Fig 1.
The IntensityCheck sensors and Android app.
(A) Electrical wiring diagram of the TCS34725 or TSL2561 light sensor circuit boards to the RFduino microcontroller. (B) Smartphone screenshot of the IntensityCheck App. (C) The IntensityCheck sensor mounted as objective lens sized device on an inverted confocal microscope stand (arrow). (D) The slide shaped sensor on the stage of an upright fluorescence microscope. The pictures in (C) and (D) were taken by the author.
Fig 2.
Comparing the spectral response and linearity of the different IntensityCheck light sensors.
(A) Spectral response of three TCS34725 light sensors (two objective lens-shaped units mounted on objective turret, one slide sensor on the microscope stage) across the visible spectrum, with significant variation among the sensors. To compare the readings from the different sensors the output was calculated relative to the readings at 561nm. (B) Spectral response of two TSL2561 light sensors. (C) Measuring the linearity of the response of TCS34725 sensors 1 and 2 –same as in (A)—as a function of the actual laser power (0.2 to 1500 μW 488nm). (D) Linearity of the responses from the two TSL2561 sensors, which are actually overlapping in this plot.
Fig 3.
Factors affecting the light sensor output.
(A) Timelapse recordings showing the effect of excessive laser light (561nm DPSS laser) on the output of a TCS34725 light sensor mounted behind an OD1 neutral density filter. At 50% AOTF the sensor detects small periodic laser fluctuations—presumably reflecting room temperature changes—but when power is raised to the maximum (100%) the signal drops rapidly. Upon reduction of laser power the sensor recovers again. (B) The TSL2561 (also with OD1 ND filter) didn’t show that response even when the AOTF was set to 100%. Even when the ND filter was removed, despite a 10fold increase in laser light (“50 OD0” setting) the sensor output didn’t drop; the 100% AOTF setting was not used as the sensor saturated just above 50%. Additionally the “FRAP-booster” option on this particular confocal model was employed to maximise the amount of laser light reaching the detector. (C,D) Effect of scan field rotation on sensor output on Leica confocal microscopes, shown for the two different sensors mounted on the objective turret (“IntensityCheck”), in comparison to power meter readings taken in the sample plane on the microscope stage. (E) On a different confocal instrument (Zeiss LMS510) rotating the scan area had no effect on the sensor output as the polarisation of the incident laser light is not changed by the rotation. (F) If the sensor itself is rotated on the same instrument at a given scan rotation angle–here 0°—the intensity output changes again depending on the angle.
Fig 4.
Using IntensityCheck to maintain constant sample illumination.
(A) The upper row shows the constant sample fluorescence achieved with IntensityCheck despite ramping the Argon laser output “power levels” (488nm) from 0% to 100%. The actual AOTF settings are also displayed. The lower row for comparison with a fixed 5% AOTF setting shows an increase in fluorescence due to the rising laser output. Pseudo-colour representation of Alexa488-phalloidin labelled bovine endothelial cells. Scale bar: 20 μm. (B) Quantification of the average image intensities from (A). (C) Measuring the actual laser power in the sample plane on the microscope stage in a separate experiment. The target for IntensityCheck here was to maintain 18.8 μW (488nm) while increasing the laser output (0 to 100% power level). Achieved was 18.85 ± 0.25 μW (n = 30; using theTCS34725 colour sensor). Shown are mean laser power and standard deviations of five attempts per laser power level to reach the target value with IntensityCheck.
Fig 5.
Imaging and correcting the day-to-day confocal laser variability.
(A,B) The same FluoCells cells were imaged over a number of days with either fixed laser excitation (using IntensityCheck) or using the fixed 5% AOTF setting for the 488nm Argon laser and 561nm laser. Plotted are the mean fluorescence image intensities relative to the intensities on day 1; AM/PM denote morning and afternoon recordings, typically taken 4–6 hours apart. The asterisks mark the timepoint on day 6/AM as panel (C) shows the corresponding fluorescence images with the red mitochondrial label barely visible in the uncorrected image (“fixed AOTF”). Scale bar: 10μm.
Fig 6.
Long-term laser power measurements using IntensityCheck.
(A) IntensityCheck was used to maintain an arbitrary sample illumination intensity of 30 μW (488 nm) despite the large variations in laser output shown in (B). Recalibration was required after battery replacement (changing IntensityCheck target value from 680 to 910), and again after the actual laser power had changed by more than 10% (arrow). The dotted line shows the laser power if the second recalibration had not been carried out. (B) At the same time points IntensityCheck readings were compared to the actual laser meter readings measured in the sample plane of a confocal microscope (488nm line of the Argon laser, laser power level: 20%; AOTF: 25%; TSL2561 light sensor). IntensityCheck followed the measured laser output well, except for a sudden change when the device was removed from the objective turret for battery replacement, probably slightly displacing or tilting the sensor in the process compared to the previous time points. The inset shows IntensityCheck recordings taken at the indicated timepoint (arrow) with abnormal laser oscillations at high tube current indicative of an aging gas laser that eventually failed requiring replacement. (C-E) Corresponding measurements of the other lasers on the system: 561nm DPSS, 594nm HeNe laser and 633nm HeNe laser. For each laserline the scan field rotation was adjusted first for maximum detector output.
Fig 7.
Comparing the performance of different confocal light detectors across multiple instruments.
(A) Effect of frame averaging on the signal-to-noise ratio of the five detectors on a Leica confocal microscope. (B) The decrease in pixel dwell time due to the increase in scan speed causes the reduction of the signal-to-noise ratio—same detectors as in (A). (C) The signal-to-noise ratio of images acquired with the stable output from the LED mounted on the colour sensor circuit board is useful to compare and characterise different confocal detector types (photomultiplier tubes, PMT, versus hybrid detectors, HyD) across multiple Leica confocal microscopes. The HyD detectors provide a significant improvement over the photomultiplier tubes.