Image Format

STEM images consist of a 4096 byte header followed by 512x512 pixels of image data consisting of two channels (8-bit each) interleaved. PCMass reads a STEM image, displaying both image channels side by side with the header information below the left image. The program also recognizes the magnification and sets the calibration automatically. Any image or portion of the display screen can be copied to the clipboard and pasted into a document or image, as will be described below. The current version of PCMass does not print directly and has limited image filtering and annotation capabilities, so a second standard image package such as PhotoShop, Canvas or Paint is recommended for publication quality output. Similarly, graphs and histograms are intended to give a rapid view of the data in hand and publication quality can be obtained by importing results files into SigmaPlot or similar software.

Image import into PhotoShop

STEM images can be displayed, annotated and printed with PhotoShop using the "Open As" and "RAW" import mode in the "File" menu with the above parameters entered into the specified fields (512, 512, 2 channels, interleaved, 4096 header). Some care is required to determine which image is the large angle detector signal (the channel most reliable for mass measurements). PhotoShop also permits saving a "flattened" image (use the "flatten" option in the "Layers" Menu) as a .jpg or .tif file using the "save as" option for distribution to other programs or users. This allows image compression to save space.

Imaging Detectors

Data in the STEM image comes from 3 detectors: Large Angle (LA), and Small Angle (SA) dark-field annular detectors (40-200 mrad and 15-40 mrad acceptance angles, respectively) and a bright-field detector (BF, 0-15 mRad). All detectors are scintillator-photomultiplier with nearly quantum efficiency. For thin specimens, both LA and SA signals are proportional to the number of atoms weighted by Z (atomic number) in the irradiated pixel. This is the basis for STEM mass measurement. The BF signal is 1-(LA+SA), so only two signals give independent information. Normally LA and SA are recorded but for thicker specimens, such as those embedded in a matrix, LA and BF may convey more information. Details of small corrections applied during signal read-in are discussed below.

Image Background

The key to reliable STEM mass measurement is accurate background determination. With even the thinnest carbon film substrate, the particle is usually less than half the mass (particle + background) within the measuring area. Two issues arise in background determination: 1) deciding which pixels are far enough away from an object to be considered background and 2) deciding whether the background measured away from the particle is valid underneath the particle. Issue 2 becomes critical if the sample contains detergent, salt, sugar, denatured protein or any material other than the specimen of interest. PCMass contains a background program, originally written by Jim Hainfeld with improvements by Tim Baker, which masks out particles efficiently as long as the specimen is not too crowded or contaminated. Several additional means are provided to test the reliability of the background determination, and will be discussed below.

Particle Selection

Actual mass measurement consists of selecting suitable particles with an adjustable circle or rectangle. A mouse positions the circle or rectangle and keystrokes adjust size and rectangle length. The program sums intensity minus background for each pixel inside the boundary, then multiplies by the STEM calibration constant. In PCMass, the displayed mass value is updated as parameters are changed. A left mouse-click, <ENTER> or '=' keystroke marks the particle on the display and saves the measurement parameters. Showing the mass value places the burden on the user to be objective in particle selection, rather than excluding particles simply because they spoil the SD. This is a deliberate decision, since focusing special attention on deviant particles is important to improving future preparations and understanding the specimen as well as its limits of interpretation.

TMV-based Mass Calibration

TMV is included in nearly all STEM samples as an internal control. In the past, we used a TMV calibration for each image. However, extensive measurements have demonstrated that the STEM mass calibration is very stable and fluctuations in TMV M/L (mass per unit length) are primarily due to variation in specimen quality. Therefore PCMass uses a standard STEM calibration value (115 Dalton/intensity unit with 1 nm pixels) instead of a TMV calibration. If extensive measurements suggest that a different TMV value should be used, we advise scaling the data in your plotting program. It is always a good idea to check all unobstructed TMV segments for M/L and shape as well as examining the squareness of ends (most sensitive to damage). Many problems which could compromise data quality can be sorted out using the known TMV structure.

Particle Models

A number of models are provided as an aid in selecting and categorizing particles. PCMass automatically adjusts the size, amplitude and orientation of the selected model, displaying image, model and difference image in separate windows. (This feature can be disabled by clicking A-Align OFF at the lower left of the image.) Rotational power spectrum, radial mass profile and density profile are also displayed. The difference image can be especially useful in identifying damaged particles or salt. Separate mass statistics are maintained for measurements made relative to each of the 80 available models. Extensive capabilities are provided for viewing mass measurements and histograms of single images and groups of images. Once a suitable set of models is defined, particles can be selected automatically on single images or groups of images. This has the advantages of speed and objectivity, although there are pitfalls. With manual particle selection, even an experienced user will tend to select a different set of particles in a repeat series of measurements.

Updated 4 Oct 2007	Security Notice   Webteam   Site Map   STEM   Biology   BNL