25 th

Anniversary
2
VOLUME 25
MAY
2023
69721

Official Partner of the EMS

© petunyia - stock.adobe.com

Light Sheet Microscopy

Virtual Microscopy

4D STEM

Atomic Force Microscopy

 13 - 15 September 2023
ICFO - The Institute of Photonic Sciences, Barcelona

Conference Topics SPM Methods

• Polymers and Composites • Nanomechanical and electrical
• Nanoelectronics, photonic and characterization
photovoltaic applications • Characterization of soft materials
• Nanomaterials and Life Science in liquid environment
• Special Session: • Advanced imaging
Using SPM to accelerate the clean energy • High resolution imaging
transition • Automation in AFM

event.nanoscientific.org/eu/2023

Keynote Speakers

• Prof. Dr. F. Pelayo García de Arquer • Prof. Dr. Lukas Eng

ICFO. Spain Technical University Dresden, Germany
• Prof. Ricardo Garcia • Dr. Moshe Ben Shalom
Materials Science Institute of Madrid Tel Aviv University, Israel
CSIC, Spain • Dr. Andrei Kholkin
• Dr. Esther Alarcon Llado University of Aveiro, Portugal
AMOLF physics of functional complex • Prof. Frederic Dubreuil
matter, The Netherlands École centrale de Lyon (ECL), France
• Dr. Deepak Venkateshvaran • Prof. Edoardo Albisetti
University of Cambridge, United Kingdom Polytechnic University of Milan, Italy Submit your
• Prof. Gustau Catalan • Dr. Christopher Kley Abstract at
Catalan Institute of Nanoscience and Helmholtz-Zentrum Berlin for Materials
Nanotechnology ICN2, Spain and Energy (HZB), Germany #NSFE2023
• Dr. Elena Gubanova • Dr. Thomas Miller and win up to
TUM School of Natural Sciences, Munich, University College London, €500
Germany United Kingdom AFM Scholarship!
Website I Registration I Abstract Submission: Deadline:
www.event.nanoscientific.org/eu/2023 June 1, 2023
Contact: info-eu@nanoscientific.org
SUPPORTED BY

REGISTER for #NSFE2023 TODAY


Combining Microscopy Techniques
with Machine Learning and AI
Five years ago, I was sitting in a conference
room in Cambridge, Massachusetts, watch-
ing Prof. Alán Aspuru-Guzik’s talk about
generative models for the inverse design of
molecules and materials and pondering the
possibilities of what the future of machine
learning could hold. I was studying new
photosensitizers for electronic/energy devic-
es, and the possibility of screening several
molecules at once before going to the lab
caught my attention. He discussed a gener-
al workflow for an automated, “self-driving”
approach to generate candidate molecules
for applications such as drug development
or renewable energy. In the heart of Biotech
startups, many questions emerged on how
this system could be used to discover poten-
tial drug candidates.
Years later, much progress has been made
in different research fields thanks to artifi-
cial intelligence (AI) and machine learning.
Recently, Aspuru-Guzik and Christine Allen,
both from the University of Toronto, pub-
lished a study in Nature Communications [1]
demonstrating how machine learning algo-
rithms can predict experimental drug release
from long-acting injectable systems. They
also described how these trained models could
guide the design of new long-acting inject-
ables, decreasing the time and cost associated
with drug formulation development.
When combined with microscopy tech-
niques, machine learning can be used to
improve sensitivity and precision or even to
solve multiple tasks at the same time. For
example, Piñeda et al. [2] designed a method
combining time-lapse microscopy and AI to
track cell motion at a spatiotemporal scale.
“The AI method uses the information in the
graph to adapt to different situations and
can solve multiple tasks in different experi-
ments. For example, our AI can reconstruct
the path that individual cells or molecules
take when moving to achieve a certain bio-
logical function. This means that research-
ers can test the effectiveness of different
medications and see how well they work as
potential cancer treatments,” said Piñeda.
Pharmaceutical companies are already using
the new framework, which can help scien-
tists develop more efficient cancer technol-
ogies and treatments.
Another interesting study reported by Sig-
mund et al. [3] describes a new genetic reporter
system that can recognize structures and their
functions within the cells. The spherically sym-
metric and concentric barcodes (EMcapsulins)
are easily identified under the electron micro-
Editorial
Editorial
© Mahyar Dahmardeh
Abb.: Through a combination of highly advanced microscope technology in iSCAT and multiple
machine learning techniques, the team from MPL and FAU continues to push the boundaries of
optical sensing.
scope and assigned like barcodes using AI. It [3] Sigmund, F., Berezin, O., Beliakova, S. et
can be used, for example, to identify electri- al.: Nat Biotechnol (2023) DOI: 10.1038/
cal synapses between nerve cells or receptors s41587-023-01713-y
that influence the interactions between can- [4] Dahmardeh, M., Mirzaalian Dastjerdi, H.,
cer and T-cells. Mazal, H. et al.: Nat Methods (2023) DOI:
From the articles in this issue, you can 10.1038/s41592-023-01778-2
gain interesting insights into the world of
microscopy, including a study developed by
Dahmardeh et al. [4] that employs interfer-
ometric scattering microscopy (iSCAT) and
AI to improve the spatiotemporal detection
of small proteins. By combining iSCAT with
iForest and FastDVDnet, proteins of only 10
kDa or less could be detected, pushing the
boundaries of optical sensing, and opening
doors to optical investigations of small traces
of biomolecules and disease markers. Fur-
thermore, you can find other articles related
to discoveries and achievements in micros-
copy, like real-time imaging of cell secre-
tions or a new objective for light microscopy
inspired by scallop eyes.
Inara Aguiar
Enjoy the read!
Inara Aguiar Inara is a science editor and writer with a
Ph.D. in Inorganic Chemistry. After a postdoc
in Computational Chemistry, she started
References working as a science editor in the fields of
Chemistry, Engineering, Bioengineering, and
[1] Bannigan, P., Bao, Z., Hickman, R.J. et Biochemistry. She has been working as a
al.: Nat Commun (2023) DOI: 10.1038/ technical writer / editor for several scientific
s41467-022-35343-w publishers and recently joined Wiley Analytical
[2] Pineda, J., Midtvedt, B., Bachimanchi, Science as a freelance content creator and
H. et al.: Nat Mach Intell (2023) DOI: microscopy news editor.
10.1038/s42256-022-00595-0
Imaging & Microscopy 2/2023 • 3
 Contents

EDITORIAL 3

NEWSTICKER 6

ANNOUNCEMENTS
SCANDEM 2023: 73rd Annual Meeting
of the Nordic Microscopy Society! 8
© Jürgen Mayer

Seeing is Believing – EMBO | EMBL Symposium 8

NEW BOOK
Light Sheet Fluorescence Microscopy 9

RMS IN FOCUS
mmc2023 (incorporating EMAG 2023): Book now! 10

NEWS FROM EMS

EMS Newsletter #81, May 2023 11
© TU Wien

COVER STORY
Raman Imaging of Polymer Materials 12
Comprehensive Insight into Plastics’ Chemistry and Structure
E. Vomhof et al.

LIGHT MICROSCOPY
© University of Antwerpi

Challenges in Building a Simple Conic-Based

Light Sheet Microscope 14
Combining Positive Axicon Lens with Further Aspherical Lenses to Obtain a Highly Localized Light Sheet
S. Saghafi et al.

Light Sheet Microscopy 17

Optical Sectioning with High Speed and Low Phototoxicity
J. Sanderson

Going Digital in Pathology 20

Introducing Virtual Microscopy and What Questions to Ask
K. Eser

Turnkey Multiphoton Microscopy for 3D Samples 23

Label-Free and 4D Imaging in Life-Science and Medicine
© Franceschi

S. Kiderlen and L. Krainer

PREVIEW:
ISSUE 3/2023

Coming out 21th August, 2023

Pushing the Boundaries of Optical Sensing
Combining State-of-the-Art Microscopy Technology in iSCAT and
Multiple Machine Learning Techniques
V. Sandoghdar

4 • Imaging & Microscopy 2/2023

 CORRELATIVE MICROSCOPY
Correlative Microscopy of a Catalytic Reaction 26
Zooming in on Chemical Patterns in Hydrogen Oxidation on Rhodium
P. Winkler and G. Rupprechter

ELECTRON AND ION MICROSCOPY

4D Scanning Transmission Electron Microscopy (STEM) with
Event-Based Electron Detection 30
How to Overcome the Bottleneck in Scanning Speed
D. Jannis et al.

Can AI Do your Ptychography? 32

A Machine Learning Approach to Real-Time Phase Imaging with 4D STEM
S. van Aert et al.

SCANNING PROBE MICROSCOPY

Imaging the Surface Ions of Pristine Muscovite Mica 36
Using Non-Contact Atomic Force Microscopy in Ultra-High Vacuum
G. Franceschi et al. COVER STORY
Raman Imaging of
Polymer Materials
IMAGE PROCESSING
Comprehensive Insight into
Jupyter Notebooks for Generating and Distributing Plastics’ Chemistry and Structure
Bioimage Analysis Workflows 38
My Favorite Image Analysis Tool, by Neubias Members Raman imaging is a powerful tool for the chemical
analysis of polymers and composite materials, as it
R. Camacho can non-destructively investigate phase separation
and defects, even beneath the sample surface.
Combining the technique with complementary
methods helps researchers achieve a comprehensive
ADVERTORIAL understanding of their plastic products.

Investigating Surface Catalytic Activities Using SECCM 41 Polymers are an important component of many
A. Klasen and A. Cerreta everyday products such as tires, plastic toys, clothing,
food packaging, glues and varnishes. Their widely
varying mechanical and chemical properties make
them popular in many industries and for diverse
applications. Thorough knowledge of the morphology
PRODUCTS42 and chemical composition of multi-component
polymeric materials on a sub-micrometer scale is
crucial for developing new materials, optimizing
INDEX / IMPRESSUM INSIDE BACK COVER their properties and assessing the quality of the final
products.

12
25 th
Anniversary
2
VOLUME 25
MAY
2023
69721

Official Partner of the EMS Welcome to the knowledge age. Wiley builds on its 200-year
heritage by partnering with universities, businesses, research
institutions, societies and individuals to develop digital content,
© petunyia - stock.adobe.com

learning, assessment and certification tools. Wiley continues to

share and deliver the answers to the world’s challenges
helping you to further your mission.
Light Sheet Microscopy

Virtual Microscopy

4D STEM

Atomic Force Microscopy

NEWSTICKER
In Situ Conductive AFM Real-Time Nanoplasmonic Imaging
Novel Correlative Microscopy Approach Unveils Catalyst Surface Features New System Allows Spatiotemporal Monitoring of Single-Cell Secretions

Researchers at the Helmholtz-Zen- Researchers from the BIOnanopho-

trum Berlin (HZB), and collaborators tonic Systems Laboratory (BIOS),
from the Fritz Haber Institute (FHI) EPFL, and the University of Ge-
of the Max Planck Society, have suc- neva have developed an optical
ceeded in analyzing electrocatalyt- imaging approach that gives a
ically active materials using in situ four-dimensional view of cell se-
conductive Atomic Force Microscopy cretions in both space and time. By
© M. Munz/HZB (CAFM). The new method enables si- placing individual cells into micro-
multaneous real-space imaging of local electrical, chemical-frictional, and scopic wells in a nanostructured
morphological properties evaluation of electrocatalysts in aqueous media gold-plated chip and inducing © BIOS EPFL
and under potential control. It can further help scientists evaluate pro- a phenomenon called plasmonic velopment as well as fundamental
cesses involved in battery electrodes, photocatalysis, or active biomaterials. resonance on the chip’s surface, studies.
they could map secretions as they
Original publication: are being produced. The research Original publication:
10.1021/jacs.2c12617 provides a detailed view of how doi: 10.1038/s41551-023-01017-1
More information: cells function and communicate More information:
https://bit.ly/IM-022023-a and can aid pharmaceutical de- https://bit.ly/IM-022023-b

Scattering-Type Scanning Near-Field Microscopy

© Mahyar Dahmardeh
Using Blue Light instead of red to measure electrons in semiconductors

Researchers at Brown University length gets shorter, this becomes

have developed a new micros-
copy technique that employs
blue light to measure electrons
in semiconductors. “There is a lot
of interest these days in studying
materials with nanoscale reso-
lution using optics,” said Prof.
Daniel Mittleman. “As the wave-
a lot harder to implement. As a
result, nobody had ever done it
with blue light until now.” The
research, recently published in
Light: Science & Applications, is
the first to provide nanoscale im-
aging to solve this longstanding
problem limiting the study of key
phenomena in various materials.
In the future, it can help scien-
tists to achieve more energy-ef-
ficient semiconductors and elec-
tronics.
Original publication:
doi: 10.1038/s41377-023-01137-y
More information:
https://bit.ly/IM-022023-e

Meta-Optics for Microscopes Interferometric Scattering Microscopy and AI

Observing Tiny Structures such as Nanoparticles or Transistors Improving Protein Detection Methods

opment are opaque to this light, Researchers at the Max-Planck-In-

© Lunghammer, TU Graz
Attosecond physics uses the
light-matter interaction phenom-
ena wherein attosecond (10−18 s)
photon pulses are applied to un-
ravel dynamical processes in mat-
ter with unprecedented time reso-
lution. Because it employs extreme
ultraviolet light, which oscillates
quickly, and all materials in the
construction kit of optics devel-
no usable imaging systems have
been reported until now. Professor
Marcus Ossiander: “I asked myself
whether the classical principle of
optics could not be reversed. Can
you use the absence of material in
small areas as the basis of an op-
tical element?” A team at Harvard
University has developed a new
lens, which was successfully tested
by researchers at TU Graz using
this design principle.
Original publication:
doi: 10.1126/science.adg6881
More information:
https://bit.ly/IM-022023-c
stitute for the Science of Light
(MPL) and Friedrich-Alexander-Uni-
versität (FAU) have combined one
of the most effective microscopy
methods with AI to detect small
proteins. The new approach uses in-
terferometric scattering microscopy
(iSCAT) and multiple machine learn- © Mahyar Dahmardeh
ing techniques to push boundaries Harald Köstler from FAU, employed
in optical sensing. The research, two machine-learning techniques—
published in Nature Methods, opens iForest and FastDVDnet—to detect
the door to optical investigations proteins at only 10 kDa or less.
of small traces of biomolecules
and disease markers. In order to Original publication:
enhance the iSCAT sensitivity, the doi: 10.1038/s41592-023-01778-2
MPL team led by Prof. Vahid San- More information:
doghdar, in collaboration with Prof. https://bit.ly/IM-022023-d

6 • Imaging & Microscopy 2/2023


Photoacoustic Imaging
Label-Free, Bond-Selective Imaging of
Biological Tissue
© Guyue Hu, PARS Mechanism 2023
Researchers from the University of Hong Kong
recently reported near-infrared photoacous-
tic remote sensing microscopy for noncontact
imaging of lipids. The new approach provides a
broader detection bandwidth and improves the
detection sensitivity and signal-to-noise ratio.
Photoacoustic imaging is a powerful technology
that uses light and sound to create biomedical
images. By scanning the sample and collecting
the corresponding photoacoustic signals, re-
searchers can reconstruct 2D or 3D images of
biological tissues. Photoacoustic remote sensing
imaging is a novel photoacoustic imaging mo-
dality that uses a different laser source to detect
acoustic signals instead of ultrasonic transduc-
ers used by conventional acoustics.
Original publication:
doi: 10.1117/1.APN.2.2.026011
More information:
https://bit.ly/IM-022023-f
Encoded Barcodes
Reporter Protein Readable by Standard Electron Microscopy Techniques
Researchers from the Technical University of
Munich (TUM) have developed a new genetic
reporter system that can recognize structures
and their functions within the cells. The gene
reporters are protein capsules with a fitting size
to be resolved by an electron microscope. The
capsules are produced by the cells themselves,
and their genetic blueprints are attached to
specific target genes. When the target genes be-
come active, the production of reporter proteins
starts. In this new approach, the researchers
Visit us at ELMI 2023 in Noordwijkerhout, NL
and MMC 2023 in Manchester, GB!
Cryo Transmission Electron
Tomography and Deep Learning
Observing of Platinum Catalyst Layers
© INE EPFL
EPFL researchers have revealed, for the first
time, the nanoscale structure of catalyst layers.
The new method combines cryogenic transmis-
sion electron tomography and deep learning to
enable observation of the platinum catalyst lay-
ers, highlighting how they could be optimized
for fuel cell efficiency. Proton-exchange mem-
brane fuel cell (PEMFC) is an electrochemical
energy conversion technology for automotive
applications that use nanocatalysts to trigger
electricity-producing reactions between hydro-
gen and oxygen. Because most PEMFC catalysts
contain platinum, a scarce and expensive metal,
there is a pressing global need to develop pow-
erful catalysts with reduced platinum content.
Original publication:
doi: 10.1038/s41929-023-00947-y
More information:
https://bit.ly/IM-022023-h
incorporated metal-binding proteins into dif-
ferently-sized capsules. The spherically symmet-
ric and concentric barcodes (EMcapsulins) are
easily identified under the electron microscope
and assigned like barcodes using artificial intel-
ligence.
Original publication:
doi: 10.1038/s41587-023-01713-y
More information:
https://bit.ly/IM-022023-j
Optical Filters
For Fluorescence Microscopy
Wide selection · Custom-specific designs
Newsticker
Objectives for Light Microscopy
Using a Mirror Instead of Lenses to Achieve
© Fabian Voigt et al.
2023 / UZH
Inspired by the anatomy of scallop eyes, neu-
roscientists at the University of Zurich have
developed an innovative objective for light mi-
croscopy. The new objective uses mirrors to pro-
duce images, enabling high-resolution imaging
of tissues and organs in a much wider variety of
immersion media than conventional microscope
lenses. Astronomical telescopes use mirrors in-
stead of lenses to create images, allowing them
to capture as much light as possible from plan-
ets, stars, and galaxies. The Schmidt telescope,
developed in the 1930s by Bernhard Schmidt
and still used in many observatories today, uses
a thin corrective lens combined with a large
spherical mirror. A single Schmidt objective can
be compatible with different clearing techniques
because it uses a mirror instead of lenses. The
spherical mirror focuses light at the same point,
whether the sample is immersed in liquid or air.
Original publication:
doi: 10.1038/s41587-023-01717-8
More information:
https://bit.ly/IM-022023-g
© Andreas Heddergott / TUM; Barth van Rossum
www.ahf.de
Imaging & Microscopy 2/2022 • 7
Announcement
SCANDEM 2023: 73rd Annual Meeting
of the Nordic Microscopy Society!
Uppsala, Sweden, June 12-15, 2023
After a four years break, this year, we will again
hold the SCANDEM meeting as a in-presence
conference (www.scandem2023.se). The meet-
ing hereby continues the tradition of inviting
and gathering international researchers in the
field of microscopy to share their latest findings
and socialize at a Nordic location- this time the
Ångström laboratory in Uppsala, Sweden.
SCANDEM is an important forum
for leading international re-
search groups in industry and
academia to meet and discuss
the latest trends and devel-
© oleg7799 - Adobe-Stock.com
Seeing Is Believing – EMBO | EMBL Symposium
Heidelberg, Germany, October 4 – 7, 2023
From its beginning in 2011, ‘Seeing is Believing’
has embraced novel imaging technologies that
open new windows for biological discovery, in-
cluding single-molecule and super-resolution,
light sheet and correlative light electron micros-
copy.
The molecular processes of life are naturally
dynamic in space and time from the atomic to
the organismal level. The new imaging methods
allow us to understand the mechanisms of life as
well as disease and is revolutionising our ability to
visualise the inner workings of macromolecular
8 • Imaging & Microscopy 2/2023
opments in microscopy focused on material nich, Xiaowei Zhuang - Harvard Med. School,
science and biology. This year the meeting will Michael Laue - Robert Koch Institute, Ute Kaiser -
feature invited talks by leading international Ulm University and Frances Ross - MIT. SCANDEM
and local researchers in microscopy and asso- 2023 is organized by Uppsala University and the
ciated areas, oral and poster presentations of Scandem Nordic Microscopy Society.
submitted papers and abstracts, as well as an
industrial exhibition, and workshops with vari-
ous focus themes.
Plenary speakers will be Xiaoqing Pan
- UC Irvine, Wolfgang Baumeister - TU Mu-
More information
and registration:
www.scandem2023.se
What past participants
say about the symposium:
“I was happy to attend this amazing meeting full
of talks showing incredible combinations of the
most recent state-of the-art techniques, some
of which even looked like being from far future
or sci-fi.” – Marketa Schmidt Cernohorska, Max
Perutz Labs Vienna, Austria.
Registration closes 23 August 2023.
complexes, organelles, cells, tissues, organs and
whole organisms. The symposium will bring to-
gether the leading developers of imaging meth-
ods with cutting-edge applications that illustrate
how imaging can answer biological questions. It
provides many opportunities for presentations,
discussions, and interactions between students,
postdocs, junior, and senior investigators.
Session topics include: dynamic super-reso-
lution imaging, probes & biosensors, organismal Registration:
and intravital imaging, cross-scale imaging and www.embl.org/events
image analysis & modelling.
 New Book

Light Sheet Fluorescence Microscopy

A Book by Emmanuel G. Reynaud and Pavel Tomancak

Destined to set the standard, the first book deal-

ing exclusively with this revolutionary and novel
imaging technology serves as an easy-to- un-
derstand introduction while offering numerous
tips and tricks. a2
0 % e:
Adopting a practical approach, the authors Get unt cod
o
disc AS23
who developed this actual technique provide W
a comprehensive, hands-on overview of the
basics of light sheet fluorescence microscopy,
instrumentation, applications, sample prepa- 1. Edition August 2023
ration, and data analysis. As a reflection of the 416 Pages, Softcover
uncompromisingly interdisciplinary nature of Practical Approach Book
the topic, they merge their expertise in physics, ISBN: 978-3-527-34135-1
biology, and computer science, giving valuable Wiley-VCH, Weinheim
insider tips taken from their work at major man-
ufacturers.
The result is in-depth information on hard- Order the book here:
ware and software solutions for a straightfor- https://bit.ly/Wiley-VCH-LSM
ward implementation of LSFM in the lab.

You Can Have It All In a Tabletop Package

A Full-featured SEM without the high cost!

At COXEM, we believe that electron microscopy doesn’t need to be expensive or complicated.
That’s why we developed the EM-30, a tabletop SEM that doesn’t skimp on features. Whether
you need 3Dimaging, low pressure, EDS, or even EBSD, the EM-30 delivers the performance
and features advanced users expect, at a price that entry-level users can afford. Call your
local agent to arrange a demonstration, or visit our website for more information.

www.coxem.com / overseas@coxem.com

Imaging & Microscopy 2/2023 • 9

RMS in Focus

mmc2023 (incorporating EMAG 2023):

Book now!
Manchester Central, UK, July 4-6, 2023

W
ith just two months to go before
mmc2023 (incorporating EMAG
2023) kicks off in Manchester,
UK, the excitement is well and truly build-
ing ahead of the RMS’s flagship event.

The Microscience Microscopy Congress (mmc)

is one of the biggest and best international
events in microscopy, imaging, and flow cytom-
etry - bringing together hundreds of people who
use microscopes for work, study and pleasure.
Alongside a huge, three-day conference, the
event boasts a world-class exhibition, workshops,
social networking opportunities and more.
mmc2023 is taking place at the superb Man-
chester Central Convention Complex, from 4 – 6
July, and this year’s Congress is shaping up to
be one of the best ever. A record number of ab-
stracts (more than 400) have already been sub-
mitted, and the event will also include a new and
improved RMS Learning Zone, plus a multi-cat-
egory, International Scientific Imaging Competi-
tion. Registration is still open, so now is the time
to book your place at this fantastic event for the
microscopy and imaging community.
Exhibition
Up to 80 companies – including many of the
biggest names in microscopy and imaging –
will be on hand to showcase their products and
give practical demonstrations. The exhibition
also provides important exposure for a number
of smaller companies keen to share their latest
technological developments.
The exhibition runs for three days alongside
the conference, and is completely FREE for any-
one to attend. Visitors can simply turn up and
register for an exhibition-only ticket, giving
them access to everything on show – including
live demonstrations, expert advice, and com-
pany workshops all under one roof.
Conference
Meanwhile the conference itself consists of
six parallel streams, with no fewer than 36
sessions covering every aspect of microscopy,
imaging and flow cytometry, including recent
and emerging applications. The blockbuster
programme will cover the full range of latest
techniques, applications and hottest emerging
topics – plus an incredible cast of speakers and
supporting poster sessions.
In addition to the academic content, the pro-
gramme includes a number of sessions sure to
be of interest to a wide range of scientific in-
dustries, including:
▪ Spatial and Imaging Cytometry
▪ Artificial Intelligence
▪ Public Health: The Impact of Microscopy
▪ Multimodal Microscopy
▪ Advances in label-free Imaging
▪ Clinical Diagnostic Imaging
Satellite Meetings and Bonus Features
mmc2023 (incorporating EMAG 2023) will bring
together a number of smaller meetings, allow-
ing you to meet with colleagues working in your
field as well as with cross-disciplinary peers, all
at the same event.
Other popular features include the RMS In-
ternational Scientific Imaging Competition, the
winners of which will be announced during
the conference. An eye-catching gallery of the
shortlisted images – covering all microscopi-
cal techniques - will be on display throughout
the event. The RMS will also be bringing its ev-
er-popular Learning Zone to mmc, with experts
on hand to share their knowledge with visitors
and provide demonstrations.
Register Now!
Registration is still open, and all the information
on rates, accommodation and transport can be
found on the official mmc-series website.
We look forward to seeing you there!
Contact
Allison Winton
Chief Executive
Royal Microscopical Society
London, UK
awinton@rms.org.uk

Book mmc2023: mmc exhibition: mmc-series website:

www.mmc-series.org.uk www.mmc-series.org.uk/exhibition.html www.mmc-series.org.uk

10 • Imaging & Microscopy 2/2023


José Maria Valpuesta,
EMS President
Virginie Serin,
EMS Secretary
Dear EMS members,
In March this year, the EMS Executive Board had
its annual winter meeting in Budapest, Hungary.
Some of the most important topics and deci-
sions were discussed during this meeting and
are mentioned in this newsletter.
The call for nominations for the EMS Out-
standing Paper Awards (OPA) 2022 has now
been closed. The EMS has received a number of
high-quality papers which are currently under
evaluation. The award-winning papers will be
announced to the EMS members in May this
year in our newsletter. The OPA winners will
receive their awards at the 20th International
Microscopy Congress to be held on 9 to 15 Sep-
tember, 2023, Bexco, Busan, Korea. We hope to
see most of you there.
Two other EMS calls have taken place
during the last two months:
▪ EMS Scholarships to support young research-
ers for the participation at the next IMC to be
held on Sept. 10-15, 2023 in Bexco, Busan, Ko-
rea; EMS aims to issue 20 travel grants for the
attendance of young European colleagues to
Busan (€ 800 each). The EMS provides valuable
support for our young colleagues to present
their results and learn from the specialists at
EMS web page:
https://bit.ly/EuroMicr
EMS Newsletter #81
international events. More information can be
found on our scholarships page.
▪ EMS Sponsored Events’ applications for EMS
sponsorships of microscopy-related events to
be held in the second half of 2023 (July 1 -
December 31).
This financial support from the EMS ensures
that also these smaller events, often organized
in the form of a school or course, can invite top
scientists to lecture our coming generations of
microscopists. You can find more information
on EMS Sponsored Events by visiting our web-
site.
EMS Yearbook
The 2022 EMS Yearbook will arrive in the fol-
lowing months. This edition will include re-
ports from EMS awards, EMC, EMS Sponsored
and Special events, EMS students reports from
EMC2020, and many other information. We
hope you will enjoy reading the reports on the
various activities of our society and members
during the past year, especially those written by
students who received a scholarship to attend
the EMS extension meetings: PICO and MCM16,
some of these contain truly inspiring notions
from our enthusiastic new generation of mi-
croscopists.
EMS sponsored events:
www.eurmicsoc.org/en/meeting-calendar/
ems-sponsored-events
News from EMS
May 2023
Finally, this year the EMS honors its 25 years
birthday, and different initiatives are being or-
ganized to celebrate it. A small symposium will
take place at IMC20 (Busan, Korea) to commem-
orate our first 25 years of existence.
As usual, the EMS continues with its support
to the activities of our members, so please have
a look at our webpage for the various meetings
and workshops that will take place over the
next year.
Please do not forget to regularly visit the
EMS webpage and the social media pages on
Facebook and Twitter which are daily updated
to efficiently provide you information from our
community.
If you have any questions, please do not hes-
itate to contact the EMS Secretary. Any sugges-
tions and ideas for future newsletters are also
welcome.
Contact
Prof. Dr. Virginie Serin
EMS Secretary
Toulouse, France
sec@eurmicsoc.org
Prof. Dr. José Maria Valpuesta
EMS President
Madrid, Spain
jmv@cnb.csic.es
EMS scholarships:
www.eurmicsoc.org/funding/en/
scholarships
Imaging & Microscopy 2/2023 • 11
Cover Story

Raman Imaging of Polymer Materials

Comprehensive Insight into Plastics’ Chemistry and Structure
Matthias Finger, Miriam Böhmler, Damon Strom, Eleni Vomhof

R
aman imaging is a powerful tool for
the chemical analysis of polymers and
composite materials, as it can non-de- a) b)
structively investigate phase separation and
defects, even beneath the sample surface.
Combining the technique with comple-
mentary methods helps researchers achieve
a comprehensive understanding of their
plastic products.

Polymers are an important component of

many everyday products such as tires, plas-
tic toys, clothing, food packaging, glues and
varnishes. Their widely varying mechan-
ical and chemical properties make them
popular in many industries and for diverse
applications. Thorough knowledge of the
morphology and chemical composition of c) d)
multi-component polymeric materials on a
sub-micrometer scale is crucial for devel-
oping new materials, optimizing their prop-
erties and assessing the quality of the final
products.

Correlative Raman
Imaging for Polymer Research

Raman analysis can help manufactur-

ers of polymers and composite materials
throughout the entire lifecycle of a product.
Based on inelastic light scattering, Raman
spectroscopy identifies molecules by their
unique Raman spectra and characterizes
their chemical properties. Polymers can be
identified, orientation and crystallinity can
be analyzed, and contaminations, defects
and stress states can be revealed [1]. In con-
focal Raman imaging, a Raman spectrum
is acquired at each pixel and the obtained
chemical information is color-coded in the
resulting Raman image. 2D scans can either
visualize the spatial distribution of polymer-
ic components on a sample surface or show
their layered structure in depth scans [1-4].
Additionally, 3D representations can be gen-
erated from stacks of 2D images [2].
For achieving a thorough understanding
of polymer samples, correlative microscopy
techniques offer many benefits. Atomic force
microscopy (AFM) acquires topographic
information and determines local mechan-
ical characteristics of a sample with high
spatial resolution. AFM and Raman imag-
ing thus yield complementary information
about plastic samples [4].
Fig. 1: Correlative Raman-AFM study of a polymer blend. A: Raman spectra of PMMA (red)
and PS (blue). B: Raman image showing the phase separation in a blend of PMMA (red) and PS
(blue). C: AFM topography image of the same sample area. D: Overlay of AFM topography and
Raman images (B and C), correlating structural and chemical information.
Raman-AFM Study of a Polymer Blend depressions in the raised matrix, but also
small, round islands protruding from the
A polymer blend was investigated by correl- depressions (Fig. 1C). The maximum height
ative Raman and AFM imaging. The mea- difference within the examined area was
surements were performed with a WITec 380 nm.
alpha300 RA, which combines AFM and Combining chemical and topographic
Raman imaging capabilities in one instru- information enables deeper insights into the
ment. The sample remained in place during sample composition. The overlay of AFM
all measurements, as switching between and Raman images reveals that the depres-
imaging modes required just a turn of the sions consist of PS, while all elevated areas
objective turret. are formed by PMMA (Fig. 1D).
Two components were identified by their
Raman spectra (Fig. 1A): poly(methyl meth-
acrylate) (PMMA) and polystyrene (PS). The Raman Analysis of Layered
Raman image shows their separation on the Plastic Packaging
sample surface (Fig. 1B). Both compounds
formed rounded islands of different sizes Plastic is a common packaging material in
within the other phase, respectively. the food industry. Packaging systems often
For structural insight, the topography of consist of multiple layers of different poly-
the same sample area was recorded by AFM. mers, depending on the type of food or drink
The AFM image shows irregularly shaped contained and its associated requirements. In

12 • Imaging & Microscopy 2/2023

 Cover Story

order to ensure safe and tightly sealed food

packing, the material composition needs to a) b)
be monitored and defects and inclusions
must be located and characterized, which is
possible with Raman imaging. The following
measurements were performed with a WITec
alpha300 series Raman microscope.
First, a cross section through the top foil
of a cheese packaging was investigated. The
foil lid is designed to tightly seal the con-
tainer while fused to it. Additionally, it will
stick to the lower part of the container after
its initial opening, so it can be easily closed
again to keep the contents fresh. The Raman
image reveals the layer structure of the foil
(Fig. 2A) and the Raman spectra of the indi-
vidual components (Fig. 2B) were identified
automatically using the TrueMatch spec- Fig. 3: Raman depth scan of a plastic foil with a defect. A: Bright-field image of a bubble in the
tral database management software. The foil. B: Raman depth scan along the blue line in A. The white line marks the bubble’s border.
foil contains layers of the polymers poly- Polyester (red) and two different PE modifications (green, pink) are separated by a layer of glue
propylene (PP) and polyethylene (PE). They (blue), which is missing in the bubble. Sample courtesy of Dr. Dieter Franke, Siegwerk Druckfar-
are separated by a thin film of polyurethane ben AG & Co. KGaA, Siegburg, Germany
(PU), on top of which several different pig-
ments were detected. The pigments formed a
printing on the foil and included, for exam- Non-Destructive Defect ed in a plastic product without cutting it,
ple, the titanium dioxide polymorphs rutile Analysis in Plastic Foils which is crucial for ensuring the product
and anatase for white color. A second thin quality. The bright-field image of a second
film between two PE layers consists of eth- With a confocal Raman microscope, it is plastic foil showed bubble inclusions (Fig.
ylene-vinyl acetate (EVA), which is respon- also possible to non-destructively inves- 3A) and a Raman depth scan characterized
sible for the stickiness of the cover foil after tigate the layer composition of transpar- this defect below the surface (Fig. 3B). The
initial opening that enables the reclosing of ent polymer samples below their surfaces. foil consists of a polyester and PE, separat-
the packaging. Inclusions and defects can thus be detect- ed by a glue layer. In the observed bubbles,
the glue layer is missing. Similar depth
scans at different positions also revealed
that the thickness of the glue layer varied
a) b) over the sample area.

Conclusion

The presented measurements demon-

strate the power of Raman microscopy for
characterizing plastic samples. AFM and
Raman imaging yielded complementa-
ry topographic and chemical information
for detailed insight into the structure and
composition of a two-component polymer
blend. Raman imaging characterized the
layer structure of a food packaging and a
depth scan revealed a sub-surface defect in
a plastic foil.

Contact
Fig. 2: Raman image of a layered plastic WITec GmbH
packaging foil. A: Raman image of a cross Ulm, Germany
section through the cover foil of a food info@witec.de
container, color coded according to the https://raman.oxinst.com
spectra in B. B: Raman spectra of the
identified components: the polymers PP
(red), PE (blue), EVA (cyan) and PU
(green); rutile, anatase (both pink) and
yellow, blue, purple and red pigments References:
(yellow, violet, brown, orange). https://bit.ly/IM-WITec0223

Imaging & Microscopy 2/2023 • 13

Light Microscopy

Challenges in Building a Simple

Conic-Based Light Sheet Microscope
Combining Positive Axicon Lens with Further Aspherical Lenses to Obtain a Highly Localized Light Sheet
Saiedeh Saghafi1, Meraaj Foroughipour1, Klaus Becker1,2, Inna Sabdyusheva-Litschauer1, Massih Foroughipour1, Eugenijus Kaniusas3, Hans Ulrich Dodt1,2

T
he laser beam profile is a most im-
portant factor in light interaction
with materials. Accordingly, con-
trolling devices reshaping an incident la-
ser beam via tailoring its amplitude and
phase have been developed. Meso-Aspher-
ic-Optic (MAO) characterizes the optical
properties of conical wavefields and of-
fers ways for building conical lenses with
high precision. MAO elements with com-
plex structures (e.g., Powell lenses and
axicons) reshape a Gaussian laser beam
into the specific distribution required by
the application. Here, we describe how
a well-structured positive axicon lens in
combination with further aspherical lens-
es enables us to obtain a highly localized
light sheet for static light-sheet micros-
copy. This affordable light sheet generator
system is perfect for revealing details in
small chemically cleared biological speci-
mens, as well as for looking deeply into a
limited region of larger transparent sam-
ples.

Introduction

Lasers have become a pivotal tool for most
applications in various fields (e.g., biotech-
nology, photonics, communications, de-
fense, and medicine) [1-3]. The laser beam
profile is one of the most critical factors in
laser beam interaction with different ma-
terials making beam shaping an important
prospect in the laser industry. Consequently,
beam-controlling devices for modifying the
laser output to get the desired irradiance dis-
tribution via tailoring the amplitude/phase
are in high demand. Using diffractive or re-
fractive optics, various elements for reshap-
ing laser beams have been developed. Dif-
fractive optical elements (DOEs) can convert
the fundamental laser beam into a beam with
specific characteristics [4,5]. These elements
are generally light in weight, and small in
size. However, they are highly expensive and
a lack of freedom and flexibility in the usage
of DOEs in any other applications except the
one that was initially intended has been re-
ported [6]. On the other hand, refractive op-
Fig.1: MAO conical elements; A) Symmetrical conic element (axicon), B) The effects of axicon
on the object pattern, C) Axial conic element (Powell lens), D) Object, E) Image of the object (D)
using Powell lens while the conical axis is vertical to the x-axis, F) Image of the object (D) using
Powell lens when the conical axis is parallel to the x-axis
tical elements (ROEs) are cost-effective while and symmetrical conic structures (e.g., Axi-
providing high-precision results and a high con lenses, meaning axis-image and Powell
degree of freedom for diverse applications lenses as shown in Fig.1) has been suggested
[7-11]. However, we might encounter optical [15-21].
problems either due to fabrication errors or Axicons have the property of a point
natural optical phenomena [12] while they source but creates a line of focused points
need regular calibration [13-14]. A funda- instead of one focal point, thus, there is
mental laser beam accommodates the max- no definite focal length. The beam gradu-
imum possible energy in a distinct region ally becomes a ring through propagation
of space forming a bell-shaped distribution [16]. This transformation is homomorphic.
approximated by a Gaussian function within An axicon is not an imaging lens but can
the scalar wave equation. Meanwhile, state- be considered a beam-shaping element. In
of-the-art beam-shaping methods enable us- an ideal axicon, the length of the line of
ers to overcome many of the limitations that focused points depends on the wedge angle,
we encountered by remaining in the Gauss- the diameter of the incoming beam, and
ian regime. To minimize these limitations the refractive index of the material. How-
using MAO optical elements of specific axial ever, fabrication errors in conic-structured

14 • Imaging & Microscopy 2/2023


lenses can cause massive alterations in these
parameters.
Material and Methods
Effects of Well-Structured Axicon
on the Incident Beam
Conical lenses of special structures can
convert a Gaussian beam into various dif-
fraction-limited beams [22-25], opening
a new horizon for their application in 3D
microscopy by generating a light sheet with
controlled expansion [21,26]. When a radi-
ally symmetric Gaussian field that is con-
sidered a bundle of parallel rays along the
propagation axis strikes an axicon with the
refractive index n (Fig.1A), the rays will re-
fract individually upon arrival at the conical
surface forming highly focused points with
a gradual change towards becoming a ring
(Fig.2).
Effects of Non-Ideal Axicon on
the Incident Gaussian Beam
As pointed out before, the precision in the
fabrication of the conic elements is a chal-
lenging task and can affect the final results
massively. An axicon with high precision is
usually expensive compared with the com-
mercially available ones. Depending on the
materials, methods, and precision utilized in
their manufacturing, the prices could differ.
In our laboratory, we used a wide variety
of axicons. Among them, a few suffered
from manufacturing errors that were not
detectable by macroscopic inspection. Nev-
ertheless, a consistent major alteration in
the beam shape and width became apparent
when replacing the reference well-struc-
tured axicon with another one of identical
characteristics. The beam profiles changed
Fig.2: Beam propagation through Axicon; A) Theoretical analysis, B) Real Axicon (Asphericon/
Germany), C) At short distance after the tip of the Axicon, D) at the middle of the line of fo-
cused points, E) Ring formation after the focus.
Light Microscopy
THE FUTURE
material. However, fabrication errors in conic-structured lenses can cause massive alte
these parameters.
Material and Methods
and beam widths varied by a factor of 1.5
- 5. Figure 3 compares the effects
Effects of two
of Well-Structured DEPENDS ON
Axicon on the Incident Beam
OPTICS
Conical(fan
commercially available axicons lenses of special structures can convert a Gaussian beam into various diffracti
angle:
170°), one with manufacturingbeamserrors
[22-25],(left:
opening a new horizon for their application in 3D microscopy by generat
sheet with controlled expansion [21,26]. When a radially symmetric Gaussian fie
Fig.3 A1 - D1) and one with a high-preci-
considered a bundle of parallel rays along the propagation axis strikes an axicon with the
sion structure (right: Fig.3 index
A2 - D2), on anthe rays will refract individually upon arrival at the conical surfac
n (Fig.1-A),
incident beam with Gaussian highlydistribution.
focused points with a gradual change towards becoming a ring (Fig.2).
The effects are shown at the focus (mini-
mum width) and 5 mm away from the focus.
Effects of Non-Ideal Axicon on the Incident Gaussian Beam
Static Light Sheet Microscopy As pointed
Setup outUsingbefore, the precision in the fabrication of the conic elements is a challe
and can affect the final results massively. An axicon with high precision is usually
Well-Structured Conic Lenses compared with the commercially available ones. Depending on the materials, met
We built the setup precision according utilizedtoin their manufacturing, the prices could differ. In our laboratory, we u
DE102010046133B4-patent variety using of200 milli-Among them, a few suffered from manufacturing errors that were not
axicons.
watts Sapphire laser (488 nm, by macroscopic
1.5 mm width, inspection. Nevertheless, a consistent major alteration in the beam shape
Coherent Inc./Germany) withbecameGaussianapparent when replacing the reference well-structured axicon with another one o
distri-
characteristics. The beam profiles change and beam widths varied by a factor of 1.5-5
bution for fluorescence excitation. The beam
demonstrates the comparison between the effects of two commercially available ax
is guided toward a beam-shapingangle: 170°),unitonede-with manufacturing error (left- Fig. 3-A1 to -D1) and one with a high
scribed by Saghafi. et. al. [26] for gradual
structure al- 3-A2 to -D2), on an incident beam with Gaussian distribution. The
(right- Fig.
teration from Gaussian to ashown at the focus (minimum width) and 5mm away from the focus.
flattened-Gauss-
ian beam. The semi-uniform beam is guided
toward a 170 º axicon (Asphericon/Germany)
Static Light Sheet Microscopy Setup Using Well-Structured Conic Lenses
placed at a 50 mm distance (element-1). A
We built the setup according to DE102010046133B4-patent using 200 milliwatts Sapp
precision-aspherized-achromatic(488nm,lens of focal
1.5mm width, Coherent Inc./Germany) with Gaussian distribution for fluores
length F1 (element-2) is placed at a distance
excitation. The beam is guided toward a beam-shaping unit described by Saghafi. et.
of 2F1 from the flat surface offorthe
gradual
axicon pro- from Gaussian toNaEflattened-Gaussian
alteration W beam. The semi-uniform
ducing a magnified axially guided towardbeam
symmetric a 170º axicon (Asphericon/Germany) placed at a 50mm distance (eleme
from the flat surface of the axicon120i Infinity
a magnified Corrected
with optimized parameters.precision-aspherized-achromatic
By guiding this lens of focal length F1 (element-2) is placed at a distan
producing axially symmetric beam with
beam toward an achromatic-cylinder lens of By guiding this beam toward an achromatic-cylinder lens of foca
optimized parameters.
focal length F2 (element-3) placed
length 𝐹𝐹at
! (element-3)
Objectives
F1+F2, a placed at F1+F2, a controlled ellipticity is introduced to the beam
Finally,toa the
controlled ellipticity is introduced positive
beamachromatic-cylinder lens of focal F3 (element-4) was placed at a dist
structure. Finally, a positive√2𝐹𝐹 ! from F2 and enables us to form
achromatic-cyl- Designed to reduce
a thin sheet of light the overall
as can be seenweight
in Figure 4. Th
detection
inder lens of focal F3 (element-4) unitplaced
includes a customized microscope equipped with corrected objectives for
was and size of an imaging system while
at a distance of from Fcompensating
2 and enables us
refractive index mismatches, a tube lens, a computer-controlled band-pas
maintaining optical performance:
wheel for blocking the fluorescence excitation light, and a scientific-grade CCD camera
to form a thin sheet of light sCMOS,
as can be seen in 6.5 µm pixel, UK).
5 Megapixel,
Figure 4. The detection unit includes a cus- • Reduces system length by up to
tomized microscope equipped with corrected 42% compared to conventional
objectives for compensating refractive index microscope systems
mismatches, a tube lens, a computer-con-
trolled band-pass filter wheel for blocking • Provides easy integration into
the fluorescence excitation light, and a sci- many machine vision systems
• Designed for use with next
generation imaging sensors
Find out more:
www.
edmundoptics.eu/
imaging
Visit us at
June 27-30, 2023 | Booth B1.415
Contact us:
EU/UK: +44 (0) 1904 788600
GERMANY: +49 (0) 6131 5700 0
FRANCE: +33 (0) 820 207 555
sales@edmundoptics.eu
Light Microscopy

Fig.4: Left-side; Schematic drawing (top-view and side-view) of the

axicon-based design for the static light sheet microscope including 1)
Axicon, 2) a precision-aspherized-achromatic lens of focal length F1,
3) achromatic-cylinder lens of focal length F2, 4) positive achromat-
ic-cylinder lens of focal F3, 5) Container filled with liquid and the
sample. Right-side; The actual system.

Fig.3: Effects of two 170° axicons on a 5mm laser beam with Gauss-
ian distribution; A1-B1) 3D laser profile at XY-plane and transversal
2D-laser beam profile normal to the propagation axis at 5mm away
from the tip of an axicon with fabrication error, respectfully, C1-D1)
3D transversal laser profile and transversal 2D-laser beam profile
normal to the propagation axis at the center of the line of focused
points at 12mm away from the tip of an axicon with fabrication error,
respectfully, A2, B2, C2, and D2 have the same expressions as given
in A1, B1, C1, and D1, just for a high-precision 170° axicon, respect-
fully.
Fig.5: The 3D-reconstructed images of a forelimb of Thy1 GFP mouse
obtained by the Axicon-based light-sheet microscopy demonstrating
fine details of the muscles, cartilage, tendons, bones, and GFP-positive
neuronal structures neuronal structure (outer and inner). For detection,
a 2X objective (NA 0.14, Olympus/Japan) equipped with a modulator
for compensating the refractive index mismatch was used.

entific-grade CCD camera (Neo 5.5 sCMOS,
5 Megapixel, 6.5 µm pixel, UK).
Experimental Results
The setup was used for imaging chemically
cleared tissues of the forelimb of a 2-week-
old Thy1 GFP mouse. We used the 3DISCO
clearing method to make the sample trans-
parent. The dehydration was done by as-
cending tetrahydrofuran (THF) series and
dibenzyl-ether (DBE) as the clearing solu-
tion. Dehydration lasted for one week. Using
Amira (Thermo Fisher Scientific, USA), 3D
images were constructed from 1600 optical
sections that were obtained by an optically
corrected 2X objective (NA 0.14, Olympus/
Japan) revealing the fine details of tissue
structure as shown in Figure 5. These imag- Affiliation
1Bioelectronics Section, FKE, TU Wien,
es were obtained from single stacks of im-
ages, no stitching was done and no axially Vienna, Austria
2Brain Research Institute-CBR, Medical Uni-
swept light sheet technique was used [27,28].
It demonstrates that this affordable system versity of Vienna, TU Wien, Vienna, Austria
3Ins. of Biomedical Electronics, TU Wien, Vi-
provedies excellent spatial resolution due to
the long width of the thin part of the light enna, Austria
sheet.
Contact
Dr. Saiedeh Saghafi
Conclusion Bioelectronics Section,
FKE, TU Wien
It can be concluded that Meso-aspheric opti- Vienna, Austria
cal lenses can be highly useful elements for saiedeh.saghafi@tuwien.ac.at
designing laser beam shaping devices that are
useful for many applications, such as non-in-
vasive 3D imaging techniques. However, the References:
results depend highly on the presence of man- https://bit.ly/IM-Saghafi
ufacturing errors limiting precision.

16 • Imaging & Microscopy 2/2023

 Light Microscopy

Light Sheet Microscopy

Optical Sectioning with High Speed and Low Phototoxicity
Jeremy Sanderson1
© Jürgen Mayer

F
luorescence labeling is an integral
technique for biomedical and life
science research, yet the light flux
required to visualize and capture images
of cellular structures is very large. Living
organisms and cell cultures cannot with-
stand high flux densities. The light sheet
microscope enables researchers to image
tissues very fast in three dimensions with
minimal photobleaching. For this reason,
it has been called “the smart and gen-
tle microscope” [1]. The light sheet
was developed in 2004 by Ernst
Stelzer’s group at EMBL to study
developing cells, small organisms,
and large volumes of tissue.

A 3-D rendering of a cleared adult mouse brain. Tyrosine hydroxylase-positive dopaminergic neurons labeled with Alexa 488, pseudo-colored in
cyan. The parvalbumin-positive interneurons, labeled with Rhodamine red-X are shown in yellow, and the choline acetyltransferase in cholinergic
neurons labeled with Alexa 647 is seen here in magenta. The image was taken using a Bruker Luxendo LCS SPIM table-top system for large
cleared samples. Scale is 1.0 mm. The file size was 828 Gb. Rendering was performed with Imaris Viewer 10.0.0 and Fiji 2.9.0. With thanks to
Dr. Jürgen Mayer, Applications Specialist at Bruker for his help in acquiring and presenting this image.

Imaging & Microscopy 2/2023 • 17

Light Microscopy
One of the reasons put forward as an advantage for lightsheet microscopy was the ability
to image whole embryos and organs, rather than preparing thin samples or growing cells
mounted on a coverslip.
Optical Sectioning
The confocal microscope is the workhorse of
fluorescence imaging, optically sectioning
the sample to create blur-free images. It is
perfectly good for many applications but is
not ideal for imaging delicate living samples.
In a conventional fluorescence widefield or
confocal microscope, the objective acts as
its own condenser with both illuminating
light and emitted signal traveling on-axis
through the objective. The light sheet micro-
scope also creates optical sections similar to
the confocal microscope but with the illumi-
nation objective forming a sheet of light set
orthogonally side-on to a separate imaging
objective (Fig. 1). By uncoupling the illumi-
nation and detection axes, the light sheet
microscope only illuminates a single layer
of the sample at a time. This makes it much
faster and less phototoxic than the classical
laser confocal microscope.
In its simplest configuration, the light
sheet is generated as a Gaussian beam with
cylindrical optics. Extra lenses, slits, and
mirrors are used to shape and steer the beam.
18 • Imaging & Microscopy 2/2023
The distance where the thickness of the light
sheet remains reasonably constant is called
the ‘confocal parameter’, which determines
the axial resolution and optical sectioning
ability. Significantly thinner light sheets can
be generated by using interference methods
to create a lattice sheet from a non-diffrac-
tive Bessel or Airy beam instead of a Gauss-
ian beam. This focused beam is scanned
across the field of view and a camera with
a global or rolling shutter is used to acquire
the image only from the vicinity of the illu-
mination beam at any instant. This approach
is called digitally scanned laser light sheet
microscopy.
Since the illumination usually penetrates
the sample from one side, obstacles lying
in the path of the light sheet can disturb
its quality by scattering or absorption of
the light. This causes unevenly-illuminated
datasets and striping artifacts [2]. The sim-
plest solution is to rotate the sample, which
is then sequentially illuminated from alter-
nate sides and the image fused into a sin-
gle dataset. Another approach is to add a
third objective lens so that the sample can
Fig. 1: Architecture of the
light sheet microscope.
be illuminated from two sides. Recently, var-
ious light sheet microscope configurations
have been developed that either use the same
objective for illumination and detection or
illuminate the sample and detect the image
from the same side, so-called ‘open top’ con-
figurations, which allow much larger speci-
mens to be investigated.
Sample Imaging
One of the reasons put forward as an ad-
vantage for lightsheet microscopy was the
ability to image whole embryos and or-
gans, rather than preparing thin samples
or growing cells mounted on a coverslip.
A range of mounting methods is possible
[3]. The classic method is to hold the sam-
ple within a chamber in a molded agarose
block or extrude a low melting-point aga-
rose cylinder into the sample chamber from
Teflon FEP tubing with a refractive index
similar to water.
Biological tissues are opaque and scat-
ter light; they require some form of chem-
 Light Microscopy

fore developed. Weiss et al. [4] is a useful

Manufacturer System(s) QR-Code for URL starting point for working with clearing
agents.
As with any other emergent technology, the
light sheet microscope was first home-built
3i Lattice LightSheet
in a few reference laboratories. OpenSPIM
[5] exists to create an open-source [6] and
community-driven group for the develop-
ment and improvement of open-source light
Single-Objective Light Sheet/
Applied Scientific Instrumentation sheet microscopy. A sophisticated wealth
iSPIM/diSPIM/ct-dSPIM/oSPIM
of material [7-10], including a full parts
list, costings, and construction guide, is
available with step-by-step instructions for
MuVi SPIM/InVi SPIM Lattice building a microscope.
Bruker
Pro/Luxendo LSM The introduction of the Zeiss Z1 in 2012
was the first commercially-available turn-
key system, taking 6 years to develop the
complex lenses to form the light sheet. Since
Evident/Olympus Alpha3 Light Sheet Microscope
then, a wide range of different light sheet
microscopes has appeared on the market
(Tab. 1).
Stellaris 5 and Stellaris 8
Leica Instruments
Digital LightSheet (DLS)
Summary

The optical sectioning confocal microscope

is widely used to create blur-free images.
LifeCanvas Technologies SmartSPIM/MegaSPIM
The light sheet microscope also creates op-
tical sections but only illuminates a single
layer of the sample at a time. It is there-
Airy Beam Light Sheet imaging
fore much faster and less phototoxic than
M Squared Lasers the classical laser confocal microscope.
system
The light sheet is generated as a Gaussian
beam with cylindrical optics in its simplest
configuration. Thinner lattice light sheets
MBF Bioscience ClearScope have evolved from the initial Gaussian
beam configuration, and open-top config-
urations allow large specimens to be inves-
tigated. Concomitant with the evolution of
the light sheet microscope has been the
Miltenyi Biotec UltraMicroscope II
development of various methods to clear
tissues and embryos. The light sheet mi-
croscope was first home-built. OpenSPIM
LS1 Live light sheet exists to create an open-source platform to
Viventis microscopy system aid the self-building of light sheet micro-
scope systems.

Affiliation
Zeiss Lattice Lightsheet 7/ 1Bioimaging Facility, MRC Harwell Institute,
Zeiss
Zeiss Lightsheet 7
Oxfordshire, UK

Table 1: Commercially-available light sheet microscopes (this list may not be exhaustive).
ical clearing to visualize components deep
within the tissue itself. The field of tissue
clearing has grown significantly alongside
the development of Light Sheet Microscopy.
Clearing reagents are principally divided into
three categories:
▪ solvent-based methods
▪ aqueous-based methods
Contact
Jeremy Sanderson
Bioimaging Facility Manager
▪ hydrogel-based methods to preserve MRC Harwell Institute
tissue morphology Oxfordshire, UK
j.sanderson@har.mrc.ac.uk
The solvent-based methods achieve a high
level of tissue transparency relatively quick-
ly within several days but cause substantial
tissue shrinkage and quenching of endoge- References:
nous fluorescence signals. Aqueous and hy- https://bit.ly/IM-Sanderson
drogel-based clearing methods were there-

Imaging & Microscopy 2/2023 • 19

Light Microscopy
Going Digital in Pathology
Introducing Virtual Microscopy and What Questions to Ask
T
he pathology laboratory as an insti-
tution is undergoing change. Even
with a time lag, this immanently
important medical discipline is turning to
digitalization: The laboratory is becoming
virtual. Part of this process is also virtual
microscopy, which supports the change to
digital pathology. Many pathologists still
look through an analog microscope while
deciding whether the small section of tis-
sue that lies in front of them as a sectional
preparation is infused with tumor cells. In
other laboratories, this task is already per-
formed by an automated system that inde-
pendently places the sectional preparations
under a scanning microscope, which scans
the sample and finally, an artificial intel-
ligence identifies, marks, and counts the
tumor cells. To take this step, not only do
you need the right equipment, but you also
need a new workflow in the lab and person-
nel trained to do it. This article will help to
highlight the challenges along the way and
the issues that arise.
20 • Imaging & Microscopy 2/2023
Katharina Eser1
Shortage of Pathologists Worldwide
Today cancer rates are increasing and at the
same time, the number of people who can
treat and detect it is decreasing. Many parts
of the world are medically underserved, but
even in the richest countries, there is a short-
age of specialists like pathologists. The rea-
sons for this range from too little education
and advertising during medical school to the
emotional factor that working in a labora-
tory is isolating and contact with patients
is often limited to looking at their tissue.
But there is also the fact that most diseases
become more complex the longer they are
looked at. The amount of data that would
be necessary to identify certain correlations
cannot be provided by human hands. There-
fore, the possibilities that the digitization of
a pathology laboratory brings are infinitely
attractive.
An important mainstay of pathology is
viewing tissue samples under a microscope.
Virtual microscopy offers users the abil-
ity to digitally microscope specimens inde-
pendent of time and location. For this pur-
Fig. 1: Virtual microscopy offers users the
ability to digitally microscope specimens
independent of time and location.
© PreciPoint GmbH
pose, microscopic preparations are digitized
and can thus be viewed and processed on
a screen later irrespective of the location
and/or workstation. These digital prepara-
tions can be stored in databases and shared
with an infinite number of users. To gener-
ate a digital image of a specimen, an analog
microscope can be used, equipped with an
additional camera. However, development in
pathology is tending toward the use of dig-
ital microscopes. Depending on the model,
these can usually not only produce a live
image of the specimen but also scan it. Dig-
ital microscopes not only show a single field
of view but scan the entire specimen.
Digitized microscopy slides can be called
virtual slides, scans, or whole slide images.
These terms describe a fully digitized micros-
copy specimen. To produce digital images,
the instrument scans the entire specimen
on the slide piece by piece. The software
merges the resulting high-resolution indi-
vidual images into a complete image. This
process is called stitching. On the computer,
users can navigate over the sample, magnify
it and analyze it.

Table 1: Having a digital workflow makes the work in a pathology lab faster, more efficient, and makes room for innovations.
Quality of the Specimen Is Essential
As with all microscopic procedures, the
quality of the specimens also plays a ma-
jor role in virtual microscopy. The speci-
mens must be cut as evenly as possible, as
the software automatically sets focal points
during the scanning process. Excessive
height differences can lead to plane jumps
and blurred areas in the finished scan and
cannot be corrected. The specimens must
also be within the fixed scan area of the
instrument. Specimens must be uniformly
stained to correctly represent all cell struc-
tures. In addition, air pockets, overlaps, and
other contamination of the samples should
be avoided. In special cases, the nature of
the sample recedes into the background. For
example, during surgery for tumors, sections
of the removed tissue are often made during
the procedure, the so-called frozen section.
Then only certain regions of the sample are
viewed under the microscope.
The quality of the digital specimen is also
dependent on the quality of the camera used.
A camera attachment on an analog micro-
scope usually does not provide high quality,
as these systems are not designed for digi-
tization processes. Digital microscopes are
designed for this process and, in addition to
the scanning function, have a live view so
that the specimen can be viewed live on the
screen. Pure slide scanner devices offer the
user the possibility to choose between speed
and resolution. A higher scanning speed
leads to a loss of image quality. However,
since these devices operate autonomously,
the time loss can also be adjusted by adjust-
ing the working time of the scanner, for
example at night.
To make full use of microscopic scans,
suitable image-viewing software is required.
Depending on the image format, only very
specialized programs can process images for
pathology slides. The so-called viewing soft-
ware offers different possibilities to evaluate
the images as well. With different annotation
tools, for example, lines and circles can be
drawn in or written notes can be attached. In
addition, it is also possible to integrate arti-
ficial intelligence into such programs. With
the help of integrated AI, an automatic eval-
uation of certain structures or cells becomes
possible. Ideally, the annotations and evalu-
ations can be stored according to the images.
It is possible to integrate viewing software
into a cloud. That way scans can not only
be shared with other users via the web server
but can also be viewed directly on the plat-
form. In addition, specific information about
the images can often be provided. In most
cloud services, image storage, image sharing,
and image viewing facilities are available.
Viewing the scans is possible with any end
device. It doesn’t matter whether it’s a large
screen, smartphone, tablet, or laptop. How-
ever, the nature of the screen is decisive for
the reproduced image quality [1].
Light Microscopy
© PreciPoint GmbH
Pathology Today Is Manual Work
Currently, in most cases, samples that need
to be examined in the pathology laboratory
arrive with a submission slip on which it is
noted by hand what is to be done with them.
This information is transferred to the lab-
oratory information system by staff. After
a macroscopic examination of the tissue by
a pathologist, medical technicians prepare
the samples for further examination. With
sometimes considerable manual effort, these
specimens are prepared, cut, fixed in kero-
sene, and stained using various histochem-
ical and immunohistological techniques;
they are cut, mounted on a glass slide, and
covered with glass. The specimens are then
sorted into folders and submitted to pathol-
ogists for examination. In some cases, the
specimens are also scanned. The specimens
must also be manually inserted and regis-
tered for this purpose. If there are quality
defects, the process must be repeated. This
workflow, which is only roughly outlined
here, involves many manuals, and small-
scale work steps with numerous sources of
error.
At the other end of the development
towards a fully digitalized pathology labora-
tory, the automatic scanning of large quan-
tities of sectional preparations and digital
provision for diagnostics in conjunction with
clinical data as well as digital report text
generation are on the horizon. The system
Imaging & Microscopy 2/2023 • 21
Light Microscopy
can prioritize and process orders after the
registration of incoming specimens, and it
handles quality control. In addition, artificial
intelligence is used to support histopatho-
logical diagnostics. Furthermore, the system
can integrate the analyzed image data and
molecular information into the workflows. In
the meantime, several research projects are
approaching the realization of this vision,
revealing the actual opportunities and chal-
lenges of such a theory.
Algorithms Open a Wide Range
of Possibilities
Although there are advantages to a digital
image, it does not solve the many questions
and requirements that users have. However,
digitization opens a wide range of possi-
bilities in image analysis using algorithms.
Classical algorithms can detect and count
well-defined structures such as tumor cells.
This enables the pathologist to quantify by
a concrete measured value. In doing so, the
algorithm proceeds efficiently and without
bias. Factors such as stress or time pressure
as well as the influence of optical illusions,
which affect humans, do not occur here.
There are now many products on the mar-
ket that can be used for different analysis
approaches. These programs can find pre-
defined structures quickly and efficiently
and quantify them reproducibly. There are
numerous studies describing the application
of algorithms in histological preparations of
different organs and various diseases [3].
Typically, these algorithms are trained so
that experts mark defined structures in a his-
tological section. The algorithm is trained
with a series of similar sections until it rec-
ognizes the marked structures on its own.
Common programs on the market are usually
Fig. 3: The correct sample preparation is key in virtual microscopy.
22 • Imaging & Microscopy 2/2023
Fig. 2: Having a digital sample makes it possible to have algorithms take over the costly job of
counting and annotating.
specialized for specific disease patterns; their
task is to recognize and quantify predefined
structures. An algorithm can only be as good
as the quality of the data sets with which it
has been trained [4]. The greater the number
and the more variable the spectrum of struc-
tures sought, the better and more reliable the
evaluation. This is where biobanks, which
are now being established worldwide, play
an important role. These provide not only
many physical samples but also numerous
samples that have already been digitized. The
next step is algorithms that are trained spe-
cifically for the application needs of users.
Here, too, an interesting range of products
is developing [2].
The challenge is the question of what for-
mat the data sets obtained are converted into
and how they are then finally integrated into
the laboratory information system and the
systems of the departments concerned. There
© PreciPoint GmbH
© PreciPoint GmbH
is also, of course, the question of personnel
and workflow in the laboratory.
Conclusion
The transformation of a pathology labora-
tory into a digitized pathology laboratory
can only be implemented step by step. At
the beginning of this process is the docu-
mentation and visualization of all processes,
which must be analyzed according to vari-
ous parameters such as personnel, machines,
and degree of development, and at the IT
and process support level. From this, mean-
ingful planning for the transformation can
emerge. Part of this is virtual microscopy,
the equipment that meets the requirements,
and the algorithms that support the work.
There are now many companies that special-
ize in helping labs with this transformation.
This is a very sensible service because this
transformation is complex, it costs time and
money, and it also must be well supported in
terms of personnel to function.
Affiliation
1PreciPoint, Freising, Germany
Contact
Katharina Eser
PreciPoint GmbH
Freising, Germany
katharina.eser@precipoint.de
www.precipoint.com
References:
https://bit.ly/IM-Eser

Turnkey Multiphoton Microscopy
for 3D Samples
Label-Free and 4D Imaging in Life-Science and Medicine
C
ompared to epi-fluorescence wide-
field (WF) and laser scanning confo-
cal techniques, multiphoton micros-
copy (MP) is highly superior when it comes
to imaging depth or the identification of
label-free biomarkers. Here, excitation
with a femtosecond laser typically between
780-1300  nm produces nonlinear optical
fluorescence. The intensity of the gener-
ated signal increases with the square of
the laser peak power (for 2-photon) or the
third power (for 3-photon). This phenom-
enon is confined to a very tight focal vol-
ume, significantly reducing the absorption
cross-section from out-of-focus planes.
Moreover, using wavelengths in the NIR
range leads to lower scattering in tissue
and thus yields higher penetration depth
and lower photodamage compared to linear
confocal techniques [1]. By engaging dif-
ferent imaging modalities, a broad range of
applications including 3D, label-free, deep
tissue, live-animal, whole organ, and whole
slide imaging can be addressed. Here, we
present high-quality imaging of 3D and 4D
samples as well as a label-free and non-de-
structive imaging of unique biomarkers to
understand structural and functional rela-
tionships in healthy and diseased tissue with
time-saving, multimodal MP microscopy
delivering orthogonal informational content
and enhanced imaging depth.
One Size Fits All:
From the Micro to the Macro Scale
3D and 4D ex  vivo and in  vivo imaging
have gained increasing importance. The
Fig. 1: One size fits all: the MPX provides a high flexibility and working space for a multiple range of samples from the
macroscale down to the microscale: key applications are 3D and 4D, label-free, deep tissue, whole slide, whole organ, and
live-animal imaging.
Stefanie Kiderlen1 and Lukas Krainer1
key element here was the understanding
that cells behave differently in a 3D sur-
rounding than in standard 2D cell culture
[2]. As a result, experiments are becoming
more complex, i.e., using awake animals or
in vitro 3D cell cultures. The future requires
a next-generation microscope that is agile
and adaptable with features that lower the
user barrier, expand on diagnostics with
multiple modes, and save resources and
time while improving results. With this
concept in mind, Prospective’s MPX was
engineered to be suitable for label-free as
well as 3D/4D imaging from the macro to
the micro-scale. The MPX-multiphoton mi-
croscope is a turnkey, compact, and fully
integrated next-generation multimodal
microscope, combining different imaging
techniques in one easy-to-use and por-
table device: non-linear MP microscopy
(two-photon, higher harmonics - SHG &
THG) and linear WF microscopy epi-fluo-
rescence and fluorescence lifetime (FLIM)
to maximize informational content and
to yield complementary data sets, rang-
ing from single cells up to living animals
as shown in figure 1 [3]. Here, the MPX
addresses a broad variety of samples from
the micro to the macro scale, starting with
ex vivo 3D cellular models comprising so-
called spheroids and organoids.
However, to mimic a healthy 3D sur-
rounding, these models are limited in size
because of arising hypoxia in the core when
growing bigger than ~400 um in diameter
[4]. To overcome size limitations because of
necrotic cores 3D models with a vasculariza-
tion system such as tumors grown on a CAM
membrane, 3D printed cell constructs, or
Light Microscopy
even 4D in vivo imaging can be used to eval-
uate samples on the macro-scale as shown
in figure 2. However, when it comes to sam-
ples in the mm-cm macro scale, additional
sample processing techniques such as tissue
clearing can be applied to enhance imaging
depth and increase the informational 3D vol-
umetric content of these samples, as shown
in figures 1 and 2.
Tissue Clearing for Whole Organ Imaging
Tissue and 3D cell constructs are highly
inhomogeneous, making them challeng-
ing to image due to light scattering and
absorption. Increasing the optical trans-
lucency by tissue clearing techniques can
enhance the imaging depth by a factor
>10. However, tissue-clearing methods
suffer from some drawbacks: organic sol-
vent-based clearing methods shrink the
sample, whereas water-based methods of-
ten don’t yield the same clearing effect [5].
Moreover, only a few reagents are known
for tissue clearing in living organisms, so
clearing protocols are mostly related to
fixed samples. 3D imaging in fixed sam-
ples provides the advantage that there are
many staining and processing methods to
improve the image quality, signal intensity,
or penetration depth, however, they only
represent a snapshot of the tissue. Figure 2
shows a whole cleared mouse brain stained
for neurons (green). The penetration depth
here was limited by the working distance
of the objective (4 mm) and could be in-
creased by tissue clearing by a factor of
>10.
Imaging & Microscopy 2/2023 • 23
Light Microscopy
Intravital 4D Deep Tissue Imaging
To investigate not only structural but also
functional relationships, living tissue and
cells must be imaged in 4D time series. How-
ever, as mentioned before, optical clearing
methods are mostly used for fixed samples.
To do 4D timelapse deep tissue imaging of
living cells and animals, the right sample
must be chosen. In figure 3 we demonstrate
zebrafish 4D time-lapse imaging. Zebrafish
have long served as invaluable subjects in
the fields of metabolic pathology, develop-
mental biology, and neuroscience. Its status
as a vertebrate and its sustained transparency
in the larval stage makes it a highly practical
and expedient live model for research. Figure
3 shows 4D time-lapse neuronal tracking of
retina cells and the migration of neurons in
the spinal cord of 4-day-old zebrafish larvae.
Another prominent example of 4D
in vivo imaging is the investigation of neu-
ronal activity in mouse brain using standard
markers such as GCaMP as shown in fig-
ure 2. GCaMP is a fluorophore-tagged and
genetically modified calcium indicator that
responds to the efflux of calcium (Ca2+) [6].
Overall, higher penetration depths combined
Fig. 2: Intravital MP mouse brain imaging of GCaMP-expressing neurons in the cortex and deep
tissue imaging of a fixed and cleared mouse brain. a) Microscope setup of live mouse in-vivo
brain imaging and b) close-up of the cortex at ~300 µm z-position in depth, c) whole cleared
mouse brain, d) volume scan of ~3.5 mm in depth, and e) one exemplary z-plane showing
stained neurons (green). Scalebars: 100 µm
24 • Imaging & Microscopy 2/2023
with lower photodamage are major advan-
tages of MP microscopy.
Label-Free Biomarkers for
Translational Diagnostics
The identification of label-free biomarkers
using multimodal MP imaging has rapidly
spread throughout biomedical research in
the past few years. A few nonlinear optical
modalities are being extensively considered
for clinical purposes. For example, two-pho-
ton, three-photon, second-harmonic gen-
eration (SHG), third-harmonic generation
(THG), coherent anti-stokes Raman spectros-
copy (CARS), stimulated Raman spectrosco-
py (SRS), and fluorescence lifetime imaging
(FLIM) are all used in research [7,8]. They all
have unique advantages and disadvantages.
However, combining modalities to maximize
informational content and yield complemen-
tary data is necessary when it comes to na-
tive biological samples since they are highly
heterogeneous [1]. As described above this
heterogeneity on the one hand making it
challenging to image these samples, but on
the other hand allows to identify label-free
intrinsic biomarkers. Here, timesaving,
depth-resolved, sample-saving and non-in-
vasive data acquisition is highly beneficial.
Label-free biomarkers can originate from op-
tical or structural properties and phenotype
of the sample. Currently, the workflow for
diagnosis from a biopsy or resection is tis-
sue fixation followed by paraffin sectioning
and histological staining e.g., H&E staining.
This is time-consuming, requires experienced
technicians, and includes toxic staining re-
agents. The use of label-free biomarkers in
pathology would provide a faster and easi-
er diagnosis. Here, we demonstrate SHG and
two-photon imaging as powerful diagnostic
modalities providing endogenous molecular
and chemical distinction for differentiating
healthy from diseased tissue as shown in
figure 4. The combination of both modali-
ties provides a label-free contrast originating
from molecules such as NAHD, FAD, elastin,
proflavine, and hemoglobin, providing in-
trinsic endogenous fluorescence or structural
properties from non-centrosymmetric mole-
cules like fibrous collagen [9]. Morphologi-
cal and structural changes in the cell or the
surrounding matrix e.g. the nucleus shape
in cancerous tissue or the alignment and
amount of collagen fibers in wound healing,
fibrosis, and tumors can serve as valuable di-
agnostic tools.
Engage Imaging Speed and Resolution –
Epi-Fluorescence Guided MP Microscopy
Even though non-linear microscopy tech-
niques are highly beneficial for deep tis-
sue and label-free imaging they are scan-
ning-based techniques with a limited ac-
quisition speed making them slower than
epi-WF microscopy. However, engaging the
speed of WF and the confocal depth resolv-
ing properties of non-linear microscopy can
be used for epi-fluorescence WF-guided MP
microscopy. Here, a WF overview fast scan
serves as guidance to the region of interest
(ROI) which is then imaged in detail using
MP microscopy [3]. The capability to per-
form multimodal analysis, on the same re-
gion of interest without the inconvenience
of moving the sample to another system,
maximizes content, provides orthogonal
data, and saves time as shown in figure 4.
Conclusion and Outlook
High-quality imaging of 3D and 4D samples
as well as label-free and non-destructive
imaging of unique biomarkers is becoming
more important in understanding cellular
mechanisms and their structural and func-
tional relationship. Engaging different im-
aging modalities, a broad range of key ap-
 Light Microscopy

Fig. 3: MP 3D volume scan (a) and timelapse imaging (b-k) of zebrafish larvae. a) 3D projection of a fixed zebrafish larvae expressing prox1:Ta-
gRFP and stained with Hoechst (cell nuclei – blue) and Phalloidin (actin – green). b-f) Migration of retina cells in a living zebrafish eye imaged by
time-lapse microscopy. 4-day-old zebrafish larvae expressing tp1:LifeAct-mCherry (red) with the respective close-ups. Scalebar: 50 µm (b) and 20
µm (c-f). g-k) Neuronal tracking of cells in the spinal cord in a living 4 day old zebrafish larvae alpha-catenin-YFP (green), tp1:mCherry-NLS
(red), and the SHG signal from the muscle (blue) with the respective close-ups. Scalebar: 100 µm (g) and 20 µm (h-k).

copy techniques such as WF-guided MP

microscopy can save time and provide addi-
tional data by delivering fast overview scans.
Future developments include improvements
for even higher imaging resolutions and
depth by implementing adaptive optics (AO)
techniques and three-phtonon microscopy.

Acknowledgments

We are thankful for the support from the fol-

lowing collaborators: Amelie Erben and Dr.
Stefanie Sudhop, UAS Munich; Dr. Kim Fer-
rari and Prof. Bruno Weber, University Zurich;
Dr. Sara Caviglia and Professor Stephan Neu-
hauss, University Zürich, Dr. Linda Waldherr,
University Graz, Dr. Branislav Zagrapan, LKFH
Feldkirch and Monika Puchalska, Nencki Insti-
tute of Experimental Biology Warsaw for pro-
viding samples, images and dedicating time
and resources to the experiments mentioned.
Fig. 4: Epi-fluorescence widefield guided multiphoton microscopy of a human label-free skin mela-
noma FFPE section. Fast WF overview scan (a) and H&E stained ground truth (b) for orientation Affiliation
1Prospective Instruments LK OG, Dornbirn,
and respective close-ups of healthy tissue (d,c and h,i), interface between healthy and diseased tis-
sue (e and j), and regions in the lesion (j,g and k,l). a) Endogenous fluorescence collected by epi-flu- Austria
orescence λex./λem. 390/432 nm (magenta), 475/515 nm (green), and 555/595 nm (orange). c-g)
Two-photon endogenous fluorescence collected at λex./λem 1040/542nm (green) and 1040/595 nm Contact
(red) and the SHG signal at 520 nm (blue). Scalebar: 1 mm (a&b) and 200 µm (c-l). Dr. Stefanie Kiderlen
Prospective Instruments LK OG
Dornbirn, Austria
plications including 3D imaging, label-free, ing depth, and save time in 3D in vivo and stefanie.kiderlen@p-inst.com
deep tissue, live-animal, whole organ, and ex  vivo samples as well as large area tis-
whole slide imaging can be addressed. sue sections. Higher penetration depth can
Here, we demonstrated the use of mul- be achieved by tissue-clearing techniques to References:
timodal MP microscopy to gain orthogo- enhance optical translucency in fixed sam- https://bit.ly/IM-Kiderlen
nal informational content, enhance imag- ples. We showed that using different micros-

Imaging & Microscopy 2/2023 • 25

 Correlative Microscopy

© TU Wien
Correlative Microscopy of a Catalytic Reaction
Zooming in on Chemical Patterns in Hydrogen Oxidation on Rhodium
Philipp Winkler1 and Günther Rupprechter1

S
tudying spatiotemporal phenome-
na can deepen the understanding of
catalytic reactions and thus enable
the designing of better catalysts. Often,
only combining several microscopy tech-
niques in a correlative approach allows for
capturing the full picture. Using three dif-
ferent electron microscopies, the forma-
tion of chemical patterns in catalytic H2
oxidation on Rh was studied and different
types of patterns and an island-mediat-
ed propagation mechanism for reaction
fronts were observed [1].
history (i.e., it exhibits multistable kinetics)
[4]. In addition, the surface can also undergo
periodic switches between these states of
catalytic activity at some constant exter-
nal parameters (i.e., it displays self-sustain-
ing kinetic oscillations) [5]. As the switches
between the two states occur via the spread-
ing of reaction fronts, this can result in the
formation of chemical patterns on the cata-
lyst surface [6]. An atomic-level understand-
ing of the mechanisms behind these phe-
nomena, which result from the interaction
of reactants with active sites on the cata-
lyst surface [7], enables improving the exist-
ing and tailored design of new catalytically
active materials.
Due to the inherently necessary lateral and
temporal resolution for such studies, micros-
copy techniques are ideally suited to tackle
these questions. Often, however, a single tech-
nique alone cannot provide all data required
for full understanding and several micros-
copies are thus combined in a correlative
approach.

Introduction

Heterogeneous catalysis is one of the key

enablers of sustainable energy generation
and storage. For example, renewable energy
can be stored in the chemical bonds of H2
and released on demand via catalytic H2 ox-
idation in a fuel cell, forming only water as
a product [2]. A fundamental understanding
of this reaction is therefore crucial for the
efficient operation of fuel cells [3].
Hydrogen oxidation on Rh can be in two
distinct states of catalytic activity, depending
on the set of external parameters: In the inac-
tive state, the surface is covered by oxygen
and water formation is inhibited, while in the
active state, the surface is nearly adsorbate
free and water is formed. In some regions
of the parameter space, the system can even
be in both states of catalytic activity at the Fig. 1: The experimental approach. Three different electron microscopies, each providing differ-
same set of external parameters, where the ent types of information, were applied in a correlative way in order to study the formation of
actual state only depends on the system pre- chemical patterns in catalytic H2 oxidation on Rh.

26 • Imaging & Microscopy 2/2023


Methods
For in situ studying pattern formation in
catalytic H2 oxidation on Rh, three different
microscopies were combined, each of them
yielding complementary pieces of informa-
tion, as shown in Figure 1 [1]. In UV pho-
toemission electron microscopy (UV-PEEM),
the sample is illuminated by UV-light and the
resulting low-energy photoelectrons are used
for imaging by an electron optical system. The
local image brightness is based on the local
surface coverage of adsorbed reactant mole-
cules and can thus serve as an indicator for
the local catalytic activity of the surface (ki-
netics by imaging [8]). When the sample is in-
stead illuminated by monochromatic X-rays
using X-ray photoemission electron micros-
copy (X-PEEM), similarly, photoelectrons are
generated. In X-PEEM, the photoelectrons,
however, possess higher energy as they origi-
nate from the atomic core levels and therefore
carry information on the surface composition
and identity. By selecting and using only pho-
toelectrons of specific energy for imaging, di-
rect local chemical information is encoded in
the image. Due to the high brilliance of the
X-ray beam required for X-PEEM, such ex-
periments can typically only be performed at
synchrotron light sources. In low-energy elec-
tron microscopy (LEEM), the sample is illumi-
nated by low-energy electrons. The electrons
elastically backscattered at the sample surface
are then used for imaging, whereby the local
Fig. 2: Correlative microscopy of kinetic oscillations in catalytic H2 oxidation on Rh at constant T = 468 K, p(O2) = 5.0 x 10-7 mbar, p(H2) = 4.5 x
10-7 mbar. The oscillations were visualized by correlative X-PEEM (a) and LEEM (b). Two types of reaction fronts were identified using the chemi-
cal sensitivity of X-PEEM and allowed the understanding of the image contrast in LEEM (frames 1, 2: slow oxygen front; frames 3, 4: fast hydro-
gen front). Within the yellow dashed rectangles in (a), the image contrast is enhanced for better front visibility; the insets show line intensity pro-
files along the A-A’ lines.
image brightness depends on the local diffrac-
tive properties of the sample surface. These
properties are influenced by several factors,
including the local atomic structure or the
presence of adsorbates on the sample surface.
All three techniques were combined for in
situ studies of the oscillating catalytic H2 oxi-
dation in the 10-7 mbar pressure range on indi-
vidual domains of a polycrystalline Rh foil.
The local atomic structure of each of these
domains (several 100 µm in size) was deter-
mined beforehand, turning the sample into a
library of different surface structures [9]. All
three microscopies were applied to the exact
same domains on the same sample under
identical external parameters (temperature,
reactant pressures), i.e., in a correlative way.
Due to being hosted in a single experimental
setup (the SMART instrument at UE49PGM
beamline of the BESSY II synchrotron light
source [10]), X-PEEM and LEEM were even
performed in a single experiment by switch-
ing between the two operating modes of the
instrument, representing an example of cor-
relative microscopy in its purest form.
UV-PEEM Studies
UV-PEEM studies of the oscillating mode of
catalytic H2 oxidation on Rh in the past few
years revealed a number of effects, which
influence the observed spatiotemporal struc-
tures [5, 11, 12]. All of these studies, how-
Correlative Microscopy
ever, had one common aspect: The observed
type of spatiotemporal behavior was always
characteristic of each domain in its entirety.
Just recently, during further UV-PEEM
studies, a curious situation was observed:
on the same sample at specific exter-
nal parameters some domains were in a
steady state of catalytic activity or exhib-
ited patterns spanning the whole domain,
while other domains appeared to be frag-
mented into multiple areas oscillating inde-
pendently from each other or not at all. Due
to the limited lateral resolution of the used
UV-PEEM, details were hard to discern and
these observations warranted further anal-
ysis by experimental techniques with higher
spatial resolution.
Correlative X-PEEM and LEEM Studies
X-PEEM and LEEM, which are examples of
such techniques, were used to zoom in on
those domains displaying multiple uncor-
related areas with a resolution down to sev-
eral nanometers. While the image contrast
in UV-PEEM and X-PEEM is a direct result
of the surface chemistry and is thus easy to
understand, the diffractive properties of the
surface responsible for the LEEM contrast
are often hard to predict.
Therefore, in the first step, the oscillating
mode of catalytic H2 oxidation on Rh was
observed using both X-PEEM and LEEM and
Imaging & Microscopy 2/2023 • 27
 Correlative Microscopy
by switching between these operating modes
after each oscillation cycle (Fig. 2). Two differ-
ent types of reaction fronts were observed and
identified by selecting a proper energy win-
dow for X-PEEM imaging: The slow reaction
front (dark in X-PEEM, bright in LEEM) was
related to spreading oxygen coverage, while
the fast one (bright in X-PEEM, dark in LEEM)
is caused by diffusing and reacting hydrogen.
Chemical Patterns
Due to the better suitability of LEEM for
studying fast processes, the formed spatio-
temporal patterns were then imaged primar-
ily by LEEM. Visualizing the ongoing oscil-
lations in a temperature range from 428 K
to 468 K at constant p(O2) and p(H2) led to
the observation of three different types of
patterns (Fig. 3): At 428 K, rotating double
spirals occurred; increasing the temperature
to 433 K resulted in the propagation of broad
reaction fronts over the whole field of view.
For oxygen fronts, an unusual island-me-
diated front propagation mechanism could
be detected, where small oxygen islands
nucleate ahead of the main front and then
grow, finally merging with the main front.
Upon increasing the temperature even more,
to 468 K, just small oxygen islands were
28 • Imaging & Microscopy 2/2023
formed, which were consumed by hydrogen
fronts before having the chance to merge
with other islands.
These observations, which can be
explained by considering the differing rel-
ative velocities of oxygen and hydrogen
fronts at different temperatures, enabled us
to rationalize the peculiarities observed in
the previous UV-PEEM experiments.
Conclusion
A correlative microscopy study of chemical
pattern formation in H2 oxidation on Rh
was performed using UV-PEEM, X-PEEM,
and LEEM, tackling important challenges
in studying catalytic surface reactions. Due
to their higher lateral resolution, X-PEEM
and LEEM allowed zooming in on process-
es just barely visible in UV-PEEM. Differ-
ent types of spatiotemporal patterns and an
island-mediated propagation mechanism of
oxygen fronts on the catalyst surface during
kinetic transitions could be detected.
Acknowledgment
This work was supported by the Austrian
Science Fund (FWF) (P 32772 N and F81
Fig. 3: Chemical patterns in
catalytic H2 oxidation on Rh at
constant p(O2) = 5.0 x 10-7
mbar, p(H2) = 4.5 x 10-7 mbar,
observed by LEEM. At 428 K
(a) rotating double spirals could
be observed, while at 433 K (b)
broad reaction fronts propagat-
ed over the whole field of view,
and at 468 K (c) small islands
are formed, which collapse be-
fore being able to merge. The
QR codes lead to corresponding
videos, published under a CC
BY 4.0 license as Supplementa-
ry Material for Ref. [1].
P08) and the Helmholtz-Center Berlin for
Materials and Energy (HZB) by the alloca-
tion of beamtime 212 10440 ST. The SMART
instrument was financially supported by the
Federal German Ministry of Education and
Research (BMBF) (05 KS4WWB/4), as well
as by the Max Planck Society. We are grate-
ful to J. Zeininger, M. Raab, Y. Suchorski,
M.J. Prieto, L.C. Ta˘nase, L. de Souza Caldas,
A. Tiwari, T. Schmidt, M. Stöger-Pollach, A.
Steiger-Thirsfeld, and B. Roldan Cuenya for
contributing to the original work.
Affiliation
1Institute of Materials Chemistry, TU Wien,
Vienna, Austria
Contact
Prof. Dr. Günther Rupprechter
Institute of Materials Chemistry
TU Wien
Vienna, Austria
guenther.rupprechter@tuwien.ac.at
References:
https://bit.ly/IM-Rupprechter
 le.tab
look-up
t-based
; gradien
ded algae
n embed
ion of resi
Visualizat
TESCAN 3D FIB-SEM
TOMOGRAPHY
Go from data acquisition to visualization without the need for laborious manual data
segmentation. Visualize data in seconds with a combination of two-dimensional
intensity AND gradient look up table.

A single click imports the dataset and the step-by-step wizard guides you through
the pre-processing steps to create a complete, detailed 3D visualization.

At TESCAN we specialize in nano-scale tomography, designing instruments with

users in mind.

What can you do with TESCAN 3D FIB-SEM Tomography?

Find out more.

www.tescan.com
Electron and Ion Microscopy

4D Scanning Transmission Electron Microscopy

with Event-Based Electron Detection
How to Overcome the Bottleneck in Scanning Speed
Daen Jannis1,2, Christoph Hofer1,2, Chuang Gao1,2, Timothy Pennycook1,2, Johan Verbeeck1,2

4
D Scanning Transmission Electron
Microscopy (4D STEM) records a
diffraction pattern at each probe
position while scanning the electron beam
over the specimen where the probe posi-
tion dependence of the electron scattering
is recorded in detail with a Timepix3 cam-
era. This method offers a wealth of infor-
mation that can be used to perform ad-
vanced analyses that are impossible with
monolithic integrating detectors. Howev-
er, very slow scans have been the norm in
4D STEM due to the speed of available
cameras. Here, this speed bottleneck is
overcome by using an event-driven camera.

Introduction
In Scanning Transmission Electron Micros-
copy (STEM), a probe formed by a beam of
electrons scans the specimen. Images are
formed by collecting the scattering as a
function of position. Conventional images
use detectors that integrate over an angu-
lar range of this scattering, providing an
intensity value at each probe position. For
example, annular dark field (ADF) images,
in which intensity is proportional to rough-
ly the square of the atomic number, are
formed using an annular detector that inte-
grates the signal from electrons scattered to
higher angles. Many different detector ge-
ometries can be used, but by capturing the
scattering information in detail with a full
2D image at each probe position, 4D STEM
enables all these to be applied to the same
scan. In addition to conventional signals,
4D STEM is used in a variety of experiments
that would not be possible with convention-
al monolithic integrating detectors, such as
structural phase mapping, strain mapping
in crystalline or amorphous materials, and
atomic resolution phase imaging [1].
However, a major drawback to 4D STEM
has been the slow scan speeds imposed
by the relatively low frame rate of cam-
eras compared to the probe dwell time in
a fast STEM scan. Slow scans are problem-
atic because they lead to high doses and
beam damage in sensitive samples, and just
as importantly, distortions in the images due
to instabilities, such as sample drift at these
Fig. 1: (a) The incoming probe is scanned over the specimen in a schematic setup, with the pix-
elated detector placed in the far field. The averaged diffraction pattern over the scan is indicated
as PACBED. (b) Illustration showing the difference between frame and event-based detection.
high magnifications. To put the problem
in perspective, most cameras for 4D STEM
in recent years have had a frame rate of
a few thousand frames per second. A fast
STEM scan will typically use a dwell time
of around µs. To match this speed, a cam-
era needs to provide a million frames per
second, which has yet to be achieved. How-
ever, this time resolution can be realized by
avoiding frame-by-frame acquisition in the
camera in the first place with event-driven
operations. Each time an electron hits an
event-driven camera the pixel location and
time are read out immediately, avoiding the
time-consuming process of reading out the
whole pixel array. In fast scans, it is often
the case that relatively few electrons strike
the detector and thus most pixels contain
only zeros and their readout wastes time in
a conventional framing camera. The event-
driven Timepix3 camera used can detect
single electrons with a time resolution of
1.56 ns, and easily provides µs 4D STEM
[2]. Here, these datasets are used to perform
dose-efficient phase imaging at atomic res-
olution using integrated center-of-mass
(iCOM) [3] and single-sideband ptychogra-
phy (SSB) [4]. The results are compared with
the conventional ADF and annular bright
field (ABF) images.
Experimental Setup
A schematic of the setup is shown in Fig-
ure 1(a). The signal driving the scan coils is
synchronized with the Timepix3 detector.
Since both the scan generator and detec-
tor are synchronized to the same clock, the
time-of-arrival of each event can be used to
determine the probe position it came from.
Hence, every detected event has four coor-
dinates, probe position (2D), and scattering
angle (2D). From this event stream, one can
determine virtual images such as ADF and
ABF by selecting the appropriate scattering
angle regions or calculating the center of
mass at each probe position. If multiple scans
are acquired, the shifts between the con-
secutive scans are calculated and this cor-
rection can be applied directly to the event
stream. Besides the speed improvement, the
event-driven operation provides natural data
compression by avoiding the readout of pix-
els containing zero counts, especially in low-
dose experiments. For example, 10 scans with
1024x1024 probe positions result in 86 Gb of
data with a 256x256 frame-based camera in
1-bit mode. Larger bit depths will of course
result in much larger file sizes. Note that file
size is independent of the incoming electron
current for a frame-based camera. For event-

30 • Imaging & Microscopy 2/2023


based detection, the file size only depends on
the electron current and the acquisition time,
which for 3 pA and a 1 µs dwell time results
in only 1.6 Gb for the same scan and pixel ar-
ray dimensions. One important aspect which
should not be overlooked in these types of
detectors is that the maximum incoming
count rate is limited. This Timepix3 setup is
limited to a maximum of 40 million hits per
second, corresponding to a probe current of
around 3 pA, an order of magnitude smaller
than conventional STEM currents (50  pA).
This can be addressed by incorporating ad-
ditional Timepix3 chips and readouts. How-
ever, small electron currents are often desir-
able, especially for beam-sensitive materials
such as organic materials, which undergo
structural changes at higher currents. There-
fore, this type of detector is ideally suited for
low-dose experiments, especially as the 4D
STEM-based phase imaging modes provide
greater contrast and SNR for a given dose.
Methods and Results
In this study, iCOM and SSB were performed as
highly efficient phase contrast imaging meth-
ods using the Timepix3 detector. The iCOM
algorithm tracks the center of mass, which is
proportional to the projected electric field of
the specimen, to reconstruct the electric po-
tential, which in turn changes the phase of the
electrons. The other reconstruction algorithm
used in this study is the SSB ptychography,
which directly retrieves the phase of the object
from the diffraction patterns.
In Figure  2, the different reconstructed
images from a silicate-1 zeolite are shown
using a dose per scan of 300 e-/A2. 10 scans
with 1024x1024 probe positions were col-
lected at 1 µs dwell time. Notably, the iCOM
and SSB images have much better contrast
than the ADF and ABF images, making them
ideal for low-dose experiments. To show
that the detector can also detect electrons
Fig. 2: A 10x1024x1024 scan at 1 µs dwell time. The acceleration voltage used during the ac-
quisition is 60 kV and the specimen is a silicate-1 zeolite which is known to be beam sensitive.
The total dose for the entire acquisition is 3,000 e-/A2.
Fig. 3: A 3x1024x1024 scan at 6 µs dwell time. The acceleration voltage used during the acquisition
is 60 kV and the specimen is a monolayer of WS2. The total dose for the experiment was 18,000 e-/A2.
at 60 kV, a 3-scan 1024x1024 probe position
dataset was collected at 6 µs from 2D WS2.
The stability of the material allowed a higher
dose of 18,000 e-\A2, which is required for
decent signal-to-noise ratio phase images
due to its lower scattering power. The dif-
ferent reconstructions are seen in Figure 3.
In the ADF images, only the tungsten atoms
are visible due to their high scattering power.
Both the tungsten and sulfide atoms can be
visualized by iCOM and SSB making these
imaging techniques suited to image simul-
taneously low and high Z atomic columns.
Summary
Event-driven hybrid pixelated detectors en-
able 4D STEM experiments with microsecond
dwell times. The limit with the event-driven
Timepix3 camera described here is not the
detector speed, which has nanosecond time
resolution, but instead the bandwidth of the
scan coils. Scanning at low dwell times is
essential for low-dose imaging to minimize
the specimen damage imparted by the elec-
Electron and Ion Microscopy
tron beam with a focused probe, where the
efficiency of 4D STEM phase imaging meth-
ods can be most effective.
Acknowledgments
The authors acknowledge the European
Union’s Horizon 2020 Research and Inno-
vation Program under Grant Agreement
No. 101017720 “eBEAM: Electron beams
enhancing analytical microscopy” (J.V. and
D.J); G042920N “Coincident event detec-
tion for advanced spectroscopy in transmis-
sion electron microscopy” (J.V. and D.J.);
n. 101094299 “The IMPRESS project which
has received funding from Horizon Europe
framework program for research and inno-
vation.”(J.V.). No. 802123HDEM funding
from the European Research Council (ERC)
under the European Union’s Horizon 2020
Research and Innovation Programme via
Grant Agreement No. 802123HDEM (C.G.,
C.H., and T.J.P.); the European Union’s Hori-
zon 2020 Research Infrastructure — Integrat-
ing Activities for Advanced Communities
under Grant Agreement No. 823717 — ES-
TEEM3; The FWO Projects G013122N “Ad-
vancing 4D STEM for atomic scale structure
property correlation in 2D materials” (C.H.).
Affiliation
1EMAT, University of Antwerp, Belgium
2NANOlab Center of Excellence, University
of Antwerp, Antwerp, Belgium
Contact
Dr. Daen Jannis
EMAT
University of Antwerp, Belgium
daen.jannis@uantwerpen.be
References:
https://bit.ly/IM-Jannis
Imaging & Microscopy 2/2023 • 31
Electron andIon
Electron and IonMicroscopy
Microscopy
Can AI Do your Ptychography?
A Machine Learning Approach to Real-Time Phase Imaging with 4D STEM
Thomas Friedrich1, Chu-Ping Yu1, Johan Verbeeck1, Sandra van Aert1
In this study, a method using deep learn-
ing for the reconstruction of phase images
from 4D Scanning Transmission Electron
Microscopy (4D STEM) data is presented.
The process retrieves electron exit waves
locally for each scan position, based on a
set of 3x3 convergent beam electron dif-
fraction patterns (CBEDs) around each probe
position, and subsequently uses the phase
object approximation to solve for the phase
object. We show that the method is capa-
ble of super-resolution imaging and good
noise robustness. The contrast depends on
the atomic column type and thickness. The
combination of these properties makes the
method unique among live imaging meth-
ods in 4D STEM.
Introduction
The development of direct electron detectors
(DED) opened new possibilities for Scanning
Transmission Electron Microscopy. Higher
frame rates and almost perfect quantum ef-
ficiency of the cameras allow a much more
efficient collection of convergent beam
electron diffraction patterns (CBEDs) in 4D
STEM experiments. This not only provides
abundant information but also poses chal-
lenges regarding the efficient extraction and
interpretation thereof. Several approaches
have been developed to reconstruct 2D im-
ages from such datasets (e.g., iDPC, iCOM,
Ptychography), with some astonishing re-
sults, pushing the resolution and dose effi-
ciency of STEM to new levels. Particularly,
32 • Imaging & Microscopy 2/2023
ptychographic methods for phase object
reconstructions gained popularity for their
super-resolution capabilities.
However, these methods are computation-
ally demanding and often iterative in nature,
which limits their use mainly to post-acqui-
sition scenarios. This may leave the micro-
scope operator practically blind during low-
dose experiments and calls for efficient,
real-time capable phase object reconstruc-
tion methods.
In this study, convolutional neural net-
works (CNN) are used to reconstruct the com-
plex electron exit wave for every probe posi-
tion and use them to compute local patches of
the phase object, based only on CBEDs at the
probe position and its immediate surround-
ing. This process can be very fast and is inher-
ently local, such that it can be performed live
during acquisition. A machine learning setup,
taking into consideration the physics at play
and mathematical constraints, ensures that
predictions are consistent and faithful.
Methodology
When fast electrons interact with electrostatic
potentials, the wavelengths of the electrons
are temporarily altered, which creates phase
differences. While traversing a specimen, the
phase shift of an electron accumulates result-
ing into a direct relationship with the poten-
tials of all the atoms it interacted with along
its path. In the phase object approximation
(POA), the sample is assumed to be very thin
such that the electron probe does not change
© University of Antwerp
within the specimen. The phase shift can then
be calculated by simply dividing the electron
exit wave by the electron probe function. The
resulting image can be interpreted as a de-
piction of the projected electrostatic potential
of the specimen. In STEM experiments, the
probe is typically focused and very small, so
this division only yields a patch of the full-
phase object, which can be constructed by a
complex summation of all patches resulting
from the many positions in a scan grid. In 4D
STEM, this requires knowledge of the probe
function and exit waves at all probe positions,
both of which are unknown, however. Here,
the incident waves based on the aperture size
for well-adjusted aberration-corrected micro-
scopes are estimated. The complex exit waves
for all CBEDs are recovered by the developed
CNN.
This CNN was designed to predict the
amplitude and phase of the exit wave at
each probe position, based on a kernel of 3x3
adjacent CBEDs. The estimated probe func-
tion is used as an input and directly added
to the network’s output, which enables global
residual learning and practically means that
the CNN learns to estimate how a specimen
alters a given incident wave function, rather
than building it from scratch. The network is
trained on millions of CBEDs simulated by
a multislice package with crystal structures
from the materials project database at various
low-index zone axis. Each piece of data con-
sists of a 3x3 kernel of CBEDs as features and
the corresponding exit wave in real space at
the center of the kernel as a label. In the sim-
ulations, we ensured the real-space step size

was at least 10% smaller than the
probe radius, effectively main-
taining some probe overlap in
real space. The resulting differ-
ences originating from the dis-
placement of the exit waves from
neighboring positions are utilized
to solve the inverse problem of
retrieving the phase.
The composite loss function
consists of terms accounting for
the exit waveforms in both real
and reciprocal space, as well
as the phase object itself. This
enforces physically meaning-
ful constraints on the CNN and
increases the robustness of its
estimations.
The trained CNN in con-
junction with the phase object
approximation and appropri-
ate weighting and shifting of
the retrieved phase patches pro-
vides a non-iterative, robust, and
fast (>4k CBED/s) phase imaging
method with good dose efficiency
and a resolution that is in princi-
ple only limited by the CBED col-
lection angle.
analyticalscience.wiley.com
Meet the
Greatest Minds
in Microscopy
Access all articles by
scanning the QR-code
Electron and Ion Microscopy
Results and Reconstruction
Characteristics
To test whether the reconstruc-
tion faithfully preserves the
physical properties of the object,
datasets of CBEDs are simulated
by using a single atom as a scat-
terer for the entire periodic table
up to atomic number Z = 102. The
reconstructed phase objects are
averaged over different ranges
around the potential center and
compared to the ground truth in
Figure 1. The phase values, al-
though not quite exact, follow
the theoretical values closely. At
larger integration ranges, the plot
in Figure 1c shows that the orbit-
al contraction at atoms with fully
filled orbitals is captured by the
reconstruction. These single-at-
om datasets were never part of
the training, and thus these re-
sults show that the CNN does not Fig. 1: Phase response of the CNN compared to ground truth transmis-
simply make up images of atoms sion functions for simulations of single atoms throughout the periodic
but reflects, to some level of ac- table at infinite dose for (a) peak intensity, (b) mean over 0.6x0.6 Å
curacy, the physical properties of around the atomic position, and (c) mean over 1x1 Å pixels around the
the observed materials. atomic position.
© agsandrew - stock.adobe.com
Wiley Analytical Science
ARTICLE COLLECTION
Electron and Ion Microscopy

Fig. 3: Reconstruction results of a USY-Zeolite dataset for 2 different

approaches. (a) Neural Network (b) SSB reconstruction and the Fourier
transforms in (c) and (d), respectively. The data preprocessing is de-
scribed in reference [1].

Fig. 2: Reconstructed images of an edge of an Au crystal using (a) Acknowledgments

SSB, (b) HAADF, and (c) CNN reconstruction. Line profiles across the
nanorod illustrate the thickness dependence of the corresponding sig-
nals in panel (d).
In Figure 2, the reconstruction of an exper-
imental gold particle dataset is presented,
and compared with other imaging methods
that are known for their sensitivity to thick-
ness (high angle annular dark field-HAADF)
or other advantages, such as dose efficiency
(single sideband ptychography-SSB). It is
shown that the phase contrast of the CNN
reconstruction resembles the HAADF result,
which scales with the predicted thickness.
This indicates that the phase value can be
used to estimate thickness variations, to
some extent even beyond the limits of the
POA, since ≈10 nm would typically not be
considered a PO. The presence of accurate
thickness contrast further shows that the
CNN can retrieve low-frequency compo-
nents of the phase signal, which are com-
monly lost in SSB.
One of the most significant advantages of
using DEDs over annular detectors is the fact
that they collect most of the transmitted elec-
trons in the beam. Methods that can exploit
this fact may yield a much stronger contrast.
This is proven true in the reconstruction of
a silicalite-1 zeolite sample, taken at a dose
that does not even permit HAADF imaging. In
Figure 3, the reconstructed results of both the
CNN and SSB show a strong contrast. The CNN
reconstruction benefits furthermore from the
dark field electrons, which directly contrib-
ute to the high-frequency features. This infor-
mation is missing in SSB, as it only utilizes
the intensity variation inside the bright field,
and the effect is reflected by the higher fre-
quency components in the result, indicated by
the Fourier transform (FT) of the images. The
comparison of the FTs further illustrates again
the presence of low frequencies in the CNN
reconstruction and a contrast transfer with an
overall wider frequency range.
Conclusion
A new reconstruction method is proposed that
uses a CNN for exit wave reconstructions and
the POA to solve for phase objects, utilizing
all the electrons in the region covered by the
detector for dose-efficient high-resolution
imaging. The reconstructed phase values can
be correlated with atoms’ projected potentials
and the resulting image contrast is dependent
on atomic species and the specimen thickness.
The reconstruction algorithm does not rely
on iteration and works on only 9 CBEDs at
a time, which makes it attractive for live im-
aging, potentially allowing users to see their
image develop in real time with lower doses,
high contrast, and improved resolution.
The Authors acknowledge funding from the
European Research Council (ERC) under the
European Union’s Horizon 2020 research and
innovation program (Grant Agreement No.
770887 PICOMETRICS) and funding from
the European Union’s Horizon 2020 research
and innovation program under grant agree-
ment No. 823717 ESTEEM3. J.V. and S.V.A.
acknowledge funding from the University
of Antwerp through a TOP BOF project. The
direct electron detector (Merlin, Medipix3,
Quantum Detectors) was funded by the Her-
cules fund from the Flemish Government.
This work was supported by the FWO and
FNRS within the 2Dto3D project of the EOS
program (grant number 30489208).
Affiliation
1EMAT, NANOlab Center of Excellence, Uni-
versity of Antwerp, Antwerp, Belgium
Contact
Sandra van Aert
EMAT
NANOlab Center of Excellence
University of Antwerp
Antwerp, Belgium
References:
https://bit.ly/IM-vanAert

34 • Imaging & Microscopy 2/2023

 Good Products
deserve an Award
our Wiley Analytical
Science Award.

Wiley

© Kzenon - stock.adobe.com
Analytical
Science
Award
2024 Do not think twice,
submit now!

APPLY!
Where? was-award.com
Until when? 9th June 2023

N O G
WDATE
Who? Everyone can vote for his favorite
product.
CLOSIN 3, 20 2 We are looking for the best products or
JUNE 9 solutions within the following categories:
A – Spectroscopy & Microscopy
B – Separation, Lab Automation
& Equipment

Your Contact:
Isabel Brenneisen
ibrenneise@wiley.com
+49 6201 606 716

Wiley Analytical Science https://analyticalscience.wiley.com was-award.com

Scanning Probe Microscopy

Imaging the Surface Ions of

Pristine Muscovite Mica
Using Non-Contact Atomic Force Microscopy in Ultra-High Vacuum
Giada Franceschi1, Martin Setvin1, 2, Ulrike Diebold1

A
luminosilicate minerals are ubiq-
uitous on Earth and play a crucial
role in many natural phenomena.
An accurate knowledge of their surface
structure is a prerequisite for a deeper
understanding of the underlying molec-
ular-level processes. Our team at the TU
Wien in Vienna (Austria) has succeeded in
imaging the intrinsic atomic-scale details
of muscovite mica – a common alumino-
silicate 2D mineral with a long history in
surface and interface research. Until now,
muscovite mica has only been investigated
when affected by air and in solution. Our

© Franceschi
work shows that it is now possible to mea-
sure mica (and related insulating minerals)
also in pristine, ultra-high vacuum con-
ditions through non-contact atomic force
microscopy.

Introduction

Mineral surface chemistry is at the core of

many natural phenomena, such as disso-
lution and weathering, adsorption of toxic
ions in groundwater, and ice nucleation on
mineral dust, which regulates cloud forma-
tion and weather patterns. To reach a com-
plete understanding of the molecular-level
processes occurring at mineral surfaces, an
accurate knowledge of their surface atom-
ic structures is needed. This can be ideally
achieved with the surface science approach
– combining atomically resolved investiga-
tions of well-defined single-crystalline sam-
ples in pristine, ultra-high vacuum (UHV,
pressures below 10−9 mbar) with ab initio
modeling.
However, minerals present some hur-
dles compared to the metals and (conduc-
tive) oxides most popular in surface science.
Often, they are electrically very insulat-
ing, preventing scanning tunneling micros-
copy; moreover, they develop strong sur-
face charges when cleaved in UHV, making
atomically resolved atomic force microscopy
(AFM) challenging.
These hurdles are exemplified by the sci-
entific history of muscovite mica (simply
“mica”, hereafter; Fig. 1a, b for views of its
bulk structure). Mica is a common 2D min-
eral that has been at the center of hundreds
of studies in diverse disciplines, including
nanotribology, electronics, geology, biol-
ogy, and environmental science [1]. Its per-
fect cleavage planes offer atomically flat,
micrometer-sized terraces, ideal for scan-
ning probe microscopies. To date, mica-fo-
cused and AFM-based studies are abundant
in air and solutions but there is a surprising
lack of studies in UHV. This is because the
most used cantilevers for UHV non-contact
(nc) AFM are relatively soft. They interact
strongly with the charges formed at the mica
surface after cleaving, snapping into contact
before atomic contrast is achieved [2].
In our study [3], we obtained the first
atomically resolved images of UHV-cleaved
mica using nc-AFM. An essential part of the
success laid in the choice of the AFM probe:
not a standard, Si-based cantilever, but the
qPlus sensor [4]. Its larger stiffness allows
the probe to oscillate with smaller ampli-
tudes, push through the strong electro-
static fields induced by surface charges, and
become sensitive to the short-range forces
that yield atomic contrast.
Insights into the Intrinsic Distribution of
the Surface K+ Ions of Mica
Images of UHV-cleaved mica obtained with
our setup are shown in Figures 1c and 1d
[3]. They represent the shift in resonance fre-
quency obtained while scanning the sample
at a constant height of the tip. The images
show a flat surface (as expected) decorated
by an array of dark dots (attractive regime).
In our work [3], we assigned these dots to K+
ions at the mica surface.
As seen from Figures 1a and 1b, mica
is expected to cleave within the K layers
that alternate with aluminosilicate layers
along the [001] direction. The aluminos-
ilicate layers are made of three sheets: in
the middle, octahedrally coordinated Al
ions with OH groups; at the sides, tetra-
hedrally coordinated Si and Al ions in a
3:1 ratio. The substitution of Si4+ for Al3+
introduces a missing charge in the tetrahe-
dral sheets, which is compensated by the
K+ ions. From electrostatic considerations,
one expects each cleaved surface to retain
50% of the K+ ions initially present in the
K layer. However, until now, the intrinsic –

36 • Imaging & Microscopy 2/2023


i.e., unaffected by the interaction with the
environment – distribution of these ions
had not been determined experimentally.
Early low-energy diffraction studies have
claimed a random distribution, whereas
theoretical papers typically assume fully
periodic arrangements to minimize com-
putational costs. Our images in Figs. 1c, d
show that the distribution of the K+ ions is
neither random nor fully periodic. Instead,
the ions arrange in short, alternating rows
and 120° kinks.
Insights into the Intrinsic Distribution
of the Subsurface Al3+ Ions
In addition to the ordering of the surface
K+ ions, the arrangement of the Al3+ ions
in the subsurface aluminosilicate layer of
mica has been another source of ambiguity:
As Al and Si have similar cross sections in
X-ray diffraction, it has been difficult to
ascertain the presence of short-range or-
dering in the subsurface as well. Striving to
understand better the short-range order of
mica’s surface K+ ions and to possibly tie it
to the arrangement of the subsurface Al3+
ions, we analyzed the experimental K+ dis-
tributions using density functional theory
calculations and Monte Carlo simulations
[3]. As for every surface, the arrangement
of the surface atoms is such that the surface
energy is minimized. Our analysis shows
that the structure of the surface K+ ions is
created to minimize the electrostatic repul-
sion among the surface ions, as well as the
repulsion of the surface K+ ions with the
subsurface Al3+ ions. Specifically, K+ ions
will prefer to sit closer to lower-charged
Al3+ than to Si4+ sites. This shows the im-
portant role of the subsurface Al3+ ions in
determining the surface properties of mica
– a role that is often neglected in the litera-
ture and could be a potential cause of data
misinterpretation.
Methodology
The experiments were carried out in a UHV
setup with base pressure <1 × 10−10 mbar.
Natural muscovite mica single crystals
[(0001) oriented disks of grade V1; 10 mm
diameter and 0.25 mm thickness; from Ted-
Pella] were glued on Omicron-style stainless
steel sample plates with UHV-compatible
epoxy glue (Epotek). They were cleaved in
UHV at room temperature before each ex-
periment.
The AFM measurements were performed
at 4.7 K using a commercial Omicron qPlus
head and a differential cryogenic ampli-
fier. Frequency-modulated non-contact
Fig. 1: (a) Side view of bulk mica. Cleaving occurs at the K layer, leaving half the K+ cations on
each side. (b) Top view of the surface after cleaving. (c) Atomically resolved constant-height
non-contact AFM image of mica after UHV cleaving, acquired with a qPlus sensor at 4.7 K.
Adapted from Franceschi et al. [3].
AFM mode. The tuning-fork-based AFM
sensors (k = 2,000−3,500 N/m, f0 ≈ 45 kHz,
Q ≈ 5,0000) had a separate contact for tun-
neling current attached to electrochemically
etched W tips. Before each measurement,
the tips were prepared on a clean Cu(110)
single crystal by repeated indentation and
voltage pulses. The coarse approach was
done with a setpoint of −1  Hz. The con-
troller was switched off, and the tip grad-
ually approached in constant-height mode
until a contrast was visible while scanning
in x and y. AFM images were acquired in
constant-height mode. At times, the abso-
lute values of frequency shifts during acqui-
sition were large (up to 100  Hz) and not
reproducible on different regions on the
same sample or different samples. This is
because cleaving can create domains of
trapped charges that can cause long-range
electrostatic interactions between the sur-
face and the tip. The corresponding electro-
static fields can be partially compensated
by applying a bias voltage between the tip
and the sample. Most measurements were
performed by applying a bias voltage such
that the surface was measured as close as
possible to the lowest local contact poten-
tial difference (LCPD) between the tip and
the sample. The bias voltage reported in the
presented images was applied to the back
of the sample plate while having the tip on
the ground.
Scanning Probe Microscopy
Conclusions and Outlook
Our results unveil new insights about an
oldie and favorite of surface science – mus-
covite mica – using nc-AFM with a qPlus
sensor in UHV. They show that its surface
K+ ions are arranged in short-range order
and that the arrangement of Al3+ ions in the
subsurface aluminosilicate layer influences
this order. The study paves the way to inves-
tigate other mineral surfaces (insulating and
possibly similarly charged) similarly, setting
the stage to unravel many mineral-mediat-
ed natural processes, such as ice nucleation,
CO2 sequestration, and ion dissolution.
Affiliation
1Institute of Applied Physics, TU Wien,
Vienna, Austria
2
Department of Surface and Plasma Science,
Charles University, Prague, Czech Republic
Contact
Giada Franceschi
Institute of Applied Physics
TU Wien
Vienna, Austria
franceschi@iap.tuwien.ac.at
References:
https://bit.ly/IM-Franceschi
Imaging & Microscopy 2/2023 • 37
 Image Processing
Jupyter Notebooks for Generating and
Distributing Bioimage Analysis Workflows
My Favorite Image Analysis Tool, by Neubias Members
T
o secure their reproducibility, much
effort is being placed into developing
norms for reporting bioimage anal-
ysis (BIAS) workflows. One view that has
gained many advocates is reporting BIAS
workflows containing all the steps used to
go from raw data to final numeric output
or figures [1]. In this context, workflow
documentation via scripting in a program-
ming language is critical.
A classic and popular example of such
scripting documentation is ImageJ macros.
Once the workflow has been written as an
ImageJ macro, it can be shared with others
via public repositories, such as GitHub [2]
and Zenodo [3]. More recently, a new play-
er has entered the field: Jupyter Notebooks.
While these Notebooks are not new in the
Python world and have matured over many
years in Data Science, they are now starting
to be used for BIAS. Jupyter Notebooks are
easy to install, have great documentation,
and can be used from quick prototyping to
the sharing of fully reproducible workflows.
For these reasons and other attractive fea-
tures explained below, Jupyter Notebooks
have become my favorite tool for gener-
ating and distributing bioimage analysis
workflows.
Python as a Scripting Language for
Bioimage Analysis
While the ImageJ macro language has
gained popularity, the language has sever-
al clear limitations in handling non-image
data: Manipulating multidimensional arrays
and tables, statistics, the use of data mining
algorithms on measured values, and train-
ing new deep learning models for image
processing (e.g., segmentation and denois-
ing). Due to these limitations, it has become
common practice to combine ImageJ with
other software packages, such as Python,
MATLAB, R, Excel, etc., to construct a BIAS
workflow.
To overcome the complexity caused by
using multiple software packages, there has
been a strong trend to move many BIAS
algorithms [4], workflows [5], and user inter-
faces [6] (multidimensional image-viewers)
38 • Imaging & Microscopy 2/2023
Rafael Camacho1
into Python implementations. This push
has been facilitated by the involvement
of analysts and developers from computer
science fields, such as Data Science, Com-
puter Vision, and Machine Learning, that
use Python as their Lingua Franca. Adopt-
ing Python allows a bioimage analyst to
stay within the same programming lan-
guage from raw image data processing to
final numerical output and figures.
Jupyter Notebooks for Developing and
Sharing Bioimage Analysis Workflows
Python programs can be executed in many
ways, including running a Python Script
using a terminal or using a Python Shell.
However, with Data Science and BIAS, it is
highly appreciated to have a more interac-
tive programming environment that permits
quick inspection of intermediate results of
data processing, cleansing, mining, and vi-
sualization during the construction of anal-
ysis workflows. Great examples of such an
environment are the MATLAB workspace in
the commercial sector, and Jupyter Note-
book and JupyterLab in the open-source
sector.
Jupyter Notebook in Practice
Here, I wish to demonstrate the usefulness
of the Jupyter Notebook by taking a simple
analysis workflow as an example. During the
scripting of a BIAS workflow, we use various
algorithms for image analysis that are al-
ready implemented and accessible. In Python,
such collections of implemented algorithms
are called packages or libraries. To use them
within the workflow we “import” the desired
module from those packages at the beginning
of the workflow’s code. By having the used
modules explicitly declared in the code, re-
sults can be reproduced among different users
without ambiguity. For example, we might
use the input/output module of scikit-image
[4] to load a .tif-file into a variable and use
another module matplotlib [7] to generate a
figure of the image we loaded (Fig. 1).
A Notebook consists of three types of cells:
Code cells with live code, output cells with
Rafael Camacho
works as a Scientific Officer at the Centre
for Cellular Imaging, Core Facilities, of Go-
thenburg university, Sweden, since 2018. He
did his PhD in the department of Chemical
Physics of Lund University at the Single
Molecule Spectroscopy Laboratory of Prof.
Ivan Scheblykin. In 2014, he moved to KU
Leuven as a Postdoctoral researcher in the
laboratory of Prof. Johan Hofkens for the
development of 3D super resolution imaging
and bioimage analysis software tools. He is
particularly interested in the development of
Smart Microscopy workflows that use image
analysis to continuously monitor the sam-
ple, give real-time feedback to the imaging
software, and program the microscope to
change its parameters during experimental
acquisition.
the output of a computation including fig-
ures, and markdown cells with written docu-
mentation formatted in Markdown [8]. Each
cell can be run multiple times, or sequential-
ly. It is common practice to change the input
parameters of a cell and run it “interactive-
ly” until the desired computational outcome
e.g., segmentation is obtained.
One common BIAS task is the pre-process-
ing of an image via different filters, and then
segmenting the objects of interest. For exam-
ple, we can use a Gaussian filter to attenu-
ate uneven illumination effects, and then run

Fig. 1: Using scikit-image to load a .tif-file and matplotlib to show the
image.
a classical segmentation algorithm such as
intensity thresholding by Otsu’s algorithm.
This can be implemented using the scikit-im-
age package as shown in Figure 2.
As a final step, we can measure the fea-
tures of the segmented objects and take
advantage of the rich availability of Data
Science packages in Python to get statisti-
cal summaries directly out of the measure-
ment results. For statistics, we can use Pan-
das [9], a Python library used for tabular
data manipulation and analysis (Fig. 3).
Jupyter Notebooks for Prototyping
Bioimage Analysis Workflows
Figure 4 shows an example of a simple BIAS
workflow that follows the steps described in
Figures 1-3. The Notebook includes interme-
diate steps that were not successful, so you
get a feeling for the advantage of interac-
tive analysis. This workflow is available in
the online documentation of this article as
a notebook file: Example.ipynb. If you visit
Fig. 3: Quantification and simple statistical summary.
Fig. 2: Background removal by Gaussian filter and segmentation using
Otsu’s algorithm.
the site, you will note the great support of
GitHub to render Jupyter Notebooks.
Python and Jupyter Notebooks
Installation
After demonstrating the usefulness of Jupy-
ter Notebooks, we would like to lead you to
a quick guide for installing Python on your
local machine. As the installation procedure
is not as simple as downloading an instal-
lation package and double-clicking the in-
staller, we provided an easy installation
guide for Jupyter Notebooks in the online
documentation [11].
One of the challenges when migrating
into Python for bioimage analysis is where to
start, which packages to install, and finding
good minimal examples that allow research-
ers with little Python knowledge to learn the
language by applying it directly to their own
projects. To help you get started, we provide
a list of key resources in our GitHub repos-
itory [11], including but not limited to the
Image Processing
Image.sc forum [12], scikit-image, napari,
and py-clEsperanto [13].
Jupyter Notebooks for
Distributing BIAS Workflows
Direct Sharing of Jupyter Notebooks
If you wish to share your Notebook as a
document with text and figures, and the re-
cipient does not need to run the code, then
you can simply “Save and Export Notebook
As…” a PDF file (Fig. 4). Further, if you wish
the recipient to run the code, then you can
send the *.ipynb-file. However, you will also
have to instruct the recipient on what pack-
ages to install, in such a way that they can
run the Notebook locally.
Sharing of Jupyter Notebooks
together with an environment.yml file
To make it easier for a 3rd party to run your
Jupyter Notebook it is advised to share the
“content” of your virtual environment via
an environment.yml-file. This file allows
Imaging & Microscopy 2/2023 • 39
 Image Processing
others to quickly reproduce your environ-
ment, with all the packages and versions.
Instructions to export your environment can
be found in the GitHub repository [11].
Sharing of Jupyter Notebooks
with the World
To share a BIAS workflow with the com-
munity in a more reproducible way, e.g.,
for a publication, then we recommend us-
ing GitHub. In GitHub, you can upload your
Notebook, together with the environment
file in a public (or private) repository, where
it can be used by others. This allows you
to keep an integral copy of the Notebook
and environment used for a publication.
The GitHub repository [11] associated with
this article is an example of this strategy. If
you want to have further control over the
version related to a publication, including a
DOI, then you can use Zenodo [3] in combi-
nation with the GitHub repository.
Sharing of Jupyter Notebooks by Giving
Access to Controlled Environments
If you are a bioimage analyst providing ser-
vices to a broad community (for example,
working at a core facility), then you might
need to share your workflows in an executable
environment, in terms of computer resources
and software installation. For this purpose, we
recommend 2 different solutions: a) giving
access to a BIAS server, and b) Google Colab.
a) Access to bioimage
analysis server.
Let us take as an example a simple Windows
server. As analysts, you can create a user
account. Install the appropriate virtual en-
40 • Imaging & Microscopy 2/2023
vironment and development environment,
e.g., JupyterLab. Test that the setup works
on example data. Finally, you can give ac-
cess to the user via Remote Desktop or other
alternatives such as NoMachine.
b) Sharing via Colab.
Google Colab is a free Jupyter Notebook envi-
ronment that runs in the cloud [14]. It allows
anyone to write and execute Python code in
the web browser. In a very real way, you are
borrowing computer resources from Google,
including GPUs. In this strategy, Notebooks
are stored in Google Drive or via GitHub. Then
you can share the Notebook with others and
let them run it in the cloud. Note that the vir-
tual machine instance used during develop-
ment will not be shared. Therefore, you must
include cells that install and load the libraries
needed for the Notebook to run.
Closing Remarks
One of the biggest limitations when migrating
to Python for BIAS is the learning curve asso-
ciated with using a new computer language.
However, Python and Jupyter Notebooks
have several key features that make this pro-
cess easier than with other languages. Python
is very user-friendly because its English-like
syntax is concise and easy to read. In combi-
nation with Jupyter Notebooks, the language
becomes very hands-on, allowing immediate
feedback on your progress. Because Python is
very versatile you can find libraries for many
purposes, and you can use the language to
solve small and complex tasks alike across
many domains. Therefore, I can highly rec-
Fig. 4: Full export of a BIAS workflow
documented as a Jupyter Notebook.
ommend that you look at some of the ex-
amples listed in this article and give Jupyter
Notebooks a try on your next BIAS challenge.
Acknowledgments
I would like to thank Kota Miura for the
invitation, proofreading, and suggestions
that made this article possible. I would
like to thank my colleagues at the Centre
for Cellular Imaging (CCI), Core Facilities,
The Sahlgrenska Academy, University of
Gothenburg, and the National Microscopy
Infrastructure, NMI (VR-RFI 2019-00022),
for their support during the preparation of
the manuscript. I would like to thank the
Swedish Foundation for Strategic Research
for their financial support via the Research
Infrastructure Fellows 2 grant (RIF21-0043).
Contact
Dr. Rafael Camacho
Centre for Cellular Imaging
Core Facilities
The Sahlgrenska Academy
University of Gothenburg
Gothenburg, Sweden
rafael.camacho@gu.se
References:
https://bit.ly/IM-Camacho
 Advertorial

Investigating Surface
Catalytic Activities Using SECCM
Alexander Klasen and Andrea Cerreta

F
rom converters that oxidize harmful
nitrogen oxides in vehicle exhaust to
fuel cells converting hydrogen into
clean energy or metal oxides that facili-
tate the transition of bulk chemicals into
target molecules, catalysts are employed
in almost every aspect of our modern
economy. Metal oxides are often used as
heterogeneous catalysts, fixed on a solid
surface with a large interface to either liq-
uid or gaseous reactants or products [1].
The performance of such systems depends
on several factors like the surface area,
chemical composition, or surface acidity/
basicity [2,3]. Scanning probe microsco-
py-based techniques are well established
to investigate surface parameters since
the physical interaction with a nanome-
ter-sized probe allows studying properties
such as the topography, work function, or
adhesion at high resolution [4,5].
In this article, we present measurements con-
ducted with Scanning Electrochemistry Cell
Microscopy (SECCM). First introduced by E.
Daviddi, P. R. Unwin et al. [6], SECCM uses
an electrolyte-filled pipette with a nm-sized
aperture to probe electrochemical characteris-
tics locally. The pipette position can be finely
tuned via piezo scanners. An electrolyte drop-
let is present at its aperture, which turns into a
meniscus when the pipette is close enough to
the sample surface. A small electrode within
the pipette, usually an AgCl-coated Ag wire,
allows applying a potential between the pi-
pette and the sample. The meniscus then rep-
resents a spatially confined electrochemical
cell in which reactions can occur, e.g. at con-
stant bias or current (Fig. 1). A series of elec-
trochemical cycles can be done by lifting and
approaching the pipette at different points on
a point grid. (Fig. 2).
TiO2 has gained much attention as a pho-
tocatalyst [7,8] since oxygen vacancies in its
lattice can play a twofold role. Oxygen vacan-
cies in the TiO2 bulk create energy states close
to the conduction band of the semiconduc-
tor and hence act as n-doping centers. How-
ever, surface defects create energy levels deep
within the band gap of a TiO2 crystal or layer
and can act as charge carrier traps that form
local catalytic centers [9-11].
For this study, a piece of Ti metal was
thermally treated to form a thin TiO2 surface
layer. A standard phosphate buffer contain-
ing 137 mM NaCl, 2.7 mM KCl and 10 mM
Na2HPO3 was used as electrolyte. On the TiO2
surface, we could identify locations with
strong catalytic properties visualized by bright
colors indicating large current flow (Fig. 2A).
In one exemplary IV curve, two distinct bumps
could be seen that mark the beginning of the
reduction of phosphate ions to phosphite ions
at -0.8 V and the start of water cleaving and
subsequent hydrogen formation at approx.
-1.3  V. Please note, that TiO2 is an n-type
semiconductor with a large bandgap that
forms a rectifying Schottky diode with the Ti
metal, resulting in unidirectional current flow.
Similar experiments conducted with the same
setup within one hour on a far more conduc-
tive HOPG sample showed an overall smaller
current flow since HOPG does not possess sur-
face catalytic properties. The observed current
flow was attributed to mere ion migration that
has a higher magnitude at positive applied
bias, probably due to higher ion mobility of
Na+, K+, and H+ ions compared to phosphate
ions, which possess a larger size and higher
charge density.
These two examples show that SECCM
is a unique and easy-to-use tool to investi-
gate the catalytic properties of surfaces on
a local scale.
Contact
Park Systems Europe GmbH
Mannheim, Germany
References:
https://bit.ly/IM-PS0223

Fig.1: A: An electrolyte-filled micropipette approaches the target (A) and
forms a small meniscus with the surface in close proximity: the electro-
chemical cell (B and D). An applied bias between the pipette electrode
and the sample lead to a spike in current, once the meniscus is formed.
(C). The approach is stopped and EC can be conducted by applying vari-
ous bias or current settings to the sample and pipette electrode.
Fig. 2: SECCM measurements conducted on a 10x10 grid on A: Ti met-
al with a thin TiO2 top layer and B: on a freshly cleaved HOPG sur-
face. Each pixel represents a single EC cell where bias was swept be-
tween -1.5 and +1.5 V. Displayed are the absolute highest currents of
the respective IV curves. Exemplary IV curves of each sample are dis-
played on the right side.

Imaging & Microscopy 2/2023 • 41

Products
Products
Correlative AFM / SEM Microscope
After a decade of research into cor-
relative microscopy, Quantum De-
sign has launched the FusionScope
that combines the power of AFM
with the benefits of SEM imag-
ing. Designed from the ground up
to seamlessly integrate these two
powerful techniques, it takes ad-
Microscope Software Upgrade
Raman microscopy specialist WITec
has introduced Suite SIX, the next
stage in software for Raman and
correlative measurement control,
data acquisition and post-process-
ing. Developed to enhance the al-
pha300 microscope series, it offers
multiple user profiles, regulatory
compliance support and advanced
sample navigation. Multi-user
Plug and Play Modulated CW Lasers
The Cobolt 06-01 series
includes lasers over a
large wavelength range
of 375 – 1064  nm in
a compact plug-and-
play format. The high
performance, fixed wave-
length diode laser modules and
diode-pumped lasers of up to
400 mW have a direct digital and
analog modulation capability, and
their ‘true off’ (>70  dB) during
modulation makes them ideal for
42 • Imaging & Microscopy 2/2023
vantage of an innovative shared co-
ordinate system that automatically
aligns both AFM and SEM opera-
tions for measurements and sample
positioning. This integrated shared
mapping allows you to easily iden-
tify an area of interest, measure a
sample, and combine the imaging
data in real time. It allows smooth
image transitions between milli-
meter, micron, and sub-nanometer
scales. Researchers can interactively
overlay AFM imaging data onto
SEM images while operating the
microscope in real time, creating
stunning 2D and 3D visualizations
with nanoscale resolution.
Quantum Design
www.fusionscope.com
management capability enables
the site administrator to determine
the instrument functionalities and
data access rights associated with
each system login. It also assists
with compliance to good practice
(GxP) guidelines such as the US
Food and Drug Administration’s
21 CFR Part 11. The software also
allows users to develop a library of
distinct hardware setups to further
streamline recurring experiments.
TrueOrigin, Suite SIX’s portable co-
ordinate system, locates measure-
ment areas in relation to markers
on a sample holder or features on
the sample itself.
WITec
www.oxinst.com
applications in bioimaging
and quantum technol-
ogy. They have an
integrated clean-up
filter on all diode
wavelengths. Fiber
pigtailed configu-
rations are available,
and they are compatible
with the C-Flex Laser Combiner.
Hübner Photonics
www.hubner-photonics.com
Non-Destructive 3D X-Ray Imaging
The versatile Zeiss Xradia Versa 3D
X-ray microscopes (XRM) provide
superior 3D image quality and data
for a wide range of materials and
working environments. They feature
dual-stage magnification based
on synchrotron-caliber optics and
‘Resolution at a Distance’ technol-
ogy for high resolution even at large
working distances. Non-destructive
imaging preserves and extends the
use of valuable samples, enabling
4D and in situ studies. The system
achieves high resolution perfor-
mance of 450 - 500 nm across the
full range of energy, from 30 kV to
160  kV, while NavX guides users
through automated workflows with
intelligent system insights. The dif-
fraction contrast tomography en-
All-In-One Solution for Widefield
and Spinning Disk Confocal Microscopy
Crest Optics has launched
its new spinning disk
confocal microscope
system called Cicero.
Focusing on the essen-
tial elements of epi-flu-
orescence and confocal
microscopy, it has been
developed to make high-end im-
aging accessible to the wider sci-
entific community by providing
a compact, cost-effective and
powerful imaging solution. The
spinning disk confocal instrument
can be easily integrated into ex-
isting workflows with maximum
configuration flexibility, enabling
a broad range of imaging appli-
Coolstage for Imaging of Biological Materials
Preparing biological samples for
imaging usually requires utilizing
a range of physical and chemical
processes, including dehydration
and fixing. These processes can
alter the structure of the sample,
and are often complex and time
consuming. The Coolstage from
Coxem simplifies this process by
rapidly lowering the temperature
of the sample stage, freezing the
sample without damaging its mi-
crostructure. Available on both the
EM30 Series tabletop SEM as well
as the CX-200Plus Compact SEM,
ables non-destructive mapping of
grain orientation and microstruc-
ture in 3D. Direct visualization of
3D crystallographic grain orienta-
tion enables the characterization of
polycrystalline materials like metal
alloys, geomaterials, ceramics, or
pharmaceuticals.
Carl Zeiss
www.zeiss.com
cations across basic
and applied research.
The compact sys-
tem enables users
to simply switch
between widefield
and confocal imaging
modes as their imaging
requirements evolve. The
speed and light efficiency of the
product provides the capability
for prolonged live imaging and
the capture of fast cellular events,
such as chromosome segregations
and organelle trafficking.
CrestOptics
www.crestoptics.com
the Coolstage has a temperature
range of -25°C to 50°C. The tech-
nique can be used on a variety of
biological samples, including food,
biological samples, and most any
natural sample containing mois-
ture. In addition, since the Cool-
stage is also able to heat a sample
up to 50°C, temperature-induced
changes in structure can be ob-
served by slowly heating the sam-
ple while imaging.
Coxem
www.coxem.com
 Index / Imprint

AHF Analysentechnik  7 Hübner Photonics 42 Quantum Design 42

Centro Nacional de Biotecnologia (CNB-CSIC) 11 Inara Aguiar 3 Royal Microscopical Society (RMS) 10

Charles University 36 Scandem Nordic Microscopy Society 8

MCO Congres Marseille 11
Coxem  9, 42 Technische Universität Wien 14, 26, 36
MRC Harwell Institute 17
CrestOptics 42 Tescan 29
NKT Photonics  Outside Back Cover
Edmund Optics 15 University of Antwerp 30, 32
Park Systems Europe Inside Front Cover, 41
EMS Secretary 11 University of Gothenburg 38

European Molecular Biology PreciPoint 20 WITec Cover, 12, 42

Laboratory (EMBL) Heidelberg 8 Prospective Instruments 23 Carl Zeiss 42

Imprint Director – Global Sales,

Sciences, Corporate Sales
Circulation
18,000 copies, 4 times per annum
Dan Nicholas Advertising price list from October 1st 2022
Published by
Tel.: +1 716 587 2181
Wiley-VCH GmbH
A Company of John Wiley & Sons, Inc dnicholas@wiley.com Subscription 2023
Four issues 40.00 €, single copy 15.30 € ca. plus postage.
Managing Directors Segment Sales Manager Pupils and students receive a discount of 50 % at sight of a
Sabine Haag, Dr. Guido F. Herrmann Laboratoty & Biotechnology valid certificate. Subscription orders can be revoked within
Vanessa Winde 1 week in writing. Dispatch complaints are possible only within
Publishing Director, Chem & Lab
Ph.: +49 6201 606-721 4 weeks after publishing date. Subscription cancellations are
Dr. Heiko Baumgartner
E-mail: vwinde@wiley.com accepted 6 weeks before end of year.
Regional Commercial Director
Media Consultant
Dr. Katja Habermüller Bank account
Dr. Stefanie Krauth
Ph.: +49 6201 606-728 J.P. Morgan AG, Frankfurt
Editor-in-chief Konto-Nr. 6161517443
E-mail: stefanie.krauth@wiley.com
Dr. Birgit Foltas BLZ: 501 108 00
Ph.: +49 6201 606-760 BIC: CHAS DE FX
E-mail: bfoltas@wiley.com Layout & Lithography IBAN: DE55501108006161517443
Ramona Scheirich
Specially identified contributions are the responsibility of the
Editorial Team E-mail: ramona.scheirich@wiley.com
author. Manuscripts should be addressed to the editorial office.
Dr. Cecilia Kruszynski de Assis We assume no liability for unsolicited, submitted manuscripts.
Ph.: +49 (0) 3047031 105 Reproduction, including excerpts, is permitted only with the
Production Manager
E-mail: ckruszynski@wiley.com permission of the editorial office and with citation of the source.
Jörg Stenger
Róisín Murtagh Ph.: +49 6201 606-742 The publishing house is granted the exclusive right, with regard
Ph.: +49 (0) 6201 606 042 E-mail: joerg.stenger@wiley.com to space, time and content to use the works/ editorial contribu-
E-mail: rmurtagh@wiley.com tions in unchanged or edited form for any and all purposes any
Sales Administrator number of times itself, or to transfer the rights for the use of
Dr. Martin Graf-Utzmann other organizations in which it holds partnership interests, as
Kerstin Kunkel
Ph.: +49 6201 606-766 well as to third parties. This right of use relates to print as well
Ph.: +49 6201 606-731 · Fax: -790
E-mail: martin.graf-utzmann@wiley.com as electronic media, including the Internet, as well as data-
E-mail: kerstin.kunkel@wiley.com bases / data carriers of any kind.
Dr. Christina Poggel
Ph.: +49 6201 606-625 Reprints All names, designations or signs in this issue, whether referred
E-mail: christina.poggel@wiley.com Dr. Stefanie Krauth to and/or shown, could be trade names of the respective owner.
Ph.: +49 6201 606-728 Printed in Germany
Isabel Brenneisen E-mail: stefanie.krauth@wiley.com
Ph.: +49 6201 606-716
E-mail: isabel.brenneisen@wiley.com Advisory Board
Prof. A. Diaspro, Italian Institute of Technology ISSN 1439-4243
Corinna Herbst
Dr. C. Durkan, Univ. of Cambridge, UK www.imaging-git.com
Ph.: +49 6201 606-752
Dr. M. Dürrenberger, Univ. of Basel, Switzerland
E-mail: corinna.herbst@wiley.com Dr. R. Fleck, King’s Collage London, UK
Prof. M. Gu, Swinburne Univ., Australia
Freelancer Prof. B. Hecht, Univ. of Wuerzburg, Germany
Dr. Martin Friedrich Prof. M. Hegner, Trinity College Dublin, Ireland
Ph.: +49 178 4596064 Prof. F.-J. Kao, National Yang-Ming University, Taipei, Taiwan
E-mail: editor@redaktionsbuero-friedrich.de Prof. N. Kruse, Washington State University, WA, USA
Prof. D. Nicastro, Thermo Fisher, The Netherlands
Wiley GIT Reader Service Dr. J. Rietdorf, Centre for Technological Development
in Healthcare (CDTS), Brasil
65341 Eltville / Germany
Dr. D. Spitzer, ISL, France
Phone: +49 6123 9238 246
Prof. G. A. Stanciu, Univ. of Bucharest, Romania
Fax: +49 6123 9238 244
E-Mail: WileyGIT@vuservice.de Prof. G. Valdré, Univ. of Bologna, Italy
Our service is available for you from Dr. Roger Wepf, EMBL, Heidelberg, Germany
Monday to Friday 8 am – 5 pm CET Dr. T. Zimmermann, ORG, Barcelona, Spain

Imaging & Microscopy 2/2023 • 43

400 450 500 550 600 650 700 750 800 850 900 950 1000

Confocal microscope image of a mouse. Credits: Leica Microsystems

SuperK CHROMATUNE
400-1000 nm

New!

World’s
broadest
tunable
laser
Use your SuperK CHROMATUNE for bioimaging,
fluorescence imaging & spectroscopy, single
molecule imaging and other microscopy applica-
tions, at any wavelength between 400-1000 nm.

Excite at any wavelength in the 400 to 1000 nm

range. Fiber delivered and diffraction limited. Zero
maintenance with a lifetime of thousands of hours
and two years warranty.

nktphotonics.com