Nanocrystal Inks without Ligands: Stable Colloids of Bare Germanium Nanocrystals
Entry by Yuhang Jin, AP225 Fall 2011
Zachary C. Holman and Uwe R. Kortshagen, Nano Lett., 2011, 11, 2133.
nanocrystal, quantum dot, colloidal stability, germanium, benzonitrile
Colloidal semiconductor nanocrystals (NCs) have received increasing attention recently for their potential in novel optical and electronic devices, and in particular inexpensive printing of quantum-confined thin films. Methods have been developed to synthesize II-VI and IV-VI NCs in solution with ligands, which provide steric stabilization in nonpolar solvents so that aggregation of NCs due to van derWaals forces does not occur. However, the ligands also hinder charge carrier transport between NCs, resulting in insulating films. Exchanging the ligands for shorter molecules or removing the ligands after film deposition seem to be alternative approaches, but they both have intrinsic difficulties due to the formation of covalent bonds or other adverse properties of the materials involved. To avoid such problems, this paper reports stable Ge NC colloids formed by dispersing ligand-free NCs in select solvents. This approach offers new opportunities for NC processing and device fabrication in which semiconducting films may be printed from solution without postprocessing.
Experimental details and results
Germanium NCs were synthesized with a nonthermal, continuous-flow plasma process: germanium tetrachloride vapor and hydrogen gas were dissociated in the plasma, allowing nanometer-sized Ge crystallites to nucleate via chemical clustering of the dissociation products. The diameter of the NCs was kept at 6 nm by controlling the duration of their residence in the plasma. The chemistry at the surface of the NCs is a mixture of H and Cl passivation. The NCs studied here were shown by ion beam measurements to have three times as much H on their surfaces as Cl.
Germanium NCs were collected as a powder downstream of the plasma reactor, and transferred air-free to vials. Solubility of Ge NCs was investigated in a host of solvents whose properties were similar to 1,2-dichlorobenzene in dielectric constant, chemical structure, or Hansen Solubility Parameters. Many of the solvents contained benzene rings, and halogens or nitrile groups. Common solvents such as toluene, hexanes, and water were also tried. Ultrasonication was performed to break up the powder and suspend the NCs after solvents were added to the vials. Of all the solvents studied, benzonitrile, and to a lesser extent acetonitrile, solubilized bare Ge NCs upon contact: a brief shake alone could yield an optically transparent colloid (shown in Fig. 1).
The stability of Ge NCs was studied by comparing the initial NC concentration to the concentration of NCs that remained suspended after filtering the dispersions through a 0.2 μm PTFE filter. The concentrations were determined from UV-vis absorption measurements. The absorption spectra of Ge NCs in benzonitrile were identical before and after filtration, indicating small (or no) aggregates. By contrast, scattering from agglomerates dominated the spectra of most suspensions before filtration, whereas many samples looked like pristine solvent after filtration. For example, NCs in 2-chloropropane, heptyl cyanide, 1,2-dibromoethane, bromoethane, dichloromethane, pyridine, 1,3-dichlorobenzene, chloroform, as well as all very nonpolar solvents tried, were unstable and quickly formed aggregates larger than the filter pore size.
The absorption of unfiltered Ge NCs in benzonitrile (with a concentration of 0.25 mg/mL) was used as a reference to determine the unknown concentrations of Ge NCs remaining in other solvents after filtration, as shown in Fig. 2a. The normalized concentration was calculated using absorbance data at 525 nm. Fig. 2b displays the relative concentrations after filtration for all solvents with nonzero postfiltration concentrations. In Fig. 2c, the postfiltration concentrations are plotted against the static dielectric constant of the solvents. Solvents that stabilize the NC dispersions tend to fall within a broad range of dielectric constants, excluding both very polar and very nonpolar solvents. However, it is worth noting that dielectric constant alone does not afford predictive power of solubility.
The data in Fig. 2 represent a metric for colloidal stability. Filtration helps find solvents that have the fewest agglomerates larger than a few tens of nanometers. Films were cast from unfiltered Ge NC colloids in several solvents to verify the quality of this metric. For solutions with no Ge NCs remaining after filtration, NC aggregates micrometers in size with areas of bare substrate were observed (Fig. 3a). However, NCs in benzonitrile produce uniform films a few nanometers to a few micrometers in thickness on glass and Si wafer chips (Fig. 3b,c), with no large agglomerates visible, confirming their absence in solution.
As for the mechanism for the stabilization, it is not evident that the dominant stabilizing forces are the same for each solvent, and the paper focuses on only benzonitrile here. With the exception of pyridine and water, NCs were not observed to undergo structural changes in any solvents: X-ray diffraction spectra show that 6 nm NCs in synthesized NC powder and NC films are indistinguishable. Also, no evidence of reaction between NCs and solvents was observed: RBS and Fourier transform infrared (FTIR) spectroscopy revealed similar NC chemistry before and after colloid formation in benzonitrile, and thermogravimetric analysis (TGA) showed that residual solvent in the film evaporates at or below the boiling point of benzonitrile, as expected for unreacted solvent. Therefore, solubility or stability depends entirely on the state of the NC surfaces.
Experiments were conducted in which Ge NC surface chemistry was controllably altered by heating to 300 C, dipping in diluted hydrofluoric (HF) acid, and oxidizing in ambient air. Fig. 4 shows the FTIR and RBS spectra for the surface chemistry of these NCs. The NC powder had a combination of H, H2, and H3 termination in addition to Cl. Annealing at 300 C removes both H and Cl species, dipping in HF leads to a purely monohydride surface with no Cl, and exposure to air gives rise to a significant oxide shell with residual Hx termination and trace amounts of Cl. Of these samples, only the as-synthesized NC powder was soluble in benzonitrile, whereas the others yielded NC concentrations of zero after filtering. This indicates that the Cl on the as-synthesized NCs plays an important role in their solubilization in benzonitrile, and that H- and Cl-free surfaces, purely H-terminated surfaces, and oxidized surfaces do not provide the colloidal force(s) necessary for stable suspension. The absence of steric forces in these ligand- and surfactant-free colloids or species present to participate in hydrogen bonding suggests that electronegative Cl surface species induce electrostatic stabilization. This is confirmed by a measured -25 mV zeta potential. The most likely routes to NC surface charging due to electrical double layer formation are Lewis acid-base interactions with the Cl surface groups acting as the acid, or dissociation of Hx groups. The degree to which a solvent can support electrostatic forces depends on its polarity, suggesting that the trend in Fig. 2c may be attributable to a single stabilization mechanism of electrical double-layer formation.
In conclusion, the paper has demonstrated the formation of stable colloids of bare Ge NCs with H- and Cl-terminated surfaces using select solvents. By moving from steric stabilization to electrostatic stabilization, NC colloids that are stable for months without the use of ligands can be cast into films with no prohibition of charge carrier transport, allowing semiconductor NC devices to be constructed from as-cast films.