Physical Review Materials,
Год журнала:
2024,
Номер
8(10)
Опубликована: Окт. 2, 2024
The
large
amount
of
powder
diffraction
data
for
which
the
corresponding
crystal
structures
have
not
yet
been
identified
suggests
existence
numerous
undiscovered,
physically
relevant
structure
prototypes.
In
this
paper,
we
present
a
scheme
to
resolve
into
with
precise
atomic
coordinates
by
screening
space
all
possible
arrangements,
i.e.,
structural
prototypes,
including
those
previously
observed,
using
pre-trained
machine
learning
(ML)
model.
This
involves
(i)
enumerating
symmetry-confined
ways
in
given
composition
can
be
accommodated
group,
(ii)
ranking
element-assigned
prototype
representations
energies
predicted
Wyckoff
representation
regression
ML
model
[Goodall
,
],
(iii)
assigning
and
perturbing
atoms
along
degree
freedom
allowed
positions
match
experimental
data,
(iv)
validating
thermodynamic
stability
material
density-functional
theory.
An
advantage
presented
method
is
that
it
does
rely
on
database
observed
prototypes
is,
therefore
capable
finding
entirely
new
symmetric
arrangements
atoms.
We
demonstrate
workflow
unidentified
x-ray
spectra
from
ICDD
identify
number
stable
structures,
where
majority
turns
out
derivable
known
However,
at
least
two
are
found
part
our
prior
sets.
Published
American
Physical
Society
2024
Ultrasonics Sonochemistry,
Год журнала:
2024,
Номер
110, С. 107030 - 107030
Опубликована: Авг. 15, 2024
Environmental
concerns
linked
to
animal-based
protein
production
have
intensified
interest
in
sustainable
alternatives,
with
a
focus
on
underutilized
plant
proteins.
Faba
beans,
primarily
used
for
animal
feed,
offer
high-quality
source
promising
bioactive
compounds
food
applications.
This
study
explores
the
efficacy
of
ultrasound-assisted
extraction
under
optimal
conditions
(123
W
power,
1:15
g/mL
solute/solvent
ratio,
41
min
sonication,
623
mL
total
volume)
isolate
faba
bean
(U-FBPI).
The
method
achieved
yield
19.75
%
and
content
92.87
%,
outperforming
control
method's
16.41
89.88
%.
Electrophoretic
analysis
confirmed
no
significant
changes
primary
structure
U-FBPI
compared
control.
However,
Fourier-transform
infrared
spectroscopy
revealed
modifications
secondary
due
ultrasound
treatment.
demonstrated
superior
water
oil
holding
capacities
isolate,
although
its
foaming
capacity
was
reduced
by
ultrasound.
Thermal
indicated
minimal
impact
protein's
thermal
properties
applied
conditions.
research
highlights
potential
improving
functional
isolates,
presenting
viable
approach
advancing
plant-based
contributing
consumption.
Journal of the American Chemical Society,
Год журнала:
2024,
Номер
146(12), С. 8098 - 8109
Опубликована: Март 13, 2024
Determining
the
structures
of
previously
unseen
compounds
from
experimental
characterizations
is
a
crucial
part
materials
science.
It
requires
step
searching
for
structure
type
that
conforms
to
lattice
unknown
compound,
which
enables
pattern
matching
process
characterization
data,
such
as
X-ray
diffraction
(XRD)
patterns.
However,
this
procedure
typically
places
high
demand
on
domain
expertise,
thus
creating
an
obstacle
computer-driven
automation.
Here,
we
address
challenge
by
leveraging
deep-learning
model
composed
union
convolutional
residual
neural
networks.
The
accuracy
demonstrated
dataset
over
60,000
different
100
types,
and
additional
categories
can
be
integrated
without
need
retrain
existing
We
also
unravel
operation
black
box
highlight
way
in
resemblance
between
compound
quantified
based
both
local
global
characteristics
XRD
This
computational
tool
opens
new
avenues
automating
analysis
unearthed
high-throughput
experimentation.
Tenside Surfactants Detergents,
Год журнала:
2024,
Номер
61(4), С. 285 - 296
Опубликована: Апрель 29, 2024
Abstract
This
review
critically
analyzes
the
incorporation
of
artificial
intelligence
(AI)
in
surface
chemistry
and
catalysis
to
emphasize
revolutionary
impact
AI
techniques
this
field.
The
current
examines
various
studies
that
using
techniques,
including
machine
learning
(ML),
deep
(DL),
neural
networks
(NNs),
catalysis.
It
reviews
literature
on
application
models
predicting
adsorption
behaviours,
analyzing
spectroscopic
data,
improving
catalyst
screening
processes.
combines
both
theoretical
empirical
provide
a
comprehensive
synthesis
findings.
demonstrates
applications
have
made
remarkable
progress
properties
nanostructured
catalysts,
discovering
new
materials
for
energy
conversion,
developing
efficient
bimetallic
catalysts
CO
2
reduction.
AI-based
analyses,
particularly
advanced
NNs,
provided
significant
insights
into
mechanisms
dynamics
catalytic
reactions.
will
be
shown
plays
crucial
role
by
significantly
accelerating
discovery
enhancing
process
optimization,
resulting
enhanced
efficiency
selectivity.
mini-review
highlights
challenges
data
quality,
model
interpretability,
scalability,
ethical,
environmental
concerns
AI-driven
research.
importance
continued
methodological
advancements
responsible
implementation
ACS Energy Letters,
Год журнала:
2024,
Номер
9(12), С. 6178 - 6214
Опубликована: Дек. 4, 2024
Batteries
based
on
sulfur
cathodes
offer
a
promising
energy
storage
solution
due
to
their
potential
for
high
performance,
cost-effectiveness,
and
sustainability.
However,
commercial
viability
is
challenged
by
issues
such
as
polysulfide
migration,
volume
changes,
uneven
phase
nucleation,
limited
ion
transport,
sluggish
redox
kinetics.
Addressing
these
challenges
requires
insights
into
the
structural,
morphological,
chemical
evolution
of
phases,
associated
changes
internal
stresses,
diffusion
within
battery.
Such
can
only
be
obtained
through
real-time
reaction
monitoring
battery's
operational
environment,
supported
molecular
dynamics
simulations
advanced
artificial
intelligence-driven
data
analysis.
This
review
provides
an
overview
