Solid catalysts are used in the manufacturing of a vast array of chemicals and fuels, and as such significantly contribute to our economy and high living standards.^[1] In addition, catalysts provide important environmental benefits, such as in catalytic converters for automobiles. However, in spite of their significance and broad utility, the discovery of new catalysts continues to be an arduous and rather unpredictable trial-and-error process. Catalysts traditionally are developed using a large variety of tedious, time-consuming and often one at a time methods, characterized and tested for activity, modified, characterized and tested again ... until no further improvements are justified. In addition, these investigations have often been conducted independently by different groups, rendering the comparison of activity information difficult. This approach, although time consuming, has been successful for the discovery of a significant number of solid-state catalysts over a time period of many decades.^[2] However, there remains a large unexplored universe of binary, ternary, quaternary and higher-order solid-state materials and other complex metal compounds that can have superior catalytic properties. Conventional approaches clearly are inadequate to rapidly synthesize and then screen this vast universe of catalytic materials; therefore, the development of new approaches that will accelerate the process of discovery, optimization and understanding of heterogeneous catalysts will have a significant impact in catalysis science and engineering.

Combinatorial chemistry, in which a large diversity of chemical compounds are rapidly produced and processed, and the libraries generated are then screened for desirable properties to identify leads in a high throughput fashion is a particularly attractive approach for the discovery and optimization of heterogeneous catalysts. Combinatorial strategies have initially been used to synthesize very large libraries of biological oligomers such as peptides and nucleotides. However, the application of combinatorial techniques solid-state materials has begun. At present two primary methods exist for the preparation of solid state libraries: 1. Thin film deposition based methods of synthesis;^[3-6] and 2. Solution based methods of synthesis of combinatorial libraries.^[7,8] These library preparation techniques were then used for the discovery of superconducting ^[3], magnetoresistive ^[4] and luminescent ^[5,6,7,9] materials, and electrochemical catalysts. ^[8]

The discovery of heterogeneous catalysts has been a challenging task because unlike superconductivity, magneto-resistivity, and electrochemical reactions, which can be tested by contact probes, or non-specific luminescence/fluorescence, the screening of heterogeneous catalysts requires the unambiguous detection of a specific molecule (i.e., a product) in the vicinity of small catalyst sites. To date three methods have been proposed to screen catalyst libraries: 1. In-situ optical screening using infrared (IR) thermography ^[10,11] or fluorescence acid-base indicator,^[8] 2. Microprobe sampling and scanning mass spectrometry (MS),^[12] and 3. In-situ resonance enhanced multiphoton ionization (REMPI) spectroscopy with microelectrode detection.^[13] The optical methods, although they provide high speed data on activity, generally provide limited information on selectivity or different reaction products. The microprobe/MS involves the sampling and transfer of very small quantities of gases containing product species that are also at low concentrations. In addition, the sampling probe must physically be moved from site to site to screen the entire library. The REMPI technique combines the speed of optical methods and the selectivity of mass spectrometry and allows the high speed screening of large solid state catalyst libraries for activity and selectivity without the need to remove samples.

In this communication, we report on the application of REMPI together with novel array microreactors for the discovery of an optimal composition in the ternary Pt-Pd-In metal catalyst system for the dehydrogenation of cyclohexane to benzene. Array microreactors allow the screening of a large number of catalysts in parallel and under identical (standard) operating conditions. Consequently, measurements become primarily sensitive to the method of preparation of the catalysts. This represents a significant advance over the current practice in catalyst research, in which different catalysts are often studied under different operating conditions, rendering comparisons difficult.

A sketch of the array microreactor system developed is shown in Figure 1. The reactor array consisted of 17 rectangular channels that were micromachined on a flat non-porous silica ceramic slab (7.5 cm x 3.75 cm x 0.63 cm). Each channel was 0.1 cm wide, 0.1cm deep and 2 cm long, and possessed a cylindrical well (0.3 cm diameter by 0.2 cm deep) to hold cylindrical catalyst pellets (0.3 cm diameter and 0.1 cm high) as shown in Figure 1. In this configuration, the flat surface of the catalyst pellet, when placed into the well, would be flush with the bottom of the reactor channel. A similar flat ceramic slab was used to cover the microreactor array slab and to form isolated channels. This design results in unobstructed gas flow rates through the channels with only wall skin friction providing flow resistance. Consequently, the establishment of nearly identical flow rates in each channel, and therefore identical contact times between the reactant gases and the catalyst pellets in each microreactor were feasible. The flat and rigid nature of the ceramic slabs also result in the formation of a good gas seal, thereby isolating each channel from one another under a broad range of temperatures. Ceramic slabs were also machined for optical access and the placement of dedicated microelectrodes as shown in Figure 1. The reactor array was cradled between two aluminum metal heating blocks. In addition, a separate heating block was used to preheat the feed gases to the desired reaction temperature. The temperature of each heating block was independently regulated using electrical heating elements embedded into the aluminum body and PID controllers (Omega, Stamford CT). Gas temperatures in the vicinity of the catalyst sites were separately monitored by small thermocouples inserted from the reactor exit to assess their uniformity across the microreactor library, and to establish the time periods for heat up and cool down. These studies revealed the presence of a nearly uniform temperature, within 1^oC, across the entire catalyst library. In addition, the reactor array was shown to heat up and/or cool down to a desired temperature within 30 minutes of the set point change.

The reactant gases were introduced into the microreactor array through a quartz feed tube and a series of flow distribution baffles that were also micromachined to the bottom ceramic slab (see Figure 1). The baffles were used to establish uniform gas flow rates through each channel. The feed stream composition was about 10% cyclohexane in helium carrier gas; this mixture was prepared by bubbling helium gas (1000 cc/min) through liquid cyclohexane in a sparger at about 25^oC. These conditions correspond to a nominal contact time of about 0.04 s to cross the flat surface of the 0.3 cm diameter catalyst pellet in each microreactor. The uniformity of gas flow rates in each channel was checked by passing smoke through the reactor array and by measuring the lengths of the smoke plume jets emanated from each channel. These studies indicated that gas flow rates were approximately the same in each channel.

Alumina pellets added into each test tube were allowed to impregnate under the competition of HCl in solution for 24 hr in order to attain homogeneous metal distribution within the particles. Subsequently, the solutions were evaporated and the pellets were dried first at 90^oC for 4 hr then at 120^oC for 4 hr. Finally, the impregnated pellet library was calcined at 500^oC for 2 hr and cooled down to ambient temperature in the furnace. The catalyst library consisted of 66 ternary combinations of Pt, Pd and In in 0.1% wt increments of each with 1 wt % total metal loading on alumina support. In addition, the final chlorine content of the pellets was also kept at about 1%. The compositions of the 66 pellets prepared are shown in Figure 2. It should be noted that all the catalyst pellets were prepared and processed in an identical manner without any human intervention.

Catalyst testing was accomplished using the following procedure: After the placement of new catalyst pellets into the wells of the array microreactor, the system was heated to 350^oC under helium flow. A blank alumina pellet was always placed into the central microreactor (9^th well from either side) in all experiments to serve as an internal reference standard. After reaching 350^oC, the helium flow was switched to pure hydrogen for 2 hrs to reduce the Pt, Pd and In chlorides into metallic form. Following the catalyst reduction step, the reactor temperature was decreased to 300^oC again under helium gas flow. Once the proper operating temperature was reached, cyclohexane feed was introduced. With the existing array microreactor system, this procedure had to be repeated 5 times to screen the entire 66 combination library. Additional experiments were also undertaken to check the reproducibility of the measurements, including the placement of catalyst pellets at different locations in the array microreactor and duplicate preparation and testing of different combinations. All these studies resulted in similar results within the limits of accuracy of the measurements, i.e. +15%.

It should be noted that all the catalysts were subjected to an identical pretreatment (i.e. reducing) process before screening in order to standardize the initial conditions. This way the activities of the catalysts can be compared to one another having been on-stream the same time period, thereby minimizing complications related to catalyst deactivation, poisoning etc. In addition, the visual inspection of the pellets immediately after the reduction steps also revealed a uniform catalyst metal distribution across the pellet.

The screening protocol demands the unambiguous detection of benzene in a cyclohexane, hydrogen and helium background, as well as other reaction byproducts. In order to accomplish this goal using the REMPI technique one must identify a suitable UV laser wavelength that selectively produces benzene REMPI ions. One such wavelength was identified to be 259.6 nm in our previous studies using time of flight mass spectrometry (TOF-MS).^[13] At this wavelength, the first photon energizes the benzene molecule to an excited electronic state, P*, and the second photon ionizes this intermediate state as shown below:

The pulsed UV laser light was passed through the optical access windows of the heating block and the ceramic slab, simultaneously crossing the exhaust streams of all the microreactors. The laser beam also passed between the dedicated microelectrodes that were embedded into the top and bottom ceramic slabs as shown in Figure 1. Microelectrodes were made from 0.1 mm diameter steel wires and their tips were flush with the ceramic slab surfaces. This design significantly simplified the process of alignment and calibration of the microelectrodes and the laser beam between catalyst changes, thus improved the speed and accuracy of the catalyst screening process. Multichannel switches were used to individually apply 100 V DC power to each microelectrode sequentially (Stanford Research Systems, PS 350). REMPI signals were detected using a digital oscilloscope (Hewlett-Packard 54540C). The screening of the catalysts in the array microreactor took 2-3 minutes.

In Figure 3 the bar chart of benzene REMPI signals, E in mV, as measured by the oscilloscope, for each catalyst is shown (66 total). These measurements correspond to an average of 10 laser shots. Signal to noise ratios were better than 20 throughout the entire screening process. As evident from Figure 3, both Pt and Pd individually exhibited catalytic activities, with Pt being more active than Pd. On the other hand In was not catalytic towards benzene production from cyclohexane. These results are in complete agreement with the previous measurements reported by other investigators.^[16,17] The reproduction of the reactivity trends of individual metals is important in providing us confidence with regard to the suitability of the system developed for the intended catalyst screening studies. As seen in Figure 3, we also made the significant discovery that a ternary mixture (0.8%Pt, 0.1%Pd and 0.1% In) exhibited superior benzene productivity compared to all other metal combinations investigated. These results demonstrate the feasibility of using array microreactors coupled with REMPI screening for the discovery and optimization of heterogeneous catalysts.

Figure 3. Array microreactor screening results for the ternary Pt-Pd-In catalyst combinations.

In Table I the approximate times associated with the execution of each step in the preparation, processing and screening of the catalyst library are given. As noted earlier, the present array microreactor accommodates a total of 17 catalyst pellets. Consequently, the complete testing of the entire library involved 5 sequential runs. As shown in Table I the preparation and screening of the Pt-Pd-In ternary catalyst library took about 2.5 days to complete. Although this represents a significant advance over conventional methods, which may take months to achieve the same objectives, further acceleration of the overall process is clearly possible. For example, increasing the number of microreactors in each array, i.e. increasing the level of parallel processing, and decreasing the times for impregnation, drying and reduction can significantly speed up the process. However, the latter issues must be carefully implemented as even minor modifications in catalyst preparation methods can dramatically alter the performance of the catalysts. On the other hand, the minimization of catalyst preparation times can be explored using the combinatorial strategies described here.

Keywords: Combinatorial chemistry, heterogeneous catalysis, high throughput screening, resonance enhanced multiphoton ionization, laser screening.

[1] National Research Council, Catalysis Looks to the Future, National Academy Press, Washington, DC 1992.

[2] H. Heinemann, A Brief History of Industrial Catalysis, in Catalysis: Science and Technology, (Eds.:J.R. Anderson and M. Boudart), Springer-Verlag, Berlin 1981, Chapter 1.

[3] X-D. Xiang, X. Sun, G. Briceno, Y. Lou, K-A. Wang, H. Chang, W.G. Wallace-Freedman, S-W. Chen, P.G. Schultz, Science 1995, 268, 1738.

[4] G. Briceno, H. Chang, X-D. Sun, P.G. Schultz, X-D. Xiang, Science 1995, 270, 273.

[5] X-D. Sun, C. Gao, J. Wang, X-D. Xiang, Appl. Phys. Lett. 1997, 70, 3353.

[6] E. Danielson, J.H. Golden, E.W. McFarland, C.M. Reaves, W.H. Weinberg, X.D. Wu, Nature 1997, 389, 944.

[7] X-D. Sun, K-A. Wang, Y. Yoo, W.G. Wallace-Freedman, C. Gao, X-D. Xiang, P.G. Schultz, Adv. Mater. 1997, 9, 1046.

[8] E. Reddington, A. Sapienza, E. Guraou, R. Viswanathan, S. Sarangapani, E.S. Smotkin, T.E. Mallouk, Science 1998, 280,1735.

[9] J. Wang, Y. Yoo, C. Gao, I. Takeuchi, X-D. Sun, H. Chang, X-D. Xiang, P.G. Schultz, Science 1998, 279,1712.

[10] F.C. Moates, M. Somani, J. Annamalai, J.T. Richardson, D. Luss, R.C. Willson, Ind. Eng. Chem. Res. 1996, 35, 4801.

[11] S.J. Taylor, J.P. Morken, Science 1998, 280, 267.

[12] P. Cong, D. Giaquinta, S. Guan, E.W. McFarland, K. Self, H. Turner, W.H. Weinberg, "A Combinatorial Chemistry Approach to Oxidation Catalyst Discovery and Optimization", 2^nd Microreaction Technology Conference, New Orleans, LA January 1998.

[13] S.M. Senkan, Nature 1998, 394, 350.

[14] R. Srinivasan, I.M. Hsing, P.E. Berger, K.F. Jensen, S.L. Firebaugh, M.A. Schmidt, M.P. Harold, J.J. Lerou, J.F. Ryley, AIChE Journal 1997, 43, 3059.

[15] S.M. Senkan, M.J. Castaldi, J. Air and Waste Mgmgt. Assoc. 1998, 48, 77.

[16] K. Ahmed, H.M. Chowdhury, Chem. Eng. J. 1992, 50,165.

[17] D.M. Rehbon, and V. Haensel, J. Catalysis 1988, 111, 397.

[**] We would like to thank Scott Deskin, Kevin Krantz, and Anna Ly in the preparation of catalyst pellets and during the catalyst screening process. We also thank LCS Inc. for the funding of this research and the use of their facilities.