Precision Measurement and Control Laboratory - Research

High-speed Atomic force microscopy for video-imaging and its application
Younkoo Jeong, Jayanth G.R, and Chia-Hsiang Menq

Introduction:

Since its introduction twenty years ago, Atomic Force Microscopy (AFM) has been used in a wide range of technologies with its superior capabilities not only to image the surface variation in sub-nanometer resolution, but also to measure the force at pico-newton scale. Among the two most commonly used modes, the dynamic (tapping) mode has a great advantage over contact mode when imaging soft materials. The most commonly used control method in the dynamic mode is a form of amplitude modulation, in typical implementations of which, due to tapping dynamics and limited bandwidth of the z-positioner, the transient response of the cantilever induced by changes of the tip-sample interaction force leads to greater variations in tip-sample interaction via feedback, causing excessive tapping forces and/or possible loss of tapping during scanning, and thus greater sample distortions and imaging errors. Therefore, while dynamic mode AFM can have many potential applications, the inability to achieve precise control of the tip-sample interaction forces has been one of the key barriers that limit imaging rate and that inhibit innovation leading to new applications. Moreover, the z-positioning loop consists of several elements that limit its response speed, and thus imaging rate.

Objective:

The objective of this research project is to investigate the design, actuation and control of a new generation AFM probing system and to create an innovative high-speed probing system that enables high-resolution and high-speed imaging upto video frame rate of samples.

Approach & Results:

1. Direct tip-sample interaction force control: Direct tip-sample interaction force control method is proposed, in which the tip-sample interaction force of each tapping cycle is directly regulated during scanning. Based on on a linear dynamic model of the cantilever along with its transient response, an estimator is designed and implemented to estimate the tip-sample interaction force of each tapping cycle as shown in Figure 1. The estimated interaction forces are then utilized by a model-based predictor to plan and control the next tapping by controlling the tip-to-sample distance. In order to attenuate the effects of modeling errors of the predictor, a feedback regulator is employed.

Since the tip-sample interaction force of each tapping cycle is directly controlled, it does not rely on the steady-state relationship between the oscillation amplitude of the cantilever and the tip-to-sample distance. Therefore, tapping dynamics in amplitude modulation is irrelevant and the time delay as the result of averaging in oscillation amplitude measurement is eliminated. Consequently, precise control of the tip-sample interaction force and high imaging rate can be achieved, independent of the quality factor Q of the cantilever. One example of the direct tip-sample interaction control is shown in Figure 2.

2. High speed tip position control: a dual-actuator tip-motion control scheme: In the dual actuator scheme, an additional magnetic mode actuator (as shown in Figure 3) is employed to achieve high bandwidth tip-motion control while the regular z-scanner provides the necessary motion range. This added actuator serves to make the entire cantilever bandwidth available for tip positioning and thus for the control of tip-to-sample distance. This block diagram of dual-actuator tip-motion control scheme is shown in Figure 4.

Two experiments were conducted to illustrate the capabilities of the dual-actuator tip-motion control in terms of response speed and travel range. In Figure 5, tip was commanded to move up by 30 nm. It is seen that while the tip motion is a rapid step response, the deflection caused by the magnetic actuator decreases gradually to zero when the z-motion achived by the z-scanner increases to 30nm. The decreasing rate is dictated by the bandwidth of the low-pass filter L(s) (Figure 4) . In the second experiment, the tip was commanded to ramp up from its initial position by 1000 nm in 10 msec and stay at the final position thereafter. The result is shown in Figure 6. It was shown that the controlled tip position (Figure 6 (c)) was able to follow the commanded input precisely while the z-scanner contributed to the large ramp and the magnetic actuator helped the tip with sharp movement.

3. HIgh speed real-time digital controller: A fast programmable electronics board was employed to realize the proposed dual-actuator control scheme, in which model cancellation algorithms were implemented to enlarge the bandwidth of the magnetic actuation and to compensate the lightly damped dynamics of the cantilever as shown in Figure 7.

4. High spatial frequency AFM imaging: The dual-actuator tip-motion control scheme was applied to image purple membrane (Halo-bacterium) patches to illustrate its ability to track high spatial-frequency surface topography when increasing imaging speed. The purple membrane patches were deposited on a freshly cleaved mica surface for 10 min. in an absorption buffer and rinced three times gently with imaging buffer before the scanning. A purple memgrane (90nmx90nm) was imaged with the regular z-scanner using amplitude modulation model while a larger membrane patch (120nmx120nm) was measured with the dual-actuator using direct force control. The results are shown in Figure 8. At the imaging rate of 20 lines/sec, the image obtained with the regular z-scanner did not show any details of the topography except the slope, (a) , while the image with the dual-actuator showed the repeating bacteriorhodopsin structures clearly, (b).

Publications

[1] Y. Jeong, G. R. Jayanth, and C. H. Menq, 'Direct tip-sample interaction force control for the dynamic mode atomic force microscopy', Applied Physics Letters, Vol. 88, 204102 (2006)

[2] G. R. Jayanth, Y. Jeong and C. H. Menq, 'Direct tip-position control using magnetic actuation for achieving fast scanning in tapping mode atomic force microscopy', Review of Scientific Instruments, Vol. 77, 053704 (2006)

[3] Y. Jeong, G. R. Jayanth and C. H. Menq, 'Control of tip-to-sample distance in atomic force microscopy: A dual-actuator tip-motion control scheme', Review of Scientific Instruments, Vol. 78, 093706 (2007)

Photonic Force Microscopy
Jingfang Wan, Yanan Huang, Ming-Chieh Cheng and Chia-Hsiang Menq

Objective:

To develop an actively controlled photonic force microscopic (PFM) system for 3D manipulation, force probing, visualization, and dynamic tracking of biological organelles and engineered objects, through optical trapping, back-focal-plane interferometry, force feedback scanning, thermal noise imaging, and active control.

System Description:

The entire PFM system has three strata (Fig. 1). The first stratum, also the bottom one, is set on a optical table where the inverted microscope sits with the confocal imaging system on the right side and a customized breadboard on the left side. The customized breadboard is the second stratum which holds most of the optical components, including the Infrared laser (1024nm , CW). The third stratum (the top one) is a home-made platform for the layout of all the QPD measurement components, which is supported by the microscope stage and the customized breadboard. An acrylic enclosure is designed to cover the PFM system with outlets for cables and confocal scanner (Fig. 2).

Application

One important application is to trap a functionalized bead to serve as a probe and detect its interaction with specific biomolecules on cell membrane. As shown in Fig. 3, a 1.87um PS bead coated with M2 antibody, specifically interacting with Flag epitope, was optically trapped and served as a probe. A live 293 cell was brought into contact with the probe to allow the interatcion (Fig. 3 left), and when the cell was gradually retreated, the interaction force initially was in balance with the trapping force, and deformed the local membrane area (Fig. 3 middle). Eventually, the probe was pulled out from the optical trap due to this interaction (Fig. 3 right).

Major Facilities

• Nikon TE2000-U Inverted Microscope with Temperature Control System
• VisiTech Infinity 2D Array Confocal Imaging System
• Physik Instrumente P-517.3CL 3-Axis Nanopositioner
• CrystaLaser IRCL-2.5W-1064 Infrared Laser
• National Instruments PXI-8186 Measurement and Automation System

Subnanometer Resolution Particle Tracking in Three Dimensions
Zhipeng Zhang and Chia-Hsiang Menq

Tracking micron-scale particles in three dimensions is an important scientific issue in many application fields, such as cell biology, fluid mechanics, and colloidal science. Many techniques have been developed to track the particle movement in all three dimensions, such as the laser detection method which utilizes a quadrant photodiode (QPD) to record the light scattered by the particle inside the laser focus, the 4D microscopy which acquires images at difference z positions and analyzes the image z-stacks, total internal reflection microscopy, and off-focus image method. Different methods have different advantages. For off-focus method, analyzing only one image, it is much faster than those methods utilizing 4D microscopy. The measurement range is as large as the microscope’s field of view in x and y axes, and the axial range is several to tens of microns depending on the magnification and numerical aperture (NA) of the objective lens. Moreover, multiple beads can be measured simultaneously as long as they are in the same field of view.

A three-dimensional (3D) particle tracking algorithm based on microscope off-focus images is developed in this research. Subnanometer resolution in all three axes at 400 Hz sampling rate is achieved using a CMOS camera. The images of a micron-scale bead appear to be very different when in focus and when out of focus. As shown in Fig. 1, the in-focus image of a micro particle has sharp edges and high contract while the out-of-focus image is blurry with concentric diffraction rings. Therefore, once the relationship between the off-focus image of the particle and its axial position being established through calibration prior to undertaking experiments, it is possible to estimate the axial positions of the particle according to the acquired off-focus images.

At each sampling, the lateral position of the spherical particle is first estimated by the centroid method. The axial position is then estimated by comparing the radius vector, which is converted from the off-focus 2D image of the particle with no information loss, with an object-specific model, calibrated automatically prior to each experiment. Because other than approximate theoretical models, object-specific calibration data and matching algorithm are used to locate the z location, the measurement precision and accuracy are both greatly improved. The computer program runs at 400 Hz, an order faster than the normal video rate of 30 frames per second. This allows measurement and analysis of particle dynamics in all three dimensions in extended frequency range. It could also be used in applications where on-line measurement is required to provide real-time feedback. The expense of faster sampling rate is larger noise in image acquired by a CMOS camera. Nevertheless, by utilizing all the information from the image, subnanometer precision is achieved in all three dimensions.

In order to illustrate measurement resolution, the developed tracking system was applied to measure nano-stepping of a 4.5µm diameter bead in three axes. The measurement system and associated experimental apparatus are set up on a modified inverted microscope (Nikon TE2000-U) equipped with a 60x water immersion objective lens (CFI Plan Apochromat VC 60X WI, NA 1.20). The bead was immobilized on a cover slip and its motion was controlled by the 3-axis piezo stage. The piezo stage was controlled to undergo 2 nm stepping along the x and y axes, and 5 nm the z axis. Fig. 2 shows the resulting position tracking of the bead. Nano-stepping can be clearly seen. The standard deviation in the x and y axes is 0.35 nm while it is 0.9 nm in the z axis. In order to illustrate measurement accuracy and range, the 4.5µm bead was controlled to follow a 10 µm triangular wave in each of the three axes, and the results are shown in Fig. 3.

Publication:

Zhipeng Zhang and Chia-Hsiang Menq, “Three-Dimensional Particle Tracking with Subnanometer Resolution Using Off-Focus Images,” Applied Optics, Vol. 47, No. 14, 2008.

Six-axis Magnetic Levitation and Motion Control

Zhipeng Zhang and Chia Hsiang Menq

Precision motion control devices are extensively used in advanced instrumentation and modern fabrication processes. Recent advances in actuator design and control electronics have significantly increased their application envelope, especially in the areas of positioning, alignment, scanning, and manipulation. Modern piezoelectric actuators are commonly used in these areas and they have internal control loops equipped with sensors that enable ultra high resolution. However, typical piezoelectric actuators travel in a linear range of micrometers, which may not be practically useful for many applications. The intrinsic nonlinear properties of piezoelectric actuators, such as hysteresis and creep, are also challenges for achieving ultra precision positioning.

Over the past two decades, the development of magnetic suspension/levitation (MS/L) technology has been undertaken by many researchers. It has been successfully implemented in many applications, such as high-speed magnetic levitation train, vibration isolation systems, and magnetic bearings. The technology uses a magnetically floated moving part to avoid contact friction and stiction. As MS/L is contact-free, there is no friction or wear, thereby enabling ultra high positioning precision. Moreover, it is also capable of multiple degrees of freedom (DOF) actuation without any mechanical compounding or cascading, thus providing high bandwidth precision motion control. Getting rid of lubrication also makes MS/L devices vacuum compatible, thus it is highly suitable for the nanotechnology and semiconductor manufacturing industries. Therefore, MS/L is considered to be a practical choice when developing high performance motion control systems.

The six-axis magnetic levitation stage (MLS), as shown in Figs. 1 &2, developed in PMCL achieved a large travel volume of 2 x 2 x 2 mm in translation and 4° x 4° x 4° in rotation. A two-axis linear actuator, based on magnetic Lorentz force law, was designed, and three actuators were implemented to achieve six-axis actuation. A high resolution laser interferometer measurement system (as seen in Fig. 3) was implemented and employed to measure the six-axis motion of the stage, facilitating real-time feedback control. Feedback linearization, based on rigid body dynamics of the levitated stage, and force distribution were implemented in a computer-controlled architecture so as to establish a decoupled dynamics between the six computed inputs and the resulting six-axis motions. Constant gain controllers were then designed and implemented, according to the concept of loop shaping, for each of the six axes, and the resolution of 0.74 nm rms has been achieved.

The range and resolution for each axis are shown in Table 1. Three experiments are conducted to demonstrate the high precision, large range and dexterity of the stage. Nano stepping of 5 nm per step in the y axis is shown in Figure 4. Three-axis contouring is illustrated in Figure 5, in which millimeter motion ranges in x, y and z are achieved. Figure 6 shows large travel motions of each of the six axes, 1 mm range in x, y and z axes and more than 2° rotation in α, β and θ axes. The large rotational motion is specially noticeable when compared with other magnetic suspension/levitation stages.

Range resolution

Publications

[1] Zhipeng Zhang and Chia-Hsiang Menq, “Six-axis magnetic levitation and motion control,” IEEE Trans. Rob., Vol. 23, No. 2, pp. 196-205, 2007.
[2] Zhipeng Zhang and Chia-Hsiang Menq, “Laser interferometric system for six-axis motion measurement,” Rev. Sci. Instrum., Vol. 78, 083107, 2007.

Visual sensing and visual servo control

Objective:

The need for high precision and robust motion control is wide spread in areas dealing with micro and nano scale phenomena where ultra precision motion control devices are extensively used in advanced instrumentations and modern fabrication processes requiring the highest precision. In the micro world, the governing mechanics are highly uncertain where gravity is no longer the dominant force and temperature fluctuations of proportions that are harmless in the macro world, can be detrimental causing structural drift, sensor drift, and surface tension variations. Therefore, manipulation of material and objects in the micro world requires schemes based on feedback information that directly and precisely conveys the objective task at hand. Our lab’s research is focusing on multi-axis and multi-object motion measurement and visual servo control in nanometer resolution.

Ongoing Project:

Six-axis visual sensing and visual servo control for nanometer resolution by using the lateral sampled white light interferometry (L-SWLI, as shown in Figure 1) method and pattern matching method.

Finished Project:

1. An Ultra Precision Six-Axis Visual SErvo Control System, a tracking of 1 um circle motion with visual feedback control was performed as shown in Figure 2.
2. Visual Servo Control Achieving Nanometer Resolution in X-Y-Z.
3. Visually Servoed 3D Alignment of Multiple Objects(system setup as shown in Figure 3) with Sub-Nanometer Precision.

Facilities:

1. Leica DMLM microscope and interferometers (20x and 50x ).
2. PI Nanocube XYZ Piezo Nanopositioning System.
3. PI High-Speed Objective Nanofocusing/Scanning Z-Drives with Direct Metrology.
4. PI LISA (Linear Stage Actuators) high-speed nanopositioner.
5. Photometric CoolSnap Monochrome CCD Camera.

Publications:

[1] J. Kim, S.K. Kuo, C.H. Menq, “An Ultra Precision Six-Axis Visual Servo Control System,” IEEE Transactions on Robotics Volume 21, Issue 5, Oct. 2005 Page(s):985 - 993.

[2] J. Kim, C.H. Menq, “Direct Visual Servo Control Achieving Nanometer Resolution in X-Y-Z,” to appear in IEEE Transactions on Robotics.
[3] J. Kim, C.H. Menq, “Visually Servoed 3D Alignment of Multiple Objects with Sub-Nanometer Precision,” to appear in IEEE Transactions on Nanotechnology.