Triphenyl phosphite as phosphorous source for the scalable and cost-effective production of transition metal phosphides.

: Transition metal phosphides have great potential to optimize a number of functionalities in several energy conversion and storage applications, particularly when nanostructured or in nanoparticle form. However, the synthesis of transition metal phosphide nanoparticles and its scalability is often limited by the toxicity, air sensitivity and high cost of the reagents used. We present here a simple, scalable and cost-effective ‘heating up’ procedure to produce metal phosphides using inexpensive, low-toxicity and air-stable triphenyl phosphite as source of phosphorous and chlorides as metal precursors. This procedure allows the synthesis of a variety of phosphide nanoparticles, including phosphides of Ni, Co and Cu. The use of carbonyl metal precursors further allowed the synthesis of Fe 2 P and MoP nanoparticles. The fact that minor modifications in the experimental parameters allowed producing nanoparticles with different compositions and even to tune their size and shape, shows the high potential and versatility of the triphenyl phosphite precursor and the presented method. We also detail here a methodology to displace organic ligands from the surface of phosphide nanoparticles which is a key step towards their application in energy conversion and storage systems.

Typically, TMP NPs are produced from the reaction of a metal precursor with substituted phosphines in the presence of aliphatic amines and acids.This strategy was initially applied to the synthesis of phosphides of manganese, iron, nickel, and cobalt using trioctylphosphine (TOP) as source of phosphorous. 36,37,48 Lter on, it was extended to a wider range of metal phosphides including Pt, Rh, Pd, Au, Ag, Ga, In, and Zn. 49,50 terwards, a combination of TOP/TOPO was employed in the synthesis of MnP 42 and FeP 37 at higher reaction temperatures >350°C, although the role of TOPO in the NP formation is yet to be elucidated.Phosphines used as a source of phosphorous has been not limited only to TOP, but triphenylphosphine and tributylphosphine have been also used. 30,31 hese and other achievements have been thoroughly reviewed. 29he exact mechanism of formation of phosphide NPs is not yet clear.Some of the publications suggest that metal NPs nucleate first and only in a second step they react with TOP (or other phosphines) to yield the corresponding phosphides.This mechanism was demonstrated in the one-pot synthesis of Ni 2 P. 50 In this direction, metal or even metal oxide NPs can be initially produced and in a second step reacted with a phosphorous source to produce the corresponding phosphides. 20,21,51 osphidation of pre-formed metallic NPs frequently results into hollow nanostructures through the Kirkendall effect. 52,53 is is for example the case for nickel phosphide 23,25,[31][32][33][34] and cobalt phosphide 20,21 NPs.However, void formation can be avoided by keeping the P/Me ratio sufficiently high (> 2.8 ). 31,54 Iportantly, the crystallinity of the phosphide NPs can be also tuned through the reaction temperature, producing amorphous NPs at low temperatures, [30][31][32] and crystalline NPs at higher.In terms of NP geometry, no general rationale has been stablished, and NPs with different shapes, including spherical, cubic, rod-like or sea urchin-like, have been produced depending on the crystal structure of the phosphide and potentially of the metal or metal oxide initially formed and the synthetic conditions.As an example, sea urchin-like cobalt-iron phosphide NPs were obtained from the reaction of cubic oxide NPs with TOP as a source of phosphorous. 55esides phosphines, other dangerous and toxic precursors such as P(SiMe 3 ) 3   38, 41, 56, 57   , white phosphorous 51,58,59 and even PH 3 gas 60,61 have been explored as phosphorous source for the preparation of phosphide NPs.On the other hand, commercially available metal precursors have been mostly limited to carbonyls, acetylacetonates and complexes with 1,5cyclooctadiene, with the exception of the use of CuCl to produce Cu 3-x P 39 and Cu 3 P 40,62 .These are relatively costly precursors compared with chlorides.Even in the case of copper, the used CuCl is much more expensive than the more common CuCl 2 .Alternative single-source precursors have been also used to produce phosphide NPs, but these are even less costeffective than commercial options.
Overall, the reported procedures to produce metal phosphides generally possess a number of significant drawbacks which limits their potential for scale up and commercialization.In particular, TOP, the most commonly used source of phosphorus, is expensive, toxic and relatively unstable as it readily oxidizes at ambient conditions in air and thus needs proper storage and manipulation under inert atmosphere.Thus, we propose here the use of triphenyl phosphite (TPOP) as a convenient alternative to the sources of phosphorous currently used.TPOP has a very moderate cost and is stable in ambient conditions.To the best of our knowledge, in only one previous report the reactivity of TPOP was studied among that of several other candidates as a source of phosphorous for the synthesis of nickel phosphide. 63Surprisingly, in spite of the promising results obtained in this previous work, no follow up report has considered TPOP as phosphorous precursor.Apart from this, TPOP has been rarely used as a stabilizer in the synthesis of Au 64 and CuInS 2 65, 66 NPs.We report here the synthesis of various transition metal phosphides using this low cost and stable phosphorous precursor, a simple and scalable heating up procedure, 67 and chlorides as low cost metal sources.were purchased from Sigma Aldrich.Chloroform, isopropanol, ethanol and tetrahydrofuran (THF) were of analytical grade and obtained from various sources.All precursors and solvents were used without further purification.
In the following, base synthetic procedures are described.All used variations are thoroughly explained and discussed in the main text.
Phosphides of cobalt, nickel and copper.Standard synthesis (Me:HDA:TPOP = 1:10:10).Metal phosphide NPs were prepared by the reaction of metal chlorides with TPOP in the presence of HDA.In a typical synthesis, 0.5 mmol of MeCl x (corresponding crystalohydrates were dissolved in a minimum amount of EtOH or THF prior addition) was mixed with 5.0 mmol (1.207 g) of HDA, 5.0 mmol (1.551 g) of TPOP and 10 g of ODE.The obtained mixture was heated under argon flow to 150 °C and maintained at this temperature for ~1 h to dissolve the metal precursor and to ensure the removal of traces of low-boiling-point impurities, ethanol (or THF) and water.After purging, the solution was heated to boiling point (which is about 280 -300 ºC, depending on the particular synthetic condition) and kept at that temperature during the desired period of time.Then, the mixture was allowed to cool naturally to 200 °C, point at which the heating mantle was removed.Finally, NPs were thoroughly purified by multiple precipitation and redispersion steps, using 2propanol and chloroform.
Phosphides of iron and molybdenum.Standard synthesis.Unlike the other cases, carbonyls instead of chlorides were used to synthesize phosphides of iron and molybdenum.0.5 mmols of Fe(CO) 5 (98 mg; 66 µL) dissolved in 0.2 mL of ODE or Mo(CO) 6 (132 mg) were added after the solvent cleaning step, just before the heating of the reaction mixture was started.Fe(CO) 5 was injected by syringe, and Mo(CO) 6 was quickly added into the opened flask under Ar flow.The rest of the synthetic procedure was as described above.
Tenfold synthesis (Me:HDA:TPOP = 1:5:5).To scale up the NP production a tenfold, the metal precursor amount was scaled-up by a factor of 10, HDA and TPOP by a factor of 5 and the amount of ODE was scaled-up only 2 times.All other synthetic conditions were the same as in the standard version of the synthesis.
Ligand exchange strategy: Organic ligands were displaced from the surface of Co 2 P NPs using HPF 6 .Note, that all the other tested compounds, such as NaN 3 and even [NBu 4 ]PF 6 , failed to displace original organic molecules.The HPF 6 ligand exchange was performed by adding a drop of HPF 6 (63 wt % in water) and 10 ml of formamide (or water) into a 10 ml solution of Co 2 P NPs in chloroform.The mixture was vigorously stirred and let stand until phase separation was observed.NPs moved from the chloroform to the formamide (or water) phase.The final formamide or water solution containing the NPs was washed several times with chloroform to drag all the remaining organic ligands surrounding the NPs.NPs were finally precipitated using acetonitrile and redispersed in N,Ndimethylformamide. Finally, the solution was cleaned by precipitation with acetone and redispersion in DMF.
Instrumentation.The morphological, chemical and structural characterization of the NPs was carried out by transmission electron microscopy (TEM) using a ZEISS LIBRA 120, operating at 120 KV and JEOL 1011 operating at 100 kV.Carbon-coated TEM grids from Ted-Pella were used as substrates.High resolution transmission electron microscopy (HRTEM) images were obtained using a FEI Tecnai F20 fieldemission gun microscope operated at 200 keV with an embedded Gatan image filter for electron energy loss spectroscopy (EELS) analyses.Images were analyzed by means of Gatan Digital Micrograph software.Powder X-ray diffraction (XRD) Figures 1-5 display representative TEM images of the 10±1 nm quasi-spherical Ni 2 P (figure 1), 5±1 nm x 22±5 nm rodshaped Fe 2 P (figure 2), 6.9±0.8 nm x 33±9 nm rod-shaped Co 2 P (figure 3), 4.5±1 nm x 17±1 nm disk-shaped Cu 3 P (figure 4) and the < 5 nm quasi-spherical MoP (figure 5) NPs prepared following the above detailed procedure.In the case of Co 2 P, both quasi-spherical and rod-like particles could be produced, depending on the synthetic conditions, as further discussed below.
Ni 2 P, Co 2 P and Cu 3 P NPs were obtained upon heating of the metal chlorides in a mixture of TPOP as a phosphorous source and possible stabilizer, HDA as stabilizer and ODE as a solvent.The formation of phosphide NPs started when the temperature was raised above ca.250 ºC as noted by a gradual darkening of the reaction mixture.The reaction yield, measured from the collected NPs after purification, was above 80% in all cases.
The syntheses of iron and molybdenum phosphides included few minor differences from the other phosphide NPs.Iron chlorides, either FeCl 3 or FeCl 2 , could not be used as metal precursor due to the exceptional stability of the formed iron complexes.Even prolonged heating did not provide iron phosphide using these precursors.Higher temperatures could make the reaction possible, but temperature was limited by TPOP by-products.MoCl 5 -based complexes are much less stable than those obtained from iron chlorides, but they neither resulted in any identifiable molybdenum phosphide phase upon reaction with TPOP (Figure S1).Subsequently, iron and molybdenum phosphides were produced using the corresponding carbonyls as metal precursors.It is important to mention that both Fe(CO) 5 and Mo(CO) 6 are low boiling point compounds and they must be introduced to the synthesis just before the final heating stage (after the solvent cleaning stage) to prevent their loss during solvent cleaning.
Generally, amorphous NPs were initially formed and prolonged reaction times were required to properly crystalize them.For all the phosphides reported here, except for molybdenum phosphide, crystalline NPs were obtained in solution at temperatures below 300 ºC.In the case of MoP, a postpreparative thermal annealing of the dried NPs at a higher temperature, 800 ºC, was needed to form the proper crystalline structure.Figure 5e displays the XRD patterns of the MoP NPs before and after a thermal process at 800 ºC for 2 h.
The stable hexagonal crystallographic structures were obtained for all the metal phosphides produced and regardless the experimental conditions and synthesis details.Only in the case of nickel, the tetragonal Ni 12 P 5 phase (space group I4/m) was also obtained during the initial stage of the synthesis, as further discussed below.In particular, Ni 2 P NPs displayed a hexagonal crystal phase, space group (SG) P 321 , with a=b=0.588nm, c=0.336 nm (Figure 1b).Similarly rod-shaped Fe 2 P NPs had a hexagonal crystal phase, SG P-62m, with a=b=0.5865nm, c=0.3456 nm (Figure 2b).Co 2 P nanorods also displayed a hexagonal phase, SG P-62m, with a= 5.7420 Å, b= 5.7420 Å, c= 3.4570 Å (Figure 3b).Cu 3 P NPs showed the hexagonal phase, SG P63cm, with a= 6.9593 Å, b= 6.9593 Å, c= 7.1430 Å (Figure 4b).

Figures 1c-5c display EELS chemical composition maps of the different phosphides (see additional data in the supporting information)
. A uniform distribution of phosphorus and metal atoms in each of the analyzed NPs was observed.However, phosphorous-and especially oxygen-rich surfaces were also detected.Surface O and P atoms most probably reflected the presence of an oxygen-containing stabilizer, probably TPOP, in a shell surrounding each NP.It is also possible that a partial surface oxidation of the NPs took place during post-synthetic washing step, where all procedures were carried out under ambient conditions.In addition, a thin oxide shell could be formed or grown during sample preparation for TEM characterization, as we used oxygen-Ar plasma cleaning to remove organic contamination.Importantly, the proposed synthetic strategy could be easily scaled up. Figure S2 shows Co 2 P NPs produced when scaling up the synthesis procedure a tenfold.Even larger amounts of TMP NPs per batch might be prepared simply by using large reaction vessels and through further concentration increase, as it has been demonstrated for a wide range of NPs. 68,69 e case of Ni was particularly helpful to provide additional information about the NP formation path.Hexagonal Ni 2 P NPs where obtained from the reaction of NiCl 2 with TPOP in the presence of HDA and at a temperature close to 300 ºC as detailed above.However, when the reaction temperature was limited to below 250 ºC, close to amorphous Ni-P NPs were obtained (figure 6e).The exact Ni:P stoichiometry of these NPs could not be determined since large amounts of unreacted TPOP were systematically left on the analyzed samples, possibly in part bound at the NP surface.Besides, when the amount of HDA was reduced by half (Me:HDA:TPOP ratio 1:5:10) and the reaction was quenched immediately before the solvent boiling point was reached, Ni 12 P 5 NPs with a tetragonal phase, SG I4/m, were obtained (figure 6).From these observations, we conclude that the reaction of NiCl 2 with TPOP in the presence of HDA initially results in amorphous Ni or Ni-P NPs.Subsequently, we hypothesize that phosphorous ions are being introduced in the NPs, which may initially crystallize in the tetragonal Ni-rich phase Ni 12 P 5 and later, when a higher amount of P has been introduced, on the hexagonal Ni 2 P.This hypothesis remains to be fully confirmed with additional in situ experiments.The variation of the reaction conditions substantially influenced the size, shape and crystallinity of the phosphide NPs produced.As an example, the evolution of the crystallinity of the Ni 2 P NPs depended on the TPOP concentration.Figure 7 shows TEM micrographs of the Ni 2 P NPs produced with different amounts of TPOP.While the size of Ni 2 P NPs increased with the concentration of TPOP, their crystallinity, in terms of the size obtained from XRD patterns through the Scherrer equation, did not change significantly (figure 7e).Similarly, the reaction time did not have a significant influence on the material crystallinity once formed the Ni 2 P phase (figure S9).It is worth to mention that the geometrical size, obtained from TEM images, of the Ni 2 P was systematically larger than the size obtained from the fitting of the XRD pattern (figure 7e).While the rod-like NP geometry was preserved, the monodispersity and overall NP quality significantly decreased when increasing the Me/HDA ratio.Reducing the amount of HDA, Co 2 P and Fe 2 P NPs lost the accurate cylindrical shape, becoming more irregular and with less defined edges.Generally, a Me:HDA:TPOP = 1:10:10 ratio resulted in highest quality NPs.However, more detailed and systematic experiments need to be done to fully understand how the NPs' morphology depends on the experimental conditions and to fully optimize the synthetic protocol, especially when scaling up and maximizing the reactant concentrations.
Besides TPOP, FTIR analysis indicated the presence of HDA at the NPs surface (figure 9).The presence of these organics may strongly reduce the NP performance in all applications involving charger transport or transfer from/to the NPs, including catalysis.Surface ligands could not be displaced using previously reported procedures with molecules such as [NBu 4 ]PF 6 , NaN 3 , NaSCN, etc.However, organic-free phosphide NPs could be obtained by displacing organic ligands with HPF 6 (figure 9).The detailed ligand displacement procedure and TEM images of Co 2 P NPs before and after ligand displacement can be found in the supporting information (figure S10).

POSSIBLE REACTION MECHANISM
We observed that the shape of Co 2 P NPs significantly changed with the argon flow rate (figure 10).Quasi-spherical Co 2 P NPs (figure 10a) were obtained when using strong argon flows during the reaction, but rod-shaped NPs (figure 10b) were prepared at low (about 30 mL min -1 ) flows of inert gas.
The key to understand this shape variation with the argon flow might be in the in situ formation of ammonium salts, which would be fully expelled from the reaction mixture only at high argon flows.Such ammonium salts are well-known shapegoverning agents.For example, Pd 2 Sn nanorods were prepared using HDA•HCl. 70Iron oxide octahedrons were synthesized with the help of trioctylammonium bromide. 71In the same direction, we hypothesize that we produced Co 2 P nanorods or quasi-spherical NPs depending on the presence or absence of HDA•HCl during the synthesis, which was related with the capacity of the used argon flow to displace this in situ formed compound.This shape control mechanism does not apply to the case of Fe 2 P nanorods, where a carbonyl instead of a chloride was used as metal precursor.Actually, Fe 2 P nanorods were obtained regardless of the argon flow rate.
The particular mechanism of in situ formation of HCl in the reaction mixture was revealed in previous studies demonstrating the reaction of TPOP with nickel 72 and copper 73 chlorides.TPOP reduced the chlorides during the complex formation and resulted in the formation of HCl.Such formed HCl, would instantly react with present amines to produce the corresponding ammonium salts.In our case, complexes might be formed during the initial degassing/cleaning step (argon bubbling at 150 ºC) and the overall reaction might be represented by the following scheme: 2MeCl x + 5x(RO) 3 P + 2xHDA+ xH 2 O 2Me[P(OR) 3 ] 2x + 2xHDA•HCl + x(RO) 3 P=O TPOP plays a double role here, as a ligand in the formed metal complexes and as reducing agent, which underlines the importance of adding this reagent in excess.The exact compositions of the formed metal complexes depend on the nature of the metal and on the particular conditions of the synthesis (e.g.amount of ODE, HDA, TPOP, temperature, etc.) and cannot be revealed at the present moment.The complexes of interest coexist with excesses of TPOP and HDA, which most proba-bly participate in the reaction as ligands, and cannot be easily separated and studied.Details of how HDA•HCl influences the shape of phosphide NPs would strongly depend on the surface chemistry in every particular case and it is going to be matter of our future work.

NEGATIVE RESULTS
Even though phosphide NPs of Ni, Co, Cu, Fe and Mo could be produced using TPOP as phosphorous source, phosphides of other tested metals, such as In, Sn, Mn, Pd and W, could not be produced in the same conditions.In some cases, e.g.In and Sn, no reaction was detected at all.The color of the solution did not change even during prolonged heat treatment and no product could be precipitated from the reaction solution.Similarly, only an unidentified, white-yellow flaky product was obtained from the reaction of TPOP with MnCl 2 under the same conditions (see XRD pattern in the supporting information).The replacement of chlorides by other metal precursors was not successful.In this direction, for Mn and Sn, MnCl 2 , Mn(acac) 2 , Mn 2 (CO) 10 , SnCl 2 , SnCl 4 •5H 2 O, Sn(acac) 2 , and SnC 2 O 4 were tested without success.On the other hand, metallic W and Pd NPs were obtained using WCl 6 and Pd(acac) 2 as metal sources.

CONCLUSION
In summary, we detailed a straightforward, cost-effective and up-scalable procedure to produce phosphide NPs.The keys to this synthesis protocol were the use of cheap, stable, easy-to-handle precursors in a 'heating up' synthetic process.Particularly, TPOP was used as an inexpensive and air-stable source of phosphorous, and mainly metal chlorides, as a similarly convenient source of metals.Stable and low-toxicity precursors not only significantly simplified the synthetic procedure but also rendered unnecessary the use of glove-box as well as many other usual precautions.Additional simplification of the synthetic procedure was achieved through the introduction of the 'heating up' methodology -the simplest and easiest to scale up, compared to the classical injection-based protocols.Both amorphous and crystalline NPs could be prepared through careful reaction temperature/time adjustment.Additionally the shape and size of the NPs could be controlled for particular materials by modifying the protocols for each individual compound.Moreover, we demonstrated the scalability of the detailed protocol and described a procedure to displace the organic ligands from the surface of the as-synthesized materials.With these possibilities at hand it will be possible to further explore the many interesting properties and applications of TMP.

ASSOCIATED CONTENT
Supporting Information.Additional HRTEM micrograph and the EELS chemical composition maps obtained from the STEM micrograph for the phospspides of the nickel, iron, cobalt and copper; TEM images and size distribution histograms (thickness and length) of the Co 2 P nanoparticles obtained from a tenfold scaled synthesis; Safety and toxicity issues of phosphorous precursors; Ligand exchange procedure; TEM images of Co 2 P nanoparticles before and after ligand exchange; XRD pattern of the unidentified material prepared with MoCl 5 ; XRD patterns of the Ni 2 P NPs, prepared during 5 and 60 minutes; TEM image, EDX-SEM spectrum and XRD pattern of the unidentified material prepared with WCl 6 .

Figure 1 .
Figure 1.TEM micrograph (a); HRTEM micrograph and corresponding power spectrum (b); ADF-STEM and EELS compositional maps (c); size distribution histogram (d); and XRD pattern of Ni 2 P NPs (e).The reference JCPDS pattern of Ni 2 P (00-003-0953) was also included in the XRD graph.The lattice fringe distances from the HRTEM micrograph were measured as 0.185 nm, 0.187 nm and 0.187 nm, at 59.41 and 120.39º which is consistent with the hexagonal Ni 2 P phase, visualized along its [0001] zone axis.

Figure 2 .
Figure 2. TEM micrograph (a); HRTEM micrograph and corresponding power spectrum (b); ADF-STEM and EELS compositional maps (c); nanorod length and thickness distribution histograms, (d); and XRD pattern of Fe 2 P NPs (e).The reference JCPDS pattern of Fe 2 P (01-076-089) was also included in the XRD graph.The lattice fringe distances from the HRTEM micrograph were measured as 0.218 nm, 0.168 nm and 0.214 nm, at 50º and 100º, respectıvely, which is consistent with the hexagonal Fe 2 P phase, visualized along its[11-23] zone axis.

Figure 3 .
Figure 3. TEM micrograph (a); HRTEM micrograph and corresponding power spectrum (b); ADF-STEM and EELS compositional maps (c); nanorod length and thickness distribution histograms (d); and XRD pattern of Co 2 P NPs (e).The reference JCPDS pattern of Co 2 P (00-054-0413) was also included in the XRD graph.From the HRTEM micrograph, the Co 2 P lattice fringe distances were measured to be 0.218 nm, 0.172 nm, 0.218 nm and 0.220 nm, at 50.78º, 101.44º, and 78.71º which is consistent with the hexagonal Co 2 P phases, visualized along its [1-100] zone axis.

Figure 4 .
Figure 4. TEM micrograph (a); HRTEM micrograph and corresponding power spectrum (b); ADF-STEM and EELS compositional maps (c); disk diameter and thickness distribution histograms (d); and XRD pattern of Cu 3 P NPs (e).The reference JCPDS pattern of Cu 3 P (01-071-2261) was also included in the XRD graph.From the HRTEM micrograph, the Cu 3 P lattice fringe distances were measured to be 0.198 nm, 0.197 nm and 0.197 nm, at 60.39º, 119.72º which is consistent with the hexagonal phase, visualized along its [0001] zone axis.

Figure 5 .
Figure 5. a) TEM micrograph of as-produced MoP NPs.b) TEM micrograph of MoP NPs after thermal treatment at 800 ºC for 2 h.c) HRTEM micrograph and corresponding power spectrum of annealed MoP NPs.d) ADF-STEM and EELS compositional maps of annealed NPs.e) XRD pattern of annealed MoP NPs.The reference JCPDS pattern of MoP (00-024-0771) was also included in the XRD graph.

Figure 6 .
Figure 6.a), b) Representative TEM micrographs of Ni 12 P 5 NPs prepared using a Me:HDA:TPOP ratio 1:5:10 and by quenching the reaction immediately before the solvent boiling point was reached.c) Size distribution histogram.d) Temperature profiles (after solvent cleaning) of the synthesis procedure at high, i.e. just below the boiling point, (red) and low, up to 250 ºC, (blue) temperature.e) XRD patterns of the materials prepared at high (red) and low (blue) temperatures.

Figure 7 .
Figure 7. TEM micrographs (a-c) and XRD patterns (d) showing the dependence of the Ni 2 P NPs with the concentration of TPOP.e) Size evolutions extracted from TEM images and XRD patterns.Numbers indicate Ni:HDA:TPOP ratios during the synthesis.All scale bars correspond to 100 nm.

Figure 8
Figure8shows the Co 2 P and Fe 2 P NPs produced with different concentrations of precursors in the synthetic mixture.While the rod-like NP geometry was preserved, the monodispersity and overall NP quality significantly decreased when increasing the Me/HDA ratio.Reducing the amount of HDA, Co 2 P and Fe 2 P NPs lost the accurate cylindrical shape, becoming more irregular and with less defined edges.Generally, a Me:HDA:TPOP = 1:10:10 ratio resulted in highest quality NPs.However, more detailed and systematic experiments need to be done to fully understand how the NPs' morphology depends on the experimental conditions and to fully optimize the synthetic protocol, especially when scaling up and maximizing the reactant concentrations.

Figure 8
Figure 8 Representative TEM micrographs of Co 2 P and Fe 2 P NPs obtained from different Me:HDA:TPOP ratios, as specified in each micrograph.All scale bars are 100 nm

Figure 9 .
Figure 9. FTIR spectra of TPOP, HDA, as-produced Ni 2 P and Co 2 P (below) and Ni 2 P and Co 2 P after ligand removal (above).

Figure 10 .
Figure 10.TEM images of the Co 2 P NPs prepared with high (left) and low (right) argon flows.All scale bars are 100 nm.