A β-ketoiminato palladium(II) complex for palladium deposition

2.1 Solid-state structure

The molecular structure of 3 in the solid state was determined by single-crystal X-ray diffraction analysis. The Ortep plot of 3 is depicted in Fig. 1 and selected bond lengths (Å), angles (°), and torsion angles (°) are given in Table 1. Suitable crystals were obtained by crystallization of 3 from a saturated dichloromethane solution at ambient temperature.

Fig. 1:

Ortep (30% probability level) plot of half of 3₂ with the atom numbering scheme. C-bonded H atoms, one disordered (0.51:0.49) molecule of hexane and the second part of the dimeric structure are omitted for clarity. Bond distances and angles are summarized in Tables 1–3.

Complex 3 crystallizes in the triclinic space group P1̅ as a dimeric species (3₂) (Figs. 2–4, Tables 2 and 3). The dimer 3₂ is formed by hydrogen bonds between the NH functionalities and the acetate ligands (N2LO5 2.838(8) Å, N5LO1 2.882(8) Å) (Table 1, Fig. 2) of the neighboring mononuclear unit resulting in a Pd–Pd distance of 3.1412(7) Å (sum of the van-der-Waals radii: 3.26 Å [46]). Dimeric structures similar to 3₂ have been reported for [Pd(OPh)₂(pyrrolidine)₂] [47], [PdCl{2-(3-C₁₂H₈-2-O)-6-CH=N(Ar)C₅H₃N}] (Ar=2,4,6-Me₃C₆H₂, 2,6-iPr₂C₆H₃, 2,4,6-iPr₃C₆H₂) [48], [49] and K[{hydroxyimio-acetate(1-)-O,N}{oxyimino-acetato(2-)-O,N}]palladate(II) [50] with comparable Pd–Pd distances.

Fig. 2:

Ortep (30% probability level) plot of the structure of 3₂ with the hetero atoms numbered. C-bonded H atoms and, the disordered (0.51:0.49) molecule of hexane are omitted for clarity. Bond distances, angles, hydrogen bond parameters and dihedral angles are summarized in Tables 1–3.

Fig. 3:

Ball and stick view along the b axis. Hydrogen atoms are omitted for clarity (Pd: pale green; O: red; N: blue).

Fig. 4:

Ortep (50% probability level) plot of the core fragment of 3₂ with the atom numbering scheme of the heteroatoms showing the [4+2] coordination at Pd. All carbon bonded hydrogen atoms, parts of the ligand and the solvent molecules are omitted for clarity. Structure parameters are summarized in Tables 1–3.

Each half of 3₂ shows a square-planar coordination sphere at Pd, whereby L acts as a tri-dendate ligand interacting via one ß-ketoiminato unit and one NH functionality for Pd(II). Together with a mono-dentately bonded acetate a distorted square-planar geometry is set up. Although the exo-C=O moiety of the acetate is -positioned within a distance defined by the sum of the van-der-Waals radii of O and Pd (3.125(7)/3.106(6) Å; Σ=3.15 Å, Fig. 4, Table 1), the O1–Pd1–O2 angles of ~45° strongly deviate from that expected for an octahedral coordination domain. Together with the Pd–Pd interaction, which is also close to the maximum of the van-der-Waals radii, a [4+2] coordination of both Pd atoms in 3₂ should be considered.

The non-complexed part of L is directed towards the perpendicularly orientated second β-ketoiminato unit in order to avoid steric interactions (Fig. 2). Within the dimer 3₂, the two ligands L are stacked with opposite directions in a C₂ fashion. Thus, the top view shows an eclipsed conformation of all Pd–X bonds (X=N, O) (Fig. 4, Table 2). The thus formed space between the two phenyl groups is occupied by a molecule of hexane, which is present as a guest solvent. The latter was refined disordered over two sets of sites (0.51:0.49). The packing motif of hexane results in a chain-type structure along [010], which is located on the midpoint of the crystallographic c axis. Hence, the 3₂ dimer is also stacked along [010], whereby adjacent units are orientated in opposite directions (Fig. 3). The non-complexed part of L shields the hexane channel.

According to NMR and IR spectroscopic investigations (vide supra), the non-complexed part of L in 3₂ crystallizes in the β-keto-enamine form, which is supported by the respective residual electron density allowing for the positioning of all NH hydrogen atoms.

The NH–C=C bond distances of the non-complexed building block of L is shortened from 1.420(10)/1.412(10) Å in a LNLCLCHL moiety (N1/N4) to 1.393(10)/1.384(11) (N3/N6) Å [51] and the C–N bond lengths are increased from 1.300(9)/1.292(9) (N1/N4) to 1.335(9)/1.340(9) Å (N3/N6) [51].

The PdN₂O planes in 3₂ form an angle of 3.8(3)° and are thus almost co-planar (Fig. 4). The Pd–N and Pd–O distances are in agreement with those in other Pd complexes, i.e. in [Pd(OAc){2-(3-C₁₂H₈-2-O)-6-(CH=NAr)C₅H₃N}] (Ar=2,6-iPr₂C₆H₃, 2,4,6-Me₃C₆H₂, 2,4,6-iPr₃C₆H₂) [47], [48], [49]. The Pd–Pd interaction between both fragments in 3₂ results in a minor shift of 0.019(3)/0.025(3) Å for the Pd atoms out of the N₂O planes of the chelating part of L towards each other. The O atom of the Pd-bonded acetate ligand is placed below these planes by 0.241(9)/0.311(9) Å, which is the result of hydrogen bonding with the NH functionality of the -non-coordinated part of L. In order to minimize steric interactions, the acetate CO₂ plane is directed away from the core fragment PdN₂O by 78.2(5)/83.0(3)° (Table 1, Fig. 4).

2.2 Thermal behavior

To determine the thermal decomposition behavior of 3₂ (Fig. 5), thermogravimetric (TG) measurements were carried out both in an atmosphere of argon (gas flow of 20 mL min⁻¹) and oxygen (gas flow of 20 mL min⁻¹) in the temperature range of 40–800°C. It has found that 3₂ decomposes with similar onset temperatures of the first decomposition steps independent of the applied gas (Fig. 5). Additional decomposition steps at >300°C are observed when oxygen was used as reactive gas. The decomposition of 3₂ under argon produces Pd ([ICDD 00-046-1043]) (Fig. 6, left), while in presence of oxygen PdO ([ICDD 00-041-1107]) is formed as evidenced by PXRD (powder X-ray diffractometry) studies (Fig. 6, right).

Fig. 5:

TG traces of 3₂ under an atmosphere of argon (solid line) and oxygen (dashed line) (gas flow, 20 mL min⁻¹, heating rate 10 K min⁻¹).

Fig. 6:

PXRD patterns of the residues of 3₂ left in TG experiments under an atmosphere of argon (left) and oxygen (right) showing the characteristic reflections of crystalline Pd and PdO, respectively.

TG-MS (thermogravimetry – mass-spectrometry) studies were carried out on 3₂ to gain a deeper insight into the decomposition behavior of this complex. The results including the TG trace, its first derivative and the ion current curves of the mass-to-charge ratios (m/z) are depicted in Fig. 7 with assignments given in the caption. It was found that at the beginning of the heating a weight loss of ca. 3% occurred between T=50 and 150°C. Within this step -fragments of m/z=12 (C⁺), 14 (CH₂⁺N⁺), 15 (CH₃⁺), 26 (C₂H₂⁺, CN⁺), 27 (C₂H₃⁺, HCN⁺), 42 (C₂H₄N⁺, C₂H₂O⁺), 43 (C₂H₅N⁺, C₂H₃O⁺), 56 (C₃H₆N⁺) and 57 (C₃H₇N⁺, C₃H₅O⁺) could be detected indicating C–C and C–N bond cleavages. The next decomposition step with an overall weight loss of 35.8% was observed from an onset of 183°C displaying the fragments m/z=14, 15, 26, 27, 30 and 57. Above 190°C, fragments such as m/z=12 (C⁺), 14 (CH₂⁺N⁺), 15 (CH₃⁺), 26 (C₂H₂⁺, CN⁺), 27 (C₂H₃⁺, HCN⁺), 30 (CH₄N⁺, CH₂O⁺), 42 (C₂H₄N⁺, C₂H₂O⁺), 43 (C₂H₅N⁺, C₂H₃O⁺), 50 (C₄H₂⁺), 51 (C₄H₃⁺), 56 (C₃H₆N⁺) and 57 (C₃H₇N⁺, C₃H₅O⁺) were detected for a weight loss of 30.9%. A mass loss of ca. 5% perceived at a temperature >350°C confirming the release of volatile components from the residue. The remaining residue of 25.7% is higher than the calculated one for the formation of Pd (18.2%) signifying the presence of impurities.

Fig. 7:

TG trace and the 1^st derivative (top) and TG-MS traces (bottom) of 3₂ (atmosphere of argon, gas flow 20 mL min⁻¹; heating rate 5 K min⁻¹; (m/z=12 (C⁺), 14 (CH₂+N⁺), 15 (CH₃⁺), 26 (C₂H₂⁺, CN⁺), 27 (C₂H₃⁺, HCN⁺), 30 (CH₄N⁺, CH₂O⁺), 42 (C₂H₄N⁺, C₂H₂O⁺), 43 (C₂H₅N⁺, C₂H₃O⁺), 50 (C₄H₂⁺), 51 (C₄H₃⁺), 56 (C₃H₆N⁺) and 57 (C₃H₇N⁺, C₃H₅O⁺).

2.3 Vapor pressure measurements

Vapor pressure measurements were carried out to prove if 3₂ is suited to be used as a CVD precursor. The methodology used is based on the mass loss of the samples as a function of increasing temperature at atmospheric pressure (Section 4) [52]. A TG system with a horizontal balance was applied to determine the weight loss in an isothermal phase at different temperatures as described previously [52]. To minimize the measurement errors and provide reliable experimental data, each study was carried out thrice.

For the measurements the temperature range of 130–190°C was chosen according to the conditions of the TG studies (Fig. 5). The results are depicted in Fig. 8. For comparison, also the vapor pressure of LH (1) is given [53]. The studies allowed the determination of the Antoine parameters, which are summarized in Table 4.

Fig. 8:

Vapor pressure traces of 1 and 3₂ in an atmosphere of nitrogen (40 mL min⁻¹).

From Fig. 8 it can be seen that 3₂ is characterized by low volatility, with the highest vapor pressure of 1.2 mbar at 190°C. Palladium(II) complexes with bis(β-diketonato) or allyl-(β-diketonato) ligands show a higher volatility at lower temperatures (e.g. 2.22 mbar at 80°C for [Pd(accp)₂] [54] (accp=2-acetylcyclopentanoate), 1.4 mbar at 60°C for [Pd(acac)₂] [55] or 3.3 mbar at 100°C for [Pd(ƞ³-2-Me-C₃H₄)(acac)] [26]). In contrast, the copper(II) complex [Cu(μ-OAc)(L)(H₂O)]₂ reported previously [53] shows a higher vapor pressure than 3₂.

As a result of the low volatility of 3₂ this complex was applied only as a spin-coating precursor for the formation of Pd and PdO deposits.

2.4 Spin-coating experiments

The spin-coating process allows the deposition of metal or metal oxide layers by utilizing non-volatile precursors [56], [57]. Hence, complex 3₂ was applied in the formation of Pd and PdO deposits exploiting this technique. Silicon wafers covered with a native SiO₂ layer were used as -substrates for the deposition studies (Section 4). The respective precursor solution was prepared by -dissolving 3₂ in acetonitrile (0.05 m solutions, Section 4). The spin coater was operated at 3000 rpm for 2 min and the -deposition procedure was repeated thrice. The respective deposition parameters are summarized in Table 5. The as-obtained deposits were then heated in a horizontal tube furnace (Pd, sample A) or in a muffle furnace (PdO, sample B) to 800°C for one (A) or 2 h (B) to leave granulated particles on the substrate surface.

2.5 Sample characterization

The morphology and the particle diameter of the as--deposited Pd clusters were studied by using SEM -(scanning electron microscopy) (Fig. 9, Figs. SI2 and SI3, Supporting Information). EDX (energy-dispersive X-ray spectroscopy) (Fig. SI4, Supporting Information) and XPS (X-ray photoelectron spectroscopy) were applied to determine the elemental composition of the respective samples (Figs. SI5 and SI6, Supporting Information).

Fig. 9:

SEM images of the Pd (sample A; top view, top left) and PdO (B, top view, bottom left and middle) residues and their cross-sectional images (A, top right; B, bottom right) deposited on silicon substrates covered with a native SiO₂ layer.

Depending on the annealing conditions, different morphologies of the as-formed deposits were observed (Fig. 9). Sample A, when annealed in an atmosphere of argon, is characterized by spherical particles at the substrate surface with a narrow size distribution (20±9 nm) (Table 5) and with the particles agglomerated to small islands. In contrast, sample B shows aspherical particles spread all over the substrate surface (Fig. 9 and SI3, left) with a higher size distribution (130±28 nm) (Table 5), probably due to the longer decomposition time compared to sample A. Next to the particles, rod like structures were found consisting of agglomerated particles. Such structures are often associated with the formation of composite materials [54]. Comparable palladium nanoparticles (NPs) were received via a CVD process using precursors like [Pd(η³-C₃H₅)(η⁵-C₅H₅)] [13], [Pd(acac)₂] [58], [Pd(hfac)₂] [59], [60] (hfac=1,1,1,5,5,5-hexafluoro-2,4-pentanedionate), [Pd(η³-2-Me-C₃H₄)(acch)] or [Pd(η³-2-^tBu-C₃H₄)(accp)] (acch=2-acetylcyclohexanoate; accp=2-acetylcyclopentanoate)) [54]. Roy and coworkers investigated the influence of the electro-less deposition of Pd NPs depending on the surface energy of the substrate [61]. Spherical-shaped particles were found at the substrate surface of Si(100) when a deposition time of 10 min was applied, while increased deposition times of 20 and 60 min led to the formation of aspherical NPs comparable to particles observed for sample B [61].

The chemical composition of the deposits was investigated by EDX spectroscopy, showing signals for Pd as well as carbon, oxygen and silicon (Fig. SI4).

Furthermore, surface sensitive ex situ XPS measurements were performed. All XPS investigations were carried out without and with sputtering of the deposits (Table SI1, see the ESI). The survey XPS spectra are depicted in Figures SI5 and SI6. Since XPS is a surface sensitive measurement method, contaminations on the surface may be overestimated. Therefore, the airborne hydrocarbon impurities and carbon surface contaminations of the precursor molecules on the surface were removed by argon sputtering (4 keV, for 60 min).

As expected, the elemental composition of the original and sputtered substrates of samples A and B differs (Table SI1, ESI), but for both deposits the palladium content increased up to approximately 1.6 at% and no carbon was detected after sputtering. Sample A exhibits a higher silicon content compared to sample B, but B also shows a significantly higher amount of oxygen after sputtering, indicating the presence of PdO and silicon oxide formation.

4.1 General procedure

All chemicals were purchased from commercial suppliers and were used without any further purification. Bis(benzoylacetone)diethylenetriamine (1) was synthesized according to a previously published method [36].

4.2 Instruments

NMR spectra (500.3 MHz for ¹H, 125.7 MHz for ¹³C{¹H}) were recorded using a Bruker Avance III 500 FT-NMR spectrometer at ambient temperature. Chemical shifts are reported in ppm downfield from tetramethylsilane with the solvent as reference signal (¹H NMR, δ (CDCl₃) 7.26 ppm; ¹³C{¹H} NMR, δ (CDCl₃) 77.16 ppm). The infrared spectrum was recorded with a Thermo Nicolet 200 FT-IR spectrometer. Elemental analysis was performed with a Thermo FLASHEA 1112 Series instrument. The high-resolution mass spectrum was recorded with a Bruker Daltonik micrOTOF-QII mass spectrometer performing in the ESI mode. The melting point was determined by using a Gallenkamp MFB 595 010 M melting point apparatus. Vapor pressure measurements were performed with a Mettler Toledo TGA/DSC1 1100 system with a UMX1 balance. The TG and TG-MS experiments were performed with a Mettler Toledo TGA/DSC1 1600 system with a MX1 balance coupled with a Pfeifer Vacuum MS Thermostar GSD 301 T2 mass spectrometer. PXRD measurements of the respective TG residues were done with a STOE-STADI P diffractometer equipped with a Ge(111) monochromator and CuKα radiation (λ=1.540598 Å, 40 kV, 40 mA). The layer formation experiments were carried out using the spin coater ws-650. The surface morphology and cross-sectional details were studied by field-emission scanning electron microscopy using a ZEISS Supra60 SEM. Energy-dispersive X-ray analysis using a Bruker Quantax 400 system attached to a SEM was applied to determine the chemical composition of the samples. The composition of the samples was investigated using a PREVAC XPS system. Monochromatic AlKα radiation (1486.6 eV) was provided by a VG Scienta MX 650 X-ray source and a monochromator. The energy distribution of the photoelectrons was measured using a VG Scienta EW3000 analyzer, which was operated at 1 eV step size and 1.0 s measurement time at each point for survey spectra. The Casa XPS 2.3.16 Pre-rel 1.4 software was used for the deconvolution of the XPS peaks. For the calculation of the atomic concentration, Scofield relative sensitivity factors (RSFs) were used. These RSFs were corrected for a monochromator-analyzer angle of 52.55°. For the escape depth correction in Casa XPS, a value of −0.75 was applied.

4.3 Synthesis of [Pd(OAc)L]₂ (3₂)

Palladium(II) acetate (2) (0.224 g, 1 mmol) was dissolved in 50 mL of acetonitrile and 0.39 g (1 mmol) of bis(benzoylacetone)diethylenetriamine (1, =LH) was added in a single portion at ambient temperature. The reaction mixture was stirred at this temperature for 16 h. Afterwards, all volatiles were removed in vacuum (1 mbar). The residue was dissolved in 20 mL of dichloromethane and the solution added dropwise to 100 mL of hexane. The precipitate was collected by filtration. After recrystallization from dichloromethane at ambient temperature, complex 3₂ could be isolated in form of yellow needles. Yield: 0.426 g (0.077 mmol, 77% based on 1). M. p. 166°C (decomposition). –Analysis calcd for C₅₂H₆₂N₆O₈Pd₂ (1110.26 g mol⁻¹): C 56.17, H 5.62, N 7.56; found C 55.74, H 5.64, N 7.62. –IR data (KBr, cm⁻¹): ν=3402 (w), 3191 (w), 3178 (w), 3095 (w), 3054 (w), 2965 (w), 2928 (w), 2877 (w), 1560 (s), 1508 (s), 1484 (m), 1458 (m), 1433 (m), 1407 (m), 1386 (m), 1347 (m), 1319 (w), 1292 (m), 1281 (w), 1260 (w), 1178 (w), 1179 (w), 1157 (w), 1143 (w), 1124 (w), 1064 (m), 1022 (w), 987 (w), 781 (m), 708 (s), 668 (m). –¹H NMR (CDCl₃): δ=2.08 (s, 6H, NCCH₃), 2.09 (s, 6H, PdNCCH₃), 2.16 (s, 6H, O₂CCH₃), 2.61–2.68 (m, 2H, CH₂), 2.89–2.98 (m, 2H, CH₂), 3.13–3.21 (m, 2H, CH₂), 3.32–3.40 (m, 2H, CH₂), 3.44–3.54 (m, 2H, CH₂), 3.57–3.65 (m, 2H, CH₂), 3.72–3.82 (m, 2H, CH₂), 3.92–4.03 (m, 2H, CH₂), 5.53 (s, 2H, CHCO), 5.71 (s, 2H, CHCOPd), 7.28–7.40 (m, 14H, ^mCH, ^pCH, C₆H₅; NH), 7.73–7.75 (m, 4H, ^oCH, C₆H₅), 7.80–7.85 (m, 4H, ^oCH, C₆H₅), 11.51 (t, ³J_HH=5.6 Hz, 2H, NH) ppm. –¹³C NMR (CDCl₃): δ=19.6 (s, O₂CCH₃), 21.1 (s, 3H, NCCH₃), 24.3 (s, PdNCCH₃), 41.7 (s, CH₂), 51.5 (s, CH₂), 54.3 (s, CH₂), 64.0 (s, CH₂), 93.1 (s, CHCOPd), 98.0 (s, CHCOH), 127.1 (s, ^oC, C₆H₅), 127.2 (s, ^oC, C₆H₅), 128.2 (s, ^mC, ^pC, C₆H₅), 128.4 (s, ^mCH, ^pCH, C₆H₅), 129.9 (s, ^mC, ^pC, C₆H₅), 130.8 (s, ^mC, ^pC, C₆H₅), 138.0 (s, ⁱC, C₆H₅) 140.3 (s, ⁱC, C₆H₅) 163.1 (s, NCCH₃) 165.0 (s, PdNCCH₃) 172.1 (s, CHCOH) 180.8 (s, CHCOPd) 188.5 (s, O₂CCH₃) ppm. –HRMS (ESI-TOF): m/z=496.1351 (calcd. 496.1220 for C₂₄H₂₈N₃O₂Pd], [M–OAc]⁺).

4.4 Single-crystal X-ray diffraction analysis of 3₂

Diffraction data was collected with an Oxford Gemini S diffractometer using graphite-monochromatized CuKα radiation (λ=1.54184 Å) at T=120 K with an oil-coated shock-cooled crystal (Table 6). The structure was solved by Direct Methods and refined by full-matrix least squares procedures on F² [62], [63], [64]. All non-hydrogen atoms were refined anisotropically, and a riding model was employed in the refinement of the hydrogen atom positions. Graphics of the molecular structures were created by using Shelxtl and Ortep [65].

CCDC 1939768 contains the supplementary crystallographic data for this paper. These data can be obtained free of charge from The Cambridge Crystallographic Data Centre via www.ccdc.cam.ac.uk/data_request/cif.

4.5 Spin coating experiments

For the spin-coating experiments 2 mL of a 0.05 m solution of 3₂ (111 mg) in acetonitrile was prepared. The deposition experiments were repeated thrice. The respective substrates of size 20 mm×20 mm were cleaned with ethanol by bath sonication and were afterwards air dried at ambient temperature.