Determination of atmospheric ozone content using a low-cost integrating sphere

Research Group: Jacob Hunter, Matthew Keiper, and Eric Reichel

Launch: Whitworth Fall 2014

''This experiment aimed to map the concentration of ozone in the atmosphere as altitude increased. Ozone has the behavior of absorbing light across the spectrum, increasing absorption as ozone concentration increases. The integrating sphere, made of two semi-spherical cake pans, utilized an orange LED light source and a photo-diode to measure the amount of light that escapes absorption by ozone. As altitude increased and ozone concentration changed, it was assumed that voltage from the photo-diode would also change. Recording this voltage change with an Arduino would allow us to determine ozone levels at the different altitudes. Major design improvements were made from the results of previous Whitworth Near Space experiments including an improved air exchange system and calibration using an airtight ozone chamber. This experiment launched on a weather balloon on October 30th 2014. Unfortunately, an equipment malfunction occurred during launch that prevented us from recording any meaningful data.''



Importance of Ozone Gas in the Atmosphere
The earth’s atmosphere is separated into five layers distinguished by chemical composition, temperature change, density, and other physical characteristics. The lowest layer, known as the troposphere, begins at the earth’s surface and rises up 6 to 20 km depending on vicinity to the equator. This layer harbors all life on earth and almost all weather takes place here. Next, the stratosphere extends above the troposphere up to around 50 km. In this region temperature increases with altitude to a maximum of about 0 degrees Celsius. This temperature increase is due to heat released in the formation process of ozone. Weather balloons normally fly in this layer. Above the stratosphere is the mesosphere (up to 85 km), the thermosphere (up to 600 km), and the exosphere (up to 10,000 km). The ozone in the stratosphere is responsible for the absorption of most of the harmful ultraviolet rays entering the atmosphere. This makes it extremely important to monitor closely.

Ozone’s absorption properties in the atmosphere have been of particular interest to scientists in the last several decades as the rising global temperature becomes increasingly threatening to our environment. A thin layer of the colorless gas in the stratosphere called the ozone layer is continuously monitored with satellite, airborne and ground-based systems by many governmental agencies including the United States NOAA. As the world began to study the effects of harmful human gas emissions on the ozone layer, scientists discovered that ozone concentrations were decreasing and even found a large hole about 20 million km$$^2$$ in the ozone above Antarctica. These findings moved countries in the UN to adopt the Montreal Protocol in an attempt to curb human action from depleting the ozone. The UN Environment Programme released a report on September 10, 2014 showing evidence of the effectiveness of the Montreal Protocol, reporting that the ozone concentration in the stratosphere is on track to recovery. Our experiment aims to determine this ozone concentration in the stratosphere.

Experimental Theory
According to the Beer-Lambert Law, the absorbance of light by a material is given by \begin{equation} A = abc = \log I_0 / I \end{equation} with \(I_0\) the intensity incident on the substance, \(I\) the intensity transmitted through thickness \(b\) of the substance, \(a\) the absorptivity being a material property that depends on the wavelength of light, and \(c\) the concentration of the material. The concentration is given by the barometric formula, ,

$$ c = c_0e^{-\frac{Mgh}{RT}} $$

with $$c$$ the atmospheric concentration at altitude $$h$$, $$c_0$$ the atmospheric concentration at sea level, $$g$$ acceleration due to gravity, $$T$$ temperature, $$R$$ the universal gas constant, and $$M$$ the molar mass of the molecule in question.

Our experimental setup utilizes an LED with a wavelength of 595 nm, as determined from the absorption spectrum of ozone for visible light. To increase the thickness through which the light travels, the light is shown into an integrating sphere with two additional ports for circulation and a port for a photodiode. A baffle was included between the ports for the LED and the photodiode to prevent light from reaching the photodiode without reflecting in the sphere.

Mechanical
The integrating sphere was constructed using two aluminum "Fat Daddio's" hemispherical cake tins. The inside of each hemisphere was painted with five light coats of white spray paint. A 1.05 cm diameter hole was drilled at the top of each hemisphere to allow air to flow, and on hemisphere a quarter inch diameter and a 0.360 inch diamter hole were drilled for the LED and photodiode, respectively. To prevent light from reaching the detector without being reflected in the integrating sphere, a baffle was constructed from sheet aluminum and spray painted in similar fashion as the hemisphere. All elements were mounted in the integrating sphere using epoxy.

The pod used was already constructed in a similar fashion to that described by the 2012 Whitworth Near Space group. To ensure circulation, holes were punched in the pod on two opposite faces. On one of these faces, a fan was bolted into the pod to blow air in; this hole was connected to one air hole of the sphere with a small piece of tubing, and the other air hole in the sphere lined up with the opposite hole in the pod. The sphere and tubing were secured with silicone caulk, and all batteries were secured with zip ties.

Testing Chamber
Two fans to aid circulation of the test chamber were powered in parallel by an LM7805 voltage regulator, which could be powered either by a 9 V power supply. The circuit diagram for the voltage regulator is shown in figure 2.



Experimental
Circuits used included a 595 nm LED, a light detection circuit, and a fan; the LED and light detection circuit were powered in parallel by two 9 V batteries in parallel, while the fan was powered in a separate circuit with an LM7805 voltage regulator and one 9 V battery like the fans in the testing chamber. Light was detected using a Hamamatsu photodiode in a light detector circuit; the signal was then amplified by a non-inverting amplifier to achieve an output signal of roughly 4 V to be read by the master radio transmitter (DAQ); see figure 3. The light detector and the amplifier each utilized a 741 op amp.



The LED was connected to a 180 ohm resistor.

Software
This research did not require any special software to be implemented.

Testing
The purpose of testing our integrating sphere was to make sure that it was properly calibrated to read the amount of ozone ( in parts per million) in air. In order to do this we needed an ozone testing chamber where we would be able to produce and read controlled amounts of concentrated ozone for equipment verification and calibration purposes. With the end goal of designing a low cost integrating sphere, our test chamber must also be low cost. A major concern in the design of the test chamber was to make it out of substances which would not be affected by ozone's corrosive effects. Because of this it was best to avoid the use of plastic or polymer based materials as they show the most wear from exposure to ozone. The testing chamber was constructed from a glass fish tank, PVC piping, scrap wood, and electrical components. Another challenge focused on making an air tight lid. The final design of the lid was made of wood, designed to fit snugly over the edges of the fish tank. Weather stripping lined the lip to ensure an air tight seal and preventing hazardous amounts of ozone from escaping. Basic design aspects of the ozone test chamber included mounted air intake and air outtake mounted on the lid powered by five volt CPU fans, and a gas release valve to ensure safe disposal of the ozone. Final aspects of the lid were basic access ports to supply power to anything in the test chamber and collect data. The process of introducing ozone inside of the test chamber was done with the use of an ozone generator. This device creates ozone though the implementation of an ultraviolet lamp, which when an oxogen molecule interacts with a photon, subsequently leading to ozone as show in the equation

$$ O_2 + \gamma \to 2O $$

followed by

$$ O + O_2 \to O_3 $$.

Because the air had to be exposed, our mounted fans and tubing allowed a steady flow of air through the ozone maker, steadily increasing the concentration of the ozone within the chamber over time.

With this test chamber our procedure was to set our entire apparatus (integrating sphere within the launch pod) within the chamber, along with an industrial use ozone detector. With our integrating sphere and Ozone detector we were then to calibrate our integrating sphere with an calibration plot.

Since we had no industrial ozone detector readings to compare our voltages from our integrating sphere, our thought was to make sure our integrating sphere’s circuitry was constructed properly so that voltage readings would drop at increasing ozone levels, rise at lowering ozone levels, and stay constant at constant ozone levels. After launch it was our plan to calibrate our voltage readings to an industrial use ozone detector once one was acquired. However since our pod was lost on the flight we were unable to correlate certain voltage readings from our integrating sphere to ozone levels. Recommendations for future testing include ample time for circuit construction. This aspect of building the integrating sphere is the most tedious and difficult to build, which will need many test runs to evaluate if properly working.

Data and Analysis
Our experiment launched on October 30th, 2014. The data we received returned a constant 5V signal beginning sometime during the launch process. We assume from this that there was a malfunction in our circuit as the balloon was being released. In addition, our pod broke away from the balloon mid flight and landed in an unknown location. Because of this, we were unable to diagnose what exactly went wrong to cause the sensor error. We postulate that probably a ground wire in the breadboard came lose, pushing the output to the 5V rail voltage. This event likely happened because of circuitry that was not fully secure. After building our working circuit in a breadboard, we did not have the time to rebuild the design in a more secure circuit board. If we were able to do this experiment again, we would try to finish our circuit design earlier to allow time to rebuild the design in a secure circuit board. We would also consider using a company that builds printed circuit boards for us to greatly minimize the probability of an electronic error from any jostling that comes along with flight on a weather balloon.

If the pod had returned a data set, each reading would have been between 0V and 5V, changing with altitude. To understand what ozone density each voltage reading corresponded to, we planned on calibrating our integrating sphere in the testing chamber we built. Using the testing chamber setup detailed in the previous section, we were to create a changing known density of ozone with an industrial ozone sensor and the UV ozone lamp and map these levels to corresponding integrating sphere voltage outputs. With this information, any voltage returned from the flight would be translated to ozone density. Unfortunately, one of the instruments in the testing chamber was malfunctioning and we were unable to get a known ozone density reading during calibration. We confirmed it was the industrial ozone sensor that was malfunctioning by creating ozone with the proven method of sparking wires in a jar repeatedly, and the sensor still returning no reading.