Logo IMG
HOME > PAST ISSUE > July-August 1998 > Article Detail


Döbereiner's Lighter

Roald Hoffmann

How Does Hydrogen Burn?

One hundred seventy-two years later, in 1995, Laurens K. Verheij and Markus B. Hugenschmidt write:

In recent years many studies on the reaction between hydrogen and oxygen on metal surfaces have been reported. Although this reaction is expected to be one of the simplest oxidation reactions, rather complex phenomena are observed which make a determination of the reaction mechanism difficult. Even for the water formation reaction on Pt(111), the system which has been most widely studied, an understanding of the reaction process seems only just emerging.

Here is a sketchy summary of what modern surface chemistry, with its expensive arsenal of ingenious spectroscopies, has taught us about the water-forming reaction on clean single-crystal surfaces of platinum under high vacuum: Both O2 and H2 undergo "dissociative chemisorption" on a platinum surface (under the conditions of Döbereiner's lighter the H2 would impinge on a platinum surface already exposed to the oxygen of the air). The following reactions summarize what happens:

O2 + Pt → Pt-O2 → Pt-O (a)

H2 + Pt → Pt-H (a)

Here the (a) stands for an atomic or molecular species adsorbed on the surface. And the notation hides much interesting detail: For instance, is there a "precursor state" in which H2 molecules bond to the surface before the H–H bond is broken? Where exactly, and in what geometry, does the H atom sit on the surface—is it above a single platinum atom, or bonded to two or three platinum atoms? The same questions may be asked for O2.

We know some, only some, of the answers. At low temperatures, way below room temperature, the O2 is bonded to the surface first as a molecule, and in no less than three different ways. As one heats up the surface, the diatomic (O2) ruptures into individual oxygen atoms, which sit bonded to triangles of platinums. At ambient temperatures, it is not likely that an O2 coming onto the surface survives very long before it breaks apart. The hydrogen molecules break apart even more readily on the surface. A consensus is emerging that on the surface, the Pt-bonded oxygen atoms and hydrogen atoms cluster, or form islands of Pt-O(a) and Pt-H(a).

Figure 3. Schematic of O<sub><a href=Click to Enlarge Image2 molecules and O atoms" float="LEFT" />

Recently, there appeared two beautiful papers illustrating the immense complexity of a step as seemingly simple as the breakdown of an oxygen molecule into individual oxygen atoms on a platinum surface. One paper comes from the group of Gerhard Ertl at the Fritz Haber Institute of the Max Planck Society in Berlin, the other from my colleague at Cornell, Wilson Ho, and his coworkers. Figure 3 shows an image from the latter group's paper of oxygen molecules and atoms adsorbed onto a platinum surface, made at atomic resolution with a scanning tunneling microscope (STM). The molecules are the light spots, and they come in two types— "pear like" (labeled F in the figure) and "clover-leaf" (labeled B) shapes. Please don't worry that the shapes don't look like the O2 dumbbell you expect. The STM technique does not really "see," for instead of sensing the position of the nuclei, it measures the flow of electrons from the STM instrument tip into available orbitals in the oxygen atoms or molecules. The orbitals (calculating them is my stock-in-trade) describe quantum mechanically where the electrons are. A clover-leaf pattern turns out to be exactly what one would expect—it confirms that these are O2 molecules. The darker individual circles (marked o) are oxygen atoms formed by molecules falling apart on the surface. Note the clustering of adsorbed molecules and atoms into islands, oriented along privileged directions of the platinum surface underneath.

Incidentally, one could have imagined that either H2 or O2 (but not both) first bond to the surface and break up. And then water might be formed directly, the Pt-H(a) or Pt-O(a) approached by a gas phase O2 or H2 molecule, respectively, without the latter bonding to the surface. This apparently does not happen; can you think of an experiment to test such a mechanism?

The next stage still remains a mystery. The most likely scenario is that on the surface a chemisorbed hydrogen atom diffuses out of its safe island, over to at Pt-O island, and reacts, forming a surface-coordinated hydroxyl Pt-OH(a). Subsequently, that species picks up a second hydrogen, and water is released. And the catalyst, Pt, gets to work on the next molecule.

Remarkable, isn't it, how old technologies link up to contemporary science? Actually behind Döbereiner's invention was cutting-edge science of its day—it was reported in several journals in Germany, France and England within weeks. And Berzelius, in a series of influential annual reports on the progress of chemistry, writes the year after:

From any point of view the most important, and, if I may use the expression, the most brilliant discovery of last year is, without doubt, that fine platinum powder has the ability to unite oxygen and hydrogen even at low temperatures.

One hundred seventy-five years later, we are still studying the intricacies of water burning. The wonder of platinum glowing in a stream of hydrogen remains. Imagine a contemporary surface chemist coming into an antique shop in Saxony, and feeling a magical connection, yet not knowing exactly why, to that mysterious object—a Döbereiner lamp in a display case.

© Roald Hoffmann

comments powered by Disqus


Subscribe to American Scientist