Full download ebook at ebookname.

com

Solid State Physics 67 1st Edition Robert E.

Camley And Robert L. Stamps (Eds.)

https://ebookname.com/product/solid-state-physics-67-1st-
edition-robert-e-camley-and-robert-l-stamps-eds/

OR CLICK BUTTON

DOWLOAD NOW

Download more ebook from https://ebookname.com

More products digital (pdf, epub, mobi) instant
download maybe you interests ...

Solid State Physics None

https://ebookname.com/product/solid-state-physics-none/

Solid State Physics 1st Edition Neil W. Ashcroft

https://ebookname.com/product/solid-state-physics-1st-edition-
neil-w-ashcroft/

Understanding Solid State Physics 1st Edition Sharon

Ann Holgate (Author)

https://ebookname.com/product/understanding-solid-state-
physics-1st-edition-sharon-ann-holgate-author/

Crystallography Applied to Solid State Physics A. R.

Verma

https://ebookname.com/product/crystallography-applied-to-solid-
state-physics-a-r-verma/
Solid State Physics Advances in Research and
Applications Vol 58 1st Edition Henry Ehrenreich

https://ebookname.com/product/solid-state-physics-advances-in-
research-and-applications-vol-58-1st-edition-henry-ehrenreich/

Physics 5th Edition Robert Resnick

https://ebookname.com/product/physics-5th-edition-robert-resnick/

Human Resource Management 13th Edition Robert L.

(Robert L. Mathis) Mathis

https://ebookname.com/product/human-resource-management-13th-
edition-robert-l-robert-l-mathis-mathis/

Solid State NMR Studies of Biopolymers 2nd Edition Anne

E. Mcdermott

https://ebookname.com/product/solid-state-nmr-studies-of-
biopolymers-2nd-edition-anne-e-mcdermott/

Chromatography Robert L. Wixom

https://ebookname.com/product/chromatography-robert-l-wixom/
Academic Press is an imprint of Elsevier
50 Hampshire Street, 5th Floor, Cambridge, MA 02139, United States
525 B Street, Suite 1800, San Diego, CA 92101-4495, United States
The Boulevard, Langford Lane, Kidlington, Oxford OX5 1GB, United Kingdom
125 London Wall, London, EC2Y 5AS, United Kingdom

First edition 2016

Copyright © 2016 Elsevier Inc. All rights reserved.

No part of this publication may be reproduced or transmitted in any form or by any means,
electronic or mechanical, including photocopying, recording, or any information storage and
retrieval system, without permission in writing from the publisher. Details on how to seek
permission, further information about the Publisher’s permissions policies and our
arrangements with organizations such as the Copyright Clearance Center and the Copyright
Licensing Agency, can be found at our website: www.elsevier.com/permissions.

This book and the individual contributions contained in it are protected under copyright by
the Publisher (other than as may be noted herein).

Notices
Knowledge and best practice in this field are constantly changing. As new research and
experience broaden our understanding, changes in research methods, professional practices,
or medical treatment may become necessary.

Practitioners and researchers must always rely on their own experience and knowledge in
evaluating and using any information, methods, compounds, or experiments described
herein. In using such information or methods they should be mindful of their own safety and
the safety of others, including parties for whom they have a professional responsibility.

To the fullest extent of the law, neither the Publisher nor the authors, contributors, or editors,
assume any liability for any injury and/or damage to persons or property as a matter of
products liability, negligence or otherwise, or from any use or operation of any methods,
products, instructions, or ideas contained in the material herein.

ISBN: 978-0-12-804796-5
ISSN: 0081-1947

For information on all Academic Press publications

visit our website at https://www.elsevier.com/

Publisher: Zoe Kruze

Acquisition Editor: Poppy Garraway
Editorial Project Manager: Shellie Bryant
Production Project Manager: Vignesh Tamil
Cover Designer: Maria Ines Cruz
Typeset by SPi Global, India
CONTRIBUTORS

T. Dumelow
Universidade do Estado do Rio Grande do Norte (UERN), Mossoró, Brazil
H. Kachkachi
PROMES, CNRS-UPR 8521, Universite de Perpignan Via Domitia, Perpignan, France
D.S. Schmool
Groupe d’Etude de la Matière Condensee GEMaC, CNRS (UMR 8635) Universite de
Versailles/Saint-Quentin, Universite Paris-Saclay, Versailles, France

vii
PREFACE

It is our great pleasure to present the 67th edition of Solid State Physics. The
vision statement for this series has not changed since its inception in 1955,
and Solid State Physics continues to provide a “mechanism … whereby inves-
tigators and students can readily obtain a balanced view of the whole field.”
What has changed is the field and its extent. As noted in 1955, the knowl-
edge in areas associated with solid state physics has grown enormously, and it
is clear that boundaries have gone well beyond what was once, traditionally,
understood as solid state. Indeed, research on topics in materials physics,
applied and basic, now requires expertise across a remarkably wide range
of subjects and specialties. It is for this reason that there exists an important
need for up-to-date, compact reviews of topical areas. The intention of these
reviews is to provide a history and context for a topic that has matured suf-
ficiently to warrant a guiding overview.
The topics reviewed in this volume illustrate the great breadth and diver-
sity of modern research into materials and complex systems, while providing
the reader with a context common to most physicists trained or working in
condensed matter. The chapter “Collective Effects in Assemblies of Mag-
netic Nanoparticles” provides an overview of emergent behavior arising
from collections of interacting magnetic particles from the perspective of
experiment, and also in terms of modeling and theory. The second chapter,
“Negative Refraction and Imaging from Natural Crystals with Hyperbolic
Dispersion,” describes aspects of material optics with a focus on the fascinat-
ing properties of hyperbolic materials whose surprising properties can be
found in naturally occurring single-phase materials, as opposed to
metamaterials in which these properties are engineered through design.
The editors and publishers hope that readers will find the introductions
and overviews useful and of benefit both as summaries for workers in these
fields, and as tutorials and explanations for those just entering.
ROBERT E. CAMLEY AND ROBERT L. STAMPS

ix
 CHAPTER ONE

Collective Effects in Assemblies

of Magnetic Nanaparticles
D.S. Schmool*,1, H. Kachkachi†
*Groupe d’Etude de la Matière Condens
ee GEMaC, CNRS (UMR 8635) Universit
e de
Versailles/Saint-Quentin, Universit
e Paris-Saclay, Versailles, France
†
PROMES, CNRS-UPR 8521, Universit
e de Perpignan Via Domitia, Perpignan, France
1
Corresponding author: e-mail address: [email protected]

Contents
1. Introduction 1
2. Magnetic Nanoparticle Assemblies: Theoretical Aspects 5
2.1 Model 6
2.2 Equilibrium Properties : Magnetization and Susceptibility 9
2.3 Dynamic Properties 17
3. Experimental Aspects 24
3.1 Magnetometry 25
3.2 AC Susceptibility 30
3.3 Magnetization Dynamics 33
3.4 Mo €ssbauer Spectroscopy 47
3.5 Neutron Scattering Experiments 56
4. Summary 85
References 90

1. INTRODUCTION
Investigating the properties of ensembles of magnetic nanoparticle is a
rich and challenging physics problem, from both the experimental and the-
oretical points of view. Indeed, one encounters the typical difficult situation
where intraparticle and interparticle effects meet into a formidable many-
body problem with both short-range and long-range interactions. The
intraparticle effects are related with the intrinsic properties of the
nanoparticles, such as the underlying material, size, shape, and energy poten-
tial. In particular, for small sizes the features of the single-nanoparticle physics
are dominated by finite-size and surface effects that drastically affect their

Solid State Physics, Volume 67 # 2016 Elsevier Inc. 1

ISSN 0081-1947 All rights reserved.
http://dx.doi.org/10.1016/bs.ssp.2016.08.001
2 D.S. Schmool and H. Kachkachi

magnetic properties, both in equilibrium and out of equilibrium. On the other

hand, assembled nanoparticles into 1D, 2D, or 3D arrays, organized or not,
reveal interesting and challenging issues related with their interactions among
themselves and with their hosting medium, a matrix or a substrate. The
ensuing collective effects show up through novel features in various measure-
ments, such as ferromagnetic resonance (FMR), AC susceptibility and
M€ ossbauer spectroscopy, to cite a few. Now, for assemblies of small particles
(
3–10 nm) one has to deal with the interplay between surface effects and
interparticle interactions whose study requires tremendous efforts. In addition,
during a few decades one had to struggle with at least two distributions,
namely that of the particles size and the anisotropy (effective) easy axes. Today,
the situation has improved owing to the huge progress in the production of
nearly monodisperse assemblies in well-organized patterns. This is one of the
reasons for which more theoretical works have appeared recently focusing on
such newly devised systems.
Needless to say that, already at equilibrium, no exact analytical treatment
of any kind is ever possible even in the one-spin approximation (OSP), i.e.,
ignoring the internal structure of the particles and thereby surface effects.
Only numerical approaches such as the Monte Carlo technique can alleviate
this frustration. Indeed, applications of this technique to the case of Ising
dipoles can be found in reference [1]. The same technique has been used
in reference [2] to study hysteretic properties of monodisperse assemblies
of nanoparticles with the more realistic Heisenberg spin model, within
the OSP approximation where each particle carries a net magnetic moment.
In reference [3], the Landau–Lifshitz thermodynamic perturbation theory
[4] is used to tackle the case of weakly dipolar-interacting monodisperse
assemblies of magnetic moments with uniformly or randomly distributed
anisotropy axes. The authors studied the influence of dipolar interactions
(DI) on the susceptibility and specific heat of the assembly. Today, the lit-
erature thrives with theoretical works on the effect of DI on the magnetic
properties of assemblies of nanoparticles, most of which make use of numer-
ical techniques [2, 5–25], because the main interest is for dense assemblies for
which experimental measurements are relatively easier to perform and the
applications more plausible. However, it is important to first build a fair
understanding of the underlying physics. This can only be done upon study-
ing model systems that are simple enough for performing analytical devel-
opments and still rich enough to capture the main qualitative features of the
targeted systems. Analytical expressions come very handy in that they allow
us to figure out what are the main relevant physical parameters and how the
Collective Effects in Assemblies of Magnetic Nanaparticles 3

physical observables of interest behave as the former are varied and the var-
ious contributions to the energy compete which other. A brief account of
our contribution will be given in the following section.
The magnetic properties of magnetic nanoparticles can be rather difficult
to measure, as we saw in the earlier chapter on single particle measurements,
where very specialized methods and adaptations are required [26]. To over-
come some of the problems with the weak experimental signals, many mea-
surements are made on assemblies of nanoparticles and elements. This means
that the results obtained are generally an average over the sample and assem-
bly and must also be interpreted taking into account the magnetic interac-
tions between the particles. There have been extensive studies using many
techniques. In the following, we aim to give a brief overview of selected
studies and techniques and will not be an exhaustive review. In particular,
we focus on well-known experimental techniques, which have been applied
to the study of nanoparticle systems.
Standard techniques, such as magnetometry and AC susceptibility, have
been applied to the study of magnetic nanoparticle systems. Measurements
can be made under the usual conditions since the material quantity is not an
issue, as stated previously. Where these techniques have shown to be of
importance is in the study of the superparamagnetic (SPM) behavior
observed in magnetic nanoparticle assemblies. This arises due to the thermal
instability introduced when the magnetic anisotropy, which usually defines
the orientation of the magnetization of the magnetic particle, is insufficient
to maintain its normal orientation. In fact the energy barrier is defined as the
product of the particles magnetic anisotropy constant K and its volume V.
Once the thermal energy is of the same order of magnitude as KV, the mag-
netization becomes unstable, switching spontaneously between the energy
minima of the system. As a result, the magnetic measurement, which has a
characteristic measurement time, will sample the magnetic state as being
(super)paramagnetic. A combination of measurements as a function of tem-
perature and applied field allows the system to be defined in terms of its
energy barrier and the blocking temperature TB, where the magnetization
is stable over the measurement time. Indeed, for AC susceptibility measure-
ments, a frequency dependence is also important. Indeed the average
switching time between magnetic easy axes is characterized as an attempt
frequency. For measurements made with lower characteristic measurement
time, such as M€ ossbauer spectroscopy and FMR, corresponding values of
the blocking temperature will be much higher due to the Arrhenius behav-
ior associated with superparamagnetism.
4 D.S. Schmool and H. Kachkachi

Ferromagnetic resonance is a very sensitive method for measuring the

magnetic properties of materials via the precessional magnetization dynamics
defined by the systems magnetic free energy. The precessional motion of
the magnetization is in general strongly influenced by magnetic anisotropies
and exchange effects in solids. This is often regarded as the internal effective
magnetic field experienced by the local magnetic spins of the system. This
can thus be separated into the various contributions to the local magnetic
field, via, magnetocrystalline anisotropy, shape anisotropy, exchange inter-
actions, etc. In magnetic nanosystems [26–29], this can be adapted to include
surface anisotropy effects as well as magnetic DI between particles. This will
produce shifts in the resonance fields and can significantly affect the
linewidth of resonance absorption lines. Once again, measurements as a
function of sample temperature can provide further information regarding
the magnetic behavior of nanoparticle assemblies as they move through dif-
ferent magnetic regimes.
Nuclear techniques provide another form of probe for the local magnetic
order in solids. When applied to magnetic nanoparticle systems, information
on the magnetic modifications at a magnetic surface can be established as can
the effects of interparticle interactions. One such technique is M€ ossbauer
spectroscopy, and this has been applied to many Fe-based nanoparticle sys-
tems. Temperature-dependent measurements provide a sensitive probe of
magnetic and SPM effects in these low-dimensional systems. It has been seen
to be particularly useful for the study of magnetic structures at the surface of
nanoparticles. M€ ossbauer spectroscopy has also been extensively used to
identify the oxide species which frequently form of metallic Fe and Fe oxide
nanoparticles. Neutron scattering is another nuclear technique which has
been broadly used as a research tool for investigating nanoparticles and mag-
netic nanoparticle assemblies. This for the most part concerns the scattering
at low angles from the incident neutron beam. Such small-angle neutron
scattering (SANS) has become a well-established technique in the study
of solids and biological samples. Here we consider how it can be applied
to provide information regarding the size and distribution of nanoparticles
in an ensemble. Indeed, information regarding the size and shape of samples
can be inferred from scattered intensity distributions. Using polarized neu-
trons allows magnetic information to be gleaned, which, as in the case of
M€ ossbauer spectroscopy, provides information on the surface of the mag-
netic particle and with care can be used to establish the spin distribution
or surface anisotropy of magnetic nanoparticles. Interparticle interactions
Collective Effects in Assemblies of Magnetic Nanaparticles 5

will also affect the magnetic scattering and thus SANS can also provide infor-
mation of magnetic interactions between the particles, where studies are fre-
quently performed as a function of particle concentration. Application of a
magnetic field to the sample is also used, where in systems of
magnetic nanoparticles dispersed in a solvent, or ferrofluid, the interaction
between the magnetic moments of the particles produces a spatial ordering
of the assembly. Core–shell models of magnetic nanoparticles can also be
established using a combination of SANS and polarized SANS measure-
ments, with and without applied magnetic fields.
In the following, we focus on some theoretical aspects related to the
treatment of assemblies of magnetic nanoparticles. This will discuss the
energy considerations for an ensemble of ferromagnetic nanoparticles,
where the individual particle energy is considered as well as the additional
energy contribution which arises from interparticle (dipolar) interactions.
This then allows the equilibrium state of the system to be evaluated and
the magnetization and susceptibility properties to be obtained. These con-
siderations are followed by a general discussion of dynamic magnetic prop-
erties and the AC susceptibility response of an assembly of weakly interacting
ferromagnetic nanoparticles. Section 3 aims to provide a brief overview of
experimental studies on magnetic nanoparticle assemblies. For each of the
methods discussed, we will give a short general introduction to the method,
where appropriate. We will cover both static and dynamic measurement
techniques.

2. MAGNETIC NANOPARTICLE ASSEMBLIES:

THEORETICAL ASPECTS
We have recently provided simple expressions for the magnetization
and susceptibility, both in equilibrium and out of equilibrium, which take
account of temperature, applied field, intrinsic properties, as well as (weak)
DI [11, 12, 21, 22, 30–35]. However, this has been done at the price of a few
simplifying assumptions, either related with the particles themselves or with
the embedding assembly. In particular, the study of the effect of DI, which is
based on perturbation theory, applies only to a dilute assembly with an inter-
particle separation thrice the mean diameter of the particles. In some cases,
we only considered monodisperse assemblies with oriented anisotropy axes.
For the calculation of the particle’s relaxation time, we only consider weak
fields, small core and surface anisotropies. A brief account of these works is
6 D.S. Schmool and H. Kachkachi

given in the following sections. For the study of interplay between surface-
dominated intrinsic properties and DI-dominated collective behavior, we
model a many-spin nanoparticle according to the effective-one-spin problem
(EOSP) proposed and studied in Refs. [34–39]. The EOSP model is a better
approximation than the OSP model in that it accounts for the intrinsic prop-
erties of the nanoparticle, such as the underlying lattice, size, and energy
parameters (exchange and anisotropy), via an effective energy potential.
In the simplest case, the latter contains a quadratic and a quartic contribu-
tions in the components of the particle’s net magnetic moment. These
two contributions should not be confused with the core and surface anisot-
ropy contributions. In fact, the effective model is a result of a competition
between several contributions to the energy, namely the spin–spin exchange
interaction inside the nanoparticle, the on-site anisotropy attributed to the
spins in the core and on the surface. The outcome of the various competitive
effects is an effective model for the net magnetic moment m of the nanopar-
ticle with a potential energy that contains terms with increasing order in its
components mα, α ¼ x, y, z. The coefficients of these terms are functions of
the atomic physical parameters, such as the constant of the on-site anisot-
ropies and exchange coupling, together with those pertaining to the under-
lying crystal structure.
In the following section, we will give a brief account of these theoretical
developments, related to the intrinsic, as well as collective features of the
nanoparticles. We will also discuss an excerpt of the main results they lead
to, for the magnetization and susceptibility.

2.1 Model
We will illustrate our theoretical developments in the simplest situation of a
monodisperse assembly and oriented anisotropy. More general situations of
polydisperse assemblies, with both oriented and random anisotropy, can be
found in the cited works, e.g., in Ref. [31]. We commence with a mono-
disperse assembly of N ferromagnetic nanoparticles carrying each a mag-
netic moment mi ¼ mi si , i ¼ 1,…, N of magnitude m and direction si,
with jsij ¼ 1. Each magnetic moment has a uniaxial easy axis e aligned along
the same z-direction. The energy of a magnetic moment mi interacting with
the whole assembly, and with a (uniform) magnetic field H ¼ Heh, reads
(after multiplying by  β 1/kBT)
ð0Þ
E i ¼ E i + E DI
i ,
(1)
Collective Effects in Assemblies of Magnetic Nanaparticles 7

ð0Þ
where the first contribution E i is the energy of the free (noninteracting)
nanocluster at site i, comprising the Zeeman energy and the anisotropy con-
tribution, i.e.
ð0Þ
E i ¼ xi si  eh + Aðsi Þ, (2)
where Aðsi Þ is a function that depends on the anisotropy model and is
given by
8
>
< σ i ðsi  ei Þ ,
2
OSP

 
Aðsi Þ ¼ ζ (3)
> 2
: σ i ðsi  ei Þ  s4i, x + s4i, y + s4i, z , EOSP:
2
The second term in Eq. (1) represents the DI between nanoclusters,
which can be written as
X
E DI
i ¼ξ si  Dij  sj (4)
j<i

where Dij is the usual DI tensor

1 
Dij  3 3eij eij  1 (5)
rij
with rij ¼ ri rj and eij ¼ rij/rij the unit vector along the i–j bond.
For later convenience, we have introduced the following dimensionless
parameters
mH K2 V K4
μ m2 =a3
x , σ , ζ , ξ¼ 0 (6)
kB T kB T K2 4π kB T


together with the DI parameter ξ  ξCð0, 0Þ , where

ð0, 0Þ 1
C ¼ 4π Dz  (7)
3
with Dz being the demagnetizing factor along the z-axis [3, 31]. K2 and K4
are the constants of the uniaxial and cubic anisotropy, respectively.
In the literature, especially for the experimental work, the magnetic
behavior of a nanoparticle is often approximated as a macrospin using
OSP or equivalently the Stoner–Wohlfarth model [40]. As discussed earlier,
in this approximation the spins of the ferromagnetic particle are considered
to be sufficiently well exchange-coupled that they move together in any
8 D.S. Schmool and H. Kachkachi

reorientation or reversal of the magnetization. In equilibrium the net mag-

netic moment is held in a direction according to the uniaxial anisotropy of
the magnetic nanoparticle, which in the absence of an applied magnetic field
will be along the easy axis. When a magnetic field is applied, the magneti-
zation will reorient in accord with the minimum of the following energy

Eðϑ,φÞ ¼ σ cos 2 ϑ + x cosðψ  ϑÞ (8)

where ϑ and ψ are the spherical angles between the easy axis and the par-
ticle’s magnetic moment and the applied magnetic field H, respectively.
Minimizing this simple energy, one finds two minima and a maximum with
two energy barriers expressed, for the case where the applied magnetic field
is parallel to the easy axis, as follows

H 2
△E ¼ σ 1  (9)
HK
where HK ¼ 2K2/Ms is the anisotropy field and corresponds to the highest
possible value of the switching field for the particle [41]. A full analysis of the
free energy of the system for a general orientation of the magnetic field with
respect to the easy axis yields a switching field which can be expressed as
HK
HSW ¼ 3=2
: (10)
ð sin 2=3 ψ + cos 2=3 ψ Þ
Again, as discussed earlier, the Stoner–Wohlfarthor OSP model is over
simplified since it ignores surface effects that arise from the reduction of
coordination at the boundaries of the magnetic system. In the literature
the anisotropy constant K2 used in HK is in fact an effective constant. Indeed,
experiments show that this model cannot account for the observed surface
effects caused by broken symmetries which significantly alter the local mag-
netic anisotropy [42, 43]. In an attempt to generalize the SW model in order
to take account of these observations, an effective anisotropy constant Keff is
derived on the basis of some arguments borrowed from 2D magnetism.
More precisely, Keff is proposed in a form that comprises uniaxial (volume)
and surface contributions, namely
KS
Keff ¼ KV + 6 (11)
D
where D is the particle diameter for a spherical particle [44]. In fact, it was
shown in Ref. [38] that Eq. (11) is only valid in elongated nanoparticles and a
more general formulation is proposed by the EOSP approach.
Collective Effects in Assemblies of Magnetic Nanaparticles 9

Then, for an assembly of randomly oriented magnetic noninteracting

OSP nanoparticles, the (angular) averaged Stoner–Wohlfarth model is used
to predict the reversal of the collective sample. However, many experiments
have shown that surface effects again have to be included and, as such the
SW model should be replaced by a more precise one. Accordingly, the effect
of finite size and surface anisotropy on the SW switching mechanisms have
been extensively studied in Ref. [45]. In particular, it was shown that for
weak surface anisotropy the spins inside the nanoparticle are nearly collinear
leading to a coherent switching of the particle’s magnetization. However, as
the surface effects become stronger, the spin switching operates via a cluster-
wise mechanism.

2.2 Equilibrium Properties : Magnetization and Susceptibility

In Ref. [31] it was shown that in a dilute assembly the magnetization of a
nanoparticle at site i (weakly) interacting with the other nanoparticles of
the assembly is given by (to first order in ξ)

XN
@ sz
szi ’ szi + ξik szk 0 Aki i 0 , (12)
0
k¼1
@xi

where Akl ¼ eh  Dkl  eh and hOi is the statistical average of the projection on
the field direction of the quantity O.
Note that Eq. (12) was obtained for an external magnetic field applied in
the z-direction leading to hsix, y i0 ¼ 0, and that this expression is only valid for
a center-to-center interparticle distance larger than thrice the mean diameter
of the nanoparticles[33]. This implies that the magnetization of an inter-
acting nanoparticle is written in terms of its “free” (with no DI) magnetiza-
tion szi 0 (and susceptibility @xi szi 0 ), with of course the contribution of the
assembly hosting matrix entering via the lattice sum in Eq. (12).
Ð Ð
Y 

The free-particle magnetization mði 0Þ  hszi i0 ¼ i
ds i sz
i exp βE
ð0Þ
i

can be either computed numerically or analytically in some limiting cases.

For the particular case of OSP, in a longitudinal magnetic field, i.e., ei
keh kez, the energy reads (dropping the particle’s index i)

E ð0Þ ¼ σs2z + xsz :

Then, one introduces the free-particle probability distribution

10 D.S. Schmool and H. Kachkachi

Z 1 Z 1
1 E ð0Þ ð0Þ E ð0Þ
P 0 ðzÞ ¼ ð0Þ
e , Zk ðσ, xÞ ¼ dsz e  dωð0Þ : (13)
Zk 1 1

ð0Þ
The free-particle partition function Zk is rewritten in terms of the
Z x
x2 2
Dawson integral DðxÞ ¼ e dt et as [46]
0

ð0Þ eσ pffiffiffiffiffiffi pffiffiffiffiffiffi

Zk ðσ,xÞ ¼ pffiffiffi ½ex Dð σ + Þ + ex Dð σ  Þ
σ

with the reduced field h ¼ x/2σ and energy barriers σ   σ ð1  hÞ2 .

Therefore, the free-particle magnetization in the presence of anisotropy
and a longitudinal magnetic field is given by
eσ
hsz i0 ðσ 6¼ 0,ξ ¼ 0Þ ¼ C1 ¼ ð0Þ
sinhx  h: (14)
σZk

There are various asymptotes that can then be derived for hszi0, see Refs.
[34, 46] for such developments.
Then, in the dilute limit, upon using Eq. (12) it is straightforward
to derive an expression for the magnetization of the assembly that takes
account of the DI. Furthermore, as it will be seen later on, an explicit expres-
sion for mð0Þ allows one to derive an approximate expression for the mag-
netization of a (weakly) interacting assembly of EOSP nanoparticles by
including the cubic anisotropy term with coefficient ζ. We recall, however,
that this applies for relatively weak surface anisotropy and thereby to an
equilibrium magnetic state with quasi-collinear spins.
Therefore, for monodisperse assemblies we have xi ¼ x, σ i ¼ σ, ξij ¼ ξ.
In this case, the magnetization of a (weakly) interacting particle, given by Eq.
(12), simplifies into the following expression
 ð0Þ

ð0Þ ð0, 0Þ @m
hs i ’ m 1 + ξC
z
: (15)
@x
This indicates that the relevant DI parameter, to this order of approximation,
is in fact the parameter introduced earlier

ξ X N
ξ  ξCð0, 0Þ ¼ Aij :
N i, j¼1, i6¼j
Collective Effects in Assemblies of Magnetic Nanaparticles 11

Note that the lattice sum Cð0, 0Þ is in fact the first of a hierarchy of lattice sums
(see Ref. [31]).
ð0Þ
Next, the longitudinal susceptibility χ k ¼ @mð0Þ =@x is given by (see,
e.g., Ref. [46] for the notation)

@mð0Þ 1 + 2S2
ð0Þ 2 ð1Þ
ð0Þ 2

¼  m ¼ a0  m
@x 3
where
8 l=2
>
> 2 ðl  1Þ!! l=2
< σ + ⋯ , σ ≪1,
ð2l + 1Þ!! ð1Þ 1 + 2S2
Sl ðσ Þ ’ , a0  : (16)
>
: 1  lðl + 1Þ + ⋯ ,
> 3
σ ≫1:
4σ
Consequently, we obtain the approximate expression for the magnetiza-
tion of a weakly interacting particle within the assembly
 
2 

ð1Þ
ð0Þ ð0Þ
hs i ’ m 1 + ξ a0  m
z
: (17)

Obviously, in the absence of any interaction and anisotropy, or at high

temperature (SPM regime), the magnetization is described, as usual, by the
Langevin function

μ HMS
hs i0 ðσ ¼ 0,ξ ¼ 0Þ ¼ L 0
z
, (18)
kB T
where μ0 is the vacuum permeability introduced so that μ0H is expressed in
tesla. LðxÞ ¼ cothx  1=x.
Now, we are ready to discuss a few examples of the main results obtained
with the help of these theoretical developments.
In Fig. 1, we present the field behavior of the magnetization for different
values of the anisotropy parameter σ, as rendered by the standard (equilib-
rium) Monte Carlo calculations [31, 47], for a noninteracting assembly
(ξ ¼ 0). In Fig. 1A, we see that in the high-field regime the higher is σ
the lower is the magnetization. This can easily be understood since randomly
distributed easy axes lead to randomly distributed equilibrium orientations of
the particles magnetic moments and thereby the projection on the field
direction of the assembly net magnetic moment decreases as the strength
of the anisotropy increases. However, in low fields this is not globally so,
12 D.S. Schmool and H. Kachkachi

Reduced magnetization (per particle)

1 1
A B
evin
Lang vin
nge
0.8 0.8 La

3 nm
0.6 0.6

s 7 nm
0.4 7 nm
17 0.4
34
68
135
T = 5K
0.2
0.2
Polydisperse assembly
Polydisperse assembly
Random anisotropy
Random anisotropy

0 10 20 30 40 0 10 20 30 40 x 50
x
Fig. 1 (A) Reduced magnetization (per particle) of an assembly of N ¼ 1024 with
lognormal-distributed magnetic moments with mean diameter Dm ¼ 7 nm and ran-
domly distributed easy axes as obtained from Monte Carlo calculations for different
anisotropy values. xm ¼ nmμBH/kBT, where nm is the mean number of Bohr magnetons
for this assembly. (B) Langevin function together with the Monte Carlo results for Dm ¼ 3
and 7 nm.

because the competition between Zeeman, thermal and anisotropy contri-

butions to the energy results in a crossing between the various magnetization
curves, as has been observed, e.g., for maghemite particles [48, 49]. In addi-
tion, we see that there is a large deviation from the Langevin law due to sev-
eral contributions to the energy, ignored by the Langevin law, especially
anisotropy. Moreover, the results in Fig. 1B show that the larger the mean
diameter of the assembly, the larger the value of σ, and thereby the larger the
expected deviation from the Langevin curve.
In Fig. 2 we plot the Langevin function (full line) and the Monte Carlo
results (symbols) for the magnetization of an interacting assembly of (N ¼
10  10  5) lognormal-distributed moments, with random anisotropy,
and for different values of the interparticle separation. Here we use the same
assemblies as in Fig. 1. The intensity of DI, or equivalently the value of ξ, is
varied by varying the lattice parameter a entering ξ [see Eq. (6)]. More pre-
cisely, the parameter a is taken as a real number k times the mean diameter
Dm of the assembly, i.e., a ¼ k  Dm. Thus, large values of k correspond to an
isotropically inflated lattice with large distances between the magnetic
Collective Effects in Assemblies of Magnetic Nanaparticles 13

Reduced magnetization
1.0

vin
nge
0.8 La
k=2
k=4
k=5
0.6 1.0

0.8

0.4 T = 5K 0.6

Polydisperse assembly
Random anisotropy 0.4
Symbols: MC
0.2 Low field
0.2 SDA (k = 2)

0.0
0.0 0.2 0.5 ζ 0.8
0.0
0.0 5.0 10.0 15.0 20.0
<x>
Fig. 2 Reduced magnetization (per particle) of an interacting assembly of N ¼ 10
10  5 lognormal-distributed magnetic moments with mean diameter Dm ¼ 7 nm
and random anisotropy. Monte Carlo in symbols and in lines the analytical expressions
(34) of Ref. [31]. In the inset, the parameters k is defined in the text while ζ ¼ x/ξ.

moments, and thereby weak DI. These results, obtained for an oblate sam-
ple, confirm the fact that in this case DI suppress the magnetization. This
result has also been obtained by perturbation theory in Ref. [34] whose
results are shown in Fig. 3, which are plots of Eq. (17) using Eq. (14) for mð0Þ .
As discussed in Ref. [31] and references therein, DI are anisotropic inter-
actions and thus contribute to the effective anisotropy. Since the anisotropy
is uniaxial and oriented, i.e., with a common easy axis, its effect leads to a
magnetization enhancement. In contrast, the DI effect depends on the sign


of ξ (or more precisely that of Cð0, 0Þ ), which is related to the sample’s shape.
For instance, in the case of oblate samples Cð0, 0Þ < 0 leading to a reduction of
the magnetization, while for prolate samples Cð0, 0Þ > 0 and thereby DI con-
tribute to enhance the assembly’s magnetization. Consequently, for oblate
samples the (oriented) uniaxial anisotropy and DI have opposite effects while
for prolate samples they play concomitant roles.
In the presence of not-too-strong surface anisotropy, one can model the
nanoparticle using the EOSP model upon which the free-particle partition
ð0Þ
function Zk is replaced by [34]
14 D.S. Schmool and H. Kachkachi

m
1

0.8
Without DDI
With DDI Oblate
0.6 With DDI prolate

0.4

0.2 D = 3 nm
T=5K

0
0 1 2 3 4 x 5
Fig. 3 Reduced magnetization of two assemblies of equivalent sizes but one is prolate
and the other oblate.

Z σζ
dφdωð0Þ e 2 Σα¼x, y, z sα :
4
Z¼

Then we assume that the cubic anisotropy remains small and proceed
with a perturbative calculation of Z. Indeed, the condition of validity for
the EOSP model [36–39] (obtained for a nanoparticle with an SC or
FCC crystal lattice) is roughly ζ ¼ K4 =K2 ≲1=4. As such, the spin non-
collinearities induced by surface anisotropy are not too strong and thereby
the anisotropy energy minima are mainly defined by the uniaxial contribu-
tion, whereas the cubic contribution only introduces saddle points. This
leads to larger relaxation rates [50] but does not affect the equilibrium
properties.
Upon performing a double expansion, with respect to x for low field and
to 1/σ for high anisotropy barriers, we obtain the following expression for
the magnetization for the EOSP particle (see Eq. (3.39) of Ref. [46] for the
case ζ ¼ 0 but arbitrary field)
     
ð0Þ 1 2 x3 ζ 2 5 x3
m ðx,σ,ζ Þ ’ 1  x  1  +  1 x+ 2 :
σ σ 3 σ σ σ 3
(19)

Next, writing this in the form

mð0Þ ’ χ ð1Þ x + χ ð3Þ x3

Collective Effects in Assemblies of Magnetic Nanaparticles 15

we can easily infer the EOSP corrections to the linear and cubic susceptibil-
ities (in the limit of a high anisotropy barrier) due to surface anisotropy of
intensity ζ
 
ð1Þ 1 ζ 2
χ ’ 1 + 1 + ,
σ σ σ
  
ð3Þ 1 2 ζ 5
χ ’ 1 + + 2 : (20)
3 σ σ σ
The competition between the uniaxial and cubic anisotropy contribu-
tions is easy to understand. As has been discussed earlier (see also Ref.
[21]), for ζ > 0 the energy minima of the cubic contribution are along
the cube diagonals ½1,  1,  1 while for ζ < 0 they are along the cube
edges ½1,0, 0, ½0,1,0, ½0,0,1. Hence, the uniaxial anisotropy with an easy
axis along the z-direction, i.e., ½0, 0,1, competes with the cubic anisotropy
when ζ > 0, whereas the two anisotropies have concomitant effects when
ζ < 0. In the former case, the particle’s magnetic moment at equilibrium
adopts an intermediate direction between the z-axis and the cube diagonal.
So, as ζ increases the magnetic moment gradually rotates away from the
z-axis and thereby its statistical average, or the magnetization, decreases.
In the case of negative ζ the two anisotropies cooperate to quickly drive
the magnetization toward saturation.
Next, using the expression (19) for the free-particle magnetization, as a
function of the applied field x, uniaxial anisotropy (and temperature) σ and
surface anisotropy ζ, in Eq. (12) or (15) we can investigate the interplay
between surface effects and DI, i.e., a competition between the terms in


ζ and ξ , respectively. This was done in Ref. [34]. The same competition
was also studied numerically in Ref. [21]. The outcome of this procedure
is the following approximate expression for the (average) magnetization
of a weakly interacting assembly of EOSP nanoparticles





m x, σ, ζ, ξ ’ χ ð1Þ x + χ ð3Þ x3 (21)

where
 

ð1Þ ð1Þ2
3 ζ
χ ’χ + ξ 1 2 1 ,
σ σ σ
  (22)

ð3Þ ð3Þ 4
3 3ζ
χ ’χ  ξ 1  ,
3 σ σ
16 D.S. Schmool and H. Kachkachi

are the linear and cubic susceptibilities (20) augmented by the DI contribu-


tion of intensity ξ .
This asymptotic expression helps understand how surface anisotropy
competes with DI. The surface contribution with intensity ζ, which plays
an important role in the magnetization curve, couples to the DI contribution



with intensity ξ via the term with coefficient ξ ζ. Hence, the overall sign of
the latter determines whether there is a competition between surface and DI
effects or if the changes in magnetization induced by the intrinsic and col-
lective contributions have the same tendency. Accordingly, plots of the
magnetization, which take into account both surface effects and DI, are
shown in Fig. 4 as a function of the field x, for an oblate sample with
Nx  Ny  Nz ¼ 20  20  5 and a prolate sample with 10  10  20,
with the respective values of Cð0, 0Þ ’ 4:0856 and 1.7293.


As discussed earlier, for oblate samples ξ < 0, DI tend to suppress mag-


netization, whereas for prolate samples ξ > 0 they enhance it. Indeed, we see
from Eqs. (22) that surface anisotropy and DI may have opposite or concom-
itant effects depending on their respective signs. In Ref. [11], it was found
that the magnetization enhancement in dilute assemblies of maghemite
nanoparticles of 3 nm in diameter is suppressed when the concentration
increases. In accordance with the present results, DI tends to smooth out
surface effects, or the other way round, the surface seems to have a screening
effect on DI.

m m

0.8 0.8

Without DDI, z = 0 Without DDI, z =0

0.6 Without DDI, z = 0.25 0.6 Without DDI, z = 0.25
With DDI, z = 0.25 With DDI, z = 0.25

0.4 0.4

20 × 20 × 5, D = 3 nm 10 × 10 × 20, D = 3 nm
0.2 0.2

0 0
0 1 2 3 4 x 5 0 1 2 3 4 x 5

Fig. 4 Left: magnetization as a function of the (dimensionless) field x for an oblate sam-
ple (20  20  5). Right: magnetization as a function of the reduced field x for a prolate
sample (10  10  5). Here ξ ’ 0.18.
Collective Effects in Assemblies of Magnetic Nanaparticles 17

2.3 Dynamic Properties

The dynamics of an assembly of magnetic nanoparticles is a rich environment
for the study of equilibrium and out-of-equilibrium many-body statistical
physics. Indeed, as discussed earlier, there are physical phenomena which
occur over wide ranges of spatial and temporal scales. The relevant length scale
can range from the Angstr€ om, through the nanometer to the millimeter, as we
go from the atoms inside of the nanocrystal, through the nanoparticle, to the
assembly thereof. On the other hand, the time scale also spans a wide range
that starts at the femtosecond timescale and ends with a duration of the order
of a few hours, as in relaxation phenomena observed in the isothermal and
thermoremanent magnetization. Obviously, these time scales are a direct
consequence of a competition between short-range and long-range interac-
tions operating at different length scales. For this reason, among others, it is not
possible to come up with a theory that covers all length and time scales. For
short-time regimes the physics is usually described with the aid of the Landau–
Lifshitz equation and its variants, deterministic or stochastic, damped or
undamped, local or macroscopic. For collective effects, occurring at the
assembly scale, the Monte Carlo technique is more appropriate, even though
the problem of an efficient algorithm for dynamical processes is not entirely
solved so far, see for instance the works in Refs. [51–56]. As for analytical
approaches, there are a very few attempts to tackle the problem, mainly
because of the tremendous difficulty to calculate the relaxation rate of a many-
spin system. The main difficulty resides in the fact that it is impossible to ana-
lyze the large number of extrema of a multivariate energy potential, in the
presence of several parameters, such as size, shape, applied fields, etc.
A way out of this difficulty was proposed in Ref. [36] where the EOSP model
was built for a spherical nanoparticle with N
eel anisotropy on the surface and
no anisotropy in the core, and in Refs. [38, 39], where it was extended to a
more general situation. Indeed, the EOSP approach makes it possible to inves-
tigate the dynamics of an interacting assembly while taking account of the
intrinsic features of the nanoparticles, since this model is a macroscopic model
whose energy potential depends on the nanoparticle’s parameters. This sim-
plification allows us to compute the relaxation time taking account of the
effect of surface anisotropy, in addition of course to that of the (effective) uni-
axial anisotropy and the applied (static) magnetic field. This was done in Ref.
[50]. Then, in Ref. [35] the AC susceptibility of a (weakly) interacting assem-
bly of EOSP nanoparticles was computed, after generalizing the calculation of
the relaxation rate of such particles.
18 D.S. Schmool and H. Kachkachi

The dynamic response of the EOSP assembly is given by the AC suscep-

tibility which, for an arbitrary angle ψ between the (common) easy axis
and the field direction, the effective susceptibility may be written as
χ ¼ χ k cos 2 ψ + χ ? sin 2 ψ. According to Debye’s model [33, 46, 57] we have

χ k ðT ,HÞ χ ðT, HÞ 2
χ ðωÞ ¼ cos 2 ψ + ? sin ψ, (23)
1 + iωτk 1 + iωτ?

where τk and τ? are the longitudinal (inter-well) and transverse (intra-well)

relaxation times and χ k(T, H) and χ ?(T, H) are, respectively, the longitudi-
nal and transverse components of the static susceptibility.
For an assembly with oriented anisotropy in a longitudinal field (ψ ¼ 0),
one assumes that the transverse response is instantaneous, i.e., τ?’ 0. In this
case the AC susceptibility is given by Eq. (23) or using τk ¼ Γ1 and
eq
eq
χ k ¼ χ eq ¼ χ free + ξ χ int ,


 χ eq
χ x,σ, ζ, ξ ,η ¼ : (24)
1 + iωΓ1
Next, we introduce the reduced frequency



η x,σ,ζ, ξ ,λ  ωτk ¼ ðωτD ÞðτD ΓÞ1 , (25)

with λ being the damping parameter. Γðx, σ, ζ, ξ , λÞ is the relaxation rate of
an EOSP nanocluster weakly interacting within the assembly. τD ¼
(λγHK)1 is the free diffusion time, HK ¼ 2K2V/M the (uniaxial) anisotropy
field, and γ ’ 1.76  1011 (Ts)1 the gyromagnetic ratio. For example, for
cobalt particles the anisotropy field is HK
0.3 T, and for λ ¼ 0.1  10, τD

2  1010  2  1012 s.
Now, if we restrict ourselves to the linear susceptibility, χ eq is equal to

ð1Þ
χ given in Eq. (22). The second quantity that needs to be calculated in
order to fully evaluate the susceptibility in Eq. (24) is the relaxation rate



Γ x, σ, ζ, ξ ,λ .
Accordingly, in Ref. [58], J€
onsson and Garcia-Palacios derived the fol-
lowing approximate expression for Γ for a weakly interacting assembly
 
1 D 2E 1
Γ ’ Γ0 1 + Ξk + FðαÞ Ξ? 0 : 2
(26)
2 0 4
Collective Effects in Assemblies of Magnetic Nanaparticles 19

D E
Ξ2k and Ξ2? are the spin averages of the longitudinal and transverse
0
0 P
components of the dipolar field Ξi ¼ ξ j6¼i Dij  sj . The subscript 0 is a
reminder of the fact that the averages are computed with the Gibbs distri-
bution of the noninteracting assembly [33]. The function F(α) is given
by [59]
1 1
FðαÞ ¼ 1 + 2ð2α2 eÞ1=ð2α Þ γð1 +
2
, Þ, (27)
2α2 2α2
Rz
with γða,zÞ ¼ 0 dt t a1 et , the incomplete gamma function, and where
pffiffiffi
α ¼ λ σ . Asymptotic expressions of F(α) are [59]
8 pffiffiffi pffiffiffi
>
> π 1 π
<  + α, α ≪ 1,
α 3 6
FðαÞ ’
>
>
:1 + 1  1 , α ≫ 1:
α 4α2
The free-particle relaxation rate Γ0 that was used in Ref. [58] is given by
2
τD Γ0 ¼ pffiffiffi σ 3=2 eσ : (28)
π
Then, the relaxation rate (28) was generalized in Ref. [35] in order to take
into account the magnetic field as well as the core and surface anisotropies.
For intermediate-to-high damping Langer’s approach allows us to com-
pute the relaxation rate Γ of a system with many degrees of freedom related
with its transition from a metastable state through a saddle point [60–65]


jκj Z s
Γ¼ , (29)
2π Z m


where Z m and Z s are, respectively, the partition functions in the vicinity of
the metastable energy minimum and the saddle point, obtained for a qua-
dratic expansion of the energy. The attempt frequency κ is computed upon
linearizing the dynamical equation around the saddle point, diagonalizing
the resulting matrix and selecting its negative eigenvalue [60, 61].
In Ref. [35] the relaxation rate Γ was calculated in various situations of
an EOSP particle including the effective uniaxial and cubic anisotropy and
the applied magnetic field. A detailed analysis of the various energy extrema
is presented in Ref. [35], and analytical expressions were given for the relax-
ation rate as a function of temperature, effective uniaxial anisotropy (σ),
20 D.S. Schmool and H. Kachkachi

surface anisotropy (ζ), and applied magnetic field. The authors of Ref. [35]
then investigated the interplay between interparticle DI and intrinsic surface
anisotropy, in the case ζ > 0 where surface (cubic) anisotropy favors the
magnetic alignment along the cube diagonals. χ 0 and χ 00 were computed
for various values of the surface anisotropy coefficient ζ, for both prolate
and oblate assemblies. Owing to the fact that the effect of increasing ζ is
to draw the particle’s magnetic moment toward the cube diagonals, it basi-
cally plays the same role in a prolate sample where the magnetization is
enhanced along the z-axis, or in an oblate sample where the magnetization
is enhanced in the xy plane.
The results in Fig. 5 show an example that illustrates the competition
between surface anisotropy and DI contribution to the real component of


the AC susceptibility. They were obtained for the finite value ξ ¼ 0:008
and an increasing (but small) surface anisotropy parameter ζ. It can be seen
that the surface anisotropy, in the present case of positive ζ, has the opposite
effect to that of DI. This again shows that there is a screening of DI by surface
effects and confirms the results of Ref. [34] for equilibrium properties for
both negative and positive ζ, as discussed earlier.
Our theoretical calculations of the AC susceptibility of magnetic
nanoparticles which accounts for the intrinsic properties (e.g., surface

∼
12 x = 0.008

10

8
c′

6 z=
0.01
4 0.05
0.1
2

0
0.1 0.15 0.2 0.25
1/s
Fig. 5 χ 0 for an interacting prolate ð10  10  20Þ assembly with a fixed DI strength


ξ ¼ 0:008 and varying surface anisotropy coefficient ζ, for the frequency


f  ωτD =ð2πÞ ¼ 0:01. h ¼ 0. Source: Reprinted figure with permission from F. Vernay,
Z. Sabsabi, H. Kachkachi, AC susceptibility of an assembly of nanomagnets: combined
effects of surface anisotropy and dipolar interactions, Phys. Rev. B 90 (2014) 094416.
Copyright (2009) by the American Physical Society.
Collective Effects in Assemblies of Magnetic Nanaparticles 21

effects) as well as the collective effects (due to DI) were then used [35] to
provide a microscopic derivation of the so-called Vogel–Fulcher law [see
also previous works [17, 66–70]]
ΔE
Γ ¼ τ1
0 e
kB ðT θVF Þ (30)

where ν0 ¼ τ1 0 ’ 10  10 Hz and θ VF represents an effective temperature

9 12

supposed to account for the DI correction; ΔE is the energy barrier, which

reads ΔE ¼ KV in the case of uniaxial anisotropy and zero field.
Our results are in full agreement with previous works [67, 68, 70] and
further extends them in that they take into account: (i) surface anisotropy,
(ii) the particles spatial distribution and shape of the assembly, and (iii) the
damping parameter. A full discussion can be found in Ref. [34]. Here we
only report the following expression found there for θVF
θVF ζ 1  2 
¼ + ξS (31)
T 4 6σ
where S is a function of the lattice and damping through the function F(α).
Expression (31) provides a somewhat microscopic description of the phe-
nomenological parameter θVF in terms of the interparticle interactions, the
surface anisotropy and damping. Indeed, the last term in (31), which is similar
to the one derived in Ref. [68], includes both the damping parameter and the
shape of the assembly, through the expression of S ðλÞ. In addition, we note
that ξ is proportional to the assembly concentration [34] CV and thereby to
a3, a being the interparticle separation. Therefore, we expect that in the
absence of surface anisotropy, θVF would scale as θVF
CV2
a6 . In Ref.
[17] experimental estimates of θVF are given for an assembly of Ni
nanoparticles with varying concentration. A comparison of Eq. (31) with
the corresponding experimental data is given in Fig. 6.
On the other hand, the first term in Eq. (31) accounts for the contribu-
tion from surface anisotropy. In practice it should be possible to adjust the
assembly characteristics (assembly shape, particles size and underlying mate-
rial) so as to achieve, to some extent, a compensation between surface effects
and the DI contribution. This could in principle suppress the dependence of
θVF on the assembly concentration. In addition, the term in ζ can also be
used to extract from the experimental data an estimate of the (effective) sur-
face anisotropy coefficient ζ by reading off the intercept from the plot in
Fig. 6. Furthermore, it is worthwhile emphasizing that θVF is not indepen-
dent of temperature, as is very often assumed in the literature. First, the
22 D.S. Schmool and H. Kachkachi

12

10 Data (Masunaga et al.)

Fit q VF = 0.5633 + 0.05405 Cv2
q VF(K) 8

0
0 2 4 6 8 10 12 14
Concentration Cv (%)
Fig. 6 θVF against the assembly concentration. Experimental data (stars) [17] and fit of
Eq. (31) (full line). Source: Reprinted figure with permission from F. Vernay, Z. Sabsabi,
H. Kachkachi, AC susceptibility of an assembly of nanomagnets: combined effects of surface
anisotropy and dipolar interactions, Phys. Rev. B 90 (2014) 094416. Copyright (2009) by the
American Physical Society.

temperature appears in the second term in (31), being related to the DI con-
tribution. Even if this term becomes negligible for very diluted assemblies, if
surface anisotropy is taken into account (ζ6¼0), e.g., for very small
nanoparticles, Eq. (31) shows that the phenomenological parameter θVF is
in fact a linear function of temperature via the term in ζ. This can be under-
stood by noting that the surface anisotropy, which is of cubic nature in the
EOSP model, drastically modifies the energy potential and thereby affects
the dynamics of the particle’s magnetization. As a consequence, the effect
of DI becomes strongly dependent on the thermal fluctuations and the ele-
mentary switching processes they induce.
Two applications of this formalism have been recently studied by one of
the authors, namely, on the one hand, the effect of DI on the FMR char-
acteristics of a 2D array of nanoparticles and, on the other, the effect of DI
and their competition with a DC magnetic field in the behavior of the spe-
cific absorption rate (SAR), which is relevant in magnetic hyperthermia.
The two corresponding works are in preparation and will be submitted
for publication elsewhere. In particular, the analytical expression of the
AC susceptibility obtained with the help of this formalism make it possible
to compute the SAR and study its behavior as a function of various param-
eters pertaining to the assembly. Indeed, it is quite easy to show that, in the
linear response, the SAR is proportional to the out-of-phase component χ 00
of the AC susceptibility.
Collective Effects in Assemblies of Magnetic Nanaparticles 23

We recall that in all of these developments involving DI, we have con-

sidered only dilute assemblies and as such we have used perturbation theory
to derive (semi-)analytically expressions for the magnetization, susceptibil-
ity, and relaxation rate that include the DI contribution. In the next section
on the experimental aspects, we will discuss the situation of more dense
assemblies.
As a final discussion it is worth addressing the issue of the relaxation rate
(29) and its use in the literature for modeling the dynamics of ensembles of
magnetic particles. In general, as can be seen in the above developments, Γ is
a function of various quantities, such as the anisotropy (core and surface), the
applied magnetic field, the DI, and so on. Its calculation has been performed
in various situations and limiting cases. However, in the experimental liter-
ature the relaxation relate is very often taken in the form of Arrhenius’ law
τD Γ0 ∝ eσ . The main reason evoked is that the behavior of Γ is dominated
by the exponential. Obviously, this is not quite so because this law ignores a
major physical phenomenon that is damping. Furthermore, if the (effective)
anisotropy is cubic, the switching mechanisms are rather different from those
of a uniaxial anisotropy.
This simple Arrhenius’ law and the ensuing simplifications are then used
to interpret, for example, the results for the hysteresis loop and, in particular,
for estimating the coercive field. More precisely, if the magnetic field is
applied parallel to the easy axis (ψ ¼ 0°), the energy minima become deeper
as the field is increased in this orientation. Then, Eq. (9) is used to obtain the
coercive field as a function of temperature
" rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi#
ln ðτm =τ0 Þ
HC ¼ H K 1  (32)
σ

where τ0 ’ 109  1010 s and τm is the measurement time and, as the name
suggests, depends on the method of measurement. For example, for a static-
like measurement such as VSM or SQUID magnetometry, this is about 102s,
while in FMR it is of the order of the inverse of the precessional frequency
and for M€ ossbauer is of the order of ns [72]. If the time window of an exper-
iment is shorter than the characteristic relaxation time at a fixed temperature,
the particles magnetization appears “blocked” in a particular state during the
measurement. However, when the measurement time is longer than the
relaxation time, the magnetization can alter state many times during
the experiment. This temperature-dependent behavior defines a tempera-
ture regime, for a specific type of measurement, below which the sample
24 D.S. Schmool and H. Kachkachi

appears ferromagnetic, while above it, it appears demagnetized and is said to

be SPM. Accordingly, within the same approximation, the so-called block-
ing temperature is introduced
KV 1
TB ¼ : (33)
kB ln ðτm =τ0 Þ
The description given here of nanoparticle systems can thus be charac-
terized as having two regimes; a low-temperature quasi-static or blocked
state below the blocking temperature, TB, which exhibits hysteresis, and
the high temperature (T > TB) where fast relaxation occurs. Rigorously, this
description only applies to an isolated nanoparticle modeled within the OSP
approach.
Obviously, in order to understand the role of surface anisotropy and its
interplay with the DI in an ensemble of magnetic nanoparticles, one has to
resort to the more general developments discussed earlier, related with the
behavior of the magnetization, the susceptibility, and the relaxation time.

3. EXPERIMENTAL ASPECTS
As discussed in Section 1, measurements on assemblies of
nanoparticles and elements are difficult to perform in general. In particular,
it is complicated, if not impossible, to disentangle the intrinsic properties of
the nanoparticles from their macroscopic properties related with their col-
lective behavior.
As outlined earlier, the proximity and intervening medium between
magnetic entities will affect the way they interact. Whatever the mechanism
of this interaction, the magnetic properties of the particles will deviate from
those of their isolated state. This adds to the already extrinsic nature of their
behavior caused by the reduced physical dimensions of the magnetic particles,
see the previous section. The coupling of magnetic objects will produce a col-
lective magnetic behavior of the ensemble of nanoparticles. Such a situation
can be considered as being produced by the effective field on an element of
the assembly due to the interaction fields of all the other elements. Indeed,
techniques, such as FMR, precisely measures this effective field.
Since an assembly of nanoparticles has no specific limit to the number
of particles measured, in addition to the methods discussed earlier, many
other techniques can be used to perform experimental studies. When
considering an ensemble of nanoparticles, the principal parameters that
characterize the sample will be particle shape, size (average) and size
Collective Effects in Assemblies of Magnetic Nanaparticles 25

distribution (mono or polydisperse), as well as the average particle separa-

tion. The matrix in which the assembly is suspended will also be of impor-
tance as interaction mechanisms depend on the intervening media between
the magnetic particles. For regular arrays of magnetic elements, we can fur-
ther consider the element shape and the periodicity and symmetry of the
nanostructure.

3.1 Magnetometry
Relaxation effects in assemblies of nanoparticles have been studied by var-
ious methods, such as DC magnetometry, AC susceptibility, as well via the
temperature-induced spontaneous magnetic noise. With conventional
magnetometry, the blocking temperature is frequently determined using
zero-field cooled (ZFC) and field cooled (FC) measurements as a function
of sample temperature. The ZFC magnetization, MZFC(T), is obtained by
heating the sample up to room temperature, then it is cooled in zero applied
field to low temperature such that the magnetic moments of the
nanoparticles are randomly oriented. Then a small magnetic field is applied
(typically of the order of 200 Oe or less) so that there is a measurable mag-
netization as the temperature is then increased. MZFC increases as the ther-
mal energy is raised, and there is sufficient energy for the particle to start
aligning with the applied field. However, as the temperature increases fur-
ther thermal fluctuations then effectively reduce the measured magnetiza-
tion. The MZFC(T) can be used to determine the blocking temperature
[71, 73]. For assemblies of nanoparticles with a size distribution (i.e., poly-
dispersion), which is frequently the case, only those particles with TB less
than the measuring temperature will contribute to the magnetization.
The size distribution typically follows a so-called log-normal distribution,
given as:
1
PðV Þ ¼ pffiffiffiffiffiffiffiffiffiffiffiffi ef½ lnðV =V0 Þ =2σ g
2 2
(34)
2πσV
where σ is the standard deviation of ln V and is related to the width of the
distribution
R and VR0 is the mean particle volume, evaluated from:
V0 ¼ VPðV ÞdV = PðV ÞdV . The log-normal distribution has a skewed
appearance as illustrated in Fig. 7.
The relative proportion of the unblocked particles which contribute to
the magnetization is proportional to MsV L(MsV H/kBT) [73, 74]. The
fitting of experimental data to this formula allows one to work backwards
26 D.S. Schmool and H. Kachkachi

Diameter d

Size fraction [%]

4

0
1 2 3 4 5 6 7
Fig. 7 Log-normal size distribution [27].

and extract the size distribution from the variation of MZFC(T) [75]. The FC
magnetization, MFC(T), is measured by initially applying a small magnetic
field to the sample at room temperature. The sample is then cooled and
the magnetization increases as the thermal fluctuations reduce. In contrast
with the ZFC magnetization, MFC(T) saturates at low temperature. The
blocking temperature is then identified as the deviation between MZFC(T)
and MFC(T), see Fig. 8.
In an isolated magnetic particle or in an assembly of identical non-
interacting particles, the magnetization will decay, due to thermal fluctua-
tions, following an exponential decay of the form:

MðtÞ ¼ M0 et=τ (35)

where M0 is the magnetization at time t0 [77]. In a an assembly with a size

distribution, there will be a subsequent distribution in the energy barrier,
which is compounded by variations in particle orientation and particle inter-
actions will further compound this distribution, broadening further the
range of relaxations times. The consideration of this problem led to the con-
cept of magnetic viscosity, S. The quantitative theory of relaxation was con-
sidered by N
eel, and is known as the N
eel model [78]. Further analysis by
Street and Woolley [79] and Barbara et al. [80] led to a magnetization
expressed as a function of time after changing the applied magnetic field as:
 Z
 
MðtÞ ¼ M0 1  exp tνeE=kB T nðEÞdE (36)
Collective Effects in Assemblies of Magnetic Nanaparticles 27

γ-Fe2O3 NP
25
ZFC
TB = 68 K FC
20
M (emu/g)

15

TB = 89 K H = 200 Oe
10

5 H = 100 Oe
TB = 101 K
H = 50 Oe
0
0 50 100 150 200 250 300
T (K)
Fig. 8 Zero-field cooled and field cooled magnetizations as a function of temperature
for γ-Fe2O3 nanoparticles of
7 nm diameter. It will be noted that the value of the block-
ing temperature TB is dependent on the applied magnetic field used in the measurement.
Source: Reprinted figure with permission from P. Dutta, A. Manivannan, M.S. Seehra, N. Shah,
G.P. Huffman, Magnetic properties of nearly defect-free maghemite nanocrystals, Phys. Rev.
B 70 (2004) 174428. Copyright (2004) by the American Physical Society.

where n(E)dE is the number of particles with an energy between E and E + dE.
For a smooth variation of n(E), it has been shown that [80]:
 B T ln ðtνÞ
MðtÞ ¼ M0 ½1  nðEÞk (37)
with E being the mean energy barrier height and then the magnetic viscosity
is expressed as:
1 dM  BT
S¼ ¼ nðEÞk (38)
M0 dð ln tÞ
While this analysis is still rather crude, it does provide a reasonable
approximation, especially over larger time scales, such as for magnetometry
measurements [81].
When interactions between nanoparticles in an assembly become very
strong, as would be the case for high concentration, then collective excita-
tions and states can be expected. This will greatly affect the magnetic
28 D.S. Schmool and H. Kachkachi

properties of the assembly as compared to low concentrations and when

interactions are negligible. The DI, for example in a system of randomly
oriented nanoparticles with a macrospin moment of μ ¼3000μB and a
center-to-center separation of R ¼ 6 nm, yields a dipolar energy of EDDI
¼ (μ0/4π)μ2/R3 [82]. With such large particle concentrations, a new mag-
netic regime can be identified which is characterized by the crossover from
single particle blocking (as described by the SPM state) to collective freezing
[72, 83, 84]. Two distinct collective states are possible. For intermediate
strength interactions, random particle spatial distribution and sufficiently
narrow size distribution, a superspin glass (SSG) state will exist. In this case
the superspins collectively freeze into a spin-glass phase below a critical
temperature, Tg [82, 85, 86]. With more concentrated systems and higher
interaction strengths a superferromagnetic (SFM) state is encountered. This
is characterized by ferromagnetic interparticle like correlations [87–89].
A consideration of the DI leads to an expected phase diagram for the order-
ing in a quasi-2D superspin system as obtained from a discontinuous
multilayer (DM) sample. This is illustrated in Fig. 9. For temperatures above

PM
Tc,bulk
Temperature

Tg,Tc
SPM

Tb

SFM
SSG

Nominal thickness
Fig. 9 Schematic phase diagram, transition temperature vs nominal thickness, with para-
magnetic (PM), superparamagnetic (SPM), superspin glass (SSG), and superferromagnetic
(SFM) phase. Relevant lines are the blocking temperature of the individual particles, Tb,
and the collective transition line, i.e., the glass transition, Tg, or SFM transition temperature,
Tc. Source: Reprinted from O. Petracic, X. Chen, S. Bedanta, W. Kleemann, S. Sahoo, S. Cardoso,
P.P. Freitas, Collective states of interacting ferromagnetic nanoparticles, J. Magn. Magn.
Mater. 300 (2006) 192–197. Copyright (2006), with permission from Elsevier.
Collective Effects in Assemblies of Magnetic Nanaparticles 29

the bulk Curie temperature, Tc, bulk, the system will be paramagnetic. Below
this temperature some form of spontaneous magnetic order will occur inside
each particle. Finite size effects can also affect the Curie temperature of a fer-
romagnet in an analogous way to that of the melting temperature of a metallic
nanoparticle. For low particle concentrations (small nominal thickness in the
DM system), the nanoparticle assembly will behave as an SPM. The effects of
any interparticle interactions for low concentrations are not significant since
blocking will disguise any transitions at low temperature. We can expect 3D
arrays of nanoparticles to behave in a similar manner.
As the concentration increases, interparticle or collective ordering can take
place where the ordering temperature will be greater than the blocking tem-
perature. For systems with random orientations and size distributions an SSG
phase will occur before and longer range ordering with a SFM phase. In terms
of magnetometry measurements, the SSG state can be observed in ZFC/FC
M vs T measurements. For the DM system of CoFe nanoparticles in an Al2O3
matrix, the experimental curve for the SSG state is shown in Fig. 10. Here the
usual peak in the magnetization is evident; however, a small minimum in
MFC(T) (as marked with an arrow) is observed for the SSG, which arises
from small paramagnetic clusters dispersed between the nanoparticles [82].

60

M FC

40
M (kA/m)

M ZFC

20

0
0 50 100
T (K)
Fig. 10 ZFC/FC magnetization vs temperature measurements in the DM [Co80Fe20
(0.9 nm)/Al2O3(3 nm)]10 measured in a field of arrow marks a dip in MFC being typical
of SSG systems. Source: Reprinted from O. Petracic, X. Chen, S. Bedanta, W. Kleemann,
S. Sahoo, S. Cardoso, P.P. Freitas, Collective states of interacting ferromagnetic
nanoparticles, J. Magn. Magn. Mater. 300 (2006) 192–197. Copyright (2006), with permis-
sion from Elsevier.
30 D.S. Schmool and H. Kachkachi

In addition to this the sample also exhibit a “memory effect,” where the
MZFC(T) curve shows a dependence on the waiting time at a temperature
below the blocking temperature [72, 83, 85, 90].
As the concentration of the nanoparticle assembly further increases, the
interaction strength between the particles will correspondingly grow. This
will eventually lead to a strong coupling and a fully collective behavior with
an SFM state. Such a system will be characterized by domain wall motion as
in a fully ferromagnetic state.

3.2 AC Susceptibility
Another measurement frequently employed in the study of magnetic nano-
particle assemblies is AC susceptibility. Theoretical aspects of this technique
were outlined in Section 2. One advantage of this method is that a static
magnetic field is not necessary to perform measurements. Ac susceptibility
measurements are usually taken in the frequency range below 100 kHz. For
nanoparticle systems, a peak in the imaginary component, χ 00 , of the AC sus-
ceptibility, is typically observed at the blocking temperature. We note that
the measurement time, being proportional to the inverse of the frequency
used in the experiment, will be significantly shorter than that used for
magnetometry, and hence will provide a larger blocking temperature, see
Eq. (33). The frequency dependence will highlight the time scale over
which the magnetization of the nanoparticle is stable. Since low magnetic
fields are used for the collection of AC susceptibility data, small rotations
of the magnetization can arise as well as thermally assisted reversal. For tem-
peratures in excess of the blocking temperature, χ 00 is small and χ 0 , the real
component of the AC susceptibility, will follow a Curie law: χ 0 ∝1=T,
indicative of paramagnetic-like behavior. The slope of 1/χ 0 vs temperature
allows the determination of the particle volume. Therefore, in polydisperse
systems the analysis is more complex, but careful fitting can also yield the size
distribution [91]. In Fig. 11A, the real and imaginary data for the AC sus-
ceptibility of Co0.1Cu0.9 alloys are shown as a function of temperature at
various frequencies. This nanosystem shows blocking behavior, however,
the effect of interactions is seen to be important since the Arrhenius behavior
yields unphysical values of the relaxation time, see Fig. 11B [92, 93].
The behavior of the blocking temperature depends on the concentration
of the nanoparticles in the assembly, where a monotonic decrease of the
blocking temperature was observed for iron-nitride nanoparticle systems
for higher concentrations [94]. At lower concentration sample showed a
Collective Effects in Assemblies of Magnetic Nanaparticles 31

A B
2.0 0.20 1
, 10 Hz
, 100 Hz
, 1 kHz 0.1
1.5 0.15 t 0 = 2.2*10−24s
c ⬘(emu/g × 10−3)

c ⬙(emu/g × 10−3)
, 10 kHz
Δ/kB = 709 K
0.01

t (s)
1.0 0.10
1E-3

0.5 0.05
1E-4

0.0 0.00 1E-5

0 50 100 150 200 0.060 0.065 0.070 0.075
T(K) 1/T (K−1)

Fig. 11 (A) Complex AC susceptibility measurements of granular Co0.1Cu0.9 alloys as a

function of temperature. Note that the χ 00 measurements are shown with filled symbols.
(B) Arrhenius plot from the χ 00 maxima in (A). Source: Reprinted from M.A. Novak, W.S.D.
Folly, J.P. Sinnecker, S. Soriano, Relaxation in magnetic nanostructures, J. Magn. Magn.
Mater. 294 (2005) 133–140. Copyright (2005), with permission from Elsevier.

much weaker dependence of TB on the applied field. However, other

researchers have found that the peak temperature of the susceptibility can
have a nonmonotonic dependence on the applied field [30, 95–97]. In
the dilute magnetic assembly of FePt nanoparticles, Zheng et al. obtain a
maximum in the peak temperature vs applied field which is due to the dis-
tribution of energy barrier sizes (and hence particle size) and the slow
decrease of the high-field magnetization above the blocking temperature
[98], see also Ref. [30]. The apparent anomaly is understood in terms of
the competition between the decrease in the moments of the SPM particles
and the increase in the moments of newly relaxed larger particles and hence
is related to the distribution of particle size [30]. The Langevin function is
used to describe the SPM behavior of unblocked particles which accounts
for the slower than expected decrease of the magnetization based on the
Curie law [30, 74].
Masunaga et al. have also performed detailed studies of the AC suscep-
tibility in Ni nanoparticles and in particular for NPs of 4–5 nm in diameter,
with mean separations of 14–21 nm [17]. The temperature-dependent mea-
surements for the real and imaginary components of χ ac are shown in Fig. 12.
A comparison of Figs. 5 and 12 shows a good qualitative agreement between
experiment and theory, where we note that 1/σ is proportional to the
sample temperature in Fig. 5. Again SPM behavior is evidenced by the fre-
quency independence of χ 0 at high temperature (T ≫ Tm0 ) while having a
frequency-dependent variation for (T ≪ Tm0 ), where the NPs are blocked
[72]. The frequency dependence can in principle also be attributed to spin
32 D.S. Schmool and H. Kachkachi

A T ⬘m 1.9% 1.5 C T ⬘m 12.8%

2
c⬘ (10−1 emu/g)

1.0

1
0.033 Hz
0.5
0.033 Hz 0.7 Hz
0.7 Hz 19 Hz
0.0
2
B T ⬙m D 12.8% T ⬙m
19 Hz 155 Hz
1.0
c⬙ (10−2 emu/g)

1895 Hz
295 Hz
9999 Hz
1895 Hz
1
9999 Hz 0.5

1.9%
0 0.0
0 20 40 60 0 20 40 60 80 100
T (K) T (K)
Fig. 12 Temperature dependence of the real and imaginary components of the AC sus-
ceptibility for samples of around 4 nm (1.9%) and 5 nm (12.8%) mean diameter, with
average separations of 21 and 14 nm, respectively. Real (A) and (C), and imaginary
(B) and (D) parts of the AC magnetic susceptibility for selected frequencies. Source:
Reprinted figure with permission from S.H. Masunaga, R.F. Jardim, P.F.P. Fichtner, J. Rivas,
Phys. Rev. B 80 (2009) 184428. Copyright (2009) by the American Physical Society.

glass-like behavior, arising from strong DI between the randomly oriented

particles. The analysis of the χ ac data for a range of samples of different con-
centrations gives rise to the Arrhenius–N
eel plots shown in Fig. 13. In this
figure, we note that the straight line plots are taken from the Arrhenius equa-
tion of the form, see Eq. (28),

τ ¼ τ0 e△E=kB T (39)

where △E corresponds to the energy barrier between minima, of an ampli-

tude defined by the anisotropy strength. Also shown is the correction to this
relation, which accounts for interparticle interactions, e.g., via dipolar
coupling, and is considered as a temperature shift T0. This relationship is
known as the Vogel–Fulcher law, see Eq. (30), which can also be expressed
in the form:

τ ¼ τ0 e△E=kB ðT T0 Þ (40)

Collective Effects in Assemblies of Magnetic Nanaparticles 33

In t 1.9%
−5
2.7%
4.0%
−10
7.9%
12.8%
−15
0.04 0.08 0.12
−1 (10−3 K−1)
T ⬙m
Fig. 13 Plots of ln τ vs 1/Tm00 for samples with 1.9, 2.7, 4.0, 7.9, and 12.8 wt% Ni. Dotted
lines are fitted to the N
eel–Arrhenius law given by Eq. (39) and arrows indicate a clear
deviation of the fitting in the limit of high frequencies. Solid lines represent the best fit of
the Vogel–Fulcher law given by Eq. (30). Source: Reprinted figure with permission from
S.H. Masunaga, R.F. Jardim, P.F.P. Fichtner, J. Rivas, Phys. Rev. B 80 (2009) 184428. Copyright
(2009) by the American Physical Society.

As discussed in Section 2.3, Vernay et al. [35] interpreted the results based
on a consideration of the surface anisotropy of the particles and the inter-
particle interactions, leading to a good agreement with experiments, see
Fig. 6.
The AC susceptibility has also been experimentally used to distinguish
between SPM, SSG, and SFM behavior in DM nanoparticle samples using
Cole–Cole plots. These display somewhat different aspects for the different
interparticle coupling regimes [82, 99].

3.3 Magnetization Dynamics

At much higher frequencies, we enter the regime of FMR, where many
studies have been performed on ferromagnetic nanoparticle assemblies.
The basis of FMR theory can be expressed from the dynamical motion
of the magnetization vector at the maximum angle of precession, where
the equation of motion is expressed in the form of the Landau–Lifshitz
equation
@M
¼ γðM ^ Heff Þ (41)
@t
where for simplicity we have ignored the damping term, see for example
Refs. [28, 100–102]. In Eq. (41), we note that the effective field, Heff, is
34 D.S. Schmool and H. Kachkachi

composed of the applied external magnetic field as well as contributions

which take account of internal effective fields due to magnetic anisotropies
and the exchange field. This can also include the dipolar field due to mag-
netic interactions between the particles of an assembly. In this case the dipo-
lar field must take into account the spatial distribution of the nanoparticles. It
is often useful to define the relation between the effective magnetic field and
the free energy of the system. This can be expressed as:
1
Heff ¼  ru E (42)
M
where E is the free energy density of the magnetic system and u ¼ M/M
defines the unit vector in the direction of the magnetization. The resonance
condition can be derived from the Landau–Lifshitz equation, which can be
expressed in the form of the Smit–Beljers equation as (a formal derivation of
this equation can be found in a number of texts, see for example Refs. [28,
100, 101]):
(   2 )
2 1=2
ω 1 @2E @2E @ E
¼  (43)
γ Ms sin ϑ @ϑ2 @φ2 @ϑ@φ

where ϑ and φ express the polar and azimuthal angles which define the ori-
entation of the magnetization and γ is the gyromagnetic ratio. It is important
to note that in order to use this equation, the equilibrium conditions must be
known, and are obtained from the minimization of the free energy with
respect to ϑ and φ. Contributions to the free energy will depend on the mag-
netic sample under consideration. In FMR, the Zeeman energy will always
be a principle component due to static and dynamic (microwave) magnetic
field that are required. Additional contributions will also be required and are
typically due to magnetostatic (or shape) energy and magneto-crystalline
anisotropies.
As we have seen, in systems of nanoparticles we generally need to take
into account the interparticle interactions. For a nonmetallic matrix, it is
usually sufficient to consider the DI. This can then be included in the free
energy of the system. Netzelmann [103] introduced the idea of separating
the magnetostatic energy into the particle demagnetization term and the
sample demagnetization term, where later corrections from Dubowik
[104] and Kakazei et al. [105] give the magnetostatic energy:
1 $ 1 $
EMS ¼ ρð1  ρÞM  N P  M + ρ2 M  N S  M (44)
2 2
Collective Effects in Assemblies of Magnetic Nanaparticles 35

$
where N p, s represents the demagnetization tensor of the particle ( p) and
sample (s), respectively, and ρ the volume fraction of magnetic nanoparticles
in the assembly, which can be specified as:
XN
Vm Vi
ρ¼ ¼ i¼1 (45)
Vs Vs

where we introduce the total volume of the sample, Vs. In the case of poly-
disperse systems we can also write Vm ¼ N hV i, with N being the total num-
ber of particles in the assembly and hV i is the average particle volume. It is
also possible to include the direct energy of the DI as [29, 106, 107]:

mi mj 
Eij ðrij Þ¼ 1  3½ sinθ sin ϑ cosðϕ  φÞ + cos θ cosϑ2 (46)
4πrij

here θ and ϕ refer to the polar and azimuthal angles of the vector, rij, between
the particles at positions i and j. This approach has been successfully applied
to planar granular media as well as 3D arrays of particles [106–108]. For the
case of DMs, the interactions can be evaluated both in the plane and between
adjacent planes. This leads to a resonance equation of the form [108]:
 2
ω sinθ cosðθ  ϑ0 Þ 2 sinθ cos 2ϑ
¼ Ha  2C Ha (47)
γ sinϑ0 sinϑ0

where θ is the polar angle of the applied magnetic field, Ha, and ϑ0 is the
equilibrium polar orientation of the magnetization, defined by the equilib-
rium condition:

C sin 2ϑ0 ¼ Ha sinðθ  ϑ0 Þ (48)

The constant C depends on sample material constants: magnetization, vol-

ume fraction, and the particle shape factor, given by

  1  IP 
C ¼ 2πρMs ð1  ρÞ Njj  N? + nΓ  ðn  1ÞΓOP (49)
M

here Njj and N? are the demagnetizing factors of the nanoparticle in the par-
allel and perpendicular orientations, respectively, ΓIP and ΓOP denote the
in-plane and out-of-plane averaged DI [108].
36 D.S. Schmool and H. Kachkachi

A 6 B 8
7
5 6
Hres (kOe)

Hres (kOe)
5
4
4
3 3
2
2 1
0 30 60 90 0 30 60 90
Angle (degrees) Angle (degrees)
C D 11

9 10
9
8
Hres (kOe)

Hres (kOe)
7 7
6
5 5
4
3 3
2
1 1
0 30 60 90 0 30 60 90
Angle (degrees) Angle (degrees)
Fig. 14 Angular variation of the resonance field from out-of-plane (0°) to in-plane (90°)
for (A) t ¼ 7 Å, (B) t ¼ 9 Å, (C) t ¼ 11 Å, and (D) t ¼ 13 Å. Points correspond to exper-
imental data and lines are theoretical fits. Source: Reprinted from D.S. Schmool, R. Rocha,
J.B. Sousa, J.A.M. Santos, G. Kakazei, J.S. Garitaonandia, L. Lezama, The role of dipolar inter-
actions in magnetic nanoparticles: ferromagnetic resonance in discontinuous magnetic
multilayers, J. Appl. Phys. 101 (2007) 103907, with the permission of AIP Publishing.

In Fig. 14 the angular variation of the resonance field for a discontinuous

CoFe/Al2O3 multilayer system illustrates the agreement with theory for dif-
ferent effective thicknesses of the nanogranular CoFe layer. It will be noted
that the difference in the in-plane to out-of-plane resonance field increases as
the effective thickness increases, meaning that the particles are on average
larger and closer together. This will increase the mean DI strength. We
observed small discrepancies for the smaller particles at angles intermediate
between the parallel and perpendicular configurations, which may arise
from deviations of the magnetization from the saturated state due to surface
anisotropy effects.
The separation of in-plane and out-of-plane interactions was also con-
sidered by Kakazei et al. [109, 110] and Majchrák et al. [111, 112]. In the
former, the authors study the transition from the continuous to the discon-
tinuous regime by studying the variation of the effective magnetic layer
Collective Effects in Assemblies of Magnetic Nanaparticles 37

(CoFe) thickness, which changes the average particle size and separation.
The effective field was obtained using a Kittel analysis of the FMR, as
expressed by:

2KV 4KS
Heff ¼ 4πMs   (50)
Ms Ms t

where KV, S denote the surface (N
eel) and volume anisotropies.

Temperature-dependent measurements of the FMR were also studied in
the discontinuous magnetic multilayer system [113]. The resonance field as a
function of the temperature for these samples is shown in Fig. 15A. Here we
see that the resonance field, Hres, reduces with T, from 300 to 100 K, in a
manner expected from classical FMR behavior of ferromagnetic materials.
At approximately 90 K there is a significant enhancement of Hres, which is
very marked for lowest effective thicknesses. In Fig. 15B we plot Henh against
1/t, where Henh ¼ Hres  Hres
expt expt
and Hres is the resonance field expected from
a normal classical dependence, extrapolating the high temperature trend to
lower temperatures. Henh(1/t) shows a linear dependence indicating that this
enhancement may have its origin in surface anisotropy, which scales as 1/t. It
is clear that the surface contribution should grow with decreasing particle
size, since the surface area to volume ratio increases. Therefore, it appears
that, although the resonance equation used at room temperature is valid,
for lower temperatures, <90 K, we need to include an extra term
corresponding to surface anisotropy to account for the effective field. From
Fig. 15A, it appears that the enhancement vanishes at 15 K.

A 3000 B 1000
2750
t=7Å 800
2500
2250
Henh (Oe)

t=9Å
Hres (Oe)

600
2000
1750 t = 11 400
1500
1250 200
1000 t = 13 Å
750 0
0 50 100 150 200 250 300 0.07 0.09 0.11 0.13 0.15
Temperature (K) 1/t (Å−1)

Fig. 15 (A) Temperature dependence of the resonance field for the [Al2O3 (40 Å)/
Co80Fe20 (t)]10/Al2O3 (30 Å) discontinuous multilayers. (B) Maximum resonance field
enhancement observed as a function of inverse thickness for the discontinuous multi-
layers. Source: Reprinted from D.S. Schmool, R. Rocha, J.B. Sousa, J.A.M. Santos, G. Kakazei,
Evidence of surface anisotropy in magnetic nanoparticles, J. Magn. Magn. Mater. 300 (2006)
e331. Copyright (2006), with permission from Elsevier.
38 D.S. Schmool and H. Kachkachi

In many studies of FMR in three-dimensional assemblies of nanoparticles,

the nanoparticle assembly is fully random in three dimensions and typically
comprises the effects of polydispersion, random orientations of particles and
SPM effects. A majority of experimental studies by FMR appear to have been
performed on γ-Fe2O3 and Co-based nanoparticle assemblies. One of the
oldest studies was performed by Valstyn et al. [114] on γ-Fe2O3, and more
recent studies of this ferrite have been reported by Hseih et al. [115], Dutta
et al. [76], Noginova et al. [116], Schmool and Schmalzl [106, 107], and
Ortega et al. [117]. Typical of these studies are the spectra presented by Dutta
et al. [76], see Fig. 16, which show the temperature-dependent spectra and
comparisons with bulk γ-Fe2O3 measurements. While in Fig. 17, we show
the temperature dependence of the resonance lines for these measurements.
For higher temperatures, the two resonance lines tend to merge producing
strongly overlapping lines. Such results are typical of NP assemblies and it

A B
g-Fe2O3 NP T = 300 K

f = 9.28 GHz f = 9.28 GHz

198 K
Bulk g-Fe2O3

132 K

g-Fe2O3 NP
98 K

46 K

Suspended g-Fe2O3 NP
14 K
A B

50 2550 5050 7550 10050 50 2550 5050 7550 10050

Magnetic field (Oe) Magnetic field (Oe)
Fig. 16 Absorption derivative FMR spectra of (A) γ-Fe2O3 NP at several temperatures
and (B) bulk γ-Fe2O3, γ-Fe2O3, and γ-Fe2O3 suspension in ethanol at room temperature.
The magnetic field scan starts from 50 Oe (rather than zero), because for H < 50 Oe the
field of the electromagnet is not stable. Source: Reprinted figure with permission from
P. Dutta, A. Manivannan, M.S. Seehra, N. Shah, G.P. Huffman, Magnetic properties of nearly
defect-free maghemite nanocrystals, Phys. Rev. B 70 (2004) 174428. Copyright (2004) by the
American Physical Society.
Collective Effects in Assemblies of Magnetic Nanaparticles 39

4200

3900

3600 High-field line B (γ-Fe2O3 NP)

3300

3000
Hr (Oe)

Bulk γ-Fe2O3
2700

2400

2100

1800 Low-field line A (γ-Fe2O3 NP)

1500 F = 9.28 GHz

0 50 100 150 200 250 300

T (K)
Fig. 17 Temperature variations of the resonance field Hr for the two EMR lines of
γ-Fe2O3 NP and for the single line of bulk γ-Fe2O. Source: Reprinted figure with permission
from P. Dutta, A. Manivannan, M.S. Seehra, N. Shah, G.P. Huffman, Magnetic properties of
nearly defect-free maghemite nanocrystals, Phys. Rev. B 70 (2004) 174428. Copyright (2004)
by the American Physical Society.

is frequently necessary to use a fitting procedure to separate the various con-

tributions. The low-field line arises from particles with their major axis aligned
parallel to the applied field, while the high-field line is due to the particles with
their major axis perpendicular to the applied field.
Using the effects of DI between nanoparticles, Schmool and Schmalzl
[109] have explained the angular dependence of FMR using Eq. (47) where
for this case of a 3D nanoparticle assembly the C constant takes the form:

π hr i3 Ms π
C¼ Vmag ¼ hr i3 Ms V (51)
6ρ 6
ρ is the volume fraction of particles which we define as; ρ ¼ Vmag/V, V being
the total volume of the sample, and hr i is the average particle radius. For
nonspherical particles this constant will have an additional term related to
the shape anisotropy, but is not considered here. The angular dependence
is shown in Fig. 18, which shows a good agreement between experiment
40 D.S. Schmool and H. Kachkachi

3200

3100
Resonance field (Oe)

3000

2900

2800

2700

2600

2500
0 20 40 60 80 100
Angle (degrees)
Fig. 18 Angular variation of the resonance field for rectangular lamina samples. Points
refer to experimental data while line is a fit to resonance, Eq. (47). Source: Reprinted from
D.S. Schmool, M. Schmalzl, Ferromagnetic resonance in magnetic nanoparticle assemblies,
J. Non-Cryst. Sol. 353 (2007) 738. Copyright (2007), with permission from Elsevier.

and theory (solid line). Further measurements at low temperature indicate

variations in the anisotropy constant, which varies from 0.238  10 5 J/m3
at room temperature to 3.034  10 5 J/m3 at 5 K. The temperature variation
of the resonance is a reflection of the variation of magnetization M in the SPM
regime, which can be expressed using a weighted Langevin function:
Z 
HMVmag
M ¼ Ms L PðV ÞdV (52)
kB T

P(V) represents the log-normal distribution (34). The comparison of the

Langevin function with the experimental data is shown in Fig. 19 for these
γ-Fe2O3 samples with hDi ¼ 4:6 nm.
Monodisperse FCC Co arrays were studied by FMR with in-plane (azi-
muthal) and out-of-plane (polar) angular measurements to the effective
magnetization and in-plane anisotropy field by Spasova et al. [118]. Regular
arrays of Co particles of about 12 nm were obtained by drying a solution of
the NP in an applied field of 0.35 T on a grid. The resulting assembly con-
sisted of stripes of regular triangular Co nanocrystals with a width of around
200–250 nm. The lowest resonance field was obtained when the external
field was applied along the direction of the stripes; (Hres)min ¼ 0.233 T which
is lower than the EPR field of ω/γ ¼ 0.3085 T, showing that an additional
Collective Effects in Assemblies of Magnetic Nanaparticles 41

1
T=5K
0.8 T = 100 K

0.6
M(×)/Ms

0.4
T = 295 K
0.2

0
0.01 0.1 1 10 100
log (x)
Fig. 19 Experimental data points at the measured temperatures with the Langevin
function (line). Source: Reprinted from D.S. Schmool, M. Schmalzl, Ferromagnetic resonance
in magnetic nanoparticle assemblies, J. Non-Cryst. Sol. 353 (2007) 738. Copyright (2007),
with permission from Elsevier.

intrinsic magnetic field due to an effective magnetization and an easy-axis

magnetization in the film plane is evident. Three considerations are made
to the interpretation of the resonance field: (i) shape anisotropy due to stri-
pes, (ii) magnetic anisotropy due to interparticle magnetostatic coupling in
an FCC-like lattice inside the stripes, and (iii) the effective magnetic anisot-
ropy of the individual particles (including shape, volume, and surface anisot-
ropy contributions). Assuming, as a first approximation, that all anisotropies
due to spin–orbit coupling vanish, with only that for shape being present, a
resonance equation is obtained [100]:
 2
ω       
¼ Hres + Nx  Nz 4πMs Hres + Ny  Nz 4πMs (53)
γ
This function is not sufficient to explain the angular dependence of the
FMR and further contributions are required. By introducing a cubic and
uniaxial symmetry, the in-plane resonance field could be obtained as:
 2 h i
ω
¼ Hres + 2Han
4jj
cos4φ  Han 2jj
cos 2ðφ  ϕÞ
γ
h   i (54)
 Hres + Heff + 2Han 2  sin 2φ  Han cos ðφ  ϕÞ
4jj 2 2jj 2

where ϕ is the angle of the applied field with the axis of the uniaxial anisot-
2jj 4jj
ropy, Han and Han are the effective uniaxial and fourfold anisotropy fields,
42 D.S. Schmool and H. Kachkachi

and the effective field is given by Heff ¼ 2K2?/Ms + 4πρMs, where ρ is the
volume fraction. The fit, shown in Fig. 20A, shows good agreement with
experiment where the following values were obtained: Han 2jj
¼ 0.037 T
and Han ¼ 0 and Heff ¼ 0.127 T, that is, only uniaxial anisotropy is observed.
4jj

For the polar dependence the following resonance equation was used:
 2
ω  
¼ Hres cos ðϑ  θÞ + Heff cos2ϑ
γ (55)
h i
 Hres cos ðϑ  θÞ  Heff cos ϑ + Han
2 2jj

The fit, Fig. 20B, yields Heff ¼ 0.13 T and Han

2jj
¼ 0.037 T, which is in good
agreement with the previous fit and experiment. The difference between the
expected 4πρMs ¼ 0.222 T for ρ ¼ 0.31 and Heff is accounted for using the
perpendicular anisotropy field 2K2?/Ms and/or the possible existence of an
antiferromagnetic CoO outer layer which would reduce Ms. Such core–shell
Co-CoO nanoparticles have been further studied by Wiedwald et al. [119]
using FMR and X-ray magnetic circular dichroism (XMCD) techniques to
study the ratio of orbital-to-spin magnetic moment. Essentially, XMCD
eff
yields a value of μL =μS ¼ 0:24  0:06, which is over three times the value
obtained from g-factor analysis of g ¼ 2.15  0.015 corresponding to
eff
μL =μS ¼ 0:075  0:008. The difference is explained as being due to the pres-
ence of uncompensated Co magnetic moments at Co-CoO core–shell inter-
face. High-resolution TEM corroborates the existence of the CoO shell.

A B
0.30 0.45

0.40
0.28
0.35
m 0HR (T)

(T)
0HR

0.26 0.30

0.25
0.24

0.20
−135° −90° −45° 0° 45° 90° 135° −90° −60° −30° 0° 30° 60° 90°
jH qH

Fig. 20 Dependence of the resonance field on the direction of the external magnetic
field: (A) in-plane ϕ and (B) out-of-plane θ (measured from the normal to the sample).
Source: Reprinted from M. Spasova, U. Wiedwald, R. Ramchal, M. Farle, M. Hilgendorff,
M. Giersig, Magnetic properties of arrays of interacting Co nanocrystals, J. Magn. Magn.
Mater. 240 (2002) 40–43. Copyright (2002), with permission from Elsevier.
Collective Effects in Assemblies of Magnetic Nanaparticles 43

Ferromagnetic resonance measurements in 1D assemblies of quasi-

monodisperse (20  1.8 nm) iron oxide nanoparticle chains, prepared
using “click” chemistry, have illustrated that such structures have an
induced uniaxial anisotropy arising from DI between individual particles
[120]. Since the 1D chains have a preferential alignment, defined by the
direction of an applied magnetic field during the nanoparticle deposition
on a substrate, the combined magnetic anisotropies of the chains will pro-
duce a cumulative effective anisotropy. This is manifest as a uniaxial var-
iation of the resonance field as a function of the applied field orientation in
the FMR experiment. A similar result was observed by Charilaou et al.
[121]. The strength of the effective anisotropy field is dependent on the
overall nanoparticle distribution, including the nanoparticle chain
length, average separation between chains and variation in chain align-
ment. It is noted that interchain DI will act to attenuate the uniaxial mag-
netic anisotropy intrinsic to the chains themselves [120]. These general
observations are confirmed by other researchers [122, 123] and is also
supported by FMR experiments and theory, as outlined earlier, as for
example in Schmool et al. [108]. The complex nature of FMR spectra
in the 1D chain assemblies, particularly those with relatively low chain
densities, implies that there is more to the magnetic behavior than a simple
homogeneous uniform resonance with uniaxial anisotropy. A more com-
plex distribution of the internal magnetic field seems to likely and prob-
ably arise from the different magnetic environments of the nanoparticles
in the chains; particles at the ends of the chains probably experience dif-
ferent effective internal fields, for example. More complex collective
excitation modes could also play a role in the dynamic magnetic properties
of the nanoparticle chains.
The existence of oxide shells in nanoparticle systems is fairly common
in pure metallic nanoparticle systems, such as Fe and Co. This is demon-
strated in Fig. 21 on Fe nanocubes, where a plasma treatment is used to
remove the oxide shell in vacuum, with FMR being performed in situ,
Trunova et al. [124]. The explanation of the lineshapes is illustrated in
Fig. 22 and essentially arises from a consideration of the distribution of par-
ticle axes and summing the various contributions in the final spectrum.
The fit shows excellent agreement with the model. Tomita et al. [125] have
studied the FMR in Fe nanogranular films, where Fe NPs are dispersed in
an SiO2 matrix with variable concentrations. For concentrations of 5% no
angular dependence was observed for the FMR, while at 15% a shift of
around 600 Oe at 9.1 GHz is observed. This could be attributed to shape
44 D.S. Schmool and H. Kachkachi

Fig. 21 Conventional FMR spectra of iron/iron oxide core–shell cubes and oxide free Fe
cubes after H-plasma treatment and covered by Ag/Pt matrix. Source: Reprinted from A.
V. Trunova, R. Meckenstock, I. Barsukov, C. Hassel, O. Margeat, M. Spasova, J. Lindner, M.
Farle, Magnetic characterization of iron nanocubes, J. Appl. Phys. 104 (2008) 093904, with
the permission of AIP Publishing.

A
B0
Hard axis
dc ″/dB [a.u.]

Easy axis

Out of plane
dc ″/dB [a.u.]

In plane

0.1 0.2 0.3 0.4 0.5 0.6

B [Tesla]
Fig. 22 (A) Theoretical simulation of crystalline anisotropy distribution from easy to
hard in-plane direction (shown as gray cubes). The black curve is the sum of all colored
resonance modes. (B) Theoretical fits (lines) to the in-plane and out-of-plane experi-
mental FMR spectra (symbols). Source: Reprinted from A.V. Trunova, R. Meckenstock, I.
Barsukov, C. Hassel, O. Margeat, M. Spasova, J. Lindner, M. Farle, Magnetic characterization
of iron nanocubes, J. Appl. Phys. 104 (2008) 093904, with the permission of AIP Publishing.
Collective Effects in Assemblies of Magnetic Nanaparticles 45

effects or more likely to interparticle interactions, which would be

expected to be significant for the more concentrated sample. The temper-
ature dependence of the FMR in these two samples was also studied, where
for the 5% sample no significant variation was observed as opposed to the
15% sample, which for the perpendicular resonance showed an increase in
value with reduced temperatures, while the parallel resonance shows a
small downward shift with decreasing temperature. The magnetic perco-
lation is expected with increased concentration, and the authors use the
Kittel equations to study the more concentrated sample. The shift in the
FMR for this sample with decreasing temperature fits to an increase in
4πMs as measured by SQUID. This could imply that, since the particles
in the two concentrations are of the same type, the stronger interparticle
DI act to stabilize the magnetic behavior, which causes a shift in the block-
ing temperature of the nanoparticle assembly, as suggest by the Vogel–
Fulcher law.
The temperature dependence of the magnetic anisotropy is seen to be an
important parameter in determining the magnetic behavior of nanoparticle
assemblies. In the study of Antoniak et al. [126], FMR and SQUID data are
jointly used in the evaluation of τ0 and Keff in FePt nanoparticles from the
two difference blocking temperatures for these two methods. From the anal-
ysis of the FMR using the Kittel equation:

ω 2
 
¼ ½Hres cos ðϑ  θÞ + HA cos 2ϑ Hres cos ðϑ  θÞ  HA cos 2 ϑ (56)
γ

HA ¼ 2Keff /Ms is the effective anisotropy field, which is the uniaxial con-
tribution due to small deviations from the spherical shape, as well as surface
and step anisotropies at the particle surface which are not averaged out.
Averaging over the angles θH of the external magnetic field gives the numer-
ical relation:
"  #
1:25 0:44
HA
Hres ¼ Hres
0
1 0
(57)
Hres

0
here Hres ¼ ℏω=gμ0 μB and g is the g-factor, which is obtained as 2.054 
0.010 from frequency-dependent measurements. Since the intensity of
the FMR line is proportional to the magnetization, the blocking tempera-
ture was evaluated analyzing the intensity vs temperature. This will give a
higher value than SQUID measurements since the time windows for
the two methods are very different; τFMR 1010 s and τSQUID 102 s.
46 D.S. Schmool and H. Kachkachi

By comparing the two blocking temperatures, see Figs. 23 and 24A, and
using the Arrhenius relationship, Eq. (39), the effective anisotropy constant
can be written:
" #1
 SQUID  27kB 1 α
Keff TB  FMR (58)
Vm TBSQUID hTB i

SQUID
Fraction

TBFMR
FMR
10 nm

0 100 200 300 400

TBSQUID
TB (K)
Fig. 23 Distribution of blocking temperatures for two different time windows. The frac-
tion of particles was calculated for the time windows τSQUID 102 s, τFMR 1010 s using
the size distribution determined from TEM image analyses. The inset shows a typical
TEM image of the Fe70Pt30 nanoparticles. Source: Reprinted from C. Antoniak, J. Lindner,
M. Farle, Magnetic anisotropy and its temperature dependence in iron-rich FexPt1x
nanoparticles, Europhys. Lett. 70 (2005) 250, http://iopscience.iop.org/journal/0295-5075,
with the permission of EDP Sciences.

A B
10
0.5 1.0
FMR-Intensity (arb. units)

8
0.4 0.8
mtot (10−6 Am2)

Keff (105 J/m3)

6
0.3 0.6

0.2 0.4 4

0.1 0.2 2

0 0 0
0 100 200 300 400 0 100 200 300 400
T (K) T (K)

Fig. 24 (A) ZFC measurements of the total magnetic moment using the SQUID (open
circles) and FMR (full circles) technique. (B) Temperature-dependent anisotropy constant
of Fe70Pt30 nanoparticles. Source: Reprinted from C. Antoniak, J. Lindner, M. Farle, Mag-
netic anisotropy and its temperature dependence in iron-rich FexPt1x nanoparticles, Euro-
phys. Lett. 70 (2005) 250, http://iopscience.iop.org/journal/0295-5075, with the permission
of EDP Sciences.
Collective Effects in Assemblies of Magnetic Nanaparticles 47

       
where α ¼ Keff TBFMR =Keff TBSQUID ¼ HA TBFMR =HA TBSQUID ,
for which a temperature-dependent Keff, α ¼ 1 and Vm is the mean volume.
A small deviation of TB from Tmax arises from a distribution of sizes. From
the FMR data TBFMR 110 K, this is about five times that obtained from
the magnetization measurements. The anisotropy constant is shown in
Fig. 24B as a function of temperature, with α ¼ 0.8 and using Eq. (58), Keff
¼ (8.4  0.9)  105 J/m3 and from Eq. (39) τ0 1.7  1012 s. The exper-
imental values of Keff are found to follow a Bloch law-like dependence, with a
h i2:1
power of 2.1; that is Keff ∝ Ms ð1  T=TB Þ3=2 [127]. The anisotropy was
found to be about an order of magnitude higher than in the bulk.

€ ssbauer Spectroscopy
3.4 Mo
M€ ossbauer spectroscopy relies on the recoil-free nuclear resonant absorption
of gamma radiation from an energy specific source to a specific absorbing
atom in a solid. It thus provides a very sensitive local probe of the immediate
environment around the absorbing atom. More specifically, it probes tiny
changes in the energy levels of an atomic nucleus in response to its environ-
ment. Typically, three types of nuclear interactions may be observed: an iso-
meric shift, also known as a chemical shift; quadrupole splitting; and
magnetic or hyperfine splitting, also known as the Zeeman effect. It is this
latter that provides local information on the magnetic state or environment
of the absorbing species. The high energy and extremely narrow linewidths
for gamma rays in the emission process makes this spectroscopy probably
the most sensitive in terms of energy resolution and is capable of detecting
changes of around 1 part in 1011[128, 129]. A very large majority of
M€ ossbauer measurements are based on the emission of gamma radiation
from a 57Co source (radioactive parent), typically in a Rh matrix and the
absorption of this radiation by the 57Fe atom in solids. As such, M€ ossbauer
spectroscopy is almost exclusively used in the study of Fe and commonly in
thin film or nanoparticle form, as well as in alloys and compounds
[128–131]. Since57Fe is the absorbing species, but is not the most abundant
isotope, frequently studies are performed by implanting 57Fe enriched atoms
in the sample.
The M€ ossbauer effect was discovered in 1957 and was quickly employed
to perform spectroscopic measurements by exploiting the Doppler effect,
which opens up an energy window in which the spectroscopy is applied.
Another random document with
no related content on Scribd:
 The Project Gutenberg eBook of Iamblichus
on the mysteries of the Egyptians, Chaldeans,
and Assyrians
 This ebook is for the use of anyone anywhere in the United
States and most other parts of the world at no cost and with
almost no restrictions whatsoever. You may copy it, give it away
or re-use it under the terms of the Project Gutenberg License
included with this ebook or online at www.gutenberg.org. If you
are not located in the United States, you will have to check the
laws of the country where you are located before using this
eBook.

Title: Iamblichus on the mysteries of the Egyptians, Chaldeans,

and Assyrians

Author: Iamblichus

Contributor: Porphyry

Translator: Thomas Taylor

Release date: January 29, 2024 [eBook #72815]

Language: English

Original publication: London: Bertram Dobell & Reeves and

Turner, 1895

Credits: Richard Tonsing, MFR, and the Online Distributed

Proofreading Team at https://www.pgdp.net (This file
was produced from images generously made available
by The Internet Archive)

*** START OF THE PROJECT GUTENBERG EBOOK IAMBLICHUS ON

THE MYSTERIES OF THE EGYPTIANS, CHALDEANS, AND ASSYRIANS
***
Transcriber’s Note:
New original cover art included with this eBook is
granted to the public domain.
 IAMBLICHUS

ON

The Mysteries

OF THE

EGYPTIANS, CHALDEANS, AND

ASSYRIANS.

TRANSLATED FROM THE GREEK

BY

THOMAS TAYLOR.

Ο δε Αριστοτελης προς Αντιπατρον περι Αλεξανδρου γραφων, εφη μη

μονον εκεινῳ προσηκειν οτι πολλων κρατει μεγαφρονειν, αλλ’ ουδεν
ηττον ει τις ορθως γινωσκει περι θεων.
 Plutarch.

Second Edition.

LONDON:
BERTRAM DOBELL,
77 CHARING CROSS ROAD, W.C.
AND
REEVES AND TURNER,
5 WELLINGTON STREET, STRAND.
MDCCCXCV.
 ADVERTISEMENT.

The various translations and original works of Thomas Taylor,

though still in request by the more zealous students of ancient
philosophy and occult science, have now become so scarce and
expensive that it is only within the power of comparatively wealthy
collectors to obtain them. This is a matter for regret, inasmuch as it
cannot be affirmed that his writings have been, or are likely to be
superseded, or that they are without value. They can hardly be
neglected without loss by those who desire to understand the
systems of philosophy which satisfied the spiritual needs of the
antique world. It is not possible, even for the most fervent believer in
modern “progress,” to dismiss the speculations of the ancient
philosophers as antiquated notions which have had their day and no
longer possess interest or value. The names of Socrates, Plato, and
Aristotle can never grow dim with age, nor is it possible to conceive a
time when men shall cease to study and reverence them. As the
disciple, the translator, and the expounder of these and of other
sages of antiquity, Thomas Taylor deserves to be held in honour and
remembrance, and it would be a misfortune if his labours remained
unknown because of the scarcity of his books. It is for this reason
that the present reprint has been undertaken; and it is hoped that it
will meet with such a measure of success as may encourage the
republication of various other works by the same author. It has been
printed in handsome style and published at a moderate price in
order that it may be regarded as a desirable addition to the scholar’s
library, while yet it will not tax severely the means of the not too
wealthy student. For the rest it is only necessary to say that this
reprint is, in size, number of pages, type, and general get-up, an
almost exact facsimile of the original edition, which was first printed
in 1821. No alterations or additions have been made in or to the
original text, as it is thought that those who care for Taylor’s writings
will prefer to have them in their integrity. Should it be found
possible, however, to continue the series it is intended to prefix to a
future volume an essay on Taylor, which will contain a biography of
him, and a critical estimate of his writings.

May, 1895.
 INTRODUCTION.

It appears to me that there are two descriptions of persons by

whom the present work must be considered to be of inestimable
worth, the lovers of antiquity and the lovers of ancient philosophy
and religion. To the former of these it must be invaluable, because it
is replete with information derived from the wise men of the
Chaldeans, the prophets of the Egyptians, the dogmas of the
Assyrians, and the ancient pillars of Hermes; and to the latter,
because of the doctrines contained in it, some of which originated
from the Hermaic pillars, were known by Pythagoras and Plato, and
were the sources of their philosophy; and others are profoundly
theological, and unfold the mysteries of ancient religion with an
admirable conciseness of diction, and an inimitable vigour and
elegance of conception. To which also may be added, as the colophon
of excellence, that it is the most copious, the clearest, and the most
satisfactory defence extant of genuine ancient theology.
This theology, the sacred operations pertaining to which called
theurgy are here developed, has for the most part, since the
destruction of it, been surveyed only in its corruptions among
barbarous nations, or during the decline and fall of the Roman
empire, with which, overwhelmed with pollution, it gradually fell,
and at length totally vanished from what is called the polished part of
the globe. This will be evident to the intelligent reader from the
following remarks, which are an epitome of what has been elsewhere
more largely discussed by me on this subject, and which also
demonstrate the religion of the Chaldeans, Egyptians, and Greeks to
be no less scientific than sublime.
In the first place, this theology celebrates the immense principle of
things as something superior even to being itself; as exempt from the
whole of things, of which it is nevertheless ineffably the source; and
does not, therefore, think fit to enumerate it with any triad[1] or order
of beings. Indeed it even apologizes for giving the appellation of the
most simple of our conceptions to that which is beyond all
knowledge and all conception. It denominates this principle
however, the one and the good; by the former of these names
indicating its transcendent simplicity, and by the latter its
subsistence as the object of desire to all beings. For all things desire
good. At the same time, however, it asserts that these appellations
are in reality nothing more than the parturitions of the soul, which,
standing as it were in the vestibules of the adytum of deity, announce
nothing pertaining to the ineffable, but only indicate her
spontaneous tendencies towards it, and belong rather to the
immediate offspring of the first God than to the first itself. Hence, as
the result of this most venerable conception of the supreme, when it
ventures not only to denominate it, though ineffable, but also to
assert something of its relation to other things, it considers this as
preeminently its peculiarity, that it is the principle of principles; it
being necessary that the characteristic property of principle, after the
same manner as other things, should not begin from multitude, but
should be collected into one monad as a summit, and which is the
principle of all principles.
The scientific reasoning from which this dogma is deduced is the
following. As the principle of all things is the one, it is necessary that
the progression of beings should be continued, and that no vacuum
should intervene either in incorporeal or corporeal natures. It is also
necessary that every thing which has a natural progression should
proceed through similitude. In consequence of this, it is likewise
necessary that every producing principle should generate a number
of the same order with itself, viz. nature, a natural number; soul, one
that is psychical (i. e. belonging to soul); and intellect an intellectual
number. For if whatever possesses a power of generating, generates
similars prior to dissimilars, every cause must deliver its own form
and characteristic peculiarity to its progeny; and before it generates
that which gives subsistence to progressions, far distant and separate
from its nature, it must constitute things proximate to itself
according to essence, and conjoined with it through similitude. It is,
therefore, necessary from these premises, since there is one unity,
the principle of the universe, that this unity should produce from
itself, prior to every thing else, a multitude of natures characterized
by unity, and a number the most of all things allied to its cause; and
these natures are no other than the Gods.
According to this theology, therefore, from the immense principle
of principles, in which all things causally subsist, absorbed in
superessential light, and involved in unfathomable depths, a
beauteous progeny of principles proceed, all largely partaking of the
ineffable, all stamped with the occult characters of deity, all
possessing an overflowing fulness of good. From these dazzling
summits, these ineffable blossoms, these divine propagations, being,
life, intellect, soul, nature, and body depend; monads suspended
from unities, deified natures proceeding from deities. Each of these
monads, too, is the leader of a series which extends from itself to the
last of things, and which, while it proceeds from, at the same time
abides in, and returns to, its leader. And all these principles, and all
their progeny, are finally centred and rooted by their summits in the
first great all-comprehending one. Thus all beings proceed from, and
are comprehended in, the first being: all intellects emanate from one
first intellect; all souls from one first soul; all natures blossom from
one first nature; and all bodies proceed from the vital and luminous
body of the world. And, lastly, all these great monads are
comprehended in the first one, from which both they and all their
depending series are unfolded into light. Hence this first one is truly
the unity of unities, the monad of monads, the principle of
principles, the God of Gods, one and all things, and yet one prior to
all.
No objections of any weight, no arguments but such as are
sophistical, can be urged against this most sublime theory, which is
so congenial to the unperverted conceptions of the human mind, that
it can only be treated with ridicule and contempt in degraded,
barren, and barbarous ages. Ignorance and impious fraud, however,
have hitherto conspired to defame those inestimable works[2] in
which this and many other grand and important dogmas can alone
be found; and the theology of the ancients has been attacked with all
the insane fury of ecclesiastical zeal, and all the imbecile flashes of
mistaken wit, by men whose conceptions on the subject, like those of
a man between sleeping and waking, have been turbid and wild,
phantastic and confused, preposterous and vain.
 Indeed, that after the great incomprehensible cause of all, a divine
multitude subsists, cooperating with this cause in the production and
government of the universe, has always been, and is still, admitted by
all nations and all religions, however much they may differ in their
opinions respecting the nature of the subordinate deities, and the
veneration which is to be paid to them by man; and however
barbarous the conceptions of some nations on this subject may be,
when compared with those of others. Hence, says the elegant
Maximus Tyrius, “You will see one according law and assertion in all
the earth, that there is one God, the king and father of all things, and
many Gods, sons of God, ruling together with him. This the Greek
says, and the Barbarian says, the inhabitant of the continent, and he
who dwells near the sea, the wise and the unwise. And if you proceed
as far as to the utmost shores of the ocean, there also there are Gods,
rising very near to some, and setting very near to others.”[3]
The deification, however, of dead men, and the worshiping men as
Gods, formed no part of this theology, when it is considered
according to its genuine purity. Numerous instances of the truth of
this might be adduced, but I shall mention for this purpose, as
unexceptionable witnesses, the writings of Plato, the Golden
Pythagoric Verses,[4] and the Treatise of Plutarch on Isis and Osiris.
All the works of Plato, indeed, evince the truth of this position, but
this is particularly manifest from his Laws. The Golden verses order
that the immortal Gods be honoured first, as they are disposed by
law; afterwards the illustrious Heroes, under which appellation the
author of the verses comprehends also angels and dæmons, properly
so called; and in the last place, the terrestrial dæmons, i. e. such good
men as transcend in virtue the rest of mankind. But to honour the
Gods as they are disposed by law, is, as Hierocles observes, to
reverence them as they are arranged by their demiurgus and father;
and this is to honour them as beings not only superior to man, but
also to dæmons and angels. Hence, to honour men, however
excellent they may be, as Gods, is not to honour the Gods according
to the rank in which they are placed by their Creator; for it is
confounding the divine with the human nature, and is thus acting
directly contrary to the Pythagoric precept. Plutarch too, in his above
mentioned treatise, most forcibly and clearly shows the impiety of
worshiping men as Gods.[5]
 “So great an apprehension indeed,” says Dr. Stillingfleet,[6] “had
the Heathens of the necessity of appropriate acts of divine worship,
that some of them have chosen to die, rather than to give them to
what they did not believe to be God. We have a remarkable story to
this purpose in Arrian and Curtius[7] concerning Callisthenes.
Alexander arriving at that degree of vanity as to desire to have divine
worship given him, and the matter being started out of design among
the courtiers, either by Anaxarchus, as Arrian, or Cleo the Sicilian, as
Curtius says; and the way of doing it proposed, viz. by incense and
prostration; Callisthenes vehemently opposed it, as that which
would confound the difference of human and divine worship, which
had been preserved inviolable among them. The worship of the
Gods had been kept up in temples, with altars, and images, and
sacrifices, and hymns, and prostrations, and such like; but it is by no
means fitting, says he, for us to confound these things, either by
lifting up men to the honours of the Gods, or depressing the Gods to
the honours of men. For if Alexander would not suffer any man to
usurp his royal dignity by the votes of men; how much more justly
may the Gods disdain for any man to take their honours to himself.
And it appears by Plutarch,[8] that the Greeks thought it a mean and
base thing for any of them, when sent on any embassy to the kings of
Persia, to prostrate themselves before them, because this was only
allowed among them in divine adoration. Therefore, says he, when
Pelopidas and Ismenias were sent to Artaxerxes, Pelopidas did
nothing unworthy, but Ismenias let fall his ring to the ground, and
stooping for that, was thought to make his adoration; which was
altogether as good a shift as the Jesuits advising the crucifix to be
held in the mandarin’s hands while they made their adorations in the
Heathen temples in China.
Conon[9] also refused to make his adoration, as a disgrace to his
city; and Isocrates[10] accuses the Persians for doing it, because
herein they showed that they despised the Gods rather than men, by
prostituting their honours to their princes. Herodotus mentions
Sperchies and Bulis, who could not with the greatest violence be
brought to give adoration to Xerxes, became it was against the law
of their country to give divine honour to men.[11] And Valerius
Maximus[12] says, “the Athenians put Timagoras to death for doing
it; so strong an apprehension had possessed them, that the manner
of worship which they used to their Gods, should be preserved sacred
and inviolable.” The philosopher Sallust also, in his Treatise on the
Gods and the World, says, “It is not unreasonable to suppose that
impiety is a species of punishment, and that those who have had a
knowledge of the Gods, and yet despised them, will in another life be
deprived of this knowledge. And it is requisite to make the
punishment of those who have honoured their kings as Gods to
consist in being expelled from the Gods.”[13]
When the ineffable transcendency of the first God, which was
considered as the grand principle in the Heathen religion by the best
theologists of all nations, and particularly by its most illustrious
promulgators, Orpheus, Pythagoras, and Plato, was forgotten, this
oblivion was doubtless the principal cause of dead men being deified
by the Pagans. Had they properly directed their attention to this
transcendency they would have perceived it to be so immense as to
surpass eternity, infinity, self-subsistence, and even essence itself,
and that these in reality belong to those venerable natures which are,
as it were, first unfolded into light from the unfathomable depths of
that truly mystic unknown, about which all knowledge is refunded
into ignorance. For, as Simplicius justly observes, “It is requisite that
he who ascends to the principle of things should investigate whether
it is possible there can be any thing better than the supposed
principle; and if something more excellent is found, the same inquiry
should again be made respecting that, till we arrive at the highest
conceptions, than which we have no longer any more venerable. Nor
should we stop in our ascent till we find this to be the case. For there
is no occasion to fear that our progression will be through an
unsubstantial void, by conceiving something about the first
principles which is greater and more transcendent than their nature.
For it is not possible for our conceptions to take such a mighty leap
as to equal, and much less to pass beyond, the dignity of the first
principles of things.” He adds, “This, therefore, is one and the best
extension [of the soul] to [the highest] God, and is, as much as
possible, irreprehensible; viz. to know firmly, that by ascribing to
him the most venerable excellences we can conceive, and the most
holy and primary names and things, we ascribe nothing to him which
is suitable to his dignity. It is sufficient, however, to procure our
pardon [for the attempt], that we can attribute to him nothing
superior.”[14] If it is not possible, therefore, to form any ideas equal to
the dignity of the immediate progeny of the ineffable, i. e. of the first
principles of things, how much less can our conceptions reach that
thrice unknown darkness, in the reverential language of the
Egyptians,[15] which is even beyond these? Had the Heathens,
therefore, considered as they ought this transcendency of the
supreme God, they would never have presumed to equalize the
human with the divine nature, and consequently would never have
worshiped men as Gods. Their theology, however, is not to be
accused as the cause of this impiety, but their forgetfulness of the
sublimest of its dogmas, and the confusion with which this oblivion
was necessarily attended.
But to return to the present work. To some who are conversant
with the writings of Porphyry, who know how high he ranks among
the best of the Platonists, and that he was denominated by them, on
account of his excellence, the philosopher, it may seem strange that
he should have been so unskilled in theological mysteries, and so
ignorant of the characteristics of the beings superior to man, as by
his epistle to Anebo he may appear to have been. That he was not,
however, in reality thus unskilful and ignorant, is evident from his
admirable Treatise on Abstinence from Animal Food, and his
Αφορμαι προς τα νοητα, or Auxiliaries to Intelligibles. His apparent
ignorance, therefore, must have been assumed for the purpose of
obtaining a more perfect and copious solution of the doubts
proposed in his Epistle, than he would otherwise have received. But
at the same time that this is admitted, it must also be observed, that
he was inferior to Iamblichus in theological science, who so greatly
excelled in knowledge of this kind, that he was not surpassed by any
one, and was equaled by few. Hence he was denominated by all
succeeding Platonists the divine, in the same manner as Plato, “to
whom,” as the acute Emperor Julian remarks, “he was posterior in
time only, but not in genius.”[16]
The difficulties attending the translation of this work into English
are necessarily great, not only from its sublimity and novelty, but
also from the defects of the original. I have, however, endeavoured to
make the translation as faithful and complete as possible; and have
occasionally availed myself of the annotations of Gale, not being able
to do so continually, because for the most part, where philosophy is
concerned, he shows himself to be an inaccurate, impertinent, and
garrulous smatterer.
 THE
EPISTLE OF PORPHYRY
TO THE
EGYPTIAN ANEBO.

Porphyry to the Prophet Anebo greeting.

I commence my friendship towards you from the Gods and good

dæmons, and from those philosophic disquisitions, which have an
affinity to these powers. And concerning these particulars indeed,
much has been said by the Grecian philosophers; but, for the most
part, the principles of their belief are derived from conjecture.
In the first place, therefore, it is granted that there are Gods. But I
inquire what the peculiarities are of each of the more excellent
genera, by which they are separated from each other; and whether
we must say that the cause of the distinction between them is from
their energies, or their passive motions, or from things that are
consequent, or from their different arrangement with respect to
bodies; as, for instance, from the arrangement of the Gods with
reference to etherial, but of dæmons to aerial, and of souls to
terrestrial, bodies?
I also ask, why, since [all] the Gods dwell in the heavens,
theurgists only invoke the terrestrial and subterranean Gods?
Likewise, how some of the Gods are said to be aquatic and aerial?
And how different Gods are allotted different places, and the parts of
bodies according to circumscription, though they have an infinite,
impartible, and incomprehensible power? How there will be a union
of them with each other, if they are separated by the divisible
circumscriptions of parts, and by the difference of places and subject
bodies?
 How do theologists, or those who are wise in divine concerns,
represent the Gods as passive, to whom on this account, it is said,
erect phalli are exhibited, and obscene language is used? But if they
are impassive, the invocations of the Gods will be in vain, which
announce that they can appease the anger of the divinities, and
procure a reconciliation with them; and still more, what are called
the necessities of the Gods, will be vain. For that which is impassive
cannot be allured, nor compelled, nor necessitated. How, therefore,
are many things, in sacred operations, performed to them as passive?
Invocations,—likewise, are made to the Gods as passive; so that not
dæmons only are passive, but the Gods also, conformably to what
Homer says,
“And flexible are e’en the Gods themselves.”[17]

But if we assert with certain persons, that the Gods are pure
intellects, but that dæmons, being psychical, participate of intellect;
in a still greater degree will pure intellects be incapable of being
allured, and will be unmingled with sensible natures. Supplications,
however, are foreign to the purity of intellect, and therefore are not
to be made to it. But the things which are offered [in sacred rites] are
offered as to sensitive and psychical essences.
Are, therefore, the Gods separated from dæmons, through the
former being incorporeal, but the latter corporeal? If, however, the
Gods are incorporeal alone, how will the sun and moon, and the
visible celestials, be Gods?
How, likewise, are some of the Gods beneficent, but others
malefic?
What is it that connects the Gods in the heavens that have bodies,
with the incorporeal Gods?
What is it that distinguishes dæmons from the visible and invisible
Gods, since the visible are connected with the invisible Gods?
In what do a dæmon, hero, and soul, differ from each other? Is it
in essence, or in power, or in energy?
What is the indication of a God, or angel, or archangel, or dæmon,
or a certain archon, or soul being present? For to speak boastingly,
and to exhibit a phantasm of a certain quality, is common to Gods
and dæmons, and to all the more excellent genera. So that the genus
of Gods will in no respect be better than that of dæmons.
Since the ignorance of, and deception about, divine natures is
impiety and impurity, but a scientific knowledge of the Gods is holy
and beneficial, the ignorance of things honourable and beautiful will
be darkness, but the knowledge of them will be light. And the former,
indeed, will fill men with all evils, through the want of erudition, and
through audacity; but the latter will be the cause to them of every
good. [I wish you, therefore, to unfold to me the truth respecting
these particulars.[18]]
[And, in the first place, I wish you to explain to me distinctly[19]]
what that is which is effected in divination? For we frequently obtain
a knowledge of future events through dreams, when we are asleep;
not being, at that time, in a tumultuous ecstasy, for the body is then
quiescent; but we do not apprehend what then takes place, in the
same manner as when we are awake.
But many, through enthusiasm and divine inspiration, predict
future events, and are then in so wakeful a state, as even to energize
according to sense, and yet they are not conscious of the state they
are in, or at least, not so much as they were before.
Some also of those who suffer a mental alienation, energize
enthusiastically on hearing cymbals or drums, or a certain
modulated sound, such as those who are Corybantically inspired,
those who are possessed by Sabazius, and those who are inspired by
the mother of the Gods. But some energize enthusiastically by
drinking water, as the priest of Clarius, in Colophon; others, by being
seated at the mouth of a cavern, as those who prophesy at Delphi;
and others by imbibing the vapour from water, as the prophetesses
in Branchidæ. Some also become enthusiastic by standing on
characters, as those that are filled from the intromission of spirits.
Others, who are conscious what they are doing in other respects, are
divinely inspired according to the phantastic part; some, indeed,
receiving darkness for a cooperator, others certain potions, but
others incantations and compositions: and some energize, according
to the imagination, through water; others in a wall, others in the
open air, and others in the sun, or in some other of the celestial
bodies. Some also establish the art of the investigation of futurity
through the viscera, through birds, and through the stars.
 I likewise ask concerning the mode of divination, what it is, and
what the quality by which it is distinguished? All diviners, indeed,
assert, that they obtain a foreknowledge of future events through
Gods or dæmons, and that it is not possible for any others to know
that which is future, than those who are the lords of futurity. I doubt,
therefore, whether divinity is so far subservient to men, as not to be
averse to some becoming diviners from meal.
But, concerning the causes of divination, it is dubious whether a
God, an angel, or a dæmon, or some other power, is present in
manifestations, or divinations, or certain other sacred energies, as is
the case with those powers that are drawn down through you
[priests] by the necessities with which invocation is attended.
Or does the soul assert and imagine these things, and are they, as
some think, the passions of the soul, excited from small incentives?
Or is a certain mixed form of subsistence produced from our soul,
and divine inspiration externally derived?
Hence it must be said, that the soul generates the power which has
an imaginative perception of futurity, through motions of this kind,
or that the things which are adduced from matter constitute
dæmons, through the powers that are inherent in them, and
especially things adduced from the matter which is taken from
animals.
For in sleep, when we are not employed about any thing, we
sometimes obtain a knowledge of the future.
But that a passion of the soul is the cause of divination, is indicated
by this, that the senses are occupied, that fumigations are
introduced, and that invocations are employed; and likewise, that
not all men, but those that are more simple and young, are more
adapted to prediction.
The ecstasy, also, of the reasoning power is the cause of divination,
as is likewise the mania which happens in diseases, or mental
aberration, or a sober and vigilant condition, or suffusions of the
body, or the imaginations excited by diseases, or an ambiguous state
of mind, such as that which takes place between a sober condition
and ecstasy, or the imaginations artificially procured by
enchantment.
 Nature, likewise, art, and the sympathy of things in the universe,
as if they were the parts of one animal, contain premanifestations of
certain things with reference to each other. And bodies are so
prepared, that there is a presignification of some by others, which is
clearly indicated by the works performed in predicting what is future.
For those who invoke the divinities for this purpose, have about
them stones and herbs, bind certain sacred bonds, which they also
dissolve, open places that are shut, and change the deliberate
intentions of the recipients, so as from being depraved to render
them worthy, though they were before depraved. Nor are the
artificers of efficacious images to be despised. For they observe the
motion of the celestial bodies, and can tell from the concurrence of
what star with a certain star or stars, predictions will be true or false;
and also whether the things that are performed will be inanities, or
significant and efficacious, though no divinity or dæmon is drawn
down by these images.
But there are some who suppose that there is a certain obedient
genus of dæmons, which is naturally fraudulent, omniform, and
various, and which assumes the appearance of Gods and dæmons,
and the souls of the deceased; and that through these every thing
which appears to be either good or evil is effected; for they are not
able to contribute any thing to true goods, such as those of the soul,
nor to have any knowledge of them, but they abuse, deride, and
frequently impede those who are striving to be virtuous. They are
likewise full of pride, and rejoice in vapours and sacrifices.
Jugglers likewise fraudulently attack us in many ways, through the
ardour of the expectations which they raise.
It very much indeed perplexes me to understand how superior
beings, when invoked, are commanded by those that invoke them, as
if they were their inferiors; and they think it requisite that he who
worships them should be just, but when they are called upon to act
unjustly, they do not refuse so to act. Though the Gods, likewise, do
not hear him who invokes them, if he is impure from venereal
connexions, yet, at the same time, they do not refuse to lead any one
to illegal venery.
[I am likewise dubious with respect to sacrifices, what utility or
power they possess in the universe, and with the Gods, and on what
account they are performed, appropriately indeed, to the powers who
are honoured by them, but usefully to those by whom the gifts are
offered.[20]]
Why also do the interpreters of prophecies and oracles think it
requisite that they should abstain from animals, lest the Gods should
be polluted by the vapours arising from them; and yet the Gods are
especially allured by the vapours of animals?
Why is it requisite that the inspector [who presides over sacred
rites] ought not to touch a dead body, though most sacred operations
are performed through dead bodies? And why, which is much more
absurd than this, are threats employed and false terrors, by any
casual person, not to a dæmon, or some departed soul, but to the
sovereign Sun himself, or to the Moon, or some one of the celestial
Gods, in order to compel these divinities to speak the truth? For does
not he who says that he will burst the heavens, or unfold the secrets
of Isis, or point out the arcanum in the adytum, or stop Baris, or
scatter the members of Osiris to Typhon, [or that he will do
something else of the like kind[21]], does not he who says this, by thus
threatening what he neither knows nor is able to effect, prove himself
to be stupid in the extreme? And what abjectness does it not produce
in those who, like very silly children, are possessed with such vain
fear, and are terrified at such fictions? And yet Chæremon, who was
a sacred scribe, writes these things, as disseminated by the
Egyptians. It is also said, that these, and things of the like kind, are of
a most compulsive nature.
What also is the meaning of those mystic narrations which say that
a certain divinity is unfolded into light from mire, that he is seated
above the lotus, that he sails in a ship, and that he changes his forms
every hour, according to the signs of the zodiac? For thus, they say,
he presents himself to the view, and thus ignorantly adapt the
peculiar passion of their own imagination to the God himself. But if
these things are asserted symbolically, being symbols of the powers
of this divinity, I request an interpretation of these symbols. For it is
evident, that if these are similar to passions of the Sun, when he is
eclipsed, they would be seen by all men who intently survey the God.
What also is the design of names that are without signification?
and why, of such, are those that are barbaric preferred to our own?
For if he who hears them looks to their signification, it is sufficient
that the conception remains the same, whatever the words may be
that are used. For he who is invoked is not of the Egyptian race; nor,
if he was an Egyptian, does he use the Egyptian, or, in short, any
human language. For either all these are the artificial contrivances of
enchanters, and veils originating from our passions, which rumour
ascribes to a divine nature; or we ignorantly frame conceptions of
divinity, contrary to its real mode of subsistence.
I likewise wish you to unfold to me, what the Egyptians conceive
the first cause to be; whether intellect, or above intellect? whether
alone, or subsisting with some other or others? whether incorporeal,
or corporeal; and whether it is the same with the Demiurgus, or prior
to the Demiurgus? Likewise, whether all things are from one
principle, or from many principles? whether the Egyptians have a
knowledge of matter, or of primary corporeal qualities; and whether
they admit matter to be unbegotten, or to be generated? For
Chæremon, indeed, and others, do not think there is any thing else
prior to the visible worlds; but in the beginning of their writings on
this subject, admit the existence of the Gods of the Egyptians, but of
no others, except what are called the planets, the Gods that give
completion to the zodiac, and such as rise together with these; and
likewise, the sections into decans, and the horoscopes. They also
admit the existence of what are called the powerful leaders, whose
names are to be found in the calendars, together with their
ministrant offices, their risings and settings, and their significations
of future events. For Chæremon saw that what those who say that the
sun is the Demiurgus, and likewise what is asserted concerning
Osiris and Isis, and all the sacred fables, may be resolved into the
stars and the phases, occultations and risings of these, or into the
increments or decrements of the moon, or into the course of the sun,
or the nocturnal and diurnal hemisphere, or into the river [Nile].
And, in short, the Egyptians resolve all things into physical, and
nothing into incorporeal and living essences. Most of them likewise
suspend that which is in our power from the motion of the stars; and
bind all things, though I know not how, with the indissoluble bonds
of necessity, which they call fate. They also connect fate with the
Gods; whom, nevertheless, they worship in temples and statues, and
other things, as the only dissolvers of fate.
Concerning the peculiar dæmon, it must be inquired how he is
imparted by the lord of the geniture, and according to what kind of
efflux, or life, or power, he descends from him to us? And also,
whether he exists, or does not exist? And whether the invention of
the lord of the geniture is impossible, or possible? For if it is possible
he is happy, who having learned the scheme of his nativity, and
knowing his proper dæmon, becomes liberated from fate.
The canons, also, of genethlialogy [or prediction from the natal
day] are innumerable and incomprehensible. And the knowledge of
this mathematical science cannot be obtained; for there is much
dissonance concerning it, and Chæremon and many others have
written against it. But the discovery of the lord, or lords, of the
geniture, if there are more than one in a nativity, is nearly granted by
astrologers themselves to be unattainable, and yet they say that on
this the knowledge of the proper dæmon depends.
Farther still, I wish to know whether the peculiar dæmon rules
over some one of the parts in us? For it appears to certain persons,
that dæmons preside over the parts of our body, so that one is the
guardian of health, another of the form of the body, and another of
the corporeal habits, and that there is one dæmon who presides in
common over all these. And again, that one dæmon presides over the
body, another over the soul, and another over the intellect; and that
some of them are good, but others bad.
I am also dubious whether this dæmon is not a certain part of the
soul, [such, for instance, as the intellectual part;] and if so, he will be
happy who has a wise intellect.
I see likewise, that there is a twofold worship of the peculiar
dæmon; the one being the worship as of two, but the other as of
three. By all men, however, the dæmon is called upon by a common
invocation.
I farther ask, whether there is a certain other latent way to felicity,
separate from the Gods? And I am dubious whether it is requisite to
look to human opinions in divine divination and theurgy? And
whether the soul does not devise great things from casual
circumstances? Moreover, there are certain other methods, which
are conversant with the prediction of future events. And, perhaps,
those who possess divine divination, foresee indeed what will
happen, yet are not on this account happy; for they foresee future
events, but do not know how to use this knowledge properly. I wish,
therefore, that you would point out to me the path to felicity, and
show me in what the essence of it consists. For with us [Greeks]
there is much verbal contention about it, because we form a
conjecture of good from human reasonings. But by those who have
devised the means of associating with beings more excellent than
man, if the investigation of this subject is omitted, wisdom will be
professed by them in vain; as they will only disturb a divine intellect
about the discovery of a fugitive slave, or the purchase of land, or, if
it should so happen, about marriage, or merchandize. And if they do
not omit this subject, but assert what is most true about other things,
yet say nothing that is stable and worthy of belief about felicity, in
consequence of employing themselves about things that are difficult,
but useless to mankind; in this case, they will not be conversant
either with Gods or good dæmons, but with that dæmon who is
called fraudulent; or, if this is not admitted, the whole will be the
invention of men, and the fiction of a mortal nature.
Iamblichus[22] on the Mysteries, &c.
 THE

ANSWER OF THE PRECEPTOR ABAMMON

TO THE

EPISTLE OF PORPHYRY TO ANEBO,

AND A

SOLUTION OF THE DOUBTS CONTAINED IN

IT.
 SECTION I.

CHAP. I.

Hermes, the God who presides over language, was formerly very
properly considered as common to all priests; and the power who
presides over the true science concerning the Gods is one and the
same in the whole of things. Hence our ancestors dedicated the
inventions of their wisdom to this deity, inscribing all their own
writings with the name of Hermes. If, therefore, we participate of a
portion of this God, adapted and commensurate to our powers, you
do well to propose your theological doubts to the priests, as friends,
and to make these doubts known to them. I also very properly
conceiving that the epistle sent to my disciple Anebo was written to
me, shall give you a true answer to your inquiries. For it would not be
becoming, that Pythagoras and Plato, Democritus and Eudoxus, and
many other of the ancient Greeks, should have obtained appropriate
instruction from the sacred scribes of their time, but that you who
are our contemporary, and think conformably to those ancients,
should be frustrated of your wish by those who are now living, and
who are called common preceptors. I, therefore, thus betake myself
to the present discussion; and do you, if you please, conceive that the
same person to whom you sent the letter returns you an answer. Or,
if it should seem fit to you, admit it to be me who discourses with you
in writing, or some other prophet of the Egyptians, for this is of no
consequence. Or, which I think is still better, dismiss the
consideration whether the speaker is an inferior or a superior
character, but direct your attention to what is said, so as readily to
excite your mind to survey whether what is asserted is true or false.
In the first place, therefore, we shall divide the genera of the
proposed problems, in order that we may know the quantity and
quality of them. And, in the next place, we shall show from what
theologies the doubts are assumed, and according to what sciences
they are investigated. For some things that are badly confused,
require a certain distinction; others are conversant with the cause
through which they subsist, and are apprehended; others, which we
propose according to a certain contrariety, draw our decision on both
sides; and some things require from us the whole development of
mystic doctrines. Such, therefore, being the nature of the subjects of
discussion, they are assumed from many places, and from different
sciences. For some things introduce animadversions from what the
wise men of the Chaldeans have delivered; others produce objections
from what the prophets of the Egyptians teach; and there are some
that, adhering to the theory of philosophers, make inquiries
conformably to them. There are now likewise some, that from other
opinions, which do not deserve to be mentioned, elicite a certain
dubitation; and others originate from the common conceptions of
mankind. These things, therefore, are of themselves variously
disposed, and are multiformly connected with each other. Hence,
through all these causes, a certain discussion is requisite for the
management of them in a becoming manner.
 CHAP. II.

We shall, therefore, deliver to you the peculiar dogmas of the

Assyrians; and also clearly develop to you our own opinions;
collecting some things from the infinite writings of the ancients, but
others from those particulars which were comprehended by the
ancients in one treatise, and pertain to the whole knowledge of divine
natures. If also you should propose any philosophic inquiry, we shall
discuss it for you, according to the ancient pillars of Hermes, which
Plato and Pythagoras knew before, and from thence constituted their
philosophy. But such things as exhibit foreign inquiries, or which are
contradictory and contentious, we shall assist mildly and aptly, or we
shall demonstrate their absurdity. Such, likewise, as proceed
conformably[23] to common conceptions, we shall endeavour to
discuss in a way perfectly known and clear. And things, indeed,
which require the experience of divine operations to an accurate
knowledge of them, we shall explain, as far as this is possible to be
effected by words alone; but such as are full of intellectual theory, we
shall develop with a view to the purification of the soul. But
indications of this theory worthy of notice may be mentioned, by
which it is possible for you, and those who resemble you, to be
conducted by intellect to the essence of [real] beings. And with
respect to such things as become known by a reasoning process, we
shall leave no one of these without a perfect demonstration. But in all
things we shall give to each that which is appropriate. And such
questions, indeed, as are theological, we shall answer theologically;
such as are theurgic, theurgically; but such as are philosophical, we
shall, in conjunction with you, philosophically explore. Of these, also,
such as extend to first causes, we shall unfold into light, by following
them conformably to first principles. But such as pertain to morals,
or to ends, we shall fitly discuss, according to the ethical mode. And,
in a similar manner, we shall examine other things methodically and