Jeffrey Perkel has been a scientific writer and editor since 2000. He holds a PhD in Cell and Molecular Biology from the University of Pennsylvania, and did postdoctoral work at the University of Pennsylvania and at Harvard Medical School.
The classic image of a biologist is of a lab coat-wearing researcher peering through microscope oculars, and many biologists do indeed log considerable time at the microscope bench. But when it comes time to document their experiments, they don’t sketch their observations on paper—modern biologists use microscope-mounted cameras instead.
In principle, microscopy digital cameras are like cell-phone cameras, though much more high-end. A lens focuses an image onto an array of sensors, which records the image. But those sensor arrays are not all alike, and microscope cameras have features that can make them more or less compatible with particular applications. Here, we review your options.
CCD, sCMOS or EMCCD?
Modern microscopy digital cameras typically fall into three categories, says Rachit Mohindra, product manager at Photometrics, a company specializing in microscopy imaging: CCD, sCMOS and EMCCD.
In all cases, a lens projects an image onto an array of pixels, just as a traditional camera would project onto film. In a film camera, the time required to take a burst of multiple pictures is defined by the shutter speed and how quickly the camera can advance the film. In digital cameras, the key variable is the speed at which pixels can be read out and reset.
The most mature technology, Mohindra says, is the CCD-based camera. Here, the pixels are read out via a single output, like a giant athletic stadium with a single exit. “If you have a very large CCD—say 10 megapixels—that’s 10 million pixels you have to transfer from one output. So you have to go extremely fast to get any sort of usable frame rate.”
For bright images, that isn’t much of a problem. But for dim images—such as those produced in certain fluorescence applications—it can be. CCD cameras, Mohindra explains, represent a trade-off between pixel-array size, sensitivity and speed. Larger pixel arrays mean more data, but potentially more time between frames, as well. “The faster you go, the higher the noise; the larger the number of pixels, the lower the frame rate. But in terms of being able to detect low-light signals, [a CCD] does quite well.”
Some newer CCDs offer additional data output ports. Sony, for instance, has developed a quad-port CCD, which allows cameras to read out four quadrants of the chip simultaneously. The Zeiss Axiocam 506, a 6-megapixel camera based on that chip, can read out 6 million pixels at 20 frames per second, says Zeiss product marketing manager Scott Olenych—almost twice as fast as the company’s earlier 1.4-megapixel cameras, yet with four times more data.
Another class of imaging device is the “scientific CMOS” (sCMOS) sensor. Here, each column of the array has its own data output. Thus, in a 1,920- x 1,080-pixel (2-megapixel) sensor, there are 1,920 data-transfer points instead of one. “Now you’re able to slow the rate of data transfer and still get extremely high frame rates. That gives the benefit of lower noise,” Mohindra says.
Search for Microscope Cameras
Find, compare and review microscope
cameras using the Biocompare search tool Search
CMOS cameras typically feature larger sensor chips than CCDs, Olenych says, providing a larger field of view and thus fewer images required to fully capture a specimen. Speed, he says, is about more than just frame rate; “It’s also about how much area you can cover per unit time.”
A final option is the EMCCD, or “electron-multiplication CCD.” Used to detect the dimmest of dim signals, EMCCDs are modified CCDs with an electron-multiplication readout ability, which amplifies signals and enables the detection of even individual photons. And they also typically are cooled (using, for instance, a Peltier element) to reduce “dark current noise.” Mohindra explains, “Cameras tend to generate heat, and the sensor can mistake this heat for light. The point of cooling is to minimize that ‘dark signal.’”
Leica Microsystems, which sells both monochrome and color cameras, including the 5-megapixel, cooled, color Leica DFC450 C, offers everything from basic to high-end imaging devices, says Karin Schwab, product manager for cameras. The company also offers third-party, high-end EMCCDs from such vendors as Hamamatsu and Andor, which cost considerably more. “These are Ferraris,” Schwab quips. “We have Volkswagens, Mercedes and Audis. But they make Ferraris.”
Key variables
According to Schwab, the first thing users must determine when deciding on a camera is what applications they hope to support. For brightfield imaging, imaged using abundant light, Schwab recommends a color camera. “You may not be so interested in sensitivity, because your specimen is bright and you can apply short exposure times.”
For low-light fluorescent applications, however—such as single-molecule imaging or total internal reflection fluorescence (TIRF) microscopy—Schwab recommends a cooled EMCCD. These cameras typically feature high dynamic ranges and larger pixel sizes, says Mohindra. “They are designed for maximal light collection,” he notes. “It’s like a bucket—the larger the bucket, the more likely you are to catch a raindrop.”
EMCCDs typically feature relatively small sensor arrays—that is, they sport large pixels, but relatively few of them. Andor’s iXon Ultra 897 EMCCD, for instance, features an array of 512 x 512 pixels measuring 16 μm on a side (The pixels in the Zeiss Axiocam 506 CCD measure 4.54 μm on a side). But image quality is a function of both the camera and the objective, Olenych notes. Low-power objectives require smaller pixels (ensuring there are sufficient pixels illuminated by the projected image to capture at good resolution), while higher magnification can be used with larger pixels. Thus, EMCCDs typically are used with high magnification (60x or 100x) objectives.
If researchers have the resources, they can buy multiple cameras to support different applications. Or they can try to find one that supports them all. Leica recently introduced its first camera (the DFC7000 T) based on a “new sensor platform” from Sony, says Schwab, which works well for both brightfield and fluorescence imaging. The new sensor, she explains, features smaller pixels, an overall optimized architecture and other advances. “It offers extremely high sensitivity and high dynamic range for fluorescence, but the pixel size and resolution are so high, and we added features for color management, so it is quite well suited for color imaging, as well.”
That said, most cameras intended for fluorescence are monochrome. Color cameras are used primarily to document such data as hematoxylin and eosin (H&E) stained slides; fluorescent images, in contrast, are taken using excitation and emission filters and then artificially colorized. “Color cameras tend to be nonquantitative in measurements,” Mohindra explains, thanks to the Bayer matrix atop the sensor that filters different wavelengths.
For those who wish to create time-lapse datasets, another key variable is frame rate. According to Mohindra, the most common CCDs typically deliver about 10 to 15 frames per second (fps). sCMOS can image at 100 fps, “and you can boost to a couple thousand fps, if you look at a few rows [of the sensor] rather than the whole field.” EMCCDs typically can handle between 30 and 70 fps, or more for smaller sensor regions.
And then there’s budget. Neither Mohindra nor Schwab would provide definitive pricing for their companies’ products. But Mohindra did estimate price ranges of $3,000 to $10,000 for CCD cameras; $10,000 to $20,000 for sCMOS devices; and $30,000 to $40,000 for EMCCDs. “That’s a very rough breakdown.”
At that price, it makes sense to try before you buy. Fortunately, you usually can. Photometrics encourages customers to test cameras in their research environment using representative samples, Mohindra says. Leica, too, offers system demos, as does QImaging, a sister company to Photometrics that sells an sCMOS camera called OptiMOS.
If nothing else, these demos—and conversations with sales consultants—can help ensure your most demanding applications are being addressed. And today, that’s more important than ever. Between new microscopy techniques and an exploding camera market, balancing the variables can be daunting. (See this essay at Nikon MicroscopyU for a detailed discussion of optical system requirements. This essay, from QImaging, compares CCD and CMOS technologies.)
Says Steven Smith, QImaging product manager, “There’s tremendous value in working with people who understand the science when evaluating new imaging technology.”
Image: iStockPhoto