Stoichiometry engineering of ternary oxide ultrathin films: ${\mathrm{Ba}}_x{\mathrm{Ti}}_2{\mathrm{O}}_3$ on Au(111)

Ternary oxide ${\mathrm{Ba}}_x{\mathrm{Ti}}_2{\mathrm{O}}_3$ ultrathin films on Au(111) substrates $(x\le 2/3)$ have been studied using a joint experimental and theoretical approach, including the use of scanning tunneling microscopy (STM), first principles calculations, and Monte Carlo simulations. The films are created by first covering the Au(111) substrate with a ${\mathrm{Ti}}_2{\mathrm{O}}_3(2\times 2)$ honeycomb (HC) network and then evaporating low concentrations of Ba atoms onto this film. STM imaging shows that the Ba atoms adsorb individually in the hollow sites of the HC network. Depending on the Ba coverage $x$, which ranges from 0 to 2/3, two ordered phases can be identified at $x=1/3$ and $x=2/3$. A disordered labyrinthlike phase is observed for values of $x$ between 1/3 and 2/3. Theoretical modeling shows that the structural character of these films is driven by the charge transfer that occurs between the electropositive Ba atoms and the electronegative ${\mathrm{Ti}}_2{\mathrm{O}}_3$/Au substrate. This results in a number of calculated effects including an increase in the film rumpling and a reduction of the film work function with increasing $x$. The evolution of the structure of the thin films as a function of Ba coverage can be described by a lattice gas model with first-, second-, and third-neighbor Ba-Ba repulsive interactions. The range of the dipolar interactions is a key factor in understanding the behavior of Ba ordering. The structural and electronic flexibility, which can be engineered through stoichiometry, temperature, or support control, makes these ultrathin films promising materials for applications related to adsorption or reactivity, or as template supports for the growth of size-selected clusters.

Figure 1

(a) STM image of the bare honeycomb ${\mathrm{Ti}}_2{\mathrm{O}}_3$ monolayer on Au(111) ($V_s=0.98$ V, $I_t=0.2$ nA, image size $2.9\times 2.9$ ${\mathrm{nm}}^2$). (b) STM image of a single Ba atom adsorbed on the honeycomb lattice ($V_s=0.72$ V, $I_t=0.2$ nA, image size $2.9\times 2.9$ ${\mathrm{nm}}^2$). (c) Top view ball and stick representation of a single Ba atom adsorbed on the ${\mathrm{Ti}}_2{\mathrm{O}}_3$/Au support, and (d) side view. Au, Ti, O, and Ba atoms are represented by yellow, blue, red, and green balls, respectively.

Figure 2

STM images (top panels) and Monte Carlo simulation snapshots obtained for $k_{B}T/J_1=0.17$, $J_3/J_2=0.666$, and $J_2/J_1=0.21$ (bottom panels). From left to right, the Ba coverage is $x=0.07$, 0.29, 0.35, 0.43, 0.49, 0.62. Only the adsorption sites are represented, which are dark red when empty and yellow when occupied. Images sizes are $10\times 10$ ${\mathrm{nm}}^2$, and tunneling conditions for the STM images are ${\mathrm{V}}_s$ ranging from 0.35 to 1.00 V and ${\mathrm{I}}_t$ ranging from 0.18 to 0.28 nA.

Figure 3

Variation of the mean number of first (top panel) and second (bottom panel) nearest neighbors as a function of Ba concentration $x$. Full lines are results of Monte Carlo simulations obtained for $k_{B}T/J_1=0.17$, $J_2/J_1=0.21$, and $J_3/J_2=0.666$. Black dots are experimental values extracted through STM image analysis.

Figure 4

Comparison of MC configurations at $x=1/2$ obtained with different sets of parameters: (a) $k_{B}T/J_1=0.17$, $J_2/J_1=0.21$, and $J_3/J_2=0.666$; (b) $k_{B}T/J_1=0.02$, $J_2/J_1=0.21$, and $J_3/J_2=0.666$; and (c) $k_{B}T/J_1=0.17$, $J_2/J_1=0.57$, and $J_3/J_2=0.87$ (see text).