From Molecules to Materials: Materials Discussion 7

---

Abstract

Paul O'Brien and Alice Sullivan describe the highlights of the discussion during the 7th Materials Discussion Meeting: From Molecules to Materials, held at Queen Mary College, University of London, 13th–15th September 2004.

---

The meeting was opened by a short presentation from Professor Don Bradley FRS; Don spent the majority of his academic career at Queen Mary College, and its successor institutions, of the University of London, on the Mile End Campus.¹ This meeting was organised to illustrate the synergy between his pioneering work on alkoxides and an enormous contemporary interest in such compounds as precursors for a wide range of materials, many of which are used in advanced electronic applications. He recounted the early isolation of zirconium alkoxides and observations on oxide thin film deposition over 40 years ago. Such methods were not properly exploited for many years and are even now being aggressively developed especially for dielectrics.

Plate1 Don Bradley

There was then a seamless transition to the first keynote lecture in which Professor Liliane Hubert-Pfalzgraf presented an overview of how the design of molecular precursors can control the preparation of advanced oxides. In particular the interplay of oxoalkoxide and alkoxide chemistry as anticipated in the early work of Bradley was discussed. In recent years the use of functional alkoxides has been important in evolving both the chemistry of these systems and in controlling deposition. Delivery methods have also had a profound influence on the design of MOCVD precursors. Strategies for the minimization of carbon incorporation by the use of reactive alkoxides prone to specific elimination reactions were discussed. Remaining challenges include developing a better understanding of the influence of precursors on the morphology of deposited material.

The first discussion session started with a description of some solvent-free reactions of aryl alcohols with lanthanides to make novel monomeric complexes supported by π-interactions; a series of adducts with ethers such as THF were reported by Glen Deacon (Monash Melbourne). Marc Henry (Université Louis Pasteur) talked about molecular recognition using titanium(IV) (oxo)alkoxides and (oxo)aryloxides. In sol–gel processing with titanium isopropoxide the (presumed) basic tetrameric core structure can lead to specific interactions; a key factor is the maintenance of bridging groups. Precursor modification and control of reactivity enables species to be well defined in solution for the first time.

Ulrich Schubert (Vienna) described a new type of molecular precursor which could be cross-linked into polymers of the type Zr₆O₄(OH)₄(carboxylate)₁₂. These molecules can lead to novel inorganic organic hybrid structures. In the initial phases hydrolytic and non-hydrolytic steps are important.

There followed a lively discussion concerning the reactivity of alkoxides in solution. Kessler (SLU Sweden) pointed out that the structure of dimeric titianium isopropoxide is known and that the distribution of charges shows that bridging ligands are more charged. There was a debate as to what is important in, for example, a sol–gel process with Henry maintaining that the first step in hydrolysis is important in controlling the sol–gel process and that there is a close link between this step and the final material produced.

John Errington commented that what we are really trying to understand is the kinetic landscape in solution. In discussing cleavage of a specific bond, e.g. at bridging positions, what are we really debating in relation to materials deposition? Henry holds that the proof comes from NMR of solutions. Errington asked if we would ever be able to compute what is exactly happening in this complex system – Henry thought that maybe we would.

Tim Leedham (Epichem) asked where does oxide come from when oxo alkoxides are formed in the absence of water and what is the driving force for the formation of oxide? Are ethers formed? Liliane Hubert-Pfzalgraf pointed out that excess alcohol is often present and zirconium isopropoxide can form oxo species; Bradley had shown the ether elimination pathway for scandium and yttrium isopropoxides and there are examples of alkene elimination from zirconium alkoxides.

Walter Klemperer (University of Illinois) commented that in Don Bradley's talk he mentioned that in the zirconium isopropoxide dimer it is the terminal alkoxide that is protonated not the bridging one, and why would this be the case if the bridge was more basic? Malcolm Chisholm (Ohio State University) commented that the difference is in the interpretation of what we say. There is a thermodynamic stability and also the kinetic stability and the terminal groups may be more reactive. In general at polyoxoanions there are examples where the terminal positions are more basic and examples where bridging positions are more basic. There is the general problem of comparing crystal structures with what goes on in solution.

Sanjay Mathur (Saarland University) opened the second session and reported on metal alkoxides as molecular scaffolds in nanocrystalline oxide synthesis but especially emphasized work on complex oxides such as Y/Al oxide alkoxide with the correct ratio of elements. The initial nanopowder can influence the nature of the final product. With careful control of hydrolysis, clean molecule-to-material conversions can be possible as in the YAlO₃ metastable phase prepared in this work.

Vadim Kessler (SLU) discussed the influence of heteroligands on the composition, structure and properties of homo- and hetero-metallic zirconium alkoxides. These complexes are very labile so how do we control reactivity? We can use Goldschmidt's rules to follow how solids might form. Using ligand volume angle rather than atomic radii can we develop rules for formation of alkoxide complexes? Indeed, Don Bradley proposed a similar idea in the 1950s.

An interesting discussion relating the stability of complexes and solid materials formed from them followed.

Matthew Hill (UNSW Sydney) reported work on dialkylcarbamato magnesium clusters as precursors to high quality MgO thin films. High molecular weight in the system leads to low intermolecular interactions and hence a more volatile precursor. Very thin and very pure films were produced with some carbon on the surface but no nitrogen incorporation.

Tony Jones (University of Liverpool and Epichem) discussed some recent developments in the MOCVD and ALD of high-κ dielectric oxides. The work is driven by the need to replace SiO₂ gate dielectric in CMOS technologies with alternative materials with high κ. This is due to the breakdown of SiO₂ at the dimensions now being approached (1–2 nm). He gave a particularly clear account of the differences in reactivity needed for the two deposition regimes. The need for stabilized monolayers in ALD methods which react when, for example, water is added to oxide layers presents a major challenge, one which was to be discussed at various times during the meeting.

Tim Leedham asked if any natural products have been used as ligands. Liliane Hubert-Pfzalgraf replied that sugars had been used. There are reports of relevant compounds but these may not have been used in materials deposition.

David Cole Hamilton asked if alkoxides stabilized to β-elimination had been used in ALD, by analogy with the use of the neopentyl group, for example, in alkyl chemistry; Tony Jones concurred that this might be a good idea.

Peter J. Wright (QinetiQ Ltd, Malvern Technology Centre) opened an interesting section of the discussion by asking about the influence of surface kinetics on crystal growth as opposed to the nature of the precursor controlling the material deposited. This comment led to lively debate about the importance of precursor design.

Sanjay Mathur pointed out that mixing Fe(II) and Fe(III) alkoxide precursors was generally ineffective in preparing Fe₃O₄ giving rise to the discrete iron(II) and iron(III) oxides. Kessler challenged the need to produce discrete molecular precursors in labile systems and cited recent work in which rare earth and aluminium alkoxides were mixed to give a solution that generated the ternary oxide. He also emphasized the importance of understanding phase diagrams in these systems.

David Cole-Hamilton made the important point that the substrate is often dictated by a device requirement and there is considerable scope in tailoring precursors for given applications. This was a point with which Gladfelter and Jones concurred, citing examples from GaN chemistry and the influence of hafnium precursors on the growth of DRAM stacks for which the chloride was an unsuitable precursor because of its low surface mobility.

The effect of the surface catalysis is important at the end of the day and can depend on what the film needs to be grown on. Gladfelter pointed out that there is room for precursor design in determining the material formed, e.g. in the GaN system.

There was discussion as to whether or not rules could be derived for the formation of materials from molecular precursors.

The second day of the meeting started with a keynote presentation by John Irven (of Air Products and Queen Mary) which provided background on the commercial importance of precursors and speciality gases. These chemicals form an approximately $3 billion market supporting part of a $200 billon market in devices. The industry road-map suggests that materials properties will need to accommodate a projected increase in transistors per chip from the current Pentium 4 HT 125 000 000 transistors to ca. 1 billion by 2016. Many thanks to John for providing the cover image for this issue of the Journal.

Wayne Gladfelter (University of Minnesota, Minneapolis) described a combinatorial CVD approach to ZrO₂/SiO₂ and HfO₂/SiO₂ compositional spreads as part of a search for high-κ dielectrics using the anhydrous nitrates and silanols. Depositions were very rapid: several thousand angstroms were grown in a few minutes at temperatures between 130 and 300 °C.

Jean Marie Nedelec (Université Blaise Pascal, France) reported on the molecular design of inorganic scintillators from alkoxide precursors. The scintillators were based on lanthanide oxides prepared from the alkoxides by sol–gel methods and powder processing or spin coating and are used in high energy physics. The oxidation state of cerium in the material was crucial in determining properties.

Stephane Parola (Universite Claude Bernard, Lyon) reported on routes to spinel materials from M^IIAl₂(acac)₃(OⁱPr)₄(OAc), M^II = Mn, Co, Zn, a new class of heterometallic heteroleptic alkoxide complexes.

In the subsequent discussion it emerged that the products of the Gladfelter deposition reactions included isobutene and t-butyl alcohol. Chemically the system is interesting as it shows the interesting stability of the anhydrous nitrates first isolated by Addison.

The question as to where the electronics industry road-map might go for the silicon industry was debated. Would there be a move from inorganic to hybrids? The electronics industry is keen on its 10 year plans but beyond 12 years we don't know about the hold-up of Moore's law; will a disruptive technology emerge?

Peter Wright pointed out that the electronics industry is very conservative and uses only three basic processes – deposition, etching and lithography – to produce devices. Looking ahead on road-maps there is a shrinking window of time to develop technology that can introduce the planned performance enhancements.

Markku Leskelä (University of Helsinki) asked how would Gladfelter's combinatorial routes be put into use? Gladfelter said that reactor design to process a specific phase would be crucial. To deposit hafnia/silica films, short residence time reactors are needed to avoid the formation of powders.

Glen Deacon queried if the advantages of lutetium-based scintillators outweigh the cost of the element. It was argued that even if the cost were high it might be possible because the expense of the scanners is so great.

Simon D. Elliott (NMRC University College Cork) asked about tri(tert-butoxy)silanol depositing on top of alumina and Roy Gordon's recent work² and the impossibility of reaching surface self-limited growth.

Chuck Winter (Wayne State, Detroit) asked John Irven about the current technology for copper deposition and how it might develop further. The PVD ‘damascene’ process is at present used and he noted a need to get through road blocks. High-κ materials are a major problem. Irven thought we would go to molecular level processing.

Glen Deacon mentioned that reverse CVD processing could be a way forward. Such a process would involve the evolution of etchng from wet to vapour and requires improving vapour methods. How do we make this more efficient? Could this be the way ahead for smaller features in copper?

George Christou (Florida) asked Irven what the future holds? He believed that the future must be molecular. Will the stage arise when molecules are the material and not the precursors? There is a great potential for chemists to make contributions. Curiosity-led research is needed to drive into these areas. John Irven discussed how things might move forward. The inevitable Catch 22 situation is in trying to get money to do adventurous research. There followed a somewhat philosophical debate on exploitation of technology and develeopment of electronics and the electronics industry.

The next session was concerned with quantitative modelling and Jean-Pierre Jolivet (CNRS Paris) reported on oxide nanoparticles by precipitation from aqueous solution. A thermodynamic model based on the interfacial properties of the system was presented.

Simon D. Elliott (NMRC) discussed simulating the atomic layer deposition of alumina. The work addressed how trimethylaluminium might react with the surface in order to give the blocked surface required for ALD.

There was a discussion of the relative reactivity of hydroxyl- and alkyl-groups at the surface. David Cole-Hamilton asked Elliott about the cleavage of Al–C followed by O–H or if there was a concerted raction.

Wayne Gladfelter was also interested to know if the calculations informed on the first, second and third methyl group eliminations. Chisholm commented that the higher the coordination number of aluminium the more difficult it is to transfer the proton of a hydroxide.

Peter J. Wright (QinetiQ Ltd, Malvern Technology Centre) discussed the role of precursors in the CVD of oxide materials, mainly perovskites for thermal imaging. The importance of matching the properties of the various precursors and newer delivery methods for less volatile compounds was emphasized.

Neil Downie (Air Products) thought that liquid injection delivery will beome more common. He asked what are the solvents and what are the problems with the solvent? High boiling point solvent is needed so the solvent does not evaporate as it enters. Things like tetraglyme are sometimes added, meaning that reaction mixtures used as precursors are complicated.

Leedham asked Wright if particulates are a problem in CVD of these materials. Wright responded that such material can be formed in the vapour phase and that a vertically aligned substrate minimizes the fall of ‘snow’ onto the device, which is not a major problem.

The Tuesday afternoon session was concerned with ‘hot topics’ in the area and Malcolm Chisholm (Ohio State University) proved to be a stimulating and enthusiastic chair. The afternoon opened with a keynote address by Walter G. Klemperer (University of Illinois) in which attempts to develop the systematic methods for polymerizing inorganic monomers was described. He pointed out that one of the problems in materials chemistry is to make materials better than can be made by engineers by the careful control of the operating parameters. He hence chose a problem that was difficult to approach using engineering methods: the deposition of continuous amorphous pin-hole-free thin films <100 Å of zirconium oxides on YBCO. These layers could act as viable tunnel barriers below the YBCO superconducting phase transition temperature. The importance of using ultrapure alkoxide precursors was emphasized.

S. Deki (Kobe University Japan) described the preparation of nanodimensional metal oxide thin films and nano-ceramics from aqueous solution. The approach used metal fluorides and a scavenger for fluoride such as H₃BO₃. By using a silicon wafer template, well defined arrays of columns or holes in TiO₂ could be prepared. Ivan Parkin (UCL London) asked if they had looked at what happens when light is shone onto the lotus leaf-like structures of TiO₂ and whether or not water spread on the film? Parkin also asked if any photonic effects were seen given that the spacing between the nano holes is of the order of the wavelength of visible light. It was confirmed that the structures with the rods produced photonic effects.

R. Bhakta (Bochum, Germany) described the MOCVD of TiO₂ thin films and studies on the nature of molecular mechanisms involved in the decomposition of [Ti(OPrⁱ)₂(tbaoac)₂]. Gladfelter thought it was the first time that ketenes had been seen in β-diketonate decomposition. Cole-Hamilton commented that in the mechanism for decomposition to the ketene that the OBu^t group of the tbaoac ligand was attatched to the metal atom but the crystallography shows that the carbonyl oxygen is attatched. There was subsequent discussion of dissociation and rotation reactions to explain this point.

Will Rees (Georgia Institute of Technology, USA) discussed the design and synthesis of heterometallic aminoalkoxides and their application in the MOCVD of zirconium-tin-titanate (ZTT). The work aimed to produce a process to materials with a dielectric constant between 20 and 30. Liquid injection was used with a mixed tin/titanium precursor and zirconium t-butoxide. Cole-Hamilton asked if there was any distillation of the precursors? Rees reported that a mixture of random samples showed that there is no appreciable change. Rees also reported the lead compound Pb₄(NSiMe₃)₄ and there was a discussion about congruent and incongruent vaporization.

Al Steigman (Florida) reported on sol–gel-derived xerogels derived from alkoxides as catalysts and more specifically as a vehicle for studying catalytic reactions by spectroscopic methods (readers should note that his paper will appear in a later issue of this journal). The cocondensation of vanadium into silica xerogels led to isolated vanadium centres which showed catalytic activity.

The final day of the meeting started with a keynote address by Malcom Chisholm (Ohio State), a distinguished graduate of the Bradley group at Queen Mary College from the 1960s. He started by discussing some of the energy and feedstock chemical problems which will occur as fossil fuel feedstocks become depleted during the present century. His stimulating lecture described copolymerization of epoxides and carbon dioxide as catalysed by well defined alkoxide catalysts.

Dominic Wright (Cambridge) discussed the intriguing idea that reductive-elimination of phosphide units could provide a general approach to several materials, in particular Zintl phases. Mike Lappert (Sussex) described the use of the 2,2,6,6-tetramethylpiperidinato (TMP) ligand in preparing lanthanide alkylamides. Several of these compounds are volatile and may be useful as precursors. He also reported on the interesting formation of a mixed ethoxide/TMP cerium complex by cleavage of the C–O bond in diethyl ether.

Markku Leskelä (University of Helsinki) described the use of novel bismuth precursors in the deposition of bismuth oxides. As discussed earlier in the meeting, ALD places unusual demands, almost contradictory demands, on precursors. Stability is needed to form the initial adsorbed layers but these layers must be chemically reactive when the feed is changed.

Will Rees asked if rigorously oxygen-free precursors could be made by the triflate route used in Lappert's work. Lappert did not think that this contamination would be a problem and Liliane Hubert-Pfalzgraf concurred. Rees, however, thought this could be a problem in materials application.

Wayne Gladfelter asked Leskelä about the difference in ratio of strontium in the as-deposited material as compared to the precursor delivery stoichiometry. This result is surprising but consistent according to the presenter.

Chuck Winter asked Mike Lappert why not use the di-t-butyl amide? Lappert thought there was no reason not to pursue this line.

Dave Hoffman (University of Houston, Texas) asked if there was a possibility of ring opening of THF to give the vinyl alkoxide complex rather than decomposition of diethyl ether to give the ethoxide complex, since THF is easier to fragment than diethyl ether. Lappert indicated that he was confident that he had prepared an ethoxide from diethyl ether and not the vinyl alkoxide from THF. He also added that the ether cleavage (which is being further investigated) appears to be accelerated by the presence of at least a trace of the alkali metal (Li) precursor.

Alice Sullivan asked Malcolm Chisholm about the advantages of the methods and why particular polymers might be used. Malcolm Chisholm responded that detailed studies of such polymers were in their early days and that although much stereochemical control is possible in these systems it is at present poorly understood.

The final session of the discussion meeting consisted of a closely linked series of papers concerning the preparation of nitrides and was chaired by Andrew Hector.

Jayaprakash Khanderi (Bochum, Germany) described some ligand-stabilized dialkyl aluminium amides as precursor for AlN films. In the presence of ammonia, polycrystalline nitride films were formed but with carbonitride contamination. Andrew Hector asked about the residue in TGA experiments which was identified as carbon-containing material.

Charles (Chuck) Winter (Wayne State, Detroit) described the preparation of molybdenum and tungsten nitride nanoparticles from molecular precursors. The goal is to develop spin-coat technology for metal nitrides. Thermolysis of bis-imido-bis-pyrazolato compounds of tungsten and molybdenum gave W₂N and Mo₂N nanoparticles of about 3 nm embedded in an amorphous carbonaceous matrix, at 800 °C. In the case of molybdenum you also get some MoO₃. At lower temperature, THF-soluble, amorphous, nitrogen-rich nanoparticles are formed.

David Hoffman (University of Houston) talked about a new precursor for the CVD of tantalum nitride films, another paper describing work driven by the the need for better methodologies for the electronic industry. In particular, Ta(IV) complexes such as Ta(NEt₂)₂(NCy₂)₂ have been developed for the aerosol-assisted CVD of TaN_x. The materials are effective in blocking copper diffusion to a silicon surface at temperatures below 500 °C and are thought to be the first example of the simple CVD of TaN_x from a Ta(IV) precursor. It is striking that the as-prepared material retains the average oxidation state of (IV) for Ta.

The final discussion paper of the meeting by David Cole-Hamilton described the decomposition of [InN₃(CH₂CH₂CH₂NMe₂)₂] to give nanoparticles of InN. Crystalline InN was formed in TOPO which luminesced.

Wayne Gladfelter asked Winter if he had tried the reactions in an inert atmosphere since the particles may be coated with carbon and if the coat on the particle would be a problem or an advantage. Chuck replied that he thought if good material could be made and spun onto a surface this might lead to useful materials as barrier layers. David Hoffman commented that the majority of useful materials to date had been amorphous.

Will Rees asked Hoffman if hydrogen in the film was leading to films which were not fully dense. Hoffman reported that the experimental density was ca. 66% of bulk, but that the films were quite stable in air.

John Errington asked Cole-Hamilton if he had tried high boiling point amines as decomposition media. D.C.-H. had tried triglyme but had not tried amines. There was a discussion of In₂O₃ and InN; Cole-Hamilton was confident that the characterization of the In in his samples ruled out the oxide.

Malcolm Chisholm asked about the use of solid state NMR to characterize such systems. Charles Winter replied this was a good idea and David Cole-Hamilton said that it had proved a useful method for studying ligands bound to the surface.

During the meeting, we remembered Professor Ram C. Mehrotra, who passed away in July 2004, for his contribution to the chemistry of metal alkoxides and in the context of his long association with Don Bradley.

In addition to the papers presented in this volume the meeting was further supported by over 40 posters. These presentations were all of a high standard and three were singled out as prize winners, an overall winner and two runners up. The overall winner was Alina Vinslava of the University of Florida and her poster concerned giant Mn₈₄ and Mn₇₀ single molecule magnets. The runners up were D. Bavot of Louvain, who reported on peroxo-tartrates of niobium and tantalum as precursors for mixed oxides, and R. A. Fischer from Bochum, who reported on the deposition of TaN films from mixed imido/hydrazido complexes. All of the poster contributors deserve thanks for their efforts and it was particularly useful that the posters were displayed for the whole meeting in the tea venue.

The meeting seemed to be enjoyed by all the participants and resulted in genuine debate between crystal growers and chemists who synthesise precursors. There were well over 100 delegates representing five of the seven modern continents; substantial attendance from both the industrial and academic sectors was particularly pleasing to the organisers.

The organisers would like to thank the sponsors of the meeting: Air Products, Elsevier, Epichem Oxides and Nitrides, RSC Publishing and Queen Mary, University of London. Penny Mohamed (RSC Conferences) and Rebecca Lavender (RSC Publishing) provided essential and cheerful support that made the meeting such a memorable success.

We hope that this report will give a flavour of the meeting to those who did not attend and will serve as an aide memoir to those who did. P. O'B. took notes of the discussions and takes responsibility for any omissions – clearly this summary has had to be somewhat selective.

Plate2 Paul O’Brien

Plate3 Alice C. Sullivan

Paul O'Brien
Chairman Materials Chemistry Forum
The Manchester Materials Science Centre and Department of Chemistry
The University of Manchester
Oxford Rd.
Manchester
UK M13 9PL

and

Alice Sullivan
The Department of Chemistry
Queen Mary University of London
Mile End Rd.
London
UK E1 4NS

References

1. P. O'Brien, Coord. Chem. Rev., 2000, 1–13.
2. D. Hausmann, J. Becker, S. L. Wand and R. G. Gordon, Science, 2002, 298, 402–406.

---