EP1149279A1

EP1149279A1 - Method for detecting fluorescence phenomena in microscope

Info

Publication number: EP1149279A1
Application number: EP00988743A
Authority: EP
Inventors: Ralf Wolleschensky; Volker Gerstner
Original assignee: Carl Zeiss Jena GmbH
Current assignee: Jenoptik AG
Priority date: 1999-11-25
Filing date: 2000-11-22
Publication date: 2001-10-31
Also published as: DE19956620A1; US6741346B1; JP2003515162A; WO2001038856A1

Abstract

According to the inventive method for detecting fluorescence phenomena in a microscope, the sample is irradiated with a modulated and/or pulsed laser light source and the fluorescence is detected in at least two different phase positions of the detector.

Description

Process for the detection of fluorescence phenomena in a microscope

The method shown concerns confocal and 2-photon fluorescence microscopy [Goeppert] [Wilson]. Both methods are assumed to be known.

Both the confocal fluorescence microscopy and the 2-photon microscopy are modified with the method shown below in such a way that an additional contrast parameter is possible: the fluorescence lifetime.

Time-resolved fluorescence or the use of the lifetime as a contrast parameter in confocal / 2-photon microscopy can be carried out in two different ways - by detection in time space (time-doamin detection) and detection in frequency space (frequency- domain detection).

In time-domain detection [Wabnitz] [Gröbler], a fluorescent sample is excited to fluoresce by means of a pulsed light source and the fluorescence emission either by means of time-correlated single-photon counting (TCSPC) [Müller] [Han ] or detection by means of an amplification time window in the detector (time-gated detection) [Schbg] [Cubeddu]] [Dowling].

For detection in the frequency domain, a fluorescent sample / preparation is excited with an actively modulated or (e.g. passively mode-coupled) pulsed light source. Since any modulation of the excitation by Fourier analysis breaks down into sinusoidal components, it is sufficient to consider a sinusoidal excitation. The detection in the frequency space is based on the delay of the fluorescence emission by a phase f and a change in the modulation depth M compared to the excitation light as a function of the

Modulation frequency ω (= 2 π f _mod ) and the lifetime τ. φ atan (ω r) (1)

M = (2)

+ ω τ

The resulting fluorescence signal oscillates at the modulation frequency, but is out of phase and demodulated. For typical fluorescence lifetimes in the range τ = 1 ... 10 ns, modulation frequencies in the range + hf _m0 d ⁼ 10 ... 100MHz are sufficient.

Since it is generally not sensible to sample the fluorescence signal at such high modulation frequencies, a frequency mixing process is used to acquire the signal. For mixing, e.g. each detection element with modulatable reinforcement.

There are essentially two different approaches - the homodyne and the heterodyne detection.

In order to understand the principle, consider generally two "signals" S _1r S ₂ (eg Si modulated excitation, S ₂ modulated amplification).

Multiplication leads to

S, S ₂ = A ₀ B ₀ + A _Q B _λ cos ω _b t + ß) + 5 ₀ _4, cos (_y _α t + a) +

(4)

A _l B _l {cos ((ω _a + ω _b ) t + (a +?)) + Cos ((ω _α - ω _b ) t + (a - ß))

Homodyne detection

If ω _a = G_ (homodyne), the _second harmonic and a frequency-independent component are generated by the mixing process of the "signals" Si and S ₂ . A low pass filter suppresses the components at ω _a and 2 ω _a . Only the DC background and the phase-dependent DC component is detected. The signal filtered using a low pass (LP = LowPass) can be written as

LP (S ₁ S ₂ ) = A ₀ B ₀ + A _χ B, cos (c. - ß) (5)

In homodyne detection, this frequency-independent (DC) signal is detected in several relative phases. In order to measure the fluorescence lifetime, at least 3 different phase positions are necessary. With just two relative phase positions, the phase shift or demodulation induced by the fluorescence lifetime can be used as a contrast parameter. [Clegg].

Heterodyne detection - cross correlation

If co = ω _a + Dw (heterodyne), the mixing process generates a high-frequency signal at the sum frequency and a signal at the cross-correlation frequency Δω. The high-frequency components are suppressed again by a low-pass filter.

LP (S _] S ₂ ) = A ₀ B ₀ + A _i B _] cos (Aωt + - ß) (6)

The difference frequency Δω is detected in the heterodyne detection. The phase position and depth of modulation of the signal at the difference frequency enable the service life to be determined. Typical cross-correlation frequencies are in the range from a few Hz to about 100 kHz.

A more comprehensive representation of the heterodyne technique can be found in publications by E. Gratton et al. [Gratton] [Gratton2]

In both the homodyne and the heterodyne detection, the high-frequency change is reflected as it were on the low-frequency range. In the case of multi-exponential decay behavior of the fluorescence emission, the lifetimes determined from the depth of modulation and phase shift are different. The excitation frequency must therefore be varied in order to measure the service life precisely [Gratton2].

On the other hand, should the service life e.g. are generally not required as contrast parameters in an imaging process. Often, e.g. the representation of the phase shift or demodulation by the service life or the service lives calculated therewith at a fixed modulation frequency.

1.1 Confocal lifetime imaging, homodyne

The confocal structure of an LSM is changed to reflect the service life in such a way that

• the excitation light source is modulated or a pulse laser is used

• the gain of the detector (e.g. PMT) is modulated.

• An electronic phase shifter is used, which allows the relative phase position of light excitation and detection to be subdivided. The diagram with the components used is shown in Figure 1.

Realization 1:

Using a synthesizer S driven by a quartz crystal, RF frequencies in the range of 10-100 MHz are generated. The output of the synthesizer S is connected on the one hand to a phase shifter / amplifier PA via a shielded high-frequency line (for example via a BNC cable). On the one hand, the phase shifter / amplifier amplifies the high-frequency input signal of the synthesizer - in the implementation shown to a power of approx. 1.5 W with a resistance of 50 W. Furthermore, the phase position of the amplified RF signal can be varied with the PA. The relative phase of the PA can be set digitally via a control line - for example, by control via a serial interface of a PC. The amplified RF signal becomes one modulatable PMT - for example the PMT module (H 6573) from Hamamatsu ¹ - supplied. The amplified RF voltage (approx. 25 V _pp / 50 W) is fed to the 2nd Dynode in the PMT and used to modulate the amplification of the PMT. To produce a defined low pass filter, the output signal of PMT ^'s using a standard low-pass (LP) filter to be smoothed. The 3dB cut-off frequency f _{g of} the LP filter should be selected so that

1 / pixel dwell time «f _g « f _laser

with f _LaSer = laser modulation frequency. The signal filtered in this way is fed to the standard detection electronics of the LRM (to understand: in principle, for example, an ADC (Analog Digital Converter) which is synchronized with the xy scanners of the LRM.)

A second identical output of the synthesizer S is also connected to the control electronics of the light modulator M via a high-frequency line. For example, an acousto-optical modulator in the laser beam path (or else an electro-optical modulator) can be used as the modulator. The light modulator is part of the laser module of the confocal laser scanning microscope. The relative phase position of the PA and the light modulator M can be varied by such a construction. The frequencies for M for modulating the laser light and the PMT boost voltage are the same. The phase position is generally different. At least the phase of an RF output can be set digitally in approximately 1 ° steps. In the case shown in Figure 1, the phase of the PMT modulation voltage is varied by the phase shifter in PA.

Description of the mode of operation and the associated advantages of the method shown.

The light from the modulated / pulsed laser light source is focused into the object plane via the scanning mirror and the LSM scanning optics using the usual LSM structure. The fluorescence is like in the confocal Zeiss LSM510 in

¹ A more detailed description of how it works is available as "Technical Information" for the H6573 from Hamamatsu. The direction of reflection is focused on the detection pinhole via a color divider, focusing lens etc. A raster-like movement of the laser focus over the object plane and synchronized detection in the - now modulated - PMT creates a confocal image of the object plane. By using homodyne technology, the PMT signal at the output of the LP filter (Figure 1) is a DC signal that only varies depending on the laser spot position (pixel). By repeating the scanning process with several different relative phases of light modulator and detector modulation, the detection of the fluorescence lifetime contrast or the measurement of the fluorescence lifetime is possible. The main advantage is that the usual data acquisition unit of the LRM can be used. The calculation of the life-time contrast is done in images.

Image acquisition to create the lifetime contrast in a confocal microscopic image or to show the confocal lifetime distribution of dyes.

• In a first step, in the implementation shown here, a first phase of the PMT detector is set digitally. The setting is done as shown above with the electronic phase shifter / amplifier PA and the resulting DC signal of PMT ^'s (according to the usual current Spannungswandiung and analog - digital conversion) in synchronism with the scanner by means of conventional (LSM) evaluation electronics registered and by means of a storage medium stored in the PC.

• In at least a second step, the procedure is repeated with a second different relative phase position.

Using e.g. On the computer screen, the at least two digital images are optionally displayed without further calculation.

The images generated with the following algorithms are also displayed on the screen for display (Fourier development) The Fourier expansion is given by

Iφ + φ _τ ) = a ₀ + a _λ • sin (<) + b _λ • cos (<p) (7)

With the fluorescence intensity of a pixel / and the corresponding Fourier coefficients a ₀ , ai, bi.

- l ri 2π α _n = I = -> / _r + n -

2 ^ (2πn \ (2π b, = -> cos / φ _τ + n

Here, ao, ai, b _{0 are} the Fourier coefficients (per pixel), N> = 2 the number of stored phase images (or pixel intensities) l = l (φ) The modulation depth _r and the phase shift due to the lifetime φ _τ can be determined by the Fourier coefficients be expressed.

φ _τ = - \ - ^ T = ωτ (10)

The modulation depth / W is calculated pixel by pixel. M = M (i, j) = M, _j . (i, j, pixel indices). Likewise the phase shift φ = φ (i, j) = φ _v

The image of the modulation depth M _tJ calculated in this way pixel by pixel and the phase shift φ _{Ψ are} shown on the monitor

Another type of representation is the τ mapping, ie the lifetimes τ (M), _j and τ (φ) calculated according to t by solving equations (9) and (10). [Clegg] :() = ^■ - 1 (11) \ M

τ {φ) = —tan {φ) (12) ω

M and φ are calculated using equations (9) and (10) and the Fourier coefficients a ₀ , ai, bi determined using equation (8). To increase the accuracy in determining the phase, optimized algorithms can also be used, which are presented in a work by Küchel, Phase Evaluation by Folding, 1989.

Realization 2:

A further realization of a confocal microscope with life time contrast or for measuring the life time distribution in a confocal sectional image can be done using pulse lasers. Suitable are, for example, pulse laser diodes or other fs laser systems (Ti: sapphire laser), for example with a frequency conversion unit connected downstream (eg frequency doubling / tripling). When using pulse lasers, the generation of the RF driver frequency by means of a synthesizer can be dispensed with. Instead, for example, an electronic diode signal of the pulse laser is made available (e.g. PD signal Out of the fs / ps NIR laser in Figure 1 or an equivalent structure by replacing / expanding the VIS laser module, e.g. with a ps diode laser etc.). The RF signal of the (sufficiently fast) photodiode generated in this way is synchronized a priori with the laser excitation. The RF signal can be used in the same way as the RF signal of a synthesizer, ie the phase shifter / amplifier PA unit is made available at the corresponding input (via an RF line). The further method is analogous to implementation 1. On the one hand, the use of a pulsed light source and the associated improvement in the signal / noise ratio in a lifetime contrast generated in this way are advantageous compared to the sinusoidal modulation of the laser excitation in implementation 1 and the omission of a synthesizer for generation the RF frequency. The fluorescence is detected after a 1-photon excitation (linear in the excitation intensity). 1.2 2-photon confocal lifetime imaging, homodyne

The subject of this chapter is the combination of 2-photon microscopy with homodyne detection for lifespan imaging.

The combination of time resolution and 2-photon microscopy has already been demonstrated by Gratton using the so-called heterodyne technique [So]. In the heterodyne technique, an fs laser is used for fluorescence excitation in a 2-photon microscope and the PMT detector is operated with a slightly different modulation frequency. The typical repetition rate of fs-Ti: Sapphire-Lasem is in the range of 80MHz. In [So] a frequency of 80MHZ + 25kHz is used for the gain modulation of the PMT detector. This creates a beat. This beat is digitized and scanned with an ADC card. This enables a pixel-by-pixel detection of the phase shift of excitation and fluorescence (heterodyne method).

In the structure shown here, in contrast to [So], the phase shift is determined not pixel by pixel but image by frame, i.e. first an image is recorded and stored at a relative phase shift φ1. In at least one further step, an image is then generated at a further different relative phase φ2. (Homodyne procedure).

The various images for the representation of the contrast are calculated analogously to Chapter 1.1.

method

A colored or self-fluorescent sample is excited to fluoresce via a two-photon absorption. The pulsed laser can be, for example, a Ti: sapphire laser (also possibly frequency doubled, frequency tripled sums or difference frequency mixed, etc.), a ps laser diode or a laser modulated, for example, by means of AOM or EOM. The physical process is in 1.1. a 1-photon excitation in 1.2 a 2-photon excitation (or generally multiphoton excitation). An fs-Ti: sapphire laser is used to excite the fluorescence in the implementation according to Figure 1. The repetition rate of the fs laser is around 80 MHz (NLO-LSM, fluorescence detection). A modulatable PMT detector from Hamamatsu (H6573) is used for the detection. The sinusoidal modulation frequency is generated via an ECL logic circuit from a photodiode signal from the laser (fs-Mira, Coherent) and is amplified to an average power of 1.5 W @ 50 W by means of an RF amplifier integrated in the phase shifters / amplifiers PA Output RF out of the PA available. The phase of the generated sine (and thus of the amplified sine) can be set to approximately 1 ° (as in Chapter 1.1). The electronic jitter is <100ps ((Figure 1), phase shifter / RF amplifier). The amplification of the PMT detector is modulated with the amplified sine at the frequency corresponding to the laser repetition rate (f.ep = fm _od ). For this purpose, the amplified RF frequency is made available to the input of the PMT module referred to in FIG. 1 as "RF modulation in". The current at the output of the PMT detector is smoothed by a passive LP filter or otherwise integrated by an integration circuit. The resulting DC signal is again made available to the detection unit of an LRM. The further data acquisition and evaluation takes place analogously to chapter 1.1.

The difference to 1.1 is that in one case a 2-photon (NLO) LSM is combined with the lifetime map - i.e. the fluorescence excitation is done with an fs or ps pulse laser. An essential difference is the type of fluorescence excitation (here 2 or more multi-photons (> 2) excitation in general.

1.3 Arrangement for phase-sensitive fluorescence detection in a laser scanning microscope

With 2 / multi photon excitation, the use of more efficient

Detection devices possible due to the deeply discriminated excitation. These are described in the literature under the collective term of non-descanned detection.

This is shown schematically in Fig.2, which on the scan lens SL and the

Attaches scanner of Fig.1 and represents a microscopic beam path, with the sample P, the objective lens OL, the beam splitter ST for coupling the Illumination / decoupling of the radiation coming from the sample and a first tube lens TL1.

A second tube lens TL2 is used in a detection beam path DE with another modulatable PMT directly, i.e. not via the scanning beam path.

This is connected to the control unit ST in Fig.1 and via this to other units in Fig.1.

A major disadvantage of these detection devices is generally the higher sensitivity to room light. Since the room light is generally not / or is modulated differently from the excitation light, the use of phase-sensitive detection devices for suppressing the room light makes sense. This eliminates the complex encapsulation of the detection unit. Arrangements with 1 and 2 / multi photon excitation have already been described above and can be completely transferred to a phase-sensitive detector. However, in the case of phase-sensitive detection, the reference signal (excitation light) and the measurement signal (fluorescence signal) have the same modulation. For the most commonly used TiSa laser systems, this modulation is 80 MHz. Thus homodyne detection takes place. For phase-sensitive detection, both signals are multiplied together in a multiplier with a fixed, adjustable phase relationship. In an advantageous arrangement, the first dynode of a modulatable photomultiplier (H6573) acts as a multiplier.

The reference signal is generated by the phase shifter / RF amplifier shown above. The clock for the phase shifter is the photodiode signal of the pulse laser. In addition, by "beating" the laser light with an optical modulator (AOM or EOM, see Figure 1), the repetition rate (repetition rate) specified by the pulse laser can be extended to a variable beat frequency range. The beat frequency is set by a synthesizer with an adjustable phase, which is controlled by trigger signals from the excitation light source (see Figure 1). The phase relationship between reference and measurement signal is set so that the phase shift is zero. The measured variable is thus demodulated. A DC or low-frequency modulated measurement signal is thus obtained. However, disturbances such as room light are strong after the multiplier modulated (like reference signal approx. 80 MHz). The highly modulated components (disturbance variables) are filtered out and thus suppressed by the downstream low-pass filter. An example of a low-pass filter is the detection electronics of a conventional LSM.

The advantage of the arrangement described is that the electronics on the detection side only have to be designed for low-frequency signals, since the demodulation already takes place in the photomultiplier.

bibliography

[Wabnitz] H. Wabnitz

Method for the determination of calcium ion concentration in cells Offenlegungsschrift DE 41 04 014, filing date 9.02.91

[Gröbler] B. Gröbler

Device for analysis and visual representation of the temporal intensity curve of the fluorescence radiation that occurs when a specimen is excited by laser light

Patent DE 36 14 359 C2

[Müller] R. Müller, M. Sauer, C. Zander

System for distinguishing fluorescent groups of molecules by time-resolved fluorescence measurement Patent WO 98/09154

[Han] K-T Han, M. Köllner, M. Sauer, S. Seeger, J. Wolfrum, C. Schmitt,

A. Schulz

Process for the optical sorting of plastics by means of time-resolved laser spectroscopy, publication DE 42 31 477 A1, filing date September 19, 1992

[Schbg] H. Schneckenburger, K. König, T. Dienersberger R. Hahn

Time-gated microscopic imaging and spectroscopy in medical diagnosis and photobiology

Optoical Engineering 33 (8) 2600 (1994)

[Cubeddu]] R. Cubeddu, G. Canti, A. Pifferi, P. Taroni, G. Valentini A real time system for fluortescence lifetime imaging SPIE 2976 (1997) 98

[Dowling] K. Dowling, S.C.W. Hyde, N.P. Barry, J.C. Dainty, P.M.W. French

Two dimensional fluorescence lifetime imaging using a 5 kHz 110 ps gated image intensifier

(WWW)

[Clegg] P: C: Schneider, R.M. Clegg

Rapid acquisition, analysis and display of fluorescence lifetime- resolved images for real-time applications Rev. Sei. Instrum. 68 (11) 4107 (1997)

[Gratton] E. Gratton, M. Limkeman

A continuously variable frequency cross-correlation phase foometer with picosecond resolution

Biophys. J. 44 (1983) 315 [Gratton2] E. Gratton, DM Jameson, RD Hall Multifrequency phase and modulation fiuorometry Ann. Rev. Biophys. Bioeng. 13 (1984) 105

[So] P.T.C So, T. French, W.M. Yu, K.K. Berland, CN. Dong, E. Gratton Time resolved fluorescence microscopy using two photon excitation Bioimaging 3 (1995) 49

[Goeppert] M. Göppert-Mayer

Ann. Physics 9, 273 (1931) [Wilson] T. Wilson

Theory and Practice of Scanning Optical Microscopy

Academic Press, 1984

Claims

claims

1.

Method for detecting fluorescence phenomena in a microscope, the sample being irradiated by a modulated and / or pulsed laser light source, and detecting the fluorescence in at least two different ones

Phase positions of the detector takes place.

Second

The method of claim 1, wherein an image is generated for each phase

Third

Method according to at least one of the preceding claims, wherein the images generated or images calculated from these images are displayed on a monitor.

4th

Method according to at least one of the preceding claims, wherein illumination and detection takes place via a laser scanning microscope.

5th

Method according to at least one of the preceding claims, wherein the detection is carried out with a modulatable PMT.

6th

Method for the detection of fluorescence phenomena in a microscope, the sample being irradiated by a modulated and / or pulsed laser light source, a reference signal corresponding to the modulated and / or pulsed laser and a measurement signal corresponding to a modulated detection being multiplied by a fixedly adjustable phase relationship and the

Result is used for visualization.

7th

The method of claim 6, wherein the phase relationship is set so that the

Phase shift is zero.

8th.

Method according to one of claims 6 or 7, with non-descanned detection.

9th

Application of a method according to one of the above claims to the mutiphoton excitation of fluorescence emission, in particular

Two-photon excitation

10th

Method according to at least one of the preceding claims, using a cw laser modulated by means of an acousto-optical modulator (AOM).

11th

Method according to at least one of the preceding claims, using a cw laser modulated by means of a Pockels cell.

12th

Method according to at least one of the preceding claims, using a pulse laser which is additionally modulated.

13th

Method according to at least one of the preceding claims, using a pulse laser which is additionally modulated by means of AOM.

14th

Method according to at least one of the preceding claims, using a pulse laser which is additionally modulated by means of a Pockels cell.

15th

Method according to at least one of the preceding claims, using for time resolution of multiphoton processes such as second

Harmony generation on surfaces or two-photon excitation.

16th

Method according to at least one of the preceding claims, using a pulsed near infrared (NIR) laser. Method according to at least one of the preceding claims, using a pulsed near infrared (NIR) laser with the following

Frequency conversion for 1-photon excitation.

18th

Method according to at least one of the preceding claims, using a pulsed near infrared (NIR) laser with the following

Frequency conversion for 1-photon excitation of fluorescence

19th

Method according to at least one of the preceding claims, using the phase-sensitive detection to improve the

Signal / noise ratio.