Ozone and UV Light

__MATHJAX_DOLLAR__ Research Group: Dan Belet, Dusty Caseria, Brian Harms, Xander Knight, Sean McGuire, and Kellen Oetgen

Launch: Whitworth Fall 2008 In this experiment, we investigated the relationship between ultraviolet radiation intensity and ozone concentration within the troposphere and stratosphere. Temperature data was also taken to further examine the relationship between ultraviolet levels and ozone concentration. Figure 1 ????below???? show the experimental data.

???Fig 1 approximate ozone concentration versus altitude???

Substantial amounts of ozone were encountered around 20 km. This is slightly above where we would expect the ozone layer to begin at around 13-15 km. This inconsistency is probably due to the primitive calibration methods used. The ultraviolet intensity plot shows a fairly linear response up to 13-15 km, at which point there is a drop in ultraviolet intensity. This roughly corresponds to the location of the tropopause where ozone concentration levels begin to rise. Again, around 20 km, the ultraviolet intensity significantly levels off which likely corresponds to UV absorption in the formation of ozone. However, due to the high level of noise in the ultraviolet data, it was hard to accurately determine where significant changes occurred in the UV levels.

???Fig 2 external temperature versus altitude???

The general trend of the external temperature data shown in figure 2 closely followed tabulated trends. Again, around 20 km, the temperature began to rise suggesting an increase in ozone concentration levels. Before this height around 13-15 km the temperature plateaus. This correctly corresponds with what is known about the location of the tropopause with respect to the ozone layer. In particular, the tropopause is known to begin around 11 km. Below this height temperature steadily decreases with altitude due to the expansion of the atmospheric gases. The exact levels of ultraviolet intensity, ozone concentration, and temperature could not be determined from our data due to limitations in the ability to calibrate our circuits. The general trends, however, were examined and appeared to correspond with tabulated trends.

Background
Earth's atmosphere is comprised of a variety of gases held within approximately 10,000 kilometers of the Earth's surface by gravity. The atmosphere is striated into five layers consisting of differing characteristics and combinations of particle and gases. The stratosphere is the layer ranging from approximately 20 kilometers to 50 kilometers. The stratosphere is particularly characterized by the ozone layer, a band of atmosphere ranging from 15 to 35 kilometers (49,000-115,000 ft) containing relatively high concentrations of the triatomic molecule $\mathrm{O}_3$. The ozone layer contains approximately 90% of the ozone found throughout the Earth's atmosphere and performs a variety of integral roles to life on Earth (see Altitude as a function of ozone concentration). The ozone layer filters out shorter length photons called Ultraviolet light that, in large doses, are harmful to most forms of life on Earth. This particular length of photon (100-400 nanometers) interacts with Oxygen ($\mathrm{O}_2$) that then produces the triatomic form of oxygen called ozone. The concentration of ozone within the stratosphere varies from approximately 2 to 8 parts per million. As a general trend, temperature decreases with altitude. However, heat is emitted as a result of oxygen absorbing photons within the ozone layer. Therefore, the stratosphere produces a temperature inversion where temperature increases as altitude increases. Absorption of UV light is greatest at the upper reaches of the ozone layer thus resulting in the inverse relationship between temperature and altitude. Although the ozone layer is present within the stratosphere throughout the globe, the thickness and concentration varies geographically and seasonally. As mentioned previously, the effects of UV light on life forms can be detrimental. Ultraviolet light has been shown to break bonds in DNA, has been linked to cancer, causes eye damage, and poses a serious threat to marine life such as plankton. Therefore, the concentration of ozone has a great effect on humans and is valuable to know for one particular area. The concentration of ozone is largely known throughout the world and is studied ad nausuem in locations where the ozone is partially depleted such as the Antarctic and Australian regions. However, the exact ozone concentration and UV absorption is not known for specific locations. This experiment measured the ozone layer as a function of UV absorption over central and eastern Washington. Temperature was also measured to provide an in-depth analysis of the ozone layer over this particular region. Three ultraviolet light sensors were placed on the outside edge of the experimental pod to measure light from three different angles. This configuration was chosen over a single sensor in order to prevent the sensor from facing away from the sun and recording no data. Three sensors placed radially equidistant apart assures that at least one sensor is recording accurate data at any one time despite the orientation of the pod. A single ozone sensor was placed on the outside of the pod measuring the concentration of ozone. A temperature sensor was placed on the inside of the pod measuring the temperature the circuits were subjected to, while an adjacent temperature sensor on the command pod measured the temperature outside. Hypothesis: Based on known levels of ozone, UV absorption, and temperature it is expected that the concentration of ozone remain steady from approximately 0 to 15 kilometers (49,000 ft). The ozone concentration should begin to increase around 15 kilometers (50,000 ft) and peak at approximately 25 kilometers (82,000 ft) before slowly decreasing until the balloon ruptures at roughly 30 kilometers (100,000 ft). The intensity of UV light is expected to begin at a low value at a low altitude and increase five percent every thousand feet. The intensity is expected to increase at a faster rated around the ozone layer. The absorption of UV light is expected to remain low until the ozone layer is reached at approximately 20 kilometers. The absorption is then expected to decrease as a function of altitude until the ozone layer is reached at approximately 15 kilometers. Temperature is expected to increase within the ozone layer as a function of altitude due to the increased absorption of UV light.

Mechanical
Assembly notes: (transcribed from raw notes)

Unique features of our pod:

Electrical:

See the circuit schematics in the electrical section for the diagrams of the five sensor circuits (one temperature, one ozone, and three UV circuits). We powered all five sensor circuits using two 9V Ultralife lithium batteries connected in parallel. We determined that we needed two batteries rather than just one because the current draw of all five circuits together was ~150 mA (determined using a multimeter). Since the batteries are rated at 1.2 A-h at only 9 mA, we determined that we would need two 9V batteries to last the entire flight (which we estimated to be 2-3 hours). The batteries were then connected to a 5V LM7805 voltage regulator. Following this, the ozone and the three UV circuits were connected to the common ground off of the 5V voltage regulator and powered by the 5V lead of the regulator. The temperature circuit was the only circuit that was powered directly by the 9V (bypassing the voltage regulator). Each of the two batteries were connected into the circuit using "quick disconnects" to allow for easy access to the batteries. (Note: during all of the trials, regular 9V alkalines were used and then during the actual flight, we replaced these with lithium batteries). The temperature sensor op-amp was powered first directly off of the 9V power source; then the lead from the positive power source on the temperature circuit was connected to the positive power input on the UV circuit op-amp, which was then connected to the positive power input on the ozone circuit op-amp (see voltage regulator schematic).

The signal outputs of all five circuits were subsequently connected to the sensor pod interface according to the diagram below: ???sensor pod interface connector legend???

Structural:

See the ???attached??? AutoCad drawings for the general layout of our pod. The unique structural features of our pod mostly included the positioning of our five sensors. We made three holes in the sides of the pod positioned so that two of the sensors were located at adjacent corners and the third sensor was on the opposite side, essentially forming a triangle. For the ozone senor, we made a hole in the bottom of the pod and inserted the ozone detector. For the temperature detector we simply left inside the pod. All of the circuit perf boards and the two batteries were strapped to the insides of the pod walls using zip ties.

Test day:

We successfully were able to detect a signal for all five of our detectors when we connected the leads to the command pod according to ???the configuration on the previous page???. Then we secured the circuits and batteries inside the pod using zip-ties to strap them to the insides of the pod. For the three UV sensors, we positioned them in the following way: ???drawing from notes???. The $\mathrm{O}_3$ detector was positioned on the bottom center of the pod. The temp detector was positioned in the same hole as one of the UV detectors. We zip-tied the sensor interface board to the inside of the lid of the pod. We were using two 9V batteries in parallel before taking the voltage down to 5V because the current draw of the circuit (all five sensors) was approximately 150 mA. Since the batteris are 1.2 A-h at 9 mA, we need two 9V to last the entire voyage. When we did a test roll of the pod, the UV detectors and $\mathrm{O}_3$ detector were not working. When we checked our circuits, we found that the $\mathrm{O}_3$ interface wire had been disconnected and one of the wires on the $\mathrm{O}_3$ bread board had been broken off. Once we re-soldered everything and checked our circuit, all five of the detectors worked well. We decided not to re-roll the pod because we only really need the data for the ascent and not the descent.

Electrical
 [[media:MQ-131.pdf|MQ-131 Datasheet]] [[media:LM335Z.pdf|LM335Z Datasheet]] [[media:TW30SX.pdf|TW30SX Datasheet]] [[media:TLC2272.pdf|TLC2272 Datasheet]]

Testing
Temperature Sensor Calibration (raw data): $\mathrm{V_{source}}$ = 8.96 V

Cold (ice water) $\mathrm{T_{official}}$ = 1.2 degrees Celcius $\mathrm{V_{measured}}$ = 2.815

Hot (boiling water) $\mathrm{T_{official}}$ = 98.1 degrees Celcius $\mathrm{V_{measured}}$ = 3.980

Room Temp $\mathrm{T_{official}}$ = 21.5 degrees Celcius $\mathrm{V_{measured}}$ = 2.967

LM335Z Temperature Sensor Validation and Calibration: After constructing the temperature circuit (see Electrical section), we needed to validate that the sensor actually was able to detect temperature and convert that into a voltage. We used a multimeter to check the voltage across the Data line and the ground line at room temperature. The resulting voltage at room temp (21.5 degrees C) was 2.967 V. The temperature of the room was recorded using a thermocouple. According to sensor documentation (see Mechanical section for data sheet), the voltage should be at 2.982 V +/- 0.05 V. Thus, the sensor was reading reasonable voltages.

The next step was to calibrate the sensor using various external temperatures. We used three situations: an ice water bath, room temperature, and boiling water. The ice water was measured at 1.2 degrees C by the thermocouple. We dipped the end of the sensor directly into the water, and the resulting voltage reading was 2.815 V. For the hot water bath, the water was raised to the temperature of 98.1 degrees C using a hot plate. This gave a sensor reading of 3.980 V. The voltage at room temp is given above.



As can be seen, the calibration gave a fairly linear relation between voltage and temperature.

UV Circuit Test Process: After constructing a working circuit, we took the circuit outside on a sunny day to test the $\mathrm{V_{out}}$ values that we were getting at Whitworth, which is about 1900 feet in elevation. After some research online, we found that the amount of UV radiation increases by about 5% per 1,000 feet increase in altitude. The rail limit that we did not want our $\mathrm{V_{out}}$ value to hit was 5 volts. We wanted to be extra careful not to hit this value because our data beyond that point would be useless if we were to hit the rail. We knew that the increase in UV radiation would therefor be 500%. When testing our circuit, we allowed for a 6% increase in radiation per 1,000 feet - so 6,000% overall - to be careful not to hit the rail limit. This meant that we wanted our $\mathrm{V_{out}}$ value when we tested our circuit outside the science building to be as close to 0.833 V, which is 5/6 of a volt, without exceeding it. We recorded $\mathrm{V_{out}}$ readings for the circuit when it was exposed to direct sunlight. (We also recorded readings for when the sensor was covered in order to verify that the circuit was responding only to sunlight and working properly. As long as these values were very small, the circuit was working properly). We changed to values of $\mathrm{R_{in}}$ and $\mathrm{R_{f}}$ in the non-inverting amplifier portion of the circuit and recorded the output voltage until we got very close to what we wanted. $\mathrm{R_{in}}$ = 27k &Omega; and $\mathrm{R_{f}}$ = 10k &Omega; gave us an output voltage of 0.793 V, which was as close to 0.833 V as we were able to get without exceeding it, so we used those resistor values in our final circuit.

Validation and Calibration of $\mathrm{{TiO}_2}$ UV Photodiodes: Based on the limitations of available instruments, we were not able to make a quantitative calibration of our three photodiodes. In particular, we did not have at our disposal a UV light source in the particular wavelength range (100-400 nm) that provided accurate photon fluxes in order to gauge what voltage resulted from a given intensity of UV light. Therefore, we were more interested in the relative change in voltage for a given change in UV intensity. In order to test the linearity of our response, we tried two different methods. The first was to actually see how all three UV sensors responded to sunlight at the elevation of Whitworth. We took the circuit outside and positioned the three sensors so that they were directly in the sunlight. The we used optical density filters to decrease the amount of light transmitted into the sensors and a multimeter to record the voltages. The table below lists the voltage readings for each of the sensors (red, green, and black) for the various OD filters.



As can be seen in the above graph, the relationship between percent transmission and voltage signal was not very linear. This was most likely due to the variability of the sunlight conditions in-between OD filter readings. The sky was somewhat overcast with cloud cover and so it was hard to ensure that the same amount of light was incident on the sensors for each data reading.

In order to remove the variability of the sunlight, we attempted to use an Hg UV lamp (Whitworth #001257). We followed the same procedure as we did outside, except this time we ensured that the sensors were held at a fixed distance away form the lamp. In particular, the sensors were mounted 2 inches form the lamp. The following data was recorded:

???graph of voltage versus percent transmission for Hg lamp???

As can be seen in the graph, the linearity is greatly improved, especially at higher transmission levels. The linearity decreases at lower transmission levels, though. This is mostly due to the fact that the Hg UV lamp was putting out very low levels of UV radiation, mostly as a safety precaution. That is why we were not reaching appreciable levels of voltage. This is of some concern, but for our purposes, we are interested in UV altitudes at higher altitudes were the signal will be much higher. Thus, this linearity will suffice for our purposes.

The question arises as to how we will analyze our data once we launch the balloon, since we will be recording three voltages from three different sensors. The placement of the three sensors was logically chosen so that the maximum amount of light can be detected at all angles. Since the pod will be twirling around rapidly during the ascent and descent, the sensors will be variably exposed to sunlight throughout the flight. In order to correct for this, we will most-likely average all three sensor readings for every 100 ft of ascent or descent. This should provide a reasonable approximation of the relative levels of UV over the 100,000 ft ascent.

Data and Analysis
The results from the ozone sensor are displayed in ???fig 1???. Using the tabulated relationship between resistance and ozone concentration from the ozone datasheet ???fig 2???, we were able to make rough approximations for the concentration of ozone. As can be seen, the detector did not pick up any significant levels of ozone until around 20 km. According to our hypothesis, we expected the ozone signal to at least begin at around 15 km and then significantly increase once it reached 20 km, in accordance with the known ozone concentrations displayed in ???fig 3???. A possible explanation will be discussed below. The approximate levels of ozone are in agreement with the known concentrations in the stratosphere (approximately 2 to 8 ppm). However, once the sensor reaches levels of about 2 ppm (the ozone concentration at which the sensor is rated) at about 23 km, the signal became digitized, most-likely due to the high levels of ozone saturating the detector. The data from the three UV sensors are shown in ???fig 4???. In this plot, voltages from all three photodiodes were first averaged together for a given altitude. This data was then averaged over every thirty seconds of flight to remove excess noise. However, even after averaging, much noise still remained. The placement of the three photodiodes as shown in ???section 3??? was chosen so that the maximum amount of light could be detected at all angles. However, since the pod was twirling around rapidly during the ascent and descent, the photodiodes were variably exposed to sunlight throughout the flight. In many cases this probably led to one diode having a high voltage reading, while the other two had significantly lower values. The highest individual reading was about three volts. However, due to the averaging process, the high readings were likely combined with two much lower values causing the average UV intensity to be much lower. Also, the random variations in the plotted signal could be due in part to the presence of clouds which, over a relatively low altitude, could sharply alter the UV intensity. Further interpretation of the UV data will follow. The data for the sensor measuring the external temperature during the flight is displayed in ???fig 5???. The external temperature did not reach values predicted in ???fig 6???. The tabulated data show the temperature dropping all the way to about 210 K. In our experiment, the external temperature from the command pod only indicated that the temperature dropped to a minimum of about 260 K. Because we did not assemble the command pod's temperature sensor nor did we calibrate it, we do not necessarily know why the data deviated from the expected data to such an extent. Most-likely though, the tabulated data in ???fig 6??? were only the average global atmospheric temperatures and may not have been an accurate depiction of the atmospheric temperature in this region. Nevertheless, the overall trend in temperature changes for our data matches quite nicely with the tabulated data. As can be seen, the temperature steadily decreased from about 12 to 13 km. Then, for about 7 km, the temperature remained fairly constant, at about 260 K. Once an altitude of about 20 km was reached the temperature began to slowly increase for the remaining ascent of the balloon. An explanation of these results will follow. To compare the external temperature with that of the pod, the data for the internal temperature within the pod is shown in ???fig 7???. In comparison with the data from the external temperature sensor, then internal temperature sensor showed very little difference in readings. Apparently, the insulation of our pod failed to retain virtually any heat, assuming that the internal and external temperature sensors were both performing correctly. This could have affected the performance of our circuitry, as discussed below.