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.

138 TIME-DOMAIN LIFETIME MEASUREMENTS<br />

Table 4.5. Resolution <strong>of</strong> a Two-Component Mixture <strong>of</strong> Anthranilic Acid and 2-Aminopurine,<br />

Observed at a Single Wavelength, Using Time-Domain Data<br />

Observa- Pre-expo- χ R 2<br />

tion nential Fractional Number <strong>of</strong><br />

wave- Lifetimes (ns) factors intensities decay times<br />

length<br />

(nm) τ 1 τ 2 α 1 α 2 f 1 f 2 2 a 1 a<br />

360 6.72 11.58 0.223 0.777 0.143 0.857 1.01 2.77<br />

380 7.51 11.51 0.404 0.596 0.306 0.694 0.85 3.81<br />

400 7.22 11.08 0.448 0.552 0.347 0.653 0.91 4.51<br />

420 7.99 11.22 0.595 0.405 0.511 0.489 0.93 2.95<br />

440 8.38 11.91 0.763 0.237 0.692 0.308 0.89 2.72<br />

aRefers to a two- or one-component fit.<br />

times were recovered at each emission wavelength. This<br />

suggests that the actual uncertainties in the recovered lifetimes<br />

are larger than expected from the ASEs, and then<br />

seem to be even larger than calculated from the χ R 2 surfaces.<br />

Furthermore, the differences in the lifetimes recovered<br />

at each emission wavelength seem to be larger than<br />

expected even from the χ R 2 surfaces. This illustrates the difficulties<br />

in recovering accurate lifetimes if the values differ<br />

by less than tw<strong>of</strong>old.<br />

Another difficulty is that the recovered amplitudes do<br />

not follow the emission spectra expected for each component<br />

(Table 4.5 and Problem 4.5). As the lifetimes become<br />

closer together, the parameter values become more highly<br />

Figure 4.51. Lifetime χ R 2 surfaces for the two-component mixture <strong>of</strong><br />

AA and 2-AP (Figure 4.50). The horizontal bars show the asymptotic<br />

standard errors. From [187].<br />

correlated, and it is difficult to know the true uncertainties.<br />

This is not intended to be a criticism <strong>of</strong> the TD measurements,<br />

but rather is meant to illustrate the difficulties inherent<br />

in the analysis <strong>of</strong> multi-exponential decays.<br />

4.10.6. Global Analysis: Multi-Wavelength<br />

Measurements<br />

One way to improve the resolution <strong>of</strong> closely spaced lifetimes<br />

is to perform measurements at additional wavelengths,<br />

and to do a global analysis (Section 4.12). The concept<br />

<strong>of</strong> global analysis is based on the assumption that<br />

decay times are independent <strong>of</strong> wavelength. The decay<br />

times are global parameters because they are the same in all<br />

data sets. The amplitudes are non-global because they are<br />

different in each data set. Global analysis <strong>of</strong> the multiwavelength<br />

data results in much steeper χ R 2 surfaces (Figure<br />

4.52), and presumably a higher probability <strong>of</strong> recovery<br />

<strong>of</strong> the correct lifetimes. The lifetimes are determined with<br />

higher certainty from the global analysis because <strong>of</strong> the<br />

steeper χ R 2 surfaces and the lower value <strong>of</strong> F χ with more<br />

degrees <strong>of</strong> freedom (more data). As shown in Problem 4.5,<br />

the amplitudes (Table 4.6) recovered from the global analysis<br />

more closely reflect the individual emission spectra than<br />

the amplitudes recovered from the single wavelength data.<br />

4.10.7. Resolution <strong>of</strong> Three Closely Spaced<br />

Lifetimes<br />

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

difficult as the number <strong>of</strong> decay times increases. This difficulty<br />

is illustrated by a mixture <strong>of</strong> indole (In), anthranilic<br />

acid (AA), and 2-amino purine (2-AP) (Figure 4.53). At<br />

380 nm all three fluorophores emit, and the decay is expected<br />

to be a sum <strong>of</strong> three exponentials.

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

Saved successfully!

Ooh no, something went wrong!