【作者】 侯玉斌；

【导师】 陆轻铀；

【作者基本信息】 中国科学技术大学 ， 凝聚态物理， 2009， 博士

【摘要】 扫描隧道显微镜和原子力显微镜是扫描探针显微镜家族里应用最为广泛的两种显微镜。它们拥有原子分辨、原子搬运、纳米微加工等共同的功能性特征,但因二者的工作原理的不同,它们得到的结果所反映的样品表面信息是完全不同的。扫描隧道显微镜测量的是样品表面的电子态,它得到的结果反映的是样品表面的电子态分布的信息。虽然具有原子分辨率,但扫描隧道显微镜并不能测量到样品的真结构。而原子力显微镜测量的是探针原子与样品原子之间的相互作用力,因而它得到的图像反映的是样品表面真实的原子排列信息,即:样品的真结构。但,原子力显微镜不能测量到可与理论比较的电子态的信息。所以扫描隧道显微镜和原子力显微镜是互补的。目前世界上商业和自制的扫描隧道显微镜和原子力显微镜非常多,也已经很成熟,但绝大部分都需要在高电压(高于100V)的条件下才能工作,原因在于他们用于粗逼近(进针)的压电步进马达需要高压才能工作。使用高电压就必须要使用诸如高压运算放大器,高压三极管等高压元器件,这就带来了漏电流大、精度低、噪音大、温漂大等一些列的缺点,同时高电压的使用也带来了高成本和高不安全因素。而且这些在高电压下工作的压电步进马达大多体积比较大、结构与控制复杂,因而基于这些步进马达而设计的扫描探针显微镜镜体的体积就比较大,也比较复杂,这些因素都不利于将他们集成为一体以及植入到各种极端环境中。为了克服以上的缺点,我们在世界上首次提出了全低压扫描隧道显微镜的概念,发明了纵横转置横向隧道结调节技术,彻底的放弃了高压,从而避免了因高压的使用而带来的各种缺点与危害。传统技术利用的是压电扫描管的轴向压电位移(单位低压产生的位移值较小)来产生惯性步进(粗逼近),这就需要用高压才能产生足够大的惯性力,从而克服摩擦力,产生步进。为了能在低压(小于工业标准低压器件的电源电压15V)下实现粗逼近,我们利用压电扫描管切向位移远大于轴向位移的特点,把传统扫描探针显微镜的探针-样品结转90o(横过来)使用,在此纵横转置横向隧道结调节技术的基础上研制成功了能在低电压(4V)下行走的惯性步进马达,从而在世界上第一次实现了全低压下工作的扫描隧道显微镜(也适用于其它扫描探针显微镜),在室温大气条件下采用剪刀剪切的铂铱合金探针就获得了可与低温超高真空条件下使用电化学腐蚀的钨针相媲美的极清晰石墨原子图像(原始数据)。在我主持开发的纵横转置扫描隧道显微镜基础上,改进版(由庞宗强完成)的惯性马达能够在全低压的环境下进行二维方向的步进,基于此马达的扫描探针显微镜能够实现大范围无间隙原子精度的搜索成像(可搜索样品表面稀少但重要的缺陷、器件等)。纵横转置横向隧道结调节技术采用两个扫描管并列固定在同一个基座上,二者之间的间距在一个毫米以内。因为两个扫描管是完全相同的,因而在压电扫描管的轴向方向的热漂移得到极大的补偿(抵消),同时两个全同扫描管的并列站立的结构也补偿了外界的干扰对隧道结的影响。本结构中决定热漂移的因素是两个扫描管之间的基座,但我们采用的是蓝宝石基座,他的导热性能非常好,热胀系数也很小,同时两个扫描管的相邻两面之间的间距不到1mm,因而基座的热胀冷缩效应对隧道结的影响可以忽略,所以我们的全低压横向隧道结扫描隧道显微镜是高度热稳定的,非常适合于在变温的场合使用。更重要的是利用本技术制作的镜体的体积非常小,能很容易的植入到各种极端环境中。基于这种纵横转置横向隧道结调节技术,我们目前正在研制能植入到?52mm 20T超强磁体中的世界最强磁场扫描隧道显微镜。在扫描隧道显微镜领域,当前有文献可查的前置放大器的电流分辨率最高只有50fA,前置放大器分辨率太低限制了STM在不良导体或绝缘体领域的应用。我们自主开发的超低噪声超高精度二级联配去偏压互阻放大器,使探针接真地以减少传统探针接虚地带来的针尖电场干扰问题,并采用极高输入阻抗的二级去偏压测量方法,配之上面提到的全低压技术,得到了在扫描隧道显微镜领域内创纪录的10fA的电流分辨率,这也将有助于将扫描隧道显微镜向绝缘体领域拓展,对于利用扫描隧道显微镜进行不良导体(如生物样品)甚至绝缘体的研究有重要的意义。电路采用两级放大的形式,前后级电路通过电桥方式连接以去除输出信号中的偏压分量,从而得到正确的输出值。利用极高的二级输入阻抗来降低二级放大对一级信号的畸变,提高整个放大器的精度。全部的测量器件均为高精度的低压器件。我们拥有发明专利权的双通道差分抗干扰电流放大电路采用完全对称的电路形式,拥有极高的抗电磁干扰的能力,有效地压制共模信号的干扰,能在完全无屏蔽的情况下测量到皮安(pA)甚至飞安(fA)级的极弱电流信号,这在国际上也未见有过报道。在微弱信号测量领域有着广泛的应用,也能极大的提高扫描隧道显微镜的抗干扰能力。在控制器方面,我们(王霁晖负责开发,我参与调试)在PXI实时操作系统上使用Labview图形化编程语言自行开发出了控制扫描探针显微镜的控制系统。和商业控制器相比,自主开发具有灵活、易用的特点,而且所消耗的费用要远低于购买商业控制器。而且,Labview的简单易学,直观形象的特点,能够简化控制软件开发的难度,加快开发的速度。在纵横转置扫描隧道显微镜的基础上,我们目前正在研制低温超高真空调频原子力显微镜。其镜体主体和20T-STM大体上相似。我们放弃传统的采用激光来构建振荡环路的机制,采用基于石英晶体的“qPlus”探针,利用石英晶体的压电效应来构建振荡环路,同时研制成功了两种自动增益控制电路,实现了“qPlus”探针的高稳定振荡。在频率测量方面,我们(由同组成员施易智建成,用于我负责建造与调试的AFM上)开发出了频率分辨率为3mHz的锁相环(PLL),优于世界著名的商业产品easyPLL。目前整个AFM系统已经搭建完毕,已经进入液氦调试阶段。虽还没有获得原子分辨率图像,但已经获得了正确而干净的原子力梯度与探针-样品间距的关系曲线。鉴于原子级别的测量对减振、隔振的超高要求,我们参考引力波测量的超级减振装置,自行设计了一种多层纵横向减振与多级悬吊减振系统。它由较重的钢砖和减振橡胶球层叠而成,每层的橡胶球都具有纵横向减振的功能,桌面由穿过钢砖叠层的吊绳吊起,多层纵横向减振效果的叠加,达到极好的减振效果。橡胶球是根据我们自己设计的配方加工的,具有很高的阻尼,低回弹率(5%)的优点,减振效果非常好。我们的显微镜就安置在这个桌面上,也就包含了悬吊减振的功能。就这样,在采用上述自主研发的多项新技术之后,我们成功地研制出国际首个全低压和首个10飞安分辨率扫描隧道显微镜(见附录中的官方查新结果),该显微镜在室温大气环境下就显示出了极高的稳定性和精度,获得了极清晰的石墨原子图像(同等测量条件下未见其它比之更清晰的原子图像)。更多还原

【Abstract】 The scanning tunneling microscope (STM) and the atomic force microscope (AFM) are the most widely applied two types of microscopes in the scanning probe microscope (SPM) family. Although they both possess the capabilities of atomic resolution, atomic manipulation, and nano-fabrication, their working principles are different, leading to different surface information they can obtain from the sample. What an STM measures is the surface density of states (DOS) of the electrons in the sample. It has atomic resolution, but an STM can not obtain the true atomic structure of the sample.However, what an AFM measures is the information of the interaction forces between the tip and the sample, hence the AFM images will reflect the true atomic arrangement of the sample surface, that is, the sample’s true structure. An AFM is nevertheless unable to obtain the DOS information which is important in the sense it can be compared with theory. Thus, the STM and AFM are complementary.Currently, there are many commercial and home-made STMs and AFMs in the world, which are very mature too. But, they need high voltages (higher than 100V) to operate, because the piezoelectric motors they use for the coarse approach need high voltage to work. Using high voltages means high voltage devices must be used such as high voltage operational amplifiers and transistors, which brings the severe issues of large leakage currents, low precisions, big noises and drifts, etc. In the meanwhile, the use of high voltages also leads to high costs and safety problems. And, the high voltage piezoelectric motors are in general large in size and complicated in structure and operation, causing the STM head to be big and complicated, which does not favor a highly integrated STM and makes it difficult to put the STM into extreme physical conditions.To solve the above problems, we have, for the first time, proposed the concept of“fully low voltage STM”and invented the 90o-flipped lateral junction regulation technology, in which high voltages are completely abandoned, thus avoiding all the drawbacks and issues associated with use of high voltages. As we know, traditional STMs utilize the axial displacement (small for unit applied voltage) of a piezoelectric tube to generate inertial stepping for coarse approach, which needs high voltage to produce an inertial force sufficiently large to overcome the friction force. To implement the coarse approach (inertial stepping) under low voltages (lower than the industrial standard±15V power supply voltages for low voltage devices), we take advantage of the fact that the lateral displacement of a piezoelectric tube scanner is much larger than its axial displacement, and flipped the traditional axially regulated junction by 90o so as to regulate the tip-sample gap laterally (by the scanner’s lateral displacement). Based on this technology, we have successfully built a piezoelectric inertial motor that can step move under 4V low voltage, and realized the first fully low voltage STM (applicable to other types of scanning probe microscopes) in the world. Using just hand cut Pt/Ir wire as the STM tip, we have obtained extremely clear atomic resolution images (raw data) of graphite in air and room temperature with the image quality being comparable with the images obtained from a low temperature ultrahigh vacuum STM with a chemically etched tungsten tip. Based on the 90o flipped junction STM of which I am in charge, Zongqiang Pang has also built the inertial motor that can step move in two dimensional plane, which gives rise to an STM that can search scarcely distributed defects (or atoms or devices) across the entire sample surface by taking atomic resolution images continuously (gapless) while the tip travels across the sample surface.Here is how the 90o flipped junction STM works: two tube scanners are mounted on the base in parallel with the gap between the scanners being less than 1 mm. Because the two scanners can be identical, the thermal drifts in the axial direction will cancel, and other external interferences (such as vibrations) will cancel too, resulting in a very stable junction structure. The thermal drift (shrinking or expansion) of the base may impact the junction gap, this impact is small since the two scanners are very close to each other (the gap is less than 1 mm) and can be even further reduced by using low drift material to make the base (such as sapphire as we used). So, the junction structure in our STM is highly stable, which is well suited to temperature varying conditions. Furthermore, the STM is very compact and can be put in various extreme conditions. This technology is currently being used in building our ?52mm 20T STM which will be the strongest magnetic field STM in the world.For the STM, the highest measurement resolution of the tunneling current that can be found in published papers is 50fA. This low resolution has limited the applications of STM in studying low conducting or insulating samples. To this end, we have invented and developed a ultra-low noise ultra-high precision two-stage bias deducted trans-impedance amplifier (ULN-UHP-TIA), in which the tip is truly grounded (thus reducing the stray field from the tip) and the bias removal from the output of the first stage is done by a second stage subtractor with extremely high input impedance (thus minimizing the signal distortion). Combined with the above mentioned 90o flipped junction STM, we have achieved a record breaking 10fA current resolution. This will help expand the STM into the measurements of weak conductors (such as biological samples) or even insulatorsWe have also patented an interference-resistant double channel differential current amplifying circuit which is a symmetric circuit with very high capability of canceling and reducing electromagnetic interferences (EMI). It can detect picoamp (even femptoamp) weak current, which is not seen published anywhere else. It may have broad applications, especially in enhancing STM’s ability to withstand EMI. For the controller, we (in charge by Jihui Wang and I have participated in the testing) use Labview programming language in the PXI real time system to develop the SPM controlling system. Compared with the commercial controller, ours is easier, more versatile and much cheaper, making the software development simplified significantly.On the basis of our lateral junction regulation method, we are currently constructing a frequency modulated AFM (FM-AFM) which is as compact as the above mentioned 20T-STM. We have abandoned the conventional reflected laser scheme for force sensing. In stead, we adopt the qPlus quarts tuning fork as the force sensor, which uses the piezoelectric effect to detect the bending of the fork. We have tried two different automatic gain control (AGC) circuits in the oscillation circuit, both leading to a very stable oscillation of the qPlus probe. To convert the oscillation frequency into a voltage signal, we have successfully developed a 3 mHz frequency resolution phase locked loop (PLL, built by Yizhi Shi and tested by me on my AFM), which is better than the commercial 5 mHz resolution Nanosurf EasyPLL. At present, the whole AFM is already completed and is being tested in liquid helium environment. Although we have not got the atomic resolution images yet, we have nevertheless measured the correct force gradient vs. tip-sample distance curves.To better isolate the vibration influence, which is critical in achieving atomic resolution, we have referenced some of the vibration isolation techniques used in gravity wave detection and designed a multi stage damped long pendulum vibration isolation platform. It consists of stacks of alternate steel block layer and damping rubber ball layer with the table top being suspended from the tops of the steel block stacks. We have also designed the rubber ball recipe ourselves. The rebound rate of the rubber balls is less than 5%. The final vibration isolation result is rather excellent. As the consequence of exploiting the above quite a few techniques we develop, we are successful in implementing the world’s first all low voltage and better than 10 fA current resolution STM (the result of an official novelty search is provided in the appendix), which has demonstrated very high stability and precision and has produced extremely clear atomic resolution images (raw data) for graphite samples in ambient conditions (comparable or better quality images have not seen previously under the same measurement conditions).更多还原