3930 XXXI International Mineral Processing Congress 2024 Proceedings/Washington, DC/Sep 29–Oct 3
and internal microstructure – often referred to as ran-
dom breakage or fracture. Liberation by detachment, on
the other hand, assumes full dependence, where breakage
occurs along grain boundaries loosely bonded in the matrix.
Despite the complexity, particle breakage mechanisms typi-
cally exhibit a superposition of both borderline cases. King
and Schneider (1998) introduced five additional condi-
tions for liberation by detachment that rule out random
breakage. In the comminution of natural minerals, random
breakage is the common assumption, therefore, additional
comminution steps downstream to ensure sufficient libera-
tion of the target phase are needed.
Typically, the assessment of the type of breakage
involves single particle breakage studies (Rumpf 2004,
Tavares 2007) and a method to image the broken state
(Maire et al. 2005, Ramandi et al. 2016). The former is
either done by single, double impact or slow compres-
sion testing. The fragments will then be photographed
and examined. Another way is to embed the fragments in
epoxy, grind and polish in a manner in which the polished
section is orthogonal to the breakage surface. While this
marks the well accepted and established method in various
studies (Popov et al. 2020), the stereological bias can still
not be avoided with irregularly or complex microstructures.
This study uses a three-dimensional approach to get insight
into the broken particle. The focus lays on breakages that
resulted from a press used for testing slag particles that can
be coupled with XCT as the imaging method.
The coupling of XCT is especially useful for analyz-
ing deformations and changes in the sample’s three-dimen-
sional structure under different loads or at different time
intervals (Van Stappen et al. 2019, Zhang et al. 2020). The
sample can remain in the comminution space and con-
tinue to be loaded until it fractures without external influ-
ence. Additionally, newly formed cracks and fractures can
be studied based on the internal, complex microstructure
of multiphase materials. This information is important,
particularly when different regions of the sample respond
differently to compression. This allows for establishing
a correlation with the breakage mechanism. The datasets
from microstructural or in-situ measurements can also be
used for breakage simulations to validate their accuracy.
MATERIAL AND METHODOLOGY
Materials and Microstructural Description of the Used
Lithium Aluminate Slag
The study utilized lithium aluminate slag produced by IME
RWTH Aachen. The production process involves homoge-
nization melting of the feed salts and a specifically designed
cooling strategy with a cooling rate of 50 K/h. The resulting
slag is composed of four minerals and an amorphous phase.
The lithium aluminate is the target phase or EnAM, which
accounts for up to 16.2 mass percent, is needed to be liber-
ated from the matrix phase.
The target phase is generally distributed throughout
the entire slag sample. In 2D, dendrites appear linear, but
in 3D, they have a layered structure with a wavy texture.
The dendrites have low connectivity, with only a thin main
branch connecting the layers. Upon closer examination of
the thin sections (Võ et al. 2024), it is evident that the den-
dritic branches consist of numerous small crystalline units,
consistently surrounded by the matrix phase. This state-
ment suggests a strong connection between the interlock-
ing and matrix phases.
The sample volume can exhibit large and fine pores,
whose dimensions and arrangement are related to crystal
growth, ion concentration in the molten state, and external
parameters such as temperature and pressure. Crystals that
grow rapidly may not achieve their final structural arrange-
ment within the cooling time, resulting in defects and voids
within the crystal lattice. These voids act as pores and are
randomly dispersed throughout the resulting slag product.
A closer and more detailed description of this particular
slag system can be found in the work of Võ et al. (2024).
Sample Preparation
To estimate the breakage force, irregularly shaped slag parti-
cles are crushed using a jaw crusher and sieving. The selected
fraction consists of slag particles small enough to fit within
the crushing space of the load cell. Specifically, slag particles
within the 4–5 mm fraction are chosen to estimate break-
age force for in-situ measurements and to adjust measure-
ment settings on the X-ray computed microscope (XCT).
A red coating on the outer surface helps with distinguish-
ing the original outer side from the newly formed fracture
surface after the successful compression and enables a clear
differentiation between the two surfaces after breakage.
Ex- Situ Single Particle Breakage
For single particle breakage studies, the Deben CT5000
5kN in-situ load cell tensile stage for XCT applications is
utilized. The crushing space is defined by the ceramic pis-
ton setup, consisting of upper and lower cylindrical ceramic
stamps and the guiding tube. Since the ceramic is made
of silicon nitride, these components are particularly suit-
able for ex- and in-situ measurements due to their high
Mohs hardness of 9–9.5, low weight and low absorption
coefficient.
In ex-situ single particle breakage experiments,
30–40 slag particles are crushed in the load cell. Multiple
and internal microstructure – often referred to as ran-
dom breakage or fracture. Liberation by detachment, on
the other hand, assumes full dependence, where breakage
occurs along grain boundaries loosely bonded in the matrix.
Despite the complexity, particle breakage mechanisms typi-
cally exhibit a superposition of both borderline cases. King
and Schneider (1998) introduced five additional condi-
tions for liberation by detachment that rule out random
breakage. In the comminution of natural minerals, random
breakage is the common assumption, therefore, additional
comminution steps downstream to ensure sufficient libera-
tion of the target phase are needed.
Typically, the assessment of the type of breakage
involves single particle breakage studies (Rumpf 2004,
Tavares 2007) and a method to image the broken state
(Maire et al. 2005, Ramandi et al. 2016). The former is
either done by single, double impact or slow compres-
sion testing. The fragments will then be photographed
and examined. Another way is to embed the fragments in
epoxy, grind and polish in a manner in which the polished
section is orthogonal to the breakage surface. While this
marks the well accepted and established method in various
studies (Popov et al. 2020), the stereological bias can still
not be avoided with irregularly or complex microstructures.
This study uses a three-dimensional approach to get insight
into the broken particle. The focus lays on breakages that
resulted from a press used for testing slag particles that can
be coupled with XCT as the imaging method.
The coupling of XCT is especially useful for analyz-
ing deformations and changes in the sample’s three-dimen-
sional structure under different loads or at different time
intervals (Van Stappen et al. 2019, Zhang et al. 2020). The
sample can remain in the comminution space and con-
tinue to be loaded until it fractures without external influ-
ence. Additionally, newly formed cracks and fractures can
be studied based on the internal, complex microstructure
of multiphase materials. This information is important,
particularly when different regions of the sample respond
differently to compression. This allows for establishing
a correlation with the breakage mechanism. The datasets
from microstructural or in-situ measurements can also be
used for breakage simulations to validate their accuracy.
MATERIAL AND METHODOLOGY
Materials and Microstructural Description of the Used
Lithium Aluminate Slag
The study utilized lithium aluminate slag produced by IME
RWTH Aachen. The production process involves homoge-
nization melting of the feed salts and a specifically designed
cooling strategy with a cooling rate of 50 K/h. The resulting
slag is composed of four minerals and an amorphous phase.
The lithium aluminate is the target phase or EnAM, which
accounts for up to 16.2 mass percent, is needed to be liber-
ated from the matrix phase.
The target phase is generally distributed throughout
the entire slag sample. In 2D, dendrites appear linear, but
in 3D, they have a layered structure with a wavy texture.
The dendrites have low connectivity, with only a thin main
branch connecting the layers. Upon closer examination of
the thin sections (Võ et al. 2024), it is evident that the den-
dritic branches consist of numerous small crystalline units,
consistently surrounded by the matrix phase. This state-
ment suggests a strong connection between the interlock-
ing and matrix phases.
The sample volume can exhibit large and fine pores,
whose dimensions and arrangement are related to crystal
growth, ion concentration in the molten state, and external
parameters such as temperature and pressure. Crystals that
grow rapidly may not achieve their final structural arrange-
ment within the cooling time, resulting in defects and voids
within the crystal lattice. These voids act as pores and are
randomly dispersed throughout the resulting slag product.
A closer and more detailed description of this particular
slag system can be found in the work of Võ et al. (2024).
Sample Preparation
To estimate the breakage force, irregularly shaped slag parti-
cles are crushed using a jaw crusher and sieving. The selected
fraction consists of slag particles small enough to fit within
the crushing space of the load cell. Specifically, slag particles
within the 4–5 mm fraction are chosen to estimate break-
age force for in-situ measurements and to adjust measure-
ment settings on the X-ray computed microscope (XCT).
A red coating on the outer surface helps with distinguish-
ing the original outer side from the newly formed fracture
surface after the successful compression and enables a clear
differentiation between the two surfaces after breakage.
Ex- Situ Single Particle Breakage
For single particle breakage studies, the Deben CT5000
5kN in-situ load cell tensile stage for XCT applications is
utilized. The crushing space is defined by the ceramic pis-
ton setup, consisting of upper and lower cylindrical ceramic
stamps and the guiding tube. Since the ceramic is made
of silicon nitride, these components are particularly suit-
able for ex- and in-situ measurements due to their high
Mohs hardness of 9–9.5, low weight and low absorption
coefficient.
In ex-situ single particle breakage experiments,
30–40 slag particles are crushed in the load cell. Multiple