本論文擬結合頻率分析技術來開發進階功能的原子力顯微鏡檢測模式，包含靜電力影像與磁力分布的分析應用。傳統上，要進行表面電位的量測主要是使用閉路式表面電位顯微鏡(Close Loop Kelvin Probe Force Microscope, CL-KPFM)的技術進行量測，這一個技術是需要一直施加一個直流偏壓在樣品上來進行回饋，並藉此得到表面電位。但有一些樣品容易受到直流電壓的影響造成樣品的形貌或者物理特性改變，或有時候會需要將樣品放在液態的環境之中進行實驗，此時就無法進行閉路式表面電位顯微鏡進行量測。因此，這種時候就需要使用開路式表面電位顯微鏡(Open Loop Kelvin Probe Force Microscope, OL-KPFM)技術，這一種量測手法不需要施加直流偏壓進行回饋來得到表面電位。本次實驗中就將開發OL-KPFM技術中的雙諧波表面電位顯微鏡(Dual Harmonic Kelvin Probe Force Microscope, DH-KPFM)技術來進行實驗，DH-KPFM是透過量測交流電壓的一倍頻振幅、一倍頻相位以及二倍頻振幅來計算出表面電位。在量測到表面電位之後，就可以利用表面電位來進行功函數的量測。
在量測樣品磁性的時候，通常使用的技術傳統的磁力顯微鏡(Magnetic Force Microscope, MFM)技術來進行量測。在使用MFM進行量測時，第一趟掃描只能將樣品的表面形貌掃描出來，要在第二趟掃描時才會得到磁力訊號。傳統MFM技術得到的磁力訊號中，我們只能判斷這個磁力是相對吸引力或排斥力不能知道確切的磁力方向，並且量測到的磁力分布解析度和靈敏度都不高。因此，我們希望透過力曲線(Peakforce)配合機器學習(Machine Learning)來進行磁力訊號的量測和分析，因為只要先將探針進行磁化的動作，在量測時力曲線便會將探針受力情況記錄下來，包含表面的凡德爾力、黏滯力、磁力等等。在透過力曲線配合機器學習的方式分析出的力曲線進行傅立葉轉換後，利用頻譜技術便可分析出每一個不同的力曲線對應到凡德瓦力、黏滯力、磁力等物理特性，藉此了解在力曲線中各自對應的作用力機制。

In this paper, we combine frequency analysis technology to develop advanced functional atomic force microscope modes, including the analysis and application of electrostatic force images and magnetic force distribution. Traditionally, the measurement of surface potential is mainly to use Close Loop Kelvin Probe Force Microscope (CL-KPFM) technology for measurement. This technology requires that a DC bias is always applied to the sample for feedback and obtain the surface potential. However, some samples are susceptible to changes in the surface or physical characteristics due to the influence of DC voltage and sometimes the sample needs to be placed in a liquid environment for experiment. At this time, CL-KPFM cannot be used for measurement. Therefore, the Open Loop Kelvin Probe Force Microscope (OL-KPFM) technology that does not apply a DC bias for feedback to obtain the surface potential is required. In this experiment, we will develop one kind of OL-KPFM technology called Dual Harmonic Kelvin Probe Force Microscope (DH-KPFM). DH-KPFM can obtain surface potential by measuring the amplitude and phase of the AC voltage frequency and the amplitude of twice the AC voltage frequency. After the surface potential is measured, the surface potential can be used to measure the work function.
When we measuring the magnetic properties of a sample, the traditional Magnetic Force Microscope (MFM) technology is usually used for measurement. In MFM technology, the first time of scanning can only scan the surface topography of the sample, and the magnetic signal will be obtained in the second scan. The magnetic signal obtained by the traditional MFM technology, we can only judge whether the magnetic force is a relative attractive or repulsive force, but the exact direction of the magnetic force cannot be known, and also the resolution and sensitivity of magnetic force distribution is not high enough. Therefore, we use the Peakforce with machine learning to measure and analyze the magnetic signal. Because as long as the probe is magnetized first, the force curve will be affected by the probe during the measurement and will be recorded, including the Van Der Waals force, adhesion and magnetic force. After Fourier transformation is performed on the force curve that analyzed by Peakforce and machine learning, the spectrum of frequency can be used to analyze the physical characteristics of each different force curve corresponding to Van Der Waals force, adhesion and magnetic force. This will help understanding the corresponding force mechanism in the force curve.

1. Collins, Liam, et al. "Open loop Kelvin probe force microscopy with single and multi-frequency excitation." Nanotechnology 24.47 (2013): 475702.
2. Kilpatrick, Jason I., et al. "Quantitative comparison of closed-loop and dual harmonic Kelvin probe force microscopy techniques." Review of Scientific Instruments 89.12 (2018): 123708.
3. Kobayashi, Naritaka, Hitoshi Asakawa, and Takeshi Fukuma. "Quantitative potential measurements of nanoparticles with different surface charges in liquid by open-loop electric potential microscopy." Journal of Applied Physics 110.4 (2011): 044315.
4. Kobayashi, Naritaka, Hitoshi Asakawa, and Takeshi Fukuma. "Nanoscale potential measurements in liquid by frequency modulation atomic force microscopy." Review of scientific instruments 81.12 (2010): 123705.
5. https://en.wikipedia.org/wiki/Work_function
6. Seo, Hosung, Dan Goo, and Gordon Jung. "How to obtain sample potential data for KPFM measurement."
7. Melitz, Wilhelm, et al. "Kelvin probe force microscopy and its application." Surface science reports 66.1 (2011): 1-27.
8. 姜政熙，鎳鐵氧/鈦酸鍶鋇/鈦酸鍶磊晶薄膜之磁電藕合特性，成功大學碩士論文(2008)。
9. B. M. Moskowitz, Hitchhiker's Guide to Magnetism (1991).
10. Agar, J. C. et al. Machine Detection of Enhanced Electromechanical Energy Conversion in PbZr0.2Ti0.8O3 Thin Films. Adv. Mater. 30, 9, doi:10.1002/adma.201800701 (2018).
11.https://mmrc.caltech.edu/AFM%20Dimension%20Icon/Bruker%20Training/004-1036-000%20PEAKFORCE%20QNM%20USERS%20GUIDE-F.pdf
12.https://mmrc.caltech.edu/AFM%20Dimension%20Icon/Bruker%20Training/Peak%20Force%20QNM_Adv%20Apps%20Training%202014.pdf
13. 王洸富，屏蔽電荷對108度域壁成核動態機制之影響，成功大學碩士論文(2010)。
14. 曾賢德、果尚志，奈米電性之掃描探針量測技術，物理雙月刊廿十五卷五期(2003)。
15. Magnetic Force Microscopy (MFM) Applicable to Dimension™ Series and MultiMode™ Systems, Digital Instruments, 1996.
16. https://stats.stackexchange.com/questions/259806/anomaly-detection-using-pca-reconstruction-error