diff --git a/joss.06028/10.21105.joss.06028.crossref.xml b/joss.06028/10.21105.joss.06028.crossref.xml
new file mode 100644
index 0000000000..d7d6b74b77
--- /dev/null
+++ b/joss.06028/10.21105.joss.06028.crossref.xml
@@ -0,0 +1,185 @@
+
+
+
+ 20240715181014-fd19125d6725cbfdec1aff836fcee30611249669
+ 20240715181014
+
+ JOSS Admin
+ admin@theoj.org
+
+ The Open Journal
+
+
+
+
+ Journal of Open Source Software
+ JOSS
+ 2475-9066
+
+ 10.21105/joss
+ https://joss.theoj.org
+
+
+
+
+ 07
+ 2024
+
+
+ 9
+
+ 99
+
+
+
+ Scanbot: An STM Automation Bot
+
+
+
+ Julian
+ Ceddia
+ https://orcid.org/0000-0003-3990-8852
+
+
+ Jack
+ Hellerstedt
+ https://orcid.org/0000-0003-2282-8223
+
+
+ Benjamin
+ Lowe
+ https://orcid.org/0000-0002-5157-7737
+
+
+ Agustin
+ Schiffrin
+ https://orcid.org/0000-0003-1140-8485
+
+
+
+ 07
+ 15
+ 2024
+
+
+ 6028
+
+
+ 10.21105/joss.06028
+
+
+ http://creativecommons.org/licenses/by/4.0/
+ http://creativecommons.org/licenses/by/4.0/
+ http://creativecommons.org/licenses/by/4.0/
+
+
+
+ Software archive
+ 10.5281/zenodo.12669343
+
+
+ GitHub review issue
+ https://github.com/openjournals/joss-reviews/issues/6028
+
+
+
+ 10.21105/joss.06028
+ https://joss.theoj.org/papers/10.21105/joss.06028
+
+
+ https://joss.theoj.org/papers/10.21105/joss.06028.pdf
+
+
+
+
+
+ Embedding human heuristics in
+machine-learning-enabled probe microscopy
+ Gordon
+ Machine Learning: Science and
+Technology
+ 1
+ 1
+ 10.1088/2632-2153/ab42ec
+ 2020
+ Gordon, O. M., Junqueira, F. L. Q.,
+& Moriarty, P. J. (2020). Embedding human heuristics in
+machine-learning-enabled probe microscopy. Machine Learning: Science and
+Technology, 1(1), 015001.
+https://doi.org/10.1088/2632-2153/ab42ec
+
+
+ Automated tip conditioning for scanning
+tunneling spectroscopy
+ Wang
+ The Journal of Physical Chemistry
+A
+ 6
+ 125
+ 10.1021/acs.jpca.0c10731
+ 2021
+ Wang, S., Zhu, J., Blackwell, R.,
+& Fischer, F. R. (2021). Automated tip conditioning for scanning
+tunneling spectroscopy. The Journal of Physical Chemistry A, 125(6),
+1384–1390.
+https://doi.org/10.1021/acs.jpca.0c10731
+
+
+ Artificial-intelligence-driven scanning probe
+microscopy
+ Krull
+ Communications Physics
+ 1
+ 3
+ 10.1038/s42005-020-0317-3
+ 2020
+ Krull, A., Hirsch, P., Rother, C.,
+& Schiffrin, A. (2020). Artificial-intelligence-driven scanning
+probe microscopy. Communications Physics, 3(1).
+https://doi.org/10.1038/s42005-020-0317-3
+
+
+ New-horizons-SPM/nanonisTCP: nanonisTCP
+v1.0.0
+ Ceddia
+ 10.5281/zenodo.7402665
+ 2022
+ Ceddia, J., jhellerstedt, &
+benlowe1. (2022). New-horizons-SPM/nanonisTCP: nanonisTCP v1.0.0
+(Version v1.0.0). Zenodo.
+https://doi.org/10.5281/zenodo.7402665
+
+
+ Autonomous scanning probe microscopy in situ
+tip conditioning through machine learning
+ Rashidi
+ ACS Nano
+ 6
+ 12
+ 10.1021/acsnano.8b02208
+ 2018
+ Rashidi, M., & Wolkow, R. A.
+(2018). Autonomous scanning probe microscopy in situ tip conditioning
+through machine learning. ACS Nano, 12(6), 5185–5189.
+https://doi.org/10.1021/acsnano.8b02208
+
+
+ Mimea nanonis
+ Specs-GmBH
+ 2015
+ Specs-GmBH. (2015). Mimea nanonis.
+https://www.specs-group.com/nanonis/products/mimea/
+
+
+
+
+
+
diff --git a/joss.06028/10.21105.joss.06028.pdf b/joss.06028/10.21105.joss.06028.pdf
new file mode 100644
index 0000000000..4f3c916cec
Binary files /dev/null and b/joss.06028/10.21105.joss.06028.pdf differ
diff --git a/joss.06028/paper.jats/10.21105.joss.06028.jats b/joss.06028/paper.jats/10.21105.joss.06028.jats
new file mode 100644
index 0000000000..89492d76af
--- /dev/null
+++ b/joss.06028/paper.jats/10.21105.joss.06028.jats
@@ -0,0 +1,345 @@
+
+
+
+
+
+
+
+Journal of Open Source Software
+JOSS
+
+2475-9066
+
+Open Journals
+
+
+
+6028
+10.21105/joss.06028
+
+Scanbot: An STM Automation Bot
+
+
+
+https://orcid.org/0000-0003-3990-8852
+
+Ceddia
+Julian
+
+
+
+*
+
+
+https://orcid.org/0000-0003-2282-8223
+
+Hellerstedt
+Jack
+
+
+
+
+
+https://orcid.org/0000-0002-5157-7737
+
+Lowe
+Benjamin
+
+
+
+
+
+https://orcid.org/0000-0003-1140-8485
+
+Schiffrin
+Agustin
+
+
+
+
+
+
+School of Physics & Astronomy, Monash University,
+Clayton, Victoria 3800, Australia
+
+
+
+
+ARC Centre of Excellence in Future Low-Energy Electronics
+Technologies, Monash University, Clayton, Victoria 3800,
+Australia
+
+
+
+
+* E-mail:
+
+
+13
+3
+2023
+
+9
+99
+6028
+
+Authors of papers retain copyright and release the
+work under a Creative Commons Attribution 4.0 International License (CC
+BY 4.0)
+2022
+The article authors
+
+Authors of papers retain copyright and release the work under
+a Creative Commons Attribution 4.0 International License (CC BY
+4.0)
+
+
+
+Python
+Scanning Tunneling Microscopy
+STM
+Automation
+
+
+
+
+
+ Summary
+
Scanning Tunnelling Microscopes (STM) are capable of capturing
+ images of surfaces with atomic-scale resolution. This is achieved by
+ scanning an atomically sharp probe across the surface of the sample
+ while monitoring an electric current. However, the quality of STM data
+ relies heavily on the atomic-scale geometry and composition of the
+ scanning probe apex, as well as the roughness and cleanliness of the
+ scanned region. For instance, blunt tips result in blurry images while
+ contaminated tips can lead to noisy images due to interactions with
+ the sample. As a result, optimal STM data acquisition commonly
+ requires time-consuming tasks such as probe conditioning—i.e.,
+ sharpening via “tip-shaping”, where the apex of the probe can be
+ refined by poking it into a clean metal surface—and identification of
+ areas of interest of the sample. Moreover, the quality of the probe
+ can vary during a scan, especially when scanning over debris or
+ excessively rough areas, necessitating additional tip-shaping.
+
Here, we present Scanbot, a program that fully automates common STM
+ data acquisition techniques, as well as tip-shaping and sample
+ surveying. Scanbot relies on a dual sample holder (DSH;
+ [fig:1]), where a sample of
+ interest is mounted alongside a clean reference metal surface, which
+ is ideal for tip preparation. Scanbot is able to analyse STM images
+ and identify when the probe requires conditioning, subsequently moving
+ it from the sample of interest to the clean reference metal, where it
+ will prepare a scanning probe capable of obtaining high-quality STM
+ images. This is accomplished using built-in piezoceramic scanners to
+ maneuver the STM tip while tracking its position through a camera
+ feed; [fig:1]b). Once
+ Scanbot determines that the probe has been conditioned adequately, it
+ moves the tip back to the sample of interest and STM data acquisition
+ resumes.
+
+
Tracking and maneuvering the STM probe above the dual
+ sample holder (DSH). a) Schematic of the STM tip over
+ the dual sample holder setup. A sample of interest is mounted next
+ to a clean reference metal substrate (e.g. Au(111)) which is ideal
+ for tip shaping. b) Image from the camera feed used by
+ Scanbot to track and maneuver the STM probe automatically from the
+ sample to the clean reference metal, where it can be refined. The
+ red (green) marker indicates the probe apex position (target
+ position, respectively). See Scanbot
+ documentation
+ for a video example.
+
+
+
+
[fig:2] demonstrates
+ Scanbot’s ability to recondition a ‘bad’ tip on a clean reference
+ metal surface. Scanbot can gently impinge the scanning probe apex onto
+ a clean, flat region of the metal surface, which results in an imprint
+ associated with the geometry of the tip. This imprint can then be
+ scanned, and the resulting image is similar to the auto-correlation
+ function of the tip’s apex. The quality of the tip can be assessed by
+ measuring the area and circularity of the imprint. If the imprint does
+ not meet the desired criteria, a more aggressive tip shaping action is
+ carried out, and the process is repeated until a high-quality tip is
+ achieved.
+
+
Successive STM images (left to right) of the tip’s
+ imprint on a clean metal surface, each following a more agressive
+ tip-shaping action in a different location. The area and circularity
+ of each imprint reflects the geometry of the apex of the scanning
+ probe. Thus the process is repeated until a desired geometry is
+ achieved.
+
+
+
+
+ Statement of need
+
To reduce the time-intensive nature of STM experiments, various
+ innovative solutions have been implemented to automate specific tasks.
+ For instance, Wang et al. created a Python package that automates
+ probe conditioning for Scanning Tunneling Spectroscopy
+ (Wang
+ et al., 2021). However, this package still requires manual
+ preparation of the tip such that it can acquire clean images. Some
+ researchers have employed the use of machine learning algorithms to
+ analyse acquired images and determine when a probe needs refining
+ (Gordon
+ et al., 2020;
+ Rashidi
+ & Wolkow, 2018), then Reinforcement Learning (RL) agents
+ can condition the probe accordingly
+ (Krull
+ et al., 2020). Although these approaches have significantly
+ advanced automation in STM experiments, they are often tailored to
+ specific surfaces and STM equipment, making it challenging to transfer
+ them directly to other labs studying different kinds of samples or
+ working with different STM systems.
+
To overcome these limitations, we have developed Scanbot, a Python
+ robot that is compatible with a broader range of STMs, specifically
+ those compatible with the Nanonis V5 software
+ (Ceddia
+ et al., 2022;
+ Specs-GmBH,
+ 2015). Additionally, our package incorporates Scanbot’s
+ distinctive approach to tip shaping, which involves monitoring the
+ tip’s motion above a dual sample holder. This method is particularly
+ beneficial in experiments where the sample’s properties might make it
+ challenging to achieve a high-quality scanning probe without needing
+ to manually switch out the sample for a clean metal on which the tip
+ can be prepared.
+
Scanbot has been developed in a modular fashion, which means its
+ functionality can easily be expanded or improved through contributions
+ from the open-source community. Furthermore, through the use of
+ hooks,
+ users can customise or replace key funcionalities that are system- or
+ lab-specific, without rewriting Scanbot’s source code. This has the
+ advantage of being able to update Scanbot to the latest version
+ without losing customised code. Such hooks can also be used to improve
+ Scanbot’s existing functionality or test potential new features. For
+ instance, Scanbot’s algorithmic approach to automated tip shaping
+ might benefit the integration of an RL agent. This could be achieved
+ by leveraging the hook
+ hk_tipShape,
+ where important parameters related to tip shaping can be adjusted
+ based on images of the tip’s imprint. Complete documentation for
+ Scanbot, including how such hooks can be leveraged, can be found at
+ https://new-horizons-spm.github.io/scanbot.
+
+
+ Acknowledgements
+
A.S. acknowledges funding support from the ARC Future Fellowship
+ scheme (FT150100426). J.C., B.L., and J.H. acknowledge funding support
+ from the Australian Research Council (ARC) Centre of Excellence in
+ Future Low-Energy Electronics Technologies (CE170100039). J.C., and
+ B.L. are supported through an Australian Government Research Training
+ Program (RTP) Scholarship.