Determination of atmospheric ozone content using absorption of red and blue light in an integrating sphere

Research Group: Maia Ketteridge, Aiden McIlraith, and Lauren Taylor

Launch: Whitworth Spring 2018

''This project continued the work done by previous Whitworth Near-Space teams to create a low-cost, light-weight integrating sphere that can effectively measure atmospheric ozone content. This prototype was focused on the improvement of circuit design and the calibration of the device. This experiment used 470 nm and 624 nm LEDs and a phototransistor to detect LED light intensity in order to correlate the absorption of red (624 nm) light to the concentration of ozone present. The red LED (624 nm) and blue LED (470 nm) were chosen based off of the premise that only the red LED will be absorbed as altitude increases and the blue LED would act as a control. However after analyzing the data, the conclusion was drawn that there was an unexpected variation in the signal from the blue LED, which was hypothesized to be attributed to the absorption of blue light by nitrogen dioxide. Unfortunately, the calibration of the integrating sphere was unsuccessful in relating the phototransistor output to the concentration of ozone. Due to the lack of ability to calibrate the sphere, only qualitative conclusions could be drawn from this experiment.''

The Importance and Effects of Ozone in the Atmosphere


The atmosphere is primarily composed of oxygen (O2) and nitrogen (N2) gas. However, ozone, which is composed of three oxygen atoms, occurs naturally at low concentrations in the troposphere and stratosphere. The troposphere is the layer of the atmosphere that is directly above Earth’s crust and reaches an altitude of approximately 7 to 20 km. The troposphere is responsible for the majority of the mass in the atmosphere (75-80%). The presence of ozone in the troposphere results in a significant number of agricultural and health related problems. Some of these issues include the development of asthma, smog in major cities, the inability to fight off bacterial infections of the respiratory system, and permanent lung damage.



Although higher concentrations of ozone in the troposphere can be damaging to the health of humans and can produce problems for major cities, ozone plays an important role in the protection from ultraviolet rays (UV-B and UV-A) when in the stratosphere. As illustrated in Figure 2, exposure to UV-B and UV-A rays have proven to be damaging to the DNA of living organisms, thus resulting in conditions such as melanoma and basal-cell carcinoma, both of which are forms of skin cancer. Ozone in the stratosphere acts as a barrier between these harmful rays and direct human exposure. This barrier is referred to as the ozone layer and is located approximately 10 to 50 km from the Earth’s surface. Due to the recent thinning of the ozone, the risk of developing melanoma has doubled since 1990. The ozone layer is being constantly depleted by human-produced chemicals. ODS, or ozone depleting sources, include chemicals such as chlorofluorocarbons (CFCs), hydrochlorofluorocarbons (HCFCs), halons, methyl bromide, carbon tetrachloride, and methyl chloroform. These ODS were used as items such as coolants or aerosol propellants. However, after the effects of these products were discovered the use of these items reduced significantly. Although prevention mechanisms have been put in place to reduce the thinning of the ozone layer, the ODS chemicals have long-lasting implications that are still currently depleting the ozone layer, specifically at polar regions. It is estimated that one chlorine atom can destroy 100,000 ozone molecules located in the stratosphere.

Measuring Ozone Concentration in the Atmosphere
There are different methods that can be used to determine the concentration of ozone in the atmosphere. The two principle categories of measurement techniques are local and remote. Remote measurements of total ozone concentration use the unique absorption of UV light in order to detect ozone at long-range distances from the instrument. An example of remote techniques being used are satellites. Satellites use the absorption of solar UV radiation by ozone in the atmosphere in order to measure the concentration of ozone over the globe on a daily basis. Some examples of sources of UV light are lasers, starlight, and the sun. Local measurement techniques require air that has a concentration of ozone to be directly drawn into an instrument. Once this air is drawn into the instrument, the concentration of ozone is determined by the absorption of UV light or by the electrical current or light produced in a chemical reaction with ozone. Lastly, there is a local measurement technique that uses the construction of ozonesondes, which are lightweight instruments that are capable of measuring ozone concentrations and are launched on weather balloons into the stratosphere. The purpose of this experiment was to improve the ongoing design of a low-cost, light-weight ozonesonde and use it to determine ozone concentration in the stratosphere as a function of altitude.

Integrating Spheres Theory
Integrating spheres are devices that are used to measure optical radiation by spatially integrating the radiant flux. There are usually four ports that are located at various locations on the sphere. Two of these are used as the input and output for gas, in our case, ozone concentrated air. The remaining two ports are used as the input and output for light in the form of LEDs and a phototransistor. The radiance of the inner surface of an integrating sphere is experimentally derived and allows for light flux to reflect multiple times before ultimately being detected by the optical radiation output, in our case, a phototransistor. These multiple reflections allow for the path length that the light takes to be larger than the diameter of the sphere itself. However, individual photons that are reflected in the sphere will take different path lengths. This has to be calibrated by using the multiplier \(M\) of a specific integrating sphere. The multiplier is used to calculate the average path length that a photon will undergo in the integrating sphere before being detected. The multiplier is derived from the equation \begin{equation} M = \frac{\rho}{1-\rho(1-a(\lambda))(1-f)}, \end{equation} where \(\rho\) is the reflectance of the spheres internal surface, \(f\) is the ratio of the area of the port openings to the total internal sphere surface area, and \(a(\lambda)\) is the wavelength dependent attenuation of the optical radiation over a single pass through the sphere due to absorption and scattering by the test gas.

The average path length that a single photon travels within the sphere is the product of the sphere multiplier and the average distance travelled by a photon for a single pass through the integrating sphere. An approximation can be made that the average path length of a photon will be two-thirds of the sphere's diameter. Due to this approximation, we can find the average path length to be given by the equation \begin{equation} L_\mathrm{effective} = \frac{2}{3} M D, \end{equation} where \(M\) represents the multiplier and \(D\) is the diameter of the integrating sphere.

Project Definition
The purpose of this experiment was to expand on the development of a low-cost, light-weight integrating sphere in order to measure the concentration of ozone in the troposphere and stratosphere. This experiment was focused on the improvement of circuit design and durability to ensure that data would be collected and returned from launch, as well as controlling for additional variables. This experiment utilized red and blue LEDs that emit photons across a large path length in the integrating sphere. After traveling across this path length, the photons were detected by a phototransistor and the voltage that the phototransistor output was recorded and graphed as a function of altitude. The specific LEDs that were used in this experiment were chosen because of ozone's ability to absorb certain wavelengths of light. The red LED (624 nm) is able to be absorbed by ozone and so it serves as the experimental variable in our experiment. The blue LED (475 nm) is unlikely to be absorbed by ozone, thus it was the control in our experiment.

We hypothesize that voltage output that is collected from the red LED will decrease as altitude increases. This is because we expect the concentration of ozone to increase with altitude, thus resulting in a greater absorption of the red LED light which causes the output of the phototransistor to be smaller in magnitude. However, the system will also be collecting data while the balloon is descending. This will result in a drop of ozone concentration so the voltage readings from the phototransistor will increase.

Originally, the voltages collected would have been converted to concentration of ozone using a mathematical relationship that would be developed in the calibration stage of this experiment. However, such a relationship was not found which allowed for only qualitative conclusions to be drawn from this experiment.

Integrating Sphere


The most important component of the project was the integrating sphere. Due to budgetary restrictions, we had to build our own low-cost and light-weight sphere. Our solution was to take two spray painted hemispheres and epoxy them together. Our integrating sphere was built from two hemispherical 4" Fat Daddio's cake pans which we purchased off of Amazon. The pans fit together almost perfectly to approximate a sphere while also providing a lightweight material to send into near space. These pans had various sized holes drilled in them at intervals to accommodate the LEDs, phototransistor, and fans. As seen in Figure 3, fan holes were cut at a 90 degree angle with each other in order to be able to fit it inside the box and acted as the gas input and output ports. The LED holes were adjacent to each other and 90 degrees from the phototransistor to provide maximum distance for light to travel within the sphere and the holes acted as the light input ports and the light output ports. Each of the printed mounts were then epoxied to the pod. The interior of the sphere was spray painted with white Rust-oleum flat protective enamel to allow for the light to dissipate properly toward the phototransistor after traveling a significant distance. The two halves of the sphere were first epoxied together at the seams without letting any epoxy get inside. Afterwards, duct-tape and electrical tape were used along the seam to ensure stability and to prevent light from entering the sphere.

Fan and LED Mountings


Two Nidec DC Brushless Fans were used and were held into place by 3D printed mounts which were designed by the 2017 near space group. They were printed using ABS plastic to reduce ozone absorption in the atmosphere. The round end of the mounts were filed down to be flush with the interior of the sphere to prevent excessive absorption of light by the ABS plastic. They were then inserted into the holes cut in the sphere and epoxied down. The flat side was inserted into holes cut in our foam box which held the sphere and our circuitry. The LEDs and phototransistor were held in place by 3D printed custom mounts which allowed for ease of replacement in case of malfunction. This became useful during testing. A 3D model of the LED mountings is shown in Figure 4. We used a blue 470 nm LED as our control light source and a red 624 nm LED as our variable light sources and a TEPT5600 phototransistor to detect the light. The mounts were both two pieces, one of which was epoxied down to the sphere while the other acted as a cap holding the LEDs and phototransistor in place. They were held together by a nut and bolt going through the center of both as well as a peg to prevent the two pieces from rotating against each other. The fans, LEDs, and phototransistor were connected to their respective circuits by wires that were external to the sphere.



Control Cylinder
In addition to the sphere, we also created a PVC tube that acted as a control which was 83.05 mm long and 33.25 mm in diameter in order to test the LEDs as temperature changed to ensure that our readings in the sphere were not affected by temperature or changes in the circuit function during flight. This cylinder was included inside the pod. The ends were capped with custom laser-cut mounts made of acetal (commonly known as Delrin). One mount had two 5 mm holes evenly spaced to accommodate one LED of each color and the other had one 5 mm hole in the center to accommodate our phototransistor. The completed control cylinder can be seen in Figure 5. We used the same LEDs and phototransistor as the sphere. The contraption was then sealed with epoxy with the phototransistor at one end and the two LEDs at the other. The tube was placed inside of the pod in order to ensure that our LEDs were not affected by the temperature of the environment. The sealed tube allowed us to detect changes in the intensity of light which would be indicative of the LEDs being affected by temperature, which would change our final data.



Pod
Our pod was custom made from foam by Foamular with dimensions of approximately 24 cm x 19 cm x 18 cm. Five of the six pieces of the box were epoxied together and the sixth became the lid after an 18.5 cm x 13.5 cm piece of foam was glued to the center to prevent the lid from falling off. A hole was cut in the foam insert in order to accommodate the launch computer as it was taller than the box. The box itself was outfitted with two square holes for the fans on adjacent sides of approximately 60 mm in length. The mountings allowed the fans to circulate air through the integrating sphere without excessive outside light contamination. The fan mounting was secured to the pod with epoxy. Additionally, the box had one hole for our launch computer LED readout. The launch computer allowed us ease of control from outside the box as well as more stable handling of the breadboard and the mbed controller. The box was then wrapped in mylar to allow for easier detection both from aircraft and on the ground. We then cut small holes in order to zip-tie all of our components to the pod to prevent any damage from landing. After our assembly, we packed our pod with packing peanuts and bubble wrap in order to prevent damage during flight and landing, as shown in Figure 6.

LEDs


Research on the absorption spectra of ozone found approximately 600 nm as the wavelength of maximum absorption of ozone within the visual light range. A 624 nm LED was chosen as the variable LED due to its low cost, accessibility, and its high absorption by ozone. The minimum wavelength of absorption was found to be approximately 390 nm, therefore a 425 nm LED was chosen as a control light source due to the fact that we suspected it would not be absorbed by ozone. The LEDs, in groups of two, were wired in parallel to a 50 Ω ceramic resistor. Ceramic resistors were chosen because they more efficiently dissipate heat to avoid changes in resistance with temperature.

Phototransistor
A phototransistor with a peak sensitivity at 570 nm and half-voltage range from 440 nm to 800 nm was chosen in order to get usable readings from both LED's. The phototransistor readings were first amplified through a dual operational amplifier, then the resulting voltage was filtered to reduce electrical noise. Two separate but identical amplifier and filter sets were constructed, one for each phototransistor. See Figure 7 for a diagram of the phototransistor and LED circuits.

Whitworth Launch Computer


The circuit was assembled on a pre-prepared Whitworth University Launch Computer. The computer contained internal circuitry to connect an SD card mount, external power switch, and external indicator LEDs. Our circuit was contained on a breadboard attached to the Launch Computer. A rechargeable LiPo battery was mounted on the underside of the computer in order to reduce the number of batteries that would have to be replaced for both testing and launch.

Fan Circuit
A separate circuit was constructed for the fan in order to reduce electrical noise. The fan circuit is powered by a single 9V battery. As seen in Figure 8, the voltage from the battery was run through a linear voltage regulator to drop the voltage to 5V which was required for the fans to operate. See Figure 8 for a circuit diagram.

Code


Here is a link to our code which contains comments with the specifics of our design. Our code was created in C++ in the mbed browser interface.

The code outputs data in a csv formatted .txt file that is easily converted to be used in Mathematica and Excel alike. This is very important for easy data analysis.

The diagram of the code, Figure 9, shows the basic layout of the code. The code is designed so that the lights switch states every 0.5 and 9.5 seconds. After each switch, the phototransistor averages 10 readings to avoid noise. The result is then added to the file with the current time from a running clock. The clock is a timer from Dr. Larkin's extended timer library. This library enabled us to run the timer for long enough that we could keep time for the entire flight. In addition, there is a ticker running in the background that saves the file every one minute regardless of where the code currently is.

Also of note, the red and blue lights next to the Launch Computer's external switch correspond to power and SD card failure respectively. The first line of the code turns on the red light, so we can know whether the power switch has been hit. In addition, the blue light is triggered if the SD does not mount at any point so that we can tell if the data is properly being saved or not. We chose these lights because they are easily viewed from outside of the pod so we know that everything works before we launch without having to plug the mbed into a computer.

Ozone Testing Chamber
An ozone testing chamber, constructed by Dr. Larkin and previous teams, was used to calibrate the integrating sphere. The testing chamber was constructed from a glass tank, PTFE (Teflon) top, UV light ozone generator, two PTFE tubes, two ABS plastic 3D printed funnels, two Insignia NS-PCP8050 computer fans, and an Eco-Sensor A-22 Ozone Sensor. In Figure 10, the tank is pictured with the Eco-Sensor A-22 Ozone Sensor, integration sphere, control cylinder, and circuits inside. The tubes and funnels on the top of the tank lead to the UV ozone generator. Fans directly underneath the funnels ensured sufficient air movement from the UVP 97-0067-01 Stable Ozone Generator into the tank.

Calibration Tests
In both calibration tests, the control cylinder used was not the final cylinder. During pod preparation, the control cylinder used in testing broke, and a new cylinder was constructed to the same specifications of the first.

Test 1: The integrating sphere, control tube, Whitworth Launch Computer, and Eco-Sensors A-22 Ozone Sensor were placed into the tank (Fig. 10). The integrating sphere, control cylinder, and test chamber fans were activated and the UV ozone generator was set at half strength. The ozone content inside the tank was manually recorded every five minutes. After three hours, the ozone content remained at 0.00 ppm. Manual recording was determined to be too time-consuming. We decided to redesign the calibration test due to the slow rate of ozone generation. A second test was conducted to attempt to calibrate our system.

Test 2: The setup was the same as the previous test, but the UV ozone generator was at full strength. The ozone content inside the tank was automatically recorded with a laptop camera approximately every ten seconds. Unfortunately, the laptop camera did not maintain a constant ten second time interval, skipping a second every 7-11 readings. This issue resulted in additional work to analyze the data, but did not affect the results. The test was run over the course of 24 hours. The ozone content in the chamber ranged between 0 ppm and 0.166 ppm. At these levels, no significant relationship between ozone concentration and phototransistor output was observed (See Figure 11). As seen in Figures 12 and 13, at approximately 0.1 ppm, the phototransistor readings rose significantly, however, this spike was also observed in our tube data, so it can be assumed the spike was not caused by the rise in the ozone concentration. The smallest voltage that the bed can detect is 0.8 mV. It can be concluded that from 0 to 0.166 ppm of ozone, the phototransistor voltage change cannot be detected by our circuit.

Durability Testing
After the pod was prepared for launch, it was tested for its ability to withstand physical shaking and impact. The pod, with top secured, was activated and then was shaken and spun vigorously by hand. The pod was also thrown short distances to ensure that our system would withstand the balloon's explosion and landing. The durability testing lasted approximately 20 minutes. Afterwards, the pod was inspected for any warning lights or damage. No damage was found. The pod took data throughout the durability test.

Data and Analysis


The weather balloon that was launched on April 19, 2018 reached an altitude of approximately 105,000 feet. The sphere and cylinder systems were able to collect data continuously through the ascent and descent of the balloon. The raw data collected by the SD card can be found here: Ozone Raw Data. The following texts and figures are based off of the data that was collected during the launch and flight of the weather balloon.

Raw Data


Figure 14 shows red LED and blue LED phototransistor readings versus the relative altitude of the sphere. Altitude was correlated with the time of our system using a beacon that was launched on the weather balloon that measures both altitude and time of flight. This is the raw data of our experiment and represents the data points that were used in further mathematical manipulations that will be discussed and shown in the following text and figures. These manipulations were done in order to compile the data in a format to which a qualitative conclusion could be drawn about our experiment. No quantitative conclusion could be drawn due the unsuccessful calibration of our system.

As we were analyzing the data, the conclusion was drawn that the intensity of the red LED was greater than that of the blue LED. In order to account for this, the ratio of the red LED to blue LED was graphed instead of the blue data outputs being subtracted by red data outputs.

Cylinder Data
The cylinder system served as the control in this experiment. This allows for the conclusion to be drawn that all of the variations seen in the data are due to natural variations rather than a result of light being absorbed by ozone. At the starting altitude of our system, the red LED has a greater magnitude of output than the blue LED. Analyzing this data allowed us to assume that this is the same for our sphere system. The data varied from a range of 0.784 V to 0.790 V.

The cylinder data also suggested that higher readings are to be expected in the sphere's system due to the large increase of output readings that was shown from altitudes 70,000 to 20,000 during the descent of the balloon.

Sphere Data
The data that was collected from the sphere contradicted the original hypothesis that this experiment was addressing. As altitude increased the ratio of the two LED voltage outputs increased to a value that is greater than one. From this we concluded that the red LED phototransistor readings were increasing as a function of altitude which was the opposite of what we predicted. Our hypothesis stated that the red readings should decrease as a function of altitude due to some of the photons that are emitted by the red LED being absorbed by ozone.

After analyzing the data further, we discovered a trend in the blue LED system that was interesting. In order to understand what was going on in our system we made three more plots on Mathematica. In Figure 15, the red sphere data is divided by the red cylinder data and that ratio was plotted versus altitude. The same manipulation was done for the blue LED systems, which can be seen in Figure 16.

Lastly, the final manipulation of data that was analyzed was the ratio of the red ratio over blue ratio plotted against altitude. This plot can be seen in Figure 17. This is where we found interesting trends that the blue LED was outputting that contradicted our hypothesis. From this ratio, we first concluded that as altitude was increasing, the red LED intensity was increasing. After a second analysis was conducted, we also concluded that the blue LED could be decreasing by a greater degree than that of the red LED.

Analysis
The atmosphere is composed of many gases that all absorb different wavelengths of particles. One thing that we did not account for in deciding the wavelength of LEDs to use was the idea that the blue LED that we were using in our system could easily be absorbed by another gas that is in the atmosphere. Based off of Figure 17, we speculated that this was in fact what was happening to our system.

The ratio of the ratios' data points were steadily increasing relative to altitude. In order for this trend to be observed one of two things could have been occurring in our system. First, the red LED could have experienced no absorption due to ozone and could have been increasing over time due to circuiting issues. The second idea is that the blue light could be being absorbed by another gas in the atmosphere at a faster rate than the red light was being absorbed by ozone. The latter of these two hypotheses was further investigated.

After researching on gases that absorb light at approximately 470 nm, we found data that demonstrates that nitrogen dioxide, a gas in the atmosphere, has a peak absorbance right around this range. The plot of this data can be seen in Figure 18.

Based off of the nitrogen dioxide absorption data, a reasonable hypothesis can be made that the blue LED was being absorbed by nitrogen dioxide at a faster rate than the red LED was being absorbed by ozone.

Conclusion
The amount of blue (470 nm) light detected decreased significantly with altitude. The 470 nm light was intended to help control for changes in the circuit function due to environmental factors. However, after analyzing the data, the decrease in phototransistor readings was hypothesized to be attributed to the absorption of photons due to nitrogen dioxide in the atmosphere. We hypothesized this because we found data that supports the idea that the peak absorbance of light by nitrogen dioxide lies around the range of 470 nm. In addition, our control cylinder, which was run on an identical circuit and subjected to the same environmental conditions, varied with altitude (purely due to environmental factors rather than gas concentration) within a very small range (0.784 V to 0.790 V). Therefore, we hypothesized that nitrogen dioxide, which increases with altitude, absorbed our blue (470 nm) light.

Due to the possible influence of the nitrogen dioxide, we were unable to gather meaningful data about the ozone concentration in the atmosphere. This could be solved by using an infrared (>650 nm) LED that would have low absorption for both gases. Although we failed to measure atmospheric ozone, we were able to determine that temperature has a small effect on phototransistor readings, and found information related to the absorption of light due to nitrogen dioxide.

We were successful in creating a mechanically sound sensor and pod. Our data was very consistent and did not have any issues related to the explosion of the balloon during flight nor the force from landing. A group continuing our research would be strongly advised to use the Launch Computers for the circuit because of the stable platform it provides for the experiment and its reliability. Also, the launch computer provides an efficient way to make sure that all systems are working before launch.

Some suggestions for future groups would be to add a second amplifier after the filter in the LED and phototransistor circuit. This would allow the data output to be a greater value and make data analysis and finding trends in the data much easier. Secondly, we would strongly suggest furthering the research about the concentration of nitrogen dioxide in the atmosphere. This could be accomplished by adding a third LED which outputs an infrared wavelength of light. Lastly, we would suggest trying to find a more efficient way to produce ozone in the ozone testing chamber for the calibration of the system. This could be done by purchasing a new dual-lamp ozone generator or by finding a chemical reaction in which ozone is produced as a by-product in large concentrations.