Frame Control mode for optimized imaging and diffraction with ClearView camera
Frame Control mode is a new feature introduced with the ClearView® camera, which was unavailable in previous generations of Gatan scintillator cameras. Frame Control mode enables ClearView users to maximize the signal-to-noise ratio by adjusting the camera's framerate, especially at lower dose conditions. The result is a high-quality image with a significantly improved ability to detect weak or low-contrast features in an image or diffraction pattern. This application note provides more detail on how Frame Control mode works and how to best leverage Frame Control to achieve the best quality data.
Image capture with CMOS cameras
CMOS cameras like ClearView are capable of imaging at much higher framerates than older CCD-based cameras thanks to their pixel architecture (see Imaging for more detail). ClearView can capture images at 50 frames per second (fps) in the full frame 4k x 4k image size, which is twice as fast as its predecessor, OneView®. When a single 4k x 4k ClearView image is acquired, the camera integrates frames at 50 fps until the user-defined total capture time is reached, and a single image is output for the user. Thus, a 2-second capture would result in a single image comprised of 100 integrated camera frames. Using different binning or sub-area readout options increases the maximum framerate. Generally, it is standard for the camera to use the maximum framerate available to capture a single image.
Imaging at high framerates is useful for several reasons. First, it provides a live View image that updates quickly and improves the user experience when using the camera to navigate their sample or using wobblers when aligning the microscope. A fast camera framerate also means that the pixels in the camera read out their charge very fast, allowing the camera to be used with more intense electron beams without saturating the pixels in the sensor. The live drift correction for single image captures requires multiple frames to be acquired and aligned. Hence, a high-quality live drift correction is not possible without a high-framerate camera.
One downside to capturing images at high framerates is the amount of noise that can be introduced into the final image. Each camera frame read-out event introduces a characteristic amount of noise into the frame, known as readout noise—all camera sensors, whether CMOS or CCD, introduce readout noise into images. The more frames that are integrated into the final output image by the camera, the more readout noise is introduced into the data. Although CMOS cameras generally have less readout noise per frame than CCD cameras, they can introduce larger total noise levels into an image because they integrate a significantly larger number of frames. Under conditions where there is a lot of signal on the camera, this noise is not a significant issue as there is a large amount of signal from the sample, and the camera can achieve a high signal-to-noise ratio in the final image; however, under conditions where there is an overall weak signal on the camera, such as when capturing a diffraction pattern or imaging a beam-sensitive material with a low dose beam, the total integrated readout noise can have a deleterious effect on image quality.
Direct detection electron counting cameras like Metro® or K3® also have readout noise, but the counting algorithm effectively eliminates this noise and dramatically improves the camera's sensitivity. ClearView and other scintillator cameras do not have sufficiently high framerates to implement the same type of counting algorithm, so the Frame Control mode was developed for ClearView as a method of minimizing the total readout noise integrated into the final captured image.
ClearView image capture in Frame Control mode
Frame Control allows ClearView users to slow down the camera framerate used for image capture, thus integrating fewer total frames and less readout noise into the final image. Frame Control is easily toggled on and off directly in the camera palette within DigitalMicrograph. Users can set the camera framerate to any value between the maximum for the selected image size, e.g., 50 fps at 4k x 4k or 200 fps at 2k x 2k, and the global minimum framerate of 1/30 fps. When Frame Control is enabled, the camera will integrate multiple frames at a given framerate until the total exposure time is reached. However, now the framerate can be user-defined for the acquisition. Using Frame Control to lower the camera framerate without changing the total exposure time can significantly improve the image's signal-to-noise ratio and the visibility of low-contrast sample features.
In the example in Figure 2, a diffraction pattern of a standard Au nanoparticle sample was acquired at 2k x 2k resolution, using various framerates with Frame Control mode and the same 10 s total capture time. At the maximum framerate of 200 fps, it is possible to resolve many diffraction rings in the pattern. However, calculating the radial intensity distribution of the diffraction pattern shows that the background noise of the image is high due to 2,000 total frames being integrated. Decreasing the camera framerate to 100 fps shows a decrease in the background noise both visually in the diffraction pattern and the radial intensity distribution. At 50 fps, the background noise decreases even further, making the higher resolution diffraction peaks (boxed region of the radial intensity distribution) more visible above the noise level.
Frame Control is also a useful tool for TEM imaging, as shown in Figure 3. In this example, a FIB-cut semiconductor sample was imaged at 4k x 4k resolution with a 2 s total capture time at 50 fps. These imaging conditions yield a good result, and when zooming in on the highlighted region, it is possible to see fine detail across the different image features. Utilizing Frame Control to adjust the framerate to 1 fps for the same 2 s capture further improves the image quality. When comparing the regions shown in the yellow highlighted boxes, differences in contrast and thin interfaces are better resolved in the 1 fps image because of the higher signal-to-noise ratio in this image, captured at the exact same beam conditions.
Using Frame Control mode during your TEM session
Frame Control mode is best used in cases where the highest signal-to-noise ratio in an image is more important than camera framerate and acquisition speed. In practice, Frame Control is useful for collecting diffraction patterns, images collected with low signal on the camera (e.g., low-dose images of beam-sensitive materials or dark-field TEM images), or any image acquisition requiring a long capture time (>10 s).
Optimizing the signal-to-noise ratio with Frame Control mode requires minimizing the total readout noise integrated into a captured image and maximizing the signal in each frame. This is done by using Frame Control to select as slow a framerate as possible without saturating the pixels in an individual frame readout. Measuring the intensity per pixel at different framerates is easily accomplished using built-in DigitalMicrograph tools to measure a live Histogram on the image or line profile over a region of interest. The ClearView user manual includes more detail for setting up image and diffraction pattern captures with Frame Control mode. For diffraction experiments and low-dose TEM imaging, using a framerate <5 – 10 fps and total exposure time <5 s will give good results.
High-framerate CMOS-based cameras like ClearView bring many new imaging capabilities to transmission electron microscopy. Still, it is not always best to collect data using a high-framerate operating mode. Microscopists can utilize Frame Control mode to reduce the camera framerate for single image captures, thus reducing the total amount of integrated read-out noise and maximizing the dynamic range of the ClearView sensor's pixels on a per-frame basis, and optimizing the camera and imaging conditions with Frame Control mode results in high signal-to-noise images and an improved ability to see subtle, low-contrast features in both images and diffraction patterns.