Mechanochemical synthesis of the NiSn, CuSn bimetallic and NiCuSn trimetallic nanocomposites using various types of additives of Solid State Chemistry

milled Cu-Sn and Ni-Sn systems using grinding additives. Several chemicals were examined like NaCl, PVP, CTAB, SDS, oleylamine, n -heptane, ethylene and polyethylene glycol. X-ray diffractometry attested the varying effects of the additives on the quality and quantity of the milling end-products: in several cases complete or partial mechanical alloying occurred, while in some instances, the segregation of the starting materials were also observed veri ﬁ ed by spatially-resolved energy dispersive X-ray analysis. Dynamic light scattering measurements revealed the ef ﬁ cacy of additive amounts used on particle size reduction and size distribution of the milled bimetallic and trimetallic solids. For the Ni-Cu-Sn nanocomposites, the average solvodynamic diameters varied in the range 180 – 700 nm with 0.21 – 0.48 polydispersity values. The application of CTAB and PVP resulted in aggregated nanoparticles under 100 nm size in signi ﬁ cant amounts veri ﬁ ed by the transmission electron mi-croscopy images. The analysis of the surface plasmon resonance bands of the nanocomposites indicated the presence of the Cu(I) oxide phase, while the calculated textural parameters increased up to 5 and 24 m 2 /g and 0.01 – 0.05 cm 3 /g speci ﬁ c surface area and total pore volume values, respectively, compared to the 0.8 m 2 /g and 0.002 cm 3 /g of the milling in the absence of additives.


Introduction
Objects, which are of 1 and 100 nm size in at least one dimension, are called nanoparticles (NPs). It is well-known that the chemical and physical properties of NPs are widely different from those of the bulk materials. Nanopowders can be defined as a subclass of NPs: they are powdered elements or compounds in which individual particles fall (or are crystalline) in the nanometer scale [1]. The advantageous properties of nanopowders and even the nanocomposites (in which at least one of the phases is in the nanometric dimension) comprising of elemental metals are associated with their nanometric size [2], as the surface to volume ratio is large, and the number of surface atoms is comparable to that of the atoms inside the NP [3,4]. This, among others, results in the alteration of surface energy [5], and in several mechanical properties [6] and increases the reactivity of the nanocomposites. Nanopowders are used in a large variety of applications including surface-enhanced spectroscopy and microscopy, light harvesting, solar cell fabrication [7][8][9][10], catalysis [11], preparation of nanocomposites and sensors [12][13][14] as well as medical diagnoses and treatments [15][16][17].
The actual selection of the preparation technique is challenging task in itself. Various preparation ways result in NPs/nanocomposites with fundamentally different morphology, i.e., structure and size, and in association with this, stability and physico-chemical properties of the products. It is known well that the major branches of NP fabrications are the bottom-up and the top-down techniques. The most commonly used top-down techniques are mechanical milling [18,19], laser ablation [20], inert gas condensation [21], physical vapour deposition [22], laser pyrolysis [23] and flame pyrolysis [24].
In the present contribution, the mechanochemical route was used for the fabrication of nanocomposites. The definition of mechanochemistry, according to Heinecke, includes chemical and/or physical transformation employing the effect of mechanical energy input [25]. During the milling process, unusual transformations can also occur in the localized spots with extreme conditions (>5000 K, several tens of atmospheres with extremely short <10 -7 s relaxation times). New materials can be mechanochemically obtained from precursors by mechanical activation or they can be produced via modifying structures, which already exist [26]. Mechanochemistry is known to be advantageous against solvent-based synthesis methods, for the latters are known to be (among others) unsustainable [27,28]. Furthermore, there are several and easily available, simple to set up, moderately priced instruments with their own specific operation characters to perform the mechano-(rolling, mixer or planetary mill) treatments making them suitable for lab-scale applications.
One can find a considerable number of examples in the literature where mechanochemistry was employed in the synthesis of metallic NPs or nanocomposites. For instance, monometallic Ni and bimetallic Ni-Cu NPs where synthesized from commercial nickel powder using high-energy ball milling [29]. From copper powder, Cu [30] and Cu 2 O [31] NPs can also be prepared this way. Examples are also found for the mechanochemical preparation of other bimetallic, e.g., Fe-Cu [32] and Ni-Fe [33], NPs. Perhaps, the most intensely studied bimetallic system is the Cu-Ni one [34][35][36][37][38], and the Cu-Sn and Ni-Sn systems are much less explored. However, Pithakratanayothin et al. could synthesize Ni/Cu-Sn intermetallics in an attrition mill [39], and similar nanocomposites were prepared investigating the short mechanical activation of Cu-Sn/In and Ni-Sn/In systems [40]. Moreover, the synthesis of Sn-rich cobalt-and nickel-containing intermetallics were published using mechanical alloying [41,42], as well as a rare solid-liquid ball milling technique for the preparation of Cu-Sn and Cu-Ni-Sn intermetallic powders [43].
Not surprisingly, the preparation and the properties of trimetallic NPs/nanocomposites are largely unexplored areas. The most probable reason of this is associated with the difficulties of creating a suitable chemical environment in which all the three metals are simultaneously present in elemental state to form nanometric size objects. One of the rare literature examples includes the deposition of Ag via reduction to the surface of a bimetallic Au-Pt NP resulting in a product with core-shell structure [44]. In the same publication, the deposition of Rh on bimetallic Pd-Ag was also successfully managed using this synthesis strategy. Employing the gas-phase method, Au-Pd-Pt NPs were synthesized [45], while Fe-Ni-Ce NPs were reported to be prepared by coprecipitation [46]. Mechanochemical method was used for the preparation of Ru-Sn-Mo and Ru-Se-Mo NPs in the presence of isopropanol additive resulting in the formation of the physical mixtures of various bimetallic alloys and pure metals [47].
Using inorganic or organic additives, the process of mechanical alloying and cold-welding can be minimized by masking metal-to-metal contacts and providing with optimized balance between the fracture and coalesce of the particles [48,49]. In spite of the received increasing attention of the use of milling additives in the last decades, the scale of the frequently applied substances are quite narrow; short-chain alcohols and stearates [50], inorganic sodium salts [51], oleic acid/oleylamine [52] and hexane-heptane-benzene compounds [49,53] are the most studied ones. Therefore, during this work, the main goal was the mechanochemical preparation of bi-(Cu-Sn and Ni-Sn) and trimetallic (Cu-Ni-Sn) nanocomposites via ball milling using commercial metal powders as starting materials. In addition, we attempted to establish the effect of the quality and the quantity of various additives (investigating the most often used surfactants in wet chemistry like CTAB, SDS and PVP) on the products obtained using a set ensemble of previously optimized instrumental parameters.

Mechanochemical process for the synthesis of the bi-and trimetallic nanocomposites
The mechanochemical treatments were performed in a Retsch MM 400 mixer mill operating with two grinding jars (stainless steel, 50 cm 3 inner volume) and a grinding ball (25 mm diameter and~60 g, stainless steel) in each jar, under air atmosphere. The ball to powder mass ratio was around 100, the applied grinding frequency (12 Hz) and time interval (2 h) were fixed. The balls impinged from the rounded ends of the jars, which moved in radial oscillations along the horizontal axis resulting in sudden collisions, therefore, the induced mechanical deformations were somewhat different from the circular motion in the commonly used rolling or planetary ball mill [54].
The amount of metal solids were 0.6 g applying 1:1 Cu:Sn, 1:1 Ni:Sn and 1:1:1 Ni:Cu:Sn molar ratios, while the amount of the added surfactants varied in the range of the 50 mm 3 and 500 mm 3 for the liquid additives. For NaCl and the solid surface active agents, the used masses were between 0.5 and 5 wt% of the 0.6 g total mass of the metal particles. After milling, the samples were washed with distilled water and ethanol, collected on 0.22 μm filters, dried at 60 C and stored under N 2 .

Applied techniques for the structural characterization
The powder X-ray diffractograms were recorded on a Rigaku Miniflex II instrument using Cu Kα (λ ¼ 1.5418 Å) radiation in the Θ ¼ 5À80 range with 4 /min scan speed. The reflections were identified on the normalized diffractograms by the JCPDSÀICDD (Joint Committee of Powder Diffraction Standards À International Centre for Diffraction Data) database. The coherently scattering domain sizes (the smallest undistorted zone À average crystallite sizes) of the Cu 6 Sn 5 crystals were calculated by the Scherrer equation fitting Gaussian curves on the first reflections applying 0.9 shape factor. The analysis of the wt% distribution of the phases were performed using the integrated reflection intensities [54].
The morphologies and the sizes of the nanocomposites were investigated by scanning (SEM, Hitachi S-4700 instrument) and transmission (TEM-FEI TECHNAI G220 X-TWIN instrument) electron microscopies at varied magnifications and acceleration voltages. The dispersion of the metal elements were analyzed by energy dispersive X-ray spectroscopy accessory coupled to the scanning electron microscope (EDX, R€ ontec QX2 spectrometer equipped with Be window).
To map the average solvodynamic diameters and heterogeneity in particle sizes of the solids, a Malvern NanoZS dynamic light scattering (DLS) apparatus operating with a 4 mW helium-neon laser light source (λ ¼ 633 nm) was used. The analyses were performed in back-scattering mode at 173 , the size measuring limit of the instrument was near 6 ultrasonic radiation getting 0.1 g/dm 3 concentration.
The registration of the surface plasmon resonance of the metallic particles were done in the range of 225-800 nm using a Shimadzu UV-1650 double beam spectrophotometer with 1 cm cuvette length and 2 nm resolution. The parameters of the analyzed dispersions were the same as used for the DLS measurements.
The N 2 adsorption-desorption isotherms were registered on a Quantachrome NOVA 3000e instrument. The samples were degassed at 200 C for 3 h in vacuum to clear away the surface-adsorbents. The specific surface areas were estimated by the Brunauer-Emmett-Teller equation from the adsorption branches, while the total pore volumes were calculated with the method of Barett-Joyner-Halenda (BJH) from the desorption branches.

Preparation of bimetallic CuSn and NiSn nanostructures
The mechanochemical synthesis of the bimetallic nanocomposites was started with the investigation of the effect of the surfactants on the end-products in the milling of Cu and Sn metal powder by X-ray diffractometry. Since the number of the known and even the generally used surfactants are extremely large, the quantity and quality of the additives were varied in wide range using liquid and solid ones (at room temperature), non-ionic (n-heptane, ethylene glycol, polyethylene glycol, PVP, oleylamine), anionic (SDS) and cationic (CTAB) surface active agents with low and high molecular weights as well. However, the NaCl is not considered as surfactant, it is generally used as inert grinding media to aid the uniform energy dissipation and to work against aggregation and cold-welding [55]. In the absence of additive and on using n-heptane and NaCl (Fig. S1, "S" is the notation used in the Supporting Information), the signal of the exclusive end-product was attributed to the η-bronze intermetallic (Cu 6 Sn 5 , monoclinic, C2/c space group, 45À1488 numbered JCPDS card), while only the reflections of the Cu (JCPDS#04À0836) and Sn (JCPDS#04À0673) starting materials could be identified after the millings with oleylamine or polyethylene glycol, save the weak sign of the partial oxidation of the Cu into the copper(I) oxide phase at lower polyethylene glycol amounts (under 1 wt%, JCPDS#78À2076) (Fig. S2). In these cases, the quantities of the additives had no influence on the outcome of mechanical treatments. Significantly different behaviour was observed on working with PVP, CTAB, SDS or ethylene glycol; the formation of η-bronze alone or the physical mixture of bronze, copper and tin occurred depending on the quantity and quality of the surfactants. At higher amounts of additives, the predominant phases were the starting metals, and lower amounts favoured to the formation of the η-bronze on using PVP, CTAB or SDS (Fig. S3). Interestingly, reverse tendency could be observed with ethylene glycol: the increasing amounts of the additive aided the generation of the bronze phase (Fig. 1). The calculated values of the Cu 6 Sn 5 content were between 22 and 95 wt% (Table S1) depending mainly on the amounts of the additives used, the Sn content diminished under the theoretically minimum (10.5 wt% assuming only the formation of Cu 6 Sn 5 phase), while the amount of the unreacted Cu particles were still detectable. These observations could indicate that part of the Sn reagents became amorphous (invisible by X-ray diffractometry) and/or the bronze intermetallic was not formed exclusively with the expected stoichiometric formula, but rather with Sn-rich compositions [56,57].
The work on the Ni-Sn system resulted in the similar tendencies. Two cases were clearly distinguishable (i) solely the reflections of the nickel (JCPDS#04À0850) and tin starting materials were observed on using oleylamine and polyethylene glycol (Fig. S4) and (ii) beside these reflections, some unidentified ones were observed for the additives PVP, CTAB, SDS, ethylene glycol, NaCl or n-heptane (Fig. S5). The predominant phases were the tin and nickel on using PVP, CTAB, SDS, ethylene glycol, and only a few signals assigned to the unknown phase/phases were recorded; but, numerous strong unknown reflections appeared on applying n-heptane and especially sodium chloride and in the absence of additive, where the reflections of tin particles were even hardly observable. The identification of the new reflections were not possible either assuming the likely formation of metal oxide (or even nitride, carbonate) secondary products or the well-known transformation of the starting face-centred cubic (Cu and Ni) into the hexagonal close packed structure or the tetragonal β phase tin to the diamond cubic α phase (tin pest).
Although, several reflections could not be assigned, the signals of the hexagonal Ni 3 Sn (35À1362) intermetallic compound were clearly separated with largely similar lattice parameters (Table S2). Under 45 2θ the weak reflections remained unidentified in spite of our best efforts, but the intense signals appearing around 30 and 42 2θ could presumably be connected to the formation of intermetallics with lower Ni content (Ni 1.52 Sn [58], Ni 2.7 Sn 2 [59], Ni 3.08 Sn 4 [60], Ni 0.62 Sn 5 [61]). The generation of these solids may be verified by the slower consumption of the nickel starting reagents compared to that the tin particles had. Using n-heptane and NaCl additives, the reflections of the nickel phase were still well-observable beside the largely disappearing signals of the tin reagent in the diffractograms (Fig. S5). Mulas et al. experienced the formation of similar Sn-rich Ni-Sn intermetallics using single-ball mixer milling without any additives [41]. Moreover, suggested by the experimental results of the Cu and Sn solids milled with NaCl and n-heptane, where the formation of the Cu 6 Sn 5 intermetallic compound was mainly favoured, the generation of a NiSn intermetallic system thought to be a highly conceivable solution. The reverse tendency between the PVPÀCTABÀSDS and ethylene glycol was also observed during the milling of Ni-Sn powder (the increasing amount of ethylene glycol aided the generation of the intermetallic compounds); however, with the other additives, the reflections of the intermetallic alloys intensified when their amounts decreased.
Interestingly, from the liquid additives, significantly larger amounts were necessary (>5%, calculated in wt% compared to total mass of the starting metal particles) than from the solids to avoid the formation of thin foil on the internal surface of grinding jars from the ductile copper and tin particles. This phenomenon could be inhibited with only 1 mm 3 oleylamine, ethylene glycol or dimethyl sulfoxide when copper was milled [60]; however, for these bimetallic systems, at least 50 mm 3 of these liquids were needed. In addition, the average crystallite sizes of the formed η-bronze intermetallic showed growing tendency from 19 to 39 nm with the mass of the additives used (23 nm for the milling without additive), and for the n-heptane and ethylene glycol (applied in higher amounts), the curves went through maxima (Fig. S6).
Since the mechanical deformations (mainly plastic deformations) are induced largely by impacts with random magnitude and intensity, the evolution of specific morphologies are quite rare, as it was attested by the electron microscopic images: only planar and highly aggregated particles (flattened by the mechanical compressive forces) with smooth edges could be observed (Fig. 2). In spite of the lower than atomic resolution of SEM, from the EDX elemental maps, one could verify visually the formation of the various end-products; the generated η-bronze intermetallic (in the presence of NaCl), the physical mixture of the intact starting materials (in the presence of oleylamine), and the mixture of the intermetallics and starting Cu and Sn particles could be distinguished well (in the presence of ethylene glycol). For the NiSn solids, SEM-EDX measurements revealed amorphous, rounded particles, the elemental distribution maps verified the formation of the NiSn intermetallic alloy/ alloys demonstrated with NaCl and PVP milling additives and the physical mixture of the Ni and Sn starting materials (oleylamine) (Fig. S7). Furthermore, in all cases, the elemental analysis did not register the presence of the iron atoms as possible contaminant from the amortization of the grinding jars or balls. TEM analysis revealed Cu/Sn grains under micron sizes in agreement with DLS measurements (detailed below). Moreover, at the highest magnifications, separate nanoparticles under 50 nm were also observable in the aggregates (Fig. 3).
For the dynamic light scattering measurements, the milled Cu-Sn systems, in the presence of additives, formed light-grey semi-transparent and non-settling dispersions after 1 h ultrasonic stirring in ethylene glycol. The number-weighed size distribution curves of the ground samples displayed mainly unimodal character (not shown), the size of the aggregated crystallites had only one maximum. The liquid and solid additives displayed well-separated behaviour independently from the quality of the evolved end-products. With NaCl, PVP, SDS and CTAB, the average solvodynamic diameters remained in small range (between 540 and 230 nm), and the polydispersity indices (PDI) were under 0.6. Increasing amounts of the additives resulted in slightly decreased sizes and PDI values, except with SDS, when the lowest diameters were measured at 1.5 and 2 wt%.
For the liquid additives, the influence on the aggregation tendency was more remarkable, the values of the average solvodynamic diameters as well as the PDI values varied from 1600 to 190 nm and from 0.76 to 0.16, respectively. Oleylamine, ethylene and polyethylene glycol aided the aggregation and/or the cold-welding of metal particles at higher amounts; the lowest diameters were measured for the 50 mm 3 volume added. n-Heptane showed different behaviour, but similar to NaCl and PVP. The smallest diameters were detected when heptane was used in the largest amount.

Synthesis of the NiCuSn trimetallic system
During the syntheses, the experimental experiences were used obtained in the works with the CuSn and the NiSn bimetallic materials. To obtain particles with the smallest extent of aggregation (highlighted in Table 1), the amount of the used additives were chosen according to the DLS measurements (Fig. 4). From ethylene glycol, 125 mm 3 were applied, for the similarly small particles, but with the lowest PDI values. The revealed behaviour of the additives was the same for the trimetallic system as well: the application of oleylamine and polyethylene glycol resulted in total segregation of the starting reagents, the XRD patterns only registered the physical mixture of the Ni, Cu and Sn solids, while with NaCl and n-heptane and in the absence of additive, the Cu and Sn particles formed the Cu 6 Sn 5 phase excluding nickel. For the rest of the surfactants, the reflections of the starting reagents and the η-bronze were also observable in various phase distributions (Table S1). However, in most instances, the evolution of the cupronickel intermetallic (Cu 0.81 Ni 0.19 JCPDS#47À1406) could not be probed, because their reflections are in the same position as those of the Cu particles. Working with NaCl, n-heptane or without additive, the absence of the copper reflection allowed us to rule out the formation of the CuNi intermetallic as well [62]. Moreover, the signals of NiSn intermetallics could not be detected either indicating that the generation of the Cu 6 Sn 5 intermetallic compound is preferred in the all cases. For the binary and ternary systems, the lattice parameters (Table S1) of the formed Cu 6 Sn 5 phases were similar to the ones reported in the literature [63], but changed in a relatively wide range due to the generated lattice stress from the disordering/fracturing effect of the mechanochemical treatments [64]. When Ni was present, these parameters did not show clear changes not registering the dissolution of the nickel atoms into the Cu 6 Sn 5  intermetallic [65,66]. In addition, the phase distribution data revealed largely uniform and same as the initial (24.3 wt% in the starting reagents) content of the unreacted Ni; nevertheless, the incorporation of the nickel atoms cannot be ruled out entirely. Due to the similar physicochemical properties of the Cu and Ni atoms, in minute amount, the formation of the NiCuSn intermetallic can be presumable.
To sum up, X-ray diffractometry results of the CuSn, NiSn and NiCuSn systems revealed that the smaller NaCl and n-heptane molecules could not prevent the mechanical alloying of the starting materials as it was always recorded using no additive. Presumably, the molecules with longer hydrocarbon chains and/or polar groups would adsorb on the surface of the particles forming protective layer and sterically/electrically inhibiting the fusion of the starting metals. Since the surface of the metal particles did not have specific charge, the ionic character of the surfactants did not have significant influence on the quality of the endproducts. Oleylamine and polyethylene glycol were extremely efficient in preserving the starting materials, while this was highly dependent on the applied amounts of the PVP, CTAB, SDS or ethylene glycol.

Dynamic light scattering and electron microscopic investigations
The DLS curves of the NiCuSn nanocomposites had unimodal distribution with average solvodynamic diameters fluctuating in wide scale; however, the PDI values remained relatively similar (Fig. 5, Table 2). Compared to the curve obtained for the blank sample, the size of the aggregates could be reduced successfully applying milling additives, except in the presence of SDS. In most cases, the particles were in the 100 and 1000 nm range, but with PVP and CTAB,~11% and~19% of the particles, respectively, were smaller than 100 nm. The TEM images verified the nanometer size of the NiCuSn powders derived from the DLS measurements. At high magnifications, several particles were recorded under 100 nm aggregated into micron sized formations (Fig. 6).
On the SEM images, the already seen rounded planar morphology was observed, and the EDX elemental maps visualize the deviation between the generation of bronze intermetallic and the simple physical mixture of the three starting materials (Fig. S8). Interestingly, a slight difference was observed regarding the position of the nickel particles around the CuSn intermetallic alloy. In the presence of SDS, EG, CTAB and PVP, where the formation of the bronze intermetallic was less favoured, and the  reflections of the Cu and Sn phases were also detected, beside the Ni particles, the signal of the latter element was located separately around the bronze on the elemental maps (Fig. S8). While the distribution of the Ni elements were more uniform covering completely the atoms of the Cu 6 Sn 5 intermetallic for the work without any and with NaCl and nheptane milling additives where the end-products consisted of largely the bronze and nickel solids, the formation of the intermetallic phase was the most favoured (Fig. 7).

UV-visible absorption, specific surface area and total pore volume analysis
The reach of the nanoscopic scale allowed us to investigate the optical properties of the milled NiCuSn powders via the identification of the localized surface plasmon resonance bands in the UV-visible spectra. Since the charge density oscillations are highly dependent on the size and shape of the metallic nanoparticles, the recorded bands were relatively broad; nevertheless, three peaks were clearly observable around 265, 490 and 615 nm. The first signal was attributed to the presence of nickel nanoparticles [67,68] for all solids (not shown), the last two were derived from the Cu 2 O and Cu nanocrystals [69,70] detected clearly for the nanocomposites milled with ethylene glycol, polyethylene glycol, NaCl and especially with CTAB or PVP (resulting in the lowest solvodynamic diamaters) (Fig. S9). The X-ray diffractograms did not show the formation of copper(I) oxide phase, hence these crystals could be largely in amorphous form and/or in minute amount generated on the surface of the copper particles or even originated from the Cu starting reagents.
Since the presence of the Cu 2 O component could be a key factor regarding the catalytic potential of the NiCuSn nanoparticles, their textural features were also determined. Even though the phase distribution of the milled solids were different, they all displayed Type IV isotherms mainly with H2 hysteresis loops according to the IUPAC classification [71] indicating disordered mesopores structure with less defined size and shape distribution of the pores (Fig. S10). The solids prepared with PVP, polyethylene glycol and oleylamine showed slightly different isotherms, the registered hysteresis loops were narrower, and their shapes were closer to the curves of H3 and H4 types indicating the evolution of the slit-like pores, too. The isotherm of the solid prepared without additive was similar to those obtained in the presence of polyethylene glycol or oleylamine, but with broader hysteresis loop, presumably as sign of the larger amount of the smaller pores. The obtained values for the specific surface areas and total pore volumes were largely similar and were in a narrow range; the powders containing the maximal four phases (Cu 6 Sn 5 , Ni, Cu, Sn particles) attested the largest textural parameters, moreover, the application of milling additives could significantly enhance these values in all instances ( Table 2).

Conclusions
The surfactant/additive-assisted mechanical treatment proved to be a simple and convenient technique to prepare CuSn, NiSn and NiCuSn metallic nanocomposites from commercially available starting metallic particles even in the large scale. While the quantity of the additives had varying influence on the size and size distribution of the frittered particles depending on the composition of the starting materials as well, the effect of their quality brought about three synthesis ways towards the different milling end-products. On the basis of the experimental data obtained with the CuSn, the NiSn and the NiCuSn systems, in all cases the following observations were made (i) independently from the amount of the additives used in the investigated range, the oleylamine and polyethylene glycol inhibited the mechanical alloying, i.e. the starting materials could be preserved in separated forms, (ii) the application of NaCl and n-heptane molecules could not hinder the formation of the intermetallic phases as it always occurred in the absence of additive, (iii) the outcome of the millings with PVP, CTAB, SDS and ethylene glycol was highly dependent on the quantity of the applied additives, increasing amounts favoured the formation of the intermetallic compounds for the ethylene glycol, while with PVP, CTAB and SDS the reverse tendency was observed: with higher amounts, the starting materials could be kept separated.

Declaration of competing interest
The authors have no conflicting interests of any kind.