AmericanChemicalSociety.com

Nanoparticles exhibit many interesting new chemical, physical, electronic, and mechanical properties and are being pursued for a number of applications.  For many of these applications, such as catalysis, the assembly or distribution of the nanoparticles may be as important as the particle size and composition.  We are carrying out fundamental studies into a novel method for growing nanoparticles on surfaces in which the size and average spacing of the nanoparticles can be controlled.  The process utilizes chemical modification of the substrate surface combined with atomic layer deposition (ALD).  ALD is a vapor phase deposition process consisting of an alternating series of self-limiting chemical reactions between gas phase precursors and the substrate.   ALD can be used to grow a large range of catalytically important materials. 

In order to achieve controlled nanoparticle growth by ALD, we have focused our efforts on using self assembled monolayers (SAM) formed from octadecyltrichlorosilane (ODTS) to modify silicon oxide-terminated silicon substrates.  We have shown in earlier work that a very well-packed SAM can resist atomic layer deposition completely, whereas less well-packed SAMs allow some ALD to occur.  Over the award period, we have investigated the effect of intentionally forming defective SAMs to allow for nucleation of nanoparticles by ALD.  Specifically, the self assembled monolayers were imparted with increasing defect concentrations by intentionally shortening the SAM formation time.  These defects, in turn, served as potential nucleation sites for ALD.  

In these studies, a series of ODTS SAMs were deposited on the silicon substrates over formation times ranging from 2 hours to 12 hours.  The SAMs were characterized by a combination of ellipsometry and water contact angle measurements.  The quality of the SAM was particularly sensitive to the surface hydrophobicity, with values of 110º suggesting the formation of a close-packed SAM and lower values indicative of a more loosely packed film.

Following SAM formation, the substrates were placed into a custom built ALD reactor to perform Pt deposition.  Pt ALD was achieved using (methylcylopentadienyl)-trimethylplatinum (IV) and oxygen precursors, delivered in separate pulses, with the substrate at approximately 300^oC.  Both precursors were pulsed into the reactor for 2 seconds and separated by 30 second purges to minimize the possibility of CVD.  Depositions were carried out for varying number of binary ALD cycles. After Pt ALD, the substrates were characterized by x-ray photoelectron spectroscopy, scanning electron microscopy, and scanning Auger electron spectroscopy.

The results show that Pt nanoparticles can be successfully deposited on the SAM-modified planar substrates.  The XPS data show that the Pt concentration at the surface after ALD decreases with increasing formation time of the surface-modifying SAM over the range of all immersion times studied.    Moreover, the SEM image results suggest a trend in which shorter SAM formation times (more defects) lead to a higher density of Pt nanoparticles, while longer SAM formation time (fewer defects) leads to a lower density of nanoparticles. However, in the case of a great number of ALD cycles, e.g. 200 cycles, the aforementioned trend breaks down.  We conclude that for high numbers of Pt ALD cycles, Pt growth occurs over the SAMs and results in almost continuous film growth regardless of immersion time.

The effect of ALD cycle number was also investigated.  For a fixed SAM formation time, the XPS data show an increase in Pt atomic percent at the surface with increasing cycle number.  SEM data show that while the Pt still grows as nanoparticles, the average size of those nanoparticles increases with cycle number.  Average particle size distribution decreased with increasing SAM immersion time.   In general, nanoparticle size ranged from less than 4 nm to greater than 40 nm.

The data collected was used to model the process of nucleation in ALD.  Using a framework of isothermal growth and nucleation as well as geometric considerations, the data was fit to a model that takes into account incubation time preceding nucleation and subsequent nucleation time constants.  The model suggests that nucleation incubation time increases with increasing SAM immersion time.

The method of using SAMs to direct the growth of nanoparticles by ALD is not limited to Pt.  In this study, both amorphous (deposited at 100 ºC) and anatase TiO₂ (deposited at 250 ºC) were also grown as nanoparticles.  We conclude that ODTS SAMs can direct the growth of nanoparticles by ALD, affording control over density and to some extent, size.  The more prohibitive a substrate was to growth—that is the greater the SAM immersion time—the lower the density of nanoparticles.  As Pt ALD cycle number increased, the average particle size increased, although the correlation was not strong.  Besides Pt, TiO₂ both in amorphous and anatase phases can be grown as nanoparticles using this method.   The data was used to gain preliminary insight into Pt ALD nucleation.

Finally, several applications of nanoparticle ALD and selective ALD were demonstrated.  Pt nanoparticles deposited into carbon aerogels using ALD were tested as oxidation catalysts; ODTS SAMs were used to direct film deposition by ALD for electrodes in solid oxide fuel cells; and, a new type of monolayer deposited on silica surfaces was shown to prevent Pt ALD.