2-(1 H - inden-1-yl)acetaldehyde with commercially available primary amines. In addition, a new two-step synthesis of 2-(1 H -inden-1-yl)acetaldehyde that uses inexpensive indene and 2-bromo-1,1-diethoxyethane as starting materials is presented. Finally, a selected indenylethylamine is used as a ligand precursor for the synthesis of a corresponding indenylethylamido titanium complex. The latter result paves the way for applications of corresponding complexes as catalysts for important chemical reactions.


Introduction
Transition metal complexes which are formed by coordination of ligands to metals are widely used as catalysts for a plethora of synthetically useful chemical transformations.However, in most cases, extensive optimization studies are necessary to finally develop an efficient and reliable new catalytic reaction.A common and powerful possibility to optimize transition metal catalysts is the variation of the ligands which are bound to the metal center and as a consequence, an ever-present demand for new ligand motifs exists in transition metal chemistry.
The titanium-catalyzed hydroaminoalkylation of alkenes (Scheme 1) 1−11 is a relatively new and promising method for a waste-free synthesis of industrially important amines.The reaction takes place by addition of the α-C−H bond of an amine across a C−C double bond and gives access to α-alkylated amines in a single step and with 100 % atom efficiency.Among the various titanium-catalysts, bis(η 5 -indenyl)dimethyltitanium (Ind 2 TiMe 2 ) 12 was found to be an excellent catalyst for the highly regioselective hydroaminoalkylation of 1-alkenes with N-methylanilines (Scheme 1). 5 In addition, the catalyst can be used for corresponding reactions of styrenes 5 and 1,3-butadienes 9 even though reduced regioselectivities are observed in these cases.On the other hand, dialkylamines and N-alkylanilines bearing alkyl groups larger than a methyl group do not undergo successful hydroaminoalkylation reactions in the presence of Ind 2 TiMe 2 .Ind 2 TiMe 2 -catalyzed hydroaminoalkylation reaction of 1-octene with N-methylaniline. 5e reaction mechanism of the titanium-catalyzed hydroaminoalkylation of alkenes has already been studied in detail 6 and it is generally accepted that titanaaziridines are the catalytically active species.These highly strained intermediates are formed under the reaction conditions from a catalyst precursor and the amine substrate.However, it is well established that in the presence of amines, bis(η 5 -cyclopentadienyl)titanium precursors (Cp 2 TiL n ) undergo an unexpected cyclopentadienide/amide ligand exchange reaction that results in the formation of mono(η 5 -cyclopentadienyl)titanium(amido) complexes (Cp(RNH)TiL n , Scheme 2). 13In analogy, it can be assumed that the catalyst precursor Ind 2 TiMe 2 used for hydroaminoalkylation reactions of alkenes undergoes a comparable ligand exchange reaction with the amine substrate and as a result, the expected catalytically active titanaaziridine formed from Ind 2 TiMe 2 possesses only one indenyl ligand and an amido ligand (Scheme 2).However, these two ligands are expected to remain bonded to the titanium center during the entire catalytic cycle and as a consequence, they both determine the activity of the catalytic species.
A major drawback of the expected ligand exchange of one indenyl ligand against an amido ligand is the fact that the reactivity of the catalytically active titanaaziridine is directly influenced by the nature of the amine substrate.Consequently, the undesirable possibility exists that a certain class of amine substrates, e.g.aromatic amines, deliver catalytic species of high activity while the use of other classes of amines, e.g.alkylamines, results in poorly active catalysts.One approach to avoid this unwanted dependency of the activity of the catalytic species on the nature of the amine substrate is to use a catalyst precursor that contains a chelating ligand with a bridged amido indenyl ligand.Corresponding silyl-linked amido cyclopentadienyl or amido indenyl ligands have extensively been used for the synthesis of various so-called constrained geometry catalysts in recent years, 14 but the notoriously labile Si-N-linkage encouraged us to focus on more stable catalyst precursors with chelating indenylethylamido ligands which are shown in Scheme 3 (general structure 1).In this context, it must also be noted that the potential new class of hydroaminoalkylation catalyst precursors of type 1 offers the additional possibility to fine-tune the catalyst by varying the substituent R bound to the nitrogen atom of the chelating ligand.From a retrosynthetic point of view, dimethyl titanium complexes of structure 1 should be easily accessible from the corresponding dichloro titanium complexes 2 and a nucleophilic methylation reagent, e.g.methyl lithium.Unfortunately, complexes of type 2 have not been described in the literature.Herein we describe the first preparative example of a general method for the synthesis of indenylethylamido dichloro titanium complexes of type 2 that relies on the reaction of titanium tetrachloride with dilithiated indenylethylamines 3 or 4 (Scheme 3).In addition, we present a new and highly flexible method for the synthesis of indenylethylamines 3/4 which are needed as ligand precursors for the planned synthesis of additional new titanium complexes.As the keystep for the formation of the indenylethylamines 3/4, a reductive amination of 2-(1H-inden-1yl)acetaldehyde (5) with widely available primary amines 6 is used.This highly flexible reaction offers a simple way to alter the nature of the substituent R bound to the nitrogen atom of the indenylethylamines 3/4.

Results and Discussion
To the best of our knowledge, only one procedure for the synthesis of the desired aldehyde 5 can be found in the literature.The corresponding report by Ipaktschi describes a photochemical ring opening of benzonorbornanone (11) using a 450 W mercury high pressure lamp that directly converts 11 into 5 (Scheme 4). 15Our synthesis of the required starting material 11 began with a Diels-Alder reaction between cyclopentadiene (8) and benzyne which can easily be generated in situ from 1,2-dibromobenzene (7) and n-butyllithium. 16After the resulting benzonorbornadiene (9) had been obtained in moderate yield of 58 % an oxymercuration and a subsequent Swern oxidation gave access to ketone 11 in 68 % combined yield over these two steps.For the final photochemical ring opening, ketone 11 was then irradiated with UV-light in diethyl ether using a 450 W mercury high pressure lamp.As monitored by 1 H NMR spectroscopy, a reaction time of 24 h was necessary to reach full conversion.Finally, distillation of the crude product delivered the pure aldehyde 5 in 75 % yield.Although the described four-step protocol can be used successfully for the multi-gram synthesis of aldehyde 5, the low overall yield of only 30 % and the time consuming experimental procedures as well as the expensive starting materials must be regarded as severe drawbacks to this synthetic approach.In search of a more convenient synthesis of aldehyde 5, we then found that lithiated indene (12) smoothly undergoes a nucleophilic substitution reaction with commercially available and inexpensive 2-bromo-1,1-diethoxyethane (13) to give the corresponding indenylethylacetal 14 in excellent yield of 98 % (Scheme 5).After aqueous work-up, acetal 14 was already obtained with high purity (97 % as determined by GC analysis) and therefore, the crude product could directly be used for the subsequent acetal cleavage.For that purpose, acetal 14 was stirred in acetone in the presence of a catalytic amount of iodine (10 mol %) at 40 °C for 2 h. 17Although aldehyde 5 could only be isolated in moderate yield of 58 % from the crude reaction mixture, the new and shorter two-step synthesis of 5 gave a significantly improved overall yield of 57 % and it is less time-consuming than the four-step alternative.With key-intermediate 5 in hand we then performed a number of reductive aminations 18 with various primary aryl-and alkylamines (Table 1).All reactions were carried out under ambient conditions and the progress of the reactions was monitored by TLC analysis.As expected, treatment of a solution of aldehyde 5 in methanol with the corresponding amines 6a-h and subsequent reduction with a solution of NaBH 3 CN and ZnCl 2 in methanol gave the amination products 3a-c and 4d-h in moderate to excellent yields (51-91 %).Interestingly, the use of arylamines always resulted in the selective formation of the corresponding 1-isomer (3a-c) of the amination products (Table 1, entries 1-3) whereas most of the more basic pyridinyl-and alkylamines were converted to the thermodynamically more stable 3-isomers (4d-g, Table 1, entries 4-7).Only the reaction of the chiral substrate (R)-1-phenylethylamine (6h) with aldehyde 5 gave a mixture of the corresponding 1-and 3-isomer 3h and 4h in a ratio of 16:84 (Table 1, entry 8).Although the products 3a-c and 4d-h were fully characterized by IR, 1 H and 13 C NMR spectroscopy and mass spectrometry it can be noted that a single-crystal X-ray structure analysis of the hydrochloride of product 4h (4h•HCl) could additionally be obtained (for details, see the supplementary material).To find out whether the reductive amination can also be used for the synthesis of multi-gram quantities of the desired products we additionally performed two selected reactions on a 30 mmol scale.For that purpose, we chose aniline (6a) and isopropylamine (6e) as typical examples of aryl-and alkylamines.Gratifyingly, both reactions proceeded smoothly and the products 3a and 4e were obtained in yields of 89 % and 59 %, respectively.These results which are even slightly better than those presented in Table 1 (entries 1 and 5) clearly prove that the new process can deliver sufficient quantities of indenylethylamines for the planned investigation of this class of compounds as ligand precursors in transition metal chemistry.The reaction was also carried out on a 30 mmol scale with a slightly better yield.c Only major isomer presented.A mixture of the 1-and the 3-isomer in a ratio of 3h/4h = 16:84 (GC analysis) was obtained.
To finally prove the general suitability of the synthesized indenylethylamines as ligand precursors for the preparation of transition metal complexes we converted compound 3a with two equivalents of n-butyllithium into the corresponding dilithium salt (15) and treated it with titanium tetrachloride in diethyl ether (Scheme 6).After separation from the formed lithium chloride, red-brown crystals of the expected new titanium complex 16 could be isolated in 33 % yield.Subsequently, the obtained solid material could be recrystallized from benzene-d 6 to give red crystals that were suitable for X-ray single-crystal analysis.The solid-state structure of 16 (Figure 1) 19 reveals a slightly distorted tetrahedral geometry around the titanium center and proves the bidentate binding mode of the new indenylethylamido ligand.[22][23]

Conclusions
In summary, it was shown that the reductive amination of 2-(1H-inden-1-yl)acetaldehyde ( 5) with commercially available primary amines represents a highly flexible and efficient method for the synthesis of various indenylethylamines and in addition, a new two-step synthesis of aldehyde 5 that uses inexpensive indene and 2-bromo-1,1-diethoxyethane as starting materials was developed.Particularly noteworthy is that a selected indenylethylamine could already be used as a ligand precursor for the synthesis of a corresponding indenylethylamido titanium complex.This result clearly proves the suitability of indenylethylamines as starting materials for the synthesis of transition metal complexes and it paves the way for the planned investigation of corresponding complexes as catalysts for hydroaminoalkylation reactions of alkenes.

Scheme 2 .Scheme 3 .
Scheme 2. Ligand exchange reactions of Cp 2 TiMe 2 and Ind 2 TiMe 2 in the presence of amines.

Table 1 .
Reductive amination of aldehyde 5 with selected primary amines was purchased from Acros Organics and used as received.DMSO was distilled from CaH 2 and stored under an argon atmosphere over molecular sieves (4 Å) prior to use.Diethyl ether, toluene and benzene-d 6 were dried by distillation from sodium.Light petroleum ether (PE) and tert-butyl methyl ether (MTBE) for flash chromatography were purified by distillation.Flash chromatography was carried out with a Büchi Sepacore ® Flash System X10 using Büchi Sepacore ® Flash Cartridges (40 g silica gel) or Büchi PLASTIGLAS ® columns with silica gel from Fluka (particle size 0.037-0.063mm).For thin layer chromatography, ALUGRAM ® TLC aluminium sheets with fluorescent indicator (254 nm) from Macherey-Nagel were used.All substances were detected with UV light.As light source for the photochemical reaction a water-cooled Hanovia L 469 A 450 W mercury high pressure lamp was used.Melting points were determined with a Schorpp-Gerätetechnik melting point MPM-H2 apparatus and are uncorrected.IR spectra were recorded with a Bruker Tensor 27 spectrometer using an attenuated total reflection (ATR) method.Absorption values are given as wavenumbers (cm −1 ).The intensities of the absorptions are given as weak(w), medium (m), strong (s), very strong (vs) and broad (br).NMR spectra were recorded with a Bruker Avance DRX or a Bruker Avance III spectrometer ( 1 H, 499.9 MHz; 13 C, 125.7 MHz; 19 F, 470.3 MHz) at 298 K.Chemical shifts (δ) are reported in ppm relative to the solvent residual peak of CDCl 3 ( 1 H, 7.26 ppm; 13 C, 77.00 ppm), benzene-d 6 ( 1 H, 7.16 ppm; 13 C, 128.00 ppm) or the signal of ferrocene ( 1 H, 4.00 ppm) or CFCl 3 ( 19 F, 0.00 ppm).Multiplicities are reported using the following abbreviations: s, singlet; d, doublet; t, triplet; q, quartet; hept, heptet; m, multiplet; br., broad signal or their combinations.Coupling constants (J) are reported in Hz.The assignments of the multiplicities are based on DEPT, COSY, HMQC and HMBC spectra.Mass spectra (MS) or high-resolution mass spectra (HRMS) were recorded on a Finnigan MAT 95 spectrometer (EI) with an optional Linden CMS source (LIFDI) or on a Waters Q-TOF Premier spectrometer (ESI).GC analyses were performed on a Shimadzu GC-2010 plus gas chromatograph with a flame ionization detector.