Since the inception of utilizing epitaxy to grow optoelectronic devices on the GaAs(001) surface in the early 1970's, both state-of-the-art theoretical and experimental techniques have been brought to bear to uncover its properties. Somewhat surprisingly, the atomic structure of the GaAs(001)-(2×4) reconstructed surface hasn't been determined until very recently (LaBella et al.,PRL:83:2989). An STM image with acompanying theoretically generated STM image which resolves all the atomic features of this surface is shown below.
Once a surface is carefully prepared it needs to be transferred in situ to the STM for imaging. Then filled state images need to be aquired at lower biases (~2.0 Volts) than what is typically used (~3 Volts). This is due to the sample sharpening mechnism, where the top dimer rows "hide" the structure within the trench from the STM tip at high biases. At lower biases, the trench widens out and the structure become visible. This is displayed in the series of images below.
Notice in the series of images shown above that the trench structure (the single dimers between the rows) becomes visible as the bias is lowered. This effect is reproduced in the the first principles generated STM images. The figure below shows two STM images at 2.1 and 3.0 volts and two theoretically calculated STM images with different energy integration windows to simulate the different bias conditions. The theoretical images show the trench widening out as the bias is lowered.
This mechanism has been dubbed sample sharpening as depicted in the figure below. It has recently been utilized to uncover the atomic structure of the InAs(001)-(2×4) surface as well which has a similar corrugated surface reconstruction.
For more information See LaBella et al., Physical Review Letters 83 2989 (1999), and Barvosa-Carter et al., Surface Science Letters 499 L129 (2002).