This protocol provides guidance for defining patterns with single-digit nanometer resolution in two common electron-beam resist using a scanning transmission electron microscope or STEM as the exposure tool. Use of an aberration-corrected STEM in this protocol allows routine patterning of lithographic features with single nanometer resolution. While very specialized and expensive tools, these instruments are sometimes available for use without cost.
The techniques that are described in this protocol can be used to transfer nanoscale pattern into a variety of materials. Thus, enable fabrication of novel devices at single-digit nanometer resolution. Demonstrating will be Na Li, a student working at The Center for Functional Nanomaterials.
To begin place two strips of double-sided carbon tape approximately equidistant from the center of the silicon holder and separated slightly less than the diameter of the TEM chip. Rinse the strips with isopropyl alcohol to reduce their adhesive strength and avoid breaking the delicate TEM chip during removal from the silicon holder. Mount the TEM chip on the silicon holder making sure that it is attached to the carbon tape strips only at two opposite edges.
To spin coat the HSQ resist, mount the silicon holder on the spinner chuck and align the center of the TEM window approximately with the center of the spinner rotor. Using a pipet, cover the entire TEM window with one drop of HSQ. Depending on the resist used follow the spin coating and baking parameters shown in the text protocol.
Following spin coating, carefully remove the TEM chip from the silicon holder. Inspect the resist uniformity over the TEM window using an optical microscope. If the film is homogenous across the central region of the membrane as shown here, proceed to the next step.
Otherwise, repeat the resist coating process on a fresh TEM window. Mount the resist coated TEM chip on the STEM sample holder. Make sure that the resist vacuum interface faces the incoming beam.
Since the beam is optimally focused at the top of the sample. Also, make sure that the sides of the TEM window are aligned approximately with the X and Y axis of the STEM stage. This will facilitate navigating to the TEM window.
Now, load the TEM chip into the microscope and pump overnight to reduce contaminates in the sample chamber. The next day, move the stage coordinates such that the beam is more than 100 microns away from the center of the TEM window to avoid accidental exposure. Set the stem probe beam current to 34 picoamps and the beam energy to 200 kilo electronvolts.
In this microscope an emission current of five microamps is equivalent to a probe beam current of 34 picoamps. In diffraction mode imaging, set the magnification to 30, 000 times with the beam out of focus. Which makes it easier to find an edge of the TEM window.
Diffraction mode is characterized by a stationary beam, z-contrast mode and mid-angle angular direct field detector. We use diffraction mode because it is faster. Since the beam does not need to be scanned to form any image.
Navigate towards the TEM window until an edge of the window is observed on the diffraction image. Then navigate along the window edges and record the X and Y coordinates of the four corners of the TEM window. In this exercise the recorded coordinates of each window are shown in this slide.
At the last window corner increase the magnification to 50, 000 times and perform rough focusing on the window membrane by moving the stage z coordinate until crossover of the diffraction pattern orientation is observed. Subsequently, perform fine focusing by adjusting the objective lens current. Now, increase magification to 180, 000 times.
Adjust focus, stigmation, and aberration-correction settings in order to obtain an aberration-corrected diffraction image of the window membrane. This focusing method is known as the Ronchigram method. Close the beam gate valve to avoid any accidental exposure of the resist when moving the stage.
Verify that the beam current is 34 picoamps and the magnification is 180, 000 times. Use the prerecorded window corner coordinates to move the stage so that the field of view center is 5 microns away from the center of the window. In this exercise, this position is represented by point A in the slide.
Open the beam gate valve and focus at this point using the Ronchigram method. Next, close the beam gate valve. Move the stage to place the field of view at the center of the TEM window.
Change the magnification to 18, 000 times. Now, transfer the beam control to pattern generator system by clicking on the NPGS command of the pattern generator user interface and position the beam anywhere away from the pattern area. Here, the top right corner is used, which is achieved with the DAC plus 10 plus 10 command.
Clicking on the Process Run File command sets the system ready for exposure which takes place when the space bar of the pattern generator computer is depressed but do not press it yet. It is critical to perform the following actions in quick succession to avoid overexposing the resist at the initial and final beam positions. Open the gate valve then verify by observing the beam diffraction pattern image.
Whether the beam is in focus at the initial beam position. Expose the pattern. When the exposure is complete.
Check if the diffraction pattern image remains in focus at the final beam position. Finally, close the gate valve and remove the TEM chip from the STEM. To develop the HSQ, stir the TEM chip in a salty deionized water solution containing 1%weight sodium hydroxide and 4%weight sodium chloride for four minutes at 24 degrees Celsius.
Then stir the chip in pure deionized water for two minutes to rinse off the salty developer. Dip the TEM chip in ACS reagent grade IPA and gently stir it for 30 seconds. Quickly place the TEM chip on a special two inch silicon wafer.
Make sure the TEM chip is always wet with IPA during the transfer. After approximately two to three minutes, close the critical point drying or CPD wafer holder assembly as diagrammed in the text protocol. Leave the whole unit soaking in ACS reagent grade IPA for an additional 15 minutes totally immersed in IPA.
Quickly transfer the complete CPD wafer holder assembly to a second container with fresh ACS reagent grade IPA and leave it for an additional 15 minutes totally immersed in IPA. Now transfer the CPD wafer holder assembly to the CPD instrument process chamber. At all times the TEM chip should be totally immersed in IPA.
Run the CPD process following the instrument operating instructions. After exposure and HSQ resist development, three to four nanometers of the ultra-thin silicon layer in the unexposed layer of the window were removed by inductive coupled plasma etching. Observing the detail of the central region of the HSQ resist.
Reveals the four lines have an average measured width of seven nanometers. Scanning electro microscopy images of the smallest pattern holes and positive tone PMMA are shown here. The average smallest isolated feature is 2.5 plus or minus 0.7 nanometers.
While the smallest pitch pattern is 17.5 nanometers. The yellow scale bar is forty nanometers. Results for negative tone PMMA are shown here.
The average smallest isolated feature is 1.7 plus or minus 0.5 nanometers. While the smallest pitch pattern is 10.7 nanometers. Again, the yellow scale bar is forty nanometers.
This protocol describes a process for patterning aberrant structures with single-digit nanometer resolution in the conventional electron electron beam resist PMMA and HSQ. It is critical to focus the electron beam before and after the exposure to achieve the highest resolution patterning and to determine whether any defocusing occurred during patterning. The use of critical point drying after developing is also critical to avoid pattern collapse due to highest variation of the pattern structures.
The results for positive and negative tone PMMA are the smallest features in the literature. The results for HSQ are not the smallest but this protocol enables obtaining reproducible sub 10 nanometer features in HSQ and demonstrates single-digit patterning of silicon structures. Additionally in agreement with previously published studies these results demonstrate that such patterns can be transferred with high fidelity to a target material of choice.
