Sohrab Andaz Lab Partner: Jon Daley

February 27, 2012 PHY 445

Abstract

Carbon-11 is one of the longest lived radioisotopes of this ubiquitous element (the longest is obviously Carbon-14). This lab will attempt to measure its half-life by producing the substance from Boron-11 using a 6 MeV proton beam from Tandem Van de Graaff accelerator. After instrument calibration and statistical fitting the half-life was found to be 20.150.08 minutes. Introduction

Since Rutherford used a beam of alpha particles to probe the fundamental truths of the nuclear structure it has been quite clear that accelerated particles are crucial to the field of experimental physics. This revelation inspired an accelerator arms race, where labs and academic institutions claimed their experimental significance by quoting the size or power of their respective accelerators. Stony Brook University got in on the game and built its own Tandem Van de Graaff accelerator that was used to do research until 2007. Unfortunately now the great machine is used for exclusively educational purposes. And so we will make use of it in this lab. In this experiment, we will try to measure the half-life of Carbon-11. We will create the isotope by firing a 6MeV proton beam, created by a Van de Graaff accelerator and a General Ionex 860A Inverted Sputter negative ion source, into Boron-11 target, which undergoes a proton in neutron out (commonly abbreviated 11B(p,n)11C) nuclear reaction forming the desired element. The purpose of this experiment is to verify the established measurements of Carbon-11s half-life and explain the operation of the necessary equipment and methodology. Materials and Method

1. Ion Source to Van de Graaff As briefly mentioned previously the ion source is a General Ionex 860A Inverted Sputter negative ion source. In a small vacuum chamber, 1 gram of Cs is baked with a 40 Watt heater which sublimes Cs off the solid. The temperature of the boiler is managed and augmented by changing current through the boiler power supply. The Cs vapor makes it way to a feed tube which steers the Cs gas toward a 1200oC tantalum coiled wire. The work function of Tantalum is 4.25eV, while the ionization potential of the Cs gas is 3.89eV. So when the Cs encounters the superhot coil, it is thermodynamically favorable for it to donate an electron to the Tantalum and form Cs+. Once the cesium gas deposits its electron, it is able to feel the -3 to -5kV potential established by a cathode power supply, which accelerates the positive ions toward a titanium hydride sample capsule. The accelerated Cs+ ions burrow into

the pellet and penetrate through several atomic layers. The impact causes the metal to sputter Hions which are pulled from the ion source and toward the beam pipe by a 15kV potential relative to the sample. The table that holds the General Ionex 860A Inverted Sputter negative ion source is held at -0.4MV relative to ground and gives the H- ions the necessary nudge toward the Tandem Van de Graaff accelerator. Along the beam pipe the ions encounter (in order) an Einzel lens, an inflection magnet, which naturally focuses in x and mass selects, an accelerating gap, which always net focuses, and electrostatic quadrupole triple lens, which unlike a double lens, has no astigmatism. To help tune the beam after the electrostatic triple lens, a low energy Faraday cup can be dropped into the beam path to measure the current. The beam then enters the Tandem Van de Graaff accelerator. 2. Acceleration and Targeting The ultimate purpose of the Tandem Van de Graaff is to take our 0.4MeV beam of Hions and produce a 6MeV beam of H+. It is clear from our object that we will need to strip electrons from the ions in our beam, but also to produce an incredible voltage using Van de Graaff generator to accelerate our beam. But because of our circumstances we can cheat a little, and ingeniously use whatever voltage we produce to accelerate our beam twice. If we hold the center of the Tandem Van de Graaff at some positive voltage, when we also strip the H-s to produce H+ at the center, the beam will accelerate toward the center, and then away from it, picking up twice the energy of the potential. This is the general scheme; but two questions demand that we provide an answer: 1) how is the voltage of the Tandem Van De Graaff produced; and 2) how are the H- ions stripped. Charge is delivered to an electrode at the center of the Van de Graaff by a 12 meter long aluminum, stainless steel, and plastic chain resembling a rope ladder called the Laddertron that revolves at 12m/s and can deliver a total current of 250A. While sections of the Laddertron chain are in contact with ground at the mouth of the Tandem Van de Graaff, they are exposed to a highly negative voltage source (50kV) and inductively charged. Then the links are moved toward the belly of the accelerator, where the charge is forced onto the center electrode by a high voltage. To try and be economical, on the way back negative charge is inductively forced from the center electrode by a high voltage, and is carried to the grounded electrode where it is pressed upon it by a highly negative voltage. To prevent over or undercharging and accurately maintain the correct voltage at the center of the Van de Graaff there are two feedback mechanisms. One obvious process to preserve the proper voltage at the center of the Van de Graaff is by correcting the speed of the Laddertron using a measured value of the voltage and comparing it with the desired voltage. This is done via the generating voltmeter circuit. But method is limited by the large time required to get accurate information because of the incredible resistance of the Laddertron, and can only provide us with a coarse control of the voltage. For a fine control we must use the corona circuit, a plaque of sharp needles placed atop the beam pipe in the Van de Graaf that is moved close enough to the high voltage electrode to pull approximately 20A away from the electrode. The current from the corona flows to ground via a vaccum triode which regulates how close the needles gets to the electrode regulating how much charge is pulled off, managing the terminal voltage.

Now we turn our attention to the method by which the H- ions are stripped of their electrons at the center of the Van de Graaff: the beam hits a thin carbon foil, and the resulting collisions produce a myriad of elementary particles. Obviously only positively charged ions will make it out of the beam pipe with 6MeVs of energy and the multifaceted beam will later be mass selected by a analyzing magnet to ensure that only protons make it to the target. On the way to the target room, the energized beam passes through a magnetic quadrupole doublet, which has an astigmatism, a high energy steerer, the aforementioned analyzing magnet, and another magnetic quadrupole doublet, which corrects the astigmatism. The high energy beam then hits a fine Boron-11 powder, encased in a tantalum foil, held in a tantalum frame, forming radioactive Carbon-11. 3. Detection of Coincidences Once we have created our source we place our source between two Bicron NaI(TI) 2 inch crystal gamma ray detectors with photomultiplier tubes and bases. Carbon-11 decays by positron emission. The positron annihilates when it comes into contact with an electron, and decays into two 511keV photons. So by orienting the NaI detectors at 180o we intend to measure positron decay by looking for the detection of two photons at the same time: a coincidence. When gamma rays hit the NaI solid, the crystal scintillates via Compton scattering. The light emitted from these detectors is forced into a photomultiplier tube, which uses the photoelectric effect to produce a highly variable current. To comprehend this signal, we need to feed the signal from the Bicron NaI(TI) 2 inch crystal gamma ray detectors into a Phillips Scientific 710 Octal Discriminator, which outputs a nice square when only when an input signal crosses a threshold voltage. The signals from both NaI gamma ray detectors both run through the octal discriminator, and then into a Lecroy 364 AK 4 fold Logic Gate. The output of this 4 fold logic gate, when set to AND will tell us when we have a coincidence, because both crystals need to detect, at the same time, a coincidence for the circuit to output. The output of the whole circuit is fed into an Ortec 871 Timer and Counter, which records the number of coincidences we have. 4. Method Before any radioactive source was made, the detection setup was carefully erected, inspected, and calibrated. Values for the respective voltages of the power source that supply the photomultiplier tubes with the necessary potential need to be set along with threshold of the octal discriminator and the width of the signal it outputs. These values should be selected so as to try and minimize unintentional coincidences (accidentals), but still preserve the integrity of the experimental setup. In this experiment, the voltages of the high voltage feeding into the photomultiplier tubes were set to 1180V, while the threshold was set as low as possible, and the width of the output signal was set to 50ns. To aid in the calibration of the setup we used a Tektronix TDS 2024B oscilloscope. To choose our values, coincidence count rate v. width, coincidence count rate v. delay, and pure count rate were measured. Then the following checks regarding the ion source were made: Deionized water level Vacuum quality

Position of TiH pellets Functionality of the power supplies for the table, magnet, and source stations.

Then the cage was closed, the LE Faraday cup dropped into the beam path, control system turned on from the control room, and values of the extraction voltage, strength of the Einzel lens, inflection magnet, and electrostatic triplet lens set to maximize the current on the LE Faraday cup. Then the following checks regarding the Tandem Van de Graaff were made: Vaccum quality of the Van de Graaff beam pipe (it should be around 10-7 Torr) Tension of the Laddertron (should be around 900lbs) The response of the AC monitor, vibration monitor, and tension monitor( This is done by quickly toggling the Laddertron belt power once between on an off once and listening to the Laddertron, and ensuring that the tension in the Laddertron drops) Ensure the position of the 11B target

The drive motor is restarted Laddertron and Tandem Van de Graaff was turned on. All individuals should evacuate the accelerator room and make their way to the control room. Here the values for the optical and accelerating elements were chosen on the LabView master control program once again to maximize the beam current at the target. The target was be bathed by the 6MeV proton beam for four minutes. Then the source was quickly whisked to the pre-calibrated detector by Rich Lefferts, where upon arrival and proper positioning, the coincidence count rate was measured in intervals of 200 seconds for 35 intervals (approximately five half-lives). Results and Discussion The results from the width v. count rate measurements are presented below: Width 200ns 100ns 50 ns 5 ns Count Rate (counts/min) 41.16 6.40 45.16 6.72 46.00 6.82 17.5 4.12

Count Rate v. Delay

60 Count Rate (counts/min) 50 40 30 20 10 0 0 50 100 Delay (ns) 150 200 250

These measurements were made with 22Na source before we exposed out 11B target so that we could use this knowledge to calibrate our detection apparatus. Our team needed to choose a good value for the width to minimize accidentals but not miss any true coincidences. Theoretically it seemed that if we decrease the output width of the discriminator, we would decrease the number of coincidences. But clearly for our experiment, 5 ns was far too small. The logic gate and counter were clearly missing valid coincidences. But given our theoretical inclination, it also seemed foolhardy to choose 100ns or 200ns. We clearly could maintain the experimental integrity of the circuit (actually improve upon it) by choosing the 50ns width. So for all other measurements this is the value we used. Also from this point on it became important to choose a good high voltage value for the photomultiplier tube. This became especially important because the uncertainty of a counting experiment goes like the square root of the number of counts. So it became imperative to choose a good value for the voltage that would provide a high number of counts, but not so high that we would count noise, and that the photomultiplier tubes would burn (we were asked to be gentle). So with the guidance of Richard Lefferts we wisely chose to keep both photomultiplier tubes at 1180V for all other parts of this experiment. The results from the delay measurements are shown below:

Count Rate v. Delay

3000 Count Rate (counts/min) 2500 2000 1500 1000 500 0 -150 -100 -50 0 Delay (ns) 50 100 150

This chart is analogous to the plot of similar data in Melissinos (Melissinos, Adrian C. Experiments in Modern Physics,. New York: Academic, 1966. Print.) and directly presents us with the number of coincidences. From this chart the most apparent glitch is that it is lopsided. We chose which direction we would call positive, and which we would call negative; the choice was absolutely arbitrary. So we would absolutely expect the graph to be symmetrical. But alas we have an asymmetry. The source of the asymmetry could only be one thing: one of the photomultiplier tubes had an inherent delay. To check whether this hypothesis was correct we measured the pure count rate of each detector. We found that detector 2 (dubbed Pinkie because of a pink tab that was streaked across the top of it) had an average count rate of 30840.894.6 (counts/min). Detector 1 had an average count rate of 29052.2147.8 (counts/min). Clearly, if both detectors were exactly the same we would expect the average of the two count rates to fall within the uncertainty of the other. But clearly that is not the case. Detector 1 obviously had a slight delay. But honestly, this imperfection would not impair our final measurement. The purpose of the experiment is to calculate the half-life of Carbon-11. We do not care about finding an accurate measurement of the number of Carbon-11 atoms we produce by smashing our 6MeV proton beam into Boron-11. We only care about the half-life. This is why for our experiment angular resolution is irrelevant: we are solely concerned with the half-life measurement. The results from the Carbon-11 measurements are summarized below:

Coincidence Counts for Produced Carbon-11

350000 300000 Number of Coincidences 250000 200000 150000 100000 50000 0 0 1000 2000 3000 4000 Time 5000 6000 7000 8000

Natural Log of the Coincidence Count Rate of Produced Carbon-11

Natural Log of the Number of Coincidences 14 12 10 8 6 4 2 0 0 1000 2000 3000 4000 Time 5000 6000 7000 8000

To try and fit the data we used the function below: ( ) Using this as our assumption we found the values of the variables to be:

sec 0.65 In each case the error in a variable was calculated by varying the respective variable while holding all other constant and allowing the 2 value to increase by one. The error for this variable was then given by subtracting the new or inaccurate value with the old or ideal value. This processes was repeated for all variables to give the error. These values are used to create a theoretical plot, which can be compared to the experimental plot (I would have actually liked to plot this function on the same graph as the experimental data; but the fit was so close that it became impossible to see the difference and the effect was lost)

Variables N1 1 N2 2

Value 2421100.673 8700 1744.635928 7.4sec 81443.59242 6200 489.7640938 56sec 0.79490581 0.65

sec

Graph of the Theoretical Fitting Functinon

350000 300000 Number of Coincidences 250000 200000 150000 100000 50000 0 0 1000 2000 3000 4000 Time 5000 6000 7000 8000

From these values the 2 was calculated to be 33.61761458 while the reduced 2 was measured to be 1.08443918. Honestly I was surprised by the goodness of the fit but it really goes to show that with quality equipment and the nudging touch of good guidance highly accurate measurements are possible. Finally the half-life was calculated using the formula below: ( ) We calculated the half-life to be seconds or minutes. The error here was given through error propagation from the and error calculation.

Conclusion We can confidently conclude that the value of the half-life of Carbon-11 is 20.15 minutes (with a reduced 2 of 1.08443918 how can you argue). This is really promising because not only is the fit good but the number seems to agree with the literature. The Atomic Mass Data Center reports the half-life of Carbon-11 to be 20.39 minutes. (http://amdc.in2p3.fr/web/nubase_en.html Atomic Mass Data Center. Atomic Mass Data Center. NUBASE.) But this lab does so much more than just present a good value of Carbon 11s half-life. It tells the story of its derivation. Along the way we discovered the inner mechanics of a Tandem Van de Graaff accelerator, the proper setup of a positron decay detector, and the subtitles of actual lab equipment. This lab is not a matter-o-fact reading. It is testament to our encounter with and discovery of physical science. List of References http://amdc.in2p3.fr/web/nubase_en.html, Atomic Mass Data Center. Atomic Mass Data Center. NUBASE Melissinos, Adrian C. Experiments in Modern Physics,. New York: Academic, 1966. Print