Water splitting by tungsten oxide prepared by atomic layer deposition and decorated with an oxygen-evolving catalyst.

When sunlight is used as direct energy input, water can be split into hydrogen and oxygen at conversion efficiencies similar to those of solar cells. This process offers a method for energy storage to address the problem that the sun does not shine continuously, and is a particularly appealing approach to solar-energy harvesting. Notwithstanding the intense research efforts, progress in this area is extremely slow. Efficient and inexpensive water splitting remains elusive. A key reason for the sluggish progress is the lack of suitable materials. The “ideal” material must absorb strongly in the visible range, be efficient in separating charges using the absorbed photons, and be effective in collecting and transporting charges for the chemical processes. Such a material has yet to be found. The difficulties in finding a suitable material stem from the competing nature of intrinsic material properties (e.g., optical depth, charge diffusion distance, and width of the depletion region, among others), which leaves limited opportunity for tunability. We recently demonstrated that heteronanostructures, a type of nanoscale material consisting of multiple components that complement each other, have a combination of properties which are not available in singlecomponent materials. For instance, we can add chargetransport components to oxide semiconductors to solve the issue of low conductivity that oxide semiconductors generally suffer. In a similar fashion, one can add an effective catalyst to address the challenge that oxygen evolution is complex and tends to be the rate-limiting step. These new materials will likely lead to significant improvement in solar watersplitting efficiencies. The success of a heteronanostructure design relies on the ability to produce high-quality components with interfaces of low defect density, and on the availability of various components. Here we show that crystalline WO3 can be synthesized by the atomic layer deposition (ALD) method in the true ALD regime. When coated with a novel Mn-based catalyst, the resulting WO3 survives soaking in H2O at pH 7 and produces oxygen by splitting H2O under illumination. We choose ALD to prepare WO3 because of the following advantages: 1) a high degree of control over the resulting materials; 2) excellent step coverage to yield conformal coatings; and 3) process versatility to tailor the composition of the deposit. WO3 was studied because it is one of the most researched compounds for water splitting. The widely available literature makes it easy to compare our results with existing reports and thus allows us to test the power of the heteronanostructure design. To avoid the production of corrosive byproducts during the ALD process and to ensure the reaction occurs in the true ALD regime, we used (tBuN)2(Me2N)2W as tungsten precursor and H2O as oxygen precursor, as described in the Experimental Section (see Supporting Information for more details). Our first goal was to verify that the growth indeed takes place in the ALD regime. The dependence of the growth rate on the precursor pulse times and on the substrate temperature unambiguously confirms this. In addition, the excellent linear dependence of the deposition thickness on the number of precursor pulses supports the ALD growth mechanism and shows the extent of control we can achieve (see Supporting Information). That a long H2O pulse time is necessary to initiate growth is a key finding of this work. Despite intentional strengthening of the oxidative conditions, as-grown WO3 exhibited a tinted color, indicating the existence of oxygen deficiencies, which was then corrected by an annealing step in O2 at 550 8C. The crystalline nature of the product is manifested in the highresolution (HR) TEM image in Figure 1a. We also synthesized WO3 on two-dimensional TiSi2 nanonets. [18,19] The uniformity and good coverage around the nanonet branches show that this deposition technique is suitable for the creation of heteronanostructures. Ready dissolution of WO3 in aqueous solutions with pH 4 is a significant challenge that impedes its widespread use. We sought to solve this problem by coating WO3 with an Mnbased catalyst. Derived from the Brudvig–Crabtree catalyst, this coating was prepared by thermally decomposing [(H2O)(terpy)Mn(O)2Mn(H2O)(terpy)](NO3)3 (terpy= 2,2’:6’,2’’terpyridine). Similar to the oxo-bridged dimanganese catalyst, the thermal decomposition product exhibits good [*] R. Liu, Y. Lin, S. W. Sheehan, Prof. Dr. D. Wang Department of Chemistry, Merkert Chemistry Center Boston College 2609 Beacon St., Chestnut Hill, MA 02467 (USA) Fax: (+1)617-552-2705 E-mail: dunwei.wang@bc.edu Homepage: http://www2.bc.edu/~dwang