Propagation losses are determined for 100 microm x 100 microm square, hollow waveguides constructed from glass capillaries. The small size makes it possible to observe optical effects not easily seen with larger waveguides. The depletion of higher-order even modes creates a large, nonlinear loss. Over a distance of a meter the loss approaches the smaller, linear value expected for the fundamental mode. Additionally, the lowest two even modes beat to produce an oscillatory loss with a period of approximately 2 cm. Making the focal radius 0.35 the waveguide width minimizes these two effects. In a related study, 50-microm waveguides embossed in polydimethylsiloxane are shown to have losses similar to glass capillaries.

Two-photon excited fluorescence detection was performed on a microfabricated electrophoresis chip. A calibration curve of the fluorescent tag beta-naphthylamine was performed, resulting in a sensitivity of 2.5 x 10(9) counts M(-1) corresponding to a detection limit of 60 nM. Additionally, leucine aminopeptidase was assayed on the chip using electrophoretically mediated microanalysis. The differential electroosmotic mobilities of the enzyme and substrate, L-leucine beta-naphthylamide, allowed for efficient mixing in an open channel, resulting in the detection of a 30 nM enzyme solution under constant potential. A zero potential incubation for 1 min yielded a calculated detection limit of 4 nM enzyme.

Two geometries of excitation/emission optics are compared for two-photon excitation of time-resolved fluorescence in capillaries. The test fluorophore is the enzymatic tag β-naphthylamine (BNA). Two 0.25 NA (numerical aperture) microscope objectives are used in the 90° geometry: This configuration has a sensitivity of 1.9 × 1010 count M−1 min−1 and a detection limit of 2.5 nM [signal-to-noise (S/N) = 3]. The epi-geometry uses single objectives of 0.25, 0.50, or 1.25 NA. This second configuration has a sensitivity of 7.6 × 1010 count M−1 min−1 and a detection limit of 1.7 nM (S/N = 3) for the 1.25 NA objective. The similar performance of the two configurations is in stark contrast to the factor of 46 enhancement predicted by geometric optics for epi-excitation. The discrepancy is attributed to large increases in collected background counts, which make detection of the signal difficult.

The loss in sensitivity of two-photon excited fluorescence, when switching from a 1 cm cell format to square and cylindrical capillaries, is described quantitatively. Sensitivities are determined with calibration curves, and beam shapes are computed with optical design software. The decreased sensitivity with square capillary (75 μm i.d.) is shown to be the simple result of a truncated pathlength resulting in a factor of ∼ 2 reduction in sensitivity from a 1 cm cell. Cylindrical capillary of the same i.d. is characterized by astigmatism caused by the curved refractive index boundary in addition to the truncation. The astigmatism is determined to be smaller in the thin-walled capillary, allowing for better performance than the thick-walled capillary. In fact, the thin-walled capillary is found to have similar performance to the square capillary.

D iffraction is usually divided into two domainsÐ near ® eld and far ® eld. The boundary between the two domains depends upon the relative sizes of the aperture and the diffraction pattern at the screen (refer to Fig. 1). Since the far ® eld is de® ned as a mathematical approximation, the location of an exact boundary is problematic. With reasonable con® dence, the far ® eld can be considered to begin at a distance where the pattern has grown by at least a factor of 10. Similarly, the near ® eld can be considered to extend from the aperture to a distance where the pattern has grown by no more than a factor of 2. Classifying intermediate distances is a problem in semantics, not optics. The conversion of relative sizes into a distance depends upon the diffraction angle (see Eqs. 33, 34, and 36). Consider an example given without proof. With visib le radiation a 10 m m square aperture will have a near ® eld that extends to ; 200 m m and a far ® eld that begins at ; 1000 m m. These boundaries scale as the square of the aperture size, so that a 100 nm NSOM (near-® eld scanning optical microscopy) tip would have the near ® eld extend to ; 20 nm. At the other extreme, light diffracted by a 1 cm square aperture will not be in the far ® eld until it has traveled a kilometer. Part I of this Focal Point series was concerned primarily with near® eld diffraction. In that discussion it was shown that diffraction from a planar aperture is given by the Fresnel±Kirchhoff equation,

Passively mode-locked titanium:sapphire (Ti:S) lasers are capable of generating a high-frequency train of transform-limited subpico-second pulses, producing peak powers near 105 W at moderate average powers. The low energy per pulse (<20 nJ) permits low fluence levels to be maintained in tightly focused beams, reducing the possibility of saturating fluorescence transitions. These properties, combined with a wavelength tunability from approximately 700 nm to 1 μm, provide excellent opportunities for studying simultaneous two-photon excitation (TPE). However, pulse formation is very sensitive to a variety of intracavity parameters, including group velocity dispersion compensation, which leads to wavelength-dependent pulse profiles as the wavelength is scanned. This wavelength dependence can seriously distort band shapes and apparent peak heights during collection of two-photon spectral data. Since two-photon excited fluorescence is proportional to the product of the peak and average powers, it is not possible to obtain source-independent spectra by using average power correction schemes alone. Continuous-wave, single-mode lasers can be used to generate source-independent two-photon data, but these sources are four to five orders of magnitude less efficient than the mode-locked Ti:S laser and are not practical for general two-photon measurements. Hence, a continuous-wave, single-mode Ti:S laser has been used to collect a source-independent excitation spectrum for the laser dye Coumarin 480. This spectrum may be used to correct data collected with multimode sources; this possibility is demonstrated by using a simple ratiometric method to collect accurate TPE spectra with the mode-locked Ti:S laser. An approximate value of the two-photon cross section for Coumarin 480 is also given.

Several common aminopeptidase probes are examined for the strength of two-photon excited fluorescence. The fluorophores 4-methoxy-β-naphthylamine and 7-amino-4-methylcoumarin have limits of detection of 930 pM and 200 pM, respectively. The coumarin derivative is shown to have a two-photon cross section at 590 nm, which is exceptionally large for an aromatic amine. Spectral data for this compound also indicate that excitation near 690 nm should permit quantitation at concentrations as low as 40 pM.

Poor detection limits of two-photon excited fluorescence in cylindrical capillaries are attributed to photothermal expansion and beam astigmatism. Photothermal expansion is demonstrated for excitation in a 1 cm cell and is inferred for the larger diameter capillaries. Astigmatism is caused by focal differences between rays in a plane Longitudinal to the capillary and rays in a plane transverse to the capillary. Data were obtained by integrating the fluorescence from variously sized cylindrical and square capillaries, and by photographing the fluorescence within a 1 cm cell.

A novel fluorimeter has been developed that combines the selectivity of time-resolved detection and the sample throughput of a 96-well plate reader, With improved background rejection via temporal resolution, a reduction in the limit of detection by a factor of 600 was attained over that of a commercially available 96-well plate fluorimeter. The excellent limit of detection was utilized to differentiate two closely related species of Pseudomonas, P. fluorescens and P. aureofaciens, at a bacterial concentration of 500,000 cells per well with only 15 min of incubation, High sample throughput was demonstrated by incubation of six bacteria samples and data collection in an amino acid 12-nutrient profile within 30 min.

Trace-level impurities, such as those found in organic solvents and aqueous buffer solutions, are known to have little contribution to the two-photon excited fluorescence blank. Previously, the major source of interference in 90° fluorescence detection has been Rayleigh scatter at the laser wavelength. Various instrumental approaches have been used to minimize the amount of scattered radiation reaching the detector. Because two-photon excited fluorescence occurs primarily at the focal point of a lens, spatial isolation with microscope objectives has been used to maximize the signal-to-scatter ratio. Modulation of the excitation beam followed by second harmonic detection of the fluorescence also was shown to reduce scatter on the basis of frequency. Photon burst spectroscopy has recently been used to differentiate between fluorescence from single molecules and scatter background, while subnanosecond excitation of fluorophores has been combined with time-filtered detection to discriminate against instantaneous scatter. The reduction in background interference from these techniques has resulted in detection limits in the picomolar range.

Fluorescence steady-state and lifetime measurements have been performed that permit the differentiation of the 2 intrinsic tryptophan residues in bovine low molecular weight phosphotyrosyl protein phosphatase (BPTP), Spectral information was obtained by use of two single-tryptophan mutant proteins, W39F and W49F, and the double mutant protein W39,49F, Fluorescence measurements show that Trp(39) is characterized by a large blue shift, a low quantum yield, and a shorter mean lifetime compared to Trp(49). Solute fluorescence quenching studies of W39F reveal that Trp(49) is, highly exposed to the aqueous environment, In contrast, Trp(39) is situated within a hydrophobic core and is only partially accessible to quenching agents such as acrylamide, iodide ion, and cesium ion, The fluorescence contributions of Trp(39),and Trp(49) are additive, and their sum is equivalent to that observed for wild type BPTP, Calculated intramolecular distances between Trp(39) or Trp(49) and, 5-[[C(acetylamino)-ethyl]amino]naphthalene-1-sulfonate group covalently bound at Cys(12) or Cys(17) of the respective protein mutants, place Trp(49) within 10 Angstrom and Trp(39) at least 20 Angstrom from the active site, The fluorescence decay of the single tryptophan mutants and, surprisingly, wild type BPTP were each adequately fitted as biexponential, The latter is a consequence of the imprecision involved in determining actual minima in a three- and four-exponential fitting, Comparison of quenching results of wild type BPTP with those of the single tryptophan mutant proteins indicates that minor fluorescence components, easily resolved using a biexponential fitting for the mutant proteins, are unresolvable for wild type BPTP, These minor components skewed the weighted magnitudes and induced perturbations in lifetimes for the tryptophan fluorescence of wild type BPTP, which directly influenced the calculated values of K-sv and k(q).

Baker's yeast cells (Saccharomyces cerevisiae) were immobilized on a CZE capillary using a midcapillary frit of novel design. The frit was constructed using a single 50-microns particle with "through pores". Cells trapped on the capillary were profiled using different amino acid beta-naphthylamide-aminopeptidase substrates. These substrates were introduced sequentially and metabolized for 1 min. After incubation the hydrolysis products were withdrawn by electrophoresis and detected by laser-induced fluorescence. An aminopeptidase profile was produced using just 500 cells. The new CZE-based method was found to have several significant advantages over the older cuvette/fluorometer method.

With the use of a cavity-dumped, synchronously pumped dye laser for excitation, two-photon fluorescence cross sections are approximately eight orders of magnitude smaller than those for one-photon excitation. Thus, examination of dilute solutions has been achieved only with great difficulty. Any successful instrumentation will require that the blank be essentially eliminated. To this end, time-filtered detection has been combined with two-photon excitation and spatial filtering to produce fluorometric detection limits of 38 pM for 9,10-diphenylanthracene and 8.6 pM for α-NPO. It is believed that this latter value is the lowest concentration yet reported for two-photon spectroscopy in fluid solution. The instrumentation and data processing are described. Additionally, a comparison is made with the performance of other recent alternative approaches involving spatial filtering and second harmonic detection.

A simple time-resolved fluorometer is constructed with an all-solid-state-based, frequency-tripled, Q-switched, diode-pumped Nd:YLF laser as the excitation source. Signal processing is accomplished with a digital oscilloscope. Simplicity of operation and applicability to trace analysis and in time-resolved spectroscopy are demonstrated with this new instrument. The laser produces 2.5-ns pulses at 349 nm and is capable of kilohertz repetition rates. For every shot of the laser, the oscilloscope collects an entire fluorescence decay at a 1-ns digitizer resolution and can average these data at the maximum laser repetition rate. When one is operating at 1 kHz and signal averaging for one second, detection limits (S/N = 3) in the 10–100 pM region are obtained. Excited-state decays are collected for several enzymatic probes and quinine sulfate, providing lifetimes consistent with those obtained by established instruments.