Ozone detection using visible light and a custom integrating sphere

__MATHJAX_DOLLAR__ Research Group: Bobby Aldridge, Peter Beck, Peter Landgren, and Jacob McCallum

Launch: Whitworth Fall 2010

The ozone concentration in the atmosphere is essential in understanding how the Earth deflects and maintains energy from light. The main purpose of ozone is to absorb harmful ultraviolet light from the sun so that it does not reach the Earth in potentially harmful amounts. In fact, the ground level concentration of ozone rarely exceeds 0.04 parts-per-million (ppm) where it reaches approximately 8ppm at a height of approximately 80,000 feet. This higher concentration prevents dangerous sunlight radiation from reaching the lower atmosphere by absorbing wavelengths of light below 320 nm.

The purpose of this balloon launch is to determine the ozone concentration as a function of height. In this experiment however, the absorption detection will take place at the wavelength of red light, ranging from 620-750 nm. This particular wavelength was decided on because ozone has a second absorption peak around 610 nm. Hereby a cheap red light could be used instead of an expensive ultraviolet light that emits 250 nm (primary peak of ozone absorption).

One problem that arises is that barely any red light will be absorbed by $\mathrm{O}_3$ over such a short path length. As such, a strategy was devised to increase the path length in order to provide more time over which the $\mathrm{O}_3$ can interact with the red light, providing a noticeable absorption detection decrease. The solution to the problem was an integrating sphere that allowed the light to bounce around, in effect lengthening the path of the light. A simple ozone detector was also placed in the sphere to be used as a reference.

Background
The overall goal of our weather balloon was to measure ozone concentration as a function of altitude. We expect to see ozone concentration increase as our balloon moves through higher altitudes of the atmosphere, and conform to known atmospheric trends in ozone concentrations with respect to altitude and atmospheric layers.

The Earth's atmosphere is composed of 5 principle layers: the troposphere, stratosphere, mesosphere and the thermosphere. The layer of atmosphere closest to the Earth is the troposphere; it spans 23,000 ft to 56,000 ft from the surface of the Earth. The troposphere is also the densest part of the atmosphere and makes up 80 percent of the mass of the atmosphere. Directly above the troposphere is the stratosphere, which extends from where the troposphere ends to about 170,000 ft above the Earth's surface. Within the stratosphere lies a large layer of ozone. This layer is responsible for scattering and absorbing a large portion of the solar ultra violet radiation entering the earth's atmosphere. We also expect to see a significant increase in the amount of ozone as our balloon inters the ozone layer in the stratosphere. Our weather balloon should only rise 96,000 ft so no data was collected from any other layers of atmosphere, making these layers less relevant to our project.

The solar ultraviolet radiation that enters the earth's atmosphere is very harmful to humans and other forms of life. So far ultraviolet radiation has been known to break bonds in DNA, cause eye damage and has been linked as a cause of skin cancer. The ozone layer plays a significant role in defending life forms on Earth from these dangers. According to NASA, the "earth's ozone molecules absorb 97-99% of the sun's high frequency Ultraviolet light, light with wavelengths between 150 and 300 nm". Ozone absorbs these frequencies of radiation because of the electron transitions within the ozone molecules. Ozone also absorbs light at 600 nm, this allowed us to construct an ozone detector using light as a source within this visible wavelength. We expect our concentration to look like the graph of concentration in figure 2.

We used two methods of ozone detection in this experiment. The first method uses a MQ-131 gas sensor which responds to the presence of ozone by changing the resistance within the device. This allowed us to construct a circuit with this ozone detector where we measured voltage to observe the change in ozone concentration.

Our second method of ozone detection utilized ozone's absorption of light to measure concentrations of ozone. In building this detector we used a light source in the 600 nm range and photodiode that would detect the intensity of light; resulting in us being able to measure the amount of light absorbed as a function of voltage. We placed the photodiode and light sensor inside an integrating sphere. The function of the integrating sphere is to increase the probability of photon absorption due to ozone before the photons are absorbed by the photodiode. The integrating sphere achieves this by reflecting the photons off the wall of the sphere in order to extend the path length. By traveling through a greater amount of space each photon has a greater chance of interaction with ozone molecules. This integrating sphere is instrumental in achieving measurable levels of ozone absorption.

Mechanical


Materials:


 * Wiffle balls (qty: 2)
 * Rubber hose (qty: 1)
 * Photodiode (from detection circuit) (qty: 1)
 * 620 nm LED (from LED circuit) (qty: 1)
 * Foil
 * Electrical Tape

Construction:

The wiffle balls we bought only have one side that is solid. We bought two wiffle balls and cut them in half. We then hot glued the halves together to create our integrating sphere. We then cut two holes, just big enough for the rubber hoses to fit in, straight across from each other. We cut the rubber hose into two pieces and hot glued each piece into the drilled holes. We then cut two holes perpendicular to each other, just big enough for the electronic devices to fit in, and mounted the electronics into the holes. In order to prevent short circuiting, we wrapped the leads of the electronics in electrical tape. Then, the integrating sphere was wrapped in foil to insure that light was not escaping or coming into the sphere.

Electrical






Components List:


 * LM7805CT Voltage Regulator (qty: 1)
 * Ozone detectors (qty: 2)[[media:MQ-131.pdf|MQ-131 Datasheet]]
 * 620 nm LED (qty: 1)[[media:BVU-5M1QN1.pdf|BVU-5M1QN1 Datasheet]]
 * Photodiode (qty: 1)
 * 741 Op-Amps (qty: 3) or 353 Dual Op-Amps (qty: 2)
 * Resistor and capacitor values as noted in schematics.

Software
N/A: No special software was necessary for this project.

Testing
to test our gas sensors we needed ozone in the lab. Ozone is created when a current is run through the air. we used a spark generator, put the leads into a jar, covered the jar, and allowed the generator to spark for a given amount of time. We then took the jar and covered the gas sensors with it. Our purpose was to see whether or not the output voltage of the sensors would respond as we expected. We expected the voltage to drop as we gave the sensor more ozone. This trend was observed and the results are shown in the table below.

From our validation results, we an see that our gas sensors do not agree. This is fine because we can compare the two sensors when we actually collect atmospheric data. The difference between the two sensors is the load resistance. Gas A has a slower response time than Gas B. We want to see which one works better. The one that agrees most with theory will be the sensor we use to present results.

__MATHJAX_DOLLAR__ In order to calculate the concentration, we interpolated the graph of ozone concentration given on the gas sensor data sheet. Since the graph is a log-log plot, we expect our concentration to have the form, $\frac{R_0}{R_S} = b{C_{\mathrm{O}_3}}^n$, where $R_0$ is the baseline resistance of the sensor in the atmosphere, $b$ is a constant to be determined, $C_{\mathrm{O}_3}$ is the concentration of ozone, and $n$ is the power the concentration is raised to. The sensor acts as voltage divider with a load resistor, so it was easy to measure the resistance using the equation $R_S = \frac{(5-V_\mathrm{out})R_L}{V_\mathrm{out}}$. We took two points off the graph and solved for $b$ and $n$. We found that $b = 16.1$ and $n = -0.865$. We then solved for the concentration and plugged in our values to obtain a numerical result.

To test our sensor, we again created ozone in the lab. Instead of creating it in a jar, we put the leads of the spark generator into the integrating sphere and let the spark generator spark for a given amount of time. This time, we were checking to see if the voltage detected by the light sensor would drop. We were unable to actually calculate a concentration because we do not have a calibration curve. We can only show the detection of ozone. We were able to successfully observe a drop in voltage as the ozone concentration increased within the integrating sphere. Gas sensor calibration during generation of ozone by means of electrical sparks.

Data and Analysis
Our data from both the ozone sensor and the light detector correlate well with each other, sowing voltage drops and thus increased ozone concentration at the same altitudes. The light detector does show more small scale variation in the ozone concentration, but this could be an artifact of its high error. Our gas sensor data is from gas sensor A, since there was a malfunction with gas sensor B and it did not return any readings.

Our data also correlates well with the known altitudes of the atmospheric layers. Our measured ozone sensor concentrations begins a steady increase starting around 10,000 meters, which is close to the start of the troposphere at 7,000 meters. Also, our measured ozone sensor concentration levels off at approximately 20,000 meters, which correlates well to the 17,000 meter start of the stratosphere. The gas detector also has some lag time, which makes sense given our data. Our light sensors show even better correlation with the atmospheric layer locations, but with less defined transitions.

Due to the manufacturer's data sheet specifications, we were able to calculate the ozone concentration for the gas sensor. This process is described in the sensor testing section. These values are about thirty times lower than expected according to established sources. However, the shape of the peak conforms very well to the established sources and increases and decreases at the appropriate altitudes, showing the data to be quantitatively accurate.

Some of our error in this experiment may have been due to the response time of the gas sensor. It took time to equilibrate and stabilize, which could introduce some error with respect to altitude. The venting of gas into the pod, where the readings actually took place, could also introduce some lag time. A third source of possible error is the temperature. We did our best to stabilize this using hand warmers, but it could have had some effect at high altitudes. In regards to the light sensor, the high amount of variation between points (average standard deviation of 0.09V for a small linear segment) could also be due to some light leaking in from outside as the pod rotated.

For the light sensor, the voltage drop we observed in the experiment was much greater than the drop we observed in the lab. Unfortunately, we were unable to obtain a calibration curve for this drop because of the uncertainties of pumping ozone into the sphere, but this still shows that there was a considerable concentration of ozone at higher altitudes.

Our measured ozone data confirms our hypothesis that ozone concentration increases with altitude. Also, our concentration data conforms in principle to known atmospheric ozone concentrations and atmospheric layers.