SCANNING VOL. 36, 554–559 (2014) © Wiley Periodicals, Inc.
An Ultra-Rigid Close-Stacked Piezo Motor for Harsh Condition Scanning Probe Microscopy YING GUO,1,2 YUBIN HOU,1 AND QINGYOU LU1,2 1
High Magnetic Field Laboratory, Chinese Academy of Sciences and University of Science and Technology of China, Hefei, Anhui, People’s Republic of China 2 Hefei National Laboratory for Physical Sciences at Microscale, University of Science and Technology of China, Hefei, Anhui, People’s Republic of China
Summary: We designed and produced a nearly closest packed stack motor with only one tiny gap of 0.15 mm in the middle of the stack. A low-voltage method of controlling the motor is introduced for the first time. Besides, the test results of the motor and the corresponding scanning tunneling microscope are also presented. To our surprise, it turns out to be so rigid that even running two oil pumps and one ultrasonic cleaner within 1 m range from a STM directly driven by this new motor cannot cause the STM to produce any visible difference in its the atomic resolution quality. This is a leap in building a true harsh condition atomic resolution SPM. SCANNING 36:554–559, 2014. © 2014 Wiley Periodicals, Inc. Key words: closest packed stack motor, low-voltage control method, minimum mechanical loop, high rigidity, STM in noise environment
Introduction Piezo motors with the advantages of compactness, rigidity, simple control and accurate positioning have aroused great attentions in our daily lives as well as in scientific researches (Wittneven et al., ’97; Wu et al., 2006; Cummings et al., 2012; Kalkan et al., 2012; Kim Contract grant sponsor: Chinese National High Magnetic Field Facilities. Contract grant sponsor: FRFCU (under Contract grant number: WK2340000035). Contract grant sponsor: NSFC (under Contract grant numbers: U1232210, 11204306, 11374278). Address for reprints: Qingyou Lu, Hefei National Laboratory for Physical Sciences at Microscale, University of Science and Technology of China, Hefei, Anhui 230026, China. E-mail: [email protected] Received 18 March 2014; Accepted with revision 21 May 2014 DOI: 10.1002/sca.21149 Published online 2 September 2014 in Wiley Online Library (wileyonlinelibrary.com).
and de Lozanne, 2012). Especially, in the field of scanning probe microscopy (SPM), piezo motors have occupied a pivotal position (Olin, ’94; Pan et al., ’99; Pertaya et al., 2004; Hanaguri, 2006; Hou et al., 2008). In an atomically resolved scanning tunneling microscope (STM) for example, the piezo motor has become the limiting factor in many aspects including the size, the symmetry and the rigidity of the most important tipsample mechanic loop, which can all impact the stability and clarity of the images. In general, an ideal piezo motor for a high performance SPM should meet the following conditions: simple structure (high reliability), low-voltage to control (i.e., low threshold voltage, which brings low noise and high precision), compact (low thermal drift and extreme condition compatible), and rigid (vibration resistant). However, most piezo motor types have nonnegligible disadvantages: Inchworm (Berndt et al., ’90; Shimizu et al., ’90; Kato et al., ’91; Cusin et al., 2000) is complicated and requires high machine tolerance; The Pan style (Pan et al., ’99) motor is somewhat complicated with lower blocking force (thus lower output force) since it relies on shear displacement; the inertial motor (Pohl, ’87; Judy et al., ’90; Howald et al., ’92; Zesch et al., ’95) is less rigid and produces small output force because the inertial force used is typically weak. Recently, several delicate tubular piezo motor types have been designed including the GeckoDrive (Wang and Lu, 2009; Wang et al., 2013), TunaDrive (Liu and Lu, 2012), and PandaDrive (Lu and Hou, 2007; Lu et al., 2012) etc., in which one or more piezo scanning tubes (PST) drive a central shaft. Their drawbacks are obvious: (1) less rigid because a small wall thickness of piezo tube simply does not have large blocking force; (2) maximum mechanical loop because the shaft is typically far away from the target initially and needs to approach a certain distance before reaching the target, resulting in a larger mechanical loop between the mount of the piezo tube and the target. (3) The piezo tube is easy to break
Y. Guo et al.: An ultra-rigid close-stacked piezo motor
since its wall thickness needs to be thin enough (0.5 mm in many cases) in order to produce a sufficient displacement. All the above issues need to be taken care of. In this article, we present a nearly closest packed piezo stack motor characterized by high compactness, rigidity and low threshold voltage. In addition to its extensive performance tests presented in this article, we will also present the high quality atomic resolution images scanned by a home built STM (which is driven by this piezo motor) in a 74 dB loud background sound harsh condition environment without using sound isolation equipment.
Materials and Methods Figure 1(a and b) shows the structure of our piezo stack motor, which consists of six piezo plates of material PZT-5H with dimensions 25 mm length 9 mm width 0.5 mm thickness. Three of the piezo plates are glued together to form a stack using conductive silver-filled epoxy (H20E from Epoxy Technology, Inc.) with each pair of the mutually contacting electrodes
555
having the same polarity. The remaining three piezo plates are glued in the same way to form another stack. Next, these two independent branch stacks (denoted by symbols S1, S2) are stacked again to form the overall stack (OS) with the same polarity and an in-between 0.15 mm thick tantalum strip being glued together at one end (called fixed end). The tantalum strip serves as the common electrode, E, of the two branch stacks. The remaining electrode of each branch stack is its outer electrode, E1þ or E2þ. Between the other ends (called free ends) of the branches is a 0.15 mm diameter round Platinum/Iridium (Pt90/Ir10) wire which makes the gap between the branches uniform but will not block their relative displacements, because the rolling of the wire reduces the blocking friction force greatly. The OS is then spring-clamped by a pair of highly polished sapphire pillars (3.5 mm diameter 50 mm length) with the two ends of OS clamped. There four thin phosphor-copper strips (9 mm 1.5 mm 0.2 mm) are set between OS and rails, two of which at one rail are bent to become flexible which serve as the springs to spring clamp the OS against the rails. The following friction force condition of the motor needs to be satisfied: The total maximum static friction force,
Fig 1. (a) Schematic of exploded view of our piezo stack motor. (b) Assembled view of motor. (c) The driving signal that our motor used. (d) Diagrammatic sketch of motor stepping using the driving signal as (c) shown. Dotted line is the initial position of motor.
556
SCANNING VOL. 36, 5 (2014)
Ffree, on the two free ends of S1 and S2 and that on the fixed end, Ffixed, should satisfy 2Ffixed > Ffree > Ffixed. This is equivalent to say that among the three maximum static friction forces: Ffixed (on the fixed end), 1/2Ffree (on the free end of one branch), and 1/2Ffree (on the free end of the other branch), the sum of any two is larger than the remaining one. The discussion on how this condition will ensure the motor to work will be given shortly after. Since the friction coefficients at the free ends and the fixed end are approximately equal, we just need to adjust the abovementioned pair of phosphorcopper spring strips so as to make Ffree slightly larger than Ffixed. An appropriate value of the total maximum static friction force needs to be set (which is about 0.4 N in our motor). A greater static friction force will lead to a larger output force but higher threshold voltage also. Finally, the sapphire pillars are glued in parallel in two mounts, which are titanium made. Figure 2(a) displays the photo of so-assembled motor, which is as narrow as 13 mm.
To control the motor with lower voltage is always desired because: (1) During coarse approach, the high operation voltage on the piezo motor can generate positioning inaccuracy, which can cause the tip to crash on the sample (the measured tunneling current becomes saturated suddenly). This uncontrolled tip crashing can reduce image quality significantly. (2) If a motor can be operated using lower voltage, it means that the size of the motor can be reduced if we allow high voltage operation. The reduced motor size has lower drift and higher stability, which is of course desired. To this end, we design a method that can drive the motor with lower voltage. Refer to Figure 1(c and d), in which one step motion cycle T is divided into 10 parts of equal time length. Starting from 0 V at T0 toward T1, an elongation signal with the amplitude of Vmax is applied to E1þ and simultaneously, a contraction signal Vmax is applied on both E2þ and E. As a result, S1 will shrink from the natural state (called L0 state) to contracted state (called L state which caused by 2Vmax) and meanwhile, S2 will
Fig 2. (a) Photo of our STM using the piezo stack motor. (b) The step size vs. Vmax with the signal frequency of 10 Hz fixed. (c) The stepping speed and (d) the step size vs. driving signal frequency with Vmax of 40 V fixed.
Y. Guo et al.: An ultra-rigid close-stacked piezo motor
remain L0 state. Such changes cause the free end of S1 moving a distance of L0 L (2Vmax) toward the fixed end (defined as “upward direction”) because the abovementioned friction force condition requires that maximum static friction force on the free end of S1, 1/2F free, be smaller than the one on the free end of S2 (also ¼ 1/2Ffree) plus the one on fixed end (¼ Ffixed). Next, from T1 to T3, E1þ and E electrodes maintain their voltages at þVmax and Vmax, respectively, while an contraction signal applied to E2þ from Vmax to Vmax. In the end, the free end of S2 alone moves a distance of L0 L(2Vmax) toward the fixed end (i.e., upward direction) as guaranteed by the friction force condition, while the free end of S1 stays at L state. The point is that from T3 to T5, the voltages applied to E1þ and E2þ simultaneously change from Vmax to Vmax, whereas the voltage on E does the opposite. These signals cause both S1 and S2 elongate from L to elongation states (called Lþ states), leading to the fixed end move on the sapphire pillars by a huge distance of Lþ(2Vmax) L(2Vmax) in the upward direction. During T4 to T5 the piezo plates are reversed biased, so they extend beyond their relaxed state. From T5 to T9, S1 and S2 move from the Lþ(2Vmax) to L0 state one by one, which makes their free ends move in the upward direction one after another. At T9, the voltages on three electrodes are all at þVmax, which directly reduce to 0 V simultaneously at T10 without changing their L0 states. Finally, the conditions of the OS at T10 become exactly the same as it at T0. These procedures can be repeated to produce a macroscopic displacement, which is limited only by the length of sapphire pillars. Also, similar principle can allow the OS to stepwise move in the opposite direction (i.e., downward direction). The advantages of this rigid three-friction motor are obvious: (1) very high density of piezo material is utilized, which provides high integrity, compactness, and rigidity. These are highly desirable in the applications of harsh condition atomic resolution scanning probe microscopes; (2) low-voltage operation achieved using the new 10-substep control mechanism, which favors low noise and high precision; (3) the output force is produced during stage T3 T5, in which the fixed end pushes upward a huge distance of Lþ(2Vmax) L(2Vmax) by both S1 and S2 simultaneously, implying that higher rigidity can be achieved; (4) smallest mechanic loop forms at the end of the approach: the clamping points on the sapphire pillars will move toward the target with the motion of the OS, which always ends up with a smallest mechanic loop between the OS and the target, leading to a more stable and less drifting final structure.
Results and Discussion Under 10 Hz driving frequency, the measured step size vs. Vmax curves at room temperature for the above
557
motor to walk upward, downward, forward, and backward (horizontal), respectively, are shown in Figure 2(b). The threshold voltages are 22 and 12 V for stepping upward and downward, respectively, and 16 V for walking forward and backward, which are rather low compared with the tubular piezo motors (Wang and Lu, 2009; Liu and Lu, 2012; Wang et al., 2013). As Vmax increases, the step sizes in all the four directions enlarge from 100 nm to several micrometer pretty linearly. The large piezo displacements make it possible for this device to work without sharp changes in voltage, which simplifies the electronics. Our stack motor contains only six piezo plates with a total thickness of just 3 mm. There is a big room to reduce the threshold voltage by using more piezo plates with reduced thickness. Under a fixed Vmax of 40 V, the measured stepping speed and step size in each of the four directions as a function of driving frequency are given in Figure 2(c and d). In general, the stepping speed grows as the driving frequency increases, which is reasonable, but the step size decreases slightly. The latter can be explained by the slow creeping nature of piezo material. In other words, when the frequency increases, the piezo material in the stack motor simply cannot respond fast enough, thus unable to displace at full potential and hence loosing step size. The reliability of the stack motor was also tested by forcing it to travel 5 mm back and forth continuously (Vmax ¼ 50 V and driving frequency ¼ 20 Hz) for more than 24 h. The result was excellent: the motor worked exactly the same as before and no new scratches were found on the sapphire pillars at all. As the most severe test of the above motor, we built a room temperature ambient condition STM using the motor to implement the tip to sample coarse approach. A tiny four-quadrant piezo tube scanner (EBL#4 from EBL Products Inc.) of 3.2 mm outer diameter 6.5 mm length 0.5 mm wall thickness is glued at the fixed end of the OS through an insulating sapphire piece (see Figure 2(a)). The tip is a hand cut Pt/Ir wire (0.15 mm diameter). It is glued at the free end of the scanner. The sample is mounted at the end of the sapphire pillar rails toward which the tip can approach by controlling the motor. The atomically resolved highly oriented pyrolytic graphite (HOPG) images scanned by this STM at room temperature inside a home built vibration sound isolation box with its lid closed and opened are presented in Figure 3(c–f), respectively, where the same sample (positively biased at 200 mV) and tip were used. Figure 3(c and d) (raw data) were scanned in constant height mode with the scanning speed of 5 lines/ s, while Figure 3(e and f) (plane fitted) were scanned in constant current mode with the scanning speed of 2 lines/s. These four images were taken in a very noisy and highly vibrational environment in which two oil pumps and an ultrasonic cleaner about one meter away
558
SCANNING VOL. 36, 5 (2014)
box with its lid closed/opened are shown in Figure 3(b). However, the STM equipped with the new motor is perfectly resistant to all these harsh conditions, since no visible difference in image quality can be seen from Figure 3(c or e) and (d or f). It is sufficient to confirm the ultra-high rigidity of our new motor. Actually, we tested the STM several times using different tips in noisy environment, and could obtain excellent atomic resolution images every time. In this article, we selected the one with the highest work function of 0.8 eV, which was measured from the I–Z curves (not given in paper) before and after scanning the images. Here, the current was measured in air while pushing the tip toward the HOPG until reaching the preset current (20 nA). The vertical drift of our motor was measured by running the constant current program to hold the tunneling current at 5 nA (scan range ¼ 0) and recording the displacement of the scanner in vertical direction every 5 min for 1.5 h. As a consequence, the average vertical drift is about 53 pm/min in the sound isolation box and 58 pm/min in the noisy environment, which are both very small.
Conclusions To summarize, a highly close-packed piezo stack motor and its low-voltage operation method have been presented. It turns out to be so rigid that even running two oil pumps and one ultrasonic cleaner within 1 m range from an STM directly driven by this new motor cannot cause the STM to produce any visible difference in its the atomic resolution quality. This is a leap in building a long dreamed harsh condition atomic resolution SPM.
References Fig 3. (a) The sound level at the point of STM within 1,000 s measured in noisy environment with and without the lid of sound isolation box opened. (b) Vibration spectra at the point of STM with two different conditions above. (c and d) Raw data of constant height STM images of HOPG with ranges of 32.4A 32.4A taken in the vibration and sound isolation box and noisy environment. (e and f) Plane fitted constant current STM images of HOPG with ranges of 37.9A 37.9A taken in the vibration and sound isolation box and noisy environment.
from the STM were simultaneously running on purpose. The average sound level in 1,000 s measured (using DT8852 of ShenZhen Everbest Machinery Industry Co.) inside the vibration and sound isolation box with its lid closed was 33.5 dB, which became 74 dB when we open the lid of sound isolation box. The detailed data are given in Figure 3(a). The measured vibration spectra (using BVM-100-2S-J, BeiJing Vibration Measurement Co.) for the above conditions of inside sound isolation
Berndt R, Schlittler RR, Gimzewski JK. 1990. Photon emission scanning tunneling microscope. J Vac Sci Technol B 9(2):573–577. Cummings ML, Chien TY, Preissner C, et al. 2012. Combining scanning tunneling microscopy and synchrotron radiation for high-resolution imaging and spectroscopy with chemical, electronic, and magnetic contrast. Ultramicroscopy 112:22– 31. Cusin P, Sawai T, Konishi S. 2000. Compact and precise positioner based on the Inchworm principle. J Mecromech Microeng 10:516–521. Hanaguri T. 2006. Development of high-field STM and its application to the study on magnetically-tuned criticality in Sr3Ru2O7. J Phys Conf Ser 51:514–521. Hou YB, Wang JT, Lu QY. 2008. All low voltage lateral junction scanning tunneling microscope with very high precision and stability. Rev Sci Instrum 79:113707. Howald L, Rudin H, Gntherodt HJ. 1992. Piezo inertial stepping motor with spherical rotor. Rev Sci Instrum 63:3909. Judy JW, Polla DL, Robbins WP. 1990. A linear piezo stepper motor with submicrometer step size and centimeter travel range. IEEE Trans Ultrason Ferroelectr Freq Control 37: 428–437.
Y. Guo et al.: An ultra-rigid close-stacked piezo motor Kalkan F, Zaum C, Morgenstern K. 2012. A scanning tunneling microscope with a scanning range from hundreds of micrometers down to nanometer resolution. Rev Sci Instrum 83:103903. Kato T, Osaka F, Tanaka I, Ohkouchi S. 1991. A scanning tunneling microscope using dual-axes inchworms for the observation of a cleaved semiconductor surface. J Vac Sci Technol B 9:1981–1984. Kim SH, de Lozanne A. 2012. Compact scanning tunneling microscope for spin polarization measurements. Rev Sci Instrum 83:103701. Liu XL, Lu QY. 2012. A piezo motor based on a new principle with high output force, rigidity and integrity: the Tuna Drive. Rev Sci Instrum 83:115111. Lu QY, Hou YB. 2007. China invention patent ZL200610161477.3. Lu QY, Wang Q, Hou YB. 2012. Comment on “A nanopositioner for scanning probe microscopy: the KoalaDrive”. Rev Sci Instrum 83:097101. Olin H. 1994. Design of a scanning probe microscope. Meas Sci Technol 5:976–984. Pan SH, Hudson EW, Davis JC. 1999. 3He refrigerator based very low temperature scanning tunneling microscope. Rev Sci Instrum 70:1459.
559
Pertaya N, Braun KF, Rieder KH. 2004. On the stability of Besocke-type scanners. Rev Sci Instrum 75:2608. Pohl DW. 1987. Dynamic piezo translation devices. Rev Sci Instrum 58:54. Shimizu N, Kimura T, Nakamura T, Umebu I. 1990. An ultrahigh vacuum scanning tunneling microscope with a new inchworm mechanism. J Vac Sci Technol A 8:333– 3335. Wang Q, Lu QY. 2009. A simple, compact, and rigid piezo step motor with large step size. Rev Sci Instrum 80:085104. Wang Q, Hou YB, Lu QY. 2013. Note: a compact, rigid, and easyto-build piezo motor: the intact-tube GeckoDrive. Rev Sci Instrum 84:056106. Wittneven C, Dombrowski R, Pan SH, Wiesendanger R. 1997. A low-temperature ultrahigh-vacuum scanning tunneling microscope with rotatable magnetic field. Rev Sci Instrum 68:3806. Wu WD, Guha A, Kim S, de Lozanne A. 2006. A compact dual-tip STM design. IEEE Transect Nanotechnol 5:77–79. Zesch W, Buchi R, Codourey A, Siegwart RY. 1995. Inertial drives for micro- and nanorobots: two novel mechanisms. Proc SPIE 2593:80–88.
An Ultra-Rigid Close-Stacked Piezo Motor for Harsh Condition Scanning Probe Microscopy YING GUO,1,2 YUBIN HOU,1 AND QINGYOU LU1,2 1
High Magnetic Field Laboratory, Chinese Academy of Sciences and University of Science and Technology of China, Hefei, Anhui, People’s Republic of China 2 Hefei National Laboratory for Physical Sciences at Microscale, University of Science and Technology of China, Hefei, Anhui, People’s Republic of China
Summary: We designed and produced a nearly closest packed stack motor with only one tiny gap of 0.15 mm in the middle of the stack. A low-voltage method of controlling the motor is introduced for the first time. Besides, the test results of the motor and the corresponding scanning tunneling microscope are also presented. To our surprise, it turns out to be so rigid that even running two oil pumps and one ultrasonic cleaner within 1 m range from a STM directly driven by this new motor cannot cause the STM to produce any visible difference in its the atomic resolution quality. This is a leap in building a true harsh condition atomic resolution SPM. SCANNING 36:554–559, 2014. © 2014 Wiley Periodicals, Inc. Key words: closest packed stack motor, low-voltage control method, minimum mechanical loop, high rigidity, STM in noise environment
Introduction Piezo motors with the advantages of compactness, rigidity, simple control and accurate positioning have aroused great attentions in our daily lives as well as in scientific researches (Wittneven et al., ’97; Wu et al., 2006; Cummings et al., 2012; Kalkan et al., 2012; Kim Contract grant sponsor: Chinese National High Magnetic Field Facilities. Contract grant sponsor: FRFCU (under Contract grant number: WK2340000035). Contract grant sponsor: NSFC (under Contract grant numbers: U1232210, 11204306, 11374278). Address for reprints: Qingyou Lu, Hefei National Laboratory for Physical Sciences at Microscale, University of Science and Technology of China, Hefei, Anhui 230026, China. E-mail: [email protected] Received 18 March 2014; Accepted with revision 21 May 2014 DOI: 10.1002/sca.21149 Published online 2 September 2014 in Wiley Online Library (wileyonlinelibrary.com).
and de Lozanne, 2012). Especially, in the field of scanning probe microscopy (SPM), piezo motors have occupied a pivotal position (Olin, ’94; Pan et al., ’99; Pertaya et al., 2004; Hanaguri, 2006; Hou et al., 2008). In an atomically resolved scanning tunneling microscope (STM) for example, the piezo motor has become the limiting factor in many aspects including the size, the symmetry and the rigidity of the most important tipsample mechanic loop, which can all impact the stability and clarity of the images. In general, an ideal piezo motor for a high performance SPM should meet the following conditions: simple structure (high reliability), low-voltage to control (i.e., low threshold voltage, which brings low noise and high precision), compact (low thermal drift and extreme condition compatible), and rigid (vibration resistant). However, most piezo motor types have nonnegligible disadvantages: Inchworm (Berndt et al., ’90; Shimizu et al., ’90; Kato et al., ’91; Cusin et al., 2000) is complicated and requires high machine tolerance; The Pan style (Pan et al., ’99) motor is somewhat complicated with lower blocking force (thus lower output force) since it relies on shear displacement; the inertial motor (Pohl, ’87; Judy et al., ’90; Howald et al., ’92; Zesch et al., ’95) is less rigid and produces small output force because the inertial force used is typically weak. Recently, several delicate tubular piezo motor types have been designed including the GeckoDrive (Wang and Lu, 2009; Wang et al., 2013), TunaDrive (Liu and Lu, 2012), and PandaDrive (Lu and Hou, 2007; Lu et al., 2012) etc., in which one or more piezo scanning tubes (PST) drive a central shaft. Their drawbacks are obvious: (1) less rigid because a small wall thickness of piezo tube simply does not have large blocking force; (2) maximum mechanical loop because the shaft is typically far away from the target initially and needs to approach a certain distance before reaching the target, resulting in a larger mechanical loop between the mount of the piezo tube and the target. (3) The piezo tube is easy to break
Y. Guo et al.: An ultra-rigid close-stacked piezo motor
since its wall thickness needs to be thin enough (0.5 mm in many cases) in order to produce a sufficient displacement. All the above issues need to be taken care of. In this article, we present a nearly closest packed piezo stack motor characterized by high compactness, rigidity and low threshold voltage. In addition to its extensive performance tests presented in this article, we will also present the high quality atomic resolution images scanned by a home built STM (which is driven by this piezo motor) in a 74 dB loud background sound harsh condition environment without using sound isolation equipment.
Materials and Methods Figure 1(a and b) shows the structure of our piezo stack motor, which consists of six piezo plates of material PZT-5H with dimensions 25 mm length 9 mm width 0.5 mm thickness. Three of the piezo plates are glued together to form a stack using conductive silver-filled epoxy (H20E from Epoxy Technology, Inc.) with each pair of the mutually contacting electrodes
555
having the same polarity. The remaining three piezo plates are glued in the same way to form another stack. Next, these two independent branch stacks (denoted by symbols S1, S2) are stacked again to form the overall stack (OS) with the same polarity and an in-between 0.15 mm thick tantalum strip being glued together at one end (called fixed end). The tantalum strip serves as the common electrode, E, of the two branch stacks. The remaining electrode of each branch stack is its outer electrode, E1þ or E2þ. Between the other ends (called free ends) of the branches is a 0.15 mm diameter round Platinum/Iridium (Pt90/Ir10) wire which makes the gap between the branches uniform but will not block their relative displacements, because the rolling of the wire reduces the blocking friction force greatly. The OS is then spring-clamped by a pair of highly polished sapphire pillars (3.5 mm diameter 50 mm length) with the two ends of OS clamped. There four thin phosphor-copper strips (9 mm 1.5 mm 0.2 mm) are set between OS and rails, two of which at one rail are bent to become flexible which serve as the springs to spring clamp the OS against the rails. The following friction force condition of the motor needs to be satisfied: The total maximum static friction force,
Fig 1. (a) Schematic of exploded view of our piezo stack motor. (b) Assembled view of motor. (c) The driving signal that our motor used. (d) Diagrammatic sketch of motor stepping using the driving signal as (c) shown. Dotted line is the initial position of motor.
556
SCANNING VOL. 36, 5 (2014)
Ffree, on the two free ends of S1 and S2 and that on the fixed end, Ffixed, should satisfy 2Ffixed > Ffree > Ffixed. This is equivalent to say that among the three maximum static friction forces: Ffixed (on the fixed end), 1/2Ffree (on the free end of one branch), and 1/2Ffree (on the free end of the other branch), the sum of any two is larger than the remaining one. The discussion on how this condition will ensure the motor to work will be given shortly after. Since the friction coefficients at the free ends and the fixed end are approximately equal, we just need to adjust the abovementioned pair of phosphorcopper spring strips so as to make Ffree slightly larger than Ffixed. An appropriate value of the total maximum static friction force needs to be set (which is about 0.4 N in our motor). A greater static friction force will lead to a larger output force but higher threshold voltage also. Finally, the sapphire pillars are glued in parallel in two mounts, which are titanium made. Figure 2(a) displays the photo of so-assembled motor, which is as narrow as 13 mm.
To control the motor with lower voltage is always desired because: (1) During coarse approach, the high operation voltage on the piezo motor can generate positioning inaccuracy, which can cause the tip to crash on the sample (the measured tunneling current becomes saturated suddenly). This uncontrolled tip crashing can reduce image quality significantly. (2) If a motor can be operated using lower voltage, it means that the size of the motor can be reduced if we allow high voltage operation. The reduced motor size has lower drift and higher stability, which is of course desired. To this end, we design a method that can drive the motor with lower voltage. Refer to Figure 1(c and d), in which one step motion cycle T is divided into 10 parts of equal time length. Starting from 0 V at T0 toward T1, an elongation signal with the amplitude of Vmax is applied to E1þ and simultaneously, a contraction signal Vmax is applied on both E2þ and E. As a result, S1 will shrink from the natural state (called L0 state) to contracted state (called L state which caused by 2Vmax) and meanwhile, S2 will
Fig 2. (a) Photo of our STM using the piezo stack motor. (b) The step size vs. Vmax with the signal frequency of 10 Hz fixed. (c) The stepping speed and (d) the step size vs. driving signal frequency with Vmax of 40 V fixed.
Y. Guo et al.: An ultra-rigid close-stacked piezo motor
remain L0 state. Such changes cause the free end of S1 moving a distance of L0 L (2Vmax) toward the fixed end (defined as “upward direction”) because the abovementioned friction force condition requires that maximum static friction force on the free end of S1, 1/2F free, be smaller than the one on the free end of S2 (also ¼ 1/2Ffree) plus the one on fixed end (¼ Ffixed). Next, from T1 to T3, E1þ and E electrodes maintain their voltages at þVmax and Vmax, respectively, while an contraction signal applied to E2þ from Vmax to Vmax. In the end, the free end of S2 alone moves a distance of L0 L(2Vmax) toward the fixed end (i.e., upward direction) as guaranteed by the friction force condition, while the free end of S1 stays at L state. The point is that from T3 to T5, the voltages applied to E1þ and E2þ simultaneously change from Vmax to Vmax, whereas the voltage on E does the opposite. These signals cause both S1 and S2 elongate from L to elongation states (called Lþ states), leading to the fixed end move on the sapphire pillars by a huge distance of Lþ(2Vmax) L(2Vmax) in the upward direction. During T4 to T5 the piezo plates are reversed biased, so they extend beyond their relaxed state. From T5 to T9, S1 and S2 move from the Lþ(2Vmax) to L0 state one by one, which makes their free ends move in the upward direction one after another. At T9, the voltages on three electrodes are all at þVmax, which directly reduce to 0 V simultaneously at T10 without changing their L0 states. Finally, the conditions of the OS at T10 become exactly the same as it at T0. These procedures can be repeated to produce a macroscopic displacement, which is limited only by the length of sapphire pillars. Also, similar principle can allow the OS to stepwise move in the opposite direction (i.e., downward direction). The advantages of this rigid three-friction motor are obvious: (1) very high density of piezo material is utilized, which provides high integrity, compactness, and rigidity. These are highly desirable in the applications of harsh condition atomic resolution scanning probe microscopes; (2) low-voltage operation achieved using the new 10-substep control mechanism, which favors low noise and high precision; (3) the output force is produced during stage T3 T5, in which the fixed end pushes upward a huge distance of Lþ(2Vmax) L(2Vmax) by both S1 and S2 simultaneously, implying that higher rigidity can be achieved; (4) smallest mechanic loop forms at the end of the approach: the clamping points on the sapphire pillars will move toward the target with the motion of the OS, which always ends up with a smallest mechanic loop between the OS and the target, leading to a more stable and less drifting final structure.
Results and Discussion Under 10 Hz driving frequency, the measured step size vs. Vmax curves at room temperature for the above
557
motor to walk upward, downward, forward, and backward (horizontal), respectively, are shown in Figure 2(b). The threshold voltages are 22 and 12 V for stepping upward and downward, respectively, and 16 V for walking forward and backward, which are rather low compared with the tubular piezo motors (Wang and Lu, 2009; Liu and Lu, 2012; Wang et al., 2013). As Vmax increases, the step sizes in all the four directions enlarge from 100 nm to several micrometer pretty linearly. The large piezo displacements make it possible for this device to work without sharp changes in voltage, which simplifies the electronics. Our stack motor contains only six piezo plates with a total thickness of just 3 mm. There is a big room to reduce the threshold voltage by using more piezo plates with reduced thickness. Under a fixed Vmax of 40 V, the measured stepping speed and step size in each of the four directions as a function of driving frequency are given in Figure 2(c and d). In general, the stepping speed grows as the driving frequency increases, which is reasonable, but the step size decreases slightly. The latter can be explained by the slow creeping nature of piezo material. In other words, when the frequency increases, the piezo material in the stack motor simply cannot respond fast enough, thus unable to displace at full potential and hence loosing step size. The reliability of the stack motor was also tested by forcing it to travel 5 mm back and forth continuously (Vmax ¼ 50 V and driving frequency ¼ 20 Hz) for more than 24 h. The result was excellent: the motor worked exactly the same as before and no new scratches were found on the sapphire pillars at all. As the most severe test of the above motor, we built a room temperature ambient condition STM using the motor to implement the tip to sample coarse approach. A tiny four-quadrant piezo tube scanner (EBL#4 from EBL Products Inc.) of 3.2 mm outer diameter 6.5 mm length 0.5 mm wall thickness is glued at the fixed end of the OS through an insulating sapphire piece (see Figure 2(a)). The tip is a hand cut Pt/Ir wire (0.15 mm diameter). It is glued at the free end of the scanner. The sample is mounted at the end of the sapphire pillar rails toward which the tip can approach by controlling the motor. The atomically resolved highly oriented pyrolytic graphite (HOPG) images scanned by this STM at room temperature inside a home built vibration sound isolation box with its lid closed and opened are presented in Figure 3(c–f), respectively, where the same sample (positively biased at 200 mV) and tip were used. Figure 3(c and d) (raw data) were scanned in constant height mode with the scanning speed of 5 lines/ s, while Figure 3(e and f) (plane fitted) were scanned in constant current mode with the scanning speed of 2 lines/s. These four images were taken in a very noisy and highly vibrational environment in which two oil pumps and an ultrasonic cleaner about one meter away
558
SCANNING VOL. 36, 5 (2014)
box with its lid closed/opened are shown in Figure 3(b). However, the STM equipped with the new motor is perfectly resistant to all these harsh conditions, since no visible difference in image quality can be seen from Figure 3(c or e) and (d or f). It is sufficient to confirm the ultra-high rigidity of our new motor. Actually, we tested the STM several times using different tips in noisy environment, and could obtain excellent atomic resolution images every time. In this article, we selected the one with the highest work function of 0.8 eV, which was measured from the I–Z curves (not given in paper) before and after scanning the images. Here, the current was measured in air while pushing the tip toward the HOPG until reaching the preset current (20 nA). The vertical drift of our motor was measured by running the constant current program to hold the tunneling current at 5 nA (scan range ¼ 0) and recording the displacement of the scanner in vertical direction every 5 min for 1.5 h. As a consequence, the average vertical drift is about 53 pm/min in the sound isolation box and 58 pm/min in the noisy environment, which are both very small.
Conclusions To summarize, a highly close-packed piezo stack motor and its low-voltage operation method have been presented. It turns out to be so rigid that even running two oil pumps and one ultrasonic cleaner within 1 m range from an STM directly driven by this new motor cannot cause the STM to produce any visible difference in its the atomic resolution quality. This is a leap in building a long dreamed harsh condition atomic resolution SPM.
References Fig 3. (a) The sound level at the point of STM within 1,000 s measured in noisy environment with and without the lid of sound isolation box opened. (b) Vibration spectra at the point of STM with two different conditions above. (c and d) Raw data of constant height STM images of HOPG with ranges of 32.4A 32.4A taken in the vibration and sound isolation box and noisy environment. (e and f) Plane fitted constant current STM images of HOPG with ranges of 37.9A 37.9A taken in the vibration and sound isolation box and noisy environment.
from the STM were simultaneously running on purpose. The average sound level in 1,000 s measured (using DT8852 of ShenZhen Everbest Machinery Industry Co.) inside the vibration and sound isolation box with its lid closed was 33.5 dB, which became 74 dB when we open the lid of sound isolation box. The detailed data are given in Figure 3(a). The measured vibration spectra (using BVM-100-2S-J, BeiJing Vibration Measurement Co.) for the above conditions of inside sound isolation
Berndt R, Schlittler RR, Gimzewski JK. 1990. Photon emission scanning tunneling microscope. J Vac Sci Technol B 9(2):573–577. Cummings ML, Chien TY, Preissner C, et al. 2012. Combining scanning tunneling microscopy and synchrotron radiation for high-resolution imaging and spectroscopy with chemical, electronic, and magnetic contrast. Ultramicroscopy 112:22– 31. Cusin P, Sawai T, Konishi S. 2000. Compact and precise positioner based on the Inchworm principle. J Mecromech Microeng 10:516–521. Hanaguri T. 2006. Development of high-field STM and its application to the study on magnetically-tuned criticality in Sr3Ru2O7. J Phys Conf Ser 51:514–521. Hou YB, Wang JT, Lu QY. 2008. All low voltage lateral junction scanning tunneling microscope with very high precision and stability. Rev Sci Instrum 79:113707. Howald L, Rudin H, Gntherodt HJ. 1992. Piezo inertial stepping motor with spherical rotor. Rev Sci Instrum 63:3909. Judy JW, Polla DL, Robbins WP. 1990. A linear piezo stepper motor with submicrometer step size and centimeter travel range. IEEE Trans Ultrason Ferroelectr Freq Control 37: 428–437.
Y. Guo et al.: An ultra-rigid close-stacked piezo motor Kalkan F, Zaum C, Morgenstern K. 2012. A scanning tunneling microscope with a scanning range from hundreds of micrometers down to nanometer resolution. Rev Sci Instrum 83:103903. Kato T, Osaka F, Tanaka I, Ohkouchi S. 1991. A scanning tunneling microscope using dual-axes inchworms for the observation of a cleaved semiconductor surface. J Vac Sci Technol B 9:1981–1984. Kim SH, de Lozanne A. 2012. Compact scanning tunneling microscope for spin polarization measurements. Rev Sci Instrum 83:103701. Liu XL, Lu QY. 2012. A piezo motor based on a new principle with high output force, rigidity and integrity: the Tuna Drive. Rev Sci Instrum 83:115111. Lu QY, Hou YB. 2007. China invention patent ZL200610161477.3. Lu QY, Wang Q, Hou YB. 2012. Comment on “A nanopositioner for scanning probe microscopy: the KoalaDrive”. Rev Sci Instrum 83:097101. Olin H. 1994. Design of a scanning probe microscope. Meas Sci Technol 5:976–984. Pan SH, Hudson EW, Davis JC. 1999. 3He refrigerator based very low temperature scanning tunneling microscope. Rev Sci Instrum 70:1459.
559
Pertaya N, Braun KF, Rieder KH. 2004. On the stability of Besocke-type scanners. Rev Sci Instrum 75:2608. Pohl DW. 1987. Dynamic piezo translation devices. Rev Sci Instrum 58:54. Shimizu N, Kimura T, Nakamura T, Umebu I. 1990. An ultrahigh vacuum scanning tunneling microscope with a new inchworm mechanism. J Vac Sci Technol A 8:333– 3335. Wang Q, Lu QY. 2009. A simple, compact, and rigid piezo step motor with large step size. Rev Sci Instrum 80:085104. Wang Q, Hou YB, Lu QY. 2013. Note: a compact, rigid, and easyto-build piezo motor: the intact-tube GeckoDrive. Rev Sci Instrum 84:056106. Wittneven C, Dombrowski R, Pan SH, Wiesendanger R. 1997. A low-temperature ultrahigh-vacuum scanning tunneling microscope with rotatable magnetic field. Rev Sci Instrum 68:3806. Wu WD, Guha A, Kim S, de Lozanne A. 2006. A compact dual-tip STM design. IEEE Transect Nanotechnol 5:77–79. Zesch W, Buchi R, Codourey A, Siegwart RY. 1995. Inertial drives for micro- and nanorobots: two novel mechanisms. Proc SPIE 2593:80–88.
