Estimation of detection limits

Kirchner, C. J. Estimation of Detection Limits for Environmental Analytical Procedures, In Currie, L. A., ed. Detection in Analytical Chemistry Importance, Theory and Practice. American Chemical Society Washington, DC, 1988. 

As shown above, limits can be estimated on two ways from blanks and from cabbration data. Modern analytical methods such as spectrometry and chromatography use a third way to an increasing extent the estimation of detection limit by means of the signal-to-noise ratio SNR. 

When a 350 ml seawater sample was spiked with 54Mn and taken through the chelation, extraction, and back-extraction procedures, the observed recovery of the radio-tracer was 100.6%. Estimates of detection limits for manganese based on sets of both shipboard and shore laboratory separations are of the order of 0.1 nmol/1. The accuracy of the technique is demonstrated by data from the ICES fifth-round intercalibration exercise for trace metals in seawater [449 ]. 

Experimental Determination of Optimum Detection Potential and Estimation of Detection Limit... 

Again, experiments as described in Experimental Determination of Optimum Detection Potential and Estimation of Detection Limit will help to determine the optimum mobile phase composition with respect to both separation and detection. The pH variations may strongly influence both the background current and the analyte electrolysis current (signal), see Figure 4-2. Therefore mobile phases should preferably be buffered. 

L.B. Kish, M. Cheng, J.U. Kim, S. Seo, M.D. King, R. Young, A. Der, G. Schmera, "Estimation of Detection Limits of the Phage-Invasion Based Identification of Bacteria", Fluct Noise Lett, 5 (2005) L105-L108. 

Estimation of Detection Limits for Environmental Analytical Procedures... 

Estimates of systematic errors, such as Interference, should be Included in the estimate of detection limit. 

Adequate sensitivity should be demonstrated and estimates of the limit of detection (LOD) and the limit of quantitation (LOQ) should be provided. The slope of the calibration line may indicate the ability of the method to distinguish the tme analyte concentration. The LOD of a method is the lowest analyte concentration that produces a reproducible response detectable above the noise level of the system. The LOQ is the lowest level of analyte that can be accurately and precisely measured. For a regulatory method, quantitation is limited by the lowest calibration standard. The techniques for these estimations should be described. 

But the main advantage of the SNR concept in modern analytical chemistry is the fact that the signal function is recorded continuously and, therefore, a large number of both background and signal values is available. As shown in Fig. 7.9, the principles of the evaluation of discrete and continuous measurement values are somewhat different. The basic measure for the estimation of the limit of detection is the confidence interval of the blank. It can be calculated from Eq. (7.52). For n = 10 measurements of both blank and signal values and a risk of error of a = 0.05 one obtains a critical signal-to-noise ratio (S/N)c = fo.95,9 = 1.83 and a = 0.01 (S/N)c = t0.99,9 = 2.82. The common value (S/N)c = 3 corresponds to a risk of error a = 0.05... 0.02 in case of a small number of measurements (n = 2... 5). When n > 6, a... 

Lee [42] determined pentachlorophenol and 19 other chlorinated phenols in sediments. Acidified sediment samples were Soxhlet extracted (acetone-hexane), back extracted into potassium bicarbonate, acetylated with acetic anhydride and re-extracted into petroleum ether for gas chromatographic analysis using an electron capture or a mass spectrometric detector. Procedures were validated with spiked sediment samples at 100,10 and lng chlorophenols per g. Recoveries of monochlorophenols and polychlorophenols (including dichlorophenols) were 65-85% and 80-95%, respectively. However, chloromethyl phenols were less than 50% recovered and results for phenol itself were very variable. The estimated lower detection limit was about 0.2ng per g. 

Unlike the preceding case where B and A were known, we now have only an estimated x-detection limit, even though y is (almost) known. ("Almost , because of the x-dependence of Oy. Though this is a second order effect when there is a high threshold, it would be preferable in this case to estimate Oy as a function of y rather than x. Then y would be quite independent of B and A.)... 

Equation 8.3 relies on the assumption that the calibration parameters hold to the limit of detection, which is not necessarily correct. If the calibration is taken over a very wide range, for example in some element analyses by inductively coupled plasma, the uncertainty is not constant but is proportional to the concentration. In this case it is not possible to estimate the detection limit by equation (8.3). 

Detection limit. Low concentrations of Ni-EDTA near the detection limit gave the following counts in a mass spectral measurement 175, 104, 164, 193, 131, 189, 155, 133, 151, 176. Ten measurements of a blank had a mean of 45 counts. A sample containing 1.00 pM Ni-EDTA gave 1 797 counts. Estimate the detection limit for Ni-EDTA. 

C. The atomic absorption signal shown below was obtained with 0.048 5 pg Fe/mL in a graphite furnace. Estimate the detection limit for Fe, defined for this problem as the concentration of Fe that gives a signal-to-noise ratio of 2. 

J. Vial and A. Jardy, Experimental Comparison of the Different Approaches to Estimate LOD and LOQ of an HPLC Method, Anal. Chem. 1999, 71, 2672 G. L. Long and J. D. Winefordner, Limit of Detection, Anal. Chem. 1983,55, 713 A W. R. Porter, Proper Statistical Evaluation of Calibration Data, Anal. Chem. 1983,55, 1290A S. Geiss and J. W. Einmax, Comparison of Detection Limits... 

A short list of rough estimates of the limits of detection (LODs) for 49 elements in ICP-AES and ICP-MS is given in Table 1. 

Raman spectroscopy provide complementary information to IR in that bands that are inactive in IR often are active in Raman, and vice versa. The Raman spectrum of 1 consists of three bands with relatively high intensity [37, 38], while the corresponding spectrum for 3 has two strong and one weaker line [37]. Thus, in contrast to IR, Raman provides a tool for fingerprinting of both molecules even without the use of isotopic labeling. In addition, the detection limit is expected to be lower in Raman than IR due to the very low intensities of the IR transitions. We have recently estimated the detection limit for 1 in liquid and solid nitrogen based on experimental measurements and computed Raman intensities for 1 and N2. The following expression was used to compute the detection limit ... 

The detection limits of individual PCAH will be determined by injecting known amounts into a flame. An estimate of the limit for pyrene can be obtained from the atmospheric pressure cell measurements. Pyreene was detected in the cell at 50°C and 1 atm of air, where its vapor pressure is about 0.1 milli-torr, or about 0.1 ppm. While this limit can be improved by optimization of the optics and electronics, it is sufficient to detect pyrene in the concentrations expected from probe measurements. 

The standard personal sampler operates at a flow rate of about 21 min-1 and newer models will sample at up to 41 min-1. Thus, in a two-hour period, 240 to 4801 or about 0.25 to 0.5 m3 can be collected. We have used a 0.25 m3 sample volume to estimate the detection limits for metals in air. The dissolution method in Section II. B. 1 should be modified for analysis of workplace samples use a 10 ml volumetric flask for the final volume, and additions of 0.2 ml of the ionization buffer and releasing agent solutions when required. 

Table 7-1. EPA Method 1668-Estimated Method Detection Limits (EMDL) and Estimated Minimal Levels (EML) of Selected PCB Congeners ...

---