Dear community members,
I need to study the fluorescence of certain analytes in an aqueous solution. Currently, I am investigating fluorescence in a physiological solution. To conduct this study, I am using an Avantes spectrophotometer.
My goal is to obtain calibration curves for specific analytes in water at different concentrations.
The experimental setup consists of a peristaltic pump drawing the aqueous solution from a beaker containing the analyte. The solution flows through a plastic tubing. A small section of the plastic tubing consists of a circular plastic cuvette housed inside a light-shielded box. Inside this box, there are two LEDs and the entrance of the optical fiber that captures the light from the LEDs and the fluorescence of the solution excited by the LED light. Once crossed the box containing the LEDs, the liquid through the plastic tube returns to the beaker where the solution is recirculated for the experiment.
As mentioned, there is an optical fiber that captures the light from the LEDs and the fluorescence of the analytes dissolved in the water. The captured light travels through the fiber and goes into the box where the conversion of the light signal to an electrical signal takes place, along with all subsequent electronic processing. The spectrophotometer is connected via Ethernet to a Raspberry Pi, and I see the software interface on the screen to manage the spectrophotometer parameters. We have two LEDs: LED 1 emits light at 445 nm, and LED 2 emits light at 340 nm. One analyte absorbs at 445 nm and emits fluorescence between 500 and 600 nm, with a peak at 523 nm. Another analyte absorbs at 340 nm and emits fluorescence between 400 nm and 600 nm with a peak at 461 nm.
Now that I have explained the operation, let me describe the optical problem I am encountering.
In short, I observe a problem of "drift" in the light intensity detected by the light-to-electric signal converter. Let me explain it better. The LEDs (unless proven otherwise) always absorb the same amount of current (the voltage across the resistor is always the same over time, and they operate in the linear region), so they always emit the same light intensity. I observe that depending on how I bend the optical fiber and how it is touched and moved, the optical fiber affects the detected light signal, accentuating or attenuating a constant increase in detected light. If the optical fiber remains bent with very pronounced curves, I observe the drift phenomenon, i.e., a weak but constant linear increase over time in the light intensity of the LEDs or the fluorescence of the analytes, as shown in the plot attached. Let me explain it better. If the fluorescence peak of an analyte at 523 nm is, for example, 250 counts at a certain concentration, if no additional concentration of the analyte is added, and the circulating solution is always the same, then the fluorescence of the analyte at that constant concentration should remain constant. Instead, I observe a linear, weak but persistent increase in the fluorescence of the analytes at all emission wavelengths. So, if at time t, I observe fluorescence at 523 nm equal to 250 counts, after, for example, t + Dt, now the entire spectrum has grown, and, for example, the peak is at 1000 counts. Obviously, this "drift" in light intensity is not due to the LEDs because I believe their light intensity remains constant over time since the current they receive remains the same. At most, the LED intensity should decrease over time. It is not due to the auto-fluorescence of the analyte from ambient light because this phenomenon is observed even when the room is completely dark. There are no chemical reactions or degradation of the analytes (I would observe a decrease in fluorescence, not an increase because the degradation products of these substances do not absorb at 445 or 340 nm). Also, this positive drift phenomenon is observed even with plain water or air. It is not due to problems with the light-to-electric signal conversion electronics or due to bugs present in the source code.
I attribute it to the optical fiber, which, if it is straight, this problem is attenuated, but if it is very bent, this effect of increasing the detected light manifests. Attached, you will also find photo and plot files showing the abnormal trend over time of the fluorescence of the analytes or the light captured by the LEDs. I do not believe this anomalous trend depends on the light source.
I'm using a FT600EMT - 0.39 NA, Ø600 µm Core Multimode Optical Fiber
In your opinion, what could this uncommon phenomenon in the optical fiber be due to? And how could it be resolved?
I would appreciate any suggestions. Thank you for your patience and the time you've taken to read my question.
Best regards,
Lorenzo
Dear Lorenzo Agostino Cadinu ,
I think the key of your drift phenomenon is the non-optimal distribution of the modes in the multimode fibers of your set up.
You mentioned above, that when touching the fibers, you see a change in the intensity. Any movement of the fiber, even a slight and slow one will give a change in intensity ouput.
In order to achieve an optimal mode distribution one uses a 'mode scrambler', which is in principle a part of the fiber being coiled or periodically bended just after the input port. In your case I think, that is overkill, because you have to have a quite short fibers to minimize intensity losses.
In your case I recommend
a) to support the fibers cables over their full length between both connectors and fix them at various points with tape or children's play dough (= the physicist's magic cure), so that no mechanical creeping can show up at all. Fix even the connectors by that play dough...,
b) wait for some time ( e.g. ~ 1/2 h) in order to achieve temperature equilibrium while your set up is switchted ON on all the time,
c) do not touch the set up...
d) switch off the ambient light or put a black cloth or a cardboard box over your set up...
Good luck and
best regards
G.M.
Dear Gerhard Martens,
Thank you very much for your detailed response and the advice you've shared. I truly appreciate your interest and willingness to contribute.
I'm curious: do you have any idea about the underlying physical phenomenon behind this drift in light intensity? In theory, bending the optical fiber, especially in an S shape as shown in the first photo, should cause mechanical stress and, consequently, a decrease in light signal. However, in my observation, the opposite occurs, with an increase in light intensity over time.
Based on your experience and knowledge, have you ever encountered a similar phenomenon? What could be the cause of this seemingly anomalous behavior, where the light intensity detected by the LED increases over time (even in the presence of only air or water without analyte and without cuvette), as shown in the bell-shaped plots of the LED and in the peak growth trend chart?
Thank you again for your time and for any additional insights you can provide on this matter.
Best regards,
Lorenzo
Dear Lorenzo Agostino Cadinu ,
thanks for feedback.
I have no explanation for that behaviour...
With respect to the drift:
does the LED has constant output apart from constant voltage and current? Rise in temperature while operating the LED(s)...
But unfortunately some papers report a decrease of LED output with increasing temperature...
Just check that possible issue for your LEDs without using a fiber cable.
Good luck and
best regards
G.M.
Dear Gerhard Martens ,
I measured with a multimeter the voltage at the ends of the LED resistor and noticed that the voltage always remains constant over time. I could infer that the current flowing in the LED is always the same. The electronic board of the LEDs does not overheat.
At this point, it occurs to me that there could be other problems, perhaps to how the light signal is converted into an electrical signal or some hidden bug in the source code. Yet, I have noticed that by stretching the fiber this slow but continuous growth phenomenon is mitigated or vanishes. That is why I see fiber as the cause.
Unfortunately, I can't find any possible explanations anywhere.
I still thank you for your advice and response.
Best regards,
Lorenzo
In the brainstorming mode, i would propose the following: perhaps LED is warming up and light intensity distrubution gradually drift in such way that outer sides emit more than the central part, that may happen due to heating in LEDs. Or even some nonuniformity of the walength can be observed. In typical optical situation, the outer layer of LED being focused into the fiber will contribute more to more out of axis rays (thinking geometric optics for such larhe fiber). Such rays more likely to escape when fiber is bent...
Dear.
Lorenzo Agostino Cadinu ,
the fakt, that the electronic board does not overheat does not mean that there is no temperature dependence with respect to the LEDs.
The temperature dependence may be due to 2 effects:
a) change of intensity and intensity distribution as Nikolay Pavlov meant,
b) change of light coupling into the fiber port due to temperature dependent relative mechanical movement of LED(s) and fiber inputport due to thermal expansion. Such movements*) take place any time when the temperature is changed, not only when the board is overheated.
So in consequence: fixing the fibers and waiting for thermal equilibrium.
I admit, that waiting for that equilibrium is a boring job. Once before I were participating in an x-ray experiment, the X-ray source of which achieved a stable intensity output at about 1,5 hours after switching on...
*) see for example the measured temperature dependence of coupling losses of fiber optic connectors as the solid lines of in Fig. 8A and 9A of:
Article Improvements of the data link dependencies of fiber optic se...
Excerpts of the figures are attached...
Best regards
G.M.
- Badges
- Science topic
- Similar topics
- Instrumentation Engineering
- Instruments