22.07.2013 Views

Principles of Fluorescence Spectroscopy

Principles of Fluorescence Spectroscopy

Principles of Fluorescence Spectroscopy

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

180 FREQUENCY-DOMAIN LIFETIME MEASUREMENTS<br />

Figure 5.29. χ R 2 surfaces for the two-component mixture <strong>of</strong> p-terphenyl<br />

and indole. The horizontal arrows show the asymptotic standard<br />

errors (ASE), and the dashed lines are at the appropriate F χ values.<br />

From [116].<br />

dom (ν = m – p). It is useful to consider some representative<br />

values for the ratio <strong>of</strong> χ R 2 values. For the two-component<br />

mixture <strong>of</strong> p-terphenyl and indole there are 29 frequencies<br />

and 58 datapoints. The two-decay-time model has three<br />

variable parameters. The F statistic for p = 3 and m = 60 can<br />

be found from Table 4.4 and is 1.19. Hence, the F P value<br />

used for the confidence interval <strong>of</strong> each parameter is 1.06.<br />

As described in Section 4.10.3, there is some disagreement<br />

in the statistics literature as to the exact equation for finding<br />

the confidence intervals. We will use eq. 5.24, but its correctness<br />

has not yet been proven.<br />

The χ R 2 surfaces for the two-component mixture are<br />

shown in Figure 5.29. The confidence intervals (CI) are<br />

determined from the intercept <strong>of</strong> the χ R 2 surfaces with the<br />

χ R 2 ratio appropriate for the number <strong>of</strong> parameters and<br />

degrees <strong>of</strong> freedom. The arrows in Figure 5.29 show the<br />

asymptotic standard errors. The ASEs are about tw<strong>of</strong>old<br />

smaller than the confidence intervals. The decay times in<br />

this mixture are widely spaced. For more closely spaced<br />

decay times it becomes even more important to consider<br />

parameter correlation in the calculation <strong>of</strong> confidence intervals,<br />

and the ASEs can grossly underestimate the true confidence<br />

intervals.<br />

5.7.2. Resolution <strong>of</strong> Two Closely Spaced Lifetimes<br />

The resolution <strong>of</strong> multi-exponential decays becomes more<br />

difficult as the decay times become more closely spaced. It<br />

was previously noted that two decay times spaced by a factor<br />

<strong>of</strong> 1.4 represents the practical resolution limit for double<br />

exponential decay. 12 It is instructive to examine data for<br />

such a mixture because the analysis illustrates the difficulties<br />

encountered at the limits <strong>of</strong> resolution.<br />

To illustrate a sample with two closely spaced decay<br />

times we have chosen the mixture <strong>of</strong> anthranilic acid (AA,<br />

τ = 8.5 ns) and 2-aminopurine (2-AP, τ = 11.3 ns). This may<br />

seem to be an easy resolution, but it is difficult to resolve<br />

decay times which are less than two-fold different. Emission<br />

spectra are shown in Figure 5.30. Frequency-domain<br />

data for the individual fluorophores and for the mixture are<br />

Figure 5.30. Emission spectra <strong>of</strong> a two-component mixture <strong>of</strong><br />

anthranilic acid (AA) and 2-aminopurine (2-AP). Also shown are the<br />

amplitudes recovered from the global analysis (Figure 5.33). From<br />

[116].

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!