Project name Main project page

## December 2013

### 18.12.13 SNARF & NBA in H$_2$O without buffer

 x mM NBA 0 1 2 4 8 H$_2$O in µl 49 42.75 36.5 24 0 NBA in µl 0 6.25 12.5 25 49 SNARF in µl 1 1 1 1 1

SNARF stock concentration: 1mM

NBA stock concentration: 8 mM

Volume: 50µl

SNARF concentration: 20 µM (V/V_sample = c/c_stock)

## January 2014

### 08.01.14 Determination of buffer impacts on NBA-induced pH drop

Preparation of buffer sequence:
Desired sample volume: 1000µl
Buffer: 0.1mol stock concentration, pH 8,03 (Sörensen phosphate-buffer)
NBA: stock concentration: 8 mmol, desired sample concentration 2mM

 x mmol buffer 0 1 2 5 10 20 30 40 50 60 70 buffer in 10 µl 0 1 2 5 10 20 30 40 50 60 70 H$_2$O in µl 750 740 730 700 650 550 450 350 250 150 50

Measuring pH of each sample yielded values around 7.9 to 8.1
Now 250µl NBA have been added to each sample.
Measuring pH again: the 20...50 samples buffer the NBA as assumed, pH drops only about 0,01
pH drop of low buffer concetration sample reasonable higher: from 8,1 down to (lowest) 7,7
By adding the right part of the buffer, the pH value can be shifted back again, so that all samples are in the same pH range.
For samples 1,2 and 5 one has to be extremely careful, otherwise it will go excess and the other part of the buffer has to be readded.
The first three samples therefore ended up with 20% more volume so they were repipetted to 1000µl again.

pH-values of samples after adding NBA

 x mmol buffer 0 1 2 5 10 20 30 40 50 60 70 pH / 7.962 7.963 7.998 8.022 8.03 8.031 8.046 8.045 8.009 7.984

Now 47.5µl-size samples are taken for testing them in the laser setup.
Adding 2.5µl SNARF, creating a SNARF concentration of 50µmol.

### 16.01.14 SNARF Calibration

1) Preparing missing Sorensen buffers with pH in the ranges of 5.4, 6.8 and 7

• ratio of 93[A]:7[B] yields 5.688, adding 45µL [A] yields final pH value of 5.675
• ratio of 96[A]:4[B] yields 5.364
• ratio of 51[A]:49[B] yields 6.750 which is fine
• ratio of 39[A]:61[B] yields 6.950, adding 2.5µL[B] yields 6.961

2) Remeasurement of remainig Sorensen buffers

 labeled pH 5.8 6 6.5 7 7.5 8 measured pH 5.71 5.915 6.43 6.961 7.46 7.96

3) Remeasurement of non-Sorensen buffers

 labeled pH 3.89 4.82 9 measured pH 3.835 34.859 8.48

4) Sample preparation
Sample volume: 50µL overall volume with usual SNARF dye @ 1mM:
We want 20µM dye concentration, so we use 1µL SNARF per sample and 49µL buffer solution
Adding a with pure water and a ferric (FeCl2) and sulfuric (Na2S) sample (pH not measured)

Halved capillaries (0.30x3.00mm) are filled with samples and put under excitation LED

5) Measurement
Background light in each wavelength channel is measured and subtracted from the absolute fluorescence values of the buffer samples (again, each channel separately).
Fluorescency in both SNARF channels is measured (approx. 10 pictures, time-evolution by bleaching etc. is not relevant) and we obtain the fluorescence ratio:

$\dfrac{\text{pointa()}-\text{background(abs,a)}}{\text{pointb()}-\text{background(abs,b)}}=\dfrac{\text{pointa()}-107}{\text{pointb()}-100}$

 measured pH ratio 3.835 1.16 4.859 1.21 5.675 1.27 5.71 1.18 5.26 1.22 5.36 1.21 5.915 1.15 6.43 0.96 6.75 0.74 6.961 0.70 7.46 0.41 7.96 0.23 8.48 0.11

6) Analysis: Calibration Curve

We fit a sigmoid-curve to the datapoints:

In blue: datapoints for ferric and sulfuric samples mit estimated pH-value (just for seeing if the points could lie on the curve somehow)

7) Analysis: pK$_{\mathrm{A}}$-Value
Originally, we wanted to obtain the coefficients for the equation which provides the pH-value of our sample by processing the relative fluoerescence intensities of SNARF.
We define $R=\dfrac{F_{\lambda 1}}{F_{\lambda 2}}$ and it is important to normalize both fluorescence intensities ($F$) separately by substracting the background illumination.
$\lambda1=580$nm and $\lambda2=640$nm (in Labview: $\lambda X$ is the upper channel).

With $A$ and $B$ noting the acidic and basic endpoints of our titration it holds:

$\text{pH}=\text{pK}_{\text{A}}-\log{\left[ \dfrac{R-R_B}{R_A-R}\cdot\dfrac{F_{B(\lambda2)}}{F_{A(\lambda2)}}\right] }$ Source: SNARF-Manual

Now we plot that equation with the fixed values for $R_A, R_B, F_{B(\lambda 2)}, F_{A(\lambda 2)}$ and receive pK$_\text{A}$ as it is the y-intercept of a linear fit:
($y=ax+b \Leftrightarrow \text{pH}=\text{pK}_{\text{A}}-x$) with slope of 1 while $x:= \log{\left[ \dfrac{R-R_B}{R_A-R}\cdot \dfrac{F_{B(\lambda2)}}{F_{A(\lambda2)}}\right] }$

We get $\text{pK}_{\text{A}}\approx 7.2$. Now we have a calibrated equation for our setup and we can calculate the pH value for the samples by only measuring the relative intensities.

### 22.01.14 Calibration with UV-Laser I

1) Remeasuring pH-Values of Buffer Samples

 number of sample 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 pH 2.87 3.821 4.856 5.25 5.368 5.673 5.707 5.915 5.92 6.44 6.978 6.756 6.983 7.512 8 8.482 8.724

2) New Routine

An improved pattern for LED & Laser exposure to simplify background normalization procedure:

### 23.01.14 Calibration with UV-Laser II

3) Data Analysis Radial dependency of the absolute fluorescence (minus background) is evaluated for LED exposure, LASER exposure and LASER+LED exposure.
Afterwards, the ratio is calculated with IGOR:

 Radial dependency of the Fluorescence-Ratio generated by LED-Excitation Radial dependency of the Fluorescence-Ratio generated by Laser-"Excitation" (Crosstalk) Radial dependency of the Fluorescence-Ratio generated by LED&Laser-Excitation

The first image clearly shows the pH-dependent emission-ratio of SNARF. The higher the ratio, the higher is the pH-value (violet curve is sample 17 with pH=8.72). As expected there is no radial distribution.
The second image shows the ratio of SNARF's fluorescent emission by (uninentional) excitation with the UV-Laser. While the ratiometric analysis helps to avoid a lot of trouble (e.g bleaching of SNARF, varying SNARF concentration,...) the crosstalk of the Laser has to be substracted because SNARF has different pH-emission-curves for different excitation wavelenghts.
The radial dependency of the ratio at the laser+led graphs is caused by the crosstalk and the spatial confinement of the laser beam.
For greater radii, the laser-caused fluorescence merges with the emission caused by LED excitation.
If the emission ratio was independent of the excitation wavelength and only depending on the pH value the curves should be approximately constant for each pH, regardless of the laser being switched on or not. In the experiment, the UV-crosstalk shifts the ratio in the area of the laser beam.

### 29.01.14 SNARF pH Measurement NBA

1) Sample Preparation

 NBA concentration in mM 0 1 2 4 6 H$_2$O in µl 45 38.75 32.5 20 7.5 NBA in µl 0 6.25 12.5 25 37.5

NBA stock concentration is 8mM. To each sample 2.5µl SNARF (@1mM Stock) and 2.5µl buffer (pH 8.0, sample #15) are added. SNARF concentration therefore is 50µM.

2) Measurement
Pattern (65s timestep): 100: LED on, 300: Shutter off (=Laser on) , 900: Shutter on (=Laser off), 1300 LED off

3) General LabView Procedure
We want to subtract the background fluorescence from the signal. Because the background and the signal data originate from different measurements, and the capillaries aren't placed onto the exact same spot every time, we have to rearrange the areas where we take our data from every time. We have to do this for both channels of each capillary (we want the same area for each of the channels, too).

These steps have to be followed:

1. Load background and signal images
2. Determine reference points (= center of laser beam) for background signal
3. Arrange one area in desired way (centered at the laser spot) and update into picture area
4. Now flip the area to the other reference point by switch reference and afterwards update
5. Remember the reference point coordinates
6. Now change loop index to view the actual signal images
7. Move reference point to the center of laser spot and flip the picture area from the background reference point to the newly defined point with Switch reference.
8. Update the new picture area
9. Readjust the moved reference point to the background image value and repeat (7)-(8) for the other channel
10. The same reference point values found for background images have to be used every time
11. Only steps (6)-(10) have to be repeated for the other signal images as the picture areas are being aligned to the correct reference points every time

After every Analyze process LabView has to be stopped and restarted (reload pictures) to clear the memory.

4) Evaluation to be continued

### 30.01.14 SNARF pH Measurement at various Buffer Concentrations

1) NBA-Mixture-Induced pH-Drop for various Buffer Concentrations
250µl NBA samples with mentioned buffer (pH 8.0, sample #15) volumes are filled up to 1000µl with water.
Impact of NBA is measured with pH-meter:

 Buffer Volume in 10 µl 0 1 2 5 10 20 30 40 50 60 70 pH 3.348 7.66 7.909 8.084 8.12 8.105 8.089 8.066 8.037 8.027 8.006

2) UV-Induced pH-Drop of NBA
For UV-induced pH-drop, 50µl subsamples are taken with 2.5µl SNARF and 47.5µl of each sample mentioned above.

## February 2014

### 10.02.14 Improving the Setup

The Thorlab Optical Chopper System MC2000 is added to the setup.
Now, the uncaging of the protons by excitation with the UV-Laser is decoupled from the measurement of the fluorescence.
Previously, the UV-Laser disturbed the fluorescence measurement:
Fluorophores were unintentionally excited by the UV-Laser, too, causing overexposure and demanding a modified calibration curve.
Now the chopper periodically switches UV-excitation on and off while the LED constantly excites the fluorophores. Whenever the UV-Laser is off, a picture is taken.

### 11.02.14 NBA Effects on Fluorescent pH-Sensitive Beads Part

The movement of specific pH-sensitive, fluorescein-marked beads is measured.
Therefore, we prepare samples with varying NBA-concentrations, firstly using the old NBA sample (to see if there are aging effects).

 NBA Concentration in mmol 1 0.64 0.32 0.064 NBA in µl 6.25 4 2 0.4 Water in µl 42.75 45 47 48.6

To each sample of 49µl, a 1µl drop of bead+fluorescein is added.

The whole procedure is repeated with a freshly mixed 1mM NBA sample (stock: 8mM).
The adjustment of the laser focus has to be done beforehand with a additional capillary containing some UV-fluorescent dye, since focussing the laser spot with the bead samples is difficult. The focus of the camera should be adjusted for every capillary, so that a nice amount of beads is in the focal plane.

 Time-resolved (green-red) Projection of Bead Movement without NBA Time-resolved (green-red) Projection of Bead Movement at 0.32 mM NBA Time-resolved (green-red) Projection of Bead Movement at 1mM (old) NBA

### 12.02.14 SNARF Calibration with Chopper Setup

With the new Setup, the calibration of SNARF has to be redone as the effects of the laser are reduced in the measurements.
At first, we want to measure the effects of different NBA concentrations.
Appearently, SNARF is phosphorescent, so even with the chopper, the laser spot is visible during the measurement because the dye shows an afterglow.

### 13.02.14 SNARF Calibration

1) NBA Concentration Array
We prepare samples with different NBA concentrations in a pH 7.5 buffer. Since the NBA concentrations we used the last times seem to be already saturating, we chose to expand the sample array in the lower NBA concentrations this time. Also, the calibration curves we got last time, recommend a smaller pH value (7.5 instead of 8.0).
We want to measure the pH value, too, so we are using 100µl sample volumes (otherwise the electrode would not fit in the sample).

 NBA concentration in mM 0 0.1 0.5 0.7 1 2 3 4 NBA in µl 0 1.25 6.25 8.75 12.5 25 37.5 50 H$_2$O in µl 90 88.75 83.75 81.25 77.5 65 52.25 40 pH value 7.54 7.57 7.53 7.55 7.55 7.52 7.53 7.49

NBA stock concentration is 8mM. To each of the subsamples with 50µl desired volume, 2.5 µl SNARF (@1mM Stock) and 2.5µl buffer (pH 7.5, sample #14) are added. SNARF concentration therefore is 50 µM.
Additonally, a capillary without dye is measured for determination of background illumination.
The chopper's phase is adjusted in that way, that the laser spot is minimum visible.

 0mM NBA 0.1mM NBA 0.5mM NBA 1mM NBA 2mM NBA 4mM NBA

Ratio of the radial average of the two channels versus time:
Approximately: 0-120: LED on, 120-420: Laser and LED on, 420-750: LED on, 750 -820: Laser on
Further evaluation on bigger radii here: 19.02.14
See pH Time Traces with SNARF

Carboxylated and fluorescent (BCECF-like) beads of various dimensions (2µm - 0.02µm diameter) will be used to determine the movement and flows of charged particles along the pH gradient.

 Bead Dilution 1:100 1:100 1:250 1:250 Bead Volume in µl 5 5 2 2 NBA in µl 3.125 1.56 3.125 1.56 Water in µl 41.875 43.44 44.875 46.44

Two samples with 0.5mM NBA and two with 0.25mM NBA were used, beause the 0.5mM NBA samples already showed a great movement. With the smaller beads being less bright, the 1:100 dilution is more convenient.

 laser off laser on

Three 50µl samples are prepared:

 Bead Dilution 1:50 1:100 1:250 1:500 Bead Volume in µl 10 5 2 1 NBA in µl 3.125 3.125 3.125 3.125 Water in µl 36.875 41.875 44,875 54,875

Each sample therefore has a NBA concentration of 1mM (NBA-Stock: 8mM). The mentioned bead volume is an already pre-diluted sample (1:10 with 20µl beads, 180 µl water).

The picture sequences show a inward stream superposed with an outward stream, appearently depending on which focal plane is observed. To determine this effect, three additional sequences are recorded. The lowest and highest focal plane and also one in between.

 top focal plane mid focal plane bottom focal plane

2µm beads with 1:100 dilution and 0.5mM NBA (timescale is black-red black-green again)
There is a strong inward (outward) movement on the bottom (top) of the capillary.

### 19.02.14 Evaluation of SNARF Calibration on 13.02.

 Time-dependent Ratio for various c[NBA] Comparison of 10px and 85px Radii at 4mM NBA

Taking a bigger area improves our signal a lot. Also, the greater radius shows a decreased pH effect (as assumed).

### 21.02.14 Evaluation of SNARF Calibration on 13.02. Part II

With the [pH-conversion recipe] the pH values can be calculated for the 85px radii.
$pH=7.7309-1.4921\cdot\log_{10}(\frac{R}{1.5354-R})+\log_{10}(0.1295)$

### 25.02.14 0.5µm & 1µm Diameter Bead Movement

 bottom focal plane mid top

 bottom focal plane mid top

### 26.02.14 NBA and SNARF with NaOH pH Adjustment

This time we want to measure the NBA induced change in pH value, without an additional buffer.
Since NBA is acidic, and pH detection with SNARF doesn't work well in pH below 7 we are adding NaOH (5mM) to achieve a pH value of around 8.
The dye will decrease the pH a little bit, so the final pH before the measurement was meant to be safely determined to be between 7 and 8 (which appeared to be very wrong!).
Sample volume for pH measurement is 10ml, because the electrode of the pH-meter won't fit in very small samples and pH regulation is easier for bigger volumes.

 NBA concentration in mM 0 0.2 0.5 0.7 1 2 4 NBA in µl 0 2.5 6.25 8.75 12.5 25 50 H$_2$O in µl 100 97.5 93.75 91.25 87.5 75 50 pH value 26/02 6.3 8.1 8 8 8 8.4 8.0 pH value 27/02 6.3 6.5 6.5 6.3 6.4 6.3 <5.5

With decreasing NBA concentration we add less NaOH (around 40µl for the 4mM sample, to <10µl for 0.2mM NBA).
Afterwards, 2.5µl SNARF are added to a 47.5µl subsample, creating a 50µM SNARF concentration.

Pure water with a 50µM SNARF concentration (475µl water, 25µl SNARF due to big pHmeter electrode) has pH 5.3,
so pH shift (-1) induced by SNARF has to be considered when measuring without a buffer.
Appearently, adjusting the pH just with NaOH molecules does not generate stable pHs.

### 27.02.14 Analysis of 26.02.14

The analyzed area is 45x90px

1) Ratios

 0.2mM NBA 0.5mM NBA 0.7mM NBA 1mM NBA 2mM NBA 4mM NBA

2) pH-Profiles

 pH at 10px pH at 15px pH at 25px pH at 40px

"Appearently, adjusting the pH just with NaOH molecules does not generate stable pHs."]
Since the pH plots differed a lot from our expectations, the pHs were measured once again and as cited above, they dropped about 2 counts over night.
Since SNARF has a pKa of ~7.5, it is best to be used in the pH range between 7 and 8.
This is probably the cause for the strange behaviour (pH increasing despite NBA uncaging..).

## March 2014

### 05.03.14 DNA

1) Sample Preparation
2mM NBA samples are prepared for 2, 5, 10, 22, 50 and 80mer. Samples are buffered with pH 8.5.
DNA-Samples are 100µM except 2- and 5mer (500µM).

 Concentration in Sample Volume NBA (@8mM) 2mM 12.5µl Buffer pH 8.5 50µM 2.5µl DNA (@100µM) 1µM 0.5µl DNA (@500µM) 1µM 0.1µl H$_2$O in µl fill up to 50µl 34.5µl

2) Measurement
Laser power is adjusted to approx. 62mA (3.1mW), LED power is 200mA, filter set Cy3.
Correct LED is 530nm!
Laser focus is adjusted with BCECF (other filter & LED needed)

### 06.03.14 DNA Analysis

The intensity of 5x5 pixel areas centered on the laser spot and 10px away from the laser spot are averaged and plotted against time.
With 0px, there is a jump when the laser is switched on. It is to determine, whether the chopper wasn't adjusted right or there is some phosphorescence,
or if DNA in fact accumulates this quick at the laser spot (and later bleaches out in the laser spot region).

 0px 10px

To account for the bleaching of the fluorescent labeled DNA(which is considerably larger with NBA, which is yet to be investigated, too)
we fit the trace (only LED on, Laser off) with a exponential function and we normalize the original waves by dividing through the fit.

 Exponential fits Normalized (without 2&22mer)

Radial profiles for different times are plotted for each -mer.

 2mer fits 5mer 10mer 80mer

Radial profiles for different -mers are plotted with set time.

 Frame 140 (~2frames after Laser ON) Frame 155 Frame 200

It seems that DNA is accumulating at the Laser spot (bleaching by laser is not accounted, but that effect would actually lead to reducing fluorescence instead of increasing).

## April 2014

### 23.04.14 DNA+NBA in Different Buffer Concentrations

A new NBA solution (50ml) is mixed with 8mM concentration (151 g/mol). New 2-component phosphate buffers (each 200ml) are made, too.
Therefore we create 0,2M solutions of NaH2PO4 (A = Acid) and Na2HPO4 (B=Base). A ratio of 16(A)/84(B) should give us a pH of 7.5
Molar weight is (A): 156.01 g/mol and (B): 358.14 g/mol. To adjust pH values, we can add small amounts of A (acidic) or B (basic).
Sample volume is 1ml to allow measurement of pH with the pH meter.
The actual buffer added is the 7.4 mixture of A and B with 0.4M (0.2+0.2) concentration.

0mM NBA

 mM Buffer 0 1 2 5 10 pH 5.8->7.4 (+0.7µl B) 7.6-> 7.4 (+0.3µl A) 7.8 -> 7.4 (+0.5µl A) 7.8 -> 7.5 (+1µl A) 7.8 -> 7.5 (+ 2µl A)

1mM NBA

 mM Buffer 0 1 2 5 10 pH 5.8 -> 7.5 (+ 0.5µl B) 7.5 -> 7.4 (+0.2µl A) 7.6 -> 7.55 (+0.3µl A) (+0.5µl A) 7.8 -> 7.5 (+1µl A)

4mM NBA

 mM Buffer 0 1 2 5 10 pH 6.6-> 7.4 (+0.3µl B) 7.5 7.8 -> 7.54 (+0.5µl A) 7.8 -> 7.55 (+1.4 µl A) 7.74 -> 7.5 (+3µl A)

### 24.04.14 DNA+NBA in Different Buffer Concentrations: Results

Exceptionally strong bleaching ocurred and actual movement of DNA seemed questionable.
Taking one step back and observing the (if existing) movement (and bleaching?) of pH sensitive beads is scheduled.

 Intensity of 0.02µm Fluorescent pH-Sensitive Beads Intensity of 0.02µm Fluorescent pH-Sensitive Beads without Bleaching

Again, unusual strong bleaching occured (see above, left). Fitting the bleach curves exponentially and dividing by them should fix the problem.
Since the cause for the bleaching still remains unknown (probably old NBA? NBA or NBA-photoproduct doing something with the dye?), another proton uncaging molecule is ordered.
The normalizing of the intensity is yet to be repeated.
Igor Procedure for exp-fit:

#pragma rtGlobals=1 // Use modern global access method.
Function Normalize(input)
Wave input
Wave W_coef
String outputName = NameOfWave(input) + "_Norm"
Duplicate/O input, $outputName // Create a wave reference for output wave Wave output =$outputName
// Substract Background Wave and normalize by third frame
End

### 28.04.14 pH-Dependencies

The intensity varies about less than 10%. The two data waves presented were made from the same sample volumes, just different capillaries.
Still, there is a variation of about 100 counts, which means the error in focussing or placement of the capillary most probably is greater than the change in intensity due to pH.
Appearance of dark or bright spots after a local pH jump therefore can't be explained with pH dependent change in fluorescence only.

2) pH-Dependencies of Cy-5 Change filter set to #1 and LED to 627nm.
Cy5 appears to be unstable in acid regions "Cy5 is physically unstable in acid conditions and should be stored and used in buffers above pH 7."

### 29.04.2014 Check: pH-Dependency of Cy5

Repeating yesterdays measurment with samples freshly mixed directly before the measurement, and 10 minutes after (dark) storage.

 pH-dependent Fluorescence Degradation of fluorescence after storage

Left picture is from yesterdays measurement. This time, pH 2.78 was measured at the beginning, but still is lower. Thererfore the fluorescence has to be pH dependent.
Degradation due to storage occurs in dark environment.

### 29.04.14 Reducing Laser Power

1) Minimum Laser Power for Sufficient Uncaging?

Checking if NBA uncaging still works sufficiently for lower laser powers, to aim for less cross-bleaching of the fluorescent dyes due to UV excitation.
Currently used standard current was 80mA at 1.6V, resulting in intensities of 3-6mW (depending on adjustment of the laser).
Reducing the voltage gives a lower current, and the resulting power has to be measured. Unfortunately, the measured intensity is strongly depending on the accuracy of the laser adjustment, and even measuring the laser power can change the initial adjustment greatly. Therefore, laser power is measured before and after the measurement to ensure accuracy.
One 50µl sample is setup with 1mM NBA (6.25µl), 2mM of the phosphate buffer (pH 7.4, 0.4mM -> 0.25µl) and 5µl of the small (0.02µm diameter) beads.
The sample is stored darkly during the measurements.
Additionally, the same measurements are done with 0mM NBA.

 Voltage Current Intensity 0.9V 45mA 15µW 1V 50mA 200µW 1.6 81mA 3.0mW

Measurements performed with 4x4 binning and 30ms exposure.

Left picture shows the fluorescence intensities over time for some laser powers without NBA.
UV induced bleaching of the fluorophores does not appear for the lower two laser intensities, which is interesting!
This means, the kinetics shown on the picture in the middle (when UV is switched on) aren't solely caused by UV-bleaching.
But still, the problem of heavy bleaching with NBA remains unresolved and might also account for the different kinetics with NBA presence.

2) Saturation of NBA Uncaging?
Another aim is to see if the UV laser already bleaches all the abundant NBA at 1.6V.
To check the existence of a saturation plateau, the laser power is reduced to slightly lower values and the signal is observed.
As we can see, a different kinetic already appears for lower intensities. Thus, at 1mM NBA there's still plenty NBA present and higher intensities might uncage even more protons.

3) Should be checked: Is there a initial loss of intensity when having NBA present?

### 30.04.2014 Dividing by Exponential Fit to Account for Bleaching

 Normalized for LED-Bleaching Exponential Fits for 1mM NBA-Sample Linear Fit for 0mM NBA (nearly constant)

The fits only considers bleaching caused by LED!
If there weren't the strange interaction between NBA or NBA residues, the UV-bleaching could be separated from actual bead movement easily by just doing a measurement with 0mM NBA at the same laser power.
In the left picture the yellow curve shows the effect of UV bleaching. There is no NBA in this sample, so the kinetics must be caused by UV bleaching.
But it is likely that not only the LED bleaching, but also UV induced bleaching is amplified when NBA is abundand which can't be easily separated by the bead flows this way.
At least it is known, that the kinetics shown in the graph are not result of UV bleaching only, since we already could watch single bigger beads being carried away (and into) the region of the laser spot.

The jump around frame 600 in the green wave is an artefact, the data points actually connect smoothly.