DOE OSTI2011
In their comment, Chen et al. try to argue that the experimentally observed controllable voltage-induced surface modification, which was attributed to a local electric field-induced atom transfer from the surface to the tip, is rather caused by either an oxidation process and/or a resistance change. In this response, we will show that we can rule out these two effects in our experiment. The statements by Chen et al. are based on two arguments: (1) the tip modification after transferring an adatom should alter the dI/dV contrast, which was not seen in our experiments and (2) the vacuum conditions in our experiment are similar to earlier reports on resistance switching. First, Chen et al. discuss that the adsorption on the tip should alter the topographic contrast, as many papers have reported. In fact, in our experiments we frequently observed tip modifications at high bias voltage. These typically result in slight changes in scanning tunneling spectroscopy data [see, for example, the spectra in Fig. 3(b) in Ref. 4 and Fig. 2(d) of Ref. 5] but only weakly affected the topographic contrast. Second, Chen et al. claim that oxidation is another possible mechanism to explain our experimental observations. To support this claim, they compare our results to an earlier publication showing resistance switching. In fact, the resistance switching mechanism is related to oxygen vacancy migration or local surface oxidation. The mechanism of oxygen vacancy migration requires a 'forming' process with a threshold current in the order of microampere or even milliampere. In our experimental setup, however, we used tunneling currents in the order of 50 pA. Even during surface modification, which was performed at open feedback loop conditions with voltage pulse of up to 3 or -5 V, the maximum transient current did not exceed a few nanoampere. Therefore, we can safely exclude oxygen vacancy migration as a potential mechanism for the observed surface modification. As a second potential mechanism Chen et al. mention a local surface oxidation process. However, the total pressure at high-vacuum conditions used in experiments, where resistance switching was observed (10{sup -7} torr in Ref. 3) is three order magnitude higher than in our experiment performed under ultrahigh vacuum (UHV) conditions (below 10{sup -10} torr). Furthermore, mass spectra measured with a residual gas analyzer show that the main residue gas in our UHV system is hydrogen ({approx} 90%). Water, oxygen, and other oxygen-related gases are negligible with a partial pressure in the order of 10{sup -12} torr range or lower. Therefore, we can also exclude that local oxidation with reactants from the residual gas causes the observed modifications. In addition, in our experiment, the refilling of the modified areas at negative bias could not be observed with fresh tip, even for bias voltages as high as -10 V. In short, the mechanism for the modification on the UHV in situ fractured Nb:SrTiO{sub 3} (Nb-doped Strontium titanate) surfaces with scanning tunneling microscope (STM) tip is different from the mechanisms such as local surface oxidation or filament formation, used to explain the largecurrent induced resistance switching works.