Monday 13 March 2017

Guest post by Weimer: Early History of Particle ALD at the University of Colorado

Please find below the first guest post ever written for the ALD History Blog: Prof. Alan Weimer from the University of Colorado continues on the topic of a previous related post, sheading light to the early particle ALD developments at the University of Colorado and ALD NanoSolutions.   

--- guest post begins ---

Early History of Particle ALD at the University of Colorado

In 1997, Steve George and I started investigating the ALD on particles – Steve investigated the surface chemistry through in-situ FTIR and I investigated the processing to coat large volumes of fine particles using a fluidized bed reactor.   At that time, Steve was an ALD chemistry expert and I was expert in the synthesis and functionalization of fine particles, having worked in the area for nearly 20 years.  Much of my background was fluidized processing.  Very intense research that year indicated that we could use ALD to coat large volumes of primary particles identically without “gluing them” to neighboring particles.   HRTEM was key as well as surface area, particle size distribution, FTIR, and XPS measurements; and of course agitated particles.   I was quite surprised that individual primary particles could be coated as I knew that CVD inherently deposited nanoparticles formed in the gas phase, hence, forming agglomerated particles.

We spent some time to formalize our patent application to the U.S. Patent Office and identified the coating with certain constraints – non-agglomerated particles with the average particle size not increasing by more than 5% as a result of the coating - apart from the particle size increases attributable to the coating itself, or, if no more than 2 wt% of the particles become agglomerated during the process of  depositing the inorganic material; conformality was defined as the thickest region on a particle being no more than 3X the thickness of the thinnest region within the limits of HRTEM at that time.  Also, we recognized that ALD films could be continuous or semi-continuous; semi-continuous essentially being not uniformly coated over the particle – including leaving exposed surface over some of the particle.  I then attempted to obtain funding from government agencies for this research over the next year or so.   Proposals were rejected as the independent panels of experts (all of them) organized by the federal government said that it was impossible to coat primary particles like we claimed, i.e. that we would agglomerate these particles.  An original small seed grant from an industrial consortium with university connections funded initial work, but, in the end none of the consortia companies supported patent filing costs and CU ended up with exclusive IP rights.  ALD NanoSolutions, Inc. was co-founded in 2001 as a way to potentially obtain industrial research sub-contract funding for ALD on particles since the gov’t agencies would not fund my purely academic research at that time.  At the same time, the U.S. Patent Office was rejecting the patent application saying it was obvious to one skilled in the art – if one could coat a flat article by ALD, then someone expert in particles could coat primary particles by ALD.  Eventually, the combination of a key article showing CVD inherently resulting in agglomerated particles due to surface deposition of nanoparticles and the rejected proposal reviews formed the basis for three U.S. Patents eventually being issued – U.S. 6,613,383; U.S. 6,713,177; and U.S. 6,913,827 as well as “sibling patents”  EP 1412175B1; JP 4507598B; and CA 2452531C.   Finally, I did receive a small discovery grant from NSF in 2002 for ALD on particles.   More grants followed in later years. 

I also vividly remember making an invited presentation at the 2006 meeting of the International Fine Particle Research Institute (IFPRI) “Functionalization of Ultrafine Particles by ALD,” presented June 26, 2006, in Santa Barbara, CA.   I was met with major skepticism (to say the least, I was blasted) by who at that time and who still today is considered one of the foremost experts in the world, probably THE foremost expert, saying that “it was impossible to coat primary particles, particularly nanoparticles, without agglomerating them”!   His reasoning was strong inter-particle forces, even when fluidized.  I don’t disclose that person’s name here as I do not have permission to do so, but they reside in Europe.  A number of years later I heard this person give a talk that was positive on ALD for particle coating.  I asked him about the IFPRI confrontation and he told me “it took a while to sink in”, but he is today a strong proponent of ALD to coat primary particles.  I have the utmost respect for this individual and his contributions to particle science.

Our group near that time had published two papers which demonstrated (my opinion) that primary nanoparticles could be uniformly coated by ALD using a fluidized bed reactor [Chemical Vapor  Deposition, 11,420-425 (2005); Nanotechnology, 16,  S375-385 (2005); a key consideration was the concept of dynamic aggregation in fluidized beds - Powder Technology, 160 (3), 149-160 (2005)].  Getting back to the semi-continuous (non-uniform) aspect of the ALD coatings on particles, we showed in Fluidization XI - Present and Future of Fluidization Engineering, ECI International, 603-610 (2004) that 10 TMA/H2O cycles on 5 micron iron particles behaved almost the same as no coating at all when these particles were oxidized in air using a TGA at 265 deg C.  It was not until 25 cycles that the particles were oxidation resistant (most likely due to a continuous film coating), and 25, 50, and 100 cycles gave almost identical results.   Likewise, reported ICP-AES results indicated 380 mg Al/kg Fe for 5 cycles, 500 mg Al/kg Fe for 10 cycles and then 1700 mg Al/kg Fe for 25 cycles.   So, broadly speaking a substantial amount of the iron particle surface was left exposed and not coated with alumina, maybe crudely 380/1700 or 22% surface coverage for 5 cycles and 500/1700, or 29% surface coverage for 10 cycles.  This aspect is nicely described in Puurunen’s famous alumina review article Journal of Applied Physics, 97, 121301 (2005), “The ALD process modifies the chemical composition of the surface through materials deposition.   The first ALD reaction cycle occurs on the surface of the original substrate material, the following cycles are usually on a surface with both the original substrate and the ALD-grown material exposed, and after several ALD reaction cycles – the exact number depending on the GPC growth and the growth mode – finally on a surface with only the ALD grown material exposed.  If the chemical composition of the surface changes, the GPC could be expected to vary with the number of cycles.”  Clearly, the particle properties are much different for when the substrate particle surface is still exposed to when the particle is completely encapsulated (after a number of ALD cycles).

Although the main focus early on for ALD on particles was for complete pinhole-free films, typically achieved with multiple ALD cycles, that was not totally the research direction.  The original application for ALD on particles at CU was both Fe oxidation resistance; and, BN surface modification to improve coupling of particles to resins so that the loading of high thermal conductivity (TC) BN particles into resins could be increased.   A clear issue was that an alumina ALD film of only 2 nm would decrease TC by 50% [Powder Technology, 142 (1), 59-69 (2004)], hence, the research direction was for nearly no film covering or semi-continuous non-uniform partial surface coverage in order to maintain high TC while improving coupling with the resin matrix.  So, CU research was directed at both compete film coverage (i.e. Fe particles for oxidation resistance – requiring as many as 25 ALD cycles, but probably less) and semi-continuous films where the surface of BN particles was still partially exposed or films were as thin as possible to achieve the desired resin adhesion since a 2 nm alumina film theoretically reduced TC by 50%.

I want to also say that, indeed, based on recently translated papers, it seems that early work was done for ALD on particles in the 1960s.   However, analytical characterization at that time could not allow one to claim primary particle (non-agglomerated) coating for all particles at the scale required.  A lot of excellent work was done in the 1990s in Finland for ALD on particles, primarily catalysis depositing active metal sites on particle surfaces (1 cycle typically), but also some papers with multiple cycles.  I have reviewed at least 20 of these excellent papers, but, I have not found (my opinion) any conclusive findings that all individual primary particles were coated with a film without agglomerating the particles.   For example, in Appl. Surf. Sci., 121/122, 286 (1997) - “samples were taken at three different radial positions and two different axial positions …, totally six samples, gave an average of 1.0 ± 0.1 wt% chromium” and “…five reaction cycles were carried out on silica and alumina…however, due to the formation of agglomerate the surface coverage was incomplete…Characterization by LEIS showed that only 50% of the silica surface was covered by the modifying oxide” (i.e. ZrO2).  This is not indicative of virtually all primary particles coated by ALD without agglomerating them.


The patents cited above have been licensed by several of the largest companies in the world for select applications after carrying out technical due diligence and right to practice studies.   I have seen various statements on other corporate web sites that Particle ALD originated in the 1960s and later that a lot of work was done for Particle ALD in Finland in the 1990s.  I don’t dispute that.  I agree.  However, I have not seen prior credible evidence (my opinion) to the work at the University of Colorado which shows that indeed virtually all primary particles are coated (i.e. not agglomerated) by ALD as described in the issued patents.   And, as I have described, skepticism that this could actually be done continued through 2006.

March 13, 2017,
Alan (Al) W. Weimer, H.T. Sears Memorial Professor
Chemical and Biological Engineering
University of Colorado
Boulder, CO USA
alan.weimer@ colorado.edu (remove space)

--- guest post ends ---

*********************
Virtual Project on the History of ALD (VPHA) - in atmosphere of Openness, Respect, and Trust

No comments:

Post a Comment