Introduction

Overview of the Combinatorial Approach
The essential steps of library creation, screening and replication of results. Idea is to define what it is that will be surveyed and to make some differentiation between lab-on-a-chip, high throughput screening, robotics and full combinatorial methodology. Introduce relationship with miniaturization, MEMS, informatics.

Historical background in Drug Discovery (and even earlier in Monoclonals)
Focus on inapplicability of biological methods of library creation and screening to materials development.

Types of Materials for which Combinatorial Methods have been Proposed:
• catalysts • polymers • magnetic materials • phosphors • colorants • adsorbents

Who will benefit from Better Methods of Materials Selection?
Traditional relationship between catalyst companies, technology owners, engineering contractors and operating companies - Development of chemical and refinery production technology classically involves co-operation among three entities, a manufacturing company, a catalyst company and and engineering and construction contractor, (E&CC). Once the technology is developed, (and usually demonstrated by the manufacturer in the group), it is the job of the E&CC to sell it to others. Once sold, again, classically, there are three money pools, the engineering profits and license fees associated with design of the unit, the ongoing catalyst sales and an annual technology access fee, or royalty. How these are assessed, negotiated and divided varies widely. Generally, however, the catalyst company is free to sell catalyst to the unit and realize the profit therefrom. Thus, their profits are deferred in time, but relatively insulated from “boom-and-bust” construction cycles.

Role of companies dedicated to combinatorial materials discovery - Of course, this role is not yet defined, but one would expect them to fit naturally into the role of the catalyst producer. Who will make the catalysts developed by these methods? The lesson of Catalytica, a company founded to exploit an earlier generation of new methods for catalyst discovery, is instructive here. In short, they had to reinvent themselves as a specialty chemical manufacturer and basically abandon their catalyst ambitions. One would also expect the catalyst companies to adopt combinatorial methodology for their own research.

Is the analogy with early biotech companies, similarly focussed on technology and not end markets, the right one in view of the different industries served?

Commercial Background
Structure of the Catalyst Business - Heterogeneous catalysts comprise a ca. $7.5 Billion annual worldwide market. (Higher figures include homogeneous catalysts and even some related co-marketed materials like polymer crosslinkers. No one, to my knowledge has publicly proposed a combinatorial method for homogeneous catalysis.) There are three segments to this heterogeneous catalyst market: automotive, refinery and commodity chemical and specialty chemical/pharmaceutical. Automotive catalysts are controlled by the specifications of the automakers. Refinery and petrochemical catalysts, which are the most interesting candidates for combinatorial screening, are intimately tied to the processes used (introduces the idea of a Process); a choice generally locked in at the time of plant construction. These markets are served by a handful of multinational horizontally integrated catalyst firms (UOP, Grace, Engelhard, Synetix, AKZO, Criterion) and a number of smaller vertically integrated firms.

Structure of Commodity Chemical and Polymer manufacturing industry - Commodity chemicals and polymers as low growth, (especially in the developed world), tight margin materials. Typically the cost of catalyst comprises one or two percent of the cost of commodity manufacture. Thus the incentive for new catalyst development lies in improving the economics of the process in which it is used, for example by increasing yield and/or throughput. This latter idea is particularly powerful in an industry where margins are eroding.

Quick overview of Specialty Chemical Industry - Unlike commodities, specialty chemicals, (which critically includes pharmaceutical intermediates), are still highly profitable. Again unlike commodities, many specialty chemicals are “speced in” to their uses and relatively insulated from competition. Thus the incentives for new catalyst applications lie, again unlike commodities, in new products and processes. Advantages sought from these new catalysts include, novelty of reaction, ease of use, product purity, (including chiral purity), and increasingly, easier environmental compliance. Relationship between high-tech material suppliers and OEMs - Most high tech materials, like polymers, and automotive catalysts are sold by long term contract to OEMs.

Opportunities to introduce new materials arise only as new contracts open, either at the expiration of an old contract or for a new product. Selling into this market requires understanding of (and ideally integration into), the product development cycle of the customer, the OEM.

Issues in Implementation of Combinatorial Methods to Materials
Relationship between library structure and screening methodology

Simultaneous vs. sequential analysis - Any sequential system is by definition slower than a simultaneous method. Simultaneous methods, however, require multiple detectors and cost and physical size constraints restrict them to a small number of simultaneous determinations. This drives the science to fast sequential methods.

Analytical response times - In any sequential method, analytical response time is the critical parameter. Only optical signals approach instantaneous readout.

Relationship between method of library generation and replication - Actually less of a problem in materials than in bilogical systems. A consequence of rationally vs. randomly assembled libraries. (See discussion below under Methods of Library Creation.)

Special Technical Considerations in Catalysis
Catalyst performance not a single variable function - Catalyst performance cannot be characterized by a single parameter, like “binding vs. non-binding. Since the catalyst may catalyze both desired and undesired reactions, one wants to maximize the desired rate and minimize all undesired rates. One usually characterizes these two issues in terms of activity, (which is a total of all rates of disappearance of reactant), and selectivity which is the ratio of desired product to the theoretical amount of desired product corresponding to the disappearance of reactant. Note that determining selectivity requires a chemical analysis. Not all organic analytical techniques are easily miniaturized nor are they all “real time.” As discussed above, optical signals are preferred, but their availability is system specific.

Significance of support in heterogeneous catalysis - Most heterogeneous catalysts are “supported”, i.e. the “active material” is deposited on an “inert” carrier, typically a refractory oxide. In almost all cases, the support is not truly inert, but affects the performance of the catalyst. At a minimum, therefore, multiple supports must also be screened. Deposition methodology also is a factor.

Selection based on performance in use requires severe, and undefined, screening conditions - Most commercial heterogeneous catalysts operate at elevated temperature; many also operate at elevated pressure. Since these conditions are not exactly known a priori, one has to screen all candidates at a series of conditions. It is also an endemic problem of catalyst research that making direct observations of catalysts at operating conditions challenges research creativity. The great white hope of miniaturizing catalyst research - a trend which has been going on independent of combinatorial development - is to solve, at least in part, this problem.

Reproducibility of results more art than science - Beside reaction conditions, catalyst performance also depends on the history of the catalyst. Many catalysts require a break-in period, sometimes at conditions different from design conditions, to stabilize at their long-term performance. Some catalysts decline rapidly with time, but can be regenerated in situ. Others decline more slowly, but cannot be regenerated; for these, catalyst life and aging becomes an important selection criteria. All of these effects mean that screening at a variety of conditions is necessary. In the limit, the library becomes secondary to the screening of operating variables.

Library Creation for Materials
Technology overview - including patent summary and literature review

Physical configurations - Because of the existence of catalyst- support interactions described above, screening multiple samples on the same support, analogous to a 96 well plate, is of limited use, unless a variety of chemically different supports is available. At the other end of this spectrum, making individual “real” catalyst particles reflecting all the variables of support choice and deposition technology moves the burden from the screening step to the robotic generation of the library. Methods of library generation - No non-biological analog to the truly random methods of library generation available in biological systems exists. Libraries must be generated by algorithms and are always therefore inherently rational. One is less likely to encounter surprising results searching a rationally assembled library than a truly random one.

Screening
Technology overview - including patent summary and literature review

Analytical methods - general

Analytical methods - catalysis

thermal - Thermal methods are particularly good because they are fast and easy. They give no information, however, on selectivity. For adsorbents and combustion catalysts, this doesn’t matter. For chemical catalysts, thermal methods may provide a first level screen to select which samples to analyze in detail.

chemical - mass spec, REMPI, etc. - Mass spec, while accurate and universally applicable is inherently slow and can’t work at actual reaction conditions. Optical methods, like REMPI, may be limited to certain sets of chemistry and require extensive analytical development.

other - Main point to make here is that chromatography, every chemists favorite analytical method, has problems being applied to combinatorial systems.

Selection techniques specific to certain classes of materials

i.e. spectrophotometric, (very effective for colorants), magnetic (described for magnetic materials)

Related Required Technologies
Each of these could be subject of its own multi-client. Idea is to give enough information for reader to follow the combi story.
• Robotics
• Microsystem design and fabrication/MEMS/Microfluidics
• Informatics
• Self Assembly