Experiment C-5

Monte Carlo simulation of the 2D Ising

model
W. I. W. Y. Boteju
s14310
Advanced Physics Laboratory II
28.01.2023

Abstract—There are three kinds of magnetic ma- I. I NTRODUCTION

terials in nature. They are Diamagnetic materials,
Paramagnetic materials, and Ferromagnetic materi- In nature, there are three kinds of magnetism,
als. Ferromagnetic materials have the capability of namely Diamagnetism, Paramagnetism, and Ferro-
retaining the induced magnetization in the direction of magnetism. The three types of ferromagnetic mate-
the applied magnetic field even after the magnetic field rials are interesting because of their behavior under
is removed. The aim of this experiment is to simulate an applied magnetic field. Rare earth elements such
the phase transitions of a ferromagnetic material
to a paramagnetic material. The boundary of this as gadolinium, dysprosium, and materials such as
magnetism transition is set by the Curie temperature. iron, nickel, and cobalt exhibit ferromagnetism.
The magnetic material acts as ferromagnetic material Ferromagnetic materials are capable of retaining
in the temperatures below the Curie point and acts induced magnetism even after the external mag-
as a paramagnetic material in the temperatures above netic field is removed, this property is referred
the Curie temperature. A Monte Carlo simulation of
the 2D Ising Model was carried out to simulate the to as the hysteresis of the material. The behavior
above-mentioned transition. The average spin of the of ferromagnetic materials is strongly dependent
system and the instantaneous energy of the system on the temperature of the system. Under neces-
were determined to generate new states of the same sary conditions, magnetic materials are capable of
ferromagnetic material via the Metropolis algorithm. exhibiting ferromagnetism and paramagnetism [1].
The thermodynamic observables such as Internal
Energy, Specific Heat Capacity, Magnetization, and The boundary value of the above phase transition is
Magnetic Susceptibility were measured with respect referred to as the Curie Temperature. The theoretical
to the system temperature. It was observed that the Curie temperature for ferromagnetic materials is
thermodynamic observables demonstrate an abrupt 2.26 K [2]. The Ising Model is a mathematical
change in their behavior about the Curie point. The model which is capable of simulating ferromagentic
periodic lattice size dependence of the thermodynamic
observables was analyzed. It was determined that only materials and determining their phase transitions.
the Internal Energy and the Specific Heat Capacity The Ising Model is available in 1D, 2D, and 3D.
of the system display a dependency. The observations The complexity of the simulation is reliant on the
concluded that the simulation is accurate when the dimension used. Monte Carlo Technique is a widely
periodic lattice size is at its highest. The Binder popular method of multiple probability simulation,
Cumulant of the system was measured for different
lattice sizes and the Curie temperature at the phase which can be used to estimate many outcomes of an
transition point was calculated to be (2.0 ± 0.5) K, uncertain event. Random numbers are used to calcu-
which is in agreement with the theoretical Curie late probabilities in Monte Carlo simulations [3]. In
temperature for a ferromagnetic material 2.26 K. this experiment, the 2D Ising Model is used to sim-
ulate the phase transitions which occur at the Curie
point temperature simulated by a periodic lattice of

1
dimensions 56x56. The Monte Carlo Technique is
used to randomly assign spin states for the atoms.
The thermodynamic observables such as Internal
Energy, Specific Heat Capacity, Magnetization, and
Magnetic Susceptibility are measured with respect
to the system temperature and by varying the peri-
odic lattice dimensions. The Curie temperature for
the ferromagnetic material is determined from the Fig. 1. The behavior of a ferromagnetic material when a
Binder Cumulant. magnetic field is applied in the +x direction [5].

II. T HEORY

A. Ferromagnetic Materials
Magnetic materials are grouped by the response
that they provide to an external magnetic field,
1. Diamagnetic Materials
2. Paramagnetic Materials
3. Ferromagnetic Materials

Diamagnetism is the property when an exter-

nal magnetic field is applied the intrinsic atomic Fig. 2. The hysteresis loop of ferromagnetic materials [6].
spins of the material are aligned anti-parallel to
the external magnetic field direction, therefore a
magnetization is resulted opposing the applied field B. The 2D Ising Model
but weaker. Para magnetism is the property when
an external magnetic field is applied the intrinsic The Ising Model is a mathematical model which
atomic spins are aligned parallel to the applied is used to study phase transitions of materials. The
magnetic field. Hence, magnetization results in the model simulates a lattice in which a cell only
direction of the applied magnetic field and it is contains either an upward or downward spin. The
proportional to the applied magnetic field. Ferro- Ising Model is capable of representing magnetic
magnetism is the property of resulting magnetiza- materials, lattice gases, and metals.
tion in the direction of the applied field, and the During this study, the Ising Model is represented
order of the magnetization can be larger than the as a magnetic material. The energy of the magnetic
applied magnetic field. Even if the magnetic field material is taken as the Hamiltonian of the system
is removed, a certain level of magnetization will be since it is capable of representing each of the
left in a ferromagnetic material. Natural elements microstates of the system. The hamiltonian in this
like Iron, Nickel, and Cobalt exhibit this behavior. system includes two terms, an external field term
Ferromagnets can be used to fabricate permanent and a term that corresponds to the interactions
magnets [1] [4]. between neighboring atoms. The external field term
Ferromagnetic materials show hysteresis as a represents the individual spins after the system is
characteristic. This occurs when a ferromagnetic kept in an external magnetic field. The interaction
material is applied with an alternating magnetic term represents the reaction of the spins due to their
field. The magnetization of the material is capable neighboring spins which acts as a mini magnetic
of tracing out a loop, by keeping a memory of its dipole [7].
behavior. The Hamiltonian of the system is given by,

2
 X X
H=− Jσi σj − hσj (1)
<i,j> j

where
J represents the Spin-Spin interactions
h represents the external field
σj are the individual spins on each of the lattice
sites.

C. Detailed Balance Condition

The instantaneous energy of the system is given
by,

X
E = −J σi σj (2)
<i,j>

A new state of the ferromagnetic material is only Fig. 3. The phase transition of a ferromagnetic material at the
Curie temperature [8].
accepted if the total energy of the current state is
lower than the total energy of the previous state.
If not the detailed balance condition must be 2) Magnetization (M): The average
satisfied, which is given by, magnetization level of the system is given
from this parameter. This is also referred to as the
X X order parameter of the system and it is calculated
Pµ P (µ → ν)) = Pν P (ν → µ)) (3) by,
µ ν

D. Curie Temperature
The Curie temperature is defined as the temper- 1
< M >= Σ<i,j> σi σj (5)
ature at which ferromagnetic materials transform N2
into paramagnetic materials. The Curie Law states
where
that the magnetic susceptibility of the material is
N is the number of elements in the system.
inversely proportional to its temperature [8].
3) Specific Heat Capacity (C): The quantity of
E. Thermodynamic Observables heat required per 1 kg of material when its temper-
1) Internal Energy: The internal energy of the ature increases from 1 K [9].
system is defined as the kinetic energy of the
molecules and it is calculated by the following
β
equation. C = − (< E 2 > − < E >2 ) (6)
T
where
Eµ = Σ<i,j> − Jσi σj (4)
β = kB1T
where kB is the Boltzmann Constant.
J represents the Spin-Spin interactions < e2 > is the average energy squared of the lattice
σi is the spin of element i in the n × n matrix. < E >2 is the square of the average energy
σj is the spin of the element j in the n × n matrix.

3
 4) Magnetic Susceptibility (χ): Magnetic sus- current system is lower than the previous system’s
ceptibility provides the degree of magnetization of total energy. If not the current state was accepted
the material when an external magnetic field is with a probability generated via random numbers
applied [10]. and the condition was set to the ’Detailed Balanced
Condition’ in numerical methods. Above steps were
repeated. The entire process was designed to operate
χ = −β(< M 2 > − < M >2 ) (7) independently as a function and was called the
where ’Metropolis Algorithm’.
< M 2 > is the average of the squared The changes in the net spin and the instantaneous
magnetization energy of the ferromagnetic lattice were observed
< M >2 is the square of the average magnetization with respect to the Monte Carlo time.

B. Measuring the thermodynamical observables as

F. Binder Cumulant (U) a function of Temperature
Binder Cumulant is a measure of the Curie The temperature of the system was significantly
temperature of a ferromagnetic material. The phase varied from 1 K to 10 K with a step size of 0.25
transition point of a magnetic material can be de- K. The net spin and the instantaneous energy of the
termined by measuring the Binder Cumulant. ferromagnetic material were observed with the tem-
perature as a function of Monte Carlo Time. Ther-
< M4 > modynamical observables such as Specific Heat
U =1− (8) Capacity, Magnetization (Order Parameter), Internal
3 < M 2 >2
Energy, and Magnetic Susceptibility of the system
III. M ETHODOLOGY were measured with respect to the temperature. The
A. Simulation of the ferromagnetic lattice standard deviation of the above observables was
The simulation of the 2D Ising Model was per- measured.
formed in the open-source scientific programming C. Measuring the thermodynamical observables as
environment, ’Spyder’ which runs in Python 3.8. a function of Temperature and Lattice dimension
First, a unit cell of the ferromagnetic material was
created with dimensions of 8 x 8. The unit lattice The ferromagnetic lattice dimensions were varied
consisted of 64 cells and 80 % of the cells were from 16x16 to 64x64 with a step size of 8x8
given a spin of -1 (spin down) and the remaining per iteration. The behavior of the thermodynami-
cells were assigned a spin of +1 (spin up). The cal observables was investigated as a function of
assignment of spins was carried out via a Monte temperature and the lattice dimension.
Carlo Simulation. IV. R ESULTS AND D ISCUSSION
A periodic lattice with boundary conditions was
simulated by repeating the unit cell to create a 7 x 7 A. Simulation of the ferromagnetic lattice
matrix, which outputs a ferromagnetic lattice with The 8x8 unit lattice which consisted of 80 % of
3136 cells. the cells having -1 spin and 20 % of the cells having
Next, random numbers were used to pick a ran- +1 spin is shown below.
dom element from the ferromagnetic lattice, and Fig. 4 represents the unit lattice of the ferromag-
its spin was changed to the opposite spin. The netic material with each cell having random spin
net change of spin due to the effects from each states. Therefore, the net spin of the unit lattice
neighboring element was calculated and the net remains zero.
energy of the lattice was determined. This process Next, the periodic lattice of the ferromagnetic
creates a new state of the ferromagnetic lattice, material was simulated by placing the same unit
which was accepted only if the total energy of the lattice in a 7x7 empty lattice.

4
 Algorithm.
Next, the average spin and the instantaneous
energy of the system were observed at 1 K.

Fig. 4. The figure of the unit lattice of the ferromagnetic

material. The majority of the cells are occupied with a -1 spin
state, which is indicated in grey color.

Fig. 6. The figure of the graph of the Average Spin of the system
with respect to Monte Carlo Time. The average spin is +1 at 1K,
indicating that the system has aligned parallel to the external
magnetic field. The systems act as a ferromagnetic material.

Fig. 5. The figure of the periodic lattice of the ferromagnetic

material. The unit lattice is repeatedly placed in the space of
7x7. This represents the macroscopic nature of the ferromagnet.

Fig. 5 represents the macroscopic image of the

ferromagnetic material. Thus far, the ferromagnetic Fig. 7. The figure of the graph of Instantaneous energy of the
material is kept in a neutral environment. Once an system with respect to Monte Carlo Time. The system reaches
external magnetic field is applied, the spins will an equilibrium when its energy is at its lowest.
orient in the direction of the external magnetic field.
Subsequently, if the field is removed, the ferromag- According to Fig. 6, the instantaneous magne-
netic materials will demonstrate a specific behavior tization of the system has reached +1 at a later
that forces them to retain the magnetization by Monte Carlo Time. This exhibits that the system
being aligned parallel to the previously applied ex- has adapted as a ferromagnetic material. All of
ternal magnetic field. The temperature of the system the atomic spins are directed toward the externally
decides whether it remains a ferromagnetic material applied field. The observation is true since the
or a paramagnetic material. This behavior of the system is kept at a temperature lower than the Curie
magnetic material is simulated via the Metropolis temperature.

5
 According to Fig. 7, the instantaneous energy
of the system is stabilized when the algorithm
time steps are increased. This indicates that the
system has reached an equilibrium. The new state
of the ferromagnetic material is only accepted if the
current system energy is lower than the previous
system energy, if not with the Detailed Balance
Condition. The temperature of the system is kept
constant at 1K during the above observations.
B. Measuring the thermodynamical observables as
a function of Temperature
Fig. 9. The figure of the graph of Instantaneous energy of
The average spin and the instantaneous energy the system with respect to Monte Carlo Time for different
of the system were measured with respect to the temperatures. From 2 K to 2.5 K the instantaneous spin of the
Monte Carlo Time while varying the temperature system finds the equilibrium at higher energy. This occurrence
is a result of system temperature travel over the Curie point.
between 1 K to 3.5 K with an increment of 0.5 K.

Fig. 8. The figure of the graph of the Average Spin of the system
with respect to Monte Carlo Time for different temperatures. Fig. 10. The figure of the graph of the Internal Energy of the
From 2 K to 2.5 K the average spin of the system exhibits a system with respect to the Temperature. Internal energy exhibits
drastic change. This occurrence is a result of system temperature a rapid increment around the Curie Temperature. This proves the
travel over the Curie point. statement, Internal Energy Temperature.

According to Fig. 8 and Fig. 9, the average

spin of the system and the instantaneous energy According to Fig.10, the internal energy at equi-
of the system demonstrate a rapid fluctuation with librium increases with the temperature. Especially,
respect to the system temperature at around 2 - about Curie Point, the internal energy demonstrates
2.5 K. The average spin of the system reaches a rapid increment. In thermodynamics, it states that
zero with increasing system temperature. The rapid with system temperature increasing, the internal
fluctuations of these variables are a result of the energy of the system also increases if the system is
Curie point temperature, theoretically 2.26 K. kept in a heat bath. This behavior can be exhibited
When the system temperature is greater than 2.26 in the simulation of the ferromagnetic lattice.
K, the system should act as a paramagnetic material, According to Fig.11, the magnetization of the
if not as a ferromagnetic material. The theoretically system drops to zero around the system tempera-
expected behaviors are seen from the simulations of ture. Magnetization is a measure of atomic spin.
the ferromagnetic periodic lattice. Therefore, if the average atomic spin tends to zero,

6
 material to an increment of temperature by a unit.
At the phase transition point, absorption of the heat
of the material is highest.

Fig. 11. The figure of the graph of the Magnetization of the

system with respect to the Temperature. A rapid decrement of
the magnetization near the Curie point is observed.

Fig. 13. The figure of the graph of the Magnetic Susceptibility

the lattice acts as a paramagnetic material. This of the system with respect to the Temperature. The maximum
demonstrates a phase transition of the system from magnetic susceptibility is obtained around Curie Point.
being a ferromagnetic material to a paramagnetic
material with increasing temperature around Curie According to Fig.13, the magnetic susceptibility
Point. Theoretically, the thermodynamic observ- of the material reaches a maximum at Curie Point.
ables should abruptly change near the Curie point. Susceptibility is a measure of the level of magne-
The expected behavior is seen from the simulation. tization possible for an applied magnetic field. The
susceptibility at a temperature greater than Curie
Point tends to zero level. The statistical uncertainty
which is obtained by calculating the standard devia-
tion of the respective variable is significantly larger
around the Curie Point and this anomaly is not
negligible for the thermodynamic observables such
as specific heat capacity and magnetic susceptibility.
This occurs due to a statistical limitation of the
simulation algorithm.
C. Measuring the thermodynamical observables as
a function of Temperature and Lattice dimension
The thermodynamic observables were measured
as a function of the temperature by varying the
Fig. 12. The figure of the graph of the Specific Heat Capacity periodic lattice size of the system from 16x16 to
of the system with respect to the Temperature. This curve 64x64.
demonstrates the derivative of the system energy with respect
to temperature as a function of temperature. The maximum of According to Fig.14, the internal energy of the
the specific heat capacity is obtained at Curie Point. system is lowest when the lattice size is the largest.
The internal energy of temperatures below the Curie
According to Fig.12, the specific heat capacity Point exhibits a sudden drop. The drop in the energy
of the system reaches a maximum at the Curie becomes larger for larger lattice sizes, indicating
point. The specific heat capacity is defined as the that the initial internal energy is lower at 1 K. This
heat absorbed by a unit mass of the ferromagnetic occurs because the randomness of the spin in the

7
 system is independent of the size of the system.
The atomic spin has only two states up and down,
therefore it is independent of the lattice size, which
confirms the observation.

Fig. 14. The figure of the graph of the Internal Energy of the
system with respect to the Temperature with varying lattice size.
At lower temperatures, the system reaches an equilibrium when
the lattice dimensions are increased. The lowest equilibrium
energy is reached when the lattice size is 64x64.

Fig. 16. The figure of the graph of the Specific Heat Capacity
simulation is more prominent when the macroscopic of the system with respect to the Temperature with varying
lattice size. The specific heat capacity of the system increases
system is bigger. The simulation is accurate when with the lattice size and the maximum is obtained for 64x64. A
the periodic lattice is larger. The size of the peri- drastic increment of the specific heat capacity is seen after the
odic lattice is constrained to a maximum of 64x64 40x40 lattice size. The Curie temperature demonstrates a slight
increment with the lattice size.
because of the computational time.
According to Fig.16, the specific heat capac-
ity of the system demonstrates a large increment
with increasing lattice size. The peak points of the
specific heat capacity exhibit two clustered sets,
the first cluster being all the lattice sizes from
16x16 to 40x40, and the second cluster is formed
from the peak points of the 48x48 lattice onwards.
This indicates that the system behaves prominently
when the macroscopic scale of the system is bigger.
The Curie temperature of the system exhibits a
slight increment when the lattice size is larger. The
lowest lattice size of 16x16 illustrates the Curie
temperature of around 2 K while the largest lattice
Fig. 15. The figure of the graph of the Magnetization of the size of 64x64 illustrates the Curie temperature of
system with respect to the Temperature with varying lattice size. around 3 K.
The magnetization of the system does not exhibit a dependence
on the lattice size. For all the lattice sizes, the magnetization of
According to Fig.17, the magnetic susceptibility
the system varies similarly with temperature. of the system displays no dependence on the lattice
size. The susceptibility corresponding to the lattice
According to Fig.15, the magnetization of the size 32x32 is an outlier. The dependence of the
system behaves as in Fig.11 for all the lattice sizes. magnetic susceptibility with temperature remains as
Therefore, it can be concluded that there is no in Fig.13.
dependence on the magnetization of the system with From the above observations, it should be noted
the lattice size, which signifies that the spin of the that the thermodynamic variables such as Internal

8
 the Curie temperature has no relationship with the
lattice dimensions which indicates that the material
is transformed into a paramagnetic material.

V. C ONCLUSIONS

Ferromagnetic materials are capable of retaining

magnetization even after the external magnetic field
is removed. In this experiment, a ferromagnetic
material is simulated via the 2D Ising Model and
Fig. 17. The figure of the graph of the Magnetic Susceptibility of Monte Carlo methods. The microscopic image of
the system with respect to the Temperature with varying lattice the ferromagnetic material is generated via a unit
size. Magnetic susceptibility does not express a relationship with
the periodic lattice size. lattice with 80 % of the spins directed downwards.
The macroscopic image of the ferromagnetic ma-
terial is represented by the periodic lattice, where
Energy and Specific Heat Capacity conclude that a unit lattice is kept in a matrix of 7x7. New
the larger lattice dimensions are preferred during states of the periodic lattice are generated and the
the simulation. states are accepted if the total energy of the current
system is lower than the previous total energy of
the system, else with the detailed balance condition.
The average spin of the system reached +1 when
the temperature of the system is below the Curie
point, else the average spin was 0 because atomic
spins were randomly aligned. The instantaneous
energy of the system reached equilibrium at lower
energy when the system temperatures were higher.
The thermodynamic observables such as Internal
Energy, Specific Heat Capacity, Magnetization, and
Magnetic Susceptibility were measured with respect
to the system temperature. The observables demon-
strated an abrupt change in their behavior around
the Curie temperature. The same set of observables
Fig. 18. The figure of the graph of the Binder Cumulant with
respect to Temperature as a function of the lattice size. The Curie was measured with respect to the temperature while
temperature of the simulated system is (2.0 ± 0.5) K. varying the periodic lattice size. The variations of
the observables such as Internal Energy and Specific
According to Fig.18, the Binder Cumulant of all Heat Capacity illustrated that the simulated results
the lattice dimension curves of the system passes are accurate when compared with the theoretically
through a single point which occurs from 2 - 4 K. predicted results when the lattice size is at its high-
This indicates a phase transition point and the Curie est. The Binder Cumulant of the system was used to
temperature of the ferromagnetic lattice is measured find the phase transition point of the ferromagnetic
to be (2.0 ± 0.5) K. Binder Cumulant of the system material. The Curie temperature of the system was
before the Curie temperature remains same for all determined to be (2.0 ± 0.5) K, which agrees
the lattice sizes since it behaves as a ferromagnetic with the theoretically predicted Curie temperature
material. The Binder Cumulant of the system after of 2.26 K.

9
 R EFERENCES
[1] Ferromagnetism. [Online]. Available: http://hyperphysics.
phy-astr.gsu.edu/hbase/Solids/ferro.html#c1. [Accessed:
20-Jan-2023].
[2] “Curie Point,” Encyclopædia Britannica. [Online]. Avail-
able: https://www.britannica.com/science/Curie-point. [Ac-
cessed: 21-Jan-2023].
[3] “What is Monte Carlo Simulation?,” IBM. [Online]. Avail-
able: https://www.ibm.com/topics/monte-carlo-simulation.
[Accessed: 21-Jan-2023].
[4] Magnetic properties of solids. [Online]. Available: http:
//hyperphysics.phy-astr.gsu.edu/hbase/Solids/magpr.html.
[Accessed: 20-Jan-2023].
[5] “Hysteresis,” Hysteresis in magnetic materials. [Online].
Available: http://hyperphysics.phy-astr.gsu.edu/hbase/
Solids/hyst.html. [Accessed: 22-Jan-2023].
[6] Admin, “Magnetic and non-magnetic met-
als with examples,” SMLease Design, 21-
Jul-2022. [Online]. Available: https://www.
smlease.com/entries/mechanical-design-basics/
magnetic-and-non-magnetic-metals/. [Accessed: 22-
Jan-2023].
[7] The Ising model. [Online]. Available: https://stanford.edu/
∼jeffjar/statmech/intro4.html. [Accessed: 25-Jan-2023].
[8] “What is the Curie temperature of magnets?,” IMA,
10-Jun-2022. [Online]. Available: https://imamagnets.
com/en/blog/what-is-the-curie-temperature-of-magnets/.
[Accessed: 24-Jan-2023].
[9] “Specific heat capacity,” Specific Heat Capacity - an
overview — ScienceDirect Topics. [Online]. Available:
https://www.sciencedirect.com/topics/engineering/
specific-heat-capacity#:∼:text=The%20specific%20heat%
20capacity%20is,J%2F(kg%20%C2%B0C). [Accessed:
25-Jan-2023].
[10] “What is magnetic susceptibility? - formula, definition,
unit,” BYJUS, 19-Jul-2022. [Online]. Available: https:
//byjus.com/physics/magnetic-susceptibility/#:∼:text=
Magnetic%20susceptibility%20is%20a%20dimensionless,
the%20externally%20applied%20magnetic%20field.
[Accessed: 26-Jan-2023].

10