Mossbauer Setup

Setup Details and Pre-Calibration Exercises

Once the apparatus is correctly assembled, there remain a number of factors that may affect the quality of data that can be obtained. First, the source, sample and detector should be arranged to maximize both count rate and data quality. Next, an appropriate PX2T gain must be selected. Finally, the SCA must be appropriately calibrated to maximize the number of real 14.4 keV events observed (see the next page on the 14.4 keV peak). Higher count rates are important because the increase the amount of data that can be collected in the available time, leading to better statistics.

Source and Sample Position

Naturally, placing the source closer to the sample and/or the sample close to the absorber will increase the count rate. Unfortunately, placing the source too close to the sample leads irregular trends in the data, and gives poor data. Why is this?

This is because the Cobalt 57 source emits photons radially in all directions. When we observe the velocity signal with the DAQ, and then use this velocity to calculate a Doppler shift and then the total energy of a photon emitted at that velocity, we are assuming that the photon is emitted parallel to the oscillation. (That is, that the photon is traveling straight at the sample.) When the source emits photons at significant angles to the horizontal, the velocity signal can no longer be used to accurately describe their total energy. Placing the source too close to the sample, such as less than an inch away from the near side of a collimating lead brick with the sample on the opposite side, generally leads to poor data, such as that seen below:

The photons emitted at large angles from the horizontal can be blocked with a collimator that only allows a narrower cross-section of source photons to hit the sample. But, blocking too many photons means that data collection takes much longer. Also, very small collimators are difficult to align. The best answer appears to be placing the sample as close as possible to the detector (if it's a foil, it can be flush with the brass muzzle on the detector), using a medium- or large-sized (about 3/4" or greater) collimating lead brick, and placing the button source at least two inches away from the side of the collimating lead brick nearest to it. This lets many 14.4 keV photons from the Co 57 source interact with the sample, but deflects photons with extreme angles. Count rates of around 80 to 100 cps should be observed with this setup when no source is present.

Effect of PX2T Gain Setting

Larger PX2T gains give slightly higher count rates. However, increasing the PX2T gain too much easily exceeds the range of the SCA (which tops out around 7.6 V). A gain setting of at least 3 but not more than 5 is recommended.

How the Ortec 550 SCA Works

The purpose of the SCA is to generate a TTL pulse that acts as a trigger signal whenever a 14.4 keV photon is detected. Therefore, the SCA window settings must be properly adjusted through a calibration process. If this is done correctly, the SCA will include only the 14.4 keV peak in the window. The SCA should be set to 'WINDOW' mode using the toggle switch on the front panel. For the actual Mossbauer experiment, the back panel toggle switch should be set to 'INT' or internal, and no external voltage should be applied. In this mode, the 'LOWER LEVEL' dial selects the lower-level voltage of the window region (range: about +20mV to +10V). The 'UPPER LEVEL OR WINDOW' dial selects the window width, ranging from about 0 to 1 V. Since the upper dial scrolls from 0 to 10 (coarse) and from 0 to 100 (fine), there is a factor of 10 difference between the upper dial setting and the real window width. For example, selecting 2 (coarse) and 50 (fine) on the upper dial results in a window width of 0.25 V.

Note: For SCA calibration (see 14.4 keV Peak), the back panel toggle switch should be set to 'EXT' or external.

How the Ortec 994 Counter Works

The Ortec 994 Counter/Timer module offers a quick and crude way to verify that the SCA is outputting pulses that generally describe the Co 57 spectrum. However, the most obvious drawback to the counter is that the count rates are NOT ACCURATE. The counter module is extremely sensitive to reflections, and to the small oscillations that may occur immediately after a logic pulse (such as from the SCA) has dropped from high to low. For example, when the detector is at a distance where the LabVIEW program reports ~40cps, the count rate on the 994 counter is largely dependent on how the input signal is (or isn't) teed and is (or isn't) terminated.

  • When the SCA output is connected directly to the 994 (isn't teed at either module, doesn't connect to the DAQ at all), and has no terminator, the 994 reports ~160 cps.

  • When the SCA output is connected directly to the 994 and does have a terminator, the 994 reports about 45 - 50 cps (MOST ACCURATE).

  • When the SCA output is teed (one side goes to the DAQ, one side goes to the 994) and there are terminators at both the DAQ and the 994, the 994 reports ~95 cps.

  • When the SCA output is teed (one side goes to the DAQ, one side goes to the 994) and there is a terminator at the DAQ only, the 994 reports ~170 - 200 cps (LEAST ACCURATE).

  • When the SCA output is teed (one side goes to the DAQ, one side goes to the 994) and there is a terminator at the 994 only, the 994 reports ~90 cps.

Additionally, how the settings on the 994 counter work is somewhat non-obvious. The 'A' channel on methods lab 994 modules has been set to a timer mode, while the 'B' channel works as a counter. All input SCA signals should go into channel 'B'! The time base should be set to seconds, not minutes. How to adjust the counting interval may not be immediately clear: DISPLAY -> PRESET, then use the PRESET advance/select controls to select 013 giving 01x10^3. With a time base of 0.01 sec, and this results in a 10-second counting interval. If the display is set to channel A (and hit 'COUNT'), the timer will count off seconds up to 10 for verification purposes. Finally, place DISPLAY -> B to view the observed input count in real time. (Also, every time the NIMBIN is powered down, these settings are erased.) The 'DWELL' dial adjusts how long the counter will display the resulting count. Lastly, the threshold at which the counter counts a signal as an event is determined by a 25-turn trimpot above the channel input. Be warned that the trimpot test points are digital. High TTL signals will be observable out of the test points when input signals begin surpassing the threshold. If any adjustment needs to be made to the counter, a pulse generator should be used to input known amplitude pulses while using an oscilloscope to monitor the test point output.

So, to reiterate, what the 994 counter IS good for is viewing relative count rates coming out of the SCA (see pre-calibration exercise #2), and not much else!

Pre-Calibration Exercise #1: Examining the Cobalt 57 Spectrum with an MCA (EASY-MCA)

Understanding the spectral features of Cobalt 57 is useful when preparing to calibrate the SCA window width and window center voltages. Importantly, the spectrum contains a number of peaks, some of which are more prominent than the 14.4keV peak, and not all peaks result from Co 57 or Fe 57 directly (some peaks come from isotopes in alternate decay pathways, and some from naturally abundant radioactive elements). It is also useful to note which of the available weak gamma sources are helpful to view in comparison; the Barium 133 and Americium 241 both contain peaks in the low keV region and are recommended.

Viewing the spectrum of the Cobalt 57 source with a NaI detector (and Digibase) allows the identification of the 122 and 136 keV peaks (red regions). This result is shown in the spectrum below - note that the 122 and 136 keV peaks overlap. However, the efficiency curve of NaI detectors show that the efficiency drops off sharply for energies below ~50 keV. Therefore, the important 14.4 keV peak is essentially not viewable using a NaI detector.

It is somewhat more useful to study the spectrum produced by the silicon detector (XR-100CR), since these are the exact signals that will eventually enter the SCA. Here, the detector signal has been amplified with the dedicated amplifier (PX2T) with a gain setting of 3 and then sent to the input of the EASY-MCA. The silicon detector is highly efficient in the low keV range, and the 14.4keV peak is easily identified (red region) in the spectrum below. Note the relative prominence of the two lower-energy peaks!

Pre-calibration Exercise #2: Using the Counter to "See" the Spectrum

You can use the counter to roughly plot the spectrum of Co 57, and familiarize yourself with the voltages where the significant peaks occur. The result should resemble the spectrum you saw using the MCA. Here's how:

1. Use the Silicon detector to view the Co 57 source. The output from the detector's PX2T amplifier should be connected to the 'DC Input' of the Ortec 550 SCA. The SCA output should be sent to the 'IN B' input of the Ortec 994 counter only using a 50 Ohm terminator.

2. The gain setting of the PX2T plays a large role in the placement and width of the 14.4 keV peak, and it is possible to use so large a gain (5 or greater) that the 14.4 keV peak is no longer easily detected. A gain setting of 3 to 5 is recommended.

3. The SCA should be placed in 'window' mode using the 4-way switch on the front panel. In this mode, the WINDOW OR UPPER LEVEL dial adjusts the window width (range: 0 to 1 V), while the LOWER LEVEL dial adjusts the center of the window (range: ~0 to 10 V).

4. For the 994 counter, a 10-second counting interval is a reasonable setting. How to adjust the counting interval may not be immediately obvious: DISPLAY -> PRESET, then use the PRESET advance/select controls to select 013 giving 01x10^3. With a time base of 0.01 sec, this results in a 10-second counting interval. Finally, place DISPLAY -> B to view the observed count in real time.

5. Beginning with the window center bottomed out near 0 V and the window width around 0.10 or 0.05 V, slowly raise the peak center while observing the count rate to get a feel for the peak locations. Recall the Co 57 MCA spectrum (Silicon detector), which has a range from 0 to 10 V. If a peak is located on the MCA at around, say, 4 V, you should expect to see a peak with the counter at roughly the same place (provided the PX2T gain setting is unchanged). As expected, you should observe one extremely large peak (i.e., very fast count increases on the 994 counter) next to a moderately-sized peak at a relatively low energy, a period of relatively low count rates, and then another moderately-sized peak. This last peak is the 14.4 keV peak - do not confuse it with the extremely high count/low-energy peak! Compare to the MCA spectrum if in doubt. Also note that the two lower-energy peaks may be difficult to distinguish from each other when the window width is larger. Narrowing the window width will provide more detail about spectrum features.

6. Once you have determined roughly where each peak is located, step the window center through small increments and record the count rates at each window center location. Plot the window center vs. count rate data in Excel; it should resemble the MCA spectrum. With a window width of 0.05 V or finer, take center vs. count rate data in the regions of the low-energy and 14.4 keV peaks. Plot the data to view the peaks in greater detail. Be sure that if your window width is very small, that your window center steps are correspondingly small. Examples are shown below; full spectrum (top), lower-energy peak detail (bottom left), and 14.4 keV peak detail (bottom right).

NOTES:

  • At low gain settings, like 2 and below, it can be difficult to distinguish the two lower-energy peaks from one another (essentially because they are 'squished' together).

  • Anomalously high count rates may be observed when the window center is either bottomed out near zero and/or saturated, which occurs at about 7.6 V. These aren't peaks and you can ignore these areas.

  • Make sure your window width and step size are appropriate to each other, and that the input to the counter has a 50 Ohm terminator.

Now that the experimenter is completely familiar with the Cobalt 57 spectrum, a 'scanning' technique with the SCA described in the next section will allow for optimum calibration.

Next, The 14.4 keV Peak