Newns-Anderson Model

The bonding of atoms and molecules to surfaces is a fundamental process that governs everything from industrial catalysis to the function of electronic devices. This phenomenon, known as chemisorption, involves the formation of strong chemical bonds, but describing this intricate quantum dance between a single entity and an infinite solid presents a significant theoretical challenge. How can we build a model that is simple enough to be insightful yet powerful enough to make predictions?

This article explores the Newns-Anderson model, an elegant theoretical framework that provides the language to understand chemisorption. It bridges the gap between the discrete quantum states of an atom and the continuous electronic bands of a metal surface. Over the following chapters, you will discover the core principles of this model and its profound consequences. The first chapter, "Principles and Mechanisms," will unpack the quantum mechanics behind the model, exploring how concepts like hybridization, level broadening, and electron correlation define the nature of a surface chemical bond. Subsequently, "Applications and Interdisciplinary Connections" will demonstrate how this single idea revolutionizes fields like catalysis, enabling the computational design of new materials and providing a unified perspective that links surface chemistry to electrochemistry and nanoscience.

Imagine you are an atom, let's say a hydrogen atom, drifting towards the vast, shimmering surface of a block of platinum. From your perspective, the surface isn't just a hard floor; it's a boundless, undulating sea of electrons. Your own electron is confined to specific, quantized energy levels, like the rungs of a ladder. The metal, however, has so many atoms packed together that their energy levels have merged into a near-infinite continuum of states—a smooth, endless ramp. What happens when your ladder meets this ramp? This is the fundamental question of chemisorption, and its answer is a beautiful story of quantum mechanics.

The interaction is not a simple collision. It's a "conversation." Your electron, once bound to you, "sees" the vast expanse of available states in the metal. It now has a choice: it can stay with you, or it can hop onto the metal and wander away. Likewise, electrons from the metal's sea can hop onto your vacant energy levels. This quantum mechanical possibility of hopping back and forth is called ​​hybridization​​. It is the very heart of the strong, ​​covalent​​ chemical bond that can form between an atom and a surface.

But there's another part to the story. If the energy levels line up just right, there might be a net flow of charge. Perhaps it's more favorable for your electron to spend most of its time in the metal, leaving you as a positive ion. Or maybe you can snatch an electron from the metal, becoming a negative ion. This charge transfer creates an ​​ionic​​ component to the bond, an electrostatic attraction that complements the covalent sharing. To build a complete picture, we need a language that can describe both. That language is the Hamiltonian.

In physics, a Hamiltonian is simply a function that represents the total energy of a system. It's a complete recipe for its behavior. The Newns-Anderson model provides an elegantly simple, yet profoundly powerful, Hamiltonian for our atom-on-a-surface problem. Let's break it down piece by piece:

1. ​​Atom's Energy ($\epsilon_a$):​​ The first term describes the energy of having an electron in the adsorbate's orbital. The energy level is $\epsilon_a$, and $n_{a\sigma}$ is the number operator that simply counts how many electrons of spin $\sigma$ are on the atom.

2. ​​Metal's Energy ($\epsilon_k$):​​ The second term is the total energy of all the electrons in the metal's sea, a sum over all of its continuous states, indexed by $k$. For now, these electrons don't interact with each other.

3. ​​The "Hop" ($V_k$):​​ This is the star of the show! It's the hybridization term. The part $a^{\dagger}_{\sigma} c_{k\sigma}$ describes an electron being destroyed in a metal state $k$ and created on the atom $a$. Its partner, $c^{\dagger}_{k\sigma} a_{\sigma}$, describes the reverse process. This term mathematically allows the electrons to be shared, forming the covalent bond. The strength of this hop is given by the matrix element $V_k$.

4. ​​The "Squeeze" ($U$):​​ This final term is a nod to reality. Electrons are all negatively charged and they repel each other. The term $U n_{a\uparrow} n_{a\downarrow}$ adds an energy penalty, $U$, only if two electrons with opposite spins ($\uparrow$ and $\downarrow$) try to occupy the single atomic orbital at the same time. It's an energy cost for squeezing two electrons into one tiny space. This is our first taste of electron-electron​​ correlation​​.

What are the consequences of this Hamiltonian? The "hopping" term has a dramatic effect. Our atom's once perfectly sharp energy level, $\epsilon_a$, gets smeared out into a ​​resonance​​.

Think about it through the lens of the uncertainty principle, $\Delta E \Delta t \gtrsim \hbar$. If we place an electron on the adsorbate, it won't stay there forever. It has a finite ​​lifetime​​, $\tau$, before it hops into the vast continuum of the metal. A finite lifetime ($\Delta t = \tau$) means the state cannot have a perfectly defined energy; there must be an energy uncertainty, or a ​​level broadening​​, $\Gamma \approx \hbar/\tau$.

This entire process is captured mathematically by a powerful tool called the ​​Green's function​​. The derivation shows that the effect of the entire metal continuum on the adsorbate can be bundled into a single complex quantity called the ​​self-energy​​, $\Sigma(\epsilon)$. The adsorbate's effective Green's function becomes:

The self-energy $\Sigma(\epsilon) = \Lambda(\epsilon) - i\frac{\Gamma(\epsilon)}{2}$ has two parts, each with a clear physical meaning:

• The imaginary part gives us the level broadening, $\Gamma(\epsilon)$. As intuition suggests, this broadening is proportional to the square of the coupling strength and the density of states in the metal available to hop into: $\Gamma(\epsilon) = 2\pi \sum_{k} |V_k|^2 \delta(\epsilon - \epsilon_k)$. More escape routes mean a shorter lifetime and a broader level.

• The real part, $\Lambda(\epsilon)$, causes an overall ​​energy shift​​. The center of the new, broadened level is not exactly at the original $\epsilon_a$. The presence of the surface itself shifts the level up or down.

The final, smeared-out energy level is called the ​​Projected Density of States (PDOS)​​. Under the simple "wide-band approximation" (assuming $\Gamma$ is constant), the PDOS takes on the elegant shape of a ​​Lorentzian​​ distribution, a bell-like curve centered at the shifted energy $\epsilon_a + \Lambda$ with a width of $\Gamma$.

So, an atom's level broadens and shifts. But why does it stick to the surface? The answer, as with any chemical bond, is that the total energy of the system is lowered.

When the atomic orbital at $\epsilon_a$ hybridizes with the metal states, they form new, mixed states. Crudely speaking, for each interacting metal state, a lower-energy ​​bonding​​ state and a higher-energy ​​antibonding​​ state are formed. To determine the final energy of the system, we need to fill these new states with electrons.

This is where the ​​Fermi level​​, $\epsilon_F$, comes in. It is the "sea level" of the electron ocean. At zero temperature, all states with energy below $\epsilon_F$ are filled, and all states above it are empty.

A strong chemical bond forms when the hybridization pushes most of the bonding character to energies below the Fermi level, where they become occupied by electrons, thus lowering the system's total energy. Simultaneously, the antibonding character is pushed to energies above the Fermi level, where the states remain unoccupied. If the antibonding states were to dip below $\epsilon_F$ and become filled, they would counteract the stabilization from the bonding states, weakening the overall bond.

The chemisorption energy, $\Delta E$, is the sum of the energies of all the occupied electrons in the final system minus their energy in the initial, separated system. A negative $\Delta E$ means the atom is happily bound to the surface.

This theory might seem abstract, but it provides the key to one of the most important technological puzzles: how to design better catalysts. Why are metals like platinum, palladium, and rhodium such excellent catalysts for so many reactions, from car exhausts to industrial chemistry?

The secret lies in their partially filled ​​d-bands​​. These electronic states are relatively localized and sit in a narrow energy range near the Fermi level, making them perfectly poised to interact with adsorbate molecules. The Newns-Anderson picture can be adapted into the powerful ​​d-band model​​, pioneered by Jens Nørskov and his colleagues.

Instead of tracking every detail of the d-band, they proposed a single, powerful ​​descriptor​​: the ​​d-band center​​, $\varepsilon_d$. This is simply the average energy of all the d-orbitals, both occupied and unoccupied, referenced to the Fermi level.

The central idea is astonishingly simple: the energy of the antibonding states, which are so crucial for determining the final bond strength, depends on the energy of the d-states they hybridize with. As the d-band center $\varepsilon_d$ shifts, so does the energy of the antibonding states. For many adsorbates, as $\varepsilon_d$ moves closer to the Fermi level, the antibonding states are pushed further up, above $\epsilon_F$. This leads to less filling of these destabilizing states and therefore a stronger chemical bond.

This simple principle explains a remarkable amount of chemistry. It predicts that moving from right to left across the transition metals in the periodic table (e.g., from Cu to Ni to Co), the d-band center rises, and the binding of many molecules becomes stronger. This often leads to a simple, approximately ​​linear scaling relationship​​ between the chemisorption energy and the d-band center. This model has transformed catalysis from a black art into a predictive science, allowing researchers to computationally screen for new catalyst materials. Of course, this simple descriptor has its limits; for complex materials like oxides and carbides where the metal d-states are already strongly mixed with other elements, the picture gets more complicated, and other descriptors may be needed.

The beauty of a great physical model is its ability to unify seemingly disparate phenomena. Let's ask one more question: can our adsorbed atom become magnetic?

Here, the "squeeze" energy, $U$, becomes the hero of the story. Imagine placing a spin-up electron on the atom. Now, bringing in a spin-down electron costs the hybridization energy, but we must also pay the Coulomb penalty $U$. This repulsion works against pairing up electrons.

In the presence of an external magnetic field, spin-up and spin-down levels split. An electron on the atom now feels a complex combination of forces: the external field trying to align its spin, the Coulomb repulsion $U$ discouraging a partner of the opposite spin, and the hybridization $V_k$ trying to delocalize it into the non-magnetic metal.

The Newns-Anderson model, solved in a mean-field approximation, allows us to calculate the resulting ​​spin susceptibility​​—a measure of how strongly the atom's-own magnetism responds to an external field. Amazingly, the model predicts that if the Coulomb repulsion $U$ is large enough compared to the level broadening $\Gamma$, the atom can develop a spontaneous magnetic moment even without an external field. This is the famous Stoner criterion for magnetism, emerging naturally from our model of surface chemistry. The same principles that determine how strongly a molecule sticks to a catalyst also govern whether that site becomes a tiny magnet. This is the inherent unity and beauty that physics strives to reveal.

Now that we have taken apart the elegant clockwork of the Newns-Anderson model, let’s see what it can do. A beautiful theory is not merely an intellectual ornament to be admired on a shelf; it is a key, a tool for unlocking the world. In the previous chapter, we saw how the interaction between a discrete atomic state and a continuous band of metal states gives rise to new, hybridized states. We saw how the energy and filling of these states determine the very nature of the chemical bond on a surface. Now, we are ready to leave the abstract realm of Hamiltonians and Green's functions and witness how this single, powerful idea breathes life into a vast range of scientific disciplines, from the high-stakes world of industrial catalysis to the subtle dynamics of a single ion glancing off a surface.

For centuries, humanity has sought a "philosopher's stone" to turn lead into gold. In modern science, one of our greatest quests is for a different kind of alchemy: the search for perfect catalysts. These materials, which speed up chemical reactions without being consumed, are the invisible engines of our civilization, responsible for everything from fertilizers that feed the world to the plastics in our homes and the clean energy technologies of our future. The traditional way to find a better catalyst was a laborious, expensive process of trial and error. But what if we could design them from scratch, using only the laws of quantum mechanics?

The Newns-Anderson model and its descendants provide the theoretical foundation for this dream. The challenge is to find a simple, calculable property of a material—a "descriptor"—that predicts its catalytic performance. For transition metals, the heroes of catalysis, an astonishingly effective descriptor has emerged: the ​​d-band center​​. This is simply the average energy of the metal's outermost, most reactive d-electrons.

Why should this one number be so powerful? Our model gives us the beautiful answer. All of catalysis is a dance of making and breaking bonds. An effective catalyst must first bind the reactant molecules, but not too weakly, or nothing will happen. Then, it must help them transform, and finally, it must release the products, but not too strongly, or the surface will become clogged and the reaction will grind to a halt. This "Goldilocks" principle of "just right" bonding was first recognized a century ago by the chemist Paul Sabatier. The d-band model gives us the quantum map to find this sweet spot.

A metal with a high-energy d-band center (one that is closer to the Fermi level, the "sea level" of electrons) has highly reactive electrons. According to the logic of the Newns-Anderson model, this leads to a strong chemical bond with an adsorbate. The reason is subtle and profound: the formation of the bond creates both stabilizing "bonding" states and destabilizing "antibonding" states. A higher d-band pushes these antibonding states to a higher energy, making it less likely that they will be filled with electrons. An unfilled antibonding state is the signature of a strong bond. Conversely, a low-energy d-band leads to weaker bonds.

This simple idea explains a wealth of chemical trends. For instance, it tells us why carbon monoxide (CO) binds very strongly to early transition metals like Titanium but weakly to late ones like Gold. As one moves across the periodic table, the d-bands become more filled and contract, sinking to lower energy. This systematically weakens the crucial "back-donation" of electrons from the metal into CO's empty $\pi^*$ orbitals, which is the primary source of the bond's strength.

The pinnacle of this concept is the celebrated ​​"volcano plot"​​. If we plot the rate of a catalytic reaction against the d-band center for a series of different metals, the points often trace out the shape of a volcano. On the left slope are the weak-binding metals, where the reaction is limited by getting molecules to stick and activate. On the right slope are the strong-binding metals, where the reaction is choked by products that won't let go. At the very peak of the volcano sits the ideal catalyst—the metal with the optimal d-band center, perfectly balancing the acts of binding and releasing. The theory doesn't just explain the volcano; it gives us a map and a compass to find our way to its summit.

The model's power extends beyond simply predicting how strongly something sticks. It can also predict what happens to it after it sticks—its chemical fate.

Consider a molecule of oxygen, $O_2$, approaching a surface. Will it land gently and remain an intact $O_2$ molecule, or will it be violently torn apart into two separate oxygen atoms? This choice between molecular and dissociative adsorption is the crucial first step in countless processes, from the rusting of iron to the generation of energy in a fuel cell.

Once again, the d-band center provides the key insight. A metal with a high d-band center, like Nickel, not only forms strong bonds with individual oxygen atoms but also actively facilitates the cleavage of the O=O bond. The strong interaction helps to stabilize the transition state where the bond is stretched to its breaking point. In contrast, a metal with a very low d-band center, like Gold, interacts so weakly that the oxygen molecule tends to adsorb (if at all) without dissociating.

This connection between binding energy and reaction barriers is captured by another profound concept in chemistry, the Brønsted–Evans–Polanyi (BEP) relation. It states that for a family of similar reactions, the activation energy (the hill that reactants must climb) is often linearly related to the final reaction energy (the overall change in elevation). Since our d-band model predicts the binding energies of the products (like the two O atoms), it also gives us a direct handle on the kinetic barrier to form them. We have made a momentous leap: from understanding structure and stability to predicting the speed of chemical reactions from first principles.

Armed with this predictive power, chemists and materials scientists can transcend explaining nature and begin to engineer it. We can become atomic-scale alchemists, designing new materials on a computer before ever setting foot in a lab.

​​Alloying:​​ What if the best catalyst for a reaction—say, Platinum—is incredibly rare and expensive? Can we do better? Yes. By mixing it with other metals, we can create alloys that precisely tune the d-band center. The effects combine electronic (ligand) and structural (strain) factors. For example, a subsurface layer of a more reactive metal like Nickel can electronically modify surface Pt atoms, causing their d-band center to shift down and weaken binding. Conversely, applying tensile strain to Pt (for example, by layering it on a different metal) can shift its d-band center up, strengthening binding. While the exact energy shifts in pedagogical examples might be illustrative, this principle is used in the real world to design catalysts that are not only more active but also more robust and less expensive.

​​Nanoscience and Strain:​​ We can play even more exotic tricks. If we shrink a catalyst down to a nanoparticle containing only a few dozen atoms, quantum mechanics introduces new effects. For a free-floating cluster, the atoms on the surface have fewer neighbors than in a bulk crystal. This reduced coordination narrows the d-band and, to conserve the number of states, pushes its center up in energy, often making the nanoparticle more reactive than its bulk counterpart. But this is not the whole story! If we then place this nanoparticle on an oxide support material, the support can pull electron density away from the metal, creating a positive charge that shifts the d-band back down. This delicate interplay of size, shape, and environment gives us a rich toolbox for tuning catalytic properties. The ultimate limit is the single-atom catalyst, where a single metal atom does the work, representing the purest test of our bonding models.

Amazingly, we don't even need to change the chemistry. We can simply take a perfect crystal of a catalyst and physically stretch it. This "tensile strain" reduces the overlap between d-orbitals, which, like the effect of reduced coordination, raises the d-band center and enhances reactivity for certain reactions, such as $O_2$ dissociation. We can literally tune a catalyst's activity by putting it on a rack.

The beauty of a truly fundamental concept is that its ripples are felt far and wide. The ideas born from the Newns-Anderson model provide a unifying language across an astonishing breadth of scientific fields.

​​Electrochemistry:​​ Instead of tuning energy levels with chemistry, what if we could use a simple knob connected to a power supply? This is the domain of electrochemistry. When a molecule adsorbs on a metal electrode, its energy levels are shifted by the applied electrical potential. The Newns-Anderson model can be adapted to this situation beautifully, showing how the adsorbate's projected density of states broadens, shifts, and changes with the applied voltage. This provides a quantum-mechanical picture of the electrochemical double layer—the microscopic region where all the action happens in batteries, fuel cells, sensors, and corrosion.

​​Organometallic Chemistry:​​ Long before surface scientists had these quantum tools, organometallic chemists developed an intuitive and powerful picture of how an alkene like ethylene binds to a single metal atom: the Dewar-Chatt-Duncanson model. It involves a synergistic dance of electron donation from the alkene to the metal and back-donation from the metal to the alkene. The Newns-Anderson formalism provides the solid-state physics counterpart to this molecular picture. It shows how the discrete orbitals of the molecule interact with the continuous bands of the solid, and it even allows us to calculate things like the partial electron population that flows into the alkene's antibonding $\pi^*$ orbital—the very act of back-donation that weakens the C=C bond and enables its transformation. It unifies the language of discrete molecular orbitals with the language of continuous bands.

​​Atomic and Surface Physics:​​ Let's end with a dynamic picture. Imagine a positively charged ion hurtling towards a metal surface. During the fleeting picoseconds of its approach and departure, does it have enough time to snatch an electron from the metal and emerge as a neutral atom? The probability of this charge transfer event depends on the rate, $\Gamma$, at which an electron can tunnel from the surface to the ion. This rate, it turns out, is determined by the same physics of hybridization that underlies chemisorption. In the language of our model, the rate $\Gamma$ is directly proportional to the "level broadening" $\Delta$, the imaginary part of the self-energy. Thus, the same model that describes the static strength of a chemical bond can also be used to calculate the survival probability of an ion in a dynamic scattering event.

From the design of industrial catalysts to the quantum description of a battery, and from the bonding in a single molecule to the fate of a scattered ion, the simple, elegant physics of a discrete state interacting with a continuum provides a powerful and unifying thread. It is a stunning reminder that in the intricate tapestry of nature, the most complex patterns are often woven from the simplest of principles.