July 26th- Day 20

Comparisons.

We set up the PET(002) with a Bragg angle of 11.81 degrees and an IFD of 121mm with a Cobalt source. The optic on the rotational stage was plugged into axis 6. We hooked up the camera with the gain and gamma settings all the way to the right. We took an image and aligned the optic. Afterward, we placed the detector back in position and hooked back up the z axis of the detector in axis 6. We placed the orange film in front of the detector and took another image. The spot was not in the center of the detector, so we moved the detector and tried again. This time the spot was closer to the center. We took the film off the detector and ran scans to find the Co peak. Scans were taken at 25kV, 30kV and 50kV and the number of counts was recorded. The solid angle of each optic was calculated using the equation below:

solid angle = area*sin(Bragg angle)

                                                                   IFD^2
Since the PET optic is comprised of several pieces, the solid angle was only an approximation. The solid angle for the Ge optic was also an approximation, but would be closer to the actual number due to the single piece. The solid angle for Ge(111) was calculated to be 0.0031 and for PET(002) is was 0.024. For each optic, the number of counts at 25kV was divided by the solid angle. This allows the optics to be compared 'apples to apples' by taking into account the differing solid angles. Dividing takes out the solid angle factor. By dividing, the flux is calculated for each of the optics. For Ge(111) it was 3.4E6 and for PET(002) is was 3.8E6. The higher the flux, the higher the efficiency making the PET optic more efficient. The flux for PET might even be higher because the solid angle could be an over approximation making it even more efficient. The optics were scanned using a different source to compare efficiencies.The new source is a Moly source. We calibrated the detector to pick up a certain peak. With this source, the peak moves as a result of varying Bragg angles. For finding the correct peak, specific ROI of 218 and 251 were set. The peak needed to be centered in this range. We tried taking an image using the film, but the detector needed to be realigned. Even after moving the detector the spot was still fairly large. A spectrum was taken but the results were not what we were looking for. Alex messed around with a few things and then took spectra once again at 25kV, 30kV and 50kV. Using the counts at 25kV the flux was calculated to be 2.2E6. Since this number is less than previously calculated, the efficiency of the Moly source was calculated to be less than the Cobalt source.
Once these calculations were complete, spot size measurements were to be taken. Unfortunately, scans were going haywire and were not producing the desired results. The detector had somehow become extremely out of place. After fixing the setup, the background ratios were lower than normal. The spot size was expected to be much larger than that of the Si(111) optic. We tried taking an image of the optic using the film, but could not find the spot. Tomorrow we will attempt to fix the problem.

July 25th- Day 19

Not so successful.

Once the new power and x-ray sources became available, Alex and Zewu were pressed for time to get data. Using the Co source, they first found an image using the camera. The image of the Ge(111) was more of a square compared to the image we obtained using the Si(111) optic. We were back to using the SDD high intensity detector that we were having problems with previously. Zewu and Alex were looking for a high flux which would mean a high efficiency. To compare sources, they would look at the intensity, from which they could calculate flux. A 127 micron Ni filter was used when taking the spectra. After they got the data they needed, we tried to minimize the spot size. We found the spot fairly quickly in the vertical, but had very different luck trying to find the spot in the horizontal direction. The spot size kept decreasing and the astigmatism would have been extremely large (~12mm minimum). Since the astigmatism should not be that large, we abandoned the scans. Alex said the tools were not right to make the optic so it would not be precise like the Si(111) optic. This spot size would have been significantly larger than the Si spot size due to the differing lattice structures. Due to being pressed for time, we would have to move on to test the PET(002) optic and come back to the Ge(111) spot size measurements.

July 22nd- Day 18

Imaging optic measurements.

Alex had finished aligning the new Si(220) optic and was ready to take the final measurements. This optic was being manufactured for Bruker for some imaging purpose. Vertical and horizontal spot size measurements were taken at 30mm, 70mm and 3mm. The OFD of 30mm produced the smallest spot (x- 152.4um, y- 217.4um) as expected. Measurements were taken at these distances to see how well aligned the source was and to see how good the optic is. A Mo source was used. This beam is used for divergence. The company has a sharp tolerance for the location of the beam. It needs to be in the center of the detector, so it can be blocked easily. If the beam was allowed to hit the detector, the detector would become over-saturated. When the beam is stopped, one can see the rings of the pattern diffraction on the beam stop.
The flux was calculated to be 4.85E7. To find the tilt, we focused and zeroed the detector at the normal OFD. Then the detector was moved back 500mm and the beam was scanned using the detector to locate the center. The difference between the two points of focus and the length of 500mm was used to construct a triangle of the vertical tilt. The angle was found to be 0.04 degrees. Calculating the tilt in x makes no sense because the error is so large.
__________calculations_________________
To calculate divergence, the optic was rotated to find the peak. While the beam was slightly focusing, the divergence was still calculable. The divergence shows the crystal diffract different energies. Doing a horizontal scans shows a shoulder off to the left of the main peak we were looking at. This was Mo K alpha2. Each angle represents a different energy. An intensity measurement was taken to observe the number of photons hitting the detector. The divergence can be taken for the 90th, 95th, or 99th percentile.
.......
The vertical divergence was calculated to be 0.66 and the horizontal divergence was 0.63.

July 21st- Day 17

Flux. Questions. Germanium. Fire drill. Heatwave. BBQ.

After fitting the intensity data we plotted in matlab, we used the equation of the line to estimate the number of counts at full power (50 keV and 0.6 mA) <2712700 counts>. Because we took 100 second spectra, we then divided this number by 100 to get the estimated counts per 1 second at full power <27127 counts>. We then took this number and divided it by the transmission of air (0.62670) found using the www.cxro.lbl.gov website <43285.46>. Then to account for varying spot sizes, we had to divide by 16 because the area of the spot size we obtained was 16 times smaller than the area of the 200 micron spot size pin hole we had taken the spectra with <2705.341>. This flux calculation is significantly smaller than the flux we had previously calculated using the optic and copper filter. This was the expected result. Because the detector has such a high efficiency, taking the detector's transmission efficiency into account has no significant contribution to the flux. This is because all of the x-rays pass through the Be window.
We got a chance to ask Alex a bunch of questions. What we had been doing for the past month or so was characterizing the optic, Si(111) to be specific. This optic would stay at XOS to be used as a reference when making other optics. However, similar optics would be sent to customers, companies like Jordan Valley. JV could be using the optic as a part of their semiconductor for some imaging purpose. Their semiconductor might use multiple beams to collect different data. Alex wasn't too sure sure exactly what they were using the optic for. We used a monochromatic convergent beam to take measurements. Applications could include anything from edxrf or wdxrf to imaging. Based on the customers' desired spot size, energy, and total working distance, XOS picks out a crystal that will fit the parameters. Crystals will accept different energies based on their lattice structure. The lower the spot size, the lower the acceptance. Graphite, if used as a crystal, accepts a much wider range of energies when compared to silicon. As a result the resolution of graphite suffers. When a crystal is bent it can accept an even greater range of energies. With our Si(111) optic, the divergence was large because of the convergent beam. A parallel beam would have a divergence of around 0.01 degrees.
We also started setting up a new set up for Ge(111). The Bragg angle is 15.89 degrees and the IFD/ OFD is 123mm. However, we would need a new source/ power supply that would not be available until next week, so we could not get very far. 

*Numbers in brackets represent the calculated number after doing the described operations.

July 19th- Day 16

Roughly perfect.


We went to XOS to work on the new optic, but upon walking into the building, Alex told us he had taken apart our new set up. He needed to test another optic and figure out why all the optics are turning out bad. So far the one we had was defected and another was broken. They aren't having very good luck making optics as of recently apparently. We went through and wrote down all the information we could get about the equipment and took pictures of the things we hadn't before. We got knife-edge scan data that we will plot in matlab in both the x and y directions. We took spectra during the knife edge scans, but they didn't turn out how we wanted. Then we took 100 second spectra with and without the lead tape so you can see the effect the tape has on the x-rays and what the detector picks up. The one without the tape shows Cu K beta, which is minimized on the spectrum with the tape barrier in place. With the data, he showed us a new program, Origin. It is like excel and using it, he plotted the data from the spectrum scan. He also showed us how to plot it in excel and tricks for importing data in various programs. He explained all the different peaks and what was causing them. This would be the energy scan. Alex also had us calculate what the energies should be using the channels the peaks were located at. Using these energies we made another graph (energy vs. channel) which produced a linear line. The equation from this line allowed us to change the x axis from channel to energy on our spectrum graph. We have all the data in excel from XOS and will enter it all into matlab between the two of us. We then took spectra for intensity to calculate flux without the crystal with a distance of 400mm. 

July 12th- Day 15

Just our luck.

Using the new software Alex showed us how to use on Friday, our goal was to minimize the spot size. We were shooting for a vertical spot size (FWHM) of either 8 or 9. He had everything aligned Friday and just had to move the z axis of the optic to try and further minimize the spot size. When we got there, we turned all the equipment on and started up the software. Our FWHM in y was 11.0. We went to move the actuator controlling axis z of the optic, but it did not want to go any further than 0.275mm in the positive direction. It had worked just fine Friday, but we did not want to force anything, so we went 1.0mm in the negative direction. We scanned axis 1 (rotation) and axis 4 (horizontal), but neither of the scans produced the desired results. We even 'went to home' with everything in between scans, so we shouls have at least gotten something. We moved the z axis actuator as far positive as it would go (0.25mm) and ran a scan. This should have produced somewhat better results because we proved the z axis needed to move further in the positive direction before we left on Friday. The FWHM was around 50. Then we notied axis 4, 5, and 6 all said they were at +8.0mm, but when we went to move them, they jumped back to 0.0mm after displaying '+Inf' and we have no idea what happened. After that the computer started running slower than normal and none of the actuators would move at all. Renee and I tried restarting the computer, but it was unsuccessful. We did not want to break anything so we called it a day and went back to the lab. We didn't do anything we hadn't done before, and hopefully Alex will know what to do.

July 8th- Day 14

New Setup!

Alex showed us the new setup and told us all about it. This new crystal was Si (220) and had a Bragg angle of 1.06 degrees. The IFD is 75mm and the OFD is 328mm. This setup has a fixed source instead of a fixed optic, so the crystal can be moved in the x, y, and z axis as well ab rotated and tilted. The detector is the same as our previous setup having mobility in the x, y, and z axis. We used a shadobox (1024) camera to view the image of the optic. The camera was placed close to the optic and we would be using it to not only make sure the optic was aligned but to find the focal point as well. Most of the day was trying to find the focal point. Once we found a good spot size, we then had to move the z axis to try to further decrease the spot size. However, the FWHM the camera gave us is not the actual spot size.

July 7th- Day 13

After a month of research all we have is... a little image!

Measuring the intensity would need a different approach rather than measuring the x-rays straight from the source because it would over saturate the detector. Instead, Alex told us to take 100 second spectra slowly building up layers in the filter. The Cu filter we had been using consisted of 8 layers. To take the spectra we needed to take apart the filter and run scans while still having a reasonable dead time (<10% optimal, but <30% is acceptable). After realizing we had been going about the wrong procedure and altering the current, we also realized our results made no sense. After getting assistance, Alex informed us that we needed to keep the current the same. However, we still were obtaining a high dead time while only using a few layers in our Cu filter. We would have needed to use an extremely low current, but at such a low current the tube is unstable so the results would not be accurate. Instead, Alex had us run different spectra using all 8 layers of the filter at a constant 50.7 kV. We were to increase the current by .05 mA intervals and take count readings. Plotting this data resulted in a linear line from which we could calculate the counts if the spectra was run at full power per 100 seconds. Dividing by 100 resulted in the number of counts per second at full power. This number divided by the transmission efficiency (taken from www.cxro.lbl.gov) resulted in the flux taking the Cu filter into account. The same mathematical procedure was then used to find the flux taking the air into account. After that, we placed an orange film really close to the optic to obtain a picture. With this image, one can see the defects of the crystal due to the various shading within the image. We then measured out 400mm from the optic at the Bragg angle of 14.22 degrees and obtained another image of the optic. This image was larger than our previous imaged and reflected with respect to a vertical axis.    

July 6th- Day 12

Fire drill!? Don't worry we still got our spot size measurements.

We continued doing more knife-edge scans at different input focal distances, just to see if we could get any better results with smaller astigmatism. Unfortunately, we weren't very successful. We could not obtain any results for the x scans at 204mm or 205mm. However, we were able to obtain results from our y scans at 204mm and 205mm that continued the pattern that we would have expected to see due to our previous results at smaller IFDs. We then graphed all of our data points in excel using our designated computer at XOS to see the trend in our data easier. According to our results, the IFD of 203mm had the least astigmatism and spot size. Then we ran official scans to get the real spot size. These scans were run at a preset time of 10 seconds and took much longer. The higher preset time was to achieve a much high background ratio. We also had to focus on a very narrow section of the the original curve when finding the correct range of values to scan at. Our official spot sizes: horizontal- 69.5 microns, vertical- 51.85 .

July 5th- Day 11

There's always something.

Luckily, today the detector worked just fine when we turned it on. Since we completed the y scan at 204mm, we started off with the x scan at 204mm. However, the values we obtained were in the 200-300 micron range. We had already proved the detector was working fine and even Alex did not know what was causing this to happen. He said it was a question for Zewu when he comes back from vacation. The only thing we could find was that the x scan was picking up both the K alpha 1 and K alpha 2 peaks in the plot. This was flattening out the graph of the counts vs. position, which was why the spot size was so large. Alex said it was very tricky to get the two separated and said it would either take too long or that he didn't know how. Instead, we scanned x and y at 203mm. Our astigmatism was slightly smaller than before, but still not optimal. Alex then told us to try the 204mm scan again.  We were able to finish the scan in the y direction and will try the scan in the x direction tomorrow.

July 1st- Day 10

Alex wasn't kidding when he said these scans take forever.

Once again when we powered up all the programs and equipment, the detector wasn't picking up anything. We checked the calibration with the Fe-55 and the gain was extremely off. We restarted the computer and everything was in working order again. To fix the astigmatism we had to move the z of the source in the positive direction. Alex suggested going by 2mm intervals and then graphing all the data to see which OFD minimized the astigmatism. The scans were run at 1.0 second preset time. We were only able to complete a full y and partial x scans at 202mm before lunch. When Dr. McColgan came with me after lunch, the x scan was completed along with a full y scan at 204mm.   

June 30th- Day 9

Repeatable results is what we aim for!

Even though, we were able to obtain a fairly small spot size yesterday, we did not realize how bad the astigmatism was until this morning after going back and looking at our results. Alex suggested increasing the input focal distance, but was not all too sure if that would help. After running a scan of x and y of the detector to try and find the spot again, it was unsuccessful. The detector was all out of wack and we could not find the peak. Nothing was changed since yesterday and we are not too sure what is causing this to happen. We even 'went to home' with all the axis before turning everything off so when the equipment was turned on today everything would be in place. We double-checked the set up and the detector, optic and source were all aligned correctly. We tried looking up potential problems and solutions, but were unsuccessful. Alex was extremely busy, so we did not want to take him away from his work. In the mean time we read through the website you sent us the other day and tried looking up some answers to the questions we need to address for our outline/ poster. When Alex came to check on us we told him the problem and he suggested trying to recalibrate the detector and to restart everything. The calibration check was unsuccessful; the gain was extremely off. It took a while to get anything to show up on the screen after restarting the software. Then we tried restarting the entire computer and it worked! Once that was all squared away we scanned at the appropriate OFDs to see if we could obtain the small spot sizes we had gotten yesterday. The vertical spot size was almost dead on. While the horizontal spot size was slightly larger, the values were acceptable due to the fact that these scans were only rough measurements. Tomorrow we will start experimenting with the input focal length and work on correcting the astigmatism.

June 29th- Day 8

Go to home. Okay!

Since we did not 'go to home' yesterday before shutting down the program, when we turned on everything the computer made the positions of the stages the new 'zero' position. We had to go back and scan the x and y of the detector to find the spot. After that we started scanning the x and y axis for the smallest spot size but were not following the correct procedure. We realized this after a few scans and by the fact that we had been obtaining fairly large spot sizes. We went back to our notes and corrected the procedure we were using. We also increased the power to obtain a higher background ratio. Changing the current would only increase the dead time (Note to self- run at 50kV but make sure dead time is <30% but preferably <10%). The z axis actuator of the detector started acting up and would not go to the correct position. We would have to move the actuator position by a millimeter opposite of the way we were trying to move it and 'go to start' or 'go to home' again. We started scanning again but again were obtaining fairly large spot sizes. To correct this, we adjusted our output focal distance (OFD), removed the pin hole, and scanned in the x axis of the detector. We found the spectrum, so we did not have to use the camera. Then the region of interest (ROI) was set and the pin hole was put back in. We scanned x and y to get everything in position and then started scanning to get spot sizes. Alex suggested scanning in y first because the spot size would change more drastically. We followed the same procedure we used during our training.  Alex said next we should be changing the IFD and take more spot size measurements. We asked him about the new set up he had been working on for us and he told us the optic keeps breaking, so we are going to have to wait until they make another one that survives the treacherous creation process.

June 28th- Day 7

Happy News!!

Right as we walked in Alex said he had 'happy news' for us. He had needed to use our detector to compare some other data to and our detector had broken. This was the reason why we were having so many problems. We had examined the detector using a Fe-55 sample. The temperature of the detector was reading to be around 230K when it was sitting at room temperature and should have been somewhere around 220K. Alex took our detector and was going to have AMPtek fix the problem. After talking to Zewu he gave us a XR-100CR Si x-ray detector. This was an analog detector unlike the digital detector we were using before. This meant it had to be hooked up differently and didn't have a USB to plug into the computer. Instead it had to be hooked up to the middle tower under where the stages were hooked up and the tv and run using a different software. The new detector was then calibrated with Fe-55 like before. After we scanned at full power without the pin hole and with the Cu filter to look for the Cu K alpha peak around 140 channel. We scanned at a lesser power in x and set to zero. We ran into some problems trying to find the spot when scanning in y and couldn't see the spot. We ended up having more alignment issues. However, we didn't realize this until after we had already attempted taking some spot size measurements. This makes sense because our spot sizes were fairly large.

June 24th- Day 6

Spot on!

With the camera anyways. We put the camera in place and almost instantly were able to see the image we were looking for. We then hooked up the detector and were able to find the spot rather quickly. That's when things started taking a turn for the worse. One of the actuators stops working for us and it had to be replaced. We aligned the old actuator with the new one to minimize the scanning that would have to be done to align everything later on for finding the spot again. At first Alex gave us a Ni filter to use but then decided it would be more efficient to use a Cu filter instead. We then scanned in x, set the ROI values and scanned in y of the detector. Then to account for the backlash when scanning the source in x had to drag the cursor to the peak > set to start > go to start > set to zero > move the stage minus about 200 microns > go home. Then we scanned the source in y and did the same procedure for the backlash. After we put the knife edge back on and started scanning for the spot. After each scan in x and y of the detector we set to zero. Then we scanned  y and x of the source and followed the backlash procedure. After that was all set we started scanning for the spot size. The vertical spot size should change more drastically according to the results obtained by XOS that we were using as a reference. Before we could get very far... another actuator decided it wasn't going to move as far as it could and we had reached a 'hard limit.'

June 23rd- Day 5

Back to square one.

When we got there we removed the detector and set up the camera again. Instantly, we got the image we were looking for and set the detector back up. At first we scanned the optic without the knife edge giving us a larger surface area to locate the spot. Still nothing. Alex concluded that there must have been communication issues between the computer and the detector. Alex worked his magic and tried a few different things. We tried calibrated the detector using an Fe-55 isotope. We switch into PHA mode and hit acquire on and looked for the peak. After reading the channel of the main peak the value and known energy of Mn K alpha were entered. The energy of Mn K alpha was used because the Fe-55 isotope decays to that energy. Next the ROI was found by switching to energy > spectrum > and look for the desired peak. In this case we were looking for the K alpha peak not the K beta one. Select the ROI with the red lines and hit 'set ROI'. Then the pin hole was taken off the detector to maximize the area of the detector. We scanned x and y and both were unsuccessful. Alex suggested calculating the difference in Bragg angles between Cu K alpha and Cu K beta. Using the Bragg equation we found the difference in angles to be 1.37 degrees which meant we had to move the detector 0.4mm in the negative x direction. After looking at the set up, we realized that we had not taken into account the distance from the anode to the outside of the x-ray source. The set up had to be realigned, yet again, but once that was done we were ready to go... for tomorrow.

June 22st- Day 4

Testing... 1... 2... 3...

We got badges and keys to the place. We're official now! Alex started things off by having us watch a safety video about the ionizing radiation we would be dealing with and the safety precautions we would have to take. We made sure everything was grounded and in place. Then it was on to the testing. We turned on the camera and... wait for it... couldn't see anything. Big surprise. We started moving the x of the source using the green '+' and '-' buttons and found a little glimpse of the crystal fairly quickly. The tricky part was trying to make the reflection of the rays off the optic a big, bright, straight, vertical line in the center of the camera. There was a lot of guessing and trying different things moving the various axis. After struggling for a while Alex came to help us. After we adjusted the height of the camera and source it turns out we weren't too far off. We knew we achieved what we were looking for when we moved x and the whole image disappeared at once. This meant the source and the optic were aligned correctly. Just trying to find the image took most of the day. Our next task was to hook up and align the detector. After that job was completed we scanned x to find the spot and then scanned x and y of the detector trying to see the peak of the Cu K alpha. At first we saw the spectrum then for whatever reason we lost the spot. Alex tried putting an orange paper(?) over the detector that would leave a mark where the spot hit it. We couldn't see anything. Alex suggested changing the set up and we tried it without any luck. By that time it was time to go home.

June 17th- Day 3

Ready, set, go!

This time it was all on us. Alex got us our own table to work at, all the parts and pieces we needed and told us to start building. We started with stacking and securing the stages on top of each other to create the 3 different axis we would need for both the source and the detector. Some assembling was easier than others, but by lunch time we had build our set up and roughly aligned it all. After a nice lunch at Paneras we hooked up all the equipment to the computer and power sources. We tested the camera and it worked! We could see the light coming through so at least it's ready for action. We'll have to wait until next time to see how well aligned our crystal is (which would be absolutely incredible)... or not at all (which is probably the case).

June 14th- Day 2

Training.


After our fist meeting, Zewu had given us data and figures for the optic so we knew what we would be dealing with. We went to XOS bright and early to be trained on how to use all their high-tech equipment. Alex had already set up a model system for us to play around with. He explained the set up and showed us what to do. Shortly after, we were let loose and told to minimize the spot size. Several scans later, we were off, but still fairly close to the values the company had achieved. 

June 9th- Day 1

The first day of the rest of our lives... well summer anyways.


Summer research with Dr. McColgan began June 6, 2011. Three days and a lot of reading later we were introduced to Zewu Chen, the rocket scientist behind all the magic at X-Ray Optical Systems, Inc., (XOS). After a fairly brief meeting with Zewu and Alex, the guy who would later be assisting us, we left with a few things (solid angle, knife edge scan) to look up before we would return.