Now that we’ve looked at some of the basic interactions between electrons and light, we can turn our focus to electronic photodetectors: devices that can sense light and respond by producing a current. A simple example is semiconductor photomultipliers, like we talked about last time, which are able to very sensitively measure the intensity of light that impacts the photomultiplier.
But what do we do if we want to record an image: a two-dimensional map of the intensity of incident light? In traditional photography, silver halide crystals on photographic film interact with incident photons, then the chemical development process causes the altered crystals to darken. Since semiconductors generate a number of electrons proportional to the number of incident photons, you might think it would be easy to develop a similar digital process. But the major issue for digital imaging was not so much sensitivity as signal processing: if you have a thin film which is electrically responding to light, how do you read out the resultant electronic signal without losing spatial resolution?
Because of these difficulties, early electronic imaging was done using vacuum tubes, a bulky but effective technology we’ve discussed several times before. Many researchers were looking for a practical means of imaging with semiconductors, but the major breakthrough came in 1969, when Boyle and Smith had the basic idea for structuring a semiconductor imaging device in what’s now called the charge-coupled device (CCD).
To retain spatial resolution in an image, the photoresponsive semiconductor in a CCD is divided into an array of capacitors, each of which can store some amount of charge. One way to picture it is as an array of wells, where a rainstorm can dump a finite amount of water into any wells under it, and that water remains separate from the water in neighboring wells. But in a CCD, as photons enter the semiconductor and generate electrons, electrodes at different voltages create potential wells to trap the electrons. These wells define what we call pixels: the smallest possible subdivisions of the image.
However, to be able to make an image we also need to be able to measure how much has accumulated in each well. In our device, this means moving the electrons to an amplifier. But how can we transfer those wells of electrons without letting them mix with each other (which would blur the image) or leak out of the device altogether? To accomplish this, we need the wells confining the electrons to be mobile! But remember that the wells themselves are defined by applying voltages to a patterned array of electrodes. This means moving the well is possible, by lowering the potential directly in front of a well and raising the potential directly behind it. This idea is illustrated below for a system with three coupled potential arrays, a silicon oxide insulating layer, and an active region of p-doped silicon.
You can imagine that, instead of our previous array of wells to map rainfall, we have an array of buckets and a brigade of volunteers to pass the buckets to a measurement point. The principle is sometimes called a bucket brigade, and the general method of moving electronic outputs forward is termed a shift register. The physical implementation in CCDs, using voltages which are cycling high and low, is called clocking.
In general, the charge in a CCD will be transferred down the columns of the detector and then read out row by row by an electronic amplifier, which converts charge to voltage. Since this is a serial process, if the efficiency of transferring charge from one pixel to the next is 99%, then after moving through 100 pixels only 36% of the electrons will be left! So for a 10 megapixel camera, where the charge may pass through as many as 6,000 pixels before being measured, the charge transfer efficiency has to be more like 99.9999%! Historically, this was first achieved by cooling the CCDs using liquid nitrogen to reduce thermal noise, a practical approach for detectors on spacecraft but one that initially limited commercial applications. But eventually CCDs were made with decent transfer efficiency at room temperature, and this has been the main technological factor behind the development of digital photography. CCDs themselves don’t distinguish between different colors of photons, but color filters can be placed over different pixels to create a red channel, a green channel, and a blue channel that are recombined to make a color image. CCDs are the image sensors in all of our digital cameras and most of our phone cameras, and it was partly for this enormous technological impact that Boyle and Smith received the Nobel Prize in 2009.