A spectrometer asks something different of its detector than an imaging system does. The light has already been spread across the sensor by a grating, so every column is a wavelength, and the signal in each one can be very faint. A detector that looks excellent in a visible-light imaging spec sheet can disappoint in a Raman instrument working out toward 1,000 nm, or in a UV system working below 300 nm.
This post is for teams building or refreshing spectroscopy instruments: Raman, near-infrared (NIR) and UV systems that use a cooled area sensor behind a spectrograph. It covers the sensor choices that matter most at the ends of the silicon range, the trade-offs that come with them, and the questions worth settling before you lock in a detector. For Finger Lakes Instrumentation (FLI), the closest fits for this work are the deep-depletion and UV-coated CCDs in our MicroLine camera line.
In an imaging camera, quantum efficiency (QE) is usually judged near its peak in the visible range. In a spectrometer, what matters is response at the specific wavelengths your instrument measures. For Raman work with a red or near-infrared laser, the scattered signal lands in the far red and NIR, right where standard silicon sensors start to lose sensitivity. For UV absorption or emission work, the opposite end of the spectrum is the challenge.
Spectroscopy also tends to mean long exposures on weak signals, so dark current and read noise matter as much as QE. And because each pixel column maps to a wavelength, any pattern the sensor adds to the signal shows up directly in the measured spectrum.
Silicon becomes more transparent as wavelength increases. In a standard thin back-illuminated CCD, much of the NIR light passes through the sensor before it's absorbed. Some of it reflects back and forth between the front and back surfaces, which creates interference fringes, often called etaloning. In a spectrum, those fringes appear as a ripple that's hard to calibrate out.
Deep-depletion CCDs use thicker silicon with a deeper active region. That gives NIR photons more material to be absorbed in, which improves NIR response and reduces fringing. It's why deep-depletion sensors are widely used in Raman and NIR spectroscopy.
FLI's MicroLine line includes several deep-depletion options:
Deep depletion comes with a cost that every spectroscopy team should plan for. Many e2v CCDs can run in what's called inverted mode (AIMO), which suppresses dark current. Deep-depletion devices can't run in inverted mode, so their dark current is 100 to 200 times higher than AIMO parts. As a rough example, where an AIMO sensor might show about 0.1 electrons per pixel per second, a deep-depletion sensor might show about 10.
That isn't a choice the camera maker makes. It's set by how the sensor itself is manufactured. What the camera maker controls is cooling, and that's why cooling matters so much for NIR work. If your measurement doesn't need NIR response, a standard back-illuminated sensor with lower dark current may be the better choice. If it does, budget for strong cooling and good dark-frame calibration.
At the short end of the spectrum, the question is how well UV photons reach the sensor and get converted. FLI offers UV-coated versions of its MicroLine CCDs, such as the ML4240UV, which use an enhanced coating for UV response. On the CMOS side, the Kepler KL400BI is described as having excellent QE at 240 nm, which is worth a look if you also need faster readout.
The window in front of the sensor matters as much as the sensor. Window materials differ in how far they transmit into the near-UV and the near-IR, so the right choice depends on the band you measure. The window also has to be chosen before the camera enters our clean-room build, so it belongs in your first conversation with us. Ask our sales team for transmission charts for the windows available on your model.
Spectroscopy formats are a balance between spectral coverage, resolution and signal per pixel. A wider sensor covers more wavelengths at a given dispersion. Larger pixels collect more signal per spectral element. A deep full well lets you measure a strong peak and a weak band in the same exposure.
The MicroLine ML4240 and ML4710 sensors both list a 100,000 electron full well and 16-bit digitization. The ML4240 gives you a 27.6 mm square sensor for wide spectral coverage or multiple fibers stacked along the slit. The smaller ML4710, at 13.3 mm square, suits compact spectrographs. If your instrument also needs very large formats, our post on large-format cameras for national labs covers the bigger sensors.
MicroLine cameras are air-cooled as standard, with optional liquid cooling. On their data sheets, the ML4240DD cools to 55 °C below ambient with air, and the ML4710DD to 60 °C below ambient. Because cooling is stated relative to ambient, the temperature inside your instrument enclosure sets how cold the sensor really runs. Our post on air versus liquid cooling for OEM cameras walks through how to plan for that.
MicroLine cameras also offer more than one digitization speed. Slower readout gives lower noise for long, faint exposures. Faster readout helps when you're aligning optics or running many short acquisitions.
If you're new to working with a camera supplier on an instrument program, our post on choosing an OEM camera partner covers the wider process.
Every FLI camera is Made in USA, and our sensor chambers are assembled in our 1,000+ sq ft ISO Class 7 clean room. See how on our in-house manufacturing page.
Tell our sales team your wavelength range, exposure times, spectrograph format and enclosure limits, and we'll help you compare deep-depletion, UV-coated and standard sensors for your instrument. Request a quote or email sales@flicamera.com.
