Multichannel Fluorescence Imaging Specifications

1. Optical path and channel configuration

• At least four channels covering DAPI, FITC, TRITC and Cy5, expandable to seven or more

• Epi illumination where excitation and emission share the objective, split by a dichroic mirror

• Independent shutter or light source switching per channel, to limit bleaching

• Switching by filter wheel, multiband filter or time multiplexed LED excitation

• Note: more channels only help when the chosen combination can actually suppress bleed through

2. Filter set parameters: the key item

• Each filter set should state excitation range, emission range and dichroic cut on wavelength

• Narrow bands of 25 nanometres or less resist bleed through, while wide bands give more light

• Bandpass emission filters are preferred for multicolour overlay, longpass only for single colour

• Peak transmission of 90 percent or more, with surface flatness and coating hardness stated

• Note: claiming support for DAPI, FITC and TRITC is not enough without wavelength ranges

3. Bleed through and channel isolation

• Adjacent channel bleed through should be specified, for example DAPI leaking 1 percent or less into FITC

• Acquiring the weakest channel first and the strongest last reduces cross talk in overlays

• Slide scanning systems should state whether channels are scanned whole slide then merged

• Autofluorescent samples such as plant tissue need autofluorescence removal or spectral unmixing

• Note: bleed through is the largest error source in quantitative fluorescence and cannot be fixed in post

4. Light source and intensity balance

• LED illumination with independently adjustable wavelengths and stability within 2 percent per hour

• Brightness adjustable per channel so weak channels are not underexposed and strong ones do not saturate

• LED lifetime of 20000 hours or more, reducing replacement and recalibration

• Intensity calibration records should be kept so data from different sessions remain comparable

• Note: quantitative multicolour comparison requires reproducible channel intensity, not an approximate match

5. Objectives and chromatic registration

• Fluorescence dedicated objectives with high transmission and low autofluorescence, aperture chosen by sample

• Apochromat objectives reduce focal shift across wavelengths and improve overlay alignment

• Focus difference between channels should stay within one depth of field to avoid colour fringes

• Immersion objectives should state medium refractive index and temperature range

• Note: ghosting in overlays usually comes from objective chromatic aberration, not from software alignment

6. Camera and exposure control

• Monochrome sCMOS or CMOS with quantum efficiency of 70 percent or more from 500 to 700 nanometres

• Bit depth of 12 bits or more, with full well capacity and read noise specified

• Exposure range from 10 microseconds to 10 seconds at least

• Hardware triggering synchronised with light source and filter wheel

• Note: for weak fluorescence, noise performance matters more than pixel count

7. Acquisition and overlay

• Support for multichannel sequences, Z stacks, time series and plate scanning

• Automatic channel registration with manual fine adjustment, accurate to one pixel

• Output keeps raw channel data, not only merged images

• Background subtraction, flat field correction and bleaching compensation provided

• Note: a system that only exports merged images cannot be re analysed, so results are not reproducible

8. Environment and light protection

• Dark room or controlled lighting, with stray light blocked between objective and sample

• Ambient 15 to 30 degrees Celsius, live imaging at 37 degrees Celsius within half a degree and 5 percent CO2

• Heat management for light source and camera, since thermal drift affects focus in long acquisitions

• Operators need wavelength appropriate eye protection, especially for ultraviolet channels

• Note: light leaking into the path can exceed sample signal in weak fluorescence channels

9. Where procurement goes wrong

• Choosing by channel count alone without checking filter wavelengths and bleed through, so colours contaminate each other

• Light source channels that cannot be dimmed independently, leaving weak channels always over or underexposed

• Systems that export only merged images, discovered at acceptance to be impossible to re analyse