Understanding Arrott Plots — Minh-Hieu Nguyen

If you work with magnetic materials, sooner or later you will encounter Arrott plots. They are one of the most powerful yet underappreciated tools in magnetism research, allowing you to determine the order of a magnetic phase transition, extract critical exponents, and estimate the Curie temperature — all from straightforward magnetization data. Yet despite their importance, many experimental papers apply them incorrectly or incompletely. This guide aims to change that.

1. What Is an Arrott Plot?

An Arrott plot is simply a graph of $M^2$ vs. $H/M$ constructed from isothermal magnetization data $M(H)$ measured at different temperatures around the Curie temperature $T_C$. The idea was first proposed by Anthony Arrott in 1957 and originates from the Landau mean-field theory of phase transitions.

The underlying physics is elegant. Near a second-order phase transition, the magnetic equation of state can be written in a generalized form. In the simplest mean-field approximation, this reduces to a linear relationship between $M^2$ and $H/M$. If the transition is truly second-order and mean-field theory applies, the isotherms in an Arrott plot should appear as a set of parallel straight lines. The isotherm that passes through the origin corresponds to the Curie temperature.

Arrott Equation (Mean-Field) $$\frac{H}{M} = a\left(T - T_C\right) + b\,M^2$$

where $a$ and $b$ are material-dependent constants. Rearranging gives the linear form: plotting $M^2$ on the $y$-axis versus $H/M$ on the $x$-axis should yield straight lines if the mean-field description is valid.

2. Why Arrott Plots Matter

The real power of Arrott plots lies in what they reveal about the nature of the magnetic phase transition. In experimental work, knowing whether a transition is first-order or second-order has profound implications for understanding the underlying physics and for potential applications, particularly in magnetocaloric refrigeration.

A second-order transition shows positive slopes in the Arrott plot. A first-order transition shows negative slopes or S-shaped curves with inflection points. This distinction is critical for magnetocaloric applications because the nature of the transition directly determines whether the entropy change is continuous or discontinuous.

Specifically, Arrott plots help you answer three fundamental questions about your magnetic material:

What is the order of the phase transition? The sign of the slope in the high-field region of the Arrott plot tells you directly. Positive slope means second-order; negative slope or S-shape means first-order. This criterion is often called the Banerjee criterion.
What is the Curie temperature? The isotherm that passes through (or closest to) the origin in the $M^2$ vs. $H/M$ plot corresponds to $T_C$.
What are the critical exponents? Through modified Arrott plots (see below), you can extract $\beta$, $\gamma$, and $\delta$, which define the universality class of your material's phase transition.

3. Step-by-Step: How to Construct an Arrott Plot

Collect Isothermal $M(H)$ Data

Measure magnetization $M$ as a function of applied field $H$ at multiple temperatures around the expected $T_C$. Typical spacing is 2–5 K, and you need data both above and below $T_C$. Ensure you go to sufficiently high fields (at least 5 T if possible) to reach the high-field linear regime. The more isotherms you measure, the more precise your analysis will be.

A typical measurement uses a VSM or SQUID magnetometer. Make sure your sample is properly aligned and demagnetization corrections are applied if the sample geometry requires it.

Calculate $M^2$ and $H/M$

For each isotherm, compute $M^2$ and $H/M$ at every data point. This is straightforward — just square the magnetization values and divide the field by magnetization. Be careful with units: if $M$ is in emu/g and $H$ is in Oe, then $H/M$ will be in Oe·g/emu.

Plot $M^2$ vs. $H/M$

Create the Arrott plot with $H/M$ on the $x$-axis and $M^2$ on the $y$-axis. Each isotherm appears as a separate curve. Color-code or label them by temperature. Below $T_C$, curves will have a positive $y$-intercept (spontaneous magnetization exists). Above $T_C$, curves will have a negative $y$-intercept (extrapolated). The curve passing through the origin marks $T_C$.

Apply the Banerjee Criterion

Examine the slope of the isotherms in the high-field region. If all slopes are positive and the curves are quasi-linear, the transition is second-order. If any isotherm shows a negative slope or an S-shaped inflection, the transition has first-order character. Document this clearly — it is one of the most important results from the analysis.

Extract $T_C$ and Spontaneous Magnetization

Perform a linear fit to the high-field portion of each isotherm. The $y$-intercept of each fit gives $M_S^2(T)$ — the square of the spontaneous magnetization at that temperature. The $x$-intercept gives the inverse initial susceptibility $\chi_0^{-1}(T)$. Plot $M_S(T)$ and $\chi_0^{-1}(T)$ versus temperature to determine $T_C$ precisely.

4. Modified Arrott Plots and Critical Exponents

Standard Arrott plots assume mean-field critical exponents ($\beta = 0.5$, $\gamma = 1.0$). However, real materials often belong to different universality classes. This is where Modified Arrott Plots (MAPs) become essential.

Instead of plotting $M^2$ vs. $H/M$, the modified version plots $M^{1/\beta}$ vs. $(H/M)^{1/\gamma}$, where $\beta$ and $\gamma$ are the critical exponents appropriate for the universality class of the material. The correct exponents will produce the best set of parallel straight lines in the high-field region.

Modified Arrott Plot (Generalized) $$\left(\frac{H}{M}\right)^{1/\gamma} = a\,\varepsilon + b\,M^{1/\beta}$$

where $\varepsilon = (T - T_C)/T_C$ is the reduced temperature. The key universality classes to test are:

Model	$\beta$	$\gamma$	$\delta$
Mean-field	0.500	1.000	3.00
3D Heisenberg	0.365	1.386	4.80
3D Ising	0.325	1.241	4.82
3D XY	0.345	1.316	4.81
Tricritical mean-field	0.250	1.000	5.00

The best-fit model is the one that produces the most linear and most parallel set of isotherms in the modified Arrott plot. This is typically assessed by the relative slope $RS = S(T)/S(T_C)$, where values closest to unity across all temperatures indicate the best model.

For manganite perovskites like Nd_0.7Sr_0.3MnO₃, the tricritical mean-field model ($\beta = 0.25$, $\gamma = 1.0$) often provides the best fit, indicating the system sits at the boundary between first-order and second-order behavior — exactly the tricritical point.

5. The Iterative Method for Precise Exponents

In practice, you rarely know the correct exponents beforehand. The standard approach is an iterative self-consistent procedure:

Start with trial exponents from one of the models in the table above.
Construct the modified Arrott plot using $M^{1/\beta}$ vs. $(H/M)^{1/\gamma}$.
Fit the linear high-field region of each isotherm.
Extract $M_S(T)$ from the $y$-intercept and $\chi_0^{-1}(T)$ from the $x$-intercept.
Fit $M_S(T)$ near $T_C$ to the power law $M_S(T) \propto (T_C - T)^{\beta}$ to obtain a new $\beta$.
Fit $\chi_0^{-1}(T)$ above $T_C$ to the power law $\chi_0^{-1}(T) \propto (T - T_C)^{\gamma}$ to obtain a new $\gamma$.
Use the new $\beta$ and $\gamma$ to reconstruct the modified Arrott plot.
Repeat until $\beta$ and $\gamma$ converge (typically 3–5 iterations are sufficient).

The final converged values of $\beta$ and $\gamma$ reveal the universality class of the transition. You can then verify self-consistency by checking the Widom relation: $\delta = 1 + \gamma/\beta$.

6. Python Implementation

Here is a minimal Python script to construct Arrott plots from raw $M(H)$ data and perform the linear fits:

Python — Arrott Plot Analysis

import numpy as np 

import matplotlib.pyplot as plt 

from scipy import stats 

# Load your isothermal M(H) data 

# Format: columns = [H, M_T1, M_T2, M_T3, ...] 

data = np.loadtxt('MH_isotherms.dat') 

H = data[:, 0] 

temperatures = [250, 255, 260, 265, 270, 275]  # K 

# Critical exponents to try (tricritical mean-field) 

beta  = 0.25 

gamma = 1.0

fig, ax = plt.subplots(figsize=(8, 6))

for i, T in enumerate(temperatures):

    M = data[:, i + 1]

    # Modified Arrott axes: M^(1/beta) vs (H/M)^(1/gamma)

    x = (H / M) ** (1 / gamma)

    y = M ** (1 / beta)

    ax.plot(x, y, 'o-', markersize=3, label=f'{T} K')

    # Linear fit to high-field region (e.g., above 1 T)

    mask = H > 10000

    slope, intercept, r, p, se = stats.linregress(x[mask], y[mask])

    print(f'T={T} K: slope={slope:.3f}, intercept={intercept:.3f}')

ax.set_xlabel(r'$(H/M)^{1/\gamma}$', fontsize=12)

ax.set_ylabel(r'$M^{1/\beta}$', fontsize=12)

ax.set_title('Modified Arrott Plot')

ax.legend(fontsize=9)

plt.tight_layout()

plt.savefig('arrott_plot.png', dpi=150)

plt.show()

7. Common Pitfalls to Avoid

⚠ Pitfall #1: Fitting the Entire Curve

Only the high-field linear region should be used for fitting. The low-field region is dominated by domain wall motion and magnetic domain effects that have nothing to do with the critical behavior. Including low-field data will give you wrong exponents.

⚠ Pitfall #2: Insufficient Temperature Range

You need isotherms both well below and well above $T_C$. If all your data is on one side, you cannot properly determine $T_C$ or extract both $\beta$ and $\gamma$. A good rule of thumb is to have at least 5 isotherms below and 5 above $T_C$.

⚠ Pitfall #3: Ignoring Demagnetization

For samples with significant demagnetization factors (thin films, irregular shapes), the internal field $H_\text{int} = H_\text{app} - NM$ can be significantly different from the applied field. Failing to correct for this will distort your Arrott plot and give incorrect exponents.

⚠ Pitfall #4: Declaring Mean-Field by Default

Many papers plot only the standard Arrott plot ($M^2$ vs. $H/M$) and declare the transition "second-order" based on positive slopes. But positive slopes alone do not confirm the mean-field model. You must construct modified Arrott plots with different exponent sets and compare the linearity to determine the correct universality class.

⚠ Pitfall #5: Not Checking Self-Consistency

After extracting $\beta$ and $\gamma$, always verify with the Widom relation ($\delta = 1 + \gamma/\beta$) and the Rushbrooke equality ($\alpha + 2\beta + \gamma = 2$). If these are violated, your exponents are unreliable, and you should revisit the fitting range or reconsider whether the system is truly in the asymptotic critical regime.

· · ·

8. Connecting to Magnetocaloric Analysis

Arrott plot analysis is not just an academic exercise — it directly feeds into magnetocaloric calculations. Once you know the order of the transition and the critical exponents, you can use them to evaluate the magnetic entropy change $\Delta S_M$ using the Maxwell relation:

Maxwell Relation $$\Delta S_M(T,\,H) = \int_0^{H} \left(\frac{\partial M}{\partial T}\right)_{\!H} dH$$

For second-order transitions, $\Delta S_M$ is fully reliable from the Maxwell relation. However, for first-order transitions, caution is needed because the discontinuity in magnetization can lead to spurious spikes in the calculated entropy change (the so-called "spike artifact"). The Arrott plot analysis tells you upfront whether you need to worry about this.

Furthermore, the critical exponents obtained from modified Arrott plots can be used to construct the universal curve of $\Delta S_M$ — collapsing all entropy change data onto a single master curve when properly rescaled. This is one of the most elegant consistency checks available in magnetocaloric research.

The beauty of Arrott plots is that they transform complex magnetization data into a visual language: straight lines mean understanding; curvature means there is more physics to uncover.

9. Summary and Checklist

Before submitting your next paper with Arrott plot analysis, make sure you can check off every item on this list:

Measured $M(H)$ isotherms at close temperature intervals spanning both sides of $T_C$
Applied demagnetization corrections if needed for your sample geometry
Constructed the standard Arrott plot ($M^2$ vs. $H/M$) and applied the Banerjee criterion
Tested multiple universality class models via modified Arrott plots
Selected the best model based on relative slope analysis ($RS$ closest to 1)
Performed the iterative procedure to converge on self-consistent $\beta$ and $\gamma$
Verified the Widom relation and Rushbrooke equality
Clearly reported $T_C$, $\beta$, $\gamma$, $\delta$ with uncertainties
Connected the results to the magnetocaloric entropy change analysis

Arrott plots are deceptively simple to construct but remarkably deep in what they reveal. Getting them right means the difference between a routine measurement and genuine physical insight. Take the time, do it properly, and your reviewer — and your science — will thank you.

Understanding Arrott Plots: A Practical Guide for Experimentalists

1. What Is an Arrott Plot?

2. Why Arrott Plots Matter

3. Step-by-Step: How to Construct an Arrott Plot

Collect Isothermal \(M(H)\) Data

Calculate \(M^2\) and \(H/M\)

Plot \(M^2\) vs. \(H/M\)