Visit https://ebookmass.

com to download the full version and

browse more ebooks or textbooks

Muon Spectroscopy: An Introduction Stephen J.

Blundell (Editor)

_____ Press the link below to begin your download _____

https://ebookmass.com/product/muon-spectroscopy-an-
introduction-stephen-j-blundell-editor/

Access ebookmass.com now to download high-quality

ebooks or textbooks
We have selected some products that you may be interested in
Click the link to download now or visit ebookmass.com
for more options!.

An Introduction to Correspondence Analysis 1st Edition

Eric J. Beh

https://ebookmass.com/product/an-introduction-to-correspondence-
analysis-1st-edition-eric-j-beh/

An Introduction to Orthodontics 5th Edtion Edition Simon

J. Littlewood

https://ebookmass.com/product/an-introduction-to-orthodontics-5th-
edtion-edition-simon-j-littlewood/

Creating the Qur’an Stephen J. Shoemaker

https://ebookmass.com/product/creating-the-quran-stephen-j-shoemaker/

Cardiology-An Integrated Approach (Human Organ Systems)

(Dec 29, 2017)_(007179154X)_(McGraw-Hill) 1st Edition
Elmoselhi
https://ebookmass.com/product/cardiology-an-integrated-approach-human-
organ-systems-dec-29-2017_007179154x_mcgraw-hill-1st-edition-
elmoselhi/
MATLAB Programming for Engineers 6th Edition Stephen J.
Chapman

https://ebookmass.com/product/matlab-programming-for-engineers-6th-
edition-stephen-j-chapman/

Ethics, Technology, and Engineering: An Introduction 1st

Edition

https://ebookmass.com/product/ethics-technology-and-engineering-an-
introduction-1st-edition/

Zero-trust – An Introduction 1st Edition Tom Madsen

https://ebookmass.com/product/zero-trust-an-introduction-1st-edition-
tom-madsen/

Handbook of Mineral Spectroscopy: X-Ray Photoelectron

Spectra: Volume 1: X-ray Photoelectron Spectra 1st Edition
J. Theo Kloprogge
https://ebookmass.com/product/handbook-of-mineral-spectroscopy-x-ray-
photoelectron-spectra-volume-1-x-ray-photoelectron-spectra-1st-
edition-j-theo-kloprogge/

Chemometrics in Spectroscopy (Second Edition) Howard Mark

https://ebookmass.com/product/chemometrics-in-spectroscopy-second-
edition-howard-mark/
 2

Muon
Spectroscopy
An Introduction

St e p h e n J. B lu n d e l l
R o b e rto D e R e n z i
To m L a n c a st e r
F r a n c i s L . P r at t
Muon Spectroscopy - An Introduction
Muon Spectroscopy - An Introduction

Edited by

Stephen J. Blundell
University of Oxford

Roberto De Renzi
University of Parma

Tom Lancaster
University of Durham

Francis L. Pratt
ISIS Neutron and Muon Source
 Great Clarendon Street, Oxford, OX2 6DP,
United Kingdom
Oxford University Press is a department of the University of Oxford.
It furthers the University’s objective of excellence in research, scholarship,
and education by publishing worldwide. Oxford is a registered trade mark of
Oxford University Press in the UK and in certain other countries
© Oxford University Press 2022
The moral rights of the authors have been asserted
Impression: 1
All rights reserved. No part of this publication may be reproduced, stored in
a retrieval system, or transmitted, in any form or by any means, without the
prior permission in writing of Oxford University Press, or as expressly permitted
by law, by licence or under terms agreed with the appropriate reprographics
rights organization. Enquiries concerning reproduction outside the scope of the
above should be sent to the Rights Department, Oxford University Press, at the
address above
You must not circulate this work in any other form
and you must impose this same condition on any acquirer
Published in the United States of America by Oxford University Press
198 Madison Avenue, New York, NY 10016, United States of America
British Library Cataloguing in Publication Data
Data available
Library of Congress Control Number: 2021936416
ISBN 978–0–19–885895–9 (hbk.)
ISBN 978–0–19–885896–6 (pbk.)
DOI: 10.1093/oso/9780198858959.001.0001
Printed and bound by
CPI Group (UK) Ltd, Croydon, CR0 4YY
Links to third party websites are provided by Oxford in good faith and
for information only. Oxford disclaims any responsibility for the materials
contained in any third party website referenced in this work.
Preface

Muons, radioactive particles that can be produced by accelerators, have

emerged as an important tool to study problems in condensed matter
physics and chemistry. Beams of muons with all their spins polarized
can be prepared and implanted in various types of material. The sub-
sequent precession and relaxation of the spins of these muons can be
used to investigate a variety of static and dynamic effects in a sample
and hence to deduce properties concerning magnetism, superconductiv-
ity, molecular or chemical dynamics, and a bewildering array of other
properties.
The technique has a slightly odd name: µSR. The µ stands, of course,
for the muon and the S is the spin of the muon, because the spin an-
gular momentum of the muon is what provides its magnetic moment,
which is the property from which the muon’s sensitivity to magnetic
fields derives, and this is the key aspect of the interaction between the
muon and the sample. But what about the R? The name was given
to make it look a bit like NMR (nuclear magnetic resonance) and ESR
(electron spin resonance), so you might guess that µSR is muon spin
resonance. However, the muon is a multifaceted probe, and for reasons
we will explain in detail in this book the technique is rarely performed
resonantly, since the dominant effects are rotation (the Larmor preces-
sion of the muon, which we can observe directly) and relaxation. Thus,
sometimes we will be dealing with muon spin rotation, sometimes with
muon spin relaxation, and just occasionally with muon spin resonance.
To avoid this ambiguity, we have called this book Muon spectroscopy:
an introduction, but note that the technique will usually be referred to,
as is conventional, by the name µSR.
The µSR technique was originally the preserve of a few specialists lo-
cated in particle physics laboratories. Today it is used by scientists from
a very wide range of science backgrounds and interests. Although there
are some very good technical monographs and review articles describing
the technique, we felt there was a need for a book designed for begin-
ners of many different types: curious undergraduates, starting graduate
students, or even experts in related fields who want to find out what all
the fuss is about. We hope this is a book that many will be able to learn
from, as well as learning to enjoy the beauty of the physics behind this
wonderful probe.
For those wishing to explore more deeply into the µSR technique, each
chapter is supplemented with further reading. More information is also
available on the websites of the international muon facilities and also
vi Preface

the International Society for Muon Spectroscopy (ISMS). The royalties

of this book are being donated to the ISMS in order to contribute to
the financial support for young scientists, particularly those from less
wealthy countries, to attend the triennial international conference on
muon spectroscopy.
This book grew out of a European workshop held during August 2019
at the Rutherford Appleton Laboratory in which scientists from across
the world gathered for 2 weeks for an advanced school on muon spec-
troscopy. The school was enabled through the major support of ISIS,
PSI and the EU SINE2020 project, along with additional support from
ISMS, TRIUMF, and JPARC. The lecturers from this school consented
to write material for the book based on their lectures. Inevitably, such an
approach produces material using a range of styles and notations, some
overlaps and some omissions. Therefore, the editors of this book resolved
to take the submitted material, mix it up and revise it, reorder and rein-
vent parts, aiming to produce a final result which would look as if it came
from a single author (the one exception being the historical appendix by
Steve Cox; with a relatively new technique, a short history is inevitably a
personal view). We are grateful for the forbearance and generosity of our
team of contributing authors (listed on page vii) for allowing their mate-
rial to be so roughly handled, but we hope that the final result is worth
it. Our thanks are primarily to them, but also to Sönke Adlung and his
team at OUP, our eagle-eyed copy editor Kate Blackham, as well as a
number of kind friends and colleagues who read through the book and
made numerous helpful comments, suggestions, and corrections: Pietro
Bonfà, Stewart Clark, Jonathan Frassineti, Elena Garlatti, George Gill,
Matjaž Gomilšek (who also produced Fig. 16.11), Alberto Hernandez
Melian, Thomas Hicken, Ben Huddart, Gianrico Lamura, Muhammad
Maikudi Isah, Ifeanyi John Onuorah, Giacomo Prando, Samuele Sanna,
Toni Shiroka, Johnny Wilkinson, and Hank Wu, as well as our contribut-
ing authors. The book is better because of their efforts. Finally, we
express our gratitude to the students who attended the school, pressed
us all to think about how to explain the various aspects of µSR, or think
again about what we thought we knew.
Stephen J. Blundell
Roberto De Renzi
Tom Lancaster
Francis L. Pratt
July 7, 2021
List of contributors

Peter J. Baker, STFC-ISIS Rutherford Appleton Laboratory, UK

Adam Berlie, STFC-ISIS Rutherford Appleton Laboratory, UK
Pabitra K. Biswas, STFC-ISIS Rutherford Appleton Laboratory, UK
Stephen J. Blundell, University of Oxford, UK
Pietro Carretta, University of Pavia, Italy
Stephen P. Cottrell, STFC-ISIS Rutherford Appleton Laboratory, UK
Stephen F. J. Cox, STFC-ISIS Rutherford Appleton Laboratory, UK
Pierre Dalmas de Réotier, CEA Grenoble, France
Roberto De Renzi, University of Parma, Italy
Adrian D. Hillier, STFC-ISIS Rutherford Appleton Laboratory, UK
Rustem Khasanov, Paul Scherrer Institute, Switzerland
Tom Lancaster, University of Durham, UK
Leandro Liborio, STFC-SCD Rutherford Appleton Laboratory, UK
James S. Lord, STFC-ISIS Rutherford Appleton Laboratory, UK
Martin Månsson, KTH Stockholm, Sweden
Iain McKenzie, TRIUMF, Vancouver, Canada
Francis L. Pratt, STFC-ISIS Rutherford Appleton Laboratory, UK
Thomas Prokscha, Paul Scherrer Institute, Switzerland
Zaher Salman, Paul Scherrer Institute, Switzerland
Simone Sturniolo, STFC-SCD Rutherford Appleton Laboratory, UK
Jun Sugiyama, CROSS-Tokai, Japan
Andreas Suter, Paul Scherrer Institute, Switzerland
Rui Vilão, University of Coimbra, Portugal
Koji Yokoyama, STFC-ISIS Rutherford Appleton Laboratory, UK

The book website is https://musr.org/muon-spectroscopy-book

Contents

1 The basics of µSR 1

1.1 The key idea 1
1.2 The principles of the experiment 2
1.3 Muon beams and spectrometers 4
1.4 Experimental geometries 6
1.5 What can we do with µSR? 6

I Elements of muon spectroscopy 9

2 Introduction 10
2.1 Discovery of the muon 10
2.2 The first muon application 11
2.3 Muon perspectives 12
2.4 The µSR experiment 17
Exercises 20

3 Muon charge and spin states 21

3.1 State formation 22
3.2 Hydrogen analogues 24
3.3 Measuring the states 26
3.4 Influencing the states 27
Exercises 28

4 The quantum muon 29

4.1 Larmor precession 29
4.2 Density matrices 33
4.3 Mixed states 34
4.4 Two spins: muonium 36
4.5 Multiple spins 44
Exercises 52

5 Polarization functions 54
5.1 Static fields 55
5.2 Dynamical fields 60
5.3 Disordered systems 64
5.4 The stretched exponential 65
Exercises 67
x Contents

II Science with µSR 69

6 Magnetism 70
6.1 The basics 70
6.2 Static magnetic order 72
6.3 The local magnetic field 76
6.4 Static field distributions 81
Exercises 86

7 Dynamic effects in magnetism 87

7.1 Correlation functions 87
7.2 Dynamics in magnets 90
7.3 Dynamics with muons 92
7.4 Relaxation as resonance 95
7.5 Dynamic magnetism 96
7.6 Coupling tensors 101
7.7 Dilute spins 103
Exercises 108

8 Measuring dynamic processes 110

8.1 Critical dynamics 110
8.2 Magnetism in metals 113
8.3 BPP relaxation 115
8.4 Mobile excitations 117
8.5 Muon diffusion 125
Exercises 129

9 Superconductors 130
9.1 The discovery 130
9.2 London penetration depth 131
9.3 Ginzburg-Landau model 134
9.4 Type-II superconductors 136
9.5 Measuring the penetration depth 140
9.6 The microscopic model 144
9.7 Example materials 146
9.8 Clean versus dirty 150
9.9 The Uemura plot 151
9.10 Spontaneous fields 154
Exercises 155

10 Semiconductors and dielectrics 157

10.1 Ubiquitous hydrogen impurities 159
10.2 Muonium 160
10.3 Silicon: the foundations 161
10.4 Shallow donor states 165
10.5 Related techniques 166
Exercises 169

11 Ionic motion 170

 Contents xi

11.1 Why use muons? 170

11.2 Science examples 173
11.3 Limitations 175
Exercises 176

12 Chemistry 177
12.1 Chemical environments 177
12.2 Muonium spectroscopy 180
12.3 Reactions of muonium 183
12.4 Muoniated radicals 186
12.5 Structure and dynamics 191
Exercises 198

III Practicalities of muon spectroscopy 199

13 Making muons 200
13.1 Muon production 200
13.2 Surface and decay muons 202
13.3 Beamline components 204
Exercises 209

14 Instrumentation 210
14.1 Spectrometer elements 210
14.2 Pulsed sources 213
14.3 Continuous sources 214
14.4 Small samples 215

15 Doing the experiment 217

15.1 Experimental setup 217
15.2 Calibrations 220
15.3 Data characteristics 221
15.4 Time domain analysis 224
15.5 Frequency domain 230
Exercises 238

IV Further topics in muon spectroscopy 239

16 Calculating muon sites 240
16.1 The site problem 241
16.2 What is DFT? 243
16.3 Methods 246
16.4 Basis sets 248
16.5 Functionals 250
16.6 Mixed methods 251
16.7 Obtaining sites 252
16.8 Quantum effects 257
16.9 Sites via experiment 258
xii Contents

Exercises 261

17 Numerical modelling 263

17.1 Beamline optimization 263
17.2 Muon range profile 265
17.3 Muon spin response 267

18 Low energy µSR 274

18.1 Generating slow muons 274
18.2 LEM facilities 277
18.3 Science examples 280

19 Stimulation methods 283

19.1 Types of stimulation 283
19.2 Case studies 285
19.3 Photoexcitation 288
19.4 Muon-spin resonance 293
Exercises 303

20 High magnetic fields 305

20.1 Why high fields? 305
20.2 Muons and high magnetic fields 307
20.3 Science at high field 310
Exercises 312

21 Muons under pressure 313

21.1 Requirements 313
21.2 The PSI setup 314
21.3 A gas-pressure setup 318
21.4 Science examples 320
21.5 Outlook 321

22 Negative muon techniques 323

22.1 µ− SR spectroscopy 323
22.2 Elemental analysis 328
Exercises 334

V Complementary techniques 335

23 µSR versus other resonance and bulk techniques 336
23.1 Magnetic resonance 336
23.2 When the muon is a plus 340
23.3 Mössbauer spectroscopy 342
23.4 Bulk techniques 344
Exercises 349

24 X-rays, neutrons, and µSR 350

24.1 X-rays 350
 Contents xiii

24.2 Neutrons 352

24.3 Where do muons fit in? 353
Exercises 355

A Fundamental constants 357

B Nuclear moments 358

C Negative muon lifetimes 360

D Answers to selected problems 361

E Muon particle physics 366

E.1 Parity violation 366
E.2 Standard Model and weak interactions 367
E.3 Muon production 368
E.4 Muon decay 369

F Quantum-mechanical polarization functions 373

F.1 Time-dependent perturbations 373
F.2 Evaluating terms 376

G The second moment of a spin distribution 379

G.1 The dipolar interaction 379
G.2 High transverse field 380
G.3 Zero field 384
G.4 Quadrupolar coupling 385

H A short history of µSR 387

Index 406
The basics of µSR 1
This is a book about a set of closely related experiments that use muons
1.1 The key idea 1
to investigate matter. They are known as µSR, which stands for muon-
1.2 The principles of the experi-
spin relaxation, rotation, or resonance. ment 2
1.3 Muon beams and spectrome-
ters 4
1.1 The key idea 1.4 Experimental geometries 6
1.5 What can we do with µSR? 6
The muon is a radioactive particle with a magnetic moment and it Chapter summary 7
can be implanted into a sample.

We will now unpack the various parts of this sentence in more detail. 1
Real experiments work with either one
muon at a time, or a pulse of 102 –103
(a) The muon is a radioactive particle
muons at a time. Therefore, accumu-
lating 2 × 107 total counts might take
The muon is radioactive and so it does not last very long. It has an ≈ 1 hour, though the time depends a lot
average lifetime of τµ = 2.2 µs, after which it decays into other particles. on the beamline and the experiment, as
The special nature of the particle’s decay allows us to read out the we will discuss in due course.
information that the muon probes.
N (t)/N (0)

Example 1.1
The radioactive decay of the muon means that the number of muons n(t) in the
sample a time t after implantation follows the equation

n(t) = n(0) e−t/τµ . (1.1)

This means that if you were able to implant twenty million muons [n(0) = 2 × 107 ] t (µs)
into the sample at t = 0 (a number of muons which is not uncommon1 in a µSR
experiment), the number remaining at t = 20 µs, is 20 × 106 × e−20/2.2 which is over
Fig. 1.1 The total counts as a func-
22,000. So even though the muon decay time is 2.2 µs, you can measure muons that
tion of time from a µSR experiment,
live a lot longer than that.
illustrating the exponential decay pre-
Note also that the number of muons N (t) dt that decay between t and t + dt is
dicted from eqn 1.2.
given by
dn(t) n(0) −t/τµ 2
N (t) = − = e = N (0) e−t/τµ , (1.2) This is because positive muons are of-
dt τ
ten more effective for investigating the
and N (t) is the quantity that is recorded in a µSR experiment (see Fig. 1.1). collective behaviour of matter, com-
pared to negative muons which, owing
to their resemblance to electrons, are
more useful in investigating the chem-
The muon is a subatomic particle and it can be positively or neg- ical elements in a material. However,
atively charged. We usually employ positively charged muons in µSR negatively charged muons are also used
measurements.2 (see Chapter 22).
2 The basics of µSR

(b) The muon has a magnetic moment

The muon is a spin- 12 particle and has a small magnetic moment. Its
magnetic moment allows it to interact magnetically with its environment
and hence investigate the properties of the sample into which it has been
implanted. A major part of this book will be to explore the nature and
consequences of these magnetic interactions.

(c) The muon is implanted in the sample

Note the word implant. This is not a scattering technique, such as X-
ray scattering or neutron scattering, in which an incoming particle is
directed towards a sample along a particular direction and then a scat-
3
Sometimes people mistakenly refer to tered particle detected.3 This is an implantation technique in which
this technique as ‘muon scattering’, the muon is directed towards the sample to be studied and it comes to
which is incorrect.
4
a complete rest somewhere inside the sample.4
The site of the implanted muon is usu-
ally an interstitial position in a crys-
talline solid. This issue will be dis-
cussed in detail in Chapter 16.
1.2 The principles of the experiment
Larmor precession
The positive muons that we implant into materials in these experiments
are always spin polarized, so all start with their spins aligned along a
known direction. If there is a magnetic field at the point where a muon
stops in a material, then the muon spin will undergo Larmor preces-
sion around the direction of the local field, at a frequency determined
by the muon gyromagnetic ratio γµ = 2π × 135.5 MHz T−1 , as shown
B in Fig. 1.2. Specifically, the Larmor (angular) frequency ω is given by
ω = γµ B, (1.3)

where B is the magnitude of the field at the muon site. Equation 1.3 is
arguably the most important one in this book.
θ
Sµ
Example 1.2
Fig. 1.2 In a magnetic field B the Calculate the precession frequency for a muon in (a) the Earth’s magnetic field (≈
muon spin S µ precesses around the 50 µT), (b) 2 mT, the magnetic field commonly used to calibrate a µSR experiment,5
field on a cone at an angular frequency and (c) 5 T, a relatively high magnetic field.
of ω = γµ B. Solution: We use the formula for precession frequency ν
ω γµ B
ν= = , (1.4)
5
Many facilities use old-fashioned cgs 2π 2π
and obtain (a) 6.8 kHz, (b) 0.27 MHz, and (c) 0.68 GHz. For experimental purposes,
units, and 2 mT translates to 20 G
it is often more useful to write the expressions in terms of frequency (ν) units, rather
(i.e. 20 Gauss, where 104 G = 1 T).
than angular frequency (ω) units, although, as we will see later, it is often more
Such a calibration is often known as a
convenient in theoretical work to use angular frequency.
‘T20’, which means a transverse field
As we have seen, a typical µSR experiment might measure times up to around 20 µs
of twenty Gauss. For more details on
after muon implantation. Precession in the Earth’s magnetic field [which would have
these units, see Appendix A.
a period of 1/(6.8 kHz) ≈ 150 µs] is therefore too slow to measure directly and one
would just observe the very beginning of a precession signal. On the other hand,
a 5 T field produces very fast precession, and to measure this one would need very
6
It is possible to do this though – see good timing resolution.6
Chapter 20.
 1.2 The principles of the experiment 3

Asymmetric emission of positrons

When a muon decays, a positron e+ and two neutrinos (ν) are emitted
via the reaction
µ+ → e+ + ν̄µ + νe . (1.5)

A property of this decay (which we will describe in more detail later)

is that the positron is not emitted isotropically, with equal chance of it Sµ
emerging in any direction. Instead it is emitted preferentially along the
direction of the muon spin at the moment of decay. (The distribution
is shown in Fig. 1.3.) Detecting the direction along which the positron
is emitted allows us to work out in which direction the muon spin was
pointing at the moment of its decay. Each muon lives for a different
amount of time (according to the radioactive decay law in eqn 1.1) and
so each muon decay gives rise to a positron whose detection contributes
to a data point at a particular time t. If we make many such positron Fig. 1.3 The probability of positron
detections, taken over many million muon decays, we can build up a emission in the direction of each grey
arrow is indicated by its length. (for
histogram. This histogram then tells us the average spin polarization of a particular value of emitted positron
the large ensemble of muons as a function of time. The data set which we energy; see Appendix E.4 for more
obtain from the experiment is simply the collection of measured average details). Positrons are preferentially
spin polarization values as a function of time after implantation. emitted along the muon spin direction
(shown in a black arrow). The figure
is axially symmetric about the muon-
spin direction (and so looks a bit like
the surface of a cherry in three dimen-
sions).
Example 1.3
A stopped spin-polarized ensemble of muons sits close to the origin of a set of Carte-
sian coordinates. Its spin initially lies along a unique direction (conventionally we
call this z) and a B-field is applied along x. Assuming no other interactions, the
muon spins will precess, undergoing circular motion in the y-z plane at an angular
frequency ω = γµ B. We place positron detectors in front of (i.e. at some value of
positive z) and behind (negative z) the origin. The average polarization P (t) of the
muon ensemble at a particular time t can be found by computing the normalized
difference in the number NF (t) of positrons counted in the forward detector and the
number NB (t) in the backward detector, both measured in a small interval centred
on t, which is to say
P (t)

NF (t) − NB (t)
P (t) ∝ . (1.6)
NF (t) + NB (t)
The polarization in this case is proportional to a cosine function P (t) ∝ cos γµ Bt,
which is simply the circular precessing motion projected along the z direction. If we
did not know the magnitude of the magnetic field we applied, we could measure it
very accurately from the frequency of the oscillations in P (t). An example of some
data showing this effect can be seen in Fig. 1.4. t (µs)

Fig. 1.4 Muon precession in a field of

2 mT.
The polarization P (t) as a function of time t can be used to tell us
about the magnetic field at the muon stopping site. If the magnetic
field at the muon site fluctuates as a function of time, then we can also
obtain information about the nature of those fluctuations. Much of this
book describes how this and other electronic and chemical properties of
matter can be extracted from the data.
4 The basics of µSR

1.3 Muon beams and spectrometers

In order to obtain enough muons to carry out one of these experiments we
must create them in a high-intensity beam of particles. For this reason
the experiments are carried out at facilities where such beams are made
using particle accelerators. Muons are born 100% spin polarized because
of the way they are produced from the decay of pions via a reaction, as
described in the following example.

Example 1.4
In a muon facility, a beam of protons is fired into a target and this makes pions. The
experiment is usually carried out using pions which are at rest in the surface of the
target and which decay quickly into muons via the reaction
π + → µ+ + νµ . (1.7)
Since this reaction must conserve linear and angular momentum, and since the pion
is initially at rest and has no spin or orbital angular momentum, we deduce that
both the muon (µ+ ) and the neutrino (νµ ) must have opposite linear and angular
momenta. Having opposite linear momenta just means the two particles head off in
opposite direction. Because neutrinos always have a negative helicity (i.e. their spin
is always observed to be antiparallel to their momentum) then the muon must also
have negative helicity (so that its spin is also antiparallel to its momentum). For this
reason, we are able to obtain a beam of fully spin-polarized muons.

cryostat

As mentioned in Section 1.1, µSR measurements are not scattering ex-

periments: the muons are always stopped in the material under inves-
tigation. The beam must therefore be of a low enough energy that the
muon can be stopped in (typically) a few hundred milligrams of solid
µ+ matter. The sample under investigation is placed in the beam, where the
muon beam muons stop in the bulk of the material without any appreciable change
sample in their spin orientation nor damage to the material. This is where the
experiment begins, with a muon at rest in the sample in a known spin
state. Muons are introduced into the sample one at a time or in a pulse.
In either case, the number of muons in the sample is so small that there
detector detector
is effectively no possibility of muons interacting with each other: they
are an ultra-dilute probe of matter.
The sample is usually mounted on a sample stick and placed in a
Fig. 1.5 A schematic of a µSR exper-
iment.
cryostat so that its temperature can be varied. The cryostat has a win-
dow to allow the muon beam to enter. The positrons that emerge after
muon decay are very energetic, travelling at a large fraction of the speed
of light, and easily exit the sample and cryostat. They are detected in
particle detectors placed around the cryostat. The apparatus comprising
this array of detectors is known as a spectrometer. The output from
these detectors is connected to the data acquisition system (which has to
receive signals at a very high rate) and then to the analysis computers.
Because the timing of the signals has to be measured to better than a
nanosecond (and sometimes to a few picoseconds) it is worth remember-
ing that the speed of light is around 0.3 m per nanosecond, so the lengths
 1.3 Muon beams and spectrometers 5

of cables and other paths for electrical signals have to be adjusted very
carefully. A very simple schematic of the experiment, showing the main
features of the spectrometer, is shown in Fig. 1.5 (we will discuss many
more details of the experimental arrangement in Chapters 14 and 15).

Example 1.5
An example of a real muon spectrometer is the HiFi instrument at ISIS, as shown
in Fig. 1.6 and Fig. 1.7. The sample is placed in a cryostat at the centre of the
spectrometer, where it stops muons from the particle beam. The sample is surrounded
by positron detectors, each of which consists of a piece of scintillator material linked
optically through a plastic light guide to a photomultiplier.

Fig. 1.6 An exterior view of the HiFi

Fig. 1.7 A cutaway drawing of the HiFi spectrometer. instrument at ISIS.

Muon spectrometers can be used at various user facilities around the

world and the main ones are shown in Fig. 1.8. Two of these have
pulsed muon beams (ISIS, based at the Rutherford Appleton Labora-
tory in Oxfordshire in the UK and J-PARC, located in Tokai, Japan)
while the other two provide continuous muon beams (PSI, in Villigen,
Switzerland and TRIUMF, in Vancouver, Canada). All four operate as
user facilities and beamtime is available through competitive applica-
tions. There is also a continuous muon beam source MuSIC in Osaka,
Japan, and further proposed muon facilities in Korea, China, and the Fig. 1.8 User facilities where µSR ex-
US are at various stages of planning and/or building. periments can be performed.
6 The basics of µSR

1.4 Experimental geometries

(a) LF-µSR We will now introduce some important pieces of jargon concerning the
type of experiment performed. When a magnetic field is applied to the
B sample, it is usually applied in one of two geometries:

S µ(0) • The magnetic field B is applied parallel to the direction of the ini-
tial muon spin polarization S µ (0). This is known as the longitu-
dinal geometry, and this is often referred to as longitudinal-field
µSR, or LF-µSR for short.
(b) TF-µSR • The magnetic field B is applied perpendicular to the direction of
the initial muon spin polarization S µ (0). This is known as the
B transverse geometry, and this is often referred to as transverse-
field µSR, or TF-µSR for short.

These two geometries are shown in Fig. 1.9(a) and (b). Sometimes,
S µ(0) however, experiments are performed with no applied magnetic field (and
most spectrometers allow the user to make the magnetic field truly zero,
nulling out the Earth’s magnetic field). This configuration is known as
(c) ZF-µSR zero-field µSR, or ZF-µSR for short [Fig. 1.9(c)].

B=0 1.5 What can we do with µSR?

S µ(0) We will spend a lot of this book looking at applications of the µSR
technique which will answer this question. However, in this introductory
chapter let us focus on a key principle: the muon is a local probe. This
means that the muon tells us about its magnetic interactions at the
Fig. 1.9 Experimental geometries for particular point in the material where it stops. The muon often stops in
(a) longitudinal field (LF-µSR), (b)
transverse field (TF-µSR), and (c) zero-
a crystallographically unique position in a solid. However, because we
field (ZF-µSR). measure the result of millions of muon implantations, there is likely to
be a difference in the behaviour probed at the position of each muon.
When we measure the polarization at each time t, we average over all of
the muons that have been implanted in the sample and have interacted
with the sample for that time t.

Example 1.6
Consider an ensemble of spin-polarized muons at rest at t = 0 sitting in slightly
different transverse magnetic fields. They will precess at slightly different angu-
lar frequencies ω = γµ Bi , depending on the value of the field Bi at each muon
site. The corresponding oscillations are progressively dephased as time elapses.
The polarization, averaged over many muons, becomes washed out as time in-
cerases and consequently the precession signal decays. An example is shown in
Fig. 1.4 for which a magnetic field of 2 mT was applied to a sample. The muons
precess at an average frequency of 0.27 MHz (see Example 1.2), but because of
these field variations the oscillations relax slightly as time increases. This relax-
ation can be used to estimate the size of the field variations inside the sample.
 1.5 What can we do with µSR? 7

The decay of polarization described in the previous example is known

as the relaxation of the polarization. There is lots of information hid-
den in the exact form that this relaxation takes, and accessing this in-
formation will allow us several insights into topics such as magnetism,
superconductivity, electronic properties, and chemistry of materials.
Uses of µSR that we will describe in this book include the following:
• The detection of magnetic order.
• The observation of the presence of spin fluctuations.
• The determination of the penetration depth in a superconductor.
• The identification of the pairing symmetry in a superconductor.
• The measurement of the local electronic environment in a semi-
conductor, where the muon simulates a proton impurity.
In Part I of the book we will go through the basic principles of µSR
in more detail. Part II describes specific science applications of the µSR
technique. Part III is devoted to particular practical considerations that
are important when performing µSR experiments. Part IV contains
some more advanced topics in muon spectroscopy, while Part V describes
the relationship between µSR and other, complementary techniques.

Chapter summary

• The muon is a radioactive particle with a magnetic moment and it

can be implanted into a sample.
• The principle of the experiment is based on Larmor precession
of the muon in a magnetic field and the asymmetric emission of
positrons.
• A spectrometer is used to detect the positrons and the experiment
can be performed in longitudinal field or transverse field geome-
tries. Measurements can also be performed in zero field.
• There are many applications of µSR, as we will show in the rest of
this book.
 Part I

Elements of muon
spectroscopy
In this introductory part of the book we set out the basic elements of
the muon spectroscopy technique.

• In Chapter 2 we provide some background to the particle physics

of the muon, as well as some perspectives of how muons can be
used in science.
• When introducing the positive muon µ+ into matter we find vari-
ous types of implanted state: the muon can exist as a bare muon
(positively charged), or can pick up an electronic charge to become
the neutral muonium state, or even pick up two electronic charges
and become a negatively charged species. These states, and the
consequences of their formation, are explored in Chapter 3.
• The muon is a quantum particle and its quantum-mechanical in-
teraction with its surroundings has important implications. In
Chapter 4 we review the basic quantum-mechanical arguments un-
derlying these interactions and discuss the consequences for muon
spectroscopy.
• The experimental observable in a muon experiment is the relax-
ation function of the muon polarization. We give an introduction
to these relaxation functions in Chapter 5.
 2 Introduction

The muon is a fundamental particle with some unique properties that

2.1 Discovery of the muon 10
have given it a special place in science. It has evolved from being a
2.2 The first muon application 11
rather unexpected curiosity, a view taken shortly after its discovery in
2.3 Muon perspectives 12
the 1930s, towards becoming a significant component of modern sci-
2.4 The µSR experiment 17
entific research, where it has a firmly established role at the forefront
Chapter summary 20 of fundamental physics, as well as providing the basis for the powerful
Further Reading 20 muon spectroscopy technique that is the topic of this book.
Exercises 20 The slightly uncomfortable early days of the muon are aptly sum-
marized by the comment “who ordered that?” that was made by the
1
I. I. Rabi (1898–1988). For details of distinguished scientist Isidor Rabi1 shortly after the muon’s discovery.
the origin of Rabi’s remark, see R. H. Rabi would go on to contribute to the founding of nuclear magnetic
Bernstein and P. S. Cooper, Phys. Rep.
532, 27 (2013).
resonance, a technique that is closely related to muon spectroscopy, for
which he was awarded a share of the 1944 Nobel Prize in Physics.
This introductory chapter looks at the muon from a number of dif-
ferent scientific viewpoints, introducing some basic ideas that will be
explored in more detail in the chapters that follow.

2.1 Discovery of the muon

The story of the muon begins with its discovery in California in 1936 by
Carl David Anderson and Seth Neddermeyer. These researchers were
studying energetic cosmic rays via the tracks left behind in a cloud
chamber. Such tracks show a curvature in magnetic field that can be
used to determine the ratio of the momentum of the particles producing
them to their charge, while the particle’s velocity can be determined
from the density of condensed water droplets in the tracks. From these
two properties the mass-to-charge ratio can be deduced. Anderson had
used the same technique a little earlier in the discovery of the positron,
charge spin mass lifetime
(µs) the first example of antimatter, that earned him a share of the 1936
Nobel Prize in Physics.
e ∓e 1/2 me ∞ The newly discovered muon particle was found to have a mass that is
µ ∓e 1/2 207 me 2.197 around 200 times greater than that of an electron and 9 times less than
p ±e 1/2 1836 me ∞
that of a proton (Table 2.1). The muon can have positive or negative
charge and is represented by the symbol µ. On the basis of this mass
Table 2.1 The basic properties of the being intermediate between the light electron, classified as a lepton, and
muon compared with those of the elec-
tron and proton. me is 0.511 MeV/c2 .
the heavy proton, classified as a baryon, the muon was initially classified
The two charge states correspond to the as a meson. The meson had been predicted by Hideki Yukawa in 1935
particle and the antiparticle. as the particle responsible for the strong nuclear interaction. The mass
 2.2 The first muon application 11

of the muon fitted the prediction for the Yukawa meson surprisingly
well, but its relatively weak interaction with matter was inconsistent
with expectations.2 Yukawa’s meson was later identified as the pion, 2
As discussed in Section 2.3, the
represented by the symbol π, which is the short-lived parent of the muon, modern view taken in the Standard
Model of particle physics, formulated
with an average lifetime of 26 ns. The muon itself is unstable with an in the latter half of the 20th cen-
average lifetime of 2.2 µs and positrons or electrons are produced in its tury, classifies hadrons as a family
decay. Thus the important sequence of events in the birth and death of strongly interacting particles formed
cycle of the muon was established: from bound states of quarks. The
hadronic particles can be divided into
baryons, which comprise an odd num-
π ± → µ± → e± . (2.1) ber of quarks (usually three), and
mesons, which comprise an even num-
The properties of the muon are compared in Table 2.1 with the more ber (usually a quark and an antiquark).
familiar charged particles of matter, the electron and the proton. Although Yukawa’s intuition about the
fundamental interactions being medi-
ated by particles was correct, and the
pion is well described by his predic-
2.2 The first muon application tions, the Standard Model identifies the
gluon as the particle ultimately respon-
The muon has the longest lifetime of any unstable elementary parti- sible for the strong force. An outline
cle. This property led to the first experimental verification of the time of the Standard Model is given in Ap-
pendix E.2.
dilation effect predicted by Einstein’s special theory of relativity, thus
providing an early demonstration of the use of the muon in resolving a
scientific question.
This study was carried out by Bruno Rossi and David Hall, who were
researchers based at the University of Chicago. In 1941 they went to
Colorado to measure the count rates for energetic muons travelling ver-
tically downwards at different altitudes. They compared the rates high
up at Echo Lake (altitude 3240 m) with those down at Denver (altitude
1616 m) and found that the lifetime of the muons increased with the
muon momentum exactly as predicted by Einstein. They deduced that
the lifetime of the muon in its inertial reference frame was 2.4(3) µs,
consistent with the more accurate value that we now have today (Ta-
ble 2.1).

Example 2.1
For a relativistic particle with velocity v = βc the time dilation is determined by the
Lorentz factor
γ = (1 − β 2 )−1/2 , (2.2)
and the lifetime of a high energy muon will be significantly increased by this Lorentz
factor, going from τµ in the rest frame of the muon, to γτµ in the rest frame of a
stationary observer. The muon energy is given by E = γmc2 and also by E 2 =
p2 c2 + m2 c4 . Combining these two equations, in units where c = 1, gives
γ = [1 + (p/m)2 ]1/2 . (2.3)
For an energetic cosmic ray muon with momentum p = 800 MeV/c and using the
muon rest mass m = 106 MeV/c2 the Lorentz factor can be evaluated as 7.6 and the
measured muon lifetime therefore increases from 2.2 to 16.7 µs.
12 Introduction

2.3 Muon perspectives

Particle physics

The particle physics view is based on the Standard Model classification of

elementary particles. In this classification the muon is not a meson, but
is itself an elementary particle. The muon and its associated neutrino are
now recognized as two leptons within the six-membered lepton family
(Fig. 2.1). The muon is subject to the electromagnetic interaction and
the weak interaction.
Its decay process is governed by the weak interaction and the decay
produces two neutrinos in addition to the positron or electron, e.g. for
the decay of a positive muon one has

µ+ → e+ + νe + ν̄µ . (2.4)

The weak force has an important characteristic, discovered in 1957 by

C. S. Wu et al., following the proposal of T. D. Lee and C. N. Yang.
This characteristic is that parity (i.e. spatial inversion symmetry) is not
conserved in weak decay processes. This parity non-conservation prop-
erty leads to the negative helicity of neutrinos (see Example 1.4) and to
the asymmetric positron emission shown in Fig. 1.3 that underpins µSR.
Another important feature of eqn 2.4 is that virtually all muons follow
this decay path. Thus muons form a very efficient source of the elusive
Fig. 2.1 The muon is one of six lep- neutrino particles. Such neutrinos can be directed towards special de-
tons within the Standard Model clas- tectors at distant sites to enable mass determination. Purpose-designed
sification of elementary particles. The high intensity muon beams can therefore be employed as ‘neutrino fac-
spin, charge, and mass are shown for
each particle.
tories’ to help address key questions in cosmology such as the origin of
Dark Matter.
The Large Hadron Collider (LHC) at CERN currently provides a ma-
jor focus for high-energy particle physics research and notably provided
the first experimental confirmation of the long-predicted Higgs boson in
2012, with subsequent studies continuing to establish further properties
of the Higgs particle and its interactions. High energy muons are sig-
nificant decay products of the Higgs particle and tracking these muons
forms an important part of these particle physics experiments.
Particle physics is looking towards an extension of the energy frontier
beyond LHC and in this context muon collider designs are actively being
pursued. Muons have an advantage over hadrons in being elementary
particles and they have less radiative energy loss compared to electrons,
which is an important advantage in designing higher energy accelerators
for future studies.
Another scientific frontier that muons are well suited to tackle is the
departure of particle properties from the precise predictions of the Stan-
dard Model. The so-called muon g − 2 experiments have detected sig-
nificant departures of the muon g factor from theory and the accuracy
of the experiments is continuously being improved. Such research may
reveal important new physics that is missing from the Standard Model.
 2.3 Muon perspectives 13

A more detailed description of the muon particle physics that is rel-

evant to implanted muon spectroscopy is given in Appendix E. The
key features are the 100% spin polarization of muons produced by pions
decaying at rest (see Appendix E.1 and E.3) and the asymmetric angu-
lar distribution of the emitted positrons with respect to the muon spin
direction at the moment of decay (Appendix E.4). Both are essential
for the µSR experiment since they give access to the initial and final
directions of the muon spin, and hence its time evolution.

Implanted muons in matter

Moving on from high energy particle physics, we switch our attention to
the study of low energy phenomena in materials with muons, using the
technique known as muon spectroscopy. This technique is also given the
label µSR, where µS stands for muon spin and R can be taken to stand
for rotation, relaxation, or resonance, the three main types of muon
spectroscopy measurement. Alternatively the R in µSR can simply be
taken to stand broadly for ‘research’. Making muons and transporting
them to the sample under study are covered in Chapter 13 and a more
detailed historical perspective of the µSR technique is given by Steve
Cox in Appendix H.
In performing µSR studies it is necessary to slow the muons down
to thermal energies in the process of implanting them in the sample of
interest. This can be done by placing a sufficiently thick sample in the
muon beam or by placing material in front of the sample to slow down
or degrade the beam. Rapid thermalization and preservation of the
muon polarization are key features of the implantation process. Pos-
itive muons are most often used here and the typical muons used in
muon spectroscopy start off with an energy of 4 MeV, whereas the ther-
mal energies associated with the sample are typically on the meV scale,
so considerable energy needs to be given up by the muon. The muon
energy loss starts with atomic ionization and excitation processes and
proceeds to charge exchange, capture, and loss processes. The stopping
process can be simulated using Monte Carlo methods, which are useful
in designing experiments. This subject is covered in Chapter 17.
Muon states are often classified using the terms diamagnetic and
paramagnetic which refer to the magnetic nature of the electrons to
which the muon are coupled (because the electronic moment is much
larger than the muon).3 Thus we have that: 3
This terminology is consistent with
the practice in NMR where the nature
• a diamagnetic muon state occurs when the muon is uncou- of the state one describes excludes the
pled to any unpaired electron density; this could be because the probe spin. In the same way, for µSR,
our diamagnetic/paramagnetic distinc-
muon is not strongly coupled to any electrons [i.e. the muon can be
tion focuses just on the electrons to
described as existing simply as µ+ , often written Mu+ (see Chap- which the muon is coupled, and not on
ter 3)] or because it is coupled to two electrons which are in a the magnetism of the muon. Thus a
singlet state (as in Mu− , to be discussed in Chapter 3). ‘diamagnetic’ state means an ‘electron-
ically diamagnetic’ state and a ‘param-
• a paramagnetic muon state occurs when the muon is strongly agnetic’ state means an ‘electronically
coupled to a single unpaired electron [as in neutral muonium (writ- paramagnetic’ state.
ten Mu0 , see below and also Chapter 3)].
14 Introduction

The nature of the final muon states obtained in the sample will depend
on the type of sample being investigated (Fig. 2.2). In metals the final
muon state usually remains a positive muon, a diamagnetic muon state.
In insulators the diamagnetic positive muon state is one possibility, but
another state retaining an unpaired electron is also possible, a param-
agnetic state. The simplest form of paramagnetic state is muonium,
4
It should be noted that states that a muon analogue of the neutral hydrogen atom. These different types
show this diamagnetic response may of state are identified by their response to a transverse field (TF), i.e.
actually be the result of a paramag- a magnetic field applied perpendicular to the initial muon polarization.
netic local environment with fast fluc-
tuations.
Diamagnetic states produce a muon precession frequency that is very
close to that of a bare muon,4 whereas paramagnetic states show sig-
nificant shift with respect to the bare muon. The various muon charge
states and the transitions between them are discussed in more detail in
Chapter 3.

a) b)

+ +

-
e

+ + 0
Positive muon: or Mu Muonium: Mu or Mu
‘diamagnetic’ ‘paramagnetic’

Fig. 2.2 a) In muon spectroscopy terminology, a positive muon state in a local en-
vironment without unpaired electrons is termed diamagnetic. This state is identified
by a precession frequency in TF that is virtually indistinguishable from that of a
bare muon. b) The case where the TF precession frequency differs significantly from
that of a bare muon is called a paramagnetic muon state. The simplest example of
a paramagnetic state is muonium, where the positive muon captures an electron and
forms a bound state analogous to atomic hydrogen.

Condensed matter physics

A condensed matter physics perspective is at the heart of much of the
muon spectroscopy work carried out at the various muon user facilities
around the world. This perspective views the thermalized, implanted
muons as local probes of static and dynamic magnetism in the sample
under investigation. For studies of static magnetism the muon can often
be regarded as a local probe with a limited set of specific stopping sites.
The data can then be analysed in terms of these particular sites and
complementary site calculations can be made to aid the analysis. These
site calculation methods are discussed in Chapter 16.
Another situation that is often encountered is where the muon can
best be viewed as randomly sampling the internal magnetic environ-
ment within the material under study. A typical example here would
be superconducting vortex lattices having a long length scale compared
to any individual sites in the lattice (see Chapter 9). Magnetic sys-
 2.3 Muon perspectives 15

tems with a large cell and many stopping sites could also fall into this
category, as could systems with a high degree of disorder. 2000
Besides observing static magnetism, the muon can also be used to 9 2000
10 1 fs
study the fluctuating magnetic fields associated with various dynamical
processes in the sample. This will be covered in detail in Chapters 5,
7, and 8. The range of fluctuation rates that can be observed using the
muon depends on a number of factors, which are illustrated in Fig. 2.3.
6 20 1 ps
One factor is the coupling constant A (measured in frequency units) 10

linking the muon to the internal magnetic fields that are fluctuating 20

Fluctuation rate ( s )
-1

Correlation time
with rate ν. As we shall derive in Chapter 5, when the fluctuations are
slow, so that ν < A, the relaxation rate λ of the muon-spin polarization
is proportional to ν. In the opposite limit of fast fluctuations where 10
3
1 ns
ν > A the muon-spin relaxation rate is proportional to A2 /ν. A sec- 0.2
ond factor determining the limits of dynamical sensitivity is the range
0.2
of muon relaxation rates that can be measured by a particular muon
spectrometer, with a maximum relaxation rate λmax determined by the 1 1 s
time resolution of the source and detectors and a minimum relaxation
rate λmin determined by the background at a continuous source and by
the muon flux and detector dead time at a pulsed source. This range
continuous
of measurable relaxation rates defines two bands of sensitivity for each -3
source
10 1 ms
value of the coupling, as shown in Fig. 2.3, one for the slow-fluctuation pulsed
source
limit and one for the fast-fluctuation limit. Notice that the gap between
the two bands is called the wipe-out region, where the fluctuation rate
leads to a relaxation that is too fast for the experimental apparatus and Fig. 2.3 Muon sensitivity to dynam-
the signal is partially or totally missing. ics compared for continuous and pulsed
sources and for three different values
It is useful to make a broad comparison of the range of fluctuation of coupling: 0.2, 20, and 2000 MHz.
rates available via µSR with those accessible using other techniques. These reflect typical values for nuclear
This is illustrated in Fig. 2.4 (on page 16) which shows that there can dipolar coupling, moderate electronic
coupling, and very strong electronic
be considerable overlap between different methods. The µSR technique
coupling. The lower band is the slow
can be seen to provide a useful bridge between characteristically fast fluctuation regime and the upper band
methods such as neutron scattering and characteristically slow methods is the fast fluctuation regime. Pulsed
such magnetic remanence and a.c. susceptibility. and continuous sources are discussed in
Chapter 13.
When comparing µSR against bulk magnetic studies several advan-
tages emerge. As a volume-averaged probe µSR is relatively immune to
low concentration impurity phases that can dominate bulk susceptibil-
ity measurements at low temperature and phase segregated samples can
easily be identified in muon data. µSR is also particularly well-suited
to studying antiferromagnetism and allows measurements to be made
in zero applied field. Another strong feature of µSR is its high sensi-
tivity to very small magnetic fields and magnetic moments, typically
being able to detect fields below 10 µT and moments below 0.01 µB .
The technique has excellent compatibility with very low temperatures
as a result of the incoming muons being able to penetrate the cryostat
windows and the outgoing positrons being able to penetrate both the
walls and windows of the cryostat. Further comparison of µSR and bulk
measurement techniques is given in Chapter 23 and the complementary
techniques of neutron and X-ray scattering are covered in Chapter 24.
16 Introduction

The many areas of condensed matter physics addressed by muon spec-

troscopy can be grouped into several major topics. Subsequent chapters
10
12
10
-12 cover magnetism (Chapters 6 and 7), dynamics (Chapter 8), supercon-
ductors (Chapter 9), semiconductors (Chapter 10), and ionic motion
a.c. susceptibility (Chapter 11).
NMR

9 -9
10 10
Fluctuation rate (s )

Correlation time (s)

-1

Magnetic resonance
neutrons
Mossbauer

10
6
10
-6
Another perspective of the muon views µSR as a member of the extended
family of magnetic resonance techniques. Properties of the muon as
a magnetic probe are summarized in Table 2.2 in comparison with those
remanence

3 -3
10 10
of the electron and the proton.
SR

As a magnetic resonance technique, µSR most closely resembles NMR.

1 1 In comparison with NMR there are many advantages of µSR, along with
several disadvantages. Some of the advantages of µSR can be listed as:

10
-3
10
3 • Having 100% spin polarization, in contrast to the very weak ther-
mal polarizations obtained for NMR nuclei.
Fig. 2.4 Illustration of typical sensitiv-
• A larger gyromagnetic ratio for the muon than for any nucleus.
ity to dynamics on different timescales • No need for specific nuclear isotopes, with muons being able to
for a range of methods in comparison take measurements in any material.
with µSR.
• Being able to use an extended field range for measurements, span-
spin magnetic γ/2π method ning from zero field (ZF) up to 10 T (in the current generation of
moment (MHz T−1 ) high field muon spectrometers).
e 1
2
658.21 µp 28025 ESR • Compatibility with low temperature sample environments (e.g. he-
µ 1
3.1833 µp 135.53 µSR lium dilution refrigerators).
2
p 1
µp 42.576 NMR • Potential access to a very wide range of correlation times, as
2
discussed in the previous section.
Table 2.2 The magnetic properties On the other hand disadvantages of µSR compared to NMR can be
and magnetic resonance method of the listed as:
muon compared with those of the elec-
tron and proton. µp is 1.4106 × • Limited potential for spin manipulation by RF pulse sequences due
10−26 J T−1 . NMR is nuclear magnetic
resonance; ESR is electron spin reso-
to the short muon lifetime.
nance. • The muon stopping sites need to be determined to fully inter-
pret the data.
• There will be some perturbation of the system under study by the
presence of the muon.
The relation between µSR and other forms of magnetic resonance is
covered in more detail in Chapter 23.

Chemistry
The chemistry perspective is based on the realization that the chem-
ical properties of Mu are analogous to those of the H atom, which is
the simplest form of free radical and is thus strongly chemically reac-
tive. Muonium therefore provides a highly reactive radioactively labelled
light isotope of H. Studying Mu via muon spectroscopy allows the time
 2.4 The µSR experiment 17

dependence of chemical reactions of a hydrogenic species to be deter-

mined and transient initial states of reactions can also be followed. The
existence of Mu significantly extends the hydrogen isotopic sequence of
H, D, and T.
A further feature of muonium chemistry is that the low mass of Mu
leads to large quantum zero-point energy so that Mu probes the up-
per reaches of the interatomic potential energy surface (PES) and can
therefore provide a thorough test of chemical reaction theory. Vibra-
tional excitation (e.g. with a laser) can be used to take the muon even
higher up the PES (at least three times higher than the ground state).
Muon spectroscopy using photoexcitation is covered in Chapter 19 and
chemistry applications of muon spectroscopy are covered in Chapter 12.

Computational science
From a computational science perspective, the implanted muon prob-
lem is a variation on the hydrogen interstitial defect problem. Well-
established computational methods such as Density Functional Theory
(DFT) can be used to calculate the structure of the interstitial muon de-
fect including the atomic relaxation around the muon site. The strong
quantum character of the muon provides an extra challenge for computa-
tional studies, since standard computational methods such as DFT only
give the classical site. Further computational procedures are required to
properly model the quantum delocalization of the muon. Various com-
putational approaches to determining muon sites and their properties
are discussed in Chapter 16.

2.4 The µSR experiment

A view of the µSR experiment could perhaps start with understand-
ing the basic features of a muon spectrometer, first introduced in the
previous chapter. The muon spectrometer comprises a set of positron
detectors placed around the sample to measure the time evolution of the
muon spin polarization. In addition, magnetic coils are arranged around
the sample position to supply external fields at the sample. Muon in-
strumentation is covered in more detail in Chapter 14, but here we
just consider the simplest detector arrangement, as shown in Fig. 2.5a).
The asymmetry of the positron emission is represented by the function
Γ(θ) = 1 + a cos θ, where a depends on the energy range of positrons
seen by the detectors (see Appendix E.4). In this simple two-detector
geometry a forward detector (F) is placed in the direction of the initial
spin direction and a backward detector (B) is placed in the opposite ori- 5
As discussed in Appendix E, this is
entation. The initial asymmetry a0 seen by each detector is the average lower than the theoretical maximum of
of Γ(θ) − 1, computed over the detector surface facing the sample. In 1/3.
practice5 we find a0 = 25%. The magnetic field direction is defined as
the z-axis and in the longitudinal field (LF) orientation the initial spin
direction is also along z. In this case the respective signals in the F and
18 Introduction

a) b) LF TF
+ F
e

Detector signal
F detector F
B

muon 0
spin
+
0.2
a0 Pz(t) a0 Px(t)

A(t)
1 + a cos 0

-0.2

B detector 0 2 4 6 0 2 4 6
t ( s) t ( s)

Fig. 2.5 a) Positron emission has an angular probability weighted towards the muon
spin direction. A pair of detectors placed forward (F) and backward (B) with respect
to the initial spin direction are used to measure the spin polarization. b) The time
evolution of the F and B signals and the asymmetry function in the case of fields at
the muon site that are either longitudinal or transverse with respect to the initial
direction of the muon spin. The dotted lines in the upper plots represent the average
signal.

B detectors, NF (t) and NB (t), are given by

NF (t) = NF0 exp(−t/τµ )[1 + a0 Pz (t)],
NB (t) = NB0 exp(−t/τµ )[1 − a0 Pz (t)], (2.5)
where NF0 and NB0
are scaling parameters for the count rates of the F
and B detectors, Pz (t) is the longitudinal polarization function which
is equivalent in this geometry to the longitudinal relaxation function
Gz (t) describing the relaxation of the muon spin in response to internal
fields in the sample. The distinction between the polarization and the
relaxation function will become clear when we discuss the transverse
field experimental configuration. The polarization function is covered in
detail in Chapter 5. In the detector signals the polarization function is
superimposed on a large, exponentially decaying signal due to the muon
lifetime [Fig. 2.5b)]. In order to extract the polarization or relaxation
function, which is the key quantity of interest in muon spectroscopy, the
forward-backward asymmetry A(t) is evaluated as
NF (t) − NB (t)
A(t) =
NF (t) + NB (t)
(NF0 − NB0 ) + (NF0 + NB0 )a0 Pz (t)
= . (2.6)
(NF0 + NB0 ) + (NF0 − NB0 )a0 Pz (t)
If the detectors are completely balanced in terms of geometry and sen-
sitivity then NF0 = NB0 and a straightforward result is obtained
A(t) = a0 Pz (t). (2.7)
 2.4 The µSR experiment 19

In practice, perfect balance is not possible and this results in different

scaling factors for the two detectors. Compensation for this difference is
then made by including a balancing factor α = NF0 /NB0 in the expression
for A(t), i.e.
NF (t) − αNB (t)
A(t) = = a0 Pz (t). (2.8)
NF (t) + αNB (t)
The balancing factor is generally estimated for muons in a diamagnetic
state by applying a small transverse field and adjusting α to give an A(t)
that oscillates symmetrically about zero. This F,B detector configura-
tion is also used in the case where no external magnetic field is applied,
6
known as zero field (ZF) mode. It should be noted that the F and B For example, the physical F detector
detectors here are logical detectors that must be mapped against the in a spectrometer is downstream and
the physical B detector is upstream,
physical array of detectors in the instrument. For a muon instrument at whereas for surface muons (see Chapter
a pulsed source these logical detectors typically represent a large group 13) the initial muon polarization points
of physical detectors. In the case of a continuous source such grouping upstream. The logical detector map-
ping in this case inverts the F and B
is not always required, but mapping from physical to logical detectors
detectors so that eqn 2.8 can be used
is still needed for dealing with different experimental configurations and to represent a positive initial asymme-
allowing eqn 2.8 to be used consistently in each case.6 try.
When in TF mode the initial polarization direction is defined as the
x-axis, which is transverse to the magnetic field BTF that is directed
along the z-axis. In this case the polarization function describing the
asymmetry becomes the transverse polarization function Px (t), which
can be represented as the product of an oscillation term and a transverse
relaxation function Gx (t)

Px (t) = cos(γµ BTF t)Gx (t). (2.9)

For TF mode it is beneficial to have a number of detectors or detector

groups distributed evenly around the field axis to maximize the signal,
=0
typically up to sixteen detector groups would be used to avoid signifi- i

cant signal loss through dephasing. The polarization function for each

i
=
45
detector then involves a phase angle φi reflecting the phase offset with
respect to the x-axis initial polarization direction (Fig. 2.6) +

BTF
Pi (t) = cos(γµ BTF t + φi )Gx (t) (2.10)

and the corresponding detector signals are given by

Ni (t) = Ni0 exp(−t/τµ )[1 + Ai (t)] (2.11)

where Ni0 is a scaling parameter for the detector count rate in the ith
detector, and the detector group asymmetry functions are defined as Fig. 2.6 Example arrangement of
Ai (t) = ai Pi (t), with the individual asymmetry parameters ai specified eight detector groups optimized for TF
to allow for variations in detector group geometry. mode, indicating the detector phase φi .
The muon polarization and relaxation functions describing the asym-
metry are the core functions of the muon spectroscopy technique and
the mathematical forms that they take in various experimental situa-
tions are described in Chapter 5. A quantum mechanical approach to
understanding the evolution of the muon polarization is also presented
in Chapter 4.
20 Exercises

Chapter summary

• The muon is a spin-1/2 lepton that interacts with matter and has
a finite average lifetime of τ ≈ 2.2 µs.
• When stopped in a sample of matter, the muon can form diamag-
netic and paramagnetic stopping states. Its spin interacts mag-
netically with its environment which, along with the nature of its
decay, forms the basis of µSR spectroscopy.
• A µSR experiment can be carried out in LF, ZF, and TF modes.
These measurements rely on the key concepts of muon polarization,
relaxation, and asymmetry.

Further Reading
• A. Yaouanc and P. Dalmas de Réotier, Muon Spin Rotation, Re-
laxation, and Resonance: Applications to Condensed Matter, OUP
(2011).
• S. J. Blundell, Contemp. Phys. 40, 175 (1999).
• S. F. J. Cox, J. Phys. C 20, 3187 (1987).
• K. Nagamine, Introductory Muon Science, CUP (2003).
• F. Halzen and A. D. Martin, Quarks and Leptons, John Wiley and
Sons (1984).
• S. L. Lee, R. Cywinski, and S. H. Kilcoyne (editors) Muon Science:
Muons in Physics, Chemistry and Materials (Scottish Graduate
Series), CRC Press (1999).

Exercises
(2.1) Calculate the average path length for the cosmic (2.3) A TF muon experiment is being carried out on a
ray muon in Example 2.1. What would the path pulsed muon source instrument with a large num-
length be if there was no time dilation? ber of detectors available in the muon spin rotation
plane. A choice of detector grouping must be made:
(2.2) A muon instrument is capable of measuring muon either four detectors at 90◦ intervals, eight detec-
relaxation rates as fast as 20 µs−1 and as slow as tors at 45◦ intervals, or sixteen detectors at 22.5◦
1 ms−1 . Calculate the fastest fluctuation rate that intervals. By considering the loss of asymmetry due
can be measured using this instrument when the to signal dephasing across each detector group, de-
coupling between the muon and the fluctuations is termine which of the groupings keep the fractional
2 MHz. asymmetry loss below 3%.
Muon charge and spin
states 3
The µSR experiment begins with a spin-polarized muon stopped in a 3.1 State formation 22
material. In this chapter we discuss the states that the stopped muon can 3.2 Hydrogen analogues 24
adopt, along with how these come to be realized during the implantation 3.3 Measuring the states 26
process. This is important as the stopped muon in its local enviroment 3.4 Influencing the states 27
is the fundamental probe that forms the basis of the measurement. A Chapter summary 28
knowledge of the muon’s stopping state (the muon fate) is the starting Further reading 28
point that allows us to compute the effect of the interactions with the Exercises 28
local environment on the muon-spin polarization, which is the subject of
the next two chapters. More precisely, the energy scale that determines
the muon fate is predominantly due to its charge, but the signature
in the µSR experiment is determined by its spin state, diamagnetic or
paramagnetic, according to Section 2.3. We shall cover mostly charge
aspects in this chapter and consider spin in more detail in Chapter 4.
Let us start with the muon at rest in the material.1 The stopped 1
Later in the book we will consider
muon is found to exist in three different charge states. These are shown cases in which the muon can hop from
site to site, see Chapter 8.
in Fig. 3.1 and summarized below.

e− e−
µ+ µ+ µ+
2 +
e− H is sometimes called hydron, but it
is simply a proton p+ .
Mu+ Mu0 Mu−
3
There is a difference between the
Fig. 3.1 The charge states of the muon: Mu+ , Mu0 , Mu− . physics and chemistry communities
with notation. Physicists, schooled in
particle physics, are more often likely
• Mu+ : The muon can exist on its own as a positively charged to write the species as µ+ . Chemists,
particle µ+ . This is analogous2 to the hydrogen cation H+ , and who are more aware of the analogy with
hydrogen, tend to write Mu+ . In this
for this reason this diamagnetic species is often written3 with the book, we will use Mu+ in the chapters
symbol Mu+ . that are concerned with semiconduc-
• Mu0 : The muon can pick up an electron, and form an electrically tors and chemistry (where the hydrogen
analogy is more relevant) but revert to
neutral bound state µ+ e− called muonium, given the symbol Mu. the physicist notation elsewhere, which
This is analogous to the hydrogen atom H. To avoid any ambiguity, is in keeping with the literature.
we can write Mu as Mu0 to emphasize that it is in the neutral,
paramagnetic, state. Note that the chemically stable version of
hydrogen is H2 , molecular hydrogen, but the muon flux is never
large enough to produce Mu2 .
22 Muon charge and spin states

• Mu− : The muon can pick up two electrons and form the nega-
tively charged bound state Mu− which is analogous to the hydride
ion H− . This state is diamagnetic (an electron singlet).
These different charge states can be thought of as distinct chemical
species, each determined by the nature of the electronic structure around
the muon. However, the three states Mu+ , Mu0 , and Mu− are idealiza-
tions. Even after complete muon thermalization, the final state may be
t (s) Ekinetic
weakly bound to neighbouring atoms or there may be screening effects in
+
production 4.1 MeV
a metallic sample. This means that the actual spin and charge densities
may depart from the simple pictures associated with these idealizations.
sample surface
But even in unsaturated molecules and magnetic solids, where the spin
-9
density becomes more and more delocalized, Mu± and Mu0 are use-
10 ionization of atoms
scattering with electrons ful proxies. We will therefore formulate our discussion in terms of the
10 keV
charge and spin state of the implanted muon, but it is important to keep
-13
10 muonium formation via
electron capture/loss
in mind that the muon is a chemical species (see Chapter 12) as well as
10 eV a magnetic entity.
-12
10 muonium-atom
collisions
0 eV
dissociation of muonium 3.1 State formation
interstitial+ thermal Implantation process
thermal muonium

Muons are usually implanted at energies of ≈ 4 MeV. A schematic of the

Fig. 3.2 The processes leading to a implantation process that leads to a stopped muon is shown in Fig. 3.2.
stopped muon following implantation
of surface muons.
The dominant energy loss process by which the muons slow down is ion-
4 ization of the sample atoms.4 The muon may be able to capture one of
This process can be described by the
Bethe-Bloch equation, using ideas de- these electrons, but it is likely to be stripped off again by a subsequent
veloped by Bethe and Bloch in the early interaction. The charge cycling will stop once the muon’s kinetic energy
1930s. drops below the ionization energy, typically of order 10 eV, with the re-
maining energy lost as phonons. The charge cycling and slowing down is
of order picoseconds and the muon’s spin is essentially unaffected during
this time because the Coulomb interaction does not couple to its mag-
netic moment, and the time spent as a neutral atom is far too short for
the hyperfine interaction to have any effect. The muon then comes to
rest slightly downstream of the track of ions, with a certain (material-
dependent) probability of having an electron so forming neutral muo-
5
The electron affinity is ≈ 0.75 eV, nium (Mu0 ). The relatively fragile negative ion5 (Mu− ) is unlikely to
and this is the (relatively small) en- be formed promptly in this way.
ergy needed to remove one of the elec-
trons from Mu− and turn it back into
Mu0 . Note that Mu0 is much more sta- Prompt processes
ble since ≈ 13.6 eV are needed to ionize
it to Mu+ . Once it has stopped, the muon may undergo rapid charge-state changes
or reactions on a timescale of order a few nanoseconds, i.e. before reliable
6
This is not uncommon: any exper- data acquisition allows us to follow its spin relaxation.6 The positive
imental detection has an initial blind ion (Mu+ ) will try to find a location with high electron density such
window.
as a lone pair on an atom (e.g. oxygen), a bond centred site, or close
to an electronegative ion such as fluorine. It may attach to an ion or
molecule, for example an oxide ion to form a singly negative hydroxide
ion, or a water molecule to form the hydronium ion H2 MuO+ (H3 O+ ). In
 3.1 State formation 23

covalent solids Mu+ may also promptly form a (spin singlet) bond, which
implies a screening charge which is spin-unpolarized. This is typically
the case in non-magnetic metals. In these cases our earlier classification
by charge becomes less stringent, although, in insulators, there may
still be a net positive charge overall. In molecular liquids and solids the
neutral atom may react with a neighbouring molecule, perhaps attacking
a double bond or lone pair site.7 Alternatively it may find an interstitial 7
See Chapter 12.
space where it can remain as a free atom (this indeed happens also in
instrinsic semiconductors8 ). In these two cases charge state changes 8
See Chapter 10.
may occur if the muon is able to capture an electron from the material,
perhaps one of those liberated along the stopping track. An electron
may also be returned to the material (e.g. to the conduction band) if
appropriate.
If the muon is in the neutral (paramagnetic) state for a time compara-
ble to the inverse hyperfine constant, before converting to diamagnetic
Mu+ or Mu− , it will dephase in transverse field.9 The result is a loss of 9
We will examine the dynamics of neu-
apparent amplitude, accompanied by a phase shift. At low field when tral muonium in more detail in Chap-
ters 4, 10, and 12.
the Mu0 fraction would give triplet preccesion, the phase shift and am-
plitude will be field-dependent, since the relevant timescale is the period
of the triplet precession. There will also be loss of amplitude in zero
field if the muon’s electron is coupled to nearby nuclear spins.10 These 10
A similar dephasing occurs if Mu+ is
effects give the missing fraction where the total oscillating amplitude slow to convert to Mu0 which is then
observed in transverse field.
in a transverse field measurement does not add up to the full asymme-
try expected of the instrument. The missing polarization is recovered in
longitudinal field repolarization measurements.11 11
Repolarization will be explored in
Section 4.4.

Slower processes

Charge state changes may continue through the muon data-taking win-
dow (which is ∼ 10µs). Reactions, site changes, and charge exchanges
may be slowed down by energy barriers, and electrons or holes from the
muon track will take time to diffuse to the muon. These changes gen-
erally result in relaxation of the muon polarization, usually exponential
in form. In the case of Mu0 converting to Mu+ or Mu− , in longitu-
dinal field, there is no immediate change of polarization. The change
can still be detected by techniques such as muon spin resonance with a
delayed RF pulse. In high longitudinal field, the conversion of Mu+ to
Mu0 causes only a small polarization change, but with the same time
constant as in low longitudinal field or transverse field.
Charge cycling can occur at higher temperatures, for example in semi-
conductors where sufficient numbers of both holes and electrons are
available for interaction. This will still be observed as an exponential
relaxation. At high longitudinal field it will take many charge cycles
to completely depolarize the muons, so a field scan will give relaxation
to zero polarization at any field, with the relaxation rate falling off at
high field. Analysis of this kind of data can give the charge cycling time
and relative instantaneous occupation of the different states, keeping in
24 Muon charge and spin states

mind that whenever charge exchange is effective also spin exchange may
take place.

Metals
Among all materials, metals stand out because muonium (Mu0 ) is never
seen in any of them. The muon’s electron orbitals overlap with the con-
duction electrons, with rapid spin exchange ensuring that any hyperfine
interaction is averaged away and the muon appears to be in a diamag-
netic state (like Mu+ ). The conductivity should ensure that the muon’s
point charge is effectively screened on a longer scale, though there may
12 be local distortion and charge redistribution around the muon site, and
Some of these will be discussed fur-
ther in Chapter 23. this has some experimental consequences.12

3.2 Hydrogen analogues

13
See Exercise 3.1 for the derivation.
The full quantum mechanical solution The previous discussion highlighted the role of paramagnetic Mu0 . It
for the hydrogen atom (the proper way will be helpful now to consider how it differs from hydrogen. The Bohr
to treat the problem) yields identical
results, but the derivation is much more model for the hydrogen atom makes the following predictions.13 The
involved. energy levels are given by E = −(~2 /2me a20 )(1/n2 ) with n = 1, 2, . . .
where the Bohr radius a0 is given by a0 = 4π0 ~2 /(me e2 ). The ionization
14
The reduced mass is conventionally energy of hydrogen is then given by the energy to take an electron from
given the symbol µ, which we use here the lowest level (n = 1) out to the continuum (n = ∞, where E = 0)
and in the following example, but in the and is equal to ~2 /(2me a20 ) = 13.6057 eV. However, in all these formulae
rest of the book we will retain the sym-
bol µ to refer to either the muon or to
we have used the electron mass me and so we have been assuming the
magnetic moment. mass of the proton is infinite. Strictly, we should be using the reduced
15 mass14 µH . This has implications15 for not only hydrogen, but all the
The implications for the basic prop-
erties are as follows: The Bohr ra- hydrogen analogues, including muonium.
dius aH is inversely proportional to
the reduced mass µH [because aH =
4π0 ~2 /(µH e2 )]. The ionization energy
is directly proportional to the reduced
mass [because the ionization energy = Example 3.1
~2 /(2µH a2H ) and aH ∝ 1/µH ]. The nu- The reduced mass µH for hydrogen is given by
merical values are listed in Table 3.1, me mp
and one can see that the variation is µH = , (3.1)
me + mp
relatively small.
and appears in the formula for the Bohr radius (a0 ∝ µ−1 H ) and in the ionization
energy (∝ µ−1H ). However, because m p  m e , µH ≈ m e (1 − me /mp ) and so the
correction for the reduced mass is small. We can carry out the calculations of these
16 various quantities for the isotopes of hydrogen (hydrogen [also called ‘protium’] 1 H,
A further property is the electron
affinity, the energy released when an deuterium 2 D, and tritium 3 T) and add in muonium as a light isotope of hydrogen
electron is added to a neutral species X and the results are shown in Table 3.1.
to form the anion X− . This quantity
is more complicated to evaluate, but
for similar reasons as we found for the
For all of the species listed in Table 3.1 the mass of the particle at
ionization energy, this takes the same
value (≈ 0.75 eV) for all the hydrogen the centre is much greater than that of the electron. They all have unit
analogues, including muonium. charge, and are essentially point particles with size much less than the
wavefunctions of electrons bound to them. Therefore the chemical prop-
erties of all these species are very similar.16 Therefore we can conclude
that a positive muon at rest inside a sample will behave very much like
 3.2 Hydrogen analogues 25

Property Mu H D T

Mass (mH ) 0.1131 1.000 1.998 2.993

Reduced Mass (me ) 0.99519 0.99946 0.99973 0.99982
Ionization energy (eV) 13.540 13.598 13.602 13.603
Bohr Radius (Å) 0.53173 0.52946 0.52932 0.52927

Nuclear moment (µN ) 8.8906 2.7928 0.8574 2.9788

1 1 1
Nuclear spin I 2 2
1 2
γ/(2π) (MHz T−1 ) 135.53 42.576 6.536 45.413
Hyperfine constant (MHz) 4463.3 1420.4 327.4 1516.7

Table 3.1 Properties of muonium, hydrogen (protium), deuterium, and tritium. The
first set of properties are related to the chemical properties (reduced mass, ionization
energy, and Bohr radius) and are all rather similar. The second set of properties
relate to the nuclear magnetism and here there are large differences. 17
The magnetic moment µ and the an-
gular momentum ~I are proportional
to each other, and the proportionality
hydrogen and attempt to reach chemical equilibrium with its surround- constant is the gyromagnetic ratio. See
ings. This equilibrium state may be a positive or negative “hydrogen” Section 4.1.
ion, a neutral atom, or chemically bound to a molecule or a lattice.
18
Where the hydrogen analogues differ strongly is in their nuclear mag- This interaction is only important for
netic properties. They have different magnetic moments µ = γ~I, where neutral muonium (Mu0 ) because it con-
tains an unpaired electron spin, and
the nuclear spin I is 12 (except for the case of deuterium where I = 1), thus has the possibility of a hyperfine
and γ is the nuclear gyromagnetic ratio.17 A quantity that will be very interaction. It is thus said to be para-
important in subsequent chapters is the contact hyperfine constant A magnetic. Both Mu+ and Mu− are
diamagnetic because neither contains
which expresses the size of the interaction between the nuclear spin and
any unpaired electron spins [Mu+ has
the electron.18 As shown in the following example, the most important no electrons and Mu− has a pair of elec-
contribution to the variation in A is the variation in γ (as well as the trons (which in the ground state will
inclusion of a factor that goes like I + 12 that is relevant for deuterium). exist in a S = 0 state)].

Example 3.2
The contact hyperfine interaction between a nuclear moment µnuc = ~γN I and an
19
electron, where19 gs µB = ~γe , can be written The g-factor of the electron gs ≈ 2.
In this example, we take the nuclear
Ĥ = −µnuc · B e = AI · S, (3.2)
spin operator I and the lectron spin op-
where B e is the magnetic field created by the electron. The electron is unpaired erator S to be dimensionless, so that A
(there is just one electron in the hydrogen atom, not two) and it has a spin. It thus has the dimensions of energy.
gives rise to a magnetization M (r) = −gs µB S|ψ(r)|2 and we are only interested in
the value this takes inside the nucleus, which is M (0) = −gs µB S|ψ(0)|2 . We can
consider the electron density inside the nucleus as a uniformly magnetized sphere
and this gives rise to a magnetic induction B e = µ0 M (0) − µ0 M (0)/3, where the
20
correction −µ0 M (0)/3 is due to the demagnetization field of a magnetized sphere.20 The demagnetization factor for a
Putting this all together yields sphere is N = 13 and the demagnetizing
2 correction to the B-field is −µ0 N M .
B e = − µ0 gs µB |ψ(0)|2 S. (3.3) This is discussed further in Chapter 6.
3
For a 1s electron, we have that ψ(r) = e−r/a0 /(πa30 )1/2 and hence |ψ(0)|2 = 1/(πa30 ).
We hence end up with an expression for the hyperfine constant as
2 1
A = µ0 gs µB ~γN 3 . (3.4)
3 πa0
26 Muon charge and spin states

Coupling together an electron spin (for a 1s orbital J = 12 ) and a nucleus (I) leads
to a total angular momentum F = I + J and since F 2 = I 2 + J 2 + 2I · J we have
that the hyperfine energy levels are
21 1 A
These are the levels F = I − 2
and E = AI · J = [F (F + 1) − I(I + 1) − J(J + 1)]. (3.5)
2
F = I + 12 .
Since J = 12 , this gives two possible levels21 separated by A(I + 12 ), so for I = 12 (rel-
22
As is often the case in spectroscopy, evant for hydrogen and muonium) the levels are separated by the hyperfine constant
we will use the same symbol A to re- A. For hydrogen, this value is 5.87 × 10−6 eV. Dividing by h gives a frequency,22
fer to the energy gap between the two which is AH = 1420.4 MHz, as written in Table 3.1. For muonium, we need to scale
levels and for the frequency that cor- up AH by the γ factor, so we might estimate AMu ≈ AH γµ /γp = 4521 MHz which
responds to this energy gap (using the is within just over 1% of the right answer, but we can get pretty much spot on23 by
formula E = hf ). also including the small change in the Bohr radius between H and Mu.
23
See Exercise 3.2.
In conclusion, we have seen that the chemical properties of muonium
are very similar to hydrogen, but the nuclear magnetic properties are
very different. Muonium will therefore have a unique spectroscopic sig-
nature which can be studied in experiments, and as we will see in Chap-
ters 10 and 12 its environment within the solid state or within a chemical
species will also modify the hyperfine interaction.

3.3 Measuring the states

For a practical experiment on a new sample, we need to know the charge
and spin state(s) of the muon in order to interpret the data, especially if
the muon is intended to observe properties such as magnetism or ionic
motion. The simplest check is to observe the amplitude of the Larmor
precession for the bare muon, at low transverse field (2 mT is typical)
and, in magnets, above any ordering transition. Any missing amplitude
must correspond to paramagnetic Mu0 either as the free atom or bound
in a radical, or a missing fraction initially formed as one of these but
converted to another state.
In favourable cases, typically samples with few or no nuclear moments,
a triplet muonium precession signal may be observed from interstitial
24
The principle behind transverse field Mu0 in low transverse field.24 If prompt conversions are suspected, a scan
measurements will be described in Sec- of low transverse fields may reveal amplitude and phase shifts. From this
tion 4.4. Typically 0.2 mT is used at a
pulsed muon source giving a frequency
one can calculate the mean lifetime of the initial state. A longitudinal
of approximately 2.8 MHz. field scan gives the repolarization25 of the radicals or muonium and will
25
Repolarization will be discussed in
give the fraction in that state plus the approximate hyperfine constant.
Section 4.4. Scans are performed typi- Repolarization also includes any missing fraction. The shape of the curve
cally from zero to around 0.2–0.5 T. will be affected by dynamics or anisotropy.
Distinguishing between Mu+ and Mu− from the muon data alone is
more difficult since neither has an electronic moment. The Mu+ will be
bound closer to other atoms so may show a larger nuclear dipole coupling
in zero field. Unfortunately this is not a good guide in oxides where a
Mu+ is likely to have oxygen as its nearest neighbour and the fraction of
magnetic 17 O is very small. Varying the doping of a semiconductor and
observing a change in the diamagnetic fraction can imply which state is
present. For example, a more n-type sample will tend to favour Mu−
over Mu+ .
 3.4 Influencing the states 27

3.4 Influencing the states

Generally the charge states are fixed for a particular material and rel-
atively similar for a series. In some cases it is possible to influence the
balance of charge states to enhance the signal from a particular state.

Example 3.3

26
• A time-independent electric field26 E applied to a sample will sweep the See also Chapter 19.
charge carriers from the muon track away from the stopped muon. In the
usual E = 0 case the carriers form a cloud around the track and diffuse to
the muon. Lower fields in the direction of implantation will sweep the carriers
past the stopped muon and may enhance the interaction. Generally, a scan of
E through zero is somewhat asymmetric.
• Illumination of an insulating or semiconducting sample can generate pho-
tocarriers which will interact with the muon. Often this causes rapid charge
exchange cycles and therefore relaxation, rather than just forming a stable
charge state.
27
• Semiconductor27 samples can be doped n- or p-type and the resulting avail- See also Chapter 10.
ability of electrons or holes will influence the prompt fractions.
28
• Solution28 experiments, with the material of interest dissolved in a solvent See also Chapter 12.
which does not react with muons, will have initial charge state fractions de-
pendent on the solvent. The radical fractions formed by reaction with the
solute scale with the initial Mu0 fraction in the solution. Similarly gas phase
experiments often use an inert buffer gas to decrease the muon stopping range,
either to study the concentration dependence of a reagent or simply that the
vapour pressure of the sample alone is too low. Choice of the buffer gas can
increase the fractions of the wanted charge state, usually Mu0 .
• Low energy muon experiments, with implantation energy of order 3–50 keV
rather than 4 MeV, will have a much shorter track and fewer charge carrier
pairs formed. The electron capture and stripping may not take place to the
same extent. As a result the relative fractions of different species in a low
energy experiment can be different to a muon experiment performed on a bulk
sample of the same material, even without any surface-related phenomena in
29
the sample.29 These will be considered in more de-
tail in Chapter 18.

Now that we have a picture of the states in which the stopped muon is
found, we shall turn to the magnetic interaction of its spin with the local
environment. This is the subject of the next chapter, where we discuss
how quantum mechanics can be used to describe the time-evolution of
the muon spin.
28 Exercises

Chapter summary

• The stopped muon can exist as the positively charged ion Mu+ ,
the neutral muonium Mu0 , or the negatively charged Mu− .
• Neutral muonium Mu0 is a paramagnetic species and has very sim-
ilar chemical properties to atomic hydrogen H0 , but has a very
different hyperfine constant.
• Both Mu+ and Mu− , the diamagnetic species, are also observed
in experiments. In some cases, the various charged species can
interconvert.

Further reading
• The following review contains a good discussion of the different
charge states of the muon observed in semiconductors: S. F. J.
Cox, Rep. Prog. Phys. 72, 116501 (2009).

Exercises
(3.1) In the Bohr model of the hydrogen atom, the energy and the minimum energy is
E of an electron is given by
L2 Ze2 Z2

1

E= 2
− , (3.6) E=− me c2 α2 , (3.8)
2me r 4π0 r n2 2
where Z is the nuclear charge (equal to 1 for hydro-
gen). The first term is the angular kinetic energy
and the second term is the electrostatic potential where α = e2 /(4π0 ~c) is the fine structure con-
energy. By setting the angular momentum L to stant.
be n~, minimize E and show that at the minimum (3.2) Verify the scaling between the hyperfine constants
value r = a0 n2 /Z, where in hydrogen, deuterium, tritium, and muonium us-
4π0 ~2 ~ ing the values of Bohr radius and nuclear spin prop-
a0 = = , (3.7) erties in Table 3.1.
me e2 mcα
The quantum muon 4
The muon is a spin- 12 particle and is therefore a quantum object. To
4.1 Larmor precession 29
understand its properties, and the properties of muonium (described in
4.2 Density matrices 33
the previous chapter), we need to use some quantum mechanics. This
4.3 Mixed states 34
chapter provides a review of the quantum mechanics of angular momen-
4.4 Two spins: muonium 36
tum and works through the problem of a muon coupled to an electron
4.5 Multiple spins 44
in muonium, followed by the problem of the muon coupled to a nucleus.
Chapter summary 51
We begin with a discussion of magnetic moments and spin precession.
Further reading 51
Exercises 52
4.1 Larmor precession
Magnetic moment and angular momentum
The muon has a magnetic moment. What does magnetic moment mean?
We can think of a classical magnetic moment as a current loop. If a
charged particle goes around in a closed orbit then we have a current
loop and the magnetic moment µ is defined by the current I multiplied
by the area of the loop A. Note that a charged particle also has mass:
not only does charge move round and round (meaning that we have a µ L
magnetic moment) but also mass moves round and round (meaning that
we have angular momentum), as shown in Fig. 4.1. This means that the
orbital magnetic moment µ and the orbital angular momentum L are q = m
connected and should be proportional to each other, the proportionality
constant being called the gyromagnetic ratio γ. Thus µ = γL.
But what if a particle doesn’t go around in an orbit? Then L = 0 but
Fig. 4.1 A massive charged particle in
the particle still can have an intrinsic magnetic moment, connected to orbit gives rise to both a magnetic mo-
its intrinsic angular momentum (also called its spin). In this case the ment µ and angular momentum L.
magnetic moment is given by1 1
To make closer contact with under-
graduate treatments of quantum me-
µ = γS, (4.1) chanics, we will begin by defining S
with the correct units of angular mo-
where S is the spin angular momentum. For a particle of mass m and mentum (i.e. including the factor of ~);
charge q the gyromagnetic ratio γ is given by later in this chapter (on page 43), we
will return to the convention adopted
gq in more advanced work for which S is
γ= , (4.2)
2m dimensionless.
where g is a constant known as the g-factor and for which for the
electron and muon is very close2 to 2. Putting the numbers in provides 2
It is about 2.00233184 for the muon.
us with a very useful number: Dirac’s theory of the electron (and the
muon is a heavy electron) gives g =
γµ 2 precisely, but quantum electrody-
= 135.5 MHz T−1 . (4.3) namics gives higher-order corrections to
2π
this.
30 The quantum muon

Note that we have used a subscript on the gyromagnetic ratio γµ to

remind us that we are talking about the muon. Also, we have chosen
to express the result divided by a factor of 2π and in units of MHz T−1
for reasons that are about to become clear. Recall from Table 2.1 that
the mass of the muon is intermediate between that of the electron and
the proton, and thus so are its magnetic moment and gyromagnetic
ratio. Because we are going to be dealing with the positive muon µ+ ,
its magnetic moment and gyromagnetic ratio are positive; note that the
negatively charged electron e− therefore has a magnetic moment and
gyromagnetic ratio which are negative.

Classical treatment
A magnetic moment µ in an applied magnetic field B has an energy E
given by
E = −µ · B. (4.4)
Thus we might think that a magnetic field would cause a magnetic mo-
ment to line up with it, so as to minimize its energy. However, because
the magnetic moment is associated with angular momentum, there is a
torque G given by
G = µ × B. (4.5)
Since torque is equal to the rate of change of angular momentum, eqn 4.5
can be rewritten as
dµ
= γµ × B. (4.6)
dt
This means that the change in µ is perpendicular to both µ and to
B. Rather than turning µ towards B, the magnetic field causes the
direction of µ to precess around B. Equation 4.6 also implies that |µ|
3
See Exercise 4.1. is time-independent.3

B, z
Example 4.1
Consider the case in which B is along the z-direction and µ is initially at an angle
of θ to B and in the xz-plane (see Fig. 4.2). Then
µ
µ̇x = γBµy
µ̇y = −γBµx

θ µ̇z = 0, (4.7)
so that µz is constant with time and µx and µy both oscillate. Solving these differ-
x ential equations leads to
µx (t) = |µ| sin θ cos ωt
Fig. 4.2 A magnetic moment µ in a µy (t) = −|µ| sin θ sin ωt
magnetic field B which lies along the µz (t) = |µ| cos θ, (4.8)
z-axis. The magnetic moment is ini-
where
tially in the xz-plane at an angle θ to
ω = γB (4.9)
B. The magnetic moment precesses
around a cone of semi-angle θ. The ro- is called the Larmor precession (angular) frequency [the frequency, measured
tation is clockwise when viewed from in Hz, is of course (γ/2π)B] and the phenomenon is called Larmor precession.
above (from +z back to the origin).
Other documents randomly have
different content
The Project Gutenberg eBook of The Nakimu
Caves, Glacier Dominion Park, B. C.
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: The Nakimu Caves, Glacier Dominion Park, B. C.

Author: Canada. National Parks Branch

Release date: January 26, 2016 [eBook #51042]

Most recently updated: October 22, 2024

Language: English

Credits: Produced by Stephen Hutcheson, Dave Morgan, and the

online
Distributed Proofreaders Canada team at
http://www.pgdpcanada.net

*** START OF THE PROJECT GUTENBERG EBOOK THE NAKIMU

CAVES, GLACIER DOMINION PARK, B. C. ***
 The Nakimu Caves
Glacier Dominion Park, B. C.
On the road to the Caves—Ross Peak in centre.
 DOMINION PARKS BRANCH
Department of the Interior
OTTAWA
1914

THE NAKIMU CAVES

GLACIER DOMINION PARK, B.C.
Near the summit of the Selkirk range of mountains in British
Columbia is one of the most curious series of subterranean caves in
the world. Being of recent discovery and, up to the present, almost
inaccessible except to the most agile and daring they remain
practically unknown to the world at large, but to those bold spirits
who have dared their black depths they have offered an experience,
weird, thrilling and wonderful. Less than a thousand visitors at
present make the journey to the Caves each year.

To the graphic description of Arthur O. Wheeler, president and

founder of the Canadian Alpine Club, together with the report of W.
S. Ayres, the mining engineer who was commissioned by this
department to report on the Caves, the substance of this pamphlet is
indebted.

The Dominion Government is at present having work done which will

render accessible to the public this unusual and somewhat gruesome
curiosity. A good road is already built between the little alplike town
of Glacier to within 6,000 feet of the Caves and a trail connects this
point with the entrance. The distance is about seven miles by saddle
pony from the beautiful Canadian Pacific Railway hotel at Glacier to
the end of the road, and the whole journey can be finished on foot in
five hours. The ride is through scenes almost incomparable, perhaps
even in the Rocky or Selkirk mountains.

The Yoho valley, in which flows the Cougar Creek, in the bed of which
are the Caves, possesses magnificent vistas of varied mountain
scenery. To quote Wheeler:—“It is difficult to express the wonder of
the colour contrasts that meet the eye in the ever changing
panorama of snow-clad peak, rock precipice dazzling névé, shining
glacier ice and bronze-green forest of pines, midst which nestle magic
lakes of changing shades of blue and green.”

Mr. Wheeler remarks that the Cougar valley, in the bed of which are
the Caves, is of special interest for two reasons other than the Caves:
—

1. “It is of that special type known as ‘a hanging valley,’ or one that

has been carved out in a U-shaped cross-section by the action of
glacial erosion, and not in the form of a V, as would have occurred
through the action of water erosion alone. It is, moreover, a very
pronounced form of its type. The numerous small glaciers that still
line the sides and head of the valley give it exceeding great beauty
and interest in summer time.”

2. “I know of no other spot in the Selkirks where alpine flora is 4

more abundant and varied, and this fact alone is an attraction
not to be despised.”

Description of the Valley of the Caves,

by A. O. Wheeler.
“The valley of Cougar creek is divided into two parts of distinctly
different characteristics. The upper valley, a great spoon-shaped
basin extending from Lookout point to Cougar pass, is a most
pronounced form of the type known as ‘hanging valley,’ or one that
has been carved out in a U-shaped cross section by the eroding
power of a glacier at one time filling up its bottom. This glacier has
now shrunk to very small proportions at the extreme head of the
valley. The floor is on a comparatively low grade and, at one point, is
covered for about half a mile by a small lake-bed in which some
water lies during the summer. The entire length of the upper valley
may be put at 2½ miles. It is inclosed by the rocky steeps of Mount
Bagheera, Catamount peak and Mount Ursus Major, on the north, and
of Cougar mountain on the south. On the sides of these massives are
small glaciers, busy at work tearing down the entire structures. At the
head of the valley, Cougar pass leads across the shrunken glacier to a
steep ravine descending to the railway at some distance below Ross
Peak station.

On the road to the Caves—Showing Cougar Valley; Mts. Ursus Major and
Bagheera in distance.
 5

Douglas Falls—Near the Caves.

“It would be difficult to find a more beautiful example of the alpine
valley. In every direction silver water-falls leap down the sides from
the glaciers and melting snows of the surrounding peaks. These
collect at the bottom of the valley in one central stream which
bounds in foaming cascades to the little lake-bed mentioned above. It
leaves this, and, continually augmented by falls from above, rushes
through luxurious meadow-lands in a second series of cascades that
have worn down to bedrock, showing where a thin veneer of soil is
overlaying it. The alpine meadows and park-lands, as well as the
open mountain slopes of the valley, are throughout the spring and
summer decked with a gorgeous array of flowers of varied hues
which, in places, are so profuse and brilliant that it seems as though
nature had spread a carpet of rainbow colours for the delight and
wonder of her visitors. In early spring, the giant Adder’s Tongue
(Erythronium gigantium) covers whole acres with a brilliant yellow.
These flowers are the first, and may be seen pushing their heads up
through the snow. Like all spring flowers in this region, they follow
the melting snows, and may be found higher up in the valley as late
as August. Almost coincident with them is the Globe flower 6
(Trollius laxus), a plant of much beauty and great wealth of
blossom. Next come the scarlet and crimson Painter’s Brush
(Castilleia), showing everywhere in the open and on the lower slopes
with a blaze of glory. Later still, the blue Larkspur (Delphinium
bicolor) and purple and pink Asters replace the earlier series, the
crimson and yellow Monkey-flower (Mimulus) are found in the beds of
the streams and where moisture is prevalent; high up in the valley
and on the alp-lands below the rocks are seen the False Heaths
(Bryanthus and Cassiope) and, highest of all, the pink-flowering moss
(Silene acaulis), found in magnificently flowered bunches directly
below the ice. There are very many other species more rare, and just
as beautiful in blossom, but of not so frequent or noticeable
occurrence.

“The timber consists, in this upper valley, chiefly of spruce and

balsam, trees which, at this elevation, in the Selkirks attain a grace
and beauty that is not noticed in the more crowded areas of lower
altitudes. Here they rise symmetrically to a great height, and their
sweeping lower boughs form shaded canopies that are most inviting
during the sultry summer weather.

“As the head of the valley is approached, a short climb will convey
the explorer to the ice of several small glaciers where he may study
with ease their formation and action, look into miniature crevasses
and see how moraines of rock detritus are formed by the downward
flow of the ice.

“At the site of the Caves, Cougar Creek valley turns from a
northeasterly to a southeasterly direction, and falls sharply 2,000 feet
from Lookout point to the tank, as compared with 1,200 feet for the
upper valley in the same distance, about two miles. The lower valley
is V-shaped and has, for the most part, been carved out by the action
of water erosion. The sides except immediately adjacent to the
railway are timbered only in patches and are for the most part
covered by closely growing alders, bracken and rank grass which
have replaced the timber where the slopes have been swept clear by
avalanches from the heights above. Through the latter half of this
section of the valley the stream flows in a narrow canyon and the fall
is steep. As soon as the corner is turned and the slopes of the
Illecillewaet valley reached, a magnificent forest growth of Douglas fir
(Pseudotsuga Douglasii), hemlock (Tsuga Mertensiana), cedar (Thuga
gigantea), and a few scattered trees of white pine (Pinus strobus) is
passed through.

7
Goat Falls—Near the entrance to the Caves.
“From a natural history point of view, the upper valley is exceptional,
due largely to the absence of visitors in the past. The Rocky Mountain
goat (Haplocerus montanus) may be seen frequently, and his tracks
are everywhere along the heights. During one of our visits a grizzly
bear (Ursus ferox) was killed by Deutschman. The black bear (Ursus
americana) is plentiful at the head of bear creek across Baloo pass,
and it may be safely assumed that he does not fail to visit Cougar
valley. Of the smaller mammals, the hoary marmot or whistler
(Arctomys Columbianus) is plentiful in both the upper and lower
valley, and is here found in large numbers, larger in size and giving
forth a louder and more shrill whistle than in the main range of the
Rockies. Its startling note is very human and resembles much the
noise that delights the heart of the small boy, produced by 8
placing the fingers between the lips. Say’s squirrel
(Spermophilus lateralis) and Parry’s marmot (Spermophilus Parryi)
are also found, the latter in great numbers. The Little Chief hare
(Logomys princeps) is frequently seen disporting itself among the
rocks, and its comical antics and quaint squeak, resembling that of a
toy rabbit, are very amusing.

“The birds are few, and in the upper valley are chiefly confined to the
Ptarmigan (Lagopus leucurus), of which a flock may nearly always be
seen, the water ousel or dipper (Cinclus Mexicanus), a funny little
dark grey chap who flits from stone to stone along the cascades and
falls of the valley, continually bobbing and dipping as though it were
the object and aim of his existence. This bird has a very sweet note.
Of other birds the black-headed jay (Cyanocitta stelleri annectens)
and the Rocky Mountain Whisky Jack (Perisoreus Canadensis
capitalis) are the most apparent.

“Taken as a whole, this wonderful valley is worthy of a visit quite

independently of the attractions offered by the Caves, and it would
be difficult to find a more representative or better type of valley to
illustrate the various phases of nature in the Selkirks—scenery,
geology, natural history and botany.”
It was in the year 1904 that Charles Henry Deutschman of
Revelstoke, a British subject, while prospecting and hunting big
game, discovered the Caves in series.

The next person to enter the Caves after Deutschman was A.

Johnston, editor of the Revelstoke Mail-Herald. Next, in May, 1905, a
party of twelve persons, including W. S. Ayres, the mining engineer,
Howard Douglas, superintendent of Rocky Mountains Park, and R. B.
Bennett, the Associated Press correspondent at Vancouver and now
member of the Dominion House representing Calgary, visited the
Caves. Most of this party, however, returned to Glacier on the
following day and left Messrs. Ayres, Deutschman, Johnston, Benson
and Field to make further explorations.

On the road to the Caves—Mt. Avalanche, Eagle Peak, Mt. Sir Donald.

10

Mr. Ayres’ Description of this Exploration.

“On the morning of May 30, the entire party entered the cave by the
opening which I have designated on the map as ‘Entrance No. 1.’
Pools of water more or less filled with ice were encountered, which
greatly impeded our progress, and finally a very deep one, at a
distance of 237 feet from the surface, barred further progress until a
raft could be prepared. Retracing our steps to the surface we sought
an entrance in ‘The Canyon’ by means of a rope. This passage was
about 70 feet below the natural surface and about 100 feet above the
bottom of ‘The Canyon,’ was very small and blocked with ice. I have
designated this opening as ‘Entrance No. 2,’ on the map.

“Another opening was found which is recorded as ‘Entrance No. 3.’

Here, by crawling through a very narrow passage on hands and
knees and then descending a steep narrow water groove for about 50
feet the brink of a very large cavern was reached that was estimated
to be 256 feet deep, but its length and breadth were, owing to their
great extent and to the insufficiency of lights at hand, inaccessible. It
was observable, however, that several openings led off from this
great cavern. The plunge and roar of a great waterfall somewhere
down in the depths of this cavern reverberates in every inch of space,
and produces in the listener sensations so weird that even those that
have elsewhere met its counterpart are at first startled.

“No further attempt was made at this time to explore the cave
through this entrance for the reason that the three hundred feet of
rope sent by Mr. Ford, resident engineer of the C.P.R., for our use,
had not yet been brought up to the cave from the water tank. On the
following day another obstacle presented itself. The rapidly melting
snow formed a sudden rush of water which poured down the
mountain side and into this opening, and made it not only
inaccessible for the time being, but proved it to be positively
dangerous to enter for any extended explorations until the snow has
practically disappeared.

“On the afternoon of May 31, and the forenoon of June 1, a raft and
additional ladders were constructed to cross over the large pool that
impeded our progress on May 30 in ‘Entrance No. 1.’ While we were
eating our noon meal on the 31st, Cougar creek overflowed into this
entrance. We made the attempt to descend but were drenched with
water, and our lights were put out. We constructed a dam to prevent
this, as far as possible, and at 6 a.m. on June 2 we again descended
into the cave by this entrance, and found at a distance of 331 feet
from the surface a large square chamber, 50 feet wide by 60 feet,
that was accessible. A large portion of its length was inaccessible
owing to Cougar creek, which flows across it, and to accumulated ice.
We named this chamber ‘The Auditorium.’

“At a point in the main entrance just as we emerge from ‘The

Auditorium,’ a branch passage was found which again joins the main
entrance about 100 feet from the surface. This branch passage 11
is marked on the map as A-B. Another branch passage was
found to connect this entrance with the surface and it is designated
on the map as B-C. A diligent search in this part of the cave disclosed
no other accessible openings.

“A descent by ropes to the bottom of the canyon at ‘Entrance No. 2’

was considered but it was deemed entirely unsafe, owing to the very
large accumulated mass of snow which might slide into it at any
moment, and without warning. Besides this, no entry can be made
until Cougar creek has subsided to a mere brook.

“The falls, designated on the map as ‘Lower Goat Falls,’ were visited
with the hope that an entrance might be effected to the very large
cavern by the passage through which the water from the fall enters;
but it was practically filled with water and ice. The falls consist of two
vertical drops, the upper one is about 30 feet and the lower one 50
feet. A large amount of water is delivered into the cave from these
falls, the place of entry being immediately at its foot. It is called
‘Entrance No. 4.’

“No other openings were discovered by which access might be had to

the large cavern, and we are barred from entering it by those
openings already described, and for the reasons given, until such
time as the spring floods may have fully subsided.”
Mr. A. O. Wheeler, D.L.S., accompanied by Deutschman, was the next
explorer, and the following is Mr. Wheeler’s description of their
experiences in the new caves which were discovered during this visit:
—

Mr. Wheeler’s Exploration and Description of

the Caves.
“Up to this time, only the Gopher Bridge series and the Mill Bridge
series (see map) had been explored. It was not until a day or two
before the close of the survey, August 4, that an attempt was made
to penetrate the huge opening seen at the north end of the bottom
of the gorge. On the date named, the writer descended by means of
a knotted rope, and was followed by Deutschman. The stream was
then at high water, and pouring into the opening with a tremendous
rush and swirl of waters. Although the creek was crossed by wading
waist deep in the icy torrent, it was found impossible to penetrate
more than a couple of hundred feet into the depth. Later in the
month Deutschman, alone and unaided, penetrated this opening and
worked his way across the stream, discovering what is now referred
to as the Gorge series, or the series of passageways reached by the
opening from the Gorge. And here the writer begs to say a word or
two concerning this remarkable man. The work of exploration he has
done without assistance, shows a character utterly devoid of fear.
The descent into depths of blackest darkness, lighted only by the dim
rays of a tallow dip, without a rope or other aid except in a case of
direst necessity, requires more than courage; it requires strength of
purpose and power of will far beyond the ordinary degree. For, 12
added to the thick darkness, there was always the fierce
vibrating roar of subterranean torrents, a sound most nerve shaking
in a position sufficiently uncanny and demoralizing without it. Huge
cracks had to be crossed, and precipitous descents made in pitch
darkness, where, it is safe to say, a misstep would have instant
death, either suddenly or through disablement. Now that ladders are
placed and ropes set, and the sure path pointed out by this intrepid
guide, it is difficult to realize how in the first place the passage could
possibly have been made without.
Typical Alpland—On Mt. Abbott about 1 mile south of Glacier.
“The first visitor after Deutschman to the Gorge series was a 13
Mr. Lang, of Golden, B.C., who very nearly met with a serious
accident when climbing up the rope from the bottom of the gorge to
the floor of the valley, 80 feet above. He had nearly reached the top
when his strength gave out, and he slid back down the rope, arriving
at the bottom in a much bruised condition, but fortunately with no
bones broken.”

Mr. Wheeler furnishes a splendid detail description of the various

series of caves, and it is thought well to reproduce this description in
extenso:—

Gopher Bridge Series.

“On leaving the lake-bed above mentioned, Cougar creek flows in a
series of cascades for a little more than half a mile through open
alpine meadowland, broken here and there by knolls crowned with
scrubby spruce. Suddenly, without warning, it drops into a cavity and,
450 feet farther down, quietly issues from its underground way.
During this disappearance the stream has only dropped 30 feet. The
intervening space between the entry into and exit from the ground
has been named ‘Gopher Bridge’ by the first party visiting the Caves,
owing to the large numbers of Parry’s marmot, which much
resembles a gopher, to be seen in the immediate vicinity, and the
underlying passages are here referred to as the Gopher Bridge series.
Directly opposite the disappearance of the creek, two striking
cascades tumble down the mountain side and uniting, flow for a
short space parallel to Cougar creek; then, vanishing into a hole in
the ground, they join the main stream by a subterranean passage.
They are known as ‘Gopher Falls,’ and the place where they disappear
as the ‘Gopher Hole’ (see map of Caves).

“The Gopher Bridge series of passageways was at first entered by the

opening shown on the map as ‘Old Entrance.’ It was a very
disagreeable operation, entailing much wriggling and squeezing
through narrow cracks over dirty rocks. Eventually, a point of vantage
was reached directly over the subterranean torrent. At the time of the
first exploration the writer took acetylene bicycle lamps, whose bull’s-
eyes enabled the pitch darkness to be pierced to some extent.
Magnesium wire also was lighted, and, by its aid, for a brief minute,
the interior was bathed in dazzling brightness. Standing on a narrow
ledge that overhangs a black abyss, the eye is first drawn by a
subterranean waterfall heard roaring immediately on the left. It
appears to pour from a dark opening above it. Below, between black
walls of rock, may be seen the foam-flecked torrent hurtling down
the incline until lost in sense shadows. Overhead, fantastic spurs and
shapes reach out into the blackness and the entire surroundings are
so weird and uncanny that it is easy to imagine Dante seated upon
one of these spurs deriving impressions for his inferno. As the
brilliant light goes out the thick darkness makes itself felt, and
instinctively you feel to see if Charon is not standing beside 14
you. This subterranean stream, with its unearthly surroundings,
is somewhat suggestive of the Styx, and incidentally supplied the
name ‘Avernus’ for the cavern of the waterfall.

“At one time the stream entered by this opening, but a natural dam
has gradually been formed causing it to find a new opening at the
spot where it is now shown disappearing on the map. About midway
between the entrance of the creek and the old entrance, Deutschman
has excavated a small natural opening to a size sufficiently large
enough to admit an average person with comfort. A small passage
joins with the underground way of the creek, and by following along
its edge you eventually come to the point of vantage previously
described. En route, several small chambers are passed, originally
carved out of the rock in the form of potholes by the swirl and swish
of the waters, but since much distorted in shape owing to
disintegration of the cleavage planes. The spot is unique and
wonderful, and the sensations it produces eerie in the extreme. It is
well worthy of a visit.
On the road 2 miles from Glacier—Ross Peak in distance.

15

Mill Bridge Series.

“On making its exit at the eastern end of Gopher Bridge, Cougar
creek pours down a narrow rock-cut for a distance of 350 feet, when
it again disappears in a whirl of flying spray below the surface of the
valley. It reappears, 300 feet farther on at the bottom of a deep
gorge, having dropped 85 feet while underground. The interval
between the exit and entrance has been named Mill bridge by the
party previously mentioned, on account of the roar of the water as it
rushes underground through the choked entrance, resembling the
noise made by a big mill in full operation.

“The rock-cut above referred to is narrow, about 8 to 10 feet wide,

and of regular appearance. The upper half presents a series of
cascades and falls, and the sides show curious small potholes that
are in the process of erosion from the soft limestone. It has been
named ‘The Flume,’ owing to its resemblance to a millrace. There are
several openings at the point where the creek disappears, and its last
spectacular leap as it vanishes underground is very striking. 70 feet
farther east is a larger opening, at one time the point where the
stream disappeared, but as the rush of water cut deeper in the rock
channel it took advantage of a handy crack and gradually carved out
for itself the opening where the full volume now descends.

“About the centre of the flume, on the eastern side, and 30 feet from
it, is the entrance to the Mill Bridge series of the caves, shown on the
map as ‘Entrance No. 1.’ This entrance is a mere cleft in the rock
strata, and is only wide enough to admit of the passage of a man’s
body. The total length of the underground passageway, at one time
accommodating a very considerable volume of water, is 400 feet. The
height varies from a minimum of 10 feet to a maximum of 25 feet,
and the width from 3 to 15 feet. At its eastern end it opens to an
irregularly-shaped chamber of approximately 60 by 70 feet, with a
greatest height of 20 feet. This chamber has been named ‘The
Auditorium’ by the first exploring party. Cougar creek in its flow
beneath Mill bridge passes through the Auditorium, and as it falls 75
feet in a distance of 200 feet, from its entrance beneath Mill bridge to
the Auditorium, the chamber is replete with its roar, and the name is
Welcome to our website – the perfect destination for book lovers and
knowledge seekers. We believe that every book holds a new world,
offering opportunities for learning, discovery, and personal growth.
That’s why we are dedicated to bringing you a diverse collection of
books, ranging from classic literature and specialized publications to
self-development guides and children's books.

More than just a book-buying platform, we strive to be a bridge

connecting you with timeless cultural and intellectual values. With an
elegant, user-friendly interface and a smart search system, you can
quickly find the books that best suit your interests. Additionally,
our special promotions and home delivery services help you save time
and fully enjoy the joy of reading.

Join us on a journey of knowledge exploration, passion nurturing, and

personal growth every day!

ebookmasss.com