What's that noise?
29 January 2008
Keith Hall of LabLogic Systems tracks down the source of background during counting of low-level radioactivity in RHPLC eluates, and describes an effective way of keeping it to a minimum that reduces counting time.
Amongst radiochromatographers, the terms 'noise' and 'background' are used interchangeably to refer to the radiation recorded by the detector that is extraneous to the samples under investigation and must therefore be eliminated from calculations.
Talk of 'background' is misleading because it implies that this detected but irrelevant radiation originates from sources distant from the focus of investigation, such as other equipment operating in the laboratory, but in fact this is far from the truth.
Noise arising from electronic equipment - even from a single photo multiplier tube in a radio flow detector - is so low as to be unmeasurable; and radioactivity present in an ambient state or from the action of cosmic rays does not register much higher up the scale. The emissions of radio isotopes are measurable of course, but only strong types such as 40K gamma are detectable at any distance; tritium, one of the most commonly used isotopes in ADME studies, travels less than 6mm in air.
Paradoxically, almost all the 'background' measured by the radio flow detector in a radio HPLC (RHPLC) system originates in the foreground and very close to home, because it is generated by the detector's own photo multiplier tubes (PMTs).
To distinguish valid data from noise, the detector has two PMTs which must respond to the same beam of light within 20 nano-seconds for the signal to be recognised as valid. Logically, it is also possible for both tubes to respond in the same time frame but to different events (a phenomenon described as cross-talk), and it is this that accounts for 90 per cent of what we call noise or (very inaccurately, as we can now see) background.
RHPLC detector manufacturers use various means of excluding cross-talk to reduce background, which in turn ultimately improves Limits Of Detection (LOD). The most successful in this respect (and the most sensitive) is the β-RAM from IN/US Systems, which has photomultiplier tubes with a diameter of just 31.75mm (1¼ ins) to maximise signal-to-noise ratios. By blanking off a proportion of the PMT phases from each other, this arrangement achieves ultra-low backgrounds - typically <4cpm for 3H and <1cpm for 14C, with corresponding counting efficiencies of >60% and >90% respectively.
To show what this means in practical terms, consider that the tables below which show the relationship between background levels and the limit of detection (in other words, the extent to which a detector with a low background reduces counting time for each sample).
Table 1
| Efficiency (%) | Background (cpm) | Desired LOD (dpm) | Residence Time (seconds) |
| 100 | 30 | 30 | 53.5 |
| 100 | 20 | 30 | 38.9 |
| 100 | 10 | 30 | 11.1 |
In Table 1, we want to achieve a LOD of 30 dpm. To achieve this with a background of 30 cpm we have to count the sample for 53 secs ('Residence Time'); but by progressively lowering the background, we can reduce the residence time to 11 secs.
Table 2
| Residence Time (seconds) | Critical Level Lc (dpm) | Limit of Detection Ld (dpm) | Limit of Quantitation Lq (dpm) |
| 7.5 | 20.84 | 63.27 | 819.52 |
| 15 | 14.74 | 40.25 | 419.09 |
| 45 | 8.51 | 20.59 | 150.99 |
Conversely, Table 2 shows that by increasing residence time we can lower the Limit of Detection. For example, with a typical residence time of 7.5 seconds (which is achieved using a 500mL flow cell with a 1mL/min eluate flow) the limit of detection is 63 dpm, but if you count for 45 seconds this can be reduced to 21 dpm. More importantly, at the longer residence time the Critical Level (at which we can see there's something there) is just over 8 dpm.