TCSPC Devices for Ultra-High Time Resolution Single-Photon Counting
More than 25 Years of Experience in TCSPC – as technology originator and leader Becker & Hickl offers TCSPC solutions with unprecedented time resolution. The SPC modules can be used for classic and multidimensional fluorescence decay applications as well as for fluorescence lifetime Imaging (FLIM), multi-wavelength FLIM, simultaneous fluorescence- and phosphorescence lifetime imaging (FLIM/PLIM), fast time-series FLIM, fluorescence correlation spectroscopy (FCS), single-molecule experiments, anti-bunching experiments (HBT) and many other multi-dimensional photon recording tasks.
Choose your solution now:
|TCSPC Channel per Module||1||3 (4)||3 (4)|
Sing. Molec. FCS
Timing Precision (Jitter, RMS)
|<20 ps||<20 ps|
|Min. Bin Width||813 fs
|4 ps||4 ps|
|Abs. Timing Stability (RMS)||<0.5 ps||<5 ps||<5 ps|
General Recording Modes
Classic TCSPC, FLIM, FCS,
|Classic TCSPC, FCS, HBT(4),
|Classic TCSPC, FLIM, FCS,
HBT(4), Time Tag
TCSPC, FCS, HBT(4),
|Detectors per Channel||Up to 16 (1)||Up to 16 (1)||Up to 16 (1)|
|Parallel Modules||Up to 4 (2) or 32 (3)||1||1|
(1) For all modules of the SPC series: up to 4 or 8 PMTs are connected using routers HRT-41 or HRT-81; for up to 8 SPADs (TTL) use HRT-82. 16 channel detection is provided by the PML-16 detector series.
(2) A single instance of SPCM software operates up to four modules. Multiple instances of SPCM can be launched to operate more than four modules or modules of different types.
(3) Using SPCM DLL with custom programming up to 32 modules can be operated in parallel.
(4) Hanbury-Brown-Twiss setup. Start-stop histogram between two detectors.
SPC-130IN TCSPC SeriesSPC-130IN, SPC-130INX, SPC-130INXX TCSPC Modules
SPC-QC-104 TCSPC ModuleNEW3 Channel TCSPC/FLIM Module
SPC-QC-004 TCSPC ModuleNEW3/4 Channel TCSPC Module
Technical Terms of TCSPC and TCSPC FLIM
TCSPC – Time-Correlated Single Photon Counting
TCSPC records the temporal profile of a repetitive optical signal by detecting single photons of the signal, determining the times of the photons after a reference (or excitation) pulse, and building up the distribution of the photons over the time after the reference pulse. The most frequent application is recording of fluorescence decay functions. TCSPC features extremely high time resolution and near-ideal photon efficiency and sensitivity. See also ‘Multidimensional TCSPC’.
Extension of the classic TCSPC technique. Multi-dimensional TCSPC detects single photons of a repetitive light signal and determines the times of the photons after a reference (or excitation) pulse along with one or several other parameters of the photons. These can be the wavelength, the spatial location within a sample from which the photons are emitted, the time after a stimulation of the sample, or any other parameter that can be associated to the photon. Examples are Multi-Wavelength TCSPC, Fluorescence Lifetime Imaging (FLIM), Phosphorescence Lifetime Imaging (PLIM), Time-Series FLIM, or a combination of these techniques.
A multidimensional TCSPC technique where the second parameter of the photons is the wavelength. Multi-wavelength TCSPC simultaneously builds up several (with current bh detectors up to 16) fluorescence decay curves for different wavelength. Please note that Multi-Wavelength TCSPC is not based on wavelength scanning. Photons of all wavelengths are detected simultaneously and directed into different decay curves of the photon distribution.
FLIM – see TCSPC FLIM
TCSPC FLIM – TCSPC-based Fluorescence Lifetime Imaging
TCSPC FLIM scans a sample with a high-frequency pulsed laser beam, detects single photons of the fluorescence light, and builds up a photon distribution over the photon times after the excitation pulses and the coordinates of the scan area. Please note that this is not a pixel-by-pixel process. When a photon is detected it is just placed in the pixel corresponding to the position of the laser beam in the moment of the photon detection and in a time channel corresponding to the time of the photon in the excitation pulse period. TCSPC FLIM can therefore work at any scan rate. FLIM is the most frequent application of multidimensional TCSPC. TCSPC FLIM delivers an array of pixels, each containing a fluorescence decay curve consisting in a large number of time channels. The technique features extremely high time resolution and near-ideal photon efficiency. With a repetitive scan, the signal-to-noise ratio of the result depends only on the photon rate and the total acquisition time. FLIM can be used as a contrast technique in microscopy, but most FLIM applications are aiming at molecular imaging, i.e. the determination of molecular parameters via their influence on the fluorescence decay functions. TCSPC FLIM was introduced by bh in 1996. The first use was in fluorescence-lifetime imaging ophthalmoscopy (FLIO). Applications in laser scanning microscopy followed in 1999.
TCSPC FLIM uses scanning. Although scanning makes the optical system more complicated than a wide-field system it has a big advantage. Since only a single pixel is excited at a time there is no scattered light from the other pixels. Scanning alone, even without the aid of confocal detection or multiphoton excitation, therefore yields a far better image contrast than wide-field imaging.
FLIO – Fluorescence Lifetime Imaging Ophthalmoscopy
Clinical FLIM at the human eye. Fluorescence lifetime images of the retina show early indications of metabolic changes. They can thus be used to discover early stages of eye diseases before these have caused permanent damage to the eye.
TCSPC FLIM, where the sample is scanned with a confocal system. The sample is scanned through an objective lens with a focused laser beam. Fluorescence photons are collected by the objective lens and projected into a pinhole in a plane conjugate with the sample. The photons passing the pinhole are recorded by TSCPC FLIM. Confocal detection suppresses out-of-focus signals and scattered light, resulting in a clean lifetime image of the focal plane.
The sample is scanned through an objective lens with a near-infrared femtosecond laser beam. Excitation occurs by a multiphoton (usually two-photon) process. Excitation occurs only in the focal plane. Therefore no pinhole is required to reject out-of-focus fluorescence. Photons collected by the objective lens are transferred directly (without a pinhole) to the FLIM detector. This principle is called ‘non-descanned detection, or ‘NDD’. The buildup if the lifetime image occurs by the usual TCSPC FLIM process. The advantage of multiphoton imaging is that the near-infrared light reaches deep sample layers. Moreover, the NDD beam path transfers photons to the detector even when they are scattered on their way out of the sample. Consequently, multiphoton FLIM delivers clean images of deep sample planes.
IRF – Instrument-Response Function
The IRF is the temporal profile a TCSPC system would record by looking directly at the exciting laser beam without a sample in between. The IRF includes the temporal width of the laser pulses, the transit-time-spread in the detector, possible synchronisation jitter, and the timing jitter of the TCSPC module itself. bh TCSPC devices currently reach an IRF width (full width at half maximum) of 18 ps with fast hybrid detectors, 9 ps with fast SPADs, and 4.4 ps with fast superconduction single-photon detectors (SSPDs).
The electrical IRF is the temporal profile the TCSPC device records when defined electrical pulses are applied to its inputs (detector and sync). The electrical IRF contains only the timing jitter in the TCSPC module. bh TCSPC modules have electrical IRF widths between <3 ps and about 6 ps, full width at half maximum.
There are different definitions of the width of the IRF of a TCSPC system. It can be the IRF width of the entire TCSPC system, including the laser, the detector, the TCSPC device or only the IRF of the TCSPC device. In the first case it is the system IRF, in the second case it is the electrical IRF. Second, there are two definition of the IRF width. Correctly, the IRF width should be given as ‘FWHM’, or ‘Full Width at Half Maximum’. However, to specify low IRF widths manufacturers more and more tend to give the IRF width as ‘RMS’, or ‘Root Mean Square’. This value is not the IRF width but the average timing jitter of the detected photons. The RMS Timing Jitter is 2 to 3 times smaller than the FWHM of the IRF.
There is no parameter that is so poorly defined as the ‘Time Resolution’ of TCSPC. It can be the FWHM of the electrical IRF width, the FWHM of the system IRF, or the RMS Timing Jitter of the electrical IRF or the system IRF. To compete in time resolution manufacturers even give the time-channel width as ‘time resolution’. The time-channel width tells nothing about the true time resolution. It is specified as ‘Time Resolution’ purely for competition reasons. Finally, ‘Time-Resolution’ can be the minimum fluorescence decay time a system can resolve by de-convolution. This resolution is more or less hypothetical and should not be used for comparison of lifetime-detection systems.
The Time-Channel Width is the width of the time channels into which the photons are collected. The parameter is often incorrectly given as ‘Time Resolution’. The time-channel width must be small enough to correctly resolve the temporal shape of the signal but there is usually no point to decrease it below about 1/10 of the FWHM of the System IRF or 1/10 of the lifetime of the fastest decay component in the signal. The bh TCSPC modules have time-channel widths down to 200 femtoseconds. This is small enough to efficiently resolve even signals detected with fast SSPDs.
The timing stability over time is important to obtain reproducible results of TCSPC and TCSPC FLIM experiments. It is also important if weak signals have to be recorded with long acquisition time. bh TCSPC devices feature timing stabilities in the sub-picosecond range. Only bh specify the timing stability for their TCSPC and TCSPC FLIM devices.
The dead time is the time the TCSPC device needs to process a photon. If another photon is detected within this time it is ignored. The dead time causes a nonlinearity of the intensity scale but not necessarily an error in the recorded fluorescence decay profile.
Loss of photons within the dead time periods (signal processing times) after the recorded photons. Counting loss causes a nonlinearity of the intensity scale of TCSPC and TCSPC FLIM but not an error in the recorded decay profiles.
Saturated Count Rate
Unless the count rate is limited by the readout speed the saturated count rate is the reciprocal dead time. It is the count rate a TCSPC device can achieve for an infinitely high photon rate at the input. A TCSPC device can be reasonably operated up to about 50% of the saturated count rate.
Pile-Up occurs when the detector of a TCSPC system detects a second photon within the same signal period with a previous one. The second photon is then lost. In contrast to counting loss, pile-up causes a distortion of the recorded decay profiles. The effect increases with increasing ratio of count rate to pulse repetition rate. Pile-up was a real problem in early TCSPC experiments with flashlamp excitation. With today’s high-repetition-rate laser light sources it is rarely a problem. There is a common misconception about the size of the pile-up effect: It is 100 times smaller than commonly believed. The often cited ‘Pile-up-Limit’ of 0.1% of the excitation pulse rate is wrong. The real pile-up limit is at a photon rate of 10% of the excitation pulse rate. The wrong estimation of the size of the pile-up effect has led to many unnecessary – and often inefficient – technical attempts to develop pile-up free fluorescence-lifetime detection principles.
Dark Count Rate
The Dark Count Rate is the pulse rate a detector delivers when no light is present at its active area. Dark counts originate from thermal emission at the photocathode of a PMT, HPD or MCP, or from thermal generation of electron-hole pairs within the depletion layer of a SPAD. The dark count rate can therefore be reduced by cooling. In TCSPC and TCSPC FLIM results dark counts cause an offset in the baseline of the recorded decay functions. However, the offset by dark counts is usually far smaller than the offset by afterpulsing, see below. Therefore cooling the detector may not always have the expected effect.
Afterpulsing is the tendency of a detector to generate a false pulse within the first microseconds after the detection of a photon. In PMTs afterpulsing is caused by ionisation of rest gas atoms in the vacuum of the tube. In SPADs afterpulsing comes from trapped carriers remaining from the carrier avalanche of a previous photon. In high-repetition rate TCSPC applications the afterpulses from photons of many excitation periods sum up and form a substantial counting background. The afterpulsing background can thus be much higher than the background from dark counts. Afterpulsing cannot be reduced by cooling. The best way to avoid it is to use hybrid detectors. Hybrid photon detectors (HPDs, HPMs) are free of afterpulsing because for every photon only a single electron is travelling in the vacuum.
Signal-to-Noise Ratio, SNR
When an optical signal is recorded the signal-to-noise ratio is limited to the square root of the number of photons recorded. This rule applies both to the recorded intensity and the recorded fluorescence lifetime. The only way to improve it is to record more photons, either by increasing the acquisition time, by increasing the photon rate at the input of the detector, or by using a detector with a higher detection quantum efficiency. In FLIM the SNR of the lifetime obtained for the individual pixels can be increased by reducing the number of pixels. For a given photon rate and a given acquisition time the photon number per pixel than becomes proportionally higher. However, reducing the pixel number goes on the expense of spatial resolution. Better results are usually obtained by overlapping binning in the FLIM data analysis.
Detection efficiency is the efficiency at which photons are detected (and recorded) by a TCSPC system. The term ‘Detection Efficiency’ is used in different context. It can be the probability that a photon at the optical input of a detector causes a useful electrical pulse at the output, the probability that the photon is recorded by the TCSPC device, or the probability that a photon emitted by the sample is recorded. In the last case ‘Detection Efficiency’ includes the probabilities that the photon is collected by the optical system, that it passes filters and lens and mirror systems or optical fibres, and that it is projected on the active area of the detector. The efficiency of the optical system is especially important in FLIM. Photons are emitted isotropically, thus the efficiency increases with the square of the numerical aperture of the objective lens.
The photon efficiency describes how many photons a lifetime detection technique has to detect to reach a given signal-to-noise ratio. It is defined as the number of photons theoretically needed to reach this SNR divided by the number of photons actually needed. When the measurement conditions are chosen correctly the photon efficiency of TCSPC and TCSPC FLIM is 1. The photon efficiency decreases if the IRF of the system is too long, if the signal is contaminated by background from afterpulsing or dark counts, and if the fluorescence decay is not entirely in the recording time interval.
There is no parameter that is discussed more controversially than the acquisition time of TCSPC and TCSPC FLIM. In early TCSPC systems the excitation-pulse rates were on the order of 10 kHz. Consequently, the useful photon rate was limited by pile-up, and the acquisition time was extremely long. Now the pile-up problem has been overcome by excitation sources with repetition rates in the range of 50 to 80 MHz. The acquisition time thus depends on the expectation to the signal-to-noise ratio of the result, the photon rate that can be obtained from the sample, the detection efficiency, and, in FLIM, on the number of pixels and the binning factor in the lifetime analysis. Depending on the desired signal-to-noise ratio, today’s TCSPC devices record fluorescence decay curves within milliseconds to tens of seconds and FLIM within a fraction of a second to several minutes. A widespread misconception is that a ‘faster’ TCSPC module (with shorter dead time) decreases the acquisition time. However, unless the pulse rate at the TCSPC input is or can be made higher than the reciprocal dead time this is not the case. Especially for biological samples the count rate available from the sample is limited, and no higher than the count rate a normal TCSPC / FLIM module can process.
DNL – Differential Nonlinearity
The DNL is the homogeneity of the time-channel width of the time-conversion in the TCSPC / FLIM module. Inhomogeneous time-channel width results in inhomogeneous distribution of the photon number in subsequent time channels. DNL thus introduces noise in the recorded signal shape. DNL was a problem is early TCSPC devices but has been largely overcome by a new TAC/ADC principle introduced by bh in 1993. In practice, DNL errors are more likely introduced by electrical crosstalk between the photon pulses and the synchronisation signal than from the internal DNL of the TCSPC module.
Channel Nonuniformity – See DNL
A method to improve the DNL of the TAC-ADC combination over that of the ADC. The output signal of the TAC is slightly shifted from one photon to another so that the time measurement occurs at a different place of the ADC characteristics. The shift is later digitally subtracted from the ADC signal to that the electrical IRF is not broadened. Introduced by bh in 1993, and basis of bh’s fast TAC-ADC timing principle.
CFD – Constant Fraction Discriminator
The photon pulses from most single-photon detectors have a considerable amplitude jitter. It must be avoided that the amplitude jitter induces a timing jitter in the discriminators of the TCSPC device. TCSPC is therefore using a ‘Constant Fraction Discriminator’. That means the discriminator triggers when the pulse reaches a given fraction of its full amplitude. Technically, this is achieved by re-shaping the photon pulse so that it gets a bipolar shape. The temporal position of the zero-cross point is independent of the pulse amplitude. By triggering on the zero-cross point, amplitude-induced timing jitter is largely avoided.
TAC – Time-to-Amplitude Converter
A TAC converts the time difference between a start and a stop pulse into the amplitude of an electrical pulse. The amplitude is then converted into a digital start-stop time by an Analog-to-Digital Converter (ADC). The digital start-stop time determines the time channel into which the current photon is added. The TAC-ADC principle is the oldest principle of time conversion in TCSPC. It has been upgraded by bh in terms of conversion speed and time resolution, and is currently the principle with the highest time resolution. bh TCSPC modules based on the TAC-ADC principle yield electrical IRF widths of less than 2 ps and RMS timing jitter of less than 1 ps. The time-channel width can be chosen as small as 200 femtoseconds. There is no other TCSPC module that comes anywhere close to this resolution.
TDC – Time-to-Digital Converter
A TDC measures the time of an event (in TCSPC this is usually the detection of a photon) by sending it through a chain of logics gates. By sending the photon pulse through one delay chain and the Sync pulse through another the time between the photon and the sync pulse is determined. The electrical IRF width is typically 40 ps to 80 ps, the minimum time-channel width a few picoseconds. An advantage of the TDC principle is that it can easily be made for several timing channels one TCSPC board. The principle is used in the bh SPC-QC-104 module.
Parallel Counter Channel
Recent bh TCSPC FLIM modules have a counter channel in parallel with the timing circuitry. The counter counts the photons with virtually no dead time. By taking the photon times from the timing circuitry and the photon numbers from the parallel counter FLIM images with a linear intensity scale are obtained.
PMT – Photomultiplier Tube
A PMT is a photon counting detector with a photocathode and a number of amplification stages, or dynodes. A photoelectron is accelerated from the photocathode to the first dynode. At the dynode it generates several secondary electrons, which are accelerated to the next dynode, where they generate more electrons. PMTs have between five and nine dynodes and reach multiplication factors up to 107. A single photon thus delivers a detectable output pulse of several mA.
HPD – Hybrid Photodetector
Photoelectrons are generated at a photocathode and accelerated towards a silicon avalanche diode. When a photoelectron hits the diode it generates a few thousand electron-hole pairs. These are further amplified by the avalanche process in the diode. The total gain is on the order of 106, which means that a single photon generates a detectable pulse at the output of the detector. Hybrid detectors have extraordinarily low transit time spread (timing jitter) and no afterpulsing. The bh HPM-100 series hybrid detectors are based on Hamamatsu HPDs.
SPAD – Single-photon Avalanche Photodiode
A SPAD is a photodiode the reverse voltage of which is increased above the breakdown level. When a photon is detected it causes an avalanche breakdown in the diode, resulting in a well detectable output pulse. After each breakdown, an electronic ‘quenching circuit’ re-establishes normal operation. SPADs have high quantum efficiency, small transit-time spread but only small active areas. Moreover, they are not free of afterpulsing.
SSPD – Superconducting Single-photon Detector
An SSPD consists of a microscopically small microstructure of superconducting material. The establish superconduction the entire structure is cooled to less than 3 Kelvin. A small electrical current is sent through the structure. When a photon hits the structure superconduction breaks down for a short period of time, causing an electrical pulse. SSPDs have extremely short transit-time spread and high quantum efficiency up to IR wavelengths. The disadvantage is the small active area and the need of cooling to extremely low temperature.
TTS – Transit-Time Spread
The TTS is the internal timing jitter of a detector. The TTS is in the sub-ns range for PMTs, in the 10 to 20 ps range for HPDs and fast SPADs, and in the range of a few ps for SSPDs. In PMTs and HPDs the electron emission at the photocathode can contribute to the TTS. Emission from the conventional bialkali and multialkali photocathodes occurs instantaneously, but emission from semiconductor cathodes (InGaAs, GaAs) happens with a random delay. Photoelectron emission from these cathodes can contribute about 100 to 200 ps to the TTS of the detectors.
SER – Single Electron Response
The SER is the electrical pulse shape a detector delivers for a single photoelectron. It closely resembles the output pulse for a single photon. The SER is normally much broader than the TTS. This is the reason that TCSPC delivers a much higher time resolution than a technique that uses the same detector as an analog electro-optical device. As an example, the SER of a hybrid detector is about 800 ps wide, but the TCSPC IRF can be as short as 18 ps.
QE – Detection Quantum Efficiency
The QE (Quantum Efficiency) of a detector is the probability that a photon causes a detectable output pulse. PMTs deliver pulses of random amplitude, therefore the effective QE depends on the threshold of the subsequent discriminator. The QE definition for PMTs is therefore sometimes replaced with a ‘Cathode Luminous Sensitivity’, which is the cathode current in mA per Watt of incident optical power. Please see bh TCSPC Handbook.
The Excited-State Lifetime is the average time an excited molecule needs from its excitation to the return to the ground state.
As far as the S1 state is considered the Fluorescence Lifetime is identical with the excited-state lifetime.
The average temporal profile of the fluorescence signal emitted by a large number of photons and/or over a large number of excitation/emission cycles. For a homogeneous assembly of molecules in a homogeneous environment the fluorescence decay function is single-exponential. The fluorescence lifetime is the time from the excitation pulse to the 1/e point of the decay function.
In practice fluorescence decay functions are rarely single-exponential. Deviation from the single-exponential decay is caused by inhomogeneity of the molecular environment, by inhomogeneity of the binding of the fluorescent molecules (fluorophores) to proteins, enzymes, and lipids, inhomogeneity of the conformation of the fluorophores themselves, or simply by presence of different fluorescent molecules. A multi-exponential decay is described by several decay components with different lifetimes and amplitudes. The composition of multi-exponential decay curve bears molecular information about the system to which the fluorophores are associated. Resolving multi-exponential decay functions is key for the use of TCSPC and TCSPC FLIM as a molecular imaging technique.