Background Considerations in Radio Flow-through Detectors

In an accompanying discussion of detectability, we demonstrate how important background counts can be. While the user of a modern instrument has little opportunity to affect background, primarily because it is inherently so low, an explanation of the factors which have made it so, should prove useful as a part of an overall understanding of the flow-through method.

We begin with a reminder that today's commercial radio flow-through detectors are conceptually coincidence-type liquid scintillation counters in which the discrete sample handling mechanism has been replaced by a transparent flat coil or several interconnected straight tubes through which sample solution continuously flows. Nowadays, the sample is almost always an HPLC eluate; rarely does it have another origin. Most often we are dealing with beta-emitters, occasionally with a soft-gamma; these instruments can also be used quite satisfactorily for alpha- and positron- emitters but such use is rather infrequent. There are other detectors dedicated to more energetic gamma-emitters, but they will not be treated here.

As in a liquid scintillation counter, flow-through detectors employ two photomultipliers which face one another. In the liquid scintillation counter the photomultipliers are separated by perhaps 1-1/2" to allow the conventional 20 mL sample vial with its 1-1/8" diameter to be interposed. However, since the photomultipliers of a flow-through detector normally sandwich a flat coil of Teflon or Tefzel tubing, they are typically separated by less than 1/2". Unfortunately, bringing the photomultipliers close together to maximize light collection, hence counting efficiency, potentially results in increased background as the spacing is reduced.

Perhaps it would be best to list the sources of "background" in these detectors and discuss how each is best minimized. But first, we had better define "background". True background derives from ambient radioactivity -- traces of long-lived gamma-emitters in the bricks of the building or the concrete floor, 40K in the glass of the photomultipliers and the sample vials, cosmic rays, traces of uranium in lead shielding, etc. Much of such background, if the source is external to the detector, can be absorbed and prevented from interfering with the detection process by the use of dense shielding material.

There are also spurious counts which are not radioactive background and for which shielding does no good. They arise from photomultiplier noise and photomultiplier "crosstalk" -- light produced within one photomultiplier that is seen and recorded by both. The effects of noise are largely eliminated by coincidence counting; crosstalk can be reduced by selecting appropriate photomultipliers and operating them under careful control. Even so, crosstalk is by far the largest component of the background of a flow-through detector.

Photomultiplier Noise -- The two photomultipliers used in modern liquid scintillation and flow-through counters each have noise rates between 200 and 1000 pulses per second, with larger tubes tending to be the noisier. For an event to be declared, the coincidence technique requires that both PMs see light from a scintillation event at about the same time -- typically within 20-50 nanoseconds. Then, pulse size is examined and, should it fall within an acceptable range, that event is recorded. Individual PM output pulses from one tube, not matched in time by a pulse from the other, are not counted.

However, the noisier the tubes, the greater the chance that both will pulse within the coincidence time giving rise to a false count ("accidental") which, when no activity is present, increases the apparent background. Two random noise generators -- for this discussion the PMs may be thought of on these terms -- generate accidentals according to the expression:

A = n1n2T
where A = accidental rate
n1, n2 = PM noise rates
T = coincidence time

Whether 200 or 1000 pps, with fast coincidence circuitry, PM noise is not a consequential problem; at 1000 pps, with a 20 ns coincidence time, the calculated accidental rate is just 2.4 pulses per minute.

Radioactive Background -- In a flow-through detector, the background contribution from all usual sources of ambient background, including cosmic rays, is also not especially significant. Depending upon operating conditions, a typical liquid scintillation counter has a background for 10- 15 mL of counting solution in a 20 mL vial from perhaps 5 to 20 cpm, about half of which is attributable to ambient activity and half to photomultiplier interactions. With the most common flow-through cells having volumes from 500 ul to 1 mL, the target for ambient radioactivity is so much less, with corresponding reduction in background contribution. Therefore, even with less shielding, unless the environment is unusually bad, ambient radioactivity only contributes a count or two to background.

Photomultiplier Crosstalk -- The principal source of "background" in a flow- through detector is a consequence of light being generated within one PM which is seen by the photocathodes of both to give a true coincidence which is then counted. Perhaps as much as 90% of the background in a flow- through detector is generated in this way. Where does this internally produced light originate and can anything be done about it?

There is more than a single source but there is a common thread; light production is more prevalent in large PMs rather than small ones. Surely the greatest problem is the 40K content of the glass envelope. When a beta-particle, such as that from a 40K- decay event (Emax = 1.34 meV), passes through a medium of high refractive index faster than the speed of light in that medium, it gives rise to Cerenkov radiation, light photons propagated perpendicular to the direction of particle motion. This blue light, known to many as the glow seen around the fuel elements of a swimming pool reactor, falls within the wavelength region where the PMs are most sensitive. When both pulse st the same time, an additional count is recorded.

In a comparison of the 1" and 2" photomyltipliers in current use in flow-through detectors, a 2" diameter photomultiplier has a faceplate and a base each with four times the area of those of the 1" PM and therefore four times the potential for Cerenkov. Also, the barrel of the 2" tube also has twice the circumference and is almost an inch longer than the smaller tube. All of this translates to a surface and a weight of glass that is nearly three times as large for the 2" tube as for the 1"; one might expect a corresponding increase in Cerenkov counts.

In addition to its beta, 40K is a hard gamma emitter (1.46 meV). These gammas too, activate the photocathodes as demonstrated by counting coincidences first with an opaque cardboard placed between the PMs and then with a Pb sheet. When the faces are close together, the decrease in counts from cardboard to lead is dramatic. The accepted explanation is that accidentals are being counted when the lead is present, accidentals plus the effects of the gammas are observed when there is no lead. Again, the larger the tube, the more 40K, and the more such counts.

And there are still other volume related sources of spurious counts. Though there is an extremely high vacuum within the PM, there still remain many gas molecules. These occasionally interact with the electron avalanche in the region of the anode and become ionized. The ions migrate to the cathode where they are discharged, sometimes producing light. With the volume of the 2" tube somewhat greater than four times that of a 1" tube, we again anticipate increased crosstalk.

While there are electronic means for eliminating some of the effects of photomultiplier crosstalk, they are rather complex and tend to reduce counting efficiency, especially for the least energetic isotopes, i.e., 3H. Dependent upon differences in the decay time of light pulses arising from true scintillations within the sample and light generated in plastic scintillator sheeting placed in front of the photomultipliers, they also impose limitations on the scintillators that can be used. It would be better to avoid such methods if possible.

These considerations lead to the conclusion that, especially for flow measurements where the PMs are deliberately placed close together, smaller tubes are to be favored over larger ones. And, the advantage is not only in the lower mass, but also in their lower operating voltage that makes for less noise and less residual gas ionization. While larger PMs do allow for larger sample cells -- and especially for larger counting vials in discrete-sample counters -- with the modern tendency for faster HPLC with slower mobile phase flows, the general trend is toward smaller cells for which there can be no argument for large photomultipliers.