Cell biologists require microscopy tools that can detect rapid movements within a cell with detail and sensitivity in low light conditions. Developers must address performance requirements, and be aware that price may be a key consideration, along with the ability to manage the large amounts of data produced by dynamic imaging experiments.
While charge-coupled device (CCD) cameras have long been the “gold standard” for imaging, scientific-grade complementary metal–oxide– semiconductor (sCMOS) cameras are gaining in popularity. Early CMOS devices suffered from image quality issues, but steady improvements in sensor design have largely erased these deficiencies.
Advanced sCMOS technology is faster than current CCD cameras and thus able to better capture time-sensitive cellular events. Increased sensitivity allows these cameras to detect low luminescence signals in short exposure times, and greater resolution results in clearer depiction of small cell structures over a large field.
Now, researchers are faced with determining which solution best suits their imaging needs. This article explores the underlying differences between CCD and sCMOS technology.
CCD versus sCMOS
Figure 1 – a, b) Basic schematics of CCD (left) and CMOS (right) sensor architectures.Both CCD and sCMOS sensors (see Figure 1) perform a similar basic function: They gather light and turn it into electronic signals. Relative strengths and weaknesses of the two technologies stem from the way they read the signal accumulated at a given pixel.
CCD cameras often use a global shutter so that every pixel in an image is exposed and captured at a precise moment in time. To package that information in a digital form, the pixels are sequentially streamed through a single output node to the analog-to-digital converter (ADC) in a process called digitization. This data is then sent to a computer for display and storage.
Because every pixel is exposed simultaneously, global shutter is particularly beneficial when the image changes drastically from frame to frame. However, CCD frame rates are limited by the rate at which individual pixels can be transferred and digitized—the more pixels that need to be transferred, the slower the total frame rate of the camera. This design causes a bottleneck at the ADC. Millions of pixels stand in a single queue waiting for conversion. Before the next exposure can begin, every pixel in the existing frame must be processed. CCD cameras are able to capture reliable static and time lapse images for studies with moderate-to-long exposure times. However, the subsequent delay in charge transfer slows the camera’s total frame rate.
This is not an issue if one’s microscopy focus is restricted to assays with longer exposures, such as slow cell migration and western blot gels. But frame rates do impact the user’s ability to rapidly study fast-moving cell phenomena, including vesicle formation, protein transport and calcium wave propagation. To capture these intracellular events, cell biologists need frame rates approaching 100 frames per second (fps) or faster. With CCD microscopy, they may be able to see tiny cell structures and measure electrochemical signals, but accurate data about direction and speed will be lost. Unwanted artifacts such as motion blur and temporal aliasing also appear when frame rates are too slow for the task.
A solution to this problem is found with sCMOS chips, which position an ADC at the end of each pixel column. With this design the number of queues for conversion is multiplied—by the thousands when large numbers of pixels are involved. With sCMOS, digital information for each frame is rapidly generated. The caveat is that only one row of pixels can be digitized at a time by the row of ADCs at the edge of the sensor.
To avoid a drag on frame rate caused by waiting for all sensor rows to be digitized at the completion of an exposure, sCMOS cameras employ a rolling shutter design. Rather than waiting for an entire frame to complete its readout, rows that are digitized first can begin exposure of the next frame while the image sensor digitizes signals from later rows. With regard to time, the camera pans across the image from top to bottom. Short time delays develop between neighboring rows in the order of the sensor readout.
The advantage of a rolling shutter design is that frames can overlap, and the overall frame rate is increased. In doing so, sCMOS sensors can provide frame rates that are 10 times faster than high-end interline CCD cameras. The downside is that the slight time difference between rows could possibly skew the data.
Some sCMOS cameras, such as the optiMOS from QImaging (B.C., Canada), offer a custom triggering mode that can achieve a global exposure with a rolling shutter readout to maximize the performance of sCMOS sensors. This triggering mode allows rapid shuttering of a high-speed light source (lasers, LEDs) so that the light source is pulsed only when all the rows in the sCMOS frame are exposing at the same moment in time, thereby achieving a global exposure. Meanwhile, the camera is kept in rolling shutter mode for digitizing charge in order to maintain high frame rates and low read noise.
Advantages of sCMOS for bioresearch
Low-light image quality
The most obvious and immediate benefit of sCMOS camera technology is improved low-light image quality. The sCMOS sensor’s low read noise and larger area provides a low-noise, large field-of-view (FOV) image that enables researchers to easily scan across a sample and capture high-quality images. For example, the optiMOS’ sCMOS sensor delivers a 45% larger FOV than traditional CCD models (see Figure 2).
Figure 2 – Image and line profile comparison of the QImaging optiMOS sCMOS camera and a scientific-grade cooled CCD camera. The image quality, signal-to-noise levels, field of view and frame rates of sCMOS cameras are superior to the scientific CCD camera.Signal-to-noise ratio
In microscopy, the term “quantum efficiency” (QE) refers to the percentage of incoming photons captured and detected during a single exposure or, in other words, a camera’s sensitivity to light. Instruments with higher QEs deliver greater clarity. Many CCD cameras perform well on the QE index, but they vary greatly in terms of camera read noise.
Because many cell labels emit low-intensity luminescence signals, captured images of these low signal levels by cameras with a high level of read noise will have a low SNR and thus low image quality. By design, sCMOS sensors have low levels of electronic read noise even at high speeds, typically one-third the levels of high-end interline CCD cameras. This allows a full range of luminescence studies with a favorable SNR.
Fluorescence imaging
Fluorescence applications that can afford longer exposure times, including immunofluorescence of fixed cells, are well served by both sCMOS and CCD cameras. Fluorescence imaging of living cells, on the other hand, is extremely sensitive to light. Such sensitivity is required both to minimize photobleaching and phototoxicity effects and to capture as much temporal information as possible. The objective is to capture images with sufficient SNRs while turning down excitation intensity and using the shortest exposure time possible. As a result, scientists conducting live-cell imaging increasingly prefer sCMOS instruments. With high SNRs they can handle short exposures and still deliver on image clarity.
Experimental
Bart Guerten, Ph.D., a postdoctoral researcher in the department of cellular neurobiology at the University of Göttingen, Germany, was working on an experiment that required rapid and sensitive cellular imaging,1 i.e., visualizing the gene activity involved in mechanosensation—the response to mechanical stimuli, notably touch, sound and changes in pressure or posture—in fruit fly larvae. Such research can help determine the role of candidate genes in mechanosensory development, function and disease.
To screen for mechanosensation defects, the team had to image green fluorescent protein (GFP) in Drosophila larvae muscle with high temporal and spatial resolution. Because this involved screening mutated lines for defects in fast muscle compressions, the researchers needed imaging technology that was easy to use and capable of high-speed image acquisition.
Classical CCD cameras in this instance would have been too slow to detect muscle compression defects in the mutated lines of larvae. An imaging solution was needed that could deliver the necessary frame rates without compromising on sensitivity and noise.
After consulting a QImaging expert, the team chose to work with the optiMOS sCMOS camera, which allowed them to screen mutant Drosophila larvae and quickly quantify the compression of single compartments of the larvae with high temporal and spatial resolution.
Long-term potential of sCMOS
Researchers have long relied on CCD cameras for scientific imaging due to their quantitative performance and sensitivity. However, the scope of these devices is severely challenged as temporal resolution requirements are increased. CCD architecture does not maintain large spatial resolutions well at high speed. On the other hand, recent advances in sCMOS sensors, including significantly lower read noise while maintaining fast frame rates, make sCMOS cameras ideal for cell biology, biophysics and ion transport physiology experiments.
James Joubert is an applications scientist at QImaging, 19535 56th Ave., Ste. 101, Surrey, B.C., V3S 6K3, Canada; tel.: 604-530-5800; e-mail: [email protected]; www.qimaging.com