Counting statistics and a method for reducing lower limits of detection in LAM-ICPMS analysis

N.J. Pearson and W.L. Griffin, GEMOC, Macquarie

In laser ablation (LA-) ICPMS, the background is measured on the nebuliser gas flow that first passes through the sample chamber and then to the ICPMS. This is routinely done immediately prior to ablation of the sample and the same quadrupole counting parameters are used to measure both the background and the signal from the sample. The precision on the background measurement is the most significant factor in the estimation of the lower limit of detection.

Two parameters that affect the lower limit of detection are the dwell time and the number of replicate measurements: these combined, give the total time each mass is counted. Typical counting protocols for LA-ICPMS involve only one sweep per reading and one reading per replicate, with a large number of replicate measurements (typically 50-150 replicates for 30 analyte masses). Signal intensity is commonly displayed and recorded as a count rate (counts/sec/replicate) rather than as the number of counts detected (counts/replicate). At low count intensities, especially backgrounds of heavy masses (>70), counts/sec/replicate will give an apparent count rate and not the true count rate. The factor relating 'apparent cps' to 'true cps' is inversely proportional to dwell time and introduces significant imprecision into the measurement. If the dwell time is set at 1000 ms and one count is detected then the count rate is 1 cps, however if the dwell time is changed to 50 ms and one count is detected in this 50 ms period, then the apparent count rate will be 20 cps for that replicate. On this basis the counting time for the background should be defined statistically by a suitable number of replicate measurements rather than by counting for a specific time.

A series of experiments was performed in which the dry plasma background was measured on the ELAN 5100 in GEMOC using combinations of dwell time and replicate measurements, in order to establish the relative importance of these two parameters.

Table1: Background precision as a function of dwell time and number of replicate measurements.
Dwell
No of
Background count rates (cps)
time (ms)
replicates
Ca44
Ni60
Sr88
Th232
sd
sd
sd
sd
40
500
314.65
99.94
31.75
31.55
2.75
9.43
1.90
7.35
40
500
309.30
99.47
35.45
34.12
2.40
8.91
1.95
7.07
100
200
310.15
59.50
34.60
22.81
2.85
6.05
1.70
4.38
100
200
314.45
63.83
35.75
22.85
2.05
5.52
2.30
5.08
500
40
319.16
30.79
36.56
9.60
2.04
2.41
2.64
2.47
500
40
311.93
32.55
32.76
8.89
2.84
2.49
2.00
1.81
1000
20
321.75
22.21
34.40
6.18
2.20
1.94
2.05
2.09
1000
20
305.35
18.16
35.16
5.29
2.95
2.44
2.90
1.94

The results of these experiments are given in Table 1 and show that for equal total count times, the precision is much better when long dwell times are used, despite the large number of replicates measured at the shorter dwell times. The calculated lower limits of dectection will be a factor of 5 times lower for the 1000 ms dwell time experiments than those for 40 ms.

The method for quantitative analysis has been modified by measuring backgrounds with extended dwell times at the start, middle and end of each analytical run. Examples of how this method reduces the lower level of detection and improves signal quality will be given. The drift stability of the ELAN 5100 contributes to the success of this procedure.

Back to the GEMOC Abstract Titles Page