sCMOS vs CMOS Cameras: What is the Difference?

• Introduction
• Back-Illuminated Sensors
• Accuracy and Range
• Scientific Domains
• sCMOS or CMOS Cameras
• Key Points

Camera technologies continue to advance and offer improved sensitivity, resolution and speeds, helping today’s researchers in many areas of study. Most modern imaging cameras are based on CMOS (Complementary Metal-Oxide-Semiconductor) sensors, with the once popular CCD (Charge Coupled Device) sensors now limited to more niche applications. CMOS cameras come in a wide range of sensor formats, and at different price-points. You will also notice that some cameras will be described as scientific CMOS, or sCMOS. In this article we will look at the similarities and differences between CMOS and sCMOS cameras and what this means in terms of their imaging capabilities and intended use.

Balor sCMOS for Astronomy

What is common to both CMOS and sCMOS?

As the names would suggest, CMOS and sCMOS are both based on the same sensor technology, namely, complementary metal-oxide-semiconductor sensor designs. In short, sCMOS can be thought of as a higher grade, higher performance CMOS.

What are the differences between CMOS and sCMOS?

The main difference between CMOS and sCMOS is that CMOS cameras are general purpose imaging cameras whereas sCMOS cameras are designed specifically for scientific research applications that require accurate and precise measurements of signal intensity. sCMOS camera achieve their superior imaging performance in a number of ways:

High QE sensors with low noise for the best sensitivity and signal detection. Sensitivity- or the ability of the sensor to detect signals is largely set by the QE and the underlying noise characteristics of the sensor. QE describes how efficient the sensor is at converting the incoming signal in the form of light (photons) to an electrical signal. Sensors will also have some level of noise present. Sources of noise are termed read noise (noise of the sensor itself when measuring the signal and reading out the value for each pixel) and dark current (thermal noise, generated in absence of any signal).

Back-Illuminated sCMOS Sensors and Thermal Noise Control

sCMOS sensors typically have a very low read noise of 1-2 e- meaning that the noise floor of the camera is very low and low level signals can be distinguished from the noise floor. CMOS sensors tend to have higher read noise that may be 6-8e- or higher. Read noise describes the variation in measurement of a signal. The lower this variance the better – i.e. a 1e- read noise means that there is only a 1e- variance in measurement whereas a 10e- read noise means that there is a 10e- variance in measurement.

sCMOS sensors are typically cooled to reduce thermal noise. Imaging sensors are subject to thermal noise, which if not addressed, would raise the noise floor of the camera and thus impact the detection limit of the camera. Cooling the sensor is an effective way to control thermal noise and this is often a key difference between CMOS and sCMOS cameras.

CMOS cameras feature passive cooling of sensors running at ambient temperatures meaning that heat generated by the sensor is dissipated through the metal casework of the camera. This will be adequate for general purpose imaging at short exposures and when the signal levels are high. For fluorescence or spectroscopy applications when signal levels are much reduced, or at longer exposures this thermal noise is detrimental to imaging performance. Furthermore, the larger the sensor, and the faster the camera operates the more thermal noise must be controlled. In higher ambient temperatures or enclosed spaces thermal noise can be especially problematic.

sCMOS cameras such as the ZL41 series offer sensor cooling to 0°C to keep thermal noise low across a range of environmental temperatures. This is achieved using thermoelectric cooling of the sensor with a fan to evacuate warm air from the camera. For the lowest thermal noise deeper cooling can be used, often with water cooling in cameras such as the Sona and Marana back-illuminated sCMOS. These camera also include true vacuum technology for elimination of condensation on either sensor or camera window.

ZL41 Wave ‘Workhorse’ sCMOS for a range of applications.

Quantitative Accuracy and Dynamic Range

A key characteristic of sCMOS cameras is that they are optimised to ensure that the sensor can be used as an analytic tool to perform measurements of light intensity. Just as one would perform calibrations for experiments, so too is the camera optimised so that the response of each pixel across the sensor is as accurate and precise as possible. As described above, each pixel of a CMOS based sensor has its own on pixel amplifier, there is analogue to digital conversion at each column and a row by row read out. This enables the high frame rates and larger sensor formats that saw them supersede the slower, serial readout process CCD based cameras. However, this means that there are very small differences to each pixel, and column in their response to light. These differences are very small so when light is abundant, or a camera is not used for measurement purposes minimal corrections are required. For applications where light levels are low, or where comparisons between experimental conditions are required, precise sensor optimisations are required.

CMOS cameras are built to provide a good level of performance but keep the price-point low and allow fast and easy high-volume production. sCMOS cameras use much more on-board FPGA functionality to correct for these small differences for each mode of operation. This increases the cost of the camera but ensures the cameras can be used for quantitative measurements. Typically, sCMOS cameras will show this specification as the quantitative accuracy or linearity of the camera in the form 99.7% linearity or 0.3% non-linearity. This specification may be omitted from CMOS camera specification sheets.

A further difference is dynamic range – the ability of a camera to capture both weak and bright parts of the image in one snap. Lower cost CMOS cameras will often compromise on dynamic range meaning that they are more prone to image saturation when handling images with dim and bright signal information. sCMOS cameras often have a much wider signal handling capacity in some modes – normally described as a 16-bit high dynamic range mode, rather than, 8, 10 or 12 bit modes.

iStar sCMOS for Gated Imaging/Spectroscopy

Scientific Domains and Camera Options

Numerous scientific domains, such as the biological sciences, microscopy, astronomy, and industrial imaging, employ both sCMOS and CMOS cameras. The requirements of the application will determine which option is best.

• Applications that require excellent image quality, low noise, and high sensitivity frequently choose sCMOS cameras. These consist of fluorescence imaging, single-molecule imaging, and super-resolution microscopy.
• Applications requiring high-speed imaging, such particle tracking, high-speed industrial inspection, or catching quick biological processes, are a good fit for CMOS cameras.

Choosing Between sCMOS and CMOS Cameras

When deciding between sCMOS and CMOS cameras for your research, the limitations of the experiment should be considered. CMOS cameras provide high-speed imaging within the limitations of a small budget, but sCMOS cameras are superior in sensitivity, low noise, and dynamic range. It is important to realise the benefits and weaknesses of each technology and to compare it on a case by case basis, to your particular experiments conditions.

Scientific cameras such as the ZL41 Cell feature a high dynamic range as mentioned but also a linear response to light. What this means is that the camera responds in a linear, proportional way to the signal as light levels reduce to the limit of sensitivity. This means that it is possible to quantify the signal and relate this to measurements. Industrial cameras are typically not configured for this and may suffer from non-linearity at both lower limits and up to point of saturation. In addition many camera will use multiple amplifiers to enable the full image information to be digitized. If these are not set-up correctly then there can be steps during the cross-over.

There is also an effect at the individual pixel level. In a sCMOS sensor each pixel will have a different response to light. Scientific cameras will have effective pixel maps that account for these variations. This means that it is possible to monitor light levels and when replacement is required.

Key Points

• Knowing the Basics: Complementary metal-oxide-semiconductor technology, which is common to both sCMOS and CMOS cameras, facilitates the conversion of light (photons) into an electrical signal.
• Sensitivity and Signal-to-Noise Ratio: sCMOS cameras have a very low noise level, great quantum efficiency, and therefore provide a significant sensitivity benefit over CMOS cameras. CMOS cameras offer a less favourable signal-to-noise ratio by having a higher noise floor and lower QE. Nevertheless, they still provide sufficient sensitivity for applications that have abundant levels of light. In addition, with less complex camera design, cooling or optimisations they are available at much lower points.
• Speed and Frame Rates: The capability of CMOS cameras to record images at high frame rates is one of its main features. Rapid reading speeds make them perfect for capturing dynamic events or subjects in motion. Despite being slower than CMOS cameras, sCMOS cameras nevertheless provide reasonable frame rates that are appropriate for a wide range of scientific applications.
• Dynamic Range: sCMOS cameras are frequently commended for their broad dynamic range, which allows them to catch details in both bright and dark portions of an experiment. They are therefore very helpful in applications that need accurate capture of a wide range of intensities. Despite having a good dynamic range, CMOS cameras might not be as capable as sCMOS cameras in this regard.
• Cost and Affordability: CMOS cameras are more affordable than sCMOS cameras, which of course, is an advantage. Because of their price, CMOS cameras are available to a broader audience of consumers, particularly those with limited resources or funding. On the other hand, because of their greater performance features, sCMOS cameras are typically more expensive.