What Are Charge-Coupled Image Sensors?

Charge-coupled image sensors are solid-state devices that convert incident light into electrical charge and then transfer that charge across an array of closely spaced capacitors to produce a digitized image. They are built on a metal-oxide-semiconductor structure in which photons generate electron-hole pairs in silicon; the electrons accumulate in potential wells beneath each pixel electrode, and a sequence of clock voltage pulses shifts the packets of charge from pixel to pixel until they reach an output register and amplifier. Invented on October 17, 1969, by Willard Boyle and George E. Smith at Bell Labs, the charge-coupled device (CCD) became the dominant image sensing technology for scientific, medical, and consumer photography through the late 2010s.

Boyle and Smith received the 2009 Nobel Prize in Physics for the invention, and in 2006 the IEEE awarded them the Charles Stark Draper Medal in recognition of the CCD's impact on imaging technology. The fundamental shift-register charge transport mechanism that gives CCDs their name distinguishes them from competing sensor architectures and directly determines their sensitivity, dynamic range, and read noise characteristics.

CCD Architecture and Operation

A CCD image sensor is organized as a two-dimensional array of photosensitive pixels, each consisting of a p-type silicon substrate beneath a thin oxide layer topped by a polysilicon gate electrode. During exposure, gate voltages are held at levels that create a potential well accumulating photogenerated electrons in proportion to incident light intensity. After exposure, a three-phase or four-phase clocking scheme shifts each row of charge packets horizontally into a vertical serial register, which then transfers them one row at a time into a horizontal output register. The output register delivers the charge packets serially to an on-chip charge-to-voltage converter. This architecture, detailed in Stanford's EE392B course notes on CCD operation, achieves very low read noise because the signal-carrying charge never passes through a noisy amplifier until the final conversion step.

Charge Readout and Signal Chain

The on-chip output amplifier in a CCD is typically a floating diffusion followed by a source-follower stage. A charge packet deposited on the floating diffusion node raises its voltage by Q/C, where C is the node capacitance, typically a few femtofarads, giving conversion gains of 1 to 10 microvolts per electron. Correlated double sampling (CDS), performed either on-chip or in the analog front-end, subtracts the reset noise from the signal voltage by sampling the node before and after charge injection, reducing the effective read noise floor to a few electrons RMS. Back-illuminated CCD variants improve quantum efficiency from roughly 40 percent in front-illuminated designs to over 90 percent by thinning the silicon substrate and illuminating from the side opposite the gate electrodes, a technique widely used in astronomical CCDs and described in detail by Teledyne e2v's technical overview of CCD operation.

Comparison with CMOS Image Sensors

CMOS image sensors differ from CCDs in that each pixel contains its own readout transistors, allowing direct random-access readout without the serial charge transfer chain. This architecture reduces power consumption, enables on-chip integration of analog-to-digital converters and image processing logic, and simplifies fabrication using standard CMOS processes. CCDs retain advantages in uniformity, fill factor, and read noise in large-format scientific sensors, where the single shared output amplifier and absence of per-pixel transistor mismatch produce highly uniform images. Since the mid-2010s, CMOS image sensor improvements at the IEEE Image Sensors and Imaging Systems conference have narrowed the noise gap considerably, and CMOS sensors now dominate consumer and mobile applications while scientific and low-light applications continue to use CCDs.

Applications

Charge-coupled image sensors have applications in a wide range of fields, including:

• Astronomical imaging and space telescope focal planes
• Medical imaging in endoscopes and fluorescence microscopy
• Industrial machine vision and quality inspection
• Digital cinema cameras requiring high dynamic range
• Scientific spectroscopy and X-ray detection