The Microscopic Mechanism of 1/f Fluctuations in the Tunneling Current of Scanning Tunneling Microscopes

通过研究超高真空中铂Ir针尖与Au膜或晶态石墨的组合以及W针尖与石墨样品的组合，实验研究了扫描隧道显微镜隧道电流中观察到的显著低频噪声(包括1/f波动)的产生机理。对于1/f噪声，固定电流下的噪声强度随着偏置电压的增加而增大。在Au膜的样品表面上，噪声功率因位置而异。噪声的空间变化与z调制法测量的隧道势垒高度Phi的大小成正相关。在z调制实验中解调的信号也表现出相对幅度的1/f谱与当前噪声中几乎相等的波动。能够以原子分辨率成像石墨表面的TIPS有时会产生带有洛伦兹光谱的随机电报噪声，这是很好地理解的，因为考虑到Shar…更多的p尖端选择了一个恰好位于窄电流路径上的单一驰豫过程。在超高真空中测量了扫描隧道显微镜隧道电流1/f起伏的温度依赖关系。在65K到室温的温度范围内，噪声谱表现出1/f^a型的频率依赖关系，α变化在1左右。100赫兹处的噪声强度随温度而变化，呈现两个可辨别的峰值。后者与声子数涨落模型预测噪声强度随温度单调增加的理论预期相矛盾。实验结果可以用假设热激活多重松弛过程的模型来很好地解释。1014赫兹量级的频率因子与实验数据符合得很好，表明1/f噪声的来源是一些散布在远离探头尖端的样品表面的移动“缺陷”。较少

The generation mechanism of conspicuous low frequency noise (including 1/f fluctuations) observed in the tunneling current of scanning tunneling microscopes was studied experimentally by investigating combinations of Pt-Ir tips and Au films or crystalline graphite and of W tips and graphite samples in ultra high vaccum. For 1/f noise, the noise intensity at a fixed current increases with increasing bias voltage. The noise power varies from place to place in a sample surface of Au film. The spatial variation of the noise has a positive correlation with the magnitude of the tunneling barrier height phi measured by the z-modulation method. The demodulated signal in the z-modulation experiments also exhibits fluctuations with a 1/f spectrum of the relative magnitude nearly equal to that in the current noise. Tips capable of imaging graphite surface in atomic resolution sometimes yield random telegraphic noise with a Lorentzing spectrum, which is well understood by considering that the shar … More p tips select a single relaxation process that happens to be located in the narrow current path.The temperature dependence of 1/f fluctuations in the tunneling current of scanning tunneling microscopes was measured for graphite smaples in ultra high vacuum. The noise spectra in the temperature range between 65K and room temperature show the frequency dependence of 1/f^a type with alpha ranging around unity. The noise intensity at 100 Hz varies with temperature exhibiting two discernible peaks. The latter fact contradicts with the theoretical expectation from the phonon number fluctuation model which predicts monotonic increase of noise intensity with temperature. The experimental results are best explained by the model assuming thermally activated multiple relaxation processes. The frequency factor of the order of 1014 Hz which yields a good fit with the experimental data suggests that the origin of the 1/f noise is some mobile "defects" dispersed over the sample surfce at positions remote from the probe tip. Less