During the past two decades, the traditional fluorescence microscopy has developed from a discipline interested in the description of static microstructures to a tool of molecular cell biology aiming to analyze physiologically relevant molecular processes on the cellular and systems network level.

Exploring the dynamics and function of macromolecular networks requires high-resolution microscope techniques below the diffraction limit of visible light [1]. With the introduction of new microscopes and optical techniques, as, for instance, the introduction of multiphoton microscopy [2–5, 4pi-microscopy [6–10], stimulated-emission-depletion microscopy (STED) [11, 12], structured illumination (SIM) [13, 14], localization microscopy (ground-state-depletion microscopy (GSD)) [15], spatial precision-distance microscopy (SPDM) [16, 17], photoactivated localization microscopy (PALM/STORM) [18, 19], and light sheet illumination [20, 21], a significant improvement of spatial resolution beyond the diffraction limit of classical microscopy was recently achieved. Most of these concepts take advantage of the fact that fluorescence can be switched between a dark state and a bright state, either in a stochastic (PALM/STORM/GSD) or in a deterministic way (STED).

Unfortunately, most fluorescence microscopes work at high illumination intensities. The strive to collect as many photons as possible necessarily results in high average and peak powers [22] (typically >1 kW/cm²), which leads to irreversible photobleaching of dyes and significant photodamage of living cells [23]. Therefore, the abovementioned visualization techniques (1) do not allow the direct observation of molecular interactions over a longer period of time, (2) induce photodynamic reactions of the molecular probes, and (3) distort the living state. Prerequisite for non-distorted visualization of dynamic processes within living cells (e.g., during synapse formation and signal transduction) by fluorescence microscopy is a minimal invasive approach [24] to excite and detect the molecular constituents at conditions that preserve the undisturbed living state. Consequently, the development of minimal invasive observation systems that yield information about protein–protein or protein–target structure interactions on the cellular, subcellular, and single molecule level, without disturbing the intracellular metabolism, is an outstanding goal in modern cell microscopy.

The improvements of the spatial resolution in modern superresolution microscopes were an important step forward to visualize the molecular organization of cells but do not exploit the full informativeness of the fluorescence process. A fundamental technique capable to grant highest resolution, sensitivity, and comprehensive information about a fluorophore’s environment is fluorescence lifetime imaging microscopy (FLIM) (Fig. 1). Measuring the lifetime of the fluorophore’s excited state reveals incidents occurring in nanometer range of the molecular environment like changes in pH, temperature, ion concentration, viscosity, or polarity among others.

At present, two methods have been evolved to measure fluorescence lifetimes: frequency domain methods [25–27] and time domain-based methods [28–30] (Fig. 2a, b).

Whereas in the frequency domain, the fluorescence lifetime is extracted from the phase shift and demodulation of the fluorescent light with respect to the phase and the modulation depth of a modulated excitation source, in time domain the decay information is acquired directly by recording the decrease of the fluorescent light intensity over time after exciting the specimen in a single-photon counting mode (Figs. 2b and 3). That means, a sample is illuminated with short laser pulses (fs to ps range), and the time difference between the emitted laser pulse and the appropriate fluorescent photon at the detector is measured resulting in a histogram of arrival times (Fig. 3). Among the other techniques, only single-photon counting is capable to utilize the detected photons as efficiently as it is physically possible [31] with short excitation pulses [32].

A large number of detection techniques are employed for FLIM applications nowadays. The most widespread detectors, implementing single-photon counting-based approach, are photomultiplier tubes (PMTs) (Fig. 4, see Note 1), microchannel plate-based PMTs (MCP-PMTs) (Fig. 5a, c, see Notes 2 and 3), and single-photon avalanche diodes (SPADs) (see Note 4). In addition there are a number of integrating (opposite to counting) detectors: gated CCD cameras (Fig. 6, see Note 5), streak cameras (see Note 6), and conventional photodiodes. Advantages and drawbacks of the individual systems are listed in Table 1.

In contrast to time only single-photon counting techniques that acquire positional information by scanning methods in a wide-field FLIM-system, a vacuum-sealed microchannel plate-based photomultiplier tube (MCP-PMT) with a positional sensitive multi-anode is used as detector.

A position-sensitive photomultiplier (see Note 3) combines a photocathode (see Note 7), a microchannel plate (MCP) stack (see Notes 2 and 3), and a position-sensitive anode system (see Note 8) (Fig. 5).

To register single photons, the detector needs an amplification element to convert single quanta of light into an electrical signal that can be processed and measured. This conversion process is performed by the photocathode. The photoelectron following an applied electrical field is subsequently multiplied by a microchannel plate stack (Fig. 5c). The resulting avalanche, depending on the number of amplifying elements, consists of 10⁵ to 10⁷ electrons. The avalanche is detected by an anode (Fig. 5a). A conventional disk or cone-like anode enables only time-resolved registration. Alternatively, the electron cloud is converted back to the visible light by a phosphor screen for CCD-based acquisition (e.g., gated CCD).

However, the latter approach cannot be used in applications where single particle counting is required because of the long readout time of a CCD. Therefore, it is necessary to employ a position-sensitive anode that provides spatial information for each individual detected event in combination with timing information.