Ortal Lidor-Shalev and
David Zitoun*
Department of Chemistry, Bar Ilan Institute of Nanotechnology and Advanced Materials (BINA), Bar Ilan University, Ramat Gan, 52900, Israel. E-mail: David.Zitoun@biu.ac.il
First published on 18th November 2014
The understanding of the chemical reaction yielding nanomaterials is often based on basic assumptions and novel syntheses are usually presented without knowledge on their reaction mechanism. Herein, the chemical reduction of metal precursors by Lewis complexes of amine–borane is fully investigated, reaction and mechanism, for the synthesis of metallic nanoparticles (NPs). In this study, Sn NPs are obtained by the chemical reduction of tin acetylacetonate complex [Sn(acac)2] using dimethylamine–borane (DMAB) as a reducing agent. The reaction yields single crystalline Sn NPs independently of the use of surfactant. The size of the NPs can be adjusted by regulating the complex/DMAB ratio. The use of DMAB as a reducing agent is extended to the synthesis of Ni, Pd, Pt and Sn/Pd NPs. We investigate the synthetic mechanism by using 11B-NMR and calorimetric analytical tools. The mechanism is found to be specific to the acetylacetonate complexes and involves the reaction of borane on the ligand.
Research on Sn NPs is not very common compared to noble metals, since it is hard to control the surface properties of the NPs. The synthesis of Sn NPs is relatively complicated, mainly due to the need for strong reducing agent or very reactive metal–organic precursors. Tin and tin oxide (SnO2) NPs and bimetallic systems can be obtained by different methods.23–25 However, some of those methods may lead to particle coalescence and growth. The most appropriate chemical method for obtaining relatively small crystalline Sn NPs is the chemical reduction of organometallic or metal–organic precursors in the presence of surfactants. The demand for Sn and SnO2 NPs is mainly for MOS (metal-oxide semiconductor) hydrogen sensors and as anodic material for Li-ion batteries26,27.
In this article, we propose to combine the use of a simple metal–organic precursor (Sn(acac)2) with amine–borane (AB) in the absence of surfactants. The importance of our synthesis is the ability to use a derivative of AB as a source for hydrogen molecules in a redox reaction leading to metallic NPs. In this research we report a new synthetic path for the synthesis of Sn NPs, with various sizes, by reducing metal precursors using DMAB as a reducing agent for obtaining the crystalline Sn NPs. The new route enables formation of Sn NPs with a thin oxide layer on their surface that have a tremendous advantage of a clean surface without using stabilizing ligands. Different conditions, such as time, concentration, surfactants, temperature, metal–organic precursor and reducing agent, influence the obtained Sn NPs and affect the yield. Various parameters are changed to obtain Sn NPs with a high homogeneity, a small size distribution, a stable colloidal system, a defined shape, and a clean surface. The same method, with DMAB as the reducing agent, is used in order to obtain monometallic and bimetallic assemblies of colloids. The use of DMAB as a reducing agent is extended to the formation of Ni, Pd, Pt and Sn–Pd NPs. In addition, the synthesis mechanism is investigated by analytical tools, like calorimetry and NMR spectroscopy.
TEM images show the round morphology of the obtained Sn NPs from a stable DMSO colloidal solution (Fig. 1a) with the occurrence of faceted nanoparticles (mainly cubic). Selected area electron diffraction (SAED) shows highly crystalline NPs with typical planes of tetragonal Sn (Fig. 1a inset), following the bulk tetragonal structure of Sn (JCPDS # 04-0673). The crystalline structure is supported by powder X-ray diffraction (XRD) results with the tetragonal cell parameters; a = b = 5.831 Å, c = 3.181 Å (Fig. 1b). No other phase, by-products, or crystalline oxides were observed, demonstrating the selectivity of the reaction.
![]() | ||
Fig. 1 (a) TEM images of Sn nanoparticles with 1![]() ![]() |
HRTEM of single nanoparticle reveals the single crystalline nature of the particles within a tetragonal phase (Fig. 2a). The NPs surface is covered with a thin amorphous layer with a thickness of 3.7 ± 1 nm (Fig. 2a). X-Ray photoelectron spectroscopy (XPS) of the nanoparticles at the Sn edge 3d5/2 can be fitted with a main component (91%) centered at 486.8 eV (SnIV) and a minor component (9%) at 484.7 eV (Sn0) (Fig. S1†). Such a spectrum is typical of metallic tin with an SnO2 oxide shell. The Sn NPs have a spherical shape with a few square particles and some faceted ones (Fig. 2c). The XPS results are supported by energy dispersive X-ray spectroscopy (EDS) (Fig. S2b and S2c†) of a single nanoparticle (Fig. S2a†). The EDS results show an increase of oxygen/tin atomic ratio from the core to the shell of the nanocrystal.
The Sn NPs polydispersity and the NPs mean size (37 ± 13 nm) were analyzed according to TEM images based on a statistical analysis of more than 200 NPs (Fig. 2d). In addition to the NPs mean size from TEM, the size of the colloids in DMSO dispersion was measured by dynamic light scattering (DLS) and shows size distribution of a Gaussian centered at 37 nm and a secondary peak at 225 nm (Fig. 2b). The main distribution of hydrodynamic diameters corresponds to the nanocrystal size distribution calculated from TEM images. The secondary peak may originate from the aggregation of the NPs that might occur in the absence of surfactants.
![]() | ||
Fig. 3 TEM images of Sn nanoparticles with different ratios Sn(acac)2/amine–borane: (a) 1![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
![]() | ||
Fig. 5 TEM images of a rapid reaction by using Sn(hfac)2 as a precursor for obtaining Sn NPs [ratio Sn(hfac)2/amine–borane 1![]() ![]() |
The DSC curves on the pristine DMAB (Fig. 7a) show an endotherm at 36 °C (melting point) and at 164 °C (boiling point), all in accordance with the literature.30,31 An exotherm can be observed almost at the same temperature (around 160 °C) which accounts for the formation of a DMAB dimer with H2 release, as already observed from NMR and reported in the literature.32 The same procedure is now applied to a mixture of DMAB with Sn(acac)2. DSC results of different DMAB/Sn(acac)2 ratios are shown (Fig. 7, graphs b–d). The spectra do not display any endotherm and the main exotherm at 70 °C is attributed to the reaction between Sn(acac)2 and DMAB, yielding Sn NPs.
![]() | ||
Fig. 7 Differential scanning calorimetry (DSC) measurements of (a) DMAB, (b) Sn(acac)2/DMAB 1![]() ![]() ![]() ![]() ![]() ![]() |
Sn/Pd precursors yield crystalline Sn NPs of 45 ± 9 nm with a crystalline structure corresponding to the bulk tetragonal Sn and crystalline Pd NPs of 5 ± 1 nm with a crystalline structure corresponding to the bulk FCC Pd. The reaction with two different precursors yields two distributions of colloids which obviously correspond to large particles of Sn and small particles of Pd. No intermetallic phase can be observed, certainly due to the low temperature of reaction. Such an approach allows for the decorrelation of the two reactions and could be extended to several bi-metallic systems. The only prerequisite seems to be the use of acetylacetonate complex.
The standard deviation remains low even for large particles, in the range of 20%. This result is noticeable for NPs synthesized in the absence of surfactants, which are usually the main pathway towards low polydispersity. The addition of hexa decylamine as steric surfactant improves the size deviation without a significant change in the mean diameter (Fig. 4). We assume that hexadecylamine can limit the particle size growth by the stronger affinity to NPs, as previously demonstrated on Sn NPs.23
In all cases, the particles are suspended in a polar solvent after washings; DMSO displays the best dispersion as the electrostatic repulsion stabilizes the colloidal solution. The use of a steric stabilizer is not necessary and the DMSO dispersions are stable for weeks.
In this research, Sn(acac)2 is used as a metal–organic precursor to yield Sn NPs in combination with amine–borane complexes. Such a combination has not been reported before. Other complexes with dimethylamine moieties have been thorough fully investigated by Chaudret et al. for the syntheses of Sn and SnO2 NPs from {Sn(NMe2)2}2 metal–organic precursors. Sn/SnO2 NPs in the range 10–20 nm are obtained by the thermal decomposition of {Sn(NMe2)2}2, followed by a controlled surface hydrolysis and finally oxidation to SnO2.33,34 The homoleptic Sn(II) amides have been singled as metal precursors because of the relative thermodynamic weakness of the Sn–N bond {Sn(NMe2)2}2 which enables reactivity even at low temperature. Noteworthy, the dimethylamine moiety resembles the Lewis base in the amino–borane complex. Therefore, the reaction we propose is an alternative to the synthesis of Sn(II) amide complexes with a final size distribution in the range 20–100 nm which has not been achieved with the {Sn(NMe2)2}2 complexes. The relative similarity of the two reactions bring out a question about the reaction mechanism. Does the reaction involve intermediate species related to the amide complexes family?
To answer this question, we could consider the reactions involving a transfer of hydrides to Sn complexes, as reported in the literature. For instance, single-crystalline Sn nanorods are synthesized using a multistep seed-mediated method involving the reaction of SnCl2 with NaBH4 in tetraethylene glycol (TEG), in the presence of poly(vinylpyrrolidone) (PVP).35,36 An addition of metals salts to the β-Sn seed solution yields intermetallic M–Sn (M = Co, Ni, Au, Cu, Ag, Ru, Pt) nanocrystals. In this study, the reaction mechanism of borohydrides at high temperature has not been elucidated. Since no investigation of reaction between amine–borane complexes and Sn complexes is reported in the literature, we shall compare the mechanism with other metallic systems.
On the one hand, the synthesis of Ni NPs from the reduction of Ni(acac)2 with tributylamine–borane (BTB) has been reported in the presence of oleylamine and oleic acid.37 These particles can catalyse amine–borane dehydrogenation. Following the same schema, Pd NPs for catalytic oxidation of formic acid in fuel cell conditions can be obtained by reducing Pd(acac)2 with BTB in the oleylamine-mediated synthesis.38 The AB derivative t-butylamine–borane complex also reduces HAuCl4 in the presence of OAm to yield Au NPs.39 Unfortunately, the very interesting results reported in the previous studies are not supported by mechanistic studies.
On the other hand, amine–borane complexes also react with amide metal–organic complexes. Diisopropylamine–borane (AeB) has been investigated as reducing agent for the synthesis of Fe NPs.40 As proposed by the authors, the reaction between AeB and Fe{N(SiMe3)2}2 complex generates diisopropylimine–borane (AoB) and H2 which reduces Fe{N(SiMe3)2}2, followed an unclear further decomposition with B incorporation in Fe NP.
The same reaction has been performed with Co{N(SiMe3)2}2 and followed by reaction with a Rh(allyl)3 complex.41 The bimetallic particles are formed in a sequential manner which allows for the directed segregation of the two metallic species in the core (Fe or Co) and on the surface (Rh).
In the present research, we additionally describe the synthesis of Ni, Pd, Pt NPs, using the same synthetic pathway of metal–acetylacetonate reaction with amine–borane. This chemical route yields metallic NPs with a size suitable for catalysis (3–5 nm) in the absence of surfactants and stabilizers, i.e. with a clean surface. Our results on bimetallic Sn/Pd are consistent with the versatility of the amine–borane reduction pathway using stable metal–acetylacetonate complexes.
As exposed in the introduction, our primary goal is to circumvent the use of H2, which poses issues of safety, by generating in situ H2 from the DMAB decomposition. The DMAB thermal decomposition starts at 130 °C to release a hydrogen molecule and dimer structure. Our DSC measurements show a decomposition temperature of DMAB which corresponds to the release of hydrogen molecules. However, the reaction with Sn(acac)2 occurs at temperature as low as 70 °C.
The analysis of the 11B-NMR spectra clearly shows the formation of a new species which differs from the dimer (di-borane) resulting from H2 generation (Fig. 9a). The new species displays a peak at 18.5 ppm (11B-NMR spectrum), which increases with the ratio of Sn(acac)2 versus amine–borane. Since acetylacetonate is the only source of oxygen in the reaction, this peak certainly corresponds to the formation of a compound with a B–O bond which results from a reaction between the borane and acetylacetonate moieties. On Fig. 9b, we propose a possible chemical pathway which accounts for the chemical reduction of Sn(II) by the hydrogen atoms of the borane and the amine, followed by the subsequent nucleophilic attack of the oxygen on the borane. The reduction mechanism goes through the formation of a cyclic intermediate. The intermediate complex coordination number increases from 4 to 6 and the coordination changes from tetragonal to octahedral. In the next step, one of the B–H bonds is released and acetylacetonate group coordinates the boron to give a possible intermediate. An additional intermediate is formed by the reaction of one more amino-borane equivalent, and subsequently forming the B–acetylacetonate bonds. This complex, presumably dioxaborine (Fig. 9b), is analyzed as the only by-product by 11B-NMR (singlet, 18.5 ppm).
![]() | ||
Fig. 9 (a) DMAB dimerization mechanism and (b) possible reaction mechanism in the presence of Sn(acac)2. |
The mechanism demonstrates the amine–borane (DMAB) coordination to the Sn(II) metal center and its reduction to yield Sn NPs. The synthetic mechanism is found to involve the reactivity of acetylacetonate ligands with borane. The chemical reduction of Sn(acac)2 is a competitive reaction for the DMAB thermal decomposition and the results obtained with Sn(hfac)2 strengthens this point. The reaction kinetics are greatly enhanced by the presence of electronegative CF3 groups, showing the pivotal role of the ligands during the reaction of Sn precursors.
The reaction mechanism is very different from the one reported on M{N(SiMe3)2}2 (M: Co or Fe) where the AB complex acts as a local source for H2.39,40 We believe our findings will help in the investigation of the more stable acetylacetonate complexes as metal precursors for nanomaterials synthesis.
The Sn(acac)2 (0.5 mmol) was stirred in a vial with 5 mL mesitylene. DMAB (2 mmol) was dissolved and stirred in a vial with 5 mL mesitylene at 40 °C for 10 min. A 3 neck-flask was connected to a Schlenk line and to a condenser with water stream. The flask was flushed with N2 and the two solutions were added simultaneously by fast injection. The solution was stirred at 150 °C for 2 hours under constant nitrogen bubbling. At the end of the synthesis, the dark colloidal (brown-black color) solution was left and cooled to room temperature. In order to separate the solid product, the dispersion was centrifuged at 4000 rpm for 10 min and redispersed in the same solvent (mesitylene, 2 mL) this procedure was repeated three times. The cleaning procedure was followed by three washes with DMSO (dimethyl sulfoxide, 2 mL) at 4000 rpm for 10 min. Finally, the obtained Sn NPs were dispersed in DMSO.
The effect of different precursor/reducing agent ratios was investigated, using the stoichiometric Sn(acac)2/DMAB ratios of; 1:
2 [0.5 mmol Sn(acac)2, 1 mmol DMAB]; 1
:
4 [0.5 mmol Sn(acac)2, 2 mmol DMAB]; 1
:
8 [0.5 mmol Sn(acac)2, 4 mmol DMAB]; 1
:
12 [0.5 mmol Sn(acac)2, 6 mmol DMAB]. The total volume of the solvent was constant at 10 mL. The solution was stirred at 150 °C for 2 hours under constant nitrogen bubbling, using a 3 neck-flask.
This procedure was repeated using Sn(hfac)2 instead Sn(acac)2 with the same experimental protocol with a ratio Sn(hfac)2/DMAB 1:
2 [0.5 mmol Sn(hfac)2 and 1 mmol DMAB].
The effect of a steric surfactant was examined by using different stoichiometric ratios of hexadecylamine (HDA) and a constant quantity of DMAB [0.5 mmol] and Sn(acac)2 [0.25 mmol] as described for the surfactant-free reaction. The addition of 0.5 mmol and 1 mmol HDA correspond to the metal precursor/surfactant ratio of 1:
2 and 1
:
4 respectively.
The reaction between M(acac)2 (M = Ni, Pd, Pt) and DMAB was done using the previously described procedure (Synthesis of NPs). The stoichiometric precursor/reducing agent ratio was kept at 1:
2. The syntheses were performed for 2 hours at 50 °C under continuous stirring. The solutions were cleaned by the previously described cleaning procedure.
Differential scanning calorimetry (DSC) was done with a Mettler Toledo DSC-822 system equipped with a liquid nitrogen cooling system, at a heating rate of 3 °C min−1, in the temperature range of 25–250 °C. The enthalpy profile of the thermal decomposition of DMAB was investigated for different stoichiometric Sn(acac)2/DMAB ratios (1:
1, 1
:
2 and 1
:
4) in a solvent free reaction.
11Boron-NMR (instrument used: Bruker Avance-400, 128.3 MHz for 11B) spectra were used in order to prove the supposed mechanism. The reactions between Sn(acac)2 and DMAB with 1:
1 and 1
:
4 Sn(acac)2/DMAB ratios were investigated in a solvent-free reaction and were compared with the typical spectrum of DMAB. In addition, the 11B-NMR spectra of DMAB annealed at 150 °C, 200 °C and 300 °C, under continuous stirring for 48 h, were investigated. At the end of the reaction, CDCl3 was added to the mixture and the filtrated supernatant was checked by NMR.
Footnote |
† Electronic supplementary information (ESI) available. See DOI: 10.1039/c4ra11483c |
This journal is © The Royal Society of Chemistry 2014 |