Troubleshooting the XRF Detector

Troubleshooting the XRF Detector

Some time ago we received donated equipment through the Laboratory Equipment Donation Program (LEDP, www.osti.gov/ledp/). It was a Kevex Omicron XRF machine. It was in very good shape and after setting up an old Windows 98 computer to run the control software, I began testing the instrument.

The x-ray source seemed to work fine, but I just couldn’t receive any signal from the detector except some spurious counts that didn’t make any sense. I therefore took a closer look at the detector and began troubleshooting. Not having been being familiar with the XRF technique I learned a lot in this process.

The detector works as an energy dispersive spectrometer (EDS). The heart of the detector is a so-called “lithium-drifted silicon, Si(Li)” crystal. Depending on the energy of the incoming X-ray photon more or less charges are created within the crystal, then separated by a high voltage bias, and subsequently counted as individual pulses by a pre-amplifier.

It turns out lithium is highly mobile in Si at room temperature. Li-drifting, however, would substantially deteriorate the detector’s performance and, therefore, the Si(Li) chip needs to be cooled continuously by liquid nitrogen (LN2) to preserve ultimate performance. Thus, Si(Li) detectors mostly (if not always) come as a LN2-dewar/detector assembly (picture).

The necessity to keep the Si(Li) detector at cryogenic temperatures leads to the requirement of vacuum around the cooled surfaces to prevent build-up of ice from atmospheric water vapor. This means the detector cannot be exposed to ambient air, but needs to be housed in a vacuum-tight enclosure. These enclosures are typically made of metal, which is not transparent to X-ray radiation. An X-ray transparent, vacuum-tight window is needed that is also strong enough to withstand rupture due to the 1 atm of ambient pressure resting on it.

Beryllium has a sufficiently high transmission for X-rays and has been long used as a material for X-ray windows. As a thin foil (>8 mu thickness) it still possesses sufficient strength to prevent rupture and also provides a tight vacuum seal. Such a thin Be foil is used as an X-ray window for the Kevex detector (picture).

Back to our Kevex machine. The dewar/detector assembly is equipped with a small vacuum ion pump (picture) to maintain the necessary vacuum (~ 10e-6 mbar or better) around the Si(Li) detector. The ion pump controller is located behind the removable panel on the right side of the instrument. The pressure reading on the controller consistently showed “over pressure” which is of course suspicious. Furthermore, when filling the dewar with LN2 the outside of the vessel became very cold and accumulated frost, indicating that the dewar’s vacuum insulation was failing . My conclusion was that both the detector and the dewar shared the same “vacuum space” and that this vacuum was either really bad or completely gone.

Leak testing was the logical next step. The detector housing has a small port sealed with a plug. This port can be used to evacuate the dewar/detector assembly. To connect a vacuum pump to this port a special connector is necessary (picture). It is a CPC-Cryolab 1/4″ Valve Operator, Model # VO8-084-F (with DN16 flange) or VO8-084-T (with nipple). It can be purchased from Circle Valve Technologies, Inc.

Helium leak testing revealed the beryllium window being the culprit. The Be foil must have suffered a tiny rupture which then had affected the detector vacuum. Beryllium is a carcinogenic material when in micro particle form and, thus, processing should only be done in appropriate facilities, following all safety and health requirements.  Currently, the detector is sent off for repair and I’m awaiting its return to proceed with my instrument testing procedure.

Leave a Reply

Your email address will not be published. Required fields are marked *