Measurements of Small Scale Light Changes in Water

[Lux Natura]

This is a work in progress, and I will be posting updates here.

In large part, this was inspired by a picture @ghetto alchemist posted in this post. Specifically, the image showing the changing measurement of illumination in an aqueous solution.



The research is from Vladimir L. Voeikov and his Russian colleagues. There are a few papers they published, you can read one of them here.

I had, in the past, while exploring the radioactivity of americium-241 found in old smoke detectors, played around with using webcams as data collection tool. With that in mind, I have written some python code that will be used to hopefully detect changing light intensities in a solution. Right now I am waiting on a shipment of luminol to begin testing.

The solution, as per Vladimir L. Voeikov, is composed of luminol and iron II sulfate in a bicarbonate and water solution, as luminol is not soluble in neutral water, but soluble in alkaline water. The amounts of luminol and iron sulfate are 10 micromolar each, places in 10ml of bicarbonate solution. The actual bicarbonate concentration is not known, so it will be less that full saturation for the test. I will still have to do a bit of research on the best concentrations of luminol to use.

In their tests, they used quite sensitive detectors, and light amplifiers to amplify the photons, which I do not have. Thus I will be scaling up the size of the vessel 50 or 100 times, using a 500ml or 1000ml of bicarbonate solution with luminol and iron sulfate.

The code is designed around the parameters of the original test - the light intensity is sampled 10 times per second, and averaged together for a reading each second. All that data is then piped to a separate file, which will be used to create the charts.

The solution and camera will be kept in a completely sealed dark box for the duration of the testing, and a baseline calibration without the solution in the box will be done before hand, to get an idea of what the values are for how noisy the camera CMOS sensor is, so it can be subtracted from any results.

Rough calculation show that the webcam I have for this project is about 10% effective, that is, for every 100 photons that would go towards the camera, ultimately only 10 end up hitting the CMOS sensor. Ultimately, this may mean the project is a failure, unless using a larger volume of liquid can allow for more photon excitations, which is entirely why I am running it on such a larger scale.

The next project update wont be a for a while, but I wanted to introduce this to others who may be interested. Perhaps if it shows results, other salts can be tested as well.

 

[ghetto alchemist]

The solution, as per Vladimir L. Voeikov, is composed of luminol and iron II sulfate in a bicarbonate and water solution

You do realise that you need to use the right kind of water to replicate the results right?
The authors of the linked paper specified that they used "hydro carbonate natural water".

Fortunately in your case you have some close by, since you're in North-west USA and you have access to ground water from a well.
IMO you should make sure to use your water immediately after it comes out of the ground.
Eg: don't leave it sitting exposed to air for any extended period of time before you use it.

Otherwise good luck in replicating the anomalous result.

 

[Lux Natura]

Eventually I'm going to try with a few different types of water, including our own well water. I have the analysis for it - I know exactly what's in there. This is just a small side project - we will see if it gives any sort of results, and what else can be done afterwards. I'm mainly going by the information as found in the paper linked above for now.

 

[Lux Natura]

Some details while I work on the setup:

I'm using a Logitech C270 webcam I had used for other experiments. It's an older camera with just under a 1 megapixel sensor. Pretty old tech by todays terms, but is 921600 pixels that can potentially capture data. Half the pixels have a green filter, and quarter have blue and red filters, so while they filter the light, I am discarding all the data except the brightness portion.

The brightness reading of each pixel is from 0 (no light) to 255 (maximum light), but with testing, I have determined that the sensor noise or transistor noise peaks just under 0.001 when averaged for the nearly 1 million sensors. Further testing will be done to verify this, but for now that is the baseline noise floor, so reading at 0.001 and below will be ignored as they are likely just sensor fluctuations. I'm actually monitoring to many more decimal places, but the noise floor right now seems to be at 3 decimal places.

If this experiment works out, I may look into much better (consumer grade) sensors. The lab grade sensors are not something I want to mess with at this time, unless aliexpress or some other such site has them for cheap.

10 times a second, the code adds up all the brightness data from the just under 1 million sensors and averages them, giving a final overall brightness reading.

There are two glass containers I will be using for preliminary testing. One is clear glass, the other has a layer of adhesive aluminum tape on the outside, burnished, in order to try to decrease any potential light loss through the sides of the glass, as the camera will be looking downwards towards the top of the inner glass. These inner containers have the ionic solution with luminol inside them. These containers sit inside of a larger glass container, which has been first painted white, to maximize the inner light reflection, and then the outside has been painted black, to limit outside light bleeding in. The outside container has a lid that screws on, to which the camera is attached, creating a setup that does not bleed in any outside light, so only the camera lens look into the container, and the rest of the camera is outside of it.

Some general thoughts about how this works in theory. The ionic (salt) solution allows for electron conductivity, and luminol requires an alkaline solution to dissolve. The luminol is able to attract free electrons, and ultimately can emit a glow. This is why in the tests in the paper linked above, they added iron sulfate, as it is an electron donor. Without an electron donor such as iron sulfate, the electrons will have to come from somewhere else, but it may still be possible to use an electron donor, and see periodic or cyclical changes occurring nonetheless, if, for example, there are environmental variables that interact with the overall system, such as time of day, gravitation, moon phase, etc, that may boost the rate of electron creation, donation, or absorption, or perhaps increase the capacity of the ionic solution to facilitate such transfers. In order to stop potentially changing pH values in the inner container due to small amounts of evaporation (and to stop the camera sensor from being in the path of the released humidity) the top of the inner chamber will be sealed with saran wrap. I have also considered using a magnifying lens to seal the top, which could help focus in any released electrons towards the camera sensor, but I do not have the right size magnifying lens currently.

Will post more info and updates as this progresses.

 

[Lux Natura]

General ideas of the ionic solution setup.

Luminol requires a solution with a ph of around 10-11 to dissolve into solution. I will be using sodium carbonate to achieve the required ph, and sodium carbonate to provide the necessary buffering, approximately 10g/L of carbonate and 15g/L of bicarbonate. This should create an ionic solution with a ph of 10.5 with a large ph buffering capacity. At this ph, luminol can dissolve around 2g/L into solution, but I will be using 0.5g/L instead, as this is the high end of the mixture for most types of chemical luminescence experiments from the data I have found. From what I have read online, increasing the luminol even up to its maximum saturation point does not result in a stronger glow.

Calibration of the camera / CMOS sensor data.

I have run 20 hours of analysis of the output of the CMOS sensor logging data in a 100% dark environment. The light value data is polled 10 times a second, averaged per second, and graphed. When the camera starts up, the overall readings increase for the first 1-2 hours, as the circuitry and silicon inside reach operating temperature. Afterwards, the data being output maintains a steady line, with only a 0.0039% variance. This variance is the averaged noise inherent in the CMOS sensors, or more specifically the transistors and amplifiers behind each pixel. When I gather data from using actual ionic solutions, that 0.0039% variance will be removed, or rather, a 0.0039% change in data is the minimum threshold at and below which no actual changes will be captured in the graph.

 

[Lux Natura]

I've completed the preliminary calibration for the camera, running it for over 6 days in a completely dark enclosure in an area that does not see much (if any) temperature fluctuations. Along with the readings, what the graph shows is local sunrise/sunset and moonrise/moonset laid over the chart. Its running on a dedicated linux laptop that has nothing else loaded than what it requires for the script to run.

There is some noise inherent in the system, but it does not seem to be directly related to either solar or lunar cycles. This is transistor noise basically from the camera. This noise is greatly amplified for the purposes of collecting data, but the chart is scaled so the lowest point still reaches close to 0. Each 1 minute sum is actually the total of all 60 seconds of data collection, and each second of data collection is the average of 10 samples, each taken at 1/10th second intervals by the code. I can go up to 30 samples per second, but I don't think that is needed. The solar / lunar cycles are geolocated for where I am, so they are as accurate as possible. I'm posting this now because I will soon begin with the actual testing, using the luminol. I've decided to use our own well water for the first test, as it does contain some trace amounts of dissolved iron, and other sulfate salts, which may aid in providing the free electrons to the luminol, hopefully returning some results.

If the well water and luminol yields results, I am open to testing it against distilled water water with carbonate, bicarbonate, and luminol, and those against spagyric preparations as well, to see if there is any difference in results.

 

[ghetto alchemist]

it does not seem to be directly related to either solar or lunar cycles.

Interesting that you say that, because when I look at the data, I see lower than average results around 6:00AM-8:00AM each day, and higher than average results between 4:00PM-8:00PM each day.

If the well water and luminol yields results, I am open to testing it against distilled water water with carbonate, bicarbonate, and luminol, and those against spagyric preparations as well, to see if there is any difference in results.

That sounds like a good idea.
Glad to hear you've decided to go with the well water after all.

 

[Lux Natura]

Interesting that you say that, because when I look at the data, I see lower than average results around 6:00AM-8:00AM each day, and higher than average results between 4:00PM-8:00PM each day.

Now that you mentioned it, yes. I see that too. Although this was just a test in a dark sealed container, and it's basically "sensor noise", those sensors are still responding to some changing variables, whether it's tiny changes in resistance due to temperature, something in the laptop, or even the wires going to the camera picking up some minute changes in the electrical or magnetic fields of the surrounding environment. I know there are ways I can insulate from these changes in order to get a more precise sensor baseline reading, but it occurred to me that, in a way, this "sensor noise" may be ultimately also picking up those things which the ionic solution will later be also reacting to, thus that sensor noise may actually be part of the useful data. Only when I have the actual readings will I know one way or another if this theory is correct.

That sounds like a good idea.
Glad to hear you've decided to go with the well water after all.

I was looking back on the most recent water analysis we had done on the water, and realized it's actually perfect in so many ways. Already contains significant carbonates, sulfates, and iron as well as manganese, which like iron is a good electron donor.