Jessica Dion
Senior Seminar

Abstract:

The rate of adsorption of molecules onto metallic surfaces depends on such things as the pressure of the gas or concentration of the molecules in solution, the number of sites available on the metal surface, and the type of adsorption which occurs. Physisorption is characterized by change in enthalpy of 20-25 ^kJ/_mole or less and doesn't involve the sharing or transfer of electrons. Chemisorption, on the other hand, is the results in chemical bond formation between the molecule and the metallic surface, often at the expense of the adsorbed molecule's structural integrity. Enthalpies of chemisorption are in the 200 ^kJ/_mole range. Many techniques have been developed for studying the structure of metallic surfaces, including STM (scanning tunneling microscopy), which gives depth profiling of a surface down to the atomic scale, LEED (low energy electron diffraction) which yields important information about the periodicity of the surface atoms, and EELS or HREELS (high resolution electron energy loss spectroscopy) which give information about the vibrational modes of adsorbed molecules, and therefore, their coordination environment. The catalytic properties of many metallic surfaces have been known for over a century. Some common reactions that use heterogeneous metallic catalysts include the hydrogenation of alkenes, polymerization reactions, including the synthesis of polyethylene and polypropylene, and the well-known and less understood Fischer-Tropsch synthesis of wide variety of organic products from carbon monoxide and hydrogen.

Theory of Adsorption

Atoms in the solid phase have much less mobility than atoms in the liquid or gaseous phases.¹⁰ Because of this, surface variations in a solid crystal are relatively permanent over time. The variety of coordination environments available on the surface a metal allow for the possibility of several types of interactions between molecules in an adjacent gaseous or liquid phase.

Adsorption is the process whereby a molecule becomes immobilized at an interphase between two phases, without being dissolved in either phase.¹¹ It is usually an exothermic process, although there are some exceptions. The surface tension of the metal often decreases as molecules become adsorbed.¹ Because there is a greater probability of adsorbed molecules interacting as the surface coverage increases, the enthalpy of adsorption depends on the amount of surface coverage.

The interaction between two adsorbed atoms or molecules on the surface may be attractive or repulsive.⁶ If the molecules attract each other, such as in the case of oxygen on tungsten, they will adsorb in small clusters, growth of the layer will occur chiefly at the boundaries of the clusters. If the adsorbed molecules repel each other, such as in the case of oxygen molecules on palladium, they will adsorb in a disordered layer, until the surface is nearly covered. The enthalpy of adsprotion in this case becomes less negative as the fractional coverage increases.

There are two main types of adsorption, characterized by the type and energy of the adsorbate-metal bond formed.^1,6,12 When molecules become weakly immobolized on a surface, due to van der Waals or dipole interactions, they are said to be physically adsorbed, or physisorbed. The enthalpies of physisorbtion are typically less than 20-25 ^kJ/_mole ⁶. The forces holding physisorbed molecules to the surface are therefore not strong enough to break chemical bonds within the adsorbed molecule. The molecule retains its identity, although it may become structurally distorted due to interactions with the surface. The energy liberated when molecules become physisorbed is often absorbed by the metal lattice as thermal motion. Because of this, it is possible to determine the amount or rate of adsorption by measuring the change in temperature of the metal with a known specific heat.

Physisorption requires no activation energy, and therefore, equilibrium between the surface and the gas or solution is achieved very quickly. Because no true chemical bonds are made between physisorbed species and the surface, an adsorbed molecule may wander or diffuse along the surface of a metal. In diffusing, it might travel across a crystal face until it gains enough kinetic energy to break free from the surface. It may also migrate onto another crystal plane, or into a step or kink along the surface, where it may become more tightly bound. This mobility of physisorbed molecules is an important part of the catalytic properties of metallic surfaces.

Sometimes electron transfer or sharing will occur between an adsorbed molecule and surface atoms. This case, where a true chemical bond is formed, is referred to as chemical adsorption, or chemisorption.^1,12 The enthalpy of chemisorption is often as high as 200 ^kJ/_mole, although there is often an activation barrier between the physically and chemically adsorbed molecule, such as with N₂ on iron^4,6. In the process of becoming chemisorbed, a molecule may be torn apart, in order to satisfy valency requirements of the surface metal atoms. Molecular fragments would then result, and the molecule would lose its former identity. This is another important property that leads to the high catalytic activity observed for many metallic surfaces.

In some cases, different degrees of chemisorption are seen. For the group I and II metals, an interesting trend is seen.² When beryllium or aluminum (metals with relatively low electron density) are exposed to atomic hydrogen at relatively low temperatures, a small charge transfer occurs, and the hydrogen atoms become chemisorbed to the surface. At higher temperatures, the equilibrium shifts, and some hydrogen desorbes from the surface, but no qualitative change in the type of H-metal bond is seen. When magnesium, a metal with an intermediate charge density, is exposed to atomic hydrogen at relatively low temperatures, a surface hydride is formed. At higher temperatures, the hydrogen begins to dissolve into the bulk metal, and the rate of desorption increases. In the case of lithium, another metal with an intermediate electron density, a surface hydride is formed at low temperatures, and this LiH is seen to extend into the bulk of the metal. At higher temperatures, the hydride extends further into the bulk of the metal. For metals with higher electron density, such as potassium and sodium, a thick layer of hydride is seen to form on the surface and extend into the bulk of the metal. At higher temperatures, the effect becomes even more pronounced.

Kinetics of Adsorption

In the case of adsorption, molecules in the free gaseous state are in a dynamic equilibrium with adsorbed molecules. The fractional coverage of the surface, s, is equal to the number of occupied adsorption sites of the metal divided by the number of unoccupied sites. The fractional coverage, therefore, depends on the pressure (concentration) of the gas, and the number of sites available.⁶

There are many models for describing the kinetics of adsorption. One of the first models was proposed in 1918 by Langmuir.³ Langmuir made two basic assumptions. The first was that in the process of adsorption, a simple monolayer of adsorbed molecules was formed, after which time the surface would become saturated, and no additional molecules could be adsorbed. The second assumption made was that none of the molecules adsorbed onto the surface had attractive or repelling interactions, even at high fractional coverages. Plots of the number of molecules adsorbed vs. the pressure (concentration) of gas present at a constant temperature are often referred to as Langmuir Isotherms. A plot of the volume (or amount) adsorbed vs. pressure will have the general line shape:

As the pressure (or concentration) increases, the volume (amount) of gas that become adsorbed to the surface increases until finally, when the surface is saturated, higher pressures will lead to no change in the amount of adsorbed gas.^5,6

Isotherms may take other forms - often it is easier to statistically interpret a relationship that approaches linearity under ideal circumstances. An isothermal plot of ¹/_volume vs. ¹/_pressure will be linear, in the case of a non-interacting monolayer.

A plot of ^pressure/_volume vs. pressure will also be linear under those conditions.
Consider the reaction:

		 ka
	G + M  <---->  GM
		 kd

In the case where the gas molecule dissociates when it becomes adsorbed, as is often the case with chemisorption of diatomic gases on metallic surfaces, the rate of adsorption will be proportional not only to the pressure, but to the probability that each dissociated fragment will find a site on the metal. The rate will be equal to kap[N(1-s)]2, and the rate of desorption will be equal to Kdp(Ns)2. At equilibrium, the overall rate will be equal to Kp1/2/1+(Kp)1/2.⁶ (NOTE: THIS IS WRONG, see Atkins for the real formulas!! *blush*)

Methods for Studying Surface Structure

There are currently hundreds of methods for studying the structure of clean metallic surfaces, and the changes that those surfaces undergo as a result of adsorption. A few of the most common techniques for studying metallic surfaces will be introduced now, including Scanning Tunneling Microscopy, Low Energy Electron Diffraction Spectroscopy, and Electron Energy Loss Spectroscopy. Most methods for studying phenomena related to adsorption of gases onto metallic surfaces take place in a UHV (ultra high vacuum) chamber, where pressures as low as 10^-9 to 10^-11 torr may be obtained. The surfaces of metals used in most studies must be free of other adsorbents before the experiments begin. This is most often accomplished by using repeated cycles of Argon sputtering (where argon atoms bombard the surface of the metal) and annealing.

Scanning Tunneling microscopy (STM) is a technique that was was invented in the 1980s by Gerd Binnig and Heinrich Rohrer, who received the Nobel Price in Chemistry for their invention.⁷ A tungsten probe is sharpened using field evaporation8 until the radius of curvature at the tip (r) is only a few angstroms. The probe is then brought to within a few angstroms of a conducting surface. As the probe scans the surface in the xy directions, a constant tunneling resistance is maintained between the tip and the surface, via a feedback mechanism. This tunneling current is usually in nanoamperes or less. The current varies exponentially with the distance between the probe tip and the sample (d), and the tunneling current typically drops by an order of magnitude for every � the tip is away from the surface. The result is a contour map, or "topograph" of the surface. It is difficult, however, to define exactly what a surface is, at the sub-atomic level. The instrument is fairly insensitive to the positions of atoms under the first layer.¹³

The apparent "depth" of the surface can be traced to within tenths of angstroms, ideally. The depth resolution and accuracy depend on the angle of the tip.^7,8 If the tip is very narrow, a more accurate scan along sharp vertical boundaries is obtained, because the tip doesn't "get in its own way". Wider tips will be able to scan a shallower area, and the depth profile will be different. Even tips with a relatively large radius of curvature (1000� or so), a resolution of 50� may be obtained along the surface. Because the tunneling current is so dependant on the gap distance, the best topographs obtained are thought to be the result of current between single atoms on the surface and single atoms on the probe. The lateral resolution is somewhat poorer, and is approximately equal to [ (2�) (r+d) ]1/2¹³

Some advantages of this technique include the extremely high depth profiling possible, the face that, since no lenses are used, noise from abberations is not a problem, radiation-induced sample changes are unlikely, because the electrons involved have very low energy (only a few electron volts at most). Also, STM can be carried out in air, other gases, or even in a liquid because there are no free elctrons involved. This technique also provides direct, real-space information about the surface, which is particulary important when examining non-periodic surfaces or surfaces with defects.

Low Energy Electron Diffraction (LEED) is a very common technique used to study the structure of metallic surfaces.^1,4 Usually a single crystal is placed inside an ultra-high vacuum chamber, and a beam of electrons is focused on the crystal, perpendicular to the plane of the surface. Because the electrons are of low energy (only a few eV), they do not penetrate the sample to a depth of more than a few atomic radiif, and only surface information is obtained. The electrons reflecting off of the surface of the metal meet three semi-circular grids. The first grid is of the same potential as the metal crystal. The second grid only admits electrons of the original energy, so that only electrons making elastic collisions with the metal contribute to the signal. The third grid accelerates the electrons onto a fluorescent screen, which may be photographed.

The diffraction pattern for a clean surface usually consists of a pattern of spots which correspond to the symmetry of the surface grid of atoms. This technique therefore gives information about the unit cell and periodicity of the surface.⁵ The resolution is actually about .01�-1 , so that the degree of order and type of symmetry can be determined over a region of about 100 � (at most)⁵ The pattern will shrink or expand depending on the energy of the incident electron beam. Surfaces containing periodic defects or arrays of ordered adsorbed molecules often exhibit more complex, but interpretable, patterns.^6,7 Disordered surfaces or surfaces with randomly adsorbed molecules often produce "blurry" diffraction patterns.

Electron Energy Loss Spectroscopy (EELS) is a similar technique to LEED, but it is more useful for surfaces with adsorbed molecules, because it yields vibrational information.^5,6 This may lead to insight as to the types of bonding which have occurred between the molecule and the metal. The instrument is the same as that used for LEED experiments, except that in this case, electrons making inelastic collisions with the metal surface contribute to the signal. In this case, instead of obtaining a "photograph" of the surface periodicity, a plot can be made with intensity vs. electron energy loss. The energy lost by the electrons colliding with the surface will correspond quantitatively with the energy absorbed by adsorbed molecules in the form of vibrations. This technique is very sensitive to light elements, and amounts of adsorbate as low as m1% of a monolayer may be detected with this method.⁶

These are just a few powerful techniques commonly used to study surface structure and phenomena. Many other forms of spectroscopy and microscopy are used for different kinds of specific surface studies. A fairly complete table defining these methods may be found in Ref. 1, pp. 330-337.

Catalytic Properties

Knowledge gained from the study of metallic surfaces and adsorption has many applications¹⁴, and can contribute to the development of better insulating layers for metals, corrision control, decorative coatings, adhesion promotion, molecular sensors, heterogeneous catalysts, internal prosthetics for living organisms¹⁵, and hundreds of others. One important reason for studying the structure and reactivity of reactions that take place on the surface of metals is to understand the origins of the catalytic properties of many metallic surfaces. Some of the best metal catalysts are those which have a face centered cubic crystal structure, such as Pt. Ni, Pd, Rh, Cu, Ag and Au.⁵ The difference in catalytic activity between each of these may be partially due to differencec in restructuring along the surface upon adsorption, as well as the valence characteristics. Each metal exhibits different types of distortions and symmetry changes along different crystal planes upon adsorption. Therefore, each are more suited for catalyzing some reactions than others.

There is always a change in enthalpy that is associated with any reaction. If the change in enthalpy is positive, some activation energy will be needed in order for the transition state to be reached and the products formed. If this activation energy is much greater than the energy available to the reactants, the reaction will proceed at a very slow rate. A catalyst is a substance which is regenerated in the reaction process (doesn't become depleted), and which lowers the activation energy needed to form products from reactants. Intermediates formed in the presence of a catalyst should not be so stable that the product is thermodynamically unfavorable.⁴ This is to be considered when developing a catalyst.

Reactions catalyzed by metals may be homogeneous, where the metal is dissolved in the reaction medium, or they may be homogeneous, where the metal is in a discrete, immiscible phase (usually a solid) from the reaction medium.¹⁶ Heterogeneous catalysts often require relatively high temperatures or pressures, and lead to a mixture of products. To increase the surface area of the heterogeneous catalyst, and therefore the number of active sites, a heterogeneous catalyst is often "finely divided" or broken into small particles, but it is still a separate phase from the reactants. In spite of the hundreds of techniques available for studying surfaces, it is much more difficult to study heterogeneous reactions than it is to study reactions where all species are in solution. Also, many heterogeneous catalysts produce a wide range of products and have no single mechanism of action. Because of this, the exact mechanisms for many seemingly simple reactions which are catalyzed by metals, such as the production of ammonia from diatomic nitrogen and hydrogen, is still poorly understood.⁴

Homogeneous catalysts are somewhat more selective, and the reactions often require lower temperatures and pressures, but the catalyst must be separated from the reaction medium and products when the reaction is complete.¹⁶ Because many advanced analytical methods require the system being studied to be homogeneous or in solition, the mechanisms for many homogeneous reactions are well understood. A hybrid form of catalyst, which attempts to take advantages of the selectivity of homogeneous catalysts and the separability of heterogeneous catalysts is sometimes seen. In these cases, an organometallic group is anchored to a solid support. The distinction between heterogeneous and homogenous catalysts is somewhat arbitrary in this case, because some reactions that occur within the coordination sphere of a metal atom in solution will also take place if the metal atom is anchored on a solid support.¹⁷

Specific Catalytic Applications

Some basic reactions that can be efficiently catalyzed by metallic surfaces include the hydrogenation of alkenes and alkynes, catalytic cracking (a process where high molecular weight (C₁₂ or higher) fractions of petroleum are converted to mixture of more highly branched, lower molecular weight alkanes (C₅-C₁₀)), the production of polymers, including the synthesis of high density polyethylene (HDPE) and polypropylene, the well-known and less understood Fischer-Tropsch synthesis of a wide variety of organic molecules from carbon monoxide and hydrogen gas and many others.

The hydrogenation of alkenes with molecular hydrogen is an exothermic process, with a (H§ ( -30 kcal/mol.¹⁸ However, the uncatalyzed reaction has a high energy of activation. Although this reaction does not occur at an appreciable rate a room temperature, it proceeds quickly when a heterogenous metal catalyst is added, such as finely divided platinum, nickel, palladium, rhodium, or ruthenium.¹⁸ When such a metal is present, molecular hydrogen becomes chemisorbed to the surface, weakening the hydrogen-hydrogen bond considerably. Alkenes also become adsorbed to the surface, and the hydrogen atoms are added to the alkene in a stepwise, sin fashion. The newly formed alkane then breaks free from the surface, and the process continues. This is one example where the stereochemistry of the product may be controlled by using a heterogeneous catalys.

The production of polypropylene using a Titanium chloride salt anchored onto a MgCl₂ support, known as the Ziegler Natta reaction, can be acheived using ambient temperatures and pressures. Ziegler and Natta received the Mobel Prize in 1963 for this development.²⁰ In this reaction, the titanium atoms on the surface do not have a filled coordination sphere, and they act as lewis acids, effectively accepting adsorbed ethylene or propylene as a ligand.

The Fischer-Tropsch method for synthesizing organic molecules from carbon monoxide and molecular hydrogen is an incredible example of the lack of specificity and complex mecanisms involved in heterogeneous catalysis.¹⁹ The overall reaction is:

nCO + 2nH₂ ---> -(CH₂)n + nH₂O.

The (H500K = 39.4 ^kcal/_mol for this reaction. The products include various straight chain alkanes, alkenes, alcohols, aldehydes, carboxylic acids, esters, and some arenes.¹⁹ The heterogeneous metal used to catalyze this reaction may be Fe, Co, Ni or Rh₂.¹

The probability of chain growth is affected by several reaction conditions.²¹ The basicity of the catalyst has a definite effect on on the proportion of products formed. K₂O is often added to the metal to increase the efficiency of the reaction. It is thought that K₂O donates some elctron density to the metal, lowering its work function and making it easier for the metal to donate electron density to adsorbed CO. The overall temperature of the reaction also affects the dominant products. The thermodynamic products for the complete reaction are carbon (graphite) and methane. At lower temperatures, kinetic factors dominate, and more complex molecules are seen as the "products" of the catalyzed reaction. At higher temperatures, the formation of methane and smaller chain alkanes increases, as thermodynamic considerations have a greater affect.

Perhaps the most important reaction condition is the relative pressures of the reactants.¹⁹ CO and H₂O are known to be strongly chemisorbed on these metals under these reaction conditions, while H2 is only weakly chemisorbed. If the partial pressure of molecular hydrogen is high relative to CO and H₂O, it will be able to compete with the latter two gases more effectly for active sites on the metal surface. Because of the increased concentration of chemisorbed hydrogen, the probability of chain terminating steps will be fairly high, and the products will have fairly low molecular weights. If the partial pressure of H2 is kept low with respect to CO and H₂O partial pressures, the surface will have less chemisorbed hydrogen, and the chain formation and growth steps will dominate.

In each of these reactions, the adsorption (chemisorption) of the reactants lowers the activation energy of the reaction considerably. These are just a few examples of heterogeneous catalysts - there are many more, composed of alloys or metal solutions, salts, oxides, and other combinations and types of surfaces.

Future Prospects

The advancements in our understanding of the forces governing adsorption onto metallic surfaces have great implications not only for academic research, but in such fields as materials science, as human demand for clothing and machined products becomes much more advanced, in the engineering, electronics, and semiconducting industries, where it is becoming more important to understand adsorption and the exact properties of certain metals and alloys, as the pressure to miniaturize increases. Other applications include the development of biocompatible artificial tissues, limbs, and prosthetics¹⁵, and the development of different or safer methods for processing of food components, such as in the partial hydrogenation of saturated fats. As our ability to probe the exact nature of clean and unclean surfaces increases, and because the surfaces of metals and alloys vary so greatly at the atomic level, the prospects for developing new heterogeneous catalysts are very great.

References:

1. Physical Chemistry of Surfaces, Adamson, Arthur W., c1990 John Wiley and Sons, Inc.

2. Sprunger, P. T., Plummer, E. W., The Interaction of Hydrogen with Simple Metal Surfaces, Surface Science, 307-309, 118-122 (1994)

3. Langmuir, Journal of the American Chemical Society, 40, 1361-1402 (1918)

4. Kinetics of Heterogeneous Catalytis Reactions, Boudart, Michel, Dj‚ga-Mariadassou, G., c1984 Princeton University Press. NJ.

5. Topics in Applied Physics: Volume 4: Interactions on Metal Surfaces, Gomer, R. Ed., c1975 Springer-Verlag, N.Y.

6. Physical Chemistry, 4th ed., Aktins, P. W., c1990 Oxford University Press.

7. Scanning Probe Microscopy and Spectroscopy:Methods and Applications, Wiesendanger, Roland, c1975 Cambridge University Press.

8. Methods of Experimental Physics, Volume 27: Scanning Tunneling Microscopy, Stroscio, Joseph A., Kaiser, William J., c1993 Academic Press, Inc.

9. Ohtani, H. Kao, C. T., Van Hove, M. A., Somorjai, G. A., Prog. Surface Science 23, 155 (1987)

10. The Physics and Chemistry of Surfaces, 3rd ed., Adam, N. K., c1941 Oxford University Press, Inc.

11. On Physical Adsorption, Ross, S., Olivier, J. P., c1964 John Wiley and Sons, Inc.

12. Thermal Accomodation Coeffiecients and Adsorption Coefficients of Gases, Saxena, S.C., Joshi, R.K., 1981 Purdue Research Foundation.

13. Tersoff, J., Hamann, D.R., Theory of the scanning tunneling microscope, Phys. Rev. B, 31(2), 805 (1985)

14. Richarson, N.V., Frederick, B.G., Unertl, W.N., El Farrash, A., Chemisorption studies related to reactive organic film growth, Surface Scienceb, 307-309 124-131 (1994)

15. Personal communication, Moira Ringo, Dept. of Chemistry, University of Vermont, 1995

16. Inorganic Chemistry: Principles of structure and reactivity, 4th ed., Huheey, J., Keiter, E.A., Keiter, R.L., c1993 Harper College Publishers.

17. Marshall, G.W., Putscher, R.E., Organometallic Chemistry and Catalysis in Industry, Journal of Chemical Education, 63, 189-191 (1986)

18. Organic Chemistry, Solomons, T.W.G., c1992 John Wiley and Sons, Inc.

19. Hermann, W.A., Organometallic aspects of the Fischer-Tropsch Synthesis, Journal of Organometallic Chemistry, 372, 117-130 (1989)

20. The History and current state of the art of polypropylene polymerization catalysis, Journal of Chemical Education, 63, 191-195 (1986)

21. Dry, M.E., Commercial conversion of carbon monoxide to fuels and chemicals, Journal of Organometallic Chemistry, 372, 117-127 (1989)