PerspectiveChemistry

Atomic Layer Electrodeposition

See allHide authors and affiliations

Science  07 Dec 2012:
Vol. 338, Issue 6112, pp. 1300-1301
DOI: 10.1126/science.1231853

The growth of ultrathin films is generally hindered by roughening and three-dimensional mound formation. Atomic layer deposition (ALD), in which atomic layer control and conformal growth are achieved through sequential, self-limiting surface reactions (1), can eliminate or reduce such roughening. One application of ALD is to deposit ultrathin layers of expensive metals such as Pt that are used, for example, as the catalyst in protonexchange membrane fuel cells (2). Besides the economic incentives to produce ultrathin films (3), there are also scientific payoffs—they often have catalytic, electronic, or magnetic properties that are not found in the bulk material (46). Although ALD processes are usually conducted in the vapor phase, Liu et al. (7) show on page 1327 of this issue that they can sequentially electrodeposit two-dimensional Pt layer by layer by simply pulsing the applied electrochemical potential in a single plating bath. The process is inexpensive and rapid. Because each layer is produced by cycling the potential rather than by exchange of reactants, electrochemical ALD could be orders of magnitude faster than vapor-phase ALD.

Electrodeposition is a bottom-up processing method, because the solid is assembled from ionic or molecular precursors in solution. It is similar in many ways to biomineralization, because solution additives and pH can be used to control the growth. Hence, the size, shape (8), crystallographic orientation, and even chirality (9) can be tuned. Compared with deposition from ultrahigh vacuum, electrodeposition is inexpensive, and the deposition rates can be much higher. It is not a line-of-sight process, so conformal films can be grown on complex shapes—for instance, in the onchip deposition of copper interconnects into submicrometer-sized features of semiconductor devices.

One step at a time.

Electrochemical atomic layer deposition of ultrathin Pt films (7) deposited one monolayer at a time by simply pulsing the electrode potential between ÷0.4 and −0.8 V. A capping layer of hydrogen is produced at −0.8 V that blocks the deposition of more than one monolayer of Pt. When the potential is stepped to ÷0.4 V, the hydrogen layer is desorbed and the cycle can begin again. The self-limiting processing method is fast because it is performed in a single plating bath, so it is not necessary to exchange reactants. The ultrathin Pt films could lower the costs of the Pt catalyst in fuel cells and provide a platform to study how the catalytic, electronic, and magnetic properties of ultrathin films evolve with film thickness (46).

CREDIT: C. SMITH/SCIENCE

What distinguishes electrodeposition from other deposition techniques is the applied potential, a single parameter that controls the departure from equilibrium and, therefore, the rate of the reaction. The electrodeposition of metals requires only that the electrode potential be driven negative of the equilibrium potential. The difference between the applied potential and the equilibrium potential is called the over-potential. Because the electrode must be poised at a potential for deposition to occur, the substrate must be a conductor or semiconductor. In contrast, vapor-phase ALD does not require a conducting substrate.

It is also possible by underpotential deposition (UPD) to produce a highly ordered monolayer or submonolayer of a metal at potentials positive of the equilibrium potential (10, 11). The substrate can be single crystalline or polycrystalline. UPD occurs because the binding of the monolayer to the foreign substrate is stronger than the binding of the monolayer to a substrate of the same material. This phenomenon is a surface-limited reaction, because only a monolayer will be deposited, regardless of how long the UPD potential is held. The self-limiting action of metal UPD allows electrochemical deposition to function as an ALD method. Electrochemical ALD has been used, for instance, to grow compound semiconductors by sequentially depositing each element in a UPD cycle (12).

Previous work on the atomic layer deposition of metals used a process called surface-limited redox replacement to produce ultrathin layers of metals such as Pt, Pd, and Ag (13, 14). The atomic layer deposition is realized by galvanic replacement of underpotentially deposited metal monolayers of less noble metals such as Pb or Cu. The surface-limited redox replacement occurs spontaneously, because the reduction potential of the less noble metal is more negative than that of the subsequent ALD layer. The two deposition steps in the surface-limited redox replacement of metals must be performed in separate solutions, so it is necessary to exchange reactants, much like the case of ALD by vapor deposition methods.

Conventional wisdom would suggest that the best way to electrodeposit ultrathin metal films would be to apply either an underpotential or a very small overpotential. Liu et al. report the surprising result that a monolayer of Pt is deposited at an overpotential of 1 V, a value at which the deposition rate of Pt should be very large. At this high overpotential, a monolayer of hydrogen is formed on the Pt surface, which completely blocks the deposition of additional Pt, thereby making the process self-limiting. Liu et al. attribute this blocking to the disruption of the electrical double layer. The trick to producing multilayers of Pt comes from the fact that the hydrogen that is deposited at −0.8 V can be desorbed at ÷0.4 V. By pulsing the potential between −0.8 and ÷0.4 V in a single plating bath, Pt is deposited on the surface monolayer by monolayer (see the figure). The deposition of a Pt monolayer is fast (complete in 1 s) and may also lead to less carryover of contaminants that occur when the reactants are exchanged in the surface-limited redox replacement scheme.

If the self-limiting growth is observed in alloys of Pt, magnetism could be imparted in the films by incorporation of magnetic metals. It would also be intriguing to deposit compositionally modulated superlattices (15) one monolayer at a time. As other metals have well-defined hydrogen adlayers, the prospects of this as a general processing method are encouraging. What about deposition of materials other than metals? Is there some potential-controlled process that can make the growth of metal oxides (8, 9, 15) or semiconductors (12) self-limiting? Ultrathin films deposited layer by layer onto single-crystalline substrates should be epitaxial. Electrochemical ALD would be ideal for in situ studies of the effect of lattice mismatch on the transition from two-dimensional to three-dimensional growth in epitaxial films. The beauty of this new electrochemical route to ALD is that it blends basic electrochemistry and surface science to unlock an important new technology.

References

Stay Connected to Science

Navigate This Article