Journal of Microscopy, Vol. 225, Pt 1 January 2007, pp. 88–95
Received 18 April 2006; accepted 22 August 2006
Characterization of uniform ultrathin layer
for z-response measurements in three-dimensional section
fluorescence microscopy
G . V I C I D O M I N I ∗ †, M . S C H N E I D E R ‡, P. B I A N C H I N I ∗ ,
S . K RO L # , T. S Z E L L A S ¶ & A . D I A S P RO ∗ §
∗ LAMBS, MicroSCoBiO Research Center, Department of Physics, University of Genoa, 16146,
Genoa, Italy
†Department of Computer Science, University of Genoa, 16146, Genoa, Italy
‡Department of Biopharmaceutics and Pharmaceutical Technology, Saarland University, D-66041,
Saarbrücken, Germany
#Department of Physics, University of Genoa, 16146, Genoa, Italy
¶Leica Microsystems, D-68165, Mannheim, Germany
§IFOM The FIRC Institute for Molecular Oncology Foundation, 20139 Milan, Italy
Key words. 4Pi microscopy, confocal microscopy, optical sectioning,
shift-variant system, two-photon excitation microscopy, ultrathin uniform
fluorescent PE layer, z-response.
Summary
Layer-by-layer technique is used to adsorb a uniform ultrathin
layer of fluorescently labelled polyelectrolytes on a glass cover
slip. Due to their thickness, uniformity and fluorescence
properties, these ultrathin layers may serve as a simple and
applicable standard to directly measure the z-response of
different scanning optical microscopes. In this work we use
ultrathin layers to measure the z-response of confocal,
two-photon excitation and 4Pi laser scanning microscopes.
Moreover, due to their uniformity over a wide region, i.e. cover
slip surface, it is possible to quantify the z-response of the system
over a full field of view area. This property, coupled with a bright
fluorescence signal, enables the use of polyelectrolyte layers
for representation on sectioned imaging property charts:
a very powerful method to characterize image formation
properties and capabilities (z-response, off-axis aberration,
spherical aberration, etc.) of a three-dimensional scanning
system. The sectioned imaging property charts method needs a
through-focus dataset taken from such ultrathin layers. Using
a comparatively low illumination no significant bleaching
occurs during the excitation process, so it is possible to achieve
long-term monitoring of the z-response of the system. All
the above mentioned properties make such ultrathin layers a
Correspondence to: Alberto Diaspro. Tel: +39 010314218; fax: +39 0103536426;
e-mail: [email protected]
suitable candidate for calibration and a powerful tool for realtime evaluation of the optical sectioning capabilities of different
three-dimensional scanning systems especially when coupled
to sectioned imaging property charts.
Introduction
Although known for a very long time (Hooke, 1665), optical
microscopy continues to gain importance, exploiting its unique
capacity to enable high resolution in studies of biological
systems from cells to tissues (Hell, 2003). In general, optical
microscopy techniques offer a comparatively simple approach
to study biological systems, combined with a suitable spatial
resolution. As well, a continuous increase in radial and
axial resolution has been achieved in the last 25 years:
confocal microscope (Brakenhoff et al., 1979; Sheppard and
Wilson, 1980; Diaspro, 2001), multiphoton fluorescence
excitation (Denk et al., 1990; Diaspro et al., 2006) and
4Pi microscope (Hell and Stelzer, 1992b; Hell and Stelzer,
1992a; Bewersdorf et al., 2006). The microscope resolution
is known to be limited as a consequence of diffraction.
For three-dimensional (3D) optical sectioning the resolution
performance along the z-direction (axial resolution) plays an
important role. Unfortunately, the axial resolution is more
sensitive to aberrations than the lateral component (Born and
Wolf, 1993).
The 3D optical sectioning performance of the microscope is
strongly related to the axial resolution of the system, which
C 2007 The Authors
C 2007 The Royal Microscopical Society
Journal compilation Z - R E S P O N S E M E A S U R E M E N T S I N 3 D S E C T I O N F L U O R E S C E N C E M I C RO S C O P Y
can be quantified by the so called 3D point spread function,
i.e.P S Fξ,η,ζ (x, y, z), that describes how the detected image
of a point-like object at a given position (ξ, η, ζ ) is blurred
in space (x, y, z). Considering the optical microscope as a
linear shift-invariant system it can be fully described in terms
of a unique PSF, i.e.P S F (x, y, z). The PSF of an optical
microscope can be experimentally estimated as the response of
the microscope to subresolution fluorescent objects. However,
the axial resolution is better quantified by the z-response
of the microscope to a subresolution fluorescent layer
imaged by scanning along the optical axis (Hell and Stelzer,
1992a).
The relationship between PSF and z-response under the shiftinvariant assumption is given by the intensity distribution
along the z-axis:
P S F (x, y, z) d xd y.
(1)
I (z) =
x
y
The full width at half-maximum (FWHM) of the intensity
profile along the axis serves as a reasonable measure for the
axial resolution of the microscope.
Due to the finite dimension of the fluorescent layers, it is not
possible to directly measure the z-response of the system but
the convolution of the z-response with the layer, according to
the following relationship:
I z − z ld z d z ,
(2)
Id (z) =
z
where l d (z) is the function that describes uniform fluorescent
layer of a given thickness d.
As demonstrated by Brakenhoff et al. (1979) for a regular
confocal microscope with a typical axial PSF width of around
600–700 nm (under high numerical aperture conditions) by
using a layer thickness d of the order of 100 nm (thin layer),
convolution effects can be neglected.
On the other hand, for novel 3D imaging methods as 4Pi
microscope, where the axial resolution decreases to around
100 nm, convolution effects must be taken into account. In
order to neglect convolution effects for such systems, it is
necessary to reduce the layer thickness d to few nanometres
(ultrathin layer).
In this paper, we report results on fluorescently labelled
polyelectrolyte (PE) uniform ultrathin layers specially tailored
and realized to be used as subresolution objects to determine
the z-response of different optical sectioning systems. To
demonstrate the properties of the proposed layer we performed
measurements using two different confocal microscopes, a
two-photon excitation (TPE) microscope and a 4Pi microscope.
The validity of such an approach was demonstrated by
Schrader et al. (1998b). Unfortunately, their preparation
was laborious and the resulting samples exhibited a limited
uniformity over a small region. Zwier et al. (2004) solved the
problem of uniformity using polyvinylalcohol layers at the
89
expense of sample thickness, which are approximately 100 nm.
The very same sample was used by Brakenhoff et al. (2005) to
propose a standard method to characterize image properties in
sectioning microscopy, in particular the possible lateral-shiftvariant behaviour of the z-response. In fact, optical systems
are not completely shift-invariant systems over the whole
object domain. So far, these thin and ultrathin layers can
be used to study lateral shift variance. A simple but effective
representation of the image formation properties is given by
the sectioned imaging property charts (SIPcharts; Brakenhoff
et al., 2005).
This paper is organized in three sections. Section 1
shows materials and methods for PE layer preparation and
an overview of the instrumentation used in this work. A
characterization of the main properties of the proposed PE layer
(thickness, uniformity over large area, spectral properties,
linearity of fluorescence intensity and bleaching behaviour)
is developed in Section 2. In Section 3, the proposed PE layer
is used for different applications: z-response measurements for
4Pi and for TPE microscopes and SIPcharts analysis to compare
two different confocal microscope systems. Comments and
discussion of the proposed PE layer conclude the paper.
Materials and methods
Preparation of PE layer
Polyethyleneimine (PEI: MW 25 kDa; Aldrich, Italy), poly
(4-styrenesulfonate sodium) (PSS: MW 70 kDa; Aldrich),
poly(allylamine hydrochloride) (PAH: MW 15 kDa; Aldrich),
fluorescein labelled poly(allylamine hydrochloride) (PAHFITC: compound was made using poly(allylamine hydrochloride), polymer base MW 15 kDa; Aldrich) were used to
build the ultrathin fluorescent layers. PSS, PAH and PAH-FITC
solution were each prepared with a concentration of 5 mg/mL
in 0.15 M NaCl solution, whereas PEI solutions were prepared
in the same solution with a concentration of 10 mg/mL. All
PEs were used without further purification.
Multilayer films were prepared on 0.17-mm-thickness glass
cover slip (Forlab Carlo Erba, Italy) or 0.22-mm-thickness
circular quartz cover slip for 4Pi microscope. In both case the
cover slips were cleaned for 30 min in NOCHROMIX solution
(Godax Laboratories, Inc., Takoma Parc, MD, U.S.A.) and
successively washed twice in Milli-Q grade pure water (MilliQ-System, Italy) under sonication (Transonic 130, ADAC
Laboratories, Milpitas, CA, U.S.A.) for 15 min.
The method used to prepare ultrathin PE film is based on the
layer-by-layer technique (Decher, 1997).
A precoating of the glass surface is mandatory since Lowman
and Buratto (2002) found that films with an initial PEI layer
produce a much smoother surface.
Glass precoating was done by adsorbing a single layer of PEI
on the cover slip. The cover slip was incubated into PEI solution
for 15 min, then washed three times by dipping (for 1 min in
C 2007 The Authors
C 2007 The Royal Microscopical Society, Journal of Microscopy, 225, 88–95
Journal compilation 90
G. VICIDOMINI ET AL.
0.15 M NaCl solution). PE layers were applied in the very same
way. No drying step was performed between each layer. Two
types of films were utilized, namely: PEI/(PSS-PAH) n /PSS/PAHFITC/PSS (n = 0 and n = 1), in which n corresponds to the
number of PSS-PAH layers. As no differences were observed
between the two types we report results only for the first one
(n = 0). For confocal and TPE measurements the coated glass
cover slip was connected by adhesive forces to a standard
microscope slide (76 × 26 mm) using a very small layer
of immersion oil (1 μL, R.I. 1.514). Under application of
a slight pressure, it was possible to obtain a very uniform
ultrathin layer of oil. Oil-mounting media is necessary to reduce
reflection.
For 4Pi measurements, a drop of 87% glycerol (20 μL
buffered with Milli-Q grade pure water) was placed between
a coated and an uncoated quartz cover slip. Under a moderate
pressure the glycerol was distributed and the cover slips were
mounted on a 4Pi sample holder (Gugel et al., 2004).
wavelength with an average power of around 200 μW, a pixel
dwell time of 4.88 μs and the fluorescence signal was detected
in the 520–550 nm spectral ranges.
The 4Pi images were acquired by a Leica TCS 4Pi (Leica
Microsystems). The 4Pi-unit was tightly mounted to the
microscope turret to maintain all the capabilities of the
scanning confocal unit in the microscope body (Gugel et
al., 2004). For TPE, the beam of a mode-locked Ti:Sapphire
ultralaser Chameleon-Ultra (Coherent) was coupled with
the confocal microscope and directed towards the 4Pi
unit (Hanninen et al., 1995). The 4Pi measurements were
performed using a glycerol immersion objective (HCX PL APO
100×/1.35 Glycerol 0.22/0.22 Pair, Leica Microsystems), an
illumination of 780-nm wavelength with an average power
of around 100 μW and fluorescence signal was detected in
the 500–550 nm spectral range using an avalanche-photodiode. To increase the signal-to-noise ratio, the imaging was
performed with a pixel dwell time of 4.88 μs, a scanning line
average of 4 and a frame accumulation of 4.
Instrumentation
SIPcharts representation for confocal analysis
The results presented in this work were obtained by using
two different confocal laser scanning microscopes, a TPE
microscope and a 4Pi (type C) microscope.
As we are not interested in comparing confocal systems, we
will refer to different confocal microscopes as confocal1 and
confocal2.
Confocal1 mounts a Plan Apochromat 60x(NA = 1.4) oil
immersion objective resulting in a maximum scan field of
212 × 212 μM. Confocal2 mounts a Plan Apochromat 63 ×
(NA = 1.4) oil immersion objective resulting in a maximum
scan field of 245 × 245 μM.
Both systems are equipped with an Argon laser operating at
λ = 488 nm and are endowed with spectral abilities.
All confocal measurements analysed in this work were
obtained by using an average power of approximately 40 μW,
measured at the back focal plane of the objective. A z-step
size of 0.05 μm, a pixel dwell time of 4.08 μs and 4.88 μs
were used for confocal1 and confocal2 systems, respectively.
For both systems the diameter pinhole dimension closest to
Airy 1 was chosen: in the case of confocal1 this diameter
corresponds to 1.17 times the diameter of the back-projected
Airy disk evaluated at the emission wavelength, in the case of
confocal2 it is reduced exactly to 1 time.
Confocal1 is equipped with a detection band-pass filter of
535/50 nm and confocal2 by means of its spectral features
detects the signal within 510–560 and 560–590 nm spectral
windows at the same time.
TPE images were acquired by a Leica SP5 (Leica
Microsystems, Germany). A Chameleon-XR (Coherent, CA,
U.S.A.) laser source was coupled directly in the scanning head
of the system. TPE measurements were performed using a HCX
PL APO 63 × 1.4 oil objective, an illumination of 750-nm
Z-response is able to reveal aberrations of practical confocal
microscope, most notably spherical aberrations which are
manifested as a pronounced asymmetry in the response.
These aberrations combined to alignment errors produce a
nonuniform imaging field of the scanning system. This can
lead to a different z-response for each imaged point in the
field of view of the system, or more general to a different
distribution of the sectioned imaging properties over the field of
view.
SIPcharts are a very powerful tool to characterize such
behaviour. It was developed by Brakenhoff et al. (2005) and
it needs a trough-focus dataset.
Here we briefly describe the main imaging properties of
the sectioning process involved in SIPchart analysis, and for
further details we refer to (Brakenhoff et al., 2005):
(1)I total , the total intensity of the z-response (integral of Eq. (1)
along z);
(2)I max , the maximum fluorescence intensity found along the
z-response;
(3)z max , the axial position at which the value I max is found;
(4)FWHM, full width at the half-maximum of the z-response
and
(5)Skew, the axial asymmetry of the z-response. It is defined
as s = (a − b) (a + b) , where a and b are the asymmetry
factors of the z-response evaluated at the half-maximum
intensity level.
All SIPcharts analysis reported in this work were evaluated
on through-focus datasets of 64 × 64 points obtained from
binning of original 3D stack images of 512 × 512 pixels. Threedimensional stacks were acquired using the maximum scan
field of view of the investigated systems.
C 2007 The Authors
C 2007 The Royal Microscopical Society, Journal of Microscopy, 225, 88–95
Journal compilation Z - R E S P O N S E M E A S U R E M E N T S I N 3 D S E C T I O N F L U O R E S C E N C E M I C RO S C O P Y
91
Ultrathin fluorescent PE layer properties
Thickness
Layer preparation as well as the concentrations used are in
complete accordance with the protocol described by Etienne
et al. (2004). There is a linear growth of the PE multilayer films,
with an increment of approx. 40 for a single layer (Ruths et al.,
2000). This is suitable for our measurements since it makes it
possible to create samples whose thickness is below the optical
resolution. We occasionally checked for film thickness using
atomic force microscope in scratching mode (data not shown).
These layers can be assumed as infinitely thin compared to
the typical axial extent of the effective PSF of the confocal and
TPE microscopes, as well for 4Pi microscope.
Uniformity
In order to use the proposed layer for SIPcharts analysis
the most important requirement is the spatial uniformity. To
demonstrate such a feature we followed the imaging analysis
made by Zwier et al. (2004) but acquiring confocal images
in open pinhole configuration (confocal2) and not like Zwier
et al. (2004) wide-field images: two fluorescence intensity
images were acquired in two different positions of the layer
(Fig. 1G). The former image was used as ‘reference image’
(Fig. 1A) and the latter as the ‘object image’ (Fig. 1B). Then we
calculated the ratio between the ‘reference image’ and ‘object
image’ (Fig. 1C). A narrow distribution of the intensity of this
image with an average value of 1 indicates a uniformity of
the sample (Fig. 1F). In our measurement, we get an average
value of 0.996 and a FWHM of 0.058 (Fig. 1). Our values
compared with the values obtained by Zwier et al. (2004) for the
100 nm layer (avg. 0.999 and FWHM 0.038) seem to indicate
Fig. 2. Emission spectrum measured using confocal1 at different UV
exposure time. Arrows indicates 10-nm shift due to UV irradiation. (Inset)
Comparison of the emission spectrum of fluorescent PE layer obtained with
both confocal systems. The emission spectrum for confocal1 was taken by
acquiring intensity values at integration intervals of 5 nm for the range
of 495 to 655 nm. The emission spectrum for confocal2 has a step size of
10 nm for the range of 490–700 nm. Background (bgr) region was
obtained by scratching a region of the PE ultrathin fluorescent layer.
Spectra are normalized at the same maximum and at the same
background.
a lower uniformity. These differences can be explained by the
lower amount of fluorescence molecules for the ultrathin layer
respect to the 100-nm layer that can cause a worse signal-tonoise ratio.
Spectral properties
As the proposed layer can be used to characterize the
properties of optical sectioning in one or more different
fluorescence spectral ranges, it is important to describe its
spectral properties. The emission spectrum of the PAH-FITC
layer (inset Fig. 2) was directly acquired on the cover slip with
an excitation wavelength of 488 nm by means of confocal1
and confocal2. Both spectra show an emission peak for FITC at
around 530 nm, shifted to 10 nm with respect to the emission
Fig. 1. (A) ‘Reference’ image. (B) ‘Object’ image.
(C) Resulting ‘calibrate’ image. (D–F) Corresponding
histograms of pixel intensity values of these images.
(G) The configuration of reference and object image.
Images were obtained with confocal2 in open pinhole
configuration.
C 2007 The Authors
C 2007 The Royal Microscopical Society, Journal of Microscopy, 225, 88–95
Journal compilation 92
G. VICIDOMINI ET AL.
Fig. 3. Linear behaviour of the average fluorescence intensity over the
whole field of view as function of excitation power (measured at the back
focal plane of the objective). Measurements were performed with confocal2
in open pinhole configuration.
maximum for FITC described in literature (molecular probes).
This is in agreement with a series of photo-reactions in the
fluorophores, which are also responsible for changes in the
bleaching rate (Song et al., 1995). This assumption was also
supported by measurements of the spectrum at different UV
exposure times (0, 30, 60, 90, 120, 150 s). It is possible to
follow the shift of the spectra from 520 to 530 nm (Fig. 2). UV
exposure was performed using a microscope 100-W mercury
lamp and a 450–490-nm excitation filter block, the objective
lens was removed so that a column of light was impinged on
the layer. Power of the source is comparable with real wide-field
imaging condition.
Fluorescence linearity intensity
To be sure to use the suitable excitation power, in particular
to exclude that we measure in saturation, we detected the
average fluorescence intensity of the proposed layer at different
excitation power (Fig. 3). For the intensity data, a linear fit is
applied to clearly show the trend. From the Beer–Lambert law,
it would be expected that an increase in the excitation power
(using an argon laser in this case) produce an increased level
of fluorescence signal. This is supported by the results found
in the experiments.
Bleaching behaviour
A very important step concerns the evaluation of the
bleaching. We have demonstrated that by using a low
illumination power (≈40 μW) and a relative short pixel dwell
time (4.08 μs) no significant bleaching occurred during the
layer observation on confocal microscopy (confocal1). It is
worth noting that these are real imaging conditions. Moreover
it is possible to iterate the observation for a long time.
Due to this property ultrathin fluorescent PE layer can be
considered a good candidate for direct investigation of axial
resolution. In fact, standard procedures hitherto comprise the
measurement of the derivative of the response to an axial edge
produced by a thick fluorescent layer or the scanning of a
Fig. 4. (A) Intensity profile through the xz image (C) for each z-response
measurements (lateral averaged across 8 pixel = 3 μM). (B) Maximum
intensity of each measure shown in (A), red line represents a typical
exponential decay, blue and green bars represent the estimated maximum
fluctuations due to non accurate repeatability of the z-scanning system,
0.7% and 2.7%, respectively. (C) xz image of the ultrathin layer, scale
bar = 5 μM. (D) position of the maximum intensity for each z-response
measurements shown in (A). (E) FWHM for each z-response measurements
shown in (A). Average value 0.63287 ± 0.03038.
plane mirror along the optic axis (Schrader et al., 1998b). Both
methods allow only an indirect view of the axial resolution of
the investigate system.
Modern microscopes are often equipped with very fast zscanning systems that make very fast acquisitions possible
along the axial direction. Thus the utilization of an ultrathin
fluorescence PE layer makes it possible to achieve ‘online’
monitoring z-response of the microscope. Figure 4(A) shows
the z-response obtained from an axial acquisition iterated for
100 times. To reduce noise each profile was obtained by the
average of eight intensity profiles.
Figure 4(B) plots the I max (see SIPcharts parameters) for
each measure, it is important to note that photo-bleaching at
the end of the process is only 15% and FWHM measurements
fluctuation are due to noise (Fig 4(E); 632 ± 30 nm).
Figure 4(C) shows a xz image obtained using axial acquisition
capability of confocal1; in this case z-response of the system
is simply the intensity profile along z. Fluctuations observed
in Fig. 4(B) can be explained not only by noise but also by the
non accurate repeatability of the z-scanning system (see also
Fig. 4D): for each iteration the distance between the focus plane
of the objective in the z max position and the real position of the
layer can change. The longer this distance is, the lower the
intensity I max of the iteration is. For example, a distance of 25
and 50 nm produce an intensity decrease of 0.7% and 2.7%,
respectively (values obtained from numerical simulation of an
ideal confocal PSF under the imaging condition used in this
paper).
C 2007 The Authors
C 2007 The Royal Microscopical Society, Journal of Microscopy, 225, 88–95
Journal compilation Z - R E S P O N S E M E A S U R E M E N T S I N 3 D S E C T I O N F L U O R E S C E N C E M I C RO S C O P Y
Fig. 5. Figure describes z-response of 4Pi system with the typical
interference pattern. (A) 4Pi xz image of the whole imaging field of view.
Spatial resolution was 93.54 μM × 3.65 μM. (B–D) Intensity profile
through the xz image in three different positions, respectively, in positions
1, 2 and 3 (all the profiles are lateral averaged across 8 pixel = 365 nm). (E)
Physically zoomed image of the area inside the red box of (A) (scale bar =
1 μM). (F) Same image of (E) after fast linear one-step deconvolution (scale
bar = 1 μM). (G) Intensity profile of the uncorrected image (E, black
line) and corrected image (F, red line) along position 4 (profiles are lateral
averaged across 8 pixel = 97 nm).
93
an axial scan over the maximum scanning field of view for the
system (93.50 μM). The intensity profiles in different positions
within the scanning field of view (Fig. 5B–D) show that it is
possible, by using ultrathin PE layer, to monitor z-response in
all these positions. Figure 5(E) shows an axial scan for a smaller
region of the scanning field of view (6.20 μm). Figure 5(G)
is a typical intensity profile of a 4Pi z-response with a sharp
maximum and two pronounced lobes. The results found with
PE layer are in complete accordance with the results obtained
by Schrader et al. (1998b): the FWHM of the sharp maximum
(Fig. 5G) is 117 nm and after fast linear one-step deconvolution
is 96 nm (Schrader et al., 1998a). These values can be improved
by a rigorous alignment of the system.
Same z-response measurements were made for TPE
microscope. Figure 6 shows a single axial scan over the
maximum scanning field of view (245.76 μM). The intensity
profiles in different positions within the scanning field of view
(Fig. 6B–D) show that by using ultrathin PE layer it is possible to
monitor z-response in all these positions. The broad FWHM of
1.5 μM and the asymmetry of the z-response shape in all three
positions may indicate a nonperfect alignment of the system.
‘Online’ z-response measurements for 4Pi and TPE
microscope can be performed for a relative long time (data
not shown) thanks also to the reduced bleaching in TPE with
respect to one-photon excitation.
Applications
SIPcharts analysis
4Pi and TPE microscope z-response
Ultrathin fluorescent layers combined with z-scanning system
were used to monitor the z-response of a 4Pi microscope
operating under TPE regime. Unlike confocal analysis, the
ultrathin thickness of the layer became a very important
condition for direct measurement of the z-response of the
system. 4Pi microscope is known for a theoretical axial
resolution of around 100 nm. For this reason layers with a
thickness of 100 nm cannot be considered as subresolution
objects. Ultrathin fluorescent layers for direct measurements
of the z-response of the 4Pi microscope were used by Schrader
et al. (1998b). Unfortunately, samples were not uniform over
large regions. Thus monitoring the z-response of the system
over the whole field of view was difficult. Figure 5(A) shows
The suitability of ultrathin fluorescent layer also for SIPcharts
analysis is demonstrated by an example of SIPcharts
representations derived from trough-focus datasets.
Figures 7 and 8 represent, respectively, SIPcharts analysis for
confocal1 and confocal2 in the 510–560-nm spectral regions
(SIPcharts representation in 560–590-nm spectral region for
confocal2 is not showed). Both figures show that it is possible
to monitor the variation of sectioned image properties over the
whole image field.
Assuming that z-response is representative for axial
resolution, by using FWHM map we noted that the resolution
performance of both systems is not uniform over the whole
imaging field. I total and I max maps show that the fluorescent
intensity in confocal images is affected by nonoptimal
Fig. 6. (A) TPE xz image through the whole imaging
field of view. Spatial resolution was 245.76 × 10 μM.
(B–D) Intensity profile through the xz image in three
different positions, respectively, 1, 2 and 3 (all the
profiles are lateral averaged across 8 pixel = 3.84 μM).
C 2007 The Authors
C 2007 The Royal Microscopical Society, Journal of Microscopy, 225, 88–95
Journal compilation 94
G. VICIDOMINI ET AL.
Fig. 7. SIPCharts for confocal1.
sectioning. This especially is true applies if the images are
acquired in off-centre positions of the imaging field. These
variations coupled with the Skew and z max map variations can
be used as measure for the lateral shift variance of the system.
Moreover, SIPcharts representation using ultrathin
fluorescent layers can be used for imaging properties
comparison of different microscope systems. For example,
comparing FWHM maps, we can infer that confocal1 has
an alignment or setup that leads to a better axial resolution
(FWHM avg 683 ± 61 nm) than confocal2 (FWHM avg 739 ±
46 nm), but to a worse uniformity distribution of the
sectioning properties over the whole field of view (I total , I max
and FWHM standard deviations are higher for confocal1
than for confocal2). Thus we demonstrated that ultrathin
fluorescent PE layers can be a suitable alternative to thin
fluorescent layer for SIPcharts analysis, despite their plausible
lower signal-to-noise ratio due to the lower amount of
fluorescence molecules.
Due to technical problems in the 3D scanning process, no
SIPcharts representations can be shown for both TPE and for
4Pi microscopes. Only xz views are reported.
Conclusion
We have demonstrated that by means of ultrathin fluorescent
layers PE the direct monitoring of the variations of z-response
Fig. 8. SIPCharts for confocal2.
C 2007 The Authors
C 2007 The Royal Microscopical Society, Journal of Microscopy, 225, 88–95
Journal compilation Z - R E S P O N S E M E A S U R E M E N T S I N 3 D S E C T I O N F L U O R E S C E N C E M I C RO S C O P Y
over the imaging field of view of a confocal system is
possible. As well, our data further enforce the use of
SIPcharts representation as an optimum tool for microscope
characterization. Elsewhere, due to higher signal-to-noise
ratio, 100-nm-thin layers continue to be preferential to
monitor section properties of regular confocal microscopes
with resolution of 600–700 nm and more. We also have shown
that the same layer can be helpful for ‘online’ investigation of
axial resolution performances during alignment procedures.
Moreover they can be used for characterization of the zresponse of 4Pi microscopes and a TPE microscope.
Ultrathin dimension and large spatial uniformity are
fundamental properties of PE layers. Thus our ultrathin PE
layers combine the two qualities of the layers demanded by
Zwier et al. (2004) and by Schrader et al. (1998b): uniformity
and ultrathin dimension.
Ultrathin fluorescent layers can also be used for repeatability
tests of the mechanical z-scanning system by repeated scans
of the same field of view. The shift of z max during the iterated
measurements of the z-response can be seen as an index of the
performances of the z-motor mounted on a scanning system
(Fig. 4D)
Because the PE can be covalently labelled with different
dyes and the easy-to-use layer-by-layer technique is useful to
create multilayer films; the development of a unique sample
containing different ultrathin layers of different excitation
and emission spectral properties is under development.
This new sample might be used to monitor the chromatic
aberrations of the optical system: short distances between
fluorescent PE layers guarantee that no spherical aberration
effects induced by mismatch in refractive index can hamper
results.
In conclusion, ultrathin PE fluorescent layers could be an
important step in the development of effective calibration
and characterization techniques for optical sectioning
microscopy.
Acknowledgements
We are greatly indebted to Rolf Borlinghaus, Martin Hoppe and
Paolo Sapuppo of Leica Microsystems for 4Pi measurements.
Thanks are owed to Mario Bertero and Patrizia Boccacci
of Computer Science Department, University of Genoa for
discussions and suggestions about linear space-invariant
and space-variant system. This work was partially granted
by MIUR, IFOM (Milan) and Compagnia di San Paolo
(Turin).
References
Bewersdorf, J., Egner, A. & Hell, S.W. (2006) 4Pi Microscopy Handbook of
biological confocal microscopy (ed. by J. B. Pawley), pp. 561–570. Springer,
New York.
95
Born, M. & Wolf, E. (1993) Principles of Optics. Pergamon Press, Oxford.
Brakenhoff, G.J., Blom, P. & Barends, P. (1979) Confocal scanning light
microscopy with high aperture immersion lenses. J. Microsc. 117, 219–
232.
Brakenhoff, G.J., Wurpel, G.W.H., Jalink, K., Oomen, L. & Zwier, J.M. (2005)
Characterization of sectioning fluorescence microscopy with thin
uniform fluorescent layers: Sectioned imaging property or SIPcharts.
J. Microsc. 19(3), 122–132.
Decher, G. (1997) Fuzzy nanoassemblies: Toward layered polymeric
multicomposites. Science 277, 1232–1237.
Denk, W., Strickler, J.H. & Webb, W.W. (1990) 2-Photon Laser Scanning
Fluorescence Microscopy. Science 248(4951), 73–76.
Diaspro, A. (2001) Confocal and Two-Photon Microscopy: Foundations,
Applications, and Advances. Wiley-Liss, New York.
Diaspro, A., Chirico, G. & Collini, M. (2005) Two-photon fluorescence
excitation and related techniques in biological microscopy. Q. Rev.
Biophys. 38, 97–166.
Etienne, O., Picart, C., Taddei, C. et al. (2004) Multilayer polyelectrolyte
films functionalized by insertion of defensin: A new approach to
protection of implants from bacterial colonization. Antimicrob. Agents
Chemother. 48(10), 3662–3669.
Gugel, H., Bewersdorf, J., Jakobs, S., Engelhardt, J., Storz, R. & Hell, S.W.
(2004) Combining 4Pi excitation and detection delivers seven-fold
sharper sections in confocal imaging of live cells. Biophys. J. 87, 4146–
4152.
Hanninen, P.E., Hell, S.W., Salo, J., Soini, E. & Cremer, C. (1995)
2-photon excitation 4Pi confocal microscope - enhanced axial
resolution microscope for biological research. Appl. Phys. Lett. 66,
1698–1700.
Hell, S.W. (2003) Toward fluorescence microscopy. Nat. Biotechnol. 21(11),
1347–1355.
Hell, S.W. & Stelzer, E.H.K. (1992a) Fundamental improvement of axial
resolution with a 4Pi-confocal fluorescence microscope using twophoton excitation. Opt. Commun. 93, 277–282.
Hell, S.W. & Stelzer, E.H.K. (1992b) Properties of a 4Pi-confocal
fluorescence microscope. J. Opt. Soc. Am. A 9, 2159–2166.
Hooke, R. (1665) Micrographia. Royal Society, London.
Lowman, G.M. & Buratto, S.K. (2002) Nanoscale morphology of
polyelectrolyte self-assembled films probed by scanning force and nearfield scanning optical microscopy. Thin Solid Films 405, 135–140.
Ruths, J., Essler, F., Decher, G. & Riegler, H. (2000) Polyelectrolytes I:
Polyanion/polycatin multilayers at the air/monolayer/water interface
as elements for quantitative polymer absorption studies and preparation
of hetero-superlattices on solid surfaces. Langmuir 16, 8871–8878.
Schrader, M., Hell, S.W. & Van De Voort, H.T.M. (1998a) Threedimensional superresolution with a 4Pi-confocal microscope using
image restoration. J. Appl. Phys. 84(8), 4033–4042.
Schrader, M., Hofmann, U.G. & Hell, S.W. (1998b) Ultrathin fluorescent
layers for monitoring the axial resolution in confocal and two-photon
fluorescence microscopy. J. Microsc. 191(2), 135–140.
Sheppard, C.J.R. & Wilson, T. (1980) Image formation in confocal scanning
microscopes. Optik 55, 331–342.
Song, L., Ennink, E.J., Young, I.T. & Tanke, H.J. (1995) Photobleaching
kinetic of fluorescein in quantitative fluorescence imaging. Biophys. J.
68, 2588–2600.
Zwier, J.M., Van Rooij, G.J., Hofstraat, J.W. & Brakenhoff, G.J. (2004) Image
calibration in fluorescence microscopy. J. Microsc. 216(1), 15–24.
C 2007 The Authors
C 2007 The Royal Microscopical Society, Journal of Microscopy, 225, 88–95
Journal compilation