Feb 2008 beamrun, day 1
22/02/08 13:04 Filed in: Work Experiments
Welcome to another "installment" of beamrun news from the APS.
Our group has a week of time at beamline 12 and will be performing a series of x-ray reflectivity experiments from various materials.
What follows is a rather dry post. However, it's one of the parts of this work and serves to justify how unglamorous some of this can be.
Currently we are suffering the virtually inevitable problem of alignment, or in other words getting the tiny beam of light conditioned and in exactly the correct place. Depending upon the kind of experiment you wish to perform, the alignment (or geometry) needs varying degrees of rigor. For some experiments, with a large sample and where the angle of the scattered light is not very important, the experiment can be done without careful alignment. Our experiments tend to be very sensitive to the various angles and positions, making us spend a good deal of effort to align everything. All of this is further complicated because every piece of equipment that interacts with the beam sits on something. Each piece may not be firmly attached, or may not be completely centered. Even the concrete floor can occasionally give a little shift (I've heard stories of the operators "detecting" earthquakes by variation in the electron beam).
What all has to be aligned and how do we do it you might ask...
The first part is largely out of control of the users (which isn't always a bad thing) and that is the light source itself. In this case we have a very large magnetic field bending the flight path of the electrons. As they follow a curve they release a continuous spectrum of light. This light tends to be brightest in the forward direction of the beam and already fairly small and tight. By changing the magnetic fields on the electron stream we could alter the initial direction of the small spotlight of photons coming out. However, direct manipulation of the electron beam by users (especially in some cases) can be a very bad idea leading to all sorts of problems. The worst of these would be total electron beam loss effecting all the other users in the facility.
So we have little control over the light source itself. It brings a beam of light down a pipe to a monochromator. This ingenious device allows us to select out a particular photon energy, or at least a very narrow band of photon energies, eliminating all the unwanted light. It also can (and does) serve a second purpose in most cases by acting as a focusing element for the beam. All of this sits down from us by about 20 meters. From there the light travels down a long pipe and into the experimental hutch.
Inside the hutch are the last few meters of the photon flight path, and incidentally the most important ones. The beam is conditioned further by a secondary focusing element and then narrowed by the use of several slits. The focusing serves to get the most light in the area of interest and the slits further work to eliminate unwanted stray photons. It also allows is to correct for minor the beam intensity variation.
So from the light source the photons travel first through the monochromator, then 20 meters and through more focusing elements and slits before all before arriving at the sample. It would be simple if we just had a place saying "put sample here," however that doesn't quite work. First there is the inevitable drift and change that occurs in the position of the beam from adjusting any of the previous instruments I mentioned. So "center" is a moving target every time (virtually) that an experiment is set up. We locate the approximate center by placing a small piece of paper that "burns" in the x-ray beam. By seeing where the make is made on the paper we can then get the diffractometer(the device that holds both the sample and the detector in place) into approximate position. From there is gets a little tricky.
Remember, we can't just see any of the photons with our eyes and being inside the hutch with the x-rays is right out. And even having the sample sitting in the center isn't enough. Each axis that the sample rotates about must be aligned and centered. We have various jigs and pins, precisely machined, that are used to position and align each axis. And since there is no definitive starting point, once we've aligned each axis and positioned everything once, it's still not fully aligned. Rather it is an iterative process with each step bringing us closer until we're within the tolerances of all the motor positions. Having such devices (which weigh well over a ton) move reliably, repeatably to within sub millimeters is something to behold... well, it is for me anyways.
At this point things should be good and all that's left is to define the detector aperture. This isn't terribly difficult as it only involves (you guessed it) more slits.
Once all this is done, we're then ready to actually move our experiment into the hutch and observe what happens. If we get lucky and things are reasonably placed from the pervious users, then it's only a few hours. Our record is around 4-5 hours I believe(and that's not because we're slow, but that's just how long it takes to systematically check everything and make minimal adjustments). At times we've gone a couple of days before working out all the problems. As that can be a couple of days out of our entire week without even placing our experiment in the hutch, it gets to be a major frustration and obstacle. I seriously wonder what the cost-benefit analysis would say regarding how much time and effort (which means money) that we waste each time because we have to take such pains (as many if not most users must do). Some of the equipment is certainly "top of the line" but much of it is "lowest bidder" equipment. How much could we have saved (and how much easier would it be for us) had the initial investment been higher?
Anyhow... We're reasonably lucky today and have an aligned diffractometer as of about 3am.
About the time I typed this we realized we might not be so lucky. I also left out a piece from the earlier writing... There is one other component in the flight path of the photons, a filter box. While for our experiments we need as many photons as possible, it is often necessary to eliminate some or most of the photons at various times for what I'll call diagnostic purposes. Usually this means checking our alignment or the orientation of a crystal or some such thing.
This little box of filters sits directly in the flight path and is capable of dropping pieces of aluminum of various thicknesses into the path. The problem is that they appear to be stuck.
The real problem is that we didn't notice this until everything else was already put together. The day shift had just finished aligning the slits down stream of this filter box when we came in. As we assumed that they'd already checked the filters up stream (everything proceeds in a down stream fashion in this business), we didn't bother attempting to use them. Or maybe they had been tested and were shown to have worked, but failed in the intervening time. Now that we're done with the alignment and ready to go we attempted to open things up and discovered this little problem. The part that really hurts and just makes me want to cry is that to fix the filterbox, we will probably end up destroying the careful positioning of the slits just downstream... and sadly, that will nuke the positioning of every subsequent piece of hardware.
It's still a little too early to imagine realigning again, so we're working as best as possible to attempt to diagnose and fix this little pneumatic filterbox. wish us luck.
Our group has a week of time at beamline 12 and will be performing a series of x-ray reflectivity experiments from various materials.
What follows is a rather dry post. However, it's one of the parts of this work and serves to justify how unglamorous some of this can be.
Currently we are suffering the virtually inevitable problem of alignment, or in other words getting the tiny beam of light conditioned and in exactly the correct place. Depending upon the kind of experiment you wish to perform, the alignment (or geometry) needs varying degrees of rigor. For some experiments, with a large sample and where the angle of the scattered light is not very important, the experiment can be done without careful alignment. Our experiments tend to be very sensitive to the various angles and positions, making us spend a good deal of effort to align everything. All of this is further complicated because every piece of equipment that interacts with the beam sits on something. Each piece may not be firmly attached, or may not be completely centered. Even the concrete floor can occasionally give a little shift (I've heard stories of the operators "detecting" earthquakes by variation in the electron beam).
What all has to be aligned and how do we do it you might ask...
The first part is largely out of control of the users (which isn't always a bad thing) and that is the light source itself. In this case we have a very large magnetic field bending the flight path of the electrons. As they follow a curve they release a continuous spectrum of light. This light tends to be brightest in the forward direction of the beam and already fairly small and tight. By changing the magnetic fields on the electron stream we could alter the initial direction of the small spotlight of photons coming out. However, direct manipulation of the electron beam by users (especially in some cases) can be a very bad idea leading to all sorts of problems. The worst of these would be total electron beam loss effecting all the other users in the facility.
So we have little control over the light source itself. It brings a beam of light down a pipe to a monochromator. This ingenious device allows us to select out a particular photon energy, or at least a very narrow band of photon energies, eliminating all the unwanted light. It also can (and does) serve a second purpose in most cases by acting as a focusing element for the beam. All of this sits down from us by about 20 meters. From there the light travels down a long pipe and into the experimental hutch.
Inside the hutch are the last few meters of the photon flight path, and incidentally the most important ones. The beam is conditioned further by a secondary focusing element and then narrowed by the use of several slits. The focusing serves to get the most light in the area of interest and the slits further work to eliminate unwanted stray photons. It also allows is to correct for minor the beam intensity variation.
So from the light source the photons travel first through the monochromator, then 20 meters and through more focusing elements and slits before all before arriving at the sample. It would be simple if we just had a place saying "put sample here," however that doesn't quite work. First there is the inevitable drift and change that occurs in the position of the beam from adjusting any of the previous instruments I mentioned. So "center" is a moving target every time (virtually) that an experiment is set up. We locate the approximate center by placing a small piece of paper that "burns" in the x-ray beam. By seeing where the make is made on the paper we can then get the diffractometer(the device that holds both the sample and the detector in place) into approximate position. From there is gets a little tricky.
Remember, we can't just see any of the photons with our eyes and being inside the hutch with the x-rays is right out. And even having the sample sitting in the center isn't enough. Each axis that the sample rotates about must be aligned and centered. We have various jigs and pins, precisely machined, that are used to position and align each axis. And since there is no definitive starting point, once we've aligned each axis and positioned everything once, it's still not fully aligned. Rather it is an iterative process with each step bringing us closer until we're within the tolerances of all the motor positions. Having such devices (which weigh well over a ton) move reliably, repeatably to within sub millimeters is something to behold... well, it is for me anyways.
At this point things should be good and all that's left is to define the detector aperture. This isn't terribly difficult as it only involves (you guessed it) more slits.
Once all this is done, we're then ready to actually move our experiment into the hutch and observe what happens. If we get lucky and things are reasonably placed from the pervious users, then it's only a few hours. Our record is around 4-5 hours I believe(and that's not because we're slow, but that's just how long it takes to systematically check everything and make minimal adjustments). At times we've gone a couple of days before working out all the problems. As that can be a couple of days out of our entire week without even placing our experiment in the hutch, it gets to be a major frustration and obstacle. I seriously wonder what the cost-benefit analysis would say regarding how much time and effort (which means money) that we waste each time because we have to take such pains (as many if not most users must do). Some of the equipment is certainly "top of the line" but much of it is "lowest bidder" equipment. How much could we have saved (and how much easier would it be for us) had the initial investment been higher?
Anyhow... We're reasonably lucky today and have an aligned diffractometer as of about 3am.
About the time I typed this we realized we might not be so lucky. I also left out a piece from the earlier writing... There is one other component in the flight path of the photons, a filter box. While for our experiments we need as many photons as possible, it is often necessary to eliminate some or most of the photons at various times for what I'll call diagnostic purposes. Usually this means checking our alignment or the orientation of a crystal or some such thing.
This little box of filters sits directly in the flight path and is capable of dropping pieces of aluminum of various thicknesses into the path. The problem is that they appear to be stuck.
The real problem is that we didn't notice this until everything else was already put together. The day shift had just finished aligning the slits down stream of this filter box when we came in. As we assumed that they'd already checked the filters up stream (everything proceeds in a down stream fashion in this business), we didn't bother attempting to use them. Or maybe they had been tested and were shown to have worked, but failed in the intervening time. Now that we're done with the alignment and ready to go we attempted to open things up and discovered this little problem. The part that really hurts and just makes me want to cry is that to fix the filterbox, we will probably end up destroying the careful positioning of the slits just downstream... and sadly, that will nuke the positioning of every subsequent piece of hardware.
It's still a little too early to imagine realigning again, so we're working as best as possible to attempt to diagnose and fix this little pneumatic filterbox. wish us luck.