• Architecture: The charge generated by photons is transferred into the readout register. Three principles are used:
  • Full frame:The image is read out directly. The fill factor is 100%, however during the readout phase of the CCD the electron beam has to be blanked by an external shutter.
  • Frame transfer: Before readout the whole image is transferred to a buffer area insensitive to light. 100% fill factor, no shutter required, but a double-sized CCD-chip is needed.
  • Interline: ~ 25% fill factor, no shutter is required.

    TVIPS provides scientific-grade full frame and frame transfer CCD cameras.

• CCD format: Number of pixels in x and y direction - typically from 1024 x 1024 to 4096 x 4096
• CCD pixel size: The pixel size determines the resolution and the full well capacity and thereby the dynamic range. A larger pixel size improves the resolution and the dynamic range.
  TVIPS produces cameras with 14, 15 and 24 µm pixel size (square).
• Field of view: Size of the sensitive CCD area, product of pixel size and format
• Cooling: Peltier cooling of the CCD chip (typically -30 - -35° C) reduces the CCD noise and the dark current.
• Readout rate: Readout in pixels per sec transferred to the analog-to-digital converter
• Digitization:The resolution of the ADC in number of bits. The digitization defines the number of grey levels and therefore roughly the possible dynamic range. But a higher digitization rate ensures not necessarily a better dynamic range, which is mainly defined by the CCD noise and the full well capacity.
  TVIPS provides CCD cameras with 12 and 16 bit digitization.
• Frame rate: In frames (images) per sec, approximately determined by readout rate, format and exposure time.
• TEM column interface: Depending on the TEM type, there are several mounting positions

Bottom-mounted (on-axis): The standard position, particularly for high resolution. The post-magnification can be increased by extension tubes and varies from 1.2 to 2.0x. All TVIPS cameras can be mounted in the on-axis position. Bottom-mounted (off-axis or near-axis): Optimal position for a fast camera parallel to an on-axis device, e.g. a slow-scan CCD camera or an energy filter (post-magnification ~1.2x). The TVIPS FastScan can be mounted in the off-axis or near position. Wide-angle (at the 35 mm port): the main advantage of this position is the large field of view. Aberrations of the projective lenses however cause visible image distortions, e.g., 1% means 10 pixels on a 1k camera! In this position, high resolution is difficult to achieve because the electron image is demagnified by a factor of about 3. Therefore higher EM magnifications are necessary. Wide-angle TVIPS cameras are not available. Retractable: For some TEMs it is possible to mount the camera to the view chamber. The post-magnification is ~0.9x.

• Electron-optical coupling: Two basic principles for transferring the photons to the CCD are used:
  • Lens coupling allows the post-magnification to be changed rather easily, but with the poor efficiency (<0.5%) single electron detection is impossible.
  • Fiber-optic coupling gives a very high efficiency (>60% of all photons are collected). For large area cameras the bonding quality of the fiber optics (FO) to the CCD is essential. TVIPS cameras are bonded with a special fiber shaping technique that guarantees an optimal resolution over the whole image are

CCD-FO coupling without adaption	CCD-FO coupling with adaption

All TVIPS cameras are fiber-optically coupled.

• Scintillator: Two common types of scintillators are used:
  • Polycrystalline phosphors are optimized for high efficiency (single electron signal) and can easily be optimized for resolution and sensitivity depending on the high tension.
  • Single crystal YAGs give about 5 times less signal than the polycrystalline phosphor, but are resistent to mechanical damage.
  TVIPS provides polycrystalline phosphors as standard. YAGs are available up to 38mm Ø.
• Resolution: The resolution of a camera system can be expressed by the point spread function (PSF) or the modulation transfer function (MTF), the Fourier transform of the PSF. The MTF defines the ratio between input and output signal as a function of the spatial frequency. It can be measured with a sharp edge or a statistical method, resulting in the noise transfer function (NTF). The MTF/NTF is determined by the pixel size, the type and quality of electron-optical coupling and the type and size of the scintillator.

Noise tranfer function (NTF) of 2k cameras with 14 µm and 24 µm pixel size (corrected for aliasing)

TVIPS specifies the camera resolution as the NTF value at Nyquist frequency.

• Binning: Adjacent pixels can be combined to a "super pixel". A faster readout of the same field of view and higher sensitivity can be achieved.
• Gain factor: Defines how the CCD electrons are converted into ADUs (analog-to-digital units = grey levels). A gain factor 1x means that the full well capacity matches the full range of the ADC.
• Full well capacity: The maximum number of CCD electrons generated by incident photons which can be hold in a single CCD pixel, typical values are 100 000 CCD electrons for a 14-µm pixel and 500 000 CCD electrons for a 24-µm pixel.
• CCD noise: There are mainly two sources of noise: the electronic readout noise, and the dark current noise.
• Dynamic range: Defined as the ratio of the maximum signal to the CCD noise. Typically >4000 for a 12-bit camera with a CCD noise of less than 1 count.
• Non-linearity: The deviation from a linear curve representing the ratio of the input to output signal should be less than 1-2%.
• Sensitivity: The number of ADUs generated by a single primary electron. It depends on the gain factor, the type and the thickness of the scintillator, and the high tension.
• Anti-blooming: Minimizes the crosstalk of adjacent pixels at high intensities.
• Gain uniformity: Slight variations of the sensitivity of the electron converter chain can be compensated by the flatfield correction.
• Flatfield correction: A method to correct the image for contributions of the CCD dark current and non-uniformities of the gain caused by fixed pattern of the fiber optics ("chicken wire"), slight variations of the CCD or scintillator sensitivity, or small dirt particles lying on the surface. For flatfield correction, a dark reference image is subtracted from the image which is then divided by a gain reference image.
  TVIPS software applies flatfield correction automatically after image acquisition.