MainPage:Nuclear:Summer2015:ChristianPbWO4
Overview
At the Jefferson Lab (JLab) in Newport News, VA there is an electron accelerator which launches high energy electrons into different experimental halls. Each hall has its own specific targets for the electrons and detecting equipment for results of the collision. In Hall C a high momentum spectrometer and a variety of detectors are used. However, since there is no detector for high energy neutral particles, The JLab must create the neutral particle spectrometer (NPS). One of the major components of the NPS will be PbWO4, a crystal which scintillates, releasing photo-electrons, when hit by particles. These photo-electrons can then be picked up by a Photomultiplier Tube (PMT) which uses high voltage to allow for the detection of individual photons. It is imperative to the precision of the detector that the characteristics between the crystals remains uniform.
Intro
PbWO4 (see picture) is a dense scintillating crystal that, due to its growth method, often varies slightly between orientation and position on the crystal. My project was conducted to test the uniformity and homogeneity of lead tungstate for its use in the NPS, an electromagnetic calorimeter which will be built starting in September of 2015 at the JLab. The detector will contain ~1100 crystals. The crystals which I worked with and which will be used in the detector are 2x2x20cm They are grown from one end until they are 20cm long which causes the variations within them. It is necessary to understand if the crystals vary, because, for the calorimeter to provide the needed resolution, the crystals must be homogenous and of good quality. The quality can be tested by investigating various optical properties, such as their refractive index, transmittance, and light yield. All of the crystals measured were from the Russian company Bogoroditsk Techno-Chemical Plant except for two crystals which were called “CRYTUR” or "Czech" since they were from the Czech company CRYTUR. Also, each crystal had a label on its storage box excluding the Czech crystals. The labels were: ##, 078, 064, and ###. For the CRYTUR crystals, since there were only two, I labeled the longer one "long" and the shorter one "short." I was not able to take refractive index and transmittance data for crystal ###.
Goals
1. Collect data on the light yield, transmittance, and refractive index of multiple crystals.
2. Check the data to confirm that the crystals meet the parameters needed for the NPS.
3. Compare the data between different crystals to analyze their uniformity.
Conventions
Due to the growth methods of the crystals, there are variances within a given crystal between position on the crystal and orientation of the crystal. In order to keep track of this, I developed several conventions. First, I added dots to the corners of the crystals to denote the orientation. The third orientation was not marked but was always the orientation through the 20cm length. To keep track of position, I always measured starting from the side OPPOSITE the dots moving toward them. When taking light yield measurements, I had to wipe off the dots to clean the crystal. I then placed a sticker on the same side as the dots (there is also an arrow pointing to the that side) and noted the crystal and orientation.
Refractive Index
The index of refraction is a property of a material which dictates how quickly light travels through that material. As light passes through a material with one index of refraction to a material with a different index of refraction, the light changes speed and "bends" as it changes mediums. When light travels from a material with a low index of refraction to one with a high index of refraction, the light is bent toward the normal, and vice versa.
An equation
Using Snell's law:
n1\sinΘ1 = n2\sinΘ2
and properties of angles and geometry, I, with the help of Marco, was able to create an equation which solved for the refractive index (n) as a function of the width of the crystal(L), the displacement of the laser(Δx), and the angle of incidence (Θ). Since the crystal reflects light well, the angle of incidence can be calculated by measuring the angle of reflectance and dividing that by two.
Measurements
To take measurements, I used calipers and image analyzation software. The calipers were used to measure the width of the crystal and the displacement of the laser, while the analyzation software was used to measure the angle. Using a turntable, I set up the crystal at a random angle. Next, I measured the displacement of the laser three times and took the average of the measurements. Then, using a level and a camera, I took a picture from nearly directly overhead of the crystal. After taking 10 measurements and pictures, I uploaded the pictures to the computer and used the analyzation software to calculate the angle. Lastly, the data was entered into excel which computed the refractive index.
Uncertainty
One concern with the equation is that it is very sensitive to fractions of a millimeter. This makes diligence in measurements very important. Unfortunately, I did not discover the my best method until near the end of my work. Therefore, for most of the refractive index measurements, the uncertainty is fairly high, around ±.25. I was able to reduce this number closer to ±.15 for my final measurements.
These uncertainties were found by taking each of the measurements numerous times to discover the uncertainty in those measurements. For the displacement it turned out to be ± .3mm for the width it was ±.05mm and for the angle it was ± .3°. Then, using excel, the highest and lowest refractive indexes based on those uncertainties were calculated. The difference between these numbers and the calculated refractive index is the uncertainty.
Results & Conclusions
When calculating the refractive index error, due to the equation being very sensitive even at variations of one tenth of a millimeter, we cannot make any larger conclusions. The measurements support that the crystal is consistent with literature, and, within the error, they do not deviate. I took ten data points for each crystal, always through orientation one. I did not test different orientations due to the large uncertainty in the measurement. The uncertainty was calculated by first taking each of the three measurements numerous times to discover the uncertainty in those measurements. For the displacement it turned out to be ± .2-.4mm (earlier methods were less exact), for the width it was ±.05mm, and for the angle it was ± .2-.5°. Then, using excel, the highest and lowest refractive indexes based on those uncertainties were calculated. The average of difference between these numbers and the calculated refractive index is the uncertainty.
Refractive index histograms
Transmittance
The transmittance of a material is the the fraction of incident electromagnetic power that is transmitted through a sample. Knowing the transmittance of the crystals tells us: a) if the crystal is of the right characteristics for the detector. The detector requires a crystal which transmits at least 60% of light at a wavelength of 420nm. b) if there is uniform characteristics between the crystals. Further, using Fresnel's equations, it can be used to double check the refractive index.
Measurements
Measurements were taken from 250nm to 800nm using a PerkinElmer Lambda 750 UV/VIS/NIR spectrometer. Due to the large size of the crystal, we had to create a stand for it which would fit inside the spectrometer. A piece of strong Styrofoam with grooves cut into it was used to hold the crystals during measurements. For the measurements, I set the spectrometer to use wavelengths of 800 nm to 250 nm in 1-5 nm steps which. I took measurements through two orientations and along the crystal in one inch increments. I then analyzed the data using Excel.
Results
The transmittance measurements proved to be fairly uniform. There is variation from crystal to crystal and orientation to orientation. I found that the deviation is ±1.5% between the crystals and the error of the measurements to be a minimum of ±.1%. Because of the number of measurements and the time it took to take them, the systematic error is more likely ±.3%. Due to this systematic error, it is hard to make conclusions on the uniformity along position; the crystals could vary from one end versus the other, but within the error the data is inconclusive. A setup which could keep the crystal consistently perpendicular to the light beam of the spectrometer is necessary for confirmation or disapproval of this hypothesis.
Transmittance Data
Error creating thumbnail: Unable to save thumbnail to destination |
Light Yield
Light yield is a property of scintillators which is used to calculate which particle hit the scintillator. Light yield is defined as the number of photoelectrons emitted per amount of energy striking the scintillator. In a detector, one can work backwards from knowing the number of photoelectrons detected, to then knowing the energy of the particle which struck it, to then knowing what the particle is. In our specific case, a PMT is attached to the PWO crystal to measure the light yield when a photon emitted from a Na-22 radioactive source strikes the crystal. This can be used to check that the yield meets the parameters required for the Neutral Particle Spectrometer.
Set Up
It is known that the light yield of lead tungstate varies with temperature. In order to keep the temperature constant, we modified a freezer that could house the Na-22 source, a Hamamatsu R4125 PMT, a collimator, a scintillator for triggering, and an ADC-based readout. Inside the freezer was also a heater, a stepper motor to control the position of the crystal, and a dehumidifier to stop ice from forming on the walls of the freezer. The heater was used to keep the temperature at ~50 degrees and to stop the freezer from displaying an on and off cycle (which saves the motor and would not have an adverse effect on food) at low temperatures.
The stepper motor was controlled remotely using a computer. Each 'step' equaled 3.175 micrometers. Therefore 3150 steps caused to motor to move the crystal 1 centimeter and 6300 for 2. However, due to a thoughtless math error, for some of the measurements 7300 steps were used which equates to 2.3 centimeters.
Keeping the temperature constant proved to be difficult. When left on overnight, sometimes the freezer would shut off or slow down and the temperature would rise into the 100 degrees fahrenheit. While here, keeping the temperature within ~1 degree was also difficult. The fight between the heater and freezer proved it could keep the temperature constant, but if ever knocked out of equilibrium, the temperature would sometimes rise, sometimes fall. The difficulty could arise from opening the freezer for a few seconds, or sometimes happen without notice.
Results
Based on my results, light yield clearly showed variance with temperature. Three of the crystals I measured showed a tight clustering of light yield, while one did not. The outlier's measurement (crystal 078) is off because of a bad LED connection which caused an erroneous PMT calibration. Based on the results from the three crystals, there was a clustering around 8 photoelectrons. Comparing the light yield accurately from crystal to crystal cannot accurately be done with this data due to varying temperatures. Since the energy of the radiation from the Na-22 source is .511 MeV, the light yield per MeV is roughly 15-16. This fits perfectly with the light yield of 10-15 ideal for the Neutral Particle Spectrometer. I also hoped to look for a correlation between position and light yield. Because I could not keep the temperature stable for a long enough periods of time, I was also unable to determine if there was a correlation.
Plots of Data
VSL Presentation
Note: This was created and presented halfway through my internship.
Research Paper
Discussion
The transmittance showed that there are variations between crystals and orientations. The deviation is about ±1.5% excluding the newer crystals from the Czech company. To more accurately observe a difference along the crystal, further, more accurate data would have to be taken. I drew a perpendicular line on the styrofoam block for the transmittance to reduce error, but this still left room for human error when determining if the crystal is perpendicular. A small wooden stake (like a toothpick) could be inserted on the line so that the crystal could be placed against the stakes. However, if there is a deviation along the crystal cannot be more than ±.5%. For the refractive index, the sensitive equation proved to be the problem. It was accurate enough to confirm that the crystals, within error, are of good quality. The results also further give confidence that these crystals are PbWO4. Developing a more accurate method which could decrease uncertainty may be possible, but due to the small width (2cm) of the crystal, this may be difficult. In order to make light yield measurements, one must calibrate the PMT. Error in the light yield came from a non-consistent measurement for the calibration. Getting more consistent results of this calibration is necessary for an accurate measurement of the light yield. Further, controlling the temperature of the freezer so that it can remain within .25° F would greatly improve results so that they could be compared from crystal to crystal. Alternatively, if each crystal proved that it varied consistently with temperature, one could calculate the light yield at different temperatures and use this data to create a function that could undo the effects of a varied temperature. The data taken proved that, even with the lowest measurements, PbWO4 has a light yield which meets the parameters required for the NPS.
The Cube
The Chinese company Shanghai institute of Ceramics (SIC) gave us a small 2x2x2cm sample cube of lead tungstate. I used this cube to become acquainted with measurement techniques for measuring the refractive index and transmittance. However, my results contrasted with literature. I measured a much lower refractive index of 1.9 and higher transmittance. This caused confusion and I hypothesized that potentially the crystal was grown with a less perfected method (there was data from studies(cite) that showed older methods had a lower refractive index). Older crystals, contrastingly, displayed a lower transmittance. Therefore we sent the cube to be analyzed by an X-ray floriencience (XRF) machine to measure the chemical makeup of a material. The machine uses x-rays to displaces electrons causing a measurable burst of energy characteristic to each element.
Results
The results measured using the XRF machine showed that the crystal had no tungstate! This means that the manufacturer, SIC, mixed up their labeling when giving us a sample of their crystal. Instead the chemical make up was primarily lead and fluorine. This lead me to believe that the cube must be lead fluoride, PbF2. For confirmation of this, I looked at literature which had transmittance and refractive index data for PbF2 and compared their results with mine. The results were very similar which further gave confidence that the cube must be PbF2. This also explains the unexpected results for the transmittance and refractive index.
Annealing with Violet Light
After testing the cube with the XRF machine, the face which was exposed to the x-rays became tainted with a brown color. We believed that this was radiation damage. To test for if it truly was radiation damage, I took transmittance measurements on the cube and compared them to measurements I had taken before using the XRF machine. Interestingly, the transmittance of all of the orientations of the cube seemed to have been altered. Next, to test a known property that lead crystals can be healed using light, I shinned a violet light at the cube and re-took transmittance measurements periodically. The voltage used for the light was was 3.4. Initially I only had one violet LED shinning on the cube, but two more over time. The results measured from the transmittance, as well as a reduction in the discoloration, prove that the light annealed, and therefore improved, the crystal. After the damage, the transmittance was altered through other orientations as well. Strangely, however, this alteration did not look like damage but rather different transmittance results entirely.
