Mechanistic insight into the formation of colloidal WS2 nanoflakes in hot alkylamine media

Developing convenient and reliable synthetic methodologies for solution processable 2D layered ultrathin nanostructures with lateral size control is one of the major challenges for practical applications. In this study, a rational understanding a long-chain amphiphilic surfactant assisted non-hydrolytic synthesis that is able to generate dimension-controllable 2D-WS2 nanocrystal flakes in a single-step protocol is proposed. The evolution of the starting soft organic–inorganic lamellar template into ultrathin few-layer 2D-WS2 nanostructures with lateral size modulation over a range between 3 and 30 nm is monitored. The initial formation of WS2 nanoseeds occurs in a self-assembled sacrificial precursor source, acting as a template, where larger two-dimensional nanostructures can grow without undergoing significant thickness variation. Overall, the chemical nature and steric hindrance of the alkylamines are essential to modulate the reactivity of such WS2 nanoclusters, which correlate with the lateral size of the resulting nanoflakes.


I. Complementary XRD simulations:
Simulations of the X-ray diffraction pattern, based on the Debye equation, show how the typical saw-tooth profile is only recovered when the number of unit cells in the crystalline domain is reduced along one axis: the a axis in the case of the 1T' phase and the c axis in the case of the 2H phase.
Therefore, in order to damp the first (highest) peak and broaden the other ones, so as to match the experimental profile, simulations with no more than 2 unit cells along the a axis (in the case of the 1T' phase) or the c axis (in the case of the 2H phase) are considered in the following.
Once suppressed the first peak, the next representative range to be considered is between 30° and 60° (2).It is then readily recognized, based on simulations comparing the 1T' and 2H phases in Fig. S2, that the peaks around 42.5° and 47.5° (2) are characteristic of the 1T' phase, whose presence is therefore established.Moreover, such peaks are quite weak and broad, resulting from the convolution of several reflections, while another characteristic peak of the 1T' phase at 54.6° is even missing; so that an even smaller domain in the (b,c) plane can be expected.

S3
On the other hand, a close inspection (Fig. S3) of the regions around the two main peaks at 32° and 57°(2), reveals a much better match between calculation and experiment for the 2H phase (Fig. S3c,d).Moreover, in the zoomed regions around the two peaks, it is clearly seen that a number of 1T' unit cells larger than 10 along the b and c axes would lead to a splitting of the first peak (Fig. S3a); whereas a number of 2H unit cells larger than 20 along the b and c axes would lead to much sharper and intense peaks (Fig. S3c,d).

S4
As a natural consequence, the experimental X-ray scattering profile can be expected to result from a superposition of the profiles relevant to 1T' and 2H two-dimensional nanocrystals with suitable weights and crystalline domain size (Fig. S4).However, no combinations of two solutions (for any size and weight fraction of the two phases) are sufficient to reproduce the experimental profile, meaning that polydispersion has to be taken into account, with smaller domains mainly contributing to the diffuse scattering in the range 32°-40°, and larger domains leading to a sharpening of the two main peaks (Fig S4).Combinations of three or more profiles are then considered; the results for the best overall agreement (Fig. S5) suggest that the main scattering contributions come from 1T' NCs with 1x10x10 unit cells and 2H NCs with 40x40x1 unit cells, being the weight of any additional 1T' contribution required to be negligible.

Figure S5
Simulated XRD profiles resulting from the combination of 1T' and 2H two dimensional nanocrystals, leading to the best agreement with experiment.The relative abundance of each nanocrystal considered in the population is reported in the legend as a multiplication factor for each term contributing to the calculated scattering profile.

II. Complementary X-Ray spectra:
To support the structural evolution described by the wide angle XRD patterns reported in Fig. 4a-e, we present wide and small angle XRD patterns in Fig. S6a,b, and Fig. S6c,d S6c, bottom spectra) and is reduced to 3.92 nm for WCl (6-x-y) :ODE:OlAm x :OctAm y (bottom spectra in Fig. S6d), thus compatible with a two or more inorganic layers spaced apart by a double organic layer arranged with a tilting angle around 50° and 33° respectively, considering the interlayer distance and carbon chain length (1,2) which compose the hybrid organic/inorganic lamellar phase.Upon injection of CS 2 precursor and heating at 130 °C, with the simultaneous formation of oleyl-thiourea and the in situ releasing of H 2 S, a characteristic interlayer distance d of 3.4 nm is detected for both systems, which can be approximately estimated as the sum of the 0.7 nm thick newly formed WS 2 nanoclusters, the two extra C-N (0.133 nm) (3) and N-S (0.171 nm) (3) bonds per OlAm molecule introduced in the structure as consequence of thiourea formation, and a double layer of OlAm constituent of formed thiourea ( 2 nm ) (4) with a tilt angle of 30°.The same periodicity of 3.4 nm has been found in the final products, where larger particles (in the stacking direction) can be expected, based on the sharpening of corresponding peaks in the wide angle range.The recovered periodicity of 3.4 nm is ascribable to a super-structure of surface capped planar nanocrystals arranged in turbostratic organization (Fig. S6c,d, top spectra).S7

IV. Complementary NMR spectra:
In order to prove that the chemical path reported in scheme 1 can be extended also for nanocrystal synthetized in OlAm/OctAm, we have reported here the reaction products involved with an experimental reference model formulated by using two equivalents of OctAm or one equivalent of OctAm and one of OlAm with one equivalent of CS 2 .No significant differences were detected in the positions and shape of the signals of α-CH 2 N protons of OctAm or OlAm/OctAm derivatives, which is concentration-affected as dictated by slight shifts in peak position (Fig. S8).In the mixture WCl 6 /CS 2 /OlAm (0.25:4:8.5) aged at 130 °C, the formation of oleylammonium species, referred as compound 2 (Scheme 1 in main text), was clearly observed, originating the low frequency shifted CH 2 -N centered at 3.04 ppm and a very broad high frequency shifted signal at 8.70 ppm (Fig. S10b).The formation of compound 2 was ascertained on the basis of the integration ratio between the two above said signals (2 to 3) and the absence of any new quaternary carbon in the 13 C NMR spectrum.The spectral pattern of such a species reproduced the one of the oleylamine hydrochloride salt (OlAm•HCl, see Fig. S10a), suggesting that WCl 6 is responsible for the formation of HCl.Ammonium salt is influenced by the presence of the metal, (5) since the relaxation times of ammonium species in nanoparticles systems are lower (T 1 = 0.31 s for CH 2 -N) if compared to pure OlAm•HCl salt (0.55 s).It is noteworthy that the amount of ammonium salt S9 formed in the final nanoparticle product is correlated to the amount of WCl 6 , as demonstrated by comparing 1 H NMR spectra of mixtures at differrent WCl 6 to amine ratios (Fig. S10b-c).

Figure S 1
Figure S 1 Simulated XRD patterns calculated for crystalline domains with different dimensions along the a axis and along the c axis for the 1T' (pattern on the left) and 2H (pattern on the right) crystal phases.The number of unit cells making the crystals along the a, b and c axes, respectively, are reported in the legend.

Figure S2
Figure S2 Simulated XRD patterns calculated for crystalline domains with different dimensions along the a and c axes for the 1T' and 2H crystal phases, showing how the two marked peaks at 42,5° and 47,5° are mainly related to the 1T' crystal phase.

Figure S3
Figure S3 Simutaled XRD patterns showing how peak profiles in the different angular regions are affected by crystalline domain size of 1T' and 2H phases.

Figure S4
Figure S4 Simulated XRD profiles for different combinations of 1T' and 2H two-dimensional nanocrystals, with suitable weights and crystalline domain size.

Figure S6
Figure S6All measurements are performed on purified sample by following the experimental procedure, then dissolved in anhydrous CHCl 3 and casted on Si-substrate in N 2 atmosphere.(a) Wide Angle (20°<2θ<80°) XRD (crystal structure) evolution recorded on time/temperature scheduled aliquots derived from synthesis of 2D-WS 2 nanoplatelets generated in only OlAm environment, withdrawn at the specific reaction condition reported in the relative legend; (b) Wide Angle (20°<2θ<80°) XRD (crystal structure) evolution recorded on time/temperature scheduled aliquots derived from synthesis of 2D-WS 2 nanoplatelets generated in OlAm/OctAm (1:2 mmol ratio) environment, withdrawn at the specific reaction condition reported in the relative legend; (c) Small Angle (0.1<q = (4/)sinθ <1.2 Å -1 ) XRD (nanoscale structure) evolution reporting inter-planar distances on x-axis, recorded on time/temperature scheduled aliquots derived from synthesis of 2D-WS 2 nanoplatelets generated in only OlAm environment, withdrawn at the specific reaction condition reported in the relative legend; (d) Small Angle (0.1<q = (4/)sinθ <1.2 Å -1 ) XRD (nanoscale structure) evolution reporting inter-planar distances on x-axis, recorded on time/temperature scheduled aliquots derived from synthesis of 2D-WS 2 , nanoplatelets generated in OlAm/OctAm (1:2 mmol ratio) environment, withdrawn at the specific reaction condition reported in the relative legend.In (c) and (d), the periodicities correlated to the small angle equally spaced diffraction peaks are reported in the legends of the individual plots with relevant colour code.In caption are reported the experimental condition where each sample has been exposed (a-d) and the corresponding periodicities of equally spaced (in reciprocal space units) peaks (c,d).

Table of Contents
respectively, in order to describe a more detailed structural evolution of recorded intermediate aliquots.Aliquots recorded at 130 °C and 160 °C for both OlAm and OlAm/OctAm mixture show featureless XRD patterns, ascribed to amorphous material or composed of too small crystalline domains.Considering the aliquots recorded at the 215 °C, the scenario completely changes because when OlAm/OctAm mixture is used (black plot), a XRD pattern clearly ascribable to the WS 2 nanosheets formation is detected, whereas an amorphous-like pattern is still characterizing the sample generated in sole OlAm (red plot) that eventually evolves in a pattern characteristic of heterophased WS 2 nanomaterials only heating up to 250 °C.Small-angle XRD patterns related to isolated OlAm x :OctAm y :CS 2 (Fig.S6c,d) give precious information on time/temperature evolution of the lamellar layered mesostructures.The d-spacing (periodicity) experimentally observed, reported on top right of every individual frame, prior to the CS 2 injection, is 4.25 nm for WCl (6-x) :ODE:OlAm x (Fig.