Nanonis STM Simulator Tutorial

Nanonis STM Simulator Tutorial

Software Version 4 Manual Version 4.0

Contents Introduction ............................................................................................................................................................................... 4 Minimum System Requirements and Installation ..................................................................................................... 5 Getting Started .......................................................................................................................................................................... 6 Session Directories............................................................................................................................................................. 6 Online Help ............................................................................................................................................................................ 6 System Startup ..................................................................................................................................................................... 6 Running the Z-Feedback .................................................................................................................................................. 7 Tip Home Position, TipLift™ ........................................................................................................................................... 8 Scan Control ............................................................................................................................................................................... 9 Setting the scan parameters ......................................................................................................................................... 10 Saving parameters with an Image ............................................................................................................................. 10 Global Counter............................................................................................................................................................... 11 Cross sectioning Scan Data ........................................................................................................................................... 12 Advanced Scan Options .................................................................................................................................................. 12 Changing the View of an Image ................................................................................................................................... 14 Using the Mouse ................................................................................................................................................................ 14 Choosing the Data to Display ....................................................................................................................................... 14 Online Image Processing ............................................................................................................................................... 14 Pasting Data into the Background ............................................................................................................................. 15 Changing the Color Scale ............................................................................................................................................... 15 Color Palettes................................................................................................................................................................. 16 Saving an Image................................................................................................................................................................. 16 Viewing Several Channels in Parallel ....................................................................................................................... 17 Quad-Scan Monitor – four Views in one Window ............................................................................................... 18 Line Monitor ....................................................................................................................................................................... 19 2D Plot Controls ................................................................................................................................................................ 19 Bias Module ......................................................................................................................................................................... 21 Understanding the Feedback Loop ................................................................................................................................ 23 Purpose of feedback ........................................................................................................................................................ 23 Working Principle............................................................................................................................................................. 23 Bandwidth Considerations ........................................................................................................................................... 24

NANONIS | STM Simulator | 2

Signal Oversampling ........................................................................................................................................................ 27 Spectroscopy ........................................................................................................................................................................... 28 Bias Spectroscopy............................................................................................................................................................. 28 Z Spectroscopy ................................................................................................................................................................... 31 Spectroscopy Locations ................................................................................................................................................. 31 Grid .................................................................................................................................................................................... 32 Line .................................................................................................................................................................................... 33 Cloud ................................................................................................................................................................................. 33 Coarse Approach .................................................................................................................................................................... 35 Automatic approach ........................................................................................................................................................ 35 Advanced automatic approach parameters ........................................................................................................... 36 Displaying Data in Graphs.................................................................................................................................................. 37 Time domain display ....................................................................................................................................................... 37 Frequency domain display............................................................................................................................................ 39 Long Term Spectrum....................................................................................................................................................... 40 Sample tilt correction .......................................................................................................................................................... 41 Thermal Drift ........................................................................................................................................................................... 43 Drift Compensation .......................................................................................................................................................... 43 Atom Tracking.................................................................................................................................................................... 43 Contact AFM............................................................................................................................................................................. 45 Defining a new Z Controller.......................................................................................................................................... 45 Detector calibration using Force Distance curves .............................................................................................. 46 Detector calibration using lockin module .............................................................................................................. 48 Atomic Manipulation............................................................................................................................................................ 50 Diagnostics and Analysis .................................................................................................................................................... 52 TCP Receiver ....................................................................................................................................................................... 52 Diving deeper .......................................................................................................................................................................... 54


Introduction Thank you for downloading the SPM Simulator package from Nanonis. This program has many powerful features that provide benefits in a variety of situations.  





It can be used to learn about the capabilities of the Nanonis controller before one is purchased. It can also be used to teach other people how to use the software without the presence of the hardware. This can be particularly useful as a presentation tool in a group meeting or anytime you are separated from the hardware. A third application is as a general purpose SPM teaching tool. Because of the accuracy and depth of the simulated microscope backend, fundamental principles such as feedback optimization, important operating parameters, and data acquisition modes can be learned when first starting out in the field of SPM. Lastly, existing Nanonis customers find the simulator an invaluable tool to debug and troubleshoot LabVIEW development using the programming interface because it does not tie up the PC used to operate the microscope but still provides the full capabilities of the environment to make sure the new program works as intended before it is placed into service on the main PC.

The document is not meant as an introduction into scanning tunneling microscope, the reader is expected to be already familiar with the basic working principle of an STM.


Minimum System Requirements and Installation Please make sure your computer meets the following requirements:    

Windows XP/Vista or Windows 2000 operating system. Minimum of 1 GB RAM Minimum of 1.5 GHz processor Screen resolution of at least 1280x1024 pixels is recommended. With a lower resolution some windows will be partly hidden and you will not be able to use them.

In the case that your computer does not fulfill these requirements, you might experience sluggish behavior. The installer is a regular Windows Installer, available at http://sim.nanonis.com. The installer will guide you through the installation procedure. If you have a previous version of the Demo software installed, the installer will uninstall that version first.


Getting Started The first time the program is started a message may appear warning you a program is trying to access a port blocked via a firewall. This port must be unblocked in order for the main program to communicate with the background process that is simulating the actual Nanonis hardware. This communication is through the internal loopback device present in all network protocol implementations. No external communication outside of the computer takes place. When operating host PC and the Nanonis standard controller connected via a network cable. Once communication is established, you will next be asked to accept the license agreement. Click the Accept button to then launch the main program.

Session Directories A dialog will appear asking to define a session directory. The concept of sessions can be very powerful in a multi-user or multi-microscope A Nanonis Session enables you to situation. Sessions allow the isolation of experimental have measurement data and the conditions, screen layouts, microscope calibration corresponding configuration factors, etc. to be easily organized. A great example is parameters in one single place. a low temperature microscope that has different piezo calibrations at room temperature and low temperature. Instead of entering the new piezo factors when the temperature changes, everything can be updated by simply saving one session file and opening a different one. To get started in the tutorial, browse into a specific directory and then click the Choose Curr Directory button in the window to place all session information in this folder.

Online Help There is a great deal of online help available throughout the program. Almost every button and entry box has a tip strip associated with it. By leaving the mouse cursor over the item, a small box appears with a short text description of the item. Context help is available within each window by pressing F1. A separate window will open with a detailed discussion about every item contained within a particular window.

System Startup When the program is launched, the main window will appear as shown in Figure 1. To get started with the program the most important modules to open are under the Modules menu. Three to open right away are Z Controller, Bias, and Scan Control.

Figure 1: Main Window


Running the Z-Feedback The next window to introduce is the Z Controller window shown in Figure 2. This is used to control the feedback bandwidth, control the working setpoint, monitor the instantaneous z piezo position, and configure the SafeTip feature. When the program is started, the feedback loop is deactivated and the tip is held fully retracted. To start with a known set of working parameters, choose File/Load Program Defaults. The setpoint, proportional gain, and integral response are all set to values that are known to work for stable control of the simulated microscope.

Figure 2: Z-Controller window.

Once the parameters are set, the tip can be engaged. Since the simulator is built to behave like the tip is just above the surface, there is no need for a coarse approach to occur. The steps required to configure coarse approach will be covered later. For now, simply click the large engage button on the right side of the window. The red pointer on the scale moves downward as the tip is ramped towards the surface. It will reach feedback around zero. The Things you can do with a digital current will change from almost zero to the requested feedback: stop the feedback setpoint value and then be reasonably steady. At this instantaneously, keep the tip at point the tip height is actively controlled by the a constant height without feedback loop. The digital readout of the z position droop, lift the tip by an exactly will start to fluctuate and you may also see small defined amount, withdraw motion of the blue indicator. Similar to the other without delay in case a signal controls, the dynamic range of the z scale can be reaches a certain threshold or changed to cover a small part of the full z piezo range control on combinations of for sensitivity or it can be left at the default scale signals. where the entire z range is displayed. The digital feedback loop of the system has proportional (P) and integral (I) gain. The speed of the loop response to a deviation from the setpoint is determined by the combination of the P and I factors. The larger the P value the faster the loop will respond. For integral gain the units can be time constant or the inverse of the time constant. If you feel more comfortable working in the units of a time constant flip the toggle switch


on the left side of the integral control. Similar to the other controls in the program, the range of the P and I gain can be changed using the arrow buttons at the end of the scale. Use this to either increase the sensitivity so small changes are possible or to provide more dynamic range for large adjustments. More details about the feedback system will be presented later. For now, the default values of the simulator will provide reasonable imaging performance. Any time while the loop is active or inactive, the Setpoint can be changed using the appropriate control. It has both a slider control and number entry box available. In the slider control is a yellow bar graph that fills to indicate the instantaneous value of the current. The scale of the bar graph is logarithmic so be sure to properly interpret the graph. The range of the bar graph is controlled using the set of buttons above the indicator. The + and – buttons decrease and increase the range of the bar graph limits, the fs buttons automatically changes the scale to range from 0 to the maximum current (when using an STM feedback signal) as determined by the preamplifier gain. The 0 button places 0 in the middle of the range and the C button places the measured value when the button is pressed in the center of the range.

Tip Home Position, TipLift™ Another important term to understand is the concept of the Home position for the tip. To park the tip at a fixed position and have the digital feedback loop then hold it there, define the location using the Home Position setting and then click the Home button. Keep in mind the defined position is on an absolute scale. To hold the tip a specific height over the surface it is preferable to use the Tip Lift feature instead of the Home button. For example, if the z position is +50 nm and you want to hold the tip 8 Angstrom away from its current position, the Home Position should be defined as 49.2 nm. If the current z position drifts to 20 nm, the Home Position would have to be redefined to be 19.2 nm. It is far easier to define the Tip Lift distance to be 8 Angstrom and for any position, the tip will be pulled back that distance from the present position when the loop is disabled. With the simulated STM, the current will drop to zero if the tip is pulled back more than a few Angstroms from its tunneling position.


Scan Control Once tunneling has been achieved, a scan can be started and images will be obtained. The image configuration is contained in the Scan Control window shown in Figure 3. There are three main groups of controls in the window. The Scan, Follow Me, and Grid buttons can be used to change the configuration display on the left side of the window. Note these can be changed while acquisition or other function is taking place without interrupting the current mode. This provides a chance to make changes to various sets of parameters in anticipation of the current operation completing and starting the next operation right away instead of waiting to configure it and then starting.

Figure 3: Scan Control, all the activity that involve moving the tip like scanning, spectroscopy or manipulation are done from within this module.

The most commonly used group is the Scan set of tabs. They will be used for image acquisition and scanning. The first step should be to activate the imaging channels to be acquired at each pixel location. The list is contained in the Scan tab. Click on the name of each channel to highlight it for activation. To acquire multiple channels during the scan, hold the Ctrl key and click on each channel to highlight it. By default the Topography (Z(m)) and Current channels are active. The pixel density is also configured in the same area of the tab. By clicking the link control next to the window, the Pixels and Lines can be forced to always equal each other for ease of configuration (only one needs


to be changed), but they can also be made independent to have a different number of pixels and lines if desired. To begin a scan, click the button with the downward facing arrow. This will commence a scan from the top of the frame towards the bottom. The image will appear on a line by line basis and a two dimensional image will be built up as shown in Figure 4.

Setting the scan parameters The position and size of the scan frame is set in the top of the tab. Values can be entered directly here or the mouse can be used to determine the size as described below and when that is done, these values will be updated to show the Figure 4: The image starts to appear as each linescan is acquired new settings. Next to the Size parameters is a control that can be set to lock the X and Y axis to be equal at all times (scanned region will be a square) or unlocked so the region will be rectangular in shape. The speed of the tip as it scans over the surface is determined by the parameters in the Speed section. The time required for one line can be fixed (Time/line) which means the actual velocity of the tip over the surface will change as the scan size is changed. Alternatively, the tip velocity can be fixed which means the time required for each scan line will vary as the scan frame size is changed. Click on the lock icon to toggle which parameter is fixed and which one varies. Note that either parameter can be changed to change the acquisition speed, the lock merely determines which is fixed as the frame is changed. As any of the parameters that effect the total time to acquire one frame are changed, the Time/frame is updated to reflect the amount of time that each image will require. Also, on the right side of the window is a countdown timer that displays the amount of time left to complete the currently acquired image.

Saving parameters with an Image

Figure 5: Save tab in the Scan Control window. Choose a basename for automatic filename generation, add a comment to each file, and choose the parameters to save in each file header

To automatically form the name of a file as data is saved, the Save tab (Figure 5) should be configured next. The Basename is the text string that every fill will begin with. The Image number is then the value incremented for each file as it is saved to insure the files are uniquely named and also sequentially numbered for easy browsing later.


There is now a long list of special codes that can be entered as part of the basename to help uniquely identify files and provide more functionality in the naming convention. Special codes that can be entered are the following %a

abbreviated weekday name

%b abbreviated month name %c locale-specific date and time %d

day of month

%H hour of current time (in 24 hour format) %I hour of current time (in 12 hour format) %m

month number

%M time

minute of current

%p

am/pm designation

%S

second of current time

%x

locale-specific date

%X

locale-specific time

%y

last two digits of year

%Y

four digits of year

Figure 6: Use the analyze tab to measure features in the image while data is still being acquired

So, using the code %Y-%m-%d would produce something like 2008-09-05 and %H%M%S would produce something like 022405 (somebody is working late in the lab). With the global counter you can generate filenames that can easily be sorted chronologically without relying on the time stamp to be handled correctly by the operating system.

Global Counter The code %N translates into a number which is given by a global counter. The counter can be set and reset in the Options window available from the main window menu. The global counter will increase by 1 every time it is accessed. For example, if you want to sort your files in the order


you acquired them, you could name them %N-ScanData, %N-BiasSpec and %N-ZSpec, which would generate a list of file with a running counter for all three file types.

Cross sectioning Scan Data The third tab present when the Scan grouping is active is the new Analyze tab (Figure 6). This can be used to take cross sections and also measure distances and angles on the surface while acquisition continues. There is no longer a need to wait for an image to be finished and saved in order to open it in a separate module that has measurement tools. When the Analyze tab is selected a line appears on the image. Click on either endpoint to Figure 7: Cross section of data extracted from the Scan Control window drag it to a new location. As the line gets longer or shorter the total length of it is displayed in the information window. This can be used as a ruler to measure features in the image. By grabbing in the middle of the line with the mouse, the line can be shifted laterally while maintaining its length and angle orientation. To view a cross section of the data covered by the line, click the Cross Section button and a separate graph window will open to display the data (Figure 7). As the line is then moved around the plot will update in real time. To determine the angle between two crystallographic axes on the surface, the angle button should be clicked. This adds a dashed line to the image area. Both of them can be moved around and the display will show the length of each one as well as the angle between them.

Advanced Scan Options Less frequently changed parameters related to scanning are contained in the Options window (Figure 8) opened from the Tools dropdown menu. Scanning can always be in the same slow direction (from the top each time or from the bottom each time) or it can alternate between scanning from top to bottom and then bottom to top. This will save the time required to return to the origin each time and also remove piezo creep from the image Under Advanced Scan Options resulting from the quick motion back to the origin. To you find many things like where activate this mode, check the Bounce Scan box so the the tip should stay at the end of tip “bounces” off the bottom of the frame when it is a scan, whether the last image reached. When the Stop button is pressed and should be pasted automatically scanning stops, the tip can be moved to a fixed in the background, or if the location every time for consistency. To use this, check forward and backward speed of the Custom end of scan position box and enter the a scan should be the same or percentage of full piezo range to place the tip. Values different. of 0,0 would place the high voltage outputs to zero and center the tip in the middle of the scanable area. If


the box is not checked, the tip motion is halted immediately and the tip remains fixed at the position when the scan is stopped until moved again. The tip position can always be determined by locating the blue dot in the viewing area. Other options in this window are used to determine the image painting of the data when various processes occur. Since the data will appear in the window, the preceding data can be erased when moving or rotating the frame to avoid confusion. If the frame is moved partway through an image a split will be evident since the frame will shift over the surface thereby producing an offset of the features. There are times where this is helpful to see but other instances where it will create confusion. Turn the feature on or off using this checkbox. It is also possible to Figure 8: Less frequently used parameters are contained in the choose to have the frame erased Options window accessed from the tools menu when the next frame starts. This can be activated by checking the box labeled starting a new scan. The data can also be pasted into the viewing area to compare to future images or also to be used as a navigational aid as the frame is zoomed and moved over the surface. To automatically paste the last image into the area choose the condition with the next set of boxes. The final parameter related to scanning in this window is the Backward linear speed. By default the reverse scan takes the same amount of time as the forward scan. This can be changed to be faster or slower using the ratio parameter or entering a fixed speed with the Custom box. This can be helpful to speed up total acquisition if the image in one direction is sufficient and the tip can be rapidly returned to start the next line. Be careful about maintaining feedback control of the tip height when using a fast return speed though. When the end of the frame is reached, the next scan will start if the adjacent Continuous Scan button is activated. This scan will either proceed in the same While a scan is paused, you can direction as the previous one or reverse the slow scan use follow me to move the tip to direction depending on the status of the Bounce Scan a new position, do spectroscope check box discussed above. The next button in the row there and then resume the scan can be used to pause the scan and hold the tip at the from where you paused. position when the button was pressed. Any number of functions can then be performed exactly at that spot (voltage pulse, spectroscopy, external equipment acquisition, etc.) and when finished the button can be released and the image acquisition will continue. The next button will halt the slow scan


increment and continue to raster the tip back and forth at the same line position. This can be very handy to alter other conditions and check the response to the change using the line scan monitor window. The final button stops the scan.

Changing the View of an Image Above the image viewing area in the right corner are a set of buttons used to change the view of the image. The paradigm used in this control is a camera floating over the image area which can be panned, zoomed, and rotated. The mouse icon changes as a visual reminder of what camera parameter will be changed. To pan (shift laterally) the view, click the hand icon. Choosing the next magnifying glass icon allows the zoom factor of the view to be altered using the scroll wheel of the mouse. The next two icons zoom in our out a fixed amount each time one of them is clicked. The next icon adjusts the view so the current scan frame exactly fills the window. Keep in mind if the frame is rotated with respect to the normal X and Y axes, the camera angle will be rotated to match the frame axes so it will no longer look rotated, but the actual scan will still be occurring at the designated angle. Any image can be pasted into the background of the area to be preserved as a navigational aid. The details of pasting an image will be covered below. The next button in the row adjusts the view to exactly coincide with the presently pasted image. The final button goes to a full scale view equal to the maximum reachable area given the current piezo calibration factors.

Using the Mouse The row of buttons on the left side above the viewing area are used to adjust the scan frame parameters using the mouse instead of entering numbers in the The mouse wheel works in the entry boxes. When the button on the left side is picked, scan control window to zoom in the scan frame can be grabbed and shifted laterally. and out! This physically moves the frame to a different part of the surface. Clicking the next button will allow the operator to rotate the frame using the mouse. The next button can be used to grab the corner of the frame and resize it to scan over a larger/smaller portion of the surface. The final button is used to also change the size, but in this case the center of the frame remains fixed and the image size increases or decreases symmetrically about the center of the current frame. When right-clicking the mouse before depressing the left button, any action is canceled. The mouse wheel zooms in and out in the viewer. The size of the scan frame of course stays constant.

Choosing the Data to Display On the right side of the window is the Scan Display section. This is used to determine what data channel is displayed in the area, which direction to display, what processing to apply to the data before showing it, and the controls to paste and image into the window. If more than one channel is acquired, all available image channels are displayed in a list when clicking the Channel control. Directly below that is a toggle switch to determine if the forward or reverse image is displayed.

Online Image Processing The data can also be processed before being displayed to remove any slope or offsets due to tip changes. Since the color map continuously adjusts to always span the largest and smallest values,


contrast might be lost if the data spreads over a large range due to artificial reasons. The preprocessing steps that can be applied are Subtract Average, Subtract Slope, Subtract Average and Slope, and Differentiate. The first choice will set the average value of each line scan to zero. This is very helpful to remove offset of scan lines due to tip changes or slow thermal drift in the microscope. Subtract Slope will fit the best line to each individual line scan and then subtract this from the data and display the result. This is most useful to remove slope along the fast scan direction and allow the color scale to be used over the small variations on the surface instead of the larger range covered by the slope. Subtract Average and Slope removes the slope of each line and also sets the average value to zero for each line. Differentiate will calculate the derivative of the line scan and display the result. This can be useful to accentuate small changes in the data which might be relevant in determining if the experiment is working as expected. The finals set of controls will place an image into the window.

Pasting Data into the Background Pressing the stamp icon copies the currently acquired data frame into the window which will remain visible as subsequent frames are acquired. This provides a You can automatically paste very convenient way to leave a large area view of the every scan in the background by surface on the screen while zooming and moving pressing the All button next to around to concentrate on specific features of the the stamp icon surface. To have a complete image pasted into the window, click the Next button. It will change color to highlight that the current frame will be pasted into the window when it is completed. Pressing the All toggle button will paste every frame into the window upon completion and erase the previous one.

Changing the Color Scale The color scale and appearance of the image can be changed using the controls in the Color Scale section of the window. The sliders can be moved to determine the physical value that corresponds to the highest color and the lowest color. Try manually adjusting their position and note their effect on the data. The small buttons next to the sliders determine the total range of the scale that the sliders can be set to. For example if the data values exceed the upper end of the scale, the upper slider will not be able to reach the value so all data above the slider position will be the same color. The scale should be reduced to access larger physical values to remove the color saturation. This is achieved using the minus button. Once the scale is changed, adjust the sliders to better use all of the colors over the entire data range. If the scale is large and the data range is very small, there will be poor control when trying to make adjustments. Use the plus button to zoom the scale so the sliders can be accurately adjusted over a small range of values to an appealing appearance of the image. To automatically have the slider range adjust to The color scale as well as the the maximum and minimum physical values color palette can be changed currently displayed, use the as button. To have even while scanning. the slider range be set to the absolute maximum


and minimum values (+/- 10 V for example) use the fs button. To set the center of the slider range to the center of the physical values, press the C button. To pick a new color palette for the image click the Palette control and select a new one from the list. Color Palettes Color Palettes can be added or changed using the palette control window accessed from the Tools/Options menu. The tool used to define and create palettes is shown in Figure 9. The buttons along the bottom can be used to add new palettes, clone an existing palette as a starting point, delete a palette, or save a redefined palette into the settings file. The small markers next to the color scale can be dragged to change the mapping and linearity of the map. New ones can be added by right clicking when over the scale and choosing Figure 9: With the color palette designer, new palettes can be created, stored and loaded. New Marker. Right click when the cursor is over a marker and choose Marker color to redefine the color at this position.

Saving an Image To permanently save a set of images, use the save buttons at the top of the window. The main save icon looks like a floppy disk. Pressing it will write the data to the file immediately, but keep in mind this will be only a partial image. This can then be opened for analysis while the rest of the image finishes which can be helpful in determining if an experiment is working properly without waiting for the entire scan to finish which often can be a long time with slow scans. To save a complete image when the scan completes, click the Next button which will change its appearance to indicate the current frame will be written to a file when the scan reaches the end. To save all images after each frame completes, depress the All button. When combined with Images can be saved individually the autonaming scheme configured in the Save tab, each or automatically at the end of file will be written to the disk and will have a unique every scan. name sequentially numbered.


Viewing Several Channels in Parallel The Scan Control window is the main navigational tool used to set the scan parameters and view the data as it is acquired. If more than one channel is acquired during the scan the image to display in the control window can be changed. This can be useful if the acquisition conditions are such that an important feature is visible in one channel but not others. The relevant channel can be loaded and the scan frame moved around relative to this visible feature instead of driving blindly and hoping to locate the frame in the correct region. To change the displayed channel, switch to the Display tab and pick the new display channel from the Channels dropdown list. All active data buffers will be present in the list and after a new one is picked, the image changes immediately to the new channel. Viewing more than one data channel during acquisition is a simple task. Additional windows can be opened to view other Figure 10: The scan monitor lets you view a second data channel, independent of what is display in the scan channels, but these are view only and cannot control window. be used for navigation. It is usually a good idea to open the images of a channel formed by both scan directions to check for reproducibility and that no scan artifacts are present. At the top of the Scan Control window is the Tools menu which can be used to open multiple image window. These are called Scan Monitor A and Scan Monitor B. An example is shown in Figure 10. Note the window looks similar to the Scan Control window with the color mapping tools. There are controls to change the active color map as well as the auxiliary colors for text, labels, etc. Once a window is opened the displayed channel can be changed using the dropdown menu similar to the Scan Control window. The same processing can also be applied to each line scan before display just like the scan control window. Images can also be pasted into these windows for storage and comparison to later scans using the Paste button as described above for the Scan Control window. Along the top of the viewing area the name of the channel and the total range of the data currently are displayed. This information can be turned on and off using the check boxes in the upper right corner. The bottom part of the viewing area also contains a scale bar to be used for estimating feature sizes as they

You can fix the camera of the scan monitor to the main camera or the scan frame.


appear in an image. The color mapping of the data is automatically adjusted after each line of data is displayed in the image. Occasionally, a glitch or excessive noise will skew the color mapping and it will not be optimal. The high and low color values can be manually adjusted for a better image appearance using the controls. The data values that correspond to color 0 and color 256 are shown in the top of the window. A slider is also presented that contains a green bar to represent the range of the data. The two small red controls can be used to adjust the physical value of color zero and color 256. If the color map is automatically adjusted, the sliders will be at the end of the green bar as shown in the figure. Similar to other parts of the program, there are a set of buttons to control the range of the displayed data values. Use this to gain finer control of the slider limits or to allow the sliders to be adjusted over a very large range of values.

Quad-Scan Monitor – four Views in one Window One additional window that can be opened to view multiple data channels is the Quad Scan Monitor which is also accessed from the Tools menu. This window has four small panels which can be individually configured to show other channels or the same data channel but with a different color map or processing applied. An example with the forward and reverse Current channel and forward and reverse Topography channel is shown Figure 11. Each panel has its own set of controls. To change the settings for any pane, click on it so the border is highlighted in red and then use the three tabs at the top of the window to change the settings and colors for this pane. Each pane can also have an image pasted into the background similar to the monitor windows. One camera angle is applied to all four panes. Use the controls on the right side of the window to change this view point for all panes simultaneously.

Figure 11: The Quad Scan Monitor is particularly useful in NC AFM to view Topography, Amplitude, Phase, and Dissipation all at once


Line Monitor Once the scan is set up, it may also be useful to open a graph window to display each line scan as it is acquired. This is accomplished by choosing Line Monitor A from the Tools menu in the Scan Control window. A window as shown in Figure 12 will appear. Up to four sets of line scans can be shown in the window. It is very helpful to display the last 1, 2, 3, or 4 line scan on the same graph to compare their appearance as the feedback loop parameters are adjusted. Features may appear sharper or more rounded and noise may appear or disappear. This information is very valuable in deciding the optimal Figure 12: Successive scan lines displayed in the Line Monitor feedback settings. Each line scan is offset along the y axis by the amount while the tip moves over a step edge selected in the lower left corner. The closer together they are, the smaller the total y range so features can be seen easier. However, confusion can occur if the line scans overlap. Do not offset the lines by a large amount in an attempt to avoid overlap as this will create a y scale with such a large dynamic range that the plots will appear almost flat and little structure will be visible. Notice there is a small triangle symbol in the upper left corner of the entry box. Whenever this symbol is present, right clicking on the box will open a quick access menu to change the value by a fixed percentage instead of absolute numerical values. Often this is a quicker and easier adjustment to make rather than entering numbers.

2D Plot Controls The forward or backward scan display can be turned off by unchecking the box in the lower right corner next to the line. Next to each direction name is a small icon showing the currently selected color and background. Click this icon for a menu which allows a great deal of configurability for the graph window. The graph style can be changed from the Common Plots menu. The plot can be lines only, data points only, points and lines combined, bar graphs, etc. Experiment with different plot styles for each type of data plot as some can provide more insight into the data compared to other styles. The color is changed in the next menu item, and the The controls for graphs and plots third one on the list is used to select the line type when are taken right from LabVIEW. a line plot is used. The line thickness can also be


changed for each plot and the graphs can also be displayed using smoother lines by choosing the anti-aliased option. Note this effects display of the screen resolution pixels only and is not the same as processing of the actual data points to change the underlying data. For bar plots, a variety of choices exist to determine how the bars are drawn. If a bar plot is chosen, the baseline of the bars is determined by the next setting called Fill Base Line. The Interpolation option is used to decide how the actual data points are connected with a line. Direct segments, right angle segments, right angle segments with the real data in the center are all choices under this menu. When the data points are displayed, the symbol for the points is chosen from the point style menu item. Finally, the axis title for the x and y axes can be turned on or off using the last items. Along the top of the plot window is a control used to choose what active channels are displayed in the window as well as the processing to apply to each line plot before it is graphed. On the right side at the You can edit the number of digits top of the window are a set of buttons used to change displayed at an axis as well as the the x and y axes. The number of significant digits, the display format of the numbers. display units, and the choice of logarithmic and linear axis can be made using these buttons. The color of the plot grid can also be changed using these buttons. There are a set of buttons for each axis. If the first button in the group shows a closed lock, the other controls are disabled and cannot be changed. To manually adjust the axis first click the icon to open the lock and then make any changes desired. To prevent accidental changing of the display, the lock can be clicked again to close it and prevent additional changes. Along the left side of the screen is a collection of buttons known as QuickScale which can be used to make changes to the range of the axis. Clicking the fs (full scale) button makes the plot automatically scale from the minimum physical value to the maximum physical value. Using the as (auto scale) button scales the range to fit the currently displayed plot in the window. The + and – buttons can be used to increase or decrease the dynamic range of the axis for each button click. The lower set of buttons adjust the DC offset of the plot. It can shift the plots within the window up or down each time the arrow buttons are clicked or the 0 button can set the center of the axis to zero. The C button is used to set the average value of QuickScale buttons are an easy the currently displayed data to be the center of way to adjust the axes of a plot. the y axis.


Bias Module The bias module is used to set the tunneling bias. This can be done by typing a value in the display window or using the slider. The parameter entry window can have values entered in multiple ways. The cursor can be placed in any digit location and new values typed. Alternatively, the cursor can be placed in any digit location and the up and down arrow keys on the keyboard can be pressed to increase/decrease the value of that number. Normally, it is the digit to the left of the cursor that is changed. Be careful when the cursor is next Figure 13: The bias pulse and calibration is to the left-most digit. As the number changes and a new hidden and can be accessed through the tools place value digit is added or one is taken away (think menu. about what happens when the up key is pressed when the current value is 9 or the down button is pressed when the digit is 0), the amount the number changes for each key press may change. The slider control has many extra tools that can help its usefulness. On each end of the slider is an arrow button. This can be used to increase or decrease the endpoints of the slider range which can help with fine control of the bias output if the slider is used instead of trying to enter values directly. There are a series of buttons above the slider that can also be You can change the calibration used to help set the slider endpoints. The 0 button is of the bias to reflect an external used to set the center of the slider range to zero. The C divider which you might have put button is used to set the center of the slider range to in the signal path. the current value of the bias. In effect, a DC offset is added to the slider range. The + button will decrease the range of the slider (zoom in) and the – button will increase the range of the control (zoom out). After making changes to the slider control, the fs button can be used to return the range to its full maximum/minimum value. In the case of the bias this would be +/- 10 V by default. The Tools menu can expand the window to include two extra sections. One is used to change the calibration of the bias output DAC. For example, if a voltage divider is installed between the bias output and the sample which attenuates the bias output, the attenuation factor can be entered here so the software can adjust for this. If the output is divided by ten so 1 V out is actually a 100 mV tipsample bias, the user would have to mentally take this into account when setting this bias. It is far easier to enter the 0.1 adjustment factor and the values entered in the software can be the real tunneling bias. If the bias output DAC is not used to control a tip-sample bias and instead controls some other signal, the units can be changed to reflect the new meaning of the Bias DAC. Many times a good way to improve the performance of the tip is to apply a high voltage bias pulse of a short duration. To perform this, the window can be expanded to open the pulse control section. A duration, peak value, and feedback on/off control are included. After setting the values to the desired amount, press the Pulse button to immediately produce a single pulse from the Bias DAC. After the pulse is done, the bias returns to its nominal tunneling value.


One point to note is the simulated microscope image does not accurately reflect the bias behavior of the real Si(111) surface. The image does not switch between the filled and empty state views when the bias polarity is reversed.


Understanding the Feedback Loop Once the image display and line scan graph are opened and configured, a deeper investigation of imaging techniques can be explored. The simulator provides a nearly ideal platform for learning basic SPM principles because complicated interactions due to poorly formed tips or sample preparation techniques are eliminated. When the image quality changes, it is due to fundamental control issues and not uncontrollable changes in the tip or sample surface.

Purpose of feedback The prime determining factor of the image quality is the feedback loop. The purpose of the loop is to maintain a constant feedback signal. This is done by moving the tip up or down to control the tunneling gap width in STM, the cantilever deflection in contact AFM, or the tip-sample separation in non-contact AFM. An STM traces out contours of constant density of electronic states which in the most basic approximation is equal to the surface. Because of the sharp tip in an STM, these contours can be localized to the orbitals of each atomic site which gives rise to its incredible resolution.

Working Principle The feedback loop is always working in the background whether the tip is stationary or scanning. It measures the feedback signal (current in this case) and compares the value to the predetermined setpoint. If the current is not equal, an error signal (current – setpoint) is calculated. If the current is larger than the setpoint, the tip is too close and needs to be pulled away from the surface. If the current is less than the setpoint, the tip is too far away and needs to be pushed towards the surface. The tip is moved by changing the voltage applied to the z piezo. No instrument is ever this perfect When the tip is stationary over the surface, once it is and the simulator can even adjusted to the correct height in an ideal world it would demonstrate this. not need to be moved again because the position would then be fixed. However, there are many factors conspiring to move the tip and sample relative to each other. The various parts of the microscope may have small temperature fluctuations and this will cause them to thermally expand and contract which will produce relative motion. The ambient vibrations in the environment will be oscillating the tip back and forth with respect to the sample which will produce fluctuations in the current as the gap distance oscillates. The electronics of the system will have a noise level both in the current conversion circuit and the high voltage amplifier circuit driving the z piezo. This noise will also create small fluctuations in the input signal. Therefore, the feedback loop has to constantly be watching the input current and making small adjustments to the tip height to maintain a constant input signal. The key factor in the feedback loop's ability to maintain a constant input signal is the system bandwidth. The faster the system can respond to changes, the more “up to date” the control voltage will be kept and the closer the input signal will remain when compared to the setpoint. However, there is an upper limit to the bandwidth and once that is exceeded the system will cross a resonance and become unstable. The simulator will also be able to demonstrate this concept. NANONIS | STM Simulator | 23

Bandwidth Considerations The loop bandwidth is determined by the P and I settings of the feedback loop. A higher P leads to faster response and a slower Time Constant will slow the response down. The I factor can be entered as time by flipping the toggle switch next to the I control to the down position so the units of the I factor is seconds. This means there are two factors to control the bandwidth and learning how to fine tune both of them will take some time. Both can be used to achieve roughly the same effect as demonstrated in Figure 14. In the left line scan plot, the P gain was increased during the scan to change from stable feedback to oscillation. Since the graph shows the last three pairs of scan lines, the older ones show nice control and the atomically resolved lattice. The most recent ones show a large, unstable oscillation. On the right side of the figure, the Time Constant was decreased until the loop started to oscillate which results in the same appearance in the line scans. Try to vary the factors to cause oscillations or stabilize the loop and see how much each needs changed to have an effect on the image.

Figure 14: Line Monitor showing how the feedback changes from stable to unstable due to a change in the proportional gain (left) or integral time constant (right).

At the other end of the possible operating conditions is having a loop response that is too slow. In this case, the loop will not respond to deviations from the setpoint so the tip will glide over the surface at approximately the same mean height. As it goes over features of various heights, the tip sample junction will change which means the tunneling current will change. When this occurs, features will be visible in the current image instead of the topography image. This condition is sometimes referred to as constant height mode as opposed to the more conventional constant current mode. Constant height mode does not need to be operated with the loop explicitly disabled. As long as the loop response is slow enough to no longer follow the small surface variations, the small scale surface features will then appear in the current image instead of the topography image. A nice illustration of this mode is illustrated in Figure 15. During the scan the Time Constant was increased dramatically so the loop cannot react as quickly. There is a very clear transition in the behavior of the two data channels visible in the graphs. As the loop slows down, the atomic resolution in the topography channel disappears and the newest line scans appear relatively featureless except for the very deep corner holes in the lattice. As the corrugation disappears in the topography channel, the variations visible in the current channel increase dramatically. In the


oldest line scan (top of the graph) the current is almost completely flat because the loop is doing an excellent job of maintaining the input signal at the setpoint value. In the newest line scans (lower plots) the current channel shows fluctuations of a few hundred picoamps.

Figure 15: The Time Constant was increased during the scan to reduce the loop bandwidth. The atomic corrugation disappears from the topography channel (left) and begins to appear in the current channel (right).

Since the loop is always active, there is a final important factor that determines the accuracy of the loop when following the surface. This is the scan speed. As the tip moves across the surface and the loop adjusts the height, it is a good idea to give the loop enough time to update the z piezo voltage at each position in order to be confident the surface is tracking correctly. As the scan speed is increased, the loop bandwidth should be increased to maintain good surface reproduction. An example of image quality suffering as the speed is changed without adjusting the loop bandwidth is shown in Figure 16. The line time was changed from almost 700 ms/line to less than 100 ms/line during the second line. The oldest line (top) still shows very nice topography with little variations i n the current channel. The newest pair of lines shows a dramatic increase in the fluctuations on the current channel which is interpreted as meaning the surface is not being followed as accurately and the loop is no longer maintaining a constant current. The topography data still shows reasonably good representation of the surface, but the data is not as high quality as when scanning slower. In most real world cases there is some variation in the feedback channel. The idea is to minimize this as best as possible so the topography channel can be interpreted as being acquired at a steady feedback condition. These three interrelated factors need to be considered at all times when trying to achieve nice images that can also be interpreted properly. The interplay between them can be a bit complicated, but mastery of it can be achieved with practice and patience.


Based on the illustrations discussed above, it would seem a slower scan speed is better because it gives the loop a better chance of following the surface reliably. However, scanning slow is not a

Figure 16: The scan speed was increased during the scan without changing the loop parameters. The decrease in quality in the topography (left) and current (right) indicates the loop cannot follow the surface as accurately

panacea for all problems and introduces a number of points to consider. The most important one is drift. Usually, the system has slow lateral motion caused by pieces expanding and contracting due to temperature changes. On an atomic scale, this drift can cause the surface to appear skewed when it is a substantial amount compared to the total time required to take an image. Other points to consider when selecting a scan speed include the surface dynamics. If the system being studied changes over time, the data needs to be acquired quickly to capture the changes in time over the course of multiple images instead of the evolution occurring within a single, slowly acquired image. Once a good understanding of the feedback control is gained, a good exercise to pursue is to add noise to the system and use the feedback parameters and scan speed to improve the image quality and remove the noise from the system. Noise can be added into the simulated microscope by opening the Nanonis STM Simulator background process. The window should be present in the Task Manager toolbar. Clicking this button will open a window as shown in Figure 17. Increase or decrease the noise in the z channel or the noise on the inputs so the feedback loop see a change in noise on the input channel. There should be a noticeable difference in the line scans as the noise is changed. Once a change can be seen, alter the feedback settings to improve the image quality. More advanced feedback loop control exercises will be presented in later chapters. Figure 17: The simulated microscope background process configuration window


Signal Oversampling With increased noise in the system, this is a good time to investigate the signal averaging algorithms in the system. As everyone knows, taking more readings and averaging them together will effectively increase the resolution past the limit normally set by the 16 bit ADCs. The underlying acquisition rate of the input channels occurs at a fixed rate. As the scan timing slows down, more readings can be taken and averaged together so the resolution increases. Change the noise level in the z direction by a factor of ten or more and look at the effect on the topography channel. The line scans should be noticeably noisier and the image quality degraded. Instead of changing the feedback loop bandwidth of the system to help clean up the line scans, slow the scan down so more readings are averaged together at each pixel. This will also decrease the noise on the topography channel and help improve image quality. The same averaging can be done for spectroscopy data, time domain oscilloscope traces, and noise power spectra introduced later.


Spectroscopy Bias Spectroscopy Point spectroscopy is a very powerful technique used to probe the state of a surface and acquire interesting information besides the simple topographic shape of the surface. In most cases the feedback loop is switched off and a voltage ramped while recording the response of various input signals. When operating an STM, An inherent assumption is that the most common technique is to ramp the the tip is the same height over bias voltage while recording the current. This the surface whenever the can be used to determine the density of states spectroscopy is performed. This of the surface under study. When a is not true and therefore spectroscopy routine disables the feedback switching off the feedback has to loop, the green button in the Z Controller be treated carefully. window will turn off to indicate the loop is disabled. One of the important factors in acquiring good spectroscopic results is to open the loop at the correct time. Since the current depends exponentially on the tip-sample separation, a small deviation can have a large effect on the acquired current. The Nanonis software and controller addresses this situation using averaging. The noise in the system on the topography signal will have a certain peak-to-peak value and when the loop is opened, the tip can be held at any position within the range of the noise. To reduce the spread in value when feedback is disabled, the system will acquire readings for a fixed amount of time and calculate the average z output value over this interval. When the loop is disabled, the tip is fixed to this specific average z value. The longer the time, the closer together the grouping of the tip-sample separation each time the loop is opened. The downside of this is that the total amount of time required for a set of measurements is increased since this averaging will take place each time the loop is disabled. Be careful to not use too Figure 18: Bias spectroscopy configuration window large of a value or the actual acquisition time might become unreasonably long given other factors. If manually switching off the feedback


the parameter that determines the averaging time is Switch off Delay in the lower right corner of the window. When disabling the loop through spectroscopy, the Switch off Delay is configured in the individual spectroscopy parameter window. The window to configure bias spectroscopy is accessed using the Experiments menu and is shown in Figure 18. The first step is selecting what input channel to acquire as the voltage is swept. Any channel listed in the window can be activated and recorded. The parameters next to the channel list determine how many steps will be in the voltage sweep and how many curves to acquire. If more than one sweep is to be obtained, the second sweep can be acquired from the final value back to the initial value. This saves time since the bias does not have to be reset to the initial value. It can also help remove anomalous current induced by capacitive coupling between the tip and sample. Since I=C(dV/dt) there will be a current offset induced by the voltage The Switch-off Delay lets you ramp. If the output is ramped in both directions, in one average the z position over some direction the offset will be positive and in the other time before switching off the direction it will be negative. If this behavior is feedback and holding the tip at observed when plotting the spectroscopy curves, they the averaged position: You will should be averaged together to cancel the two offset get a reproducible tip-height, signals and leave purely the current due to the tipand reproducible spectroscopy sample bias. The loop will be disabled during the results. sweep if the Z controller hold checkbox is marked. When disabling the loop, the output can be averaged for a fixed amount of time and the digital output set to the average value instead of the simultaneous value just as the loop is disabled. The appropriate length of time will depend on the noise fluctuations present in the feedback loop output as well as the desired accuracy in holding the tip over the surface the same distance each time the loop is opened. After the loop is opened, the tip can be moved a fixed amount closer or further from the surface by setting the Z Offset parameter to a non-zero value. Negative values will move the tip closer to the surface and positive values will move the tip away from the surface. By combining this feature with a programming loop using the LabVIEW interface it is possible to perform a series of spectroscopy measurements at various heights over the surface. This can be used to The spectroscopy module can be increase the dynamic range of the spectroscopic combined with the LabVIEW data by pushing the tip closer to the surface Programming Interface to after each curve and exploiting the increase in perform a series of tunneling current as the tip-sample distance measurements at various heights decreases. Dividing by the amplification due to above the surface. the smaller gap will produce normalized data as if it was taken at a fixed distance over the surface, but with better resolution. This method allows as many as 7 orders of magnitude in dynamic range to be measured on the current channel. After the voltage has been ramped to the initial value, it is usually a good idea to wait a fixed amount of time for the system to reach a new equilibrium. Since this jump to the initial value may


be quite large, it may take some time to stabilize. By looking at the signal using an oscilloscope, a good measurement of the required time can be obtained. This delay should be entered in the Initial Settling Time box. Once the output is ramped, the sweep timing determines the acquisition during the ramp. The Slew Rate will be used to move the voltage from one The generic sweeper lets you output value to the next. A higher rate will move the sweep setpoints, lockin voltage quicker, but is more likely to induce artifacts parameters like phase, unless sufficient time is provided for them to amplitude and frequency, or disappear. When each output voltage is reached, the simple things as the bias or a system waits for Settling Time before starting to generic output. sample the input channels in order for the system to equilibrate. Two common examples of when this is needed is the capacitive coupling which induces a DC offset on the current channel proportional to the dV/dt rate and also if a lockin amplifier is in use it will require a few time constant intervals to reach its equilibrium output each time the voltage is changed. The number of readings to average together to form the data point is determined by the Integration Time parameter. A higher number reduces noise, but will increase the amount of time the acquisition requires. Once the ramp is complete, the output returns to its normal value (in the case of bias, it is returned to the tunneling bias value) and the loop is re-enabled. When acquiring spectroscopy during image acquisition, it is a good idea to let the feedback loop adjust the tip height before starting to move the tip to the next image pixel. The amount of time to wait is set by the Z Control Time parameter. Any time a parameter is changed that effects the amount of time required to measure a complete curve, the Sweep time box is updated. The upper and lower limits of the measurement are set in the Limits section. Values can be typed in the entry boxes or the slider control can be used to move the endpoints. To start a measurement at the current location of the Figure 19: Generic Sweep module used for I(z) spectroscopy


tip, press the Start button in the Control section. A sweep can be interrupted once started using the Stop button. Each graph will appear in the lower part of the display after it is finished. Various smoothing algorithms can be applied using the Filter type drop down. After choosing a filter type the parameters of the function are controlled by the Order and Cutoff frq controls directly below the drop down list. To show/hide/delete curves from the graph change to the Display and Save tab. The check box on the far right side can be used to hide or show curves. Clicking a line in the list will allow that curve to be written to a file or deleted. All of the plots can be deleted using the Delete All button. The plots can also be changed to display the numerical derivative of the data by choosing dY/dV from the Function list. Also, the density of states plot can be displayed when plotting I-V data by choosing (dY/dV)/(Y/V) entry from the list.

Z Spectroscopy Another common STM experiment is Z spectroscopy also located under the Experiments section. The interface is identical to the Bias Spectroscopy window, the only difference is the loop is opened and the z piezo is ramped instead of the bias. This produces a curve with an exponential shape due to the nature of the current-z dependence in STM.

The Z Spectroscopy module allows you to quickly sweep the tip position and simultaneously acquire spectroscopy data.

When the spectroscopy data is acquired very slowly and the data needs to be displayed as it is acquired instead of at the end as with normal spectroscopy, the Approach-Retract module should be used. This was originally created to ramp the z piezo and record the cantilever deflection at each point. It is commonly used to measure the surface adhesion and elasticity and to also calibrate the detector when operating in contact AFM mode. To sweep any signal and record the response of other channels, the Generic Sweep (Figure 19) can be used. This module has more timing parameters at your disposal and can be more effective for I-z curves in most cases. First, choose an output channel to be swept and select which channels to record. Then pick an upper and lower limit and after this is done, set the number of points in the curve or the step size. Since one determines the other, choose the parameter to set using the toggle switch on the right side of the control. As one is changed, the grayed out one will be calculated and displayed. The Measurement period determines the number of points averaged together at each step and the Settling time is the interval to wait after the ramp reaches the next step level before data is measured. This provides ample time for any slowly responding signal to reach its new equilibrium value. This module has nearly unlimited potential, bounded only by the imagination of the user. Since there are undedicated outputs, they can be connected to external equipment to ramp any controllable signal (temperature, magnetic field, frequency, etc.) and record a response.

Spectroscopy Locations


The spectroscopy measurements discussed up to now have all occurred wherever the tip happens to be when In grid spectroscopy mode you the acquisition is started. Many times it is desirable to can run all of the spectroscopy acquire spectroscopic maps over the surface to experiments automatically on investigate the spatial dependency of the electronic multiple locations: on a grid, states. For this application the Experiments on a Grid along a line or at arbitrary module should be used. The tip is moved across the locations n the sample. surface and when it reaches a predefined location the motion is stopped and spectroscopic sweeps are performed. The curves acquired at each pixel are indexed so after the complete acquisition is finished they can be individually viewed and associated with the correct location. Additionally, slices can be taken across the three dimensional data to form images at a specific step along the sweep. The first step in setting up the experiment is to determine the size of the grid and the pixel spacing. This is done in the Scan Control window. Grid When the Grid button is first clicked, a grid is overlaid on the currently configured scan area, the size and location of the grid can be changed by typing values into the parameter entry boxes or by using the mouse. Similar to resizing the scan frame in the Scan Control window, the mouse action can be defined to be Scale, Rotate, Move, or Resize. The same buttons are used as when the Scan button is active, but when the Grid button is active the mouse is used to change the size of the grid instead of the image frame. Keep in mind in the new software, the grid can be changed while image acquisition is still going on. Once this is done, click the mouse anywhere in the image area and drag the cursor to perform the desired action. Change the number of pixels within the grid using the X and Y Points control. If the pixel density becomes too large, the individual grid lines disappear and all that remains is the outer border of the grid. Also, note it is possible to have the pixels along the line not be equal to the lines in the frame. An example of a configured grid is shown in Figure 20. After configuring the grid, the spectroscopic experiment to perform at each location must be chosen. This is done using the drop down list in the Experiment section of the window. Be sure the sweep is properly configured in its respective module window before starting the scan. When the start button in the Experiment section is pressed, the software checks the configuration of that module and pulls the

Figure 20: Grid of spectroscopy locations


parameters from there to acquire the data. Be sure to understand the spectroscopy experiment is started using the Start button in the lower left corner of the Grid section and NOT using the Start button at the top of the window. It is also possible to call an external VI at each location which again provides an unparalleled amount of experimental flexibility and control limited only by creativity and imagination. The external VI can communicate with an external piece of equipment for sophisticated integration of other techniques with SPM or it can be a VI written using the programming interface to perform a specialized acquisition not provided by the base package. Unless the spectroscopy is very fast, the total amount of time required to finish one image may be quite long. There is an obvious tradeoff between the pixel density Be sure to activate autosave and total time required. As the number of pixels is within the spectroscopy module increased to get better spatial resolution the used so all of the data is saved as acquisition time also grows. Try using a sufficient it is acquired density to have resolution of the silicon lattice and acquire a full set of data running over night if required. Once this completes and the data is saved, investigate it with the data browser and slice the curves to form images at a specific value along the sweep. Be sure to activate autosave within the spectroscopy module used so all of the data is saved as it is acquired, otherwise it will be lost when the scan is done. Line Instead of performing spectroscopy on a grid of points, it can be very interesting to acquire a family of curves along a line. This provides a nice way to study the spatial dependence of electronic states with respect to a feature on the surface such as a step edge or defect. To choose this, click the Line radio button in the window and then a pattern of x symbols will appear on the image. The number of locations, length of the line, and orientation can all be changed to match the desired spatial location. Cloud The third option for defining spectroscopy locations is the Cloud option. When this is chosen, an arbitrary set of points can be defined by clicking in the viewing area. Each click adds an additional spectroscopy location. This can be particularly useful is there is only a specific region of the surface where spectroscopy is important. Instead of acquiring data over the entire image which would waste quite a bit of time Figure 21: A cloud of arbitrary spectroscopy locations and acquire extra curves, the tip


will stop when it reaches the configure location so the curves are concentrated only in the area of interest on the surface. When the Cloud option is picked three new tools appear in the tool bar above the viewing area. These are used to determine if mouse clicks add points remove already present points or laterally shift an already present point. Also, the section on the left side of the window will present a numerical list of X,Y point pairs as each one is added. An example with a set of points along the step edge is shown in Figure 21.


Coarse Approach When starting the simulator for the first time, the tip was already close enough to the surface that activating the feedback loop created feedback. In a real microscope, of course, this is not usually the case. After placing a tip and sample in the system some means must be available to close the macroscopic gap and get the tip and sample microscopically close. This is achieved many different ways, but usually involves some type of motor that can create very tiny steps in a controlled fashion. In order to avoid tip damage the control software must be able to induce individual steps in the motor and test for a feedback condition between steps. To exercise more caution, the z piezo can be fully retracted before each step and then slowly extended after the step to see if it can reach the surface. If it cannot achieve a feedback signal, the tip is pulled back again and the cycle repeated. The specific interface to drive the motor does not matter for the purposes of this simulator, but most of the time this is achieved by toggling one or more digital TTL lines which communicate with an external piece of equipment. To practice a real-life coarse approach it is necessary to first retract the tip from the surface. Before performing this step, stop any data acquisition that may be taking place. Open the motor control window using the menu item Modules/Motor Control which is shown in Figure 22. Use the Z retract button a few times to fully disengage from the surface.

Figure 22: Manual control of the coarse approach motor is in the Motor Control window

Automatic approach Once the gap is large enough that the feedback controller has the tip fully extended, a coarse approach can be started to simulate the real situation. Open the automatic approach control from the tools menu of the Z Controller window. The window is shown in Figure 23. The first time it is opened, only part of it will be visible. Use the checkbox to view the full size. If the motor takes extremely small steps relative to the full z piezo range (step size ~1/100 of the z piezo range), there is no need to test for a stop condition after each step. Multiple steps can be taken between each testing cycle to speed up the entire approach time required. This is controlled by the Number of Pulses parameter. After each step it is usually a good idea to wait a short amount of time for any large mechanical vibrations caused by the coarse motion to decay away before Figure 23: Computer controlled approach window ramping the tip towards the surface. The interval is set by the Delay after moving parameter. The shorter this time, the faster the approach can be. Use an oscilloscope to look for transients on the feedback signal after each step and use a delay longer than the transient.


Advanced automatic approach parameters The Stop Condition is what z position should be considered in range. It makes little sense to stop the approach the first time the surface is reached because this usually occurs with the z piezo still substantially extended. The first action a user will take after this is to step the motor forward a few more times so the z piezo is in the middle of its range. This manual adjustment can be performed automatically by the software by having it not stop the approach until the z piezo is approximately centered. How closely the condition can be set to zero depends on the coarse step size compared to the z piezo range. The approach stops when the feedback loop output crosses this threshold. If the condition is zero and the n-1 step placed the z piezo at -5 nm so one more step is needed whatever the distance covered by one motor step will be the z position when the approach is stopped after the next step. It is best to have this threshold set to a slightly negative value so when it is crossed the tip may end very close to zero. For unattended approach with some measure of safety, it is recommended to pull the tip back from the surface after the approach is done. This reduces the risk of damage to the tip if an adverse event occurs that would cause tip-sample contact when it is in feedback just above the surface. For example, if the approach occurs during lunch break or class and the tip would be left over the surface for an hour or two with nobody around it is far safer to fully retract the z piezo so nothing bad can happen while nobody is around using the microscope. To utilize this feature make sure the Withdraw when tip is landed box is checked.


Displaying Data in Graphs In effect, there are two distinct types of diagnostic and analysis tools available: a set that displays the data in the time domain and a set that displays the data in the frequency domain. They each have their strengths and appropriate applications. To view a signal in the time domain for a quick measurement of its amplitude there are four different graph modules which are accessed via the Graphs menu. In order of decreasing time resolution the choices are Oscilloscope, Signal Chart, History, and Long Term Signal Chart. The oscilloscope can be used to capture transients, etc. because it features a real triggering function comparable to modern oscilloscopes. Interestingly, all of these modules can be opened and acquiring data at The system has various the same time to provide a view of the same signal on possibilities to display data in different time scales. All data channels are always both the time and frequency acquired at a steady rate and what it is processed and domain. displayed is simply a matter of which windows are opened and how they are configured.

Time domain display The oscilloscope is shown in Figure 24. The controls specific to it are located on the right side of the window. The triggering can be configured along with the time base of the display. Note that changing the Time Base also alters the sampling rate of the data. The Time Base can be changed from 256 ms to a maximum of 12.8 seconds in power of 2 increments. If the length of data displayed is more than desired, the plot can be zoomed into a smaller section using the standard tools discussed above, but keep in mind the time resolution will still be determined by the sampling rate of the original time base. There are three triggering modes, Immediate, Level, and Auto. Use level mode to pick a signal value that should trigger the acquisition to occur and the time trace to be displayed. Use this to capture an event that is known to occur at a specific value. If the trigger occurs and it is to be analyzed and saved, a subsequent event will be ignored if the Hold button is depressed. Auto triggering will continually update the display, but the software will adjust the trigger level to accomplish this. Immediate is more like free running mode where the latest data is always displayed with no trigger taken into account. At any time (whether a trace is held or the scope is in free run mode, the Save Trace button can be pushed to save the data that is displayed

Figure 24: Oscilloscope window


on the screen to a file. The time domain display with the next best resolution is the Signal Chart shown in Figure 25. It can show two channels and also contains a control for additional averaging to be performed. As the Averaging parameter is increased, the time resolution will decrease because more readings from the FPGA card are averaged together by the program, but the rms noise level of the channel will decrease. To change the channel displayed in each pane, click the channel name to produce a list with all available signals. The third display is the History graph which also has two panes for simultaneous display of two signals. At any time the record shown in the window can be saved by clicking the Save button. It can also be cleared and started over again by pressing the Clear button.

Figure 25: Signal chart window

The final display which can acquire a signal over the longest time interval is the Long Term Chart as shown in Figure 26. This will record 10000 data points so the total length of time is determined by the sampling rate. If a signal changes very slowly and needs to be recorded this is the best choice of the four modules. Examples where this could be useful are a temperature or pressure input to an open analog channel, the z controller output to record thermal drift of the head in the z direction, or the X and Y output signals if tracking is activated to record the lateral drift as a function of time. As the points are acquired, the display counts down the remaining time before the buffer fills up. The interval entered in the Delay box determines the spacing between readings.

Figure 26: The Long Term Chart can act as a strip chart recorder and acquire data over a lengthy period of time.


Frequency domain display When searching for noise sources, many times it is advantageous to study the signal in Fourier space (frequency domain) instead of the time domain so the precise frequency components can be seen. This can provide strong clues to their origin. For example, if the peaks are at harmonics of the power line (60 Hz in USA and some other parts of the world, 50 Hz in much of Europe) this indicates grounding or shielding issues that should be investigated. If peaks in the feedback signal are nonexistent when out of range, but appear when in range this is usually caused by vibrations reaching the tip-sample junction. The bandwidth of the spectrum is determined by the sampling range. In the case of the realtime system this depends on the RT Engine Frequency and the Oversampling in use. For example, using the default values of 10 kHz and oversampling ten times means the overall sampling rate is 1 kHz. Since this is reduced by a factor of 2 due to the nature of the FFT algorithm the maximum detectable frequency is 500 Hz. To increase the bandwidth of the Figure 27: The fully functional spectrum analyzer can help pinpoint noise measurement either increase the sources engine frequency or decrease the sampling rate. An example of a spectrum from the simulator (and therefore devoid of any specific peaks) is shown in Figure 27. The maximum frequency is 500 Hz due to the default conditions. Feel free to alter the engine frequency and oversampling and note the difference in the spectrum. A very useful feature for comparing before and after data is to use the Paste function. When this button is clicked the current data is stored in memory and left on the graph using a red line. As the acquisition continues, differences between the previous data and the instantaneously acquired data can be easily seen when comparing the two graphs. Use this to test a set of conditions and then make changes and see if things improve. When testing the simulator, acquire a spectrum and then increase the noise level of the simulated microscope by using the simulator configuration window in the toolbar discussed earlier. As the The spectrum analyer is ideal to noise level is increased or decreased the find ground loops and 50/60Hz baseline of the spectrum will rise or fall. An noise in your system. Any glitch example of this is also seen in the figure. becomes immediately visible.


Long Term Spectrum The spectrum graph can only display two curves at once, a pasted one and the actively acquired one. To look at the spectrum over a longer time scale and notice trends in peak location and peak height, the Long Term Spectrum should be employed. This unique presentation of noise power displays the last 50 curves acquired. The total length of time covered by these curves depends on the bandwidth of the measurement. The longer the time required to obtain one curve, the longer the time covered by the display. An example from the simulator is shown in Figure 28. The bright streak shows a noise peak at 50 Hz that was eliminated about 2/3 of the way into acquisition. If a peak shifted to a different frequency, the band would not be horizontal and instead the bright spots corresponding to the peaks would shift up and down in the image. To see more distinctly the peaks compared to the background level the spectra can be averaged together before presenting. This is determined by the parameters in the Averaging section. There is even a peak hold mode that can be quite Figure 28: Long term spectrum view can be used to observe the interesting and help locate evolution of a signal over time transients that otherwise may be averaged away and lost.


Sample tilt correction Another feature that is useful in some cases is the sample tilt correction. Under normal circumstances when scanning with feedback on, this has no real benefit. If the sample and tip are misaligned, the overall surface slope will be followed by the feedback loop just like it follows the small variations in surface height. If scanning or tip motion will take place with the feedback loop disabled, then it is important to remove the sample tilt in order to minimize the possibility of tip damage. When the sample tilt is corrected, a ratio of the x and y position is digitally added to the z output piezo. For any given position the tip will also be pulled back or pushed forward as it moves in order to follow the first order tilt. When the tip is moved without feedback control and no tilt correction, it will contact the surface as demonstrated in Figure 29. Both the feedback and the tilt correction produce the same tip trajectory (assuming reasonable feedback parameters) so it does not matter which accomplishes the task of following the linear tilt unless one is missing and then the other must be present to avoid tip-surface contact. Two good examples of scanning with the loop open are multi-pass mode where the tip is lifted when the line is repeated in order to study electrical or magnetic effects and lithography routines which require a disabled loop for proper conditions. To understand the process of correcting for sample tilt using the simulator it is necessary to first introduce tilt into the ideal surface produced by the virtual STM. This is accomplished by opening the Simulator window as shown in Figure 31 which is present in the taskbar at the bottom of the screen. Change the Plane x and Plane y values from zero to any arbitrary amount and start a scan. The images will now appear to have slope from one side to the other. Be sure to have the preprocessing mode set to Raw in order to see this. If the preprocessing is set to Slope Subtract, the slope that is present in the real data will be subtracted before display and therefore hidden. It will be in Figure 29: Removing the sample slope automatically allows trouble free scanning with the actual data, but not the displayed images. This feedback disabled is why it is usually a good idea to use either the Raw display mode or only subtract the average z value from each line in order to avoid missing an important point that should be realized about the data. Once the slope is visible in the line scans and the images, start removing the slope by opening the window using Modules/Piezo Calibration (Figure 31). As the scan progresses adjust the Tilt X and Tilt Y parameters until the image looks flat. The best method to do this is use the standard data entry method of placing the cursor next to a digit and use the up/down arrow keys to increase/decrease the value until it is optimized. An example of how it should appear is shown in Figure 30. It is usually easier to only adjust the x factor


and when done, rotate the scan by 90 degrees and now perform the correction for the y axis. The best display to use when judging the removal of tilt is the Line Scan Monitor. Each scan line will appear flatter and flatter as the slope is removed and at some point it will start to tilt the other way indicating overcompensation. Find the values which make the scans appear as horizontal as possible. The top part of the image shows a gradient from Figure 31: Automatic sample tilt correction one side to the other since the color scale is showing the large change in z value over the width of the scan area due to the tilt. As the scan progressed, the correction parameters were changed and at the bottom of the image it appears to be one uniform color because the feedback loop values are now approximately the same value since there is little variation other than the small atomic corrugation. In subsequent images the automatic color optimization will narrow the range of the color scale to span the smaller z values and a high contrast image will appear.

Figure 30: Image appearance with tilt removed automatically (middle section) compared to using the feedback loop to correct the tilt


Thermal Drift The materials used when building a microscope will expand and contract when the temperature of the system changes. Because data acquisition takes place on the nanometer scale, small changes in temperature will be visible in the images by the slow shift in the lateral position of features. If the lateral drift rate is constant, this can be nullified by adding a small correction voltage to the x and y voltages at a constant rate. The simulator can provide experience with this technique.

Drift Compensation Since the simulated microscope works as an ideal system there will be no lateral drift present. By adding a compensating voltage to the scan signals, this will introduce drift into the system. The compensating voltage can be entered by using the Atom Tracking module shown in Figure 32. Begin by setting the Drift X and Drift Y parameters to non-zero values and clicking the activate button. When this is done, the lattice will immediately appear distorted as seen in Figure 33. The observed effect is symmetric to not compensating for drift that is present in a microscope. Try changing the amount of drift and notice the skewing of the lattice increases or decreases. When using a physical microscope, if the lattice appears distorted then the most likely cause is lateral drift that should be compensated.

Atom Tracking The drift can also be measured by the acquisition software. Place the tip directly over an atom on the simulated surface. In a Figure 32: Atom tracking module to null drift real microscope it is better to choose a more distinct feature than an atomic lattice site. Something like a vacancy or an adsorbate will provide an easier feature to follow because the local gradient is much larger. That type of site is relatively steeper given its lateral dimensions are about one atomic space, but it will appear taller or deeper than an atom within the lattice.


The tracking works by orbiting the tip over a site and measuring the derivative of the signal. When the tip encounters a locally steep location it knows it is falling off the atom and moves in the opposite direction in an effort to stay on top of a local maxima (when on top of an atom or adsorbate) or minima (when in a hole). Enter an orbit frequency of a few Hz up to maybe 100 Hz. Given a single atomic site is tracked the orbiting radius should not be very large, a value of 200 pm will work well on the silicon surface. The time constant of the gradient measurement should be slow enough to average over a few oscillations, but not so slow as to not update the tip location very often. This can cause the tip not follow the feature very accurately if it is moving rather quickly. An Integral value of 10 pm/deg/sec is a good starting point. Start the orbit and then adjust the phase until the measurement seems to be tracking well. This can be judged using the red cross in the lower graph. When the phase is well tuned, it will move only along the x axis back and forth. Figure 33: The upper part of the image As the tracking proceeds it will update the measured rate of shows lattice distortion due to drift turned on as the scan was acquired motion along the x and y axis. After some time when small from the bottom towards the top variations have been averaged away, the measured drift should be equal to the value entered above which artificially introduced drift. If this was an actual working microscope that had no drift rate entered in the software to begin with, the measured rate would be entered and the skewed lattice appear like more of the expected result from crystallographic considerations. In the exercise here, drift was introduced artificially and the software was then used to measure the drift rate. The values should match when the The atom tracking lets you measurement was done properly. To return to a automatically compensate for proper lattice appearance, set the drift rates drift and sample tilt! back to zero or simply disable drift compensation by turning off the activate button on the left side of the section.


Contact AFM The controller is flexible enough to operate any microscope in any mode. Another popular mode is contact AFM. A cantilever touches the surface and its deflection measured by a laser beam bouncing off the back of it into a position detector. This mode can also be simulated in the existing software. The back end produces signals like a cantilever is touching the surface at the same time as the STM tip is tunneling. The signal related to this deflection is Input 3 and can be seen in all channel lists. To be convinced it simulates a cantilever signal, retract the tip from the surface and note the magnitude of the signal. Then approach again and notice the signal level changes due to the cantilever contacting the surface and being bent up.

Defining a new Z Controller Create a new feedback signal by opening the Controller Configuration window using the tools menu of the Z controller window. The window is displayed in Figure 34. First enter an arbitrary string for the controller name. In the example shown, it has been named Contact AFM. The control signal should also be named using a meaningful string; this was given the name Force Error in the example. The deflection of the cantilever can be both a positive and negative voltage so it has to be redefined to bipolar using the dropdown list. With the window Figure 34: Configuration of a new feedback source for contact AFM operation open, it is also evident the control signal does not have to be a single analog input. Instead, it can be a complicated mathematical expression derived from two inputs and adding, subtracting, multiplying, or dividing them and using the feedback loop to maintain this relationship constant. A SafeTip condition can also be defined for each controller. This feature allows the system to monitor a second signal from any source and if this signal meets some predefined condition, corrective action is automatically taken to avoid possible tip damage. For example, if the total signal from the quadrant photodiode falls below a certain threshold this usually means the cantilever alignment has developed a problem and the system should pull the tip back because it is no longer correcting to a reliable deflection signal. Another example would be the excitation signal required NANONIS | STM Simulator | 45

to drive the cantilever at its resonance. If this increases substantially this means a large amount of energy is injected into the system to maintain the cantilever resonance which should not be necessary. This probably indicates something else has happened to the oscillating system and it should be investigated because the validity of the acquired data is questionable. Once the new controller has been configured close the window and change the Z Controller to the new choice. The change takes effect immediately so it is a good idea to have the tip retracted when making the change. After the approach is finished make sure the feedback loop is stable and under control. This can be judged by the stability of the Force Error bar graph in the Z Controller section, or by using one of the time domain graphs. A set of parameters that provide pretty good control are a P gain of 35 nm/V and an I Time Constant of 43 msec. An image can be acquired that will show atomic resolution on silicon though this is impossible in contact mode with a real microscope. If everything appears to be working stably, force distance curves should be acquired as the next step. The tip can be pulled back while recording the deflection signal. As a bonus, the current will also be acquired to illustrate the simulator works like a simultaneous STM and contact AFM.

Detector calibration using Force Distance curves The Generic Sweeper module found under the Experiment menu will Figure 35: Generic sweeper module configured for Force-distance work very well for this acquisition. curves It is shown in Figure 35 with reasonable parameters set. Choose appropriate limits for the minimum and maximum z values keeping in mind these are relative to the feedback loop position when the loop is opened. A negative value pushes the tip closer to the surface and a positive value pulls the tip away. The step size from the minimum to maximum can be determined by choosing how many points to acquire


along the curve. When any of the three parameters are changed (min, max, and # of points) the increment is recalculated and displayed. The speed of the sweep is entered using the Measurement Period parameter and the slew rate. The period determines how many readings to acquire and average together at each voltage step and the slew rate determines how quickly to change the voltage from one step to the next. When using the generic sweeper any output signal can be swept and any input signal can be measured. There may be some experiments where the loop should remain on during the acquisition. This is set using the Z Controller off check box. Once everything is set, start the measurement by clicking one of the two start buttons in the control section. The data can be acquired from the minimum to maximum or vice versa. The direction of the sweep is selected by the start button. When the sweep starts, the slider in the middle section will move to illustrate the progression of the measurement. The actual output value will also be displayed numerically in the window. Data similar to the graphs of Figure 35 should be seen. The sloped part of the curve is produced by the z piezo moving the cantilever when in contact with the surface. Since the surface is modeled as infinitely hard, as it moves the cantilever must bend to accommodate the push. This bending produces motion of the reflected laser spot across the PSD which produces a change in the voltage signal out of the detector. This linear relationship can be used to determine the detector sensitivity which is the change in voltage from the detector for a given displacement of the cantilever. The flat part of the curve is because after the cantilever loses contact with the surface, it returns to its equilibrium deflection and maintains that for the rest of the acquisition. In this simulated microscope there is no adhesion, so the large v-shaped part of the curve commonly seen when the cantilever sticks to the surface in an actual microscope is absent. Note that just about when the cantilever loses contact with the surface and the force error becomes flat, the current rapidly drops from a saturated value to zero. The decrease in current does indeed follow an exponential shape as expected for a tunneling gap simulator. With Input 3 recorded as volts, the detector sensitivity can be measured. This is the overall amplifier gain in the circuit to convert motion of the reflected laser spot into a voltage. Once the voltage change as a function of calibrated cantilever deflection is known,

Figure 36: Beam deflection window to enter the calibration values of the AFM detector system


the overall conversion from voltage out of the PSD to applied force can be determined. When performing the spectroscopy outlined above, the cantilever is moved a specific amount and assuming an infinitely stiff material means the distance the z piezo moves is identical to the distance the cantilever is deflected. The linear slope of the Input 3 curve supports this assumption. Measure the slope of the curve to obtain the output volts per unit distance factor (Volts/meter). Divide the known cantilever spring constant (Newton/meter) by this factor to obtain the conversion of how many volts correspond to a given applied force (Newtons/Volt). This can be entered in the Beam Deflection window so the force setpoint and all measured data have the physically meaningful units of force. The label can be changed to Force or something similar. The units should be changed to N for Newtons and then the appropriate calibration factor entered in the box below the label.

Detector calibration using lockin module Another method to calibrate the AFM detector is to use the internal lockin detector module. Once the cantilever is in contact with the surface the z piezo can be modulated a small amount. This will modulate the cantilever deflection up and down (keep in mind we assume the surface is infinitely hard) which will appear as a modulation on Input 3. By using the lockin to measure the amplitude of the voltage swing for a given z modulation the conversion factor of voltage output as a function of cantilever displacement can be determined. The lockin module is shown in Figure 37 configured for the experiment. The output is applied to the z channel which simply means the z DAC is not only outputting the feedback loop response, but internally the controller is Figure 37: Lockin amplifier module for AFM detector calibration also mathematically adding a dither to the DAC value before it is actually set. This mathematical summation should produce a far cleaner signal than outputting the z signal on a DAC and then using analog circuitry to sum it with the analog output of an external lockin as conventionally done. Choose the signal to be modulated from the dropdown list. Note the flexibility provided by the Nanonis system. Virtually any input or output can be modulated. This makes a wide variety of experiments possible that previously would have been much more difficult (consider modulating the setpoint to measure the response of the feedback loop in the time domain or modulating the bias to determine dI/dV during spectroscopy ramps). Then choose a frequency for the excitation signal and the amplitude. The signal to measure as the response is chosen using the Demodulate list. The Harmonic control can be set to look at the fundamental frequency or higher order harmonics. NANONIS | STM Simulator | 48

Once everything is configured, click the green square next to the modulate list to activate the output. By opening one of the time domain graph windows, the modulation output and its response should be visible as shown in Figure 38. As the piezo is moved, the cantilever is clearly also moving up and down as expected. Instead of trying to measure the peak to peak variation on the force as the piezo is Figure 38: Modulating the z voltage while in contact with the surface moved, let the lockin module measure the amplitude. Click the Auto button to be sure the lockin has the correct phase and the actual response measured in volts is then displayed in the lower window. This can be divided by the z amplitude to obtain the voltage change from the PSD for a given displacement of the cantilever.


Atomic Manipulation A very popular technique used with SPMs is to manipulate atoms or small molecules and build structures on a surface to study their properties. The functionality to draw features on the surface or move items around is shown accesses by the Follow Me button. The window is shown in Figure 39. There are two ways to designate a destination for the tip. Either enter an X,Y coordinate pair and click the Move button or use the mouse to point to a spot and click and the tip will start to move as soon as the mouse is clicked. A small red circle will indicate the destination of the tip. A particularly powerful feature of the software is the ability to record a signal on two channels while the tip is in motion. This can be used to capture changes in the signals as the tip moves. The acquisition is configured in the Data acquisition section. Choose Figure 39: Follow Me group of controls in the Scan Control window the amount of oversampling to apply for smoother traces but a loss of time resolution when capturing transient events. Click the Show Graph button to open the time domain window shown in Figure 40. The number of samples to collect can be set as well as the two channels to display. To apply different feedback conditions during the motion activate the Alternate settings section by clicking the button. There are two choices that can be made here. A different feedback condition can be used to move the tip closer or farther away (if needed clone the existing controller and simply change the setpoint or use a totally different control signal during motion) and a different bias condition can also be used.

Figure 40: Tip move recorder can acquire data as the tip moves along an arbitrary path

The second tab under the Follow Me family is for Lithography. A series of lines can be drawn in the window and overlaid on an image. If NANONIS | STM Simulator | 50

they are saved, they can be recalled at any point in time. There is also the possibility of using the script language itself to type in a series of movements instead of drawing the figure. Once the pattern has been drawn, the Execute button can be clicked to perform the routine. The pattern can be erased using the Delete button and they can also be saved and recalled using the Save and Load button respectively. Keep in mind, as stated earlier, the patterns and lithography recipes can be created and arranged while scanning is active in preparation for the completion of the frame. This saves valuable time compared to a system where the mouse can only perform one Figure 41: Lithography can draw features on the surface by connecting a set of function at a time straight line segments and one experimental step has to finish before preparation for the next thing can start.


Diagnostics and Analysis TCP Receiver The next step is to explore the diagnostics included in the program. Some of these features can also be used for clever experimental data acquisition but primarily they are meant to be used for monitoring signals and diagnosing possible issues. One item that can be varied to check its effect on acquisition is the overall data acquisition rate. This is determined by a group of parameters in the TCP Receiver shown in Figure 42. It can be opened by choosing System/TCP Receiver from the main window. The large green LED indicator is lit when there is a connection established between the software and the electronics (a simulated controller in this software). The RT Engine Frequency and Signals Oversampling couple together to determine the overall acquisition rate of the software. The frequency setting determines the rate of the control loops inside the real time hardware or simulated hardware. By default the engine runs at 10 kHz, for STM it is Figure 42: The TCP Receiver window determines the sampling rate and averaging of the realtime engine usually safe to increase this to 20 kHz or even 30 kHz with no adverse effects. If an AFM is operated especially in non-contact mode it is advisable to leave this at 10 kHz to not overload the system. Increasing this will speed up the system and let everything run faster and the maximum acquisition speed limit will also be raised. In thorough testing of the simulator package it was routinely operated at 30 kHz and no problems were ever experienced. The RT-sampling Period displays the inverse of the RT Engine frequency. The Signals Oversampling is used to determine how many readings are averaged together in the real time controller before sending the measurement value to the software. Lowering the sampling will increase the noise of the data, but allow it to be acquired faster. It can also be used to raise the Oversampling is an important frequency window of the spectrum analyzer concept used in the Nanonis discussed below. This should be investigated to system. Every signal goes understand the effect between noise level of the through one or more stages or data and getting the data as quickly as possible. oversampling to achieve the best All of the sliders, limits on controls, and graphs possible resolution require at are updated at a rate determined by the that point. Animations period. There is little benefit to increasing this to a rapid value because the eye


will not be able to read the display anyway if it updates too quickly. The default is 20 msec which means the values are updated 50 times/second. The Indicators period determines the update of the numerical display boxes. This should be slow enough so the eye can process the readings and make decisions based on them. The default is 200 msec which is 5 updates/second. Measurements period is the time to average readings together for very precise measurements where low noise is desired. By averaging together more readings, the noise decreases by the square root of the number of readings (assuming broadband noise) so the sensitivity improves. This needs to be adjusted depending on the nature of the experiment being performed. The activity graph helps diagnose communication problems. The white line is the current time difference between a pair of packets and the red line is the moving average. This time difference should be equal to the Signals Period discussed above. In the example shown in Figure 42, the average is very close to 1.0 msec as expected. The white graph shows large fluctuations because of the speed of the graph update. If the average is not equal to the expected period, this would usually indicate an overloaded RC4 and the RT Frequency should be decreased or too much data is trying to be acquired.


Diving deeper After following this tutorial there are many ideas you probably thought of as modules were unveiled and discussed. Feel free to return to those parts of the software and explore them deeper to gain a better appreciation for the freedom of creativity the software permits. A variety of experiments can be tried and meaningful results can be obtained given the depth of the simulated STM background process. Here are some ideas for learning sections that were not covered in detail.    

  



 

   

Configure a SafeTip condition for the Z feedback and see if you can trigger it and if the software takes the correct action. Experiment with different layouts and microscope parameters to simulate a multi-user, multi- scan head environment and see how the sessions feature takes care of this. The TCP Received window has a large number of parameters. How do they effect the data appearance in various display windows? Try using atomic manipulation with varying conditions and use the time domain tools to see if the change in conditions occurs during motion as expected. (What channel can you look at to detect when the tip is in motion?) Look into the Piezo Calibration window and artificially add drift to the microscope. What does it do to the images? Have you seen this before when using a real microscope? Use the Atom Tracking window to try to measure the drift and then compensate for it. Can you return the image to the expected result? Add noise to the system in the STM Controller background process trying some of the other parameters in the window. What does it do to the data? What are various techniques you can use to reduce its appearance in the images? Does the best technique for removing the noise depend on which parameter was changed to introduce the noise? Work with the spectrum analyzer and long term spectrum and see how the noise appears as well as when a signal is modulated using the lock in amplifier module operating independently. Measure dI/dV(V) with the spectroscopy module. Measure the transfer function of the z-feedback (dZ/dI as a function of the modulation frequency while the feedback is running). There is a special module available in the tools menu of the Lock-In detector – or you can use the generic sweeper to do this – What are the differences? Explore the possibilities of a second controller in the Follow-Me mode. Look at the spectrum of a modulated signal in the Long Term Spectrum chart. Learn what is behind the scaling icons that can be found next to almost every graph. Listen to the topography of your sample with the Signal-to-Sound graph.

Let your imagination run wild and enjoy investigating the software, but also enjoy having a “perfect” microscope that acquires very beautiful data day after day. It is also nice to know it has a tip resilient enough to withstand unimagined abuse!
