A Student-Built Scanning Tunneling Microscope

Tom Ekkens

Citation: The Physics Teacher 53, 539 (2015); doi: 10.1119/1.4935765

View online: http://dx.doi.org/10.1119/1.4935765
View Table of Contents: http://aapt.scitation.org/toc/pte/53/9
Published by the American Association of Physics Teachers
A Student-Built Scanning Tunneling
Microscope
Tom Ekkens, Walla Walla University, College Place, WA

M
any introductory and nanotechnology textbooks Mechanical subsystem of the student STM
discuss the operation of various microscopes in-
cluding atomic force (AFM), scanning tunneling To build the mechanical subsystem, minimal tooling is
(STM), and scanning electron microscopes (SEM).1,2 In a required—a hand drill, a ¾-inch bit, a hand saw, and a 2-56
nanotechnology laboratory class, students frequently utilize bit and tap. All gluing is done with hot glue. This limited ma-
microscopes to obtain data without a thought about the de- chining list is possible because many hardware stores stock an
tailed operation of the tool itself. I wanted to give my students inexpensive sample approach system in the form of the B&K
a deeper appreciation for the physics by having them build a EZ-Span ¾-inch slider repair coupling. This PVC part, made
simple scanning tunneling microscope. Initially, 15 hours of to repair breaks in pipe, consists of three pieces: a housing, a
an upper-division laboratory class were devoted to building threaded end cap, and an inner slider. In the student STM de-
and operating the STM. As the build process was refined, the sign, the sample attaches to the inside end of the inner slider,
time commitment for this project has shrunk to nine hours. the piezo and tip attach to the inside of the end cap, and the
Using the method described in this paper, the project is now housing keeps the entire system aligned. The slider comes
simple enough that it can be built and operated by students in with lubricated o-rings in place that work perfectly right out
the introductory class. of the box.
The starting place for the development of this lab was the Completing the mechanical build of the student STM in
excellent article, “A Simple Scanning Tunneling Microscope a three-hour lab session requires the work of two students.
(STM)” by John D. Alexander.3 An STM uses quantum tun- One builds the scanner by cutting the conducting layer on
neling of electrons between a conducting probe and a con- a 27-mm piezo disc transducer into the quadrant pattern
ducting sample to construct a surface map of the sample. shown in the inset of Fig. 1. He solders five small magnet
The ability to control the position of the probe relative to the wires to the disc. He then glues a phenolic standoff to the
sample at the subatomic scale is one of the most critical de- brass side of the disc, cuts a tip holder from a dual in-line
sign considerations when building one of these microscopes. package socket, and glues it to the standoff. The completed
The Alexander system used a piezo disc for scanning and scanner is glued to the threaded end cap.
feedback motion. The scanner and sample were attached to
metal plates separated by threaded rods. Turning the threaded
rods allowed for coarse approach. The analog electronics were
responsible for the integral-based feedback loop, generating
the bias voltage, and amplification of all signals. A function
generator was used to generate the scanning signal and a stor-
age oscilloscope collected data.
Alexander’s system was replicated by our senior under-
graduate students, but there were limitations with the design.
Fig. 1. The assembled STM with the stepper motor on the left.
First, testing the complicated electronics circuit took several
The reducer bushing is the gray piece just to the right of the
hours. Second, machining of the metal plates was tedious, motor. Sample and tip are visible through the access hole. The
time consuming, and required complicated tooling. Third, inset at upper right pictures the connections to the piezo.
the coarse-approach method of turning a screw crashed
the tunneling tip into the sample on nearly every approach. The other student builds the approach system. She drills a
Fourth, upon entering tunneling range with the system, the ¾-inch hole in the house for access to the tip and sample, cuts
metal design lacked sufficient damping to make the system the slider to a length of two inches, and presses a ½ PVC plug
noise tolerant. On the positive side, the piezo design worked. into each end of the slider. Next, she assembles the sample
Piezo disks like these have been used in other types of micro- holder by gluing a piece of circuit board and a small magnet
scopes as well.4 to the sample end PVC plug. Finally, she taps a 2-56 threaded
In an attempt to design a system that lessened these flaws, hole into the PVC plug on the other end of the slider.
the student STM was developed. Guiding principles in the Once the pieces are assembled and wires connected to
new approach were to minimize electrical complexity, reduce the sample holder and the tip holder, the coarse-approach
tedious steps of the build, automate the approach process so system is attached. The coarse approach is accomplished by a
that tip integrity did not depend on the patience of the stu- 200-step-per-revolution stepper motor with a 5-mm shaft.
dent, and make the system robust. A 2- to 5-mm coupler connects a 2-56 threaded rod to the

DOI: 10.1119/1.4935765 The Physics Teacher ◆ Vol. 53, December 2015 539
 tion with a +10 voltage range.
For this project, a National Instruments PCIe-6323 data
acquisition card was used. Two cables and two breakout
boards connected the card to the STM electronics. Supplied
National Instruments drivers were utilized to program the
card in C++. Without an on-board processor, update speed
for the feedback loop is 1.8 kHz since each read and subse-
quent write request is handled by the host CPU. This speed
has not proved a limitation except when scanning images
smaller than 50 nm. The PCIe-6323 also has a mode where a
string of data can be written or read at speeds up to 250 kHz if
no feedback is required. This mode, which scans small areas
rapidly, is used when imaging graphite.

• Programming
A simple C++ program, developed before the laboratory
was taught, allows data to be taken in feedback mode only.
The executable file or the source code can be downloaded
from the author’s website.5
LabVIEW or open-source programs such as Gwyddion or
GXSM could be used if desired.

Electrical subsystem of the student STM

In the student STM, the computer performs most of the
Fig. 2. In this schematic of the entire electrical system, the con-
nectors to the National Instruments card are shown at left. The work, thus the analog electrical system can be simplified to
stepper motor controller is shown at the bottom. The analog three amplifiers. A transimpedance amplifier converts the
electronics are in the middle. tunnel current to voltage. The wire length from the sample
to the amplifier should be less than two inches to minimize
motor. The other end of the 2-56 rod threads into the PVC noise pickup. Two inverting amplifiers generate signals for
plug. Twenty millimeters of total slider travel is available. the -X and -Y quadrants of the piezo. All of these are built
Each motor step moves the slider forward by about 2 μm if from a single TL074 IC. The tunnel current amplifier uses a
the controller is using full-step mode. Half-step or quarter- 100-MΩ feedback resistor. All the other resistors are 10 kΩ.
step mode is recommended for best results. A 1¼- to 1½-in Four 9-V batteries power the TL074. The stepper motor con-
PVC reducer bushing holds the motor the appropriate dis- troller receives two digital inputs from the computer, one for
tance from the housing. A hose clamp keeps the entire me- step frequency and the other for step direction.
chanical system together. The completed system is shown in The entire system is shown in Fig. 2. Not shown is a dc
Fig. 1. power supply for the motor controller and an oscilloscope to
During operation the system is placed on a piece of foam monitor the z-axis voltage and the tunnel current. The electri-
about one inch thick to isolate it from large mechanical vibra- cal system can be built, connected to a computer, and tested
tions. Loud talking and people moving about the room are in a single three-hour lab period.
discouraged during image acquisition. While the system is With the piezo divided into four equal parts, a total mo-
reliable and vibration tolerant, it has several limitations. First, tion of 3 μm in both horizontal axes and 2 μm in the vertical
PVC is not as rigid as metal so at very high magnifications axis is available when powered by +10 V. This is ideal for im-
there is measurable drift in the image. Second, there is at least aging samples of 10 nm or larger.
2 μm of backlash in the coarse approach. Third, there is lim-
ited motion in the horizontal plane. Results
For student use, the imaging sample should be inexpen-
Data acquisition subsystem of the sive and easy to image. A DVD+R disc can be split along the
student STM glue line into two layers. The nonlabel-side layer is a perfect
• Acquisition card calibration standard with pitch spacing of 740 nm.6 The silver
The data acquisition card requirements include four digi- foil provides an excellent conducting layer soft enough that
tal-to-analog converter channels to drive three axes of motion when the STM slider is crashed into the tip, a large dent will
and the bias, one analog-to-digital converter channel to read be visible in the image but the tip will continue to function.
the tunnel current, and two digital channels to control the For other samples, commercial calibration standards can be
stepper motor. All analog channels should have 16-bit resolu- purchased from many vendors.7

540 The Physics Teacher ◆ Vol. 53, December 2015

 The student STM system can image graphite with certain critical dimensions using the atomic force microscope,” J. Vac.
modifications. The horizontal scan voltages must be reduced Sci. Technol. A 17, 1457 (1999).
by a factor of 10 so that an image size of 1 nm is available. 7. For example, http://www.lightsmyth.com/products/standard_
The software must be modified to run in feedback off mode reflection_gratings/index.php , accessed Sept. 2014.
at high scan speed. Usually the system needs to settle for
Tom Ekkens is a professor of physics and chair of the physics depart-
about 10 minutes after approach to maintain the stability ment at Walla Walla University. He built his first STM as part of his PhD
required to stay in tunneling range for the entire scan with- program at the University of Notre Dame. He built scanning electron
out drifting out of tunneling range. Working parameters for microscopes at a company in Silicon Valley before returning to academia
imaging graphite on this design are a bias voltage of 0.060 V, in 2004. tom.ekkens@wallawalla.edu
scan time per line of 15–50 mm, and scan size of 1–5 nm.

Conclusion
The student STM can be built for under
$100, excluding the cost of the data acquisition
card and tooling. The mechanical system can
be constructed in one three-hour lab period.
The electrical system can be built in the second
three-hour lab period. The final three-hour lab
period is used for troubleshooting and learning
to take pictures.
Student response to this project has been
positive. Several students were so excited by
the project that they wrote an article about
their lab experience in the student paper and Fig. 3. In this picture, the limits of the scan range are explored. On the left is
arranged a short presentation to the entire a DVD disc scanned with a field of view of 3560 nm. On the right is graphite
student body. All the groups in the most recent scanned with a field of view of 3.1 nm.
class reported resolution better than 300 nm. Table I. Bill of Materials - Mechanical Subsystem
The most valuable lesson students learn in this
lab is that scientific discovery is less about buy- Description Part Unit Cost Source
Number
ing the perfect tool and more about creativity
with available resources. • ¾ Slide Repair Coupling 032888605046 $3.98 Home Depot
The website http://physics.wallawalla.edu/ • 2x ½-in Plug MPT 049081143145 $0.77 Home Depot
studentstm contains more information regard- • Reducer Bushing 1¼ to 1½ 034481062271 $2.41 Home Depot
ing STM operation, this build process, and the
microscope control software. • 5–7-in Hose Clamp 078575179957 $1.98 Home Depot
• 27-mm Brass Piezo with $1.00 EBay
References 20 mm PZT (4.6-kHz
1. Douglas C. Giancoli, Physics: Principles with resonant frequency,
Applications, 7th ed. (Addison-Wesley, Boston, 16,000 pF at 100 Hz)
2014), pp. 785–786.
• Phenolic Standoff ¾-in $1.00 EBay
2. Frank J. Owens and Charles P. Poole Jr., The long x ¼-in diameter
Physics and Chemistry of Nanosolids (Wiley-
Interscience, Hoboken, NJ, 2008), pp. 54–66. • Tip holder (DIP Socket) $1.00 EBay
3. John D. Alexander, A Simple Scanning Tunnel- • Pt/Ir tip wire 20105 $2.00/cm nanoscience
ing Microscope (STM), http://www.reocities. • Circuit board material $5.00 Ebay
com/spm_stm/Project.html , accessed Aug. ROB-09238
• Stepper Motor $14.95 Sparkfun
2014. This website was originally hosted at
http://www.geocities.com/spm_stm/Project. • Stepper Motor Controller ROB-11699 $19.95 Sparkfun
html before Geocites closed in 2009.
• 2-56 x 1-in Threaded Rod $1.00 Ebay
4. Maxim Shusteff, The teaching AFM, http://
www.media.mit.edu/nanoscale/courses/AFM- • 2-56 Metal Nut $1.00 Ebay
site/positioning.html , accessed Sept. 2014. • 2x 2-56 Nylon Bolt $1.00 Ebay
5. http://physics.wallawalla.edu/studentstm .
• 2–5-mm Coupler D9L25 $3.78 Ebay
6. Donald A. Chernoff and David L. Burkhead,
“Automated, high precision measurement of • 30-gauge magnet wire $5.00 Ebay

The Physics Teacher ◆ Vol. 53, December 2015 541