U.S. patent application number 12/444290 was filed with the patent office on 2009-10-08 for method and arrangement for collimated microscopic imaging.
Invention is credited to Michael Kempe, Gerhard Krampert, Matthias Wald, Ralf Wolleschensky.
Application Number | 20090250632 12/444290 |
Document ID | / |
Family ID | 38963139 |
Filed Date | 2009-10-08 |
United States Patent
Application |
20090250632 |
Kind Code |
A1 |
Kempe; Michael ; et
al. |
October 8, 2009 |
Method and Arrangement for Collimated Microscopic Imaging
Abstract
A method and arrangement for collimated microscopic imaging,
including a first illumination of a sample in at least one region
for exciting fluorescence, and a spatially resolving detection of
the sample light by detector elements, the detection being
associated with the region, wherein by means of a second
illumination a sub-division of the region into separate fluorescent
partial regions occurs, which are associated with the detector
elements. The separation of the partial regions is carried out by
the spatial separation of the fluorescent regions by means of
intermediate regions having reduced fluorescence or no
fluorescence, and/or by means of different spectral properties of
the fluorescence from the partial regions.
Inventors: |
Kempe; Michael; (Jena,
DE) ; Krampert; Gerhard; (Jena, DE) ; Wald;
Matthias; (Kunitz, DE) ; Wolleschensky; Ralf;
(Jena, DE) |
Correspondence
Address: |
DUANE MORRIS LLP - NY;PATENT DEPARTMENT
1540 BROADWAY
NEW YORK
NY
10036-4086
US
|
Family ID: |
38963139 |
Appl. No.: |
12/444290 |
Filed: |
October 2, 2007 |
PCT Filed: |
October 2, 2007 |
PCT NO: |
PCT/EP07/08556 |
371 Date: |
April 3, 2009 |
Current U.S.
Class: |
250/459.1 ;
250/237G; 250/458.1 |
Current CPC
Class: |
G02B 21/0072 20130101;
G02B 21/16 20130101; G01N 21/6458 20130101; G02B 21/0032 20130101;
G02B 21/0076 20130101 |
Class at
Publication: |
250/459.1 ;
250/458.1; 250/237.G |
International
Class: |
G01N 21/64 20060101
G01N021/64 |
Foreign Application Data
Date |
Code |
Application Number |
Oct 6, 2006 |
DE |
10 2006 047 912.2 |
Claims
1. A method for collimated microscopic imaging with a first
illumination of a sample in at least one region for exciting
fluorescence and a spatially resolving detection of the sample
light by detector elements which is associated with the region,
comprising separating said region by a second illumination into
separate fluorescing partial regions which are associated with said
detector elements, wherein said separating of said region into
partial regions comprises spatially dividing said fluorescing
regions by means of intermediate regions with reduced fluorescence
or no fluorescence and/or by different spectral properties of the
fluorescence from said partial regions.
2. The method according to claim 1, wherein said second
illumination is spatially structured or patterned.
3. The method according to claim 1, wherein the different spectral
properties are a fluorescence of different wavelength or a
different spectral distribution.
4. The method according to claim 1, wherein a reduction of the
extent of the partial regions by de-excitation (depopulation)
and/or switching the fluorescence state are/is carried out by said
second illumination.
5. The method according to claim 1, wherein the reduction of the
extent of the partial regions is carried out by means of nonlinear
interactions of said second illumination with at least one
substance capable of fluorescing which is present in the sample,
and wherein a) a de-excitation of an excited level is carried out
by stimulated emission (STED) and/or b) a depopulation of a ground
state is carried out by triplet occupation (GSD) and/or c) the at
least one substance capable of fluorescing is driven from a state
with first fluorescence characteristics into a state with second
fluorescence characteristics or reduced fluorescence by reversible
or irreversible optical switching.
6. The method according to claim 1, wherein a definitive
correlation of the partial regions to the detector elements is
carried out by optical imaging of the fluorescence emitted by the
said partial regions onto the detector.
7. The method according to claim 1, wherein the partial regions are
associated with the detector elements in such a way that the center
of gravity of the partial regions is imaged substantially in the
center of the detector elements.
8. The method according to claim 1, wherein the samples and
patterns of the second illumination are moved relative to one
another for recording an image with enhanced resolution, and
different partial regions are accordingly scanned in the sample,
and the fluorescent light emitted by these partial regions is
detected.
9. The method according to claim 1, further comprising measuring
said second illumination of region A being carried out without a
pattern, and detecting fluorescence excited by the first
illumination in a spatially resolved manner.
10. The method according to claim 9, wherein the image obtained by
said measuring step is applied to the image recorded with enhanced
resolution in order to achieve a reduction in the background signal
in the image with enhanced resolution.
11. The method according to claim 1, wherein said region extends in
a line-shaped manner.
12. The method according to claim 11, wherein the extension
perpendicular to the line is diffraction-limited.
13. The method according to claim 1, wherein the second
illumination has a patterned light distribution which comprises a
periodic pattern along the direction of the illumination line and
at least two laterally and/or axially limiting outer lines.
14. The method according to claim 1, wherein the fluorescent light
which is excited outside the focus is suppressed by a variably
adjustable slit diaphragm in front of the detector elements.
15. The method according to claim 1, wherein said region is formed
of a plurality of line-shaped sub-regions which are spatially
separated from one another.
16. The method according to claim 1, wherein the fluorescent light
which is excited outside the focus is suppressed in an adjustable
manner by a variable selection and a variable combination of
detector elements.
17. Method for operating a fluorescence microscope, wherein a
structured illumination of the sample is carried out, comprising a
first detecting of luminescing sample points, spatially offsetting
of the illumination distribution on the sample, a second detecting
of luminescing sample points, and determining and storing the
position of the detected sample points on a compiled sample image
from the values of the spatial offset.
18. An arrangement for collimated microscopic imaging comprising at
least two light sources for respectively at least a first
illumination and second illumination of a sample, means for
generating a patterned light distribution on said sample, at least
one detector which is spatially resolving in at least one direction
and has a plurality of detector elements, wherein fluorescence is
excited by a patterned or spatially structured light distribution
in at least N partial regions of said sample which are separated
from one another, and wherein said partial regions are imaged on
the detector in such a way that a definitive association of partial
regions and detector elements is carried out.
19. The arrangement according to claim 18, wherein said pattern is
moved by moving a beam-deflecting element in the illumination beam
path and/or the sample is moved by a scanning table, and wherein a
detection of the fluorescent light emitted by the sample is carried
out so as to be synchronized with said movement.
20. The arrangement according to claim 18, wherein a patterned or
spatially structured second illumination is superimposed with an
unpatterned first illumination in the sample.
21. The arrangement according to claim 18, wherein said association
of partial regions and detector elements can be adjusted by means
of an optical zoom and by tilting a transparent, plane-parallel
plate inserted in the illumination beam path and/or detection beam
path.
22. The arrangement according to claim 18, wherein said
illumination is line-shaped.
23. The arrangement according to claim 18, wherein the pupil of the
illumination beam path is completely filled in one spatial
direction.
24. The arrangement according to claim 18, wherein the patterned or
spatially structured second illumination is split, wherein one part
generates a line-shaped, periodically patterned light distribution
in the samples and another part forms double lines which limit the
periodic light distribution axially and/or laterally.
25. The arrangement according to claim 18, further comprising phase
masks are provided in the illumination beam path for generating the
limiting double lines, which phase masks have a phase jump in their
center and/or in a symmetrical edge area.
26. The arrangement according to claim 18, further comprising an
optical grating for at least partial splitting of the partial beams
of the second illumination, wherein the splitting is carried out by
the zeroth and +/- first diffraction order of the grating.
27. The arrangement according to claim 26, wherein said optical
grating is designed so as to be variable perpendicular to the
structuring direction in frequency and/or in the magnitude of the
phase jump so that the direction and/or the intensity of the
diffraction orders can be adjusted by translation of the
grating.
28. The arrangement according to claim 25, wherein said phase masks
are arranged after the zeroth diffraction order of the grating.
29. The arrangement according to claim 25, wherein said phase masks
have a phase jump in their center or in a symmetrical edge
area.
30. The arrangement according to claim 25, wherein said phase masks
are designed so as to be variable in position and/or in the
magnitude of the phase jump in a direction perpendicular to the
line-shaped illumination so that different characteristics of the
influencing of the light can be adjusted by translation of the
phase masks.
31. The arrangement according to claim 18, wherein the axially and
laterally limiting double lines and the periodically patterned
light distribution overlap one another incoherently in the
sample.
32. The arrangement according to claim 18, wherein the incoherence
of the overlapping is affected by polarization orthogonal to one
another or by different wavelengths or by a retardation of the
overlapping partial beams.
33. The arrangement according to claim 18, wherein the spatially
resolving detector is a row detector in front of which is arranged
a variable slit diaphragm.
34. The arrangement according to claim 18, wherein an illumination
is provided in parallel with a plurality of lines and detection by
a spatially resolving, planar detector.
Description
RELATED APPLICATIONS
[0001] The present application is a U.S. National Stage application
of International PCT Application No. PCT/EP2007/008556 filed on
Oct. 2, 2007, which claims benefit of German Application No. DE 10
2006 047 912.2 filed on Oct. 6, 2006, the contents of each are
incorporated by reference in their entirety.
FIELD OF THE INVENTION
[0002] The present invention relates to methods and arrangements
for microscopic imaging with structured illumination. Increased
spatial resolution is achieved by means of nonlinear interactions
between samples, which is known from the prior art. By means of the
method and the corresponding arrangements of this invention, this
sample interaction is made use of for imaging in such a way that
confocal imaging is possible with a resolution that goes beyond the
diffraction boundary in all spatial directions with parallel data
acquisition.
PRIOR ART
[0003] The prior art discloses various nonlinear sample
interactions which can lead to enhanced spatial resolution when
suitably implemented, specifically:
1. De-excitation of the excited level through stimulated emission
(Stimulated Emission Deletion--STED), Klar and Hell, Opt. Lett. 24
(1999), 954-956 2. Depopulation of the ground state by triplet
occupation (Ground State Depletion--GSD), Hell and Kroug, Appl.
Phys. B 60 (1995). 495-497 3. Reversible or irreversible dye
switching between a fluorescing state, a non-fluorescing state, a
reduced fluorescing state, or a fluorescing state characterized by
other features (such as a different emission wavelength), Hell,
Jakobs and Kastrup, Appl. Phys. A 77 (2003), 859-860.
OBJECTS
[0004] The basic idea behind the method of the invention is that
the dye is either
[0005] a) brought to a non-fluorescing or slightly fluorescing
state in a spatially limited region so that a subsequent excitation
can lead to fluorescence only, or predominantly, in a limited
region (second and third method), or
[0006] b) is de-excited in a spatially limited region following
excitation so that fluorescence is carried out only from a limited
region (first method).
[0007] In both cases, the intention is to achieve a fluorescing
region which is smaller than that which can be achieved by
diffraction-limited excitation.
[0008] The basic principle is illustrated in FIG. 1 with the steps
and schematically depicted spatial distributions which are used or
which result. The arrangements known from the prior art are based
on point scanning methods in which light distributions with zero
settings in the center (doughnut modes, as they are called) which
are preferred for depopulation, switching and de-excitation are
applied. The laser beam is scanned over the sample and the steps
shown in FIG. 1 are carried out sequentially in every spatial
point. Previous arrangements according to the first method (e.g.,
Klar et al, PNAS 97 (2000), 8206-8210) and third method (Hoffmann
et al., PNAS 102 (2005), 17565-17569), all of which function
according to the scheme mentioned above, have been described. A
decisive disadvantage of this arrangement is the sequential data
acquisition. Owing to the fact that the increased spatial
resolution results in reduced excitation volumes, the fluorescence
emission is reduced so that the pixel integration time must
generally be longer (with a fivefold increase in resolution only
laterally, the expected fluorescence emission is about twenty-five
times lower, e.g., with a homogeneous dye distribution over the
extent of diffraction-limited excitation). Further, some known
switchable dyes such as Dronpa can be switched only with limited
light power so as to afford many switching cycles (Habuchi et al.,
PNAS 102 (2005) 9511-9516). In this case, the third method
mentioned above requires a considerably longer exposure time per
pixel for switching off than that required for fluorescence
excitation alone.
[0009] Therefore, it is clear that a parallel data acquisition is
needed in order to arrive at acceptable image compilation times. In
this connection, it must be taken into consideration that confocal
imaging is required in order to avoid a loss of contrast and
reduced resolution enhancement due to extrafocal fluorescence. A
possibility offered by the prior art for parallelizing is
simultaneous excitation with a plurality of spots, each of whose
fluorescence characteristics are modified by a doughnut-shaped
illumination. However, the doughnut-shaped illumination must lead
to a nonlinear sample interaction, i.e., a saturation, in order to
achieve increased resolution. This means that a relatively large
area around the individual spot interacts with this light
distribution, which only allows a low density of the spot and,
therefore, meager parallelization. Further, a confocal detection in
a multi-spot arrangement of this kind is cumbersome in technical
respects.
SUMMARY OF THE INVENTION
[0010] A method and an arrangement in which a high-resolution image
is achieved directly (without image processing) with parallel data
acquisition is described in the following. The extrafocal
background can be eliminated by variable confocal detection. This
method and arrangement prevent the above-mentioned disadvantages of
the multi-spot arrangement according to the prior art. In
particular, an optimal parallelization (maximum density of the
simultaneously excited regions of the sample) and a relatively
simple technical implementation are made possible by the
invention.
BRIEF DESCRIPTION OF THE DRAWINGS
[0011] The present inventions will be described in greater detail,
using examples with reference to the annexed drawings, in
which:
[0012] FIG. 1 is a diagram schematically illustrating sequential
steps of known point scanning methods;
[0013] FIG. 2 is a diagram illustrating the switching beam
distribution;
[0014] FIG. 3 illustrates the camera pixels and light distribution,
object point allocation in the scanning steps, the resulting line
image, and the resulting two-dimensional image after a Y-scanning
step;
[0015] FIG. 4 shows the lateral (X/X) and axial (optical axis Z)
light distribution in the object plane;
[0016] FIG. 5 schematically shows the axial light distribution
which would occur without axially structured switching light,
together with the axial structuring of the switching light;
[0017] FIG. 6 is a schematic optical diagram which shows splitting
in only n=3 lines;
[0018] FIG. 7 schematically shows a modified line scanning
arrangement;
[0019] FIG. 8 illustrates a phase mask for lateral structuring;
[0020] FIG. 9(a) illustrates an achromatic color splitter which
results from the pupil light distributions in FIG. 8;
[0021] FIG. 9(b) illustrates the achromatic color splitter located
in the beam in a conjugate pupil plane;
[0022] FIG. 10(a) shows the distribution without a mask
(corresponds to excitation distribution);
[0023] FIG. 10(b) shows the distribution with a mask according to
FIG. 8(b);
[0024] FIG. 10(c) shows the distribution with a mask according to
FIG. 8(a);
[0025] FIG. 10(d) shows incoherent overlapping results in the
distribution shown;
[0026] FIG. 11(a) is a graph illustrating intensity versus time
showing switching on and off,
[0027] FIG. 11(b) is a graph illustrating intensity versus time
showing switching off;
[0028] FIG. 12 is a schematic view showing a preferred illumination
unit for a laser scanning microscope for use with Dronpa;
[0029] FIG. 13 shows the integration of an illumination unit
according to the invention in a line-scanning microscope;
[0030] FIGS. 14(a)-(c) illustrate adjustment gratings for
masks;
[0031] FIG. 15 shows the intensity distribution of the illumination
in the object plane using the phase mask from FIG. 8(a).
[0032] FIG. 16 shows a molecule population in the excited state
with saturated switching off;
[0033] FIG. 17 illustrates an embodiment example of the
illumination;
[0034] FIG. 18(a) illustrates an intensity structure; and
[0035] FIG. 18(b) illustrates an axial structuring of the
illumination light in the object plane.
DESCRIPTION OF EMBODIMENTS
[0036] Basic Principles
[0037] An arrangement which meets the demands mentioned above is
based on a line-shaped excitation and confocal detection by means
of a line camera behind a slit diaphragm. By means of a suitable
exposure pattern in combination with the nonlinear sample
interaction, a de-excitation/depopulation/switching (always
referred to hereinafter, by way of example, as switching) is
realized in such a way that only a series of spots along a line can
emit fluorescent light. The distance between these spots is at
least as great as (but advantageously no greater than) the
diffraction-limited resolution of the optical system. A possibility
for generating an illumination of this kind (which is
simultaneously also axially structured) consists in the diffraction
of coherent light at a periodic structure and interference of the
diffraction orders in the object plane, as is described in the
application DE102004034962A1 which is hereby incorporated in the
present disclosure. This results in diffraction-limited light
distributions on the camera line which are separated from one
another, impinge on corresponding pixels and are detected
separately at the latter (see FIG. 2).
[0038] FIG. 2 shows, from top to bottom, the switching beam
distribution in the object plane, the excitation distribution in
the object plane, the superposition of excitation and switching
light in the object plane, the camera pixels, and the light
distribution in the camera plane. In the switching beam
distribution, X represents the zeroth diffraction order referring
to the diagram and description in FIG. 9 a) and FIG. 11. XX is the
+/- first diffraction order according to the diagram and
description in FIG. 9 a) and FIG. 11.
[0039] A highly resolved image is now obtained by scanning the
object along line (x) and perpendicular to line (y). This scanning
can be carried out by moving the illumination distributions, e.g.,
by means of a galvanometer scanner and/or by moving the object. In
the former case, the illumination beam and the detection beam must
pass over the same beam deflection elements in order to obtain a
stationary light distribution on the detector (descanned
detection). In case the object is simply moved (object scanning),
the light distributions are stationary in every case.
[0040] The information about the correlation of pixel to
measurement point on the sample is derived in a definitive manner
from the scanning movement and the corresponding optics. This will
be described more exactly: the image field raster observed by the
detector is determined by the imaging of the camera pixels on the
sample. For example, 512 pixels are imaged on the sample in such a
way that every pixel detects an image surface of 1.5-times the
diffraction-limited resolution (for an objective with an NA of 1.4,
this corresponds at a wavelength 488 nm to a field of 488
nm/(2*1.4)*1.5=260 nm edge length). Adjustment of the illumination
pattern (switching light distribution) ensures that there is a
unique correlation between the luminous spots in the sample and the
camera pixels. By means of the relative movement of the object and
illumination pattern, different sample points are imaged. For this
purpose, the scanning movement and the detection of the spot size
in the sample which is synchronized with the scanning movement must
be adapted. Typically, scanning is carried out with one half of the
spot diameter (Nyquist theorem). Thus when a fivefold increase in
resolution (N=5) is achieved in the example given above by mean of
the nonlinear sample interaction, a sample scanning must be carried
out at a distance of 260 nm/(5*2)=26 nm. Therefore, N*2
scan/detection steps (10 in the example above) are needed to
completely scan a line. Therefore, 512*2*N highly resolved data
points are obtained from 512 pixels in x-direction (see FIG.
3).
[0041] FIG. 3 shows, one below the other, the camera pixels and
light distribution, object point allocation in scanning steps 1 to
10, the resulting line image, and the resulting two-dimensional
image after a Y-scanning step.
[0042] The discussion has so far been limited to one object plane.
However, objects are always three-dimensional. This has several
consequences. For one, aside from the fluorescence in the focus
with increased resolution, fluorescence is also excited outside of
the focus. This fluorescence interferes with the imaging described
above and can make it impossible to correlate the pixels to
high-resolution object points. A remedy is afforded by confocal
detection which effectively suppresses extrafocal fluorescent
light. In the case presently under consideration, this is realized
by means of an (ideally) variable slit diaphragm in front of the
camera parallel to the line. On the other hand, the
three-dimensional point spread function (PSF) must be taken into
consideration for imaging. This will depend upon the switching
light distribution as well as on the excitation PSF and detection
PSF. In axial direction, the PSF can correspond to the
diffraction-limited PSF (structured switching light only in lateral
direction). FIG. 4 shows the lateral (X/X) and axial (optical axis
Z) light distribution in the object plane. However, an axial
resolution enhancement can also be achieved (FIG. 5) by means of
additional structuring of the switching light in axial direction.
FIG. 5 shows schematically the axial light distribution which would
occur without axially structured switching light, together with the
axial structuring of the switching light. The switching light
distribution results, e.g., from the use of an optical element from
FIG. 8 b) or FIG. 8 c) in an arrangement according to FIG. 12
(including the associated descriptions). The light distribution
actually obtained in this case is shown at the bottom in FIG. 5.
Depending upon the application, a distribution such as that in FIG.
4, bottom, or FIG. 5, bottom, can be useful. Therefore, it is
advisable to be able to implement both scenarios in one device.
[0043] In principle, all of the above considerations can also be
applied to a two-dimensionally structured switching beam
distribution and detection with an area detector. In this case, the
excitation and the switching on which may possibly be necessary are
carried out with (far-field) area illumination. An arrangement of
this kind would have the advantage of increased parallelism (e.g.,
512.sup.2 points could be exposed simultaneously), but also has the
disadvantage of non-confocal detection. The occurrence of
extrafocal fluorescence can pose a severe problem depending on the
sample (particularly with thick, strongly dyed samples).
[0044] An option for the use of increased parallelism while
simultaneously preserving a certain confocality of the imaging is
to image a plurality of separate lines on the sample and to carry
out detection by means of an area detector. When the lines on the
area detector are sufficiently separated (e.g., by about 10
pixels), a confocal detection can be achieved by selectively
reading out the pixels associated with the lines. The confocality
can be adjusted by also taking into account adjacent pixels. For
example, by reading out the two pixel rows adjacent to the line and
summing the corresponding associated pixel elements, the effective
"slit diaphragm" can be increased by a factor of 3 with a
corresponding decrease in confocality. The same elements in the
vicinity of the pupil as those shown in FIGS. 8 and 9 can be used
for an arrangement of this kind in the illumination beam path.
However, the light must be split into n partial beams impinging in
the pupil at different angles by means of an element arranged in
front of the main color splitter. This results in n lines which are
separated from one another in the object plane, every line being
shaped in the same manner as shown in FIG. 4 and FIG. 5. An
element, mentioned above, placed in front of the main color
splitter can be a suitably shaped diffractive element (or a
combination of a plurality of elements) with corresponding imaging
or an arrangement for the geometric splitting and beam deflection
of the partial beams. An example of this is shown in FIG. 6 which
shows splitting in only n=3 lines for the sake of clarity.
[0045] Preferred Arrangement
[0046] A preferred arrangement uses a modified line-scanning system
as shown in FIG. 7 and, for example, in DE 10257237A1 and EP1617271
A2 which are hereby incorporated in the present disclosure.
[0047] The modifications involve the following aspects:
[0048] The AchroGate achromatic color splitter is replaced by one
permitting a structuring of the switching light in the object
plane--this can be a suitable achromatic design or a dichroic
beamsplitter.
[0049] The illumination unit is replaced by one which enables the
required nonlinear sample interactions and the structuring of the
switching light.
[0050] The illumination and the data acquisition are adapted in
such a way that the exposure of the sample and the detection with
the suitable sequence of switching/excitation/detection/switching
on (see FIG. 1) is carried out within a suitable period of
time.
[0051] The data evaluation is adapted so that a correlation of
pixels to object points is carried out as shown schematically in
FIG. 3.
[0052] Structuring of the Switching Light
[0053] A preferred way to generate a lateral structuring is made
possible by means of a phase mask according to FIG. 8 a) which
should be imaged in the vicinity of the pupil of the objective. In
this connection, reference is had to the beam path shown in FIG.
12. The mask is divided in the middle, the right half and the left
half producing different phase shifts in the light passing through.
A double line--as is shown at the top in FIG. 2--is achieved by
means of a phase jump of .pi. for half of the center line. A
sine-shaped modulation of the line in the intermediate space of the
double line (to produce a modulation of the excitation in the
intermediate space of the double line, see FIG. 2, top) is
generated by the interference of the two line distributions at the
edge of the pupil. To achieve a modulation of the line at 2/3 of
the limiting frequency, for example, the lines must have a spacing
equal to 2/3 of the diameter of the pupil radius. It is important
that the components of the light which generate the double line and
those which generate the modulated center line overlap incoherently
in the sample. This can be ensured by delaying the arrival of the
corresponding light components in the object by a time greater than
the coherence length. However, it is preferable simply to adjust
the polarization of the light components orthogonal to one another
as is illustrated schematically by arrows in FIG. 8. This can be
accomplished, e.g., by means half-wave plates which selectively
rotate the light in the outer area of the pupil (see FIG. 8). A
suitable phase grating can be used in the vicinity of a conjugate
image plane in front of the mask in order to obtain the
distribution of the light on the phase mask shown in FIG. 8a). The
center line then corresponds to the zeroth diffraction order, and
the two lines at the edge of the pupil correspond to the first
diffraction order of the corresponding grating. The optimal
position of the light distributions in the object plane relative to
one another is ensured at the same time through the use of
diffraction orders of a grating. Because of the incoherence of the
first diffraction orders relative to the zeroth diffraction order
that is achieved by the steps described above, it is also ensured
that the center line in the object plane can be modulated up to the
limiting frequency of the imaging of incoherent light (which is
twice as great compared to imaging with coherent light). Another
consequence consists in that no structuring of the light is carried
out in axial direction, in contrast to the arrangement described in
DE102004034962 A1.
[0054] If an axial structuring of the switching light--as is shown
schematically at the top of FIG. 5--is also desired, another
partial beam of the switching light is modified with a phase mask
according to FIG. 8 b) located in the vicinity of a plane conjugate
to the objective pupil.
[0055] Masks 8 b) and 8 c) have a central region symmetric to the
optical axis and to an axis perpendicular to the optical axis which
produces a different phase delay of the passing light compared to
the outer regions to the right and left sides. A phase delay of
.pi. of an outer portion of the line-shaped pupil illumination
relative to an inner portion of the line-shaped pupil illumination
is carried out. In theory, the optimal radius of the phase jump is
1/ {square root over (2)} of the pupil radius of the objective. In
practice, a somewhat smaller radius is often optimal. A phase plate
such as that shown in FIG. 8 c) is suitable for optimizing the
radius. Adaptation of the radius can be carried out by
translation.
[0056] FIG. 9 a) shows an achromatic color splitter which results
from the pupil light distributions in FIG. 8 and which is also
compatible with the distributions for excitation and switching on
(lines). This achromatic color splitter is located in the beam in a
conjugate pupil plane as is shown in FIG. 9 b). The fluorescent
light emitted by the sample fills the pupil and therefore passes
the splitter (except at the reflecting strips). The outer mirror
surfaces (X) are used for transmitting the +/- first diffraction
order of S1 in FIG. 11, the inner mirror surface (XX) is used to
transmit the zeroth diffraction order of S1 in FIG. 12 and the beam
paths S2, S3. It must be ensured, as the case may be, that the
mirror surfaces do not all lie in the same focal plane. In case the
Rayleigh range of the light focused in this plane is less than the
focus deviation of the outer mirror surfaces, the defocusing due to
the magnification of the outer mirror surfaces (which are not in
the focus) compared to the inner mirror surface must be taken into
account.
[0057] FIG. 9 shows experimental results (sections through the
light distributions in the sample) with the corresponding masks
with an oil-immersion objective having a numerical aperture (NA) of
1.4 at 488 nm. FIG. 10 a) shows the distribution without a mask
(corresponds to excitation distribution), FIG. 10 b) shows the
distribution with a mask according to FIG. 8 b), and FIG. 10 c)
shows the distribution with a mask according to FIG. 8 a) (only the
center portion). Finally, an incoherent overlapping results in the
distribution shown in FIG. 10 d) which shows a minimum which is
surrounded in the y- and z-directions. This minimum extends to
about 170 nm (y) and 400 nm (z), which corresponds roughly to the
inverse of the diffraction-limited limiting frequencies (lateral:
.lamda./(2NA).apprxeq.170 nm, axial:
.lamda./(n- {square root over (n.sup.2-NA.sup.2)}).apprxeq.520
nm).
[0058] The point pattern of intensity minima occurring in the focus
in the sample caused by a corresponding nonlinear sample
interaction is a point pattern of excitable regions (GSD/switching)
or of regions in which there is still significant excitation
(STED). These regions are appreciably smaller after suitable
exposure optimized for the dye (wavelength, intensity, exposure
time) than a diffraction-limited light distribution and therefore
allow a scanning of the sample with increased resolution. The
adjustment of the illumination is carried out based on knowledge of
the dye and can be optimized at the device based on the resolution
of test structures (e.g., beads). Since this point pattern must be
imaged exactly on the pixels of the line camera, a variable optical
system (zoom) which makes it possible to adapt the periods of the
pattern to the periods of the detector elements is advantageously
provided in the illumination beam path or detection beam path. The
position of the imaged spots (i.e., their center of gravity) can be
centered within the pixels by means of another element such as a
tiltable, thick glass plate in the illumination beam path or
detection beam path (see, e.g., DE 102004034960A1).
[0059] Excitation and Switching On
[0060] The excitation and switching on are carried out with
diffraction-limited light distributions which are generated without
a mask. In the preferred embodiment example, these are lines which
have the most homogeneous possible intensity curve along the line
and extend in a diffraction-limited manner perpendicular to the
line.
[0061] The excitation is carried out either before the
switching/de-excitation (STED) or after the dye has been switched
off (GSD/switching) by the switching light in a structured manner.
Only in the case of switching is it generally necessary to
subsequently reverse the switching off by a switch-on laser,
because the other processes spontaneously revert to the initial
state of the dye (see also FIG. 1 with respect to the timing). In
case pulsed lasers are used (which is generally required in STED),
the time sequence can be achieved by retarding the corresponding
partial beams. When using continuously emitting (cw) lasers, the
time sequence must be achieved by means of fast switches
(preferably AOTFs).
[0062] The optimal parameters for excitation and switching on are
determined beforehand from knowledge of the dye characteristics.
For switching on, the aim is to reverse the switching off as
completely as possible; optimized excitation and simultaneous
detection of the emitted fluorescence signal is carried out by
maximizing the signal-to-noise ratio. Different boundary conditions
must be taken into account for the different interaction
mechanisms: [0063] With STED, only the number of excited molecules
remaining after excitation and subsequent de-excitation can
contribute to fluorescence. For a detection of photons going beyond
the one-time emission of these molecules, it is necessary to repeat
the process (excitation--de-excitation--fluorescence detection).
[0064] With GSD, excitation can be carried out and the resulting
fluorescence detected only until the molecules spontaneously revert
back from the triplet level. This is a comparatively short time
period on the order of 1 .mu.s. [0065] When switching dyes, the
excitation wavelength and the switching wavelength are identical in
some cases. If this switching happens to be the switching off of
the dye (as with Dronpa), the detection of fluorescence photons
must be prevented during the targeted switching with the switching
light distribution or separated from the detection of the photons
from the spatially limited region. Further, the fluorescence signal
can be detected only until the dye is switched off even in the
spatially limited region. Otherwise, there are no other
restrictions provided the dye remains stable in the switched off
state. Even when the excitation wavelength and the wavelength for
switching on the molecule are identical (as is known, e.g., for the
asFP585 protein, Hoffmann et al., PNAS 102 (2005), 17565-17569),
this does not pose any problems provided the fluorescence
excitation is more efficient than the switching process and the
excitation outputs, switching outputs and exposure times are
adapted in a corresponding manner.
[0066] Generally, not all light-emitting molecules in a sample take
part in the nonlinear sample interaction (STED/switching/GSD). This
leads to a background signal which is not subjected to the
nonlinear beam shaping and accordingly limits the high resolution
that can be achieved. In case of long-lived states (i.e.,
approximately >1 .mu.s) as in GSD and switching, the magnitude
of the background relative to the signal of the molecules
participating in the nonlinear sample interaction can be gauged by
a further recording before switching on again (switching) or before
the spontaneous relaxation into the ground state (GSD). This is
because the unstructured excitation (S2, FIG. 11), e.g., in
photoswitches such as Dronpa, in addition to the fluorescence
excitation also leads to the switching off of the Dropna molecules
(similarly the excitation also leads to an occupation of the
triplet state in GSD). Therefore, in further (unstructured)
excitation and detection, only those molecules which are not
subjected to the photoswitching process (or GSD) emit light. The
procedure is similar with switchable molecules in which the
fluorescence excitation leads to the switching on of the molecule
(as in asFP, Hoffmann et al., PNAS 102 (2005), 17565-17569).
However, the dye must initially be switched off in an unstructured
manner in an additional step in this case and the fluorescence
excited by unstructured illumination is then detected.
EXAMPLE
Switching of the Dronpa Protein
[0067] A suitable candidate for switching is the Dronpa protein
(Habuchi et al., PNAS 102 (2005) 9511-9516). In this case, the
switching off and the excitation are carried out with a wavelength
in the range of about 450 nm to 520 nm. Switching on can be carried
out in the range of 350 nm to 420 nm. This makes it possible to
work, e.g., with the lines of the argon laser at 477 nm and 488 nm
and to use a laser diode at 405 nm for switching on. The mutual
switching on and off is shown in FIG. 11 a). A detailed view of
switching off is represented in FIG. 11 b). This illustrates the
long exposure times of about 5 ms for achieving a sufficient
switching off. While the exposure time can be shortened by
increasing the intensity, this would reduce the number of possible
switching cycles compared to FIG. 11 a) in which about 100
switching cycles are shown.
[0068] FIG. 12 shows a schematic view of a preferred illumination
unit for a laser scanning microscope for use with Dronpa. S1
represents the beam path for generating the switching light
distribution (see FIGS. 2, 7 and 5). The dye is switched on in this
case either before switching off or after excitation. S2 is the
excitation beam path, S3 is another beam path for switching on a
switched off dye, if necessary. Accordingly, S1 (switching off) is
carried out first, followed by S2 (excitation) and S3 (switching
on) either at the very start or at the very end, wherein detection
is carried out at S2. The wavelength 477 nm at S2 can also be
replaced with another excitation wavelength (e.g., 488 nm).
[0069] STED and GSD (see the introductory part of the
specification) can be realized by means of S1 and S2, STED can be
realized by S2 followed by S1, and detection can be realized in S1,
GSD by S1 followed by S2, and detection in S2. In S1, after the
beam shaping (e.g., by means of a Powel lens) and after passing
through a splitter DC in the zeroth diffraction order of a grating
followed by the mask 7 a) (phase jump), the two laterally limiting
outer lines of the switching beam are generated in the sample and
the structuring of the center line is generated by means of the +/-
first diffraction order.
[0070] The mask 7 b) or c) (after reflection at the splitter DC and
separate beam path until after the mask 7c)) generates the limiting
lines of the switching beam in axial direction by means of the
phase jump described above.
[0071] Slightly different wavelengths (in this case, 488 nm, 477
nm) can be used for the beam paths through 7 a) and 7 c), but they
must both lie within the range of the response behavior of the
respective dye.
[0072] The integration of an illumination unit according to the
invention in a line-scanning microscope is shown in FIG. 13. It
contains the above-mentioned adjusting plate in the detection beam
path for positioning the spot on the row of detectors. An
additional zoom in the system serves to adjust the imaging of the
illuminated line in the object on the row detectors. Accordingly, a
selected portion of the line can be imaged on the row detectors.
The definitive correlation of the illumination spot and camera
pixels must be ensured in every case by the interaction of the two
zooms.
[0073] In order to achieve a flexible system for switching
additional dyes (also simultaneously with Dronpa), additional
lasers are connected and the design of the dichroic splitter DC is
provided for a plurality of colors (or as a neutral splitter).
Further, it may be necessary (depending upon the wavelength
difference) to change masks (although simultaneous operation would
not then be possible).
[0074] A simple adaptation consists in changing the grating
constant so that the diffraction orders occur at the same location
in the pupil. This can be achieved by a translation of a grating as
is shown in FIG. 14 a). With moderate wavelength changes (on the
order of 5%), the masks can also be used simultaneously with a
plurality of wavelengths.
[0075] Also, when a wavelength is used, the use of different
objectives, for example, creates the need for adjustment
possibilities in the system. In this connection, a basic adaptation
of the illumination to the different pupils of the objectives is
carried out by means of the zoom system directly following the main
color splitter (FIG. 8). Another specific adaptation is the
relative proportioning of the intensity in the diffraction orders,
which can result from the different losses of the orders when
changing objectives. Further, this makes it possible to adapt
specifically to the object. FIG. 14 b) shows a preferred form of
adapting by changing the effective gap ratio at the grating while
maintaining the same grating constant. Adapting the radius of the
phase jump of the mask in FIG. 8 b) is particularly critical for
optimal structuring in axial direction. A precision optimization of
this radius, e.g., based on the recording of PSF, with the
objective to be used is possible with a mask design such as that
shown in FIG. 12 c). The effective radius of the phase edge can be
adapted by translation of the mask.
[0076] Simulation of the Intensity Distribution for Dronpa
[0077] FIG. 15 shows the intensity distribution of the illumination
in the object plane using the phase mask from FIG. 8 a). The
quantity n of excited molecules decreases exponentially with the
illumination intensity and the exposure duration. Using Dronpa as
an example, FIG. 16 shows the resulting molecule population in the
excited stated with saturated switching off, which provides a
measurement for the PSF.
[0078] An Optical Embodiment Example
[0079] An embodiment example of the illumination is shown in FIG.
17. Side views of the beam path are shown at left and at the
top.
[0080] A light bundle which is adapted to a Powell lens and has a
Gaussian intensity profile is provided by fiber coupling. The light
bundle has two wavelengths, 477 nm and 488 nm. The Powell lens
(divergence 30.degree.) generates a line-shaped illumination with a
homogeneous intensity distribution along the line in the focal
plane of the following achromatic objective with a focal distance
of 10 mm. This plane is conjugate to the object plane of the
microscope. A first color splitter splits the two homogenized laser
lines. A light path, advantageously that with 488 nm, has a line
grating with 55 l/mm which generates the zeroth and .+-. first
diffraction orders in a determined intensity ratio of 8:1. These
laser lines are imaged by 2f-imaging with 150 mm focal distance
onto a phase element which is located in a plane conjugate to the
microscope pupil. The phase element is described in FIG. 8 a). In
the object plane, this phase element causes an intensity structure
according to FIG. 18 a). The grating is advantageously formed as a
phase grating. For precise adjustment of the intensity ratio
between the zeroth and first diffraction order, the grating has a
variable groove depth as is shown in FIG. 14 b). Since the
illumination of the grating is line-shaped, the efficiency of the
diffraction orders can be adjusted by lateral displacement of the
grating. The second light path only has the same 2f imaging as the
first, but without a grating. The phase procedure in the pupil of
the second light path is implemented by a phase mask according to
FIG. 8 c). This enables an axial structuring of the illumination
light in the object plane (FIG. 18 b)). The lateral extension of
the masks is about 12.times.12 mm.sup.2. The diffraction orders in
the first light path are at -4 mm, 0 mm and 4 mm. The two light
paths are recombined after these procedures. The illumination pupil
is adapted to the pupil of the microscope by means of relay optics.
The interface is the main color splitter. The imaging scale of the
relay optics is 4.35. The variability of the relay optics serves
for precision adjustment of the imaging scale. The beam combiner is
followed by a neutral splitter which serves to couple in the light
for fluorescence excitation (477 nm) and also the light for erasing
the structure impressed in the object (405 nm).
[0081] While the invention has been illustrated and described in
connection with currently preferred embodiments shown and described
in detail, it is not intended to be limited to the details shown
since various modifications and structural changes may be made
without departing in any way from the spirit of the present
invention. The embodiments were chosen and described in order to
best explain the principles of the invention and practical
application to thereby enable a person skilled in the art to best
utilize the invention and various embodiments with various
modifications as are suited to the particular use contemplated.
* * * * *