Oscillation
A coherent muon-spin precession signal appears whenever the muon ensemble experiences a well-defined local field — an applied transverse field, or a spontaneous internal field set up by magnetic order. The components here are undamped: damping is intentionally separated, so a physical line shape is built by multiplying with a relaxation envelope,
Oscillatory * Exponential + Constant
for a Lorentzian-broadened line (dynamic disorder, dilute static moments) or
Oscillatory * Gaussian + Constant
for a Gaussian-broadened line (dense static field distribution). A bare
Oscillatory + Constant fits only a perfectly coherent signal and will
absorb the inevitable line shape into spurious phase and frequency residuals.
When a signal contains several inequivalent muon sites, a sum of two or three
Oscillatory components is generally preferable to one component with a
broadened envelope; if the field distribution is genuinely continuous, look
at the Fourier spectrum first (Fourier analysis), and for an
incommensurate distribution use Bessel. Bessel is the bare
precessing line; for a powder of the same incommensurate, single-site
structure, where a non-precessing ⅓ tail is also present, use
OverhauserPowder instead. When the field distribution of a single-q
structure has two non-zero cut-offs rather than running to zero, choose by
how the local field moves: HelicalPowder when the field vector rotates on
an ellipse (helical or cycloidal order), and the
OverhauserPowderCutoff/OverhauserPowderCentre pair when the field
keeps one direction and only its size is modulated. HelicalCrystal is the
single-crystal form of HelicalPowder.
Oscillatory
Coherent precession parameterised by frequency. In zero field on an ordered magnet the spontaneous frequency \(f = \gamma_\mu B_{\mathrm{int}}/2\pi\) acts as an order parameter and is the natural quantity to trend versus temperature (Parameter trending); in transverse field the frequency calibrates the local field, \(f\,[\mathrm{MHz}] \simeq 0.01355\,B\,[\mathrm{G}]\).
Name |
Symbol |
Unit |
Description |
|---|---|---|---|
|
\(A\) |
% |
Component asymmetry amplitude. |
|
\(f\) |
MHz |
Precession frequency. |
|
\(\phi\) |
rad |
Phase offset. |
The frequency is bounded non-negative; the phase is unrestricted. In a
well-tuned spectrometer the phase of the first component should sit close to
0; large fitted phases usually indicate an instrumental phase offset that
should be calibrated out, or a wrong model. Frequencies approaching the
inverse binned time step alias: seed frequency from a Fourier peak
(Fourier analysis) rather than letting the minimiser search. A fit
window covering fewer than two or three periods leaves the phase uncertain at
the \(\sim\pi\) level — extend the range or fix the phase.
A standalone damped-cosine Oscillatory model (with Lambda and
baseline) is available in the MODELS registry.
References
Blundell, Contemp. Phys. 40, 175 (1999).
OscillatoryField
The same precession parameterised by the local field \(B\) (Gauss), with
\(f = \gamma_\mu B/2\pi\). Use when the physically interesting quantity
is the field itself — extracting the temperature dependence of a sublattice
magnetisation, or comparing internal fields across runs in a
parameter-trending workflow. For a transverse-field muonium experiment,
model the central diamagnetic Mu⁺ line with this component and add
MuoniumTF for the Mu⁰ satellites. Mathematically equivalent to
Oscillatory.
Name |
Symbol |
Unit |
Description |
|---|---|---|---|
|
\(A\) |
% |
Component asymmetry amplitude. |
|
\(B\) |
G |
Local magnetic field at the muon site. |
|
\(\phi\) |
rad |
Phase offset. |
Bessel
The polarisation of an incommensurate magnet, such as a spin-density-wave state. When the ordering wavevector is incommensurate with the lattice, the implanted muons uniformly sample the phase of the modulation and hence the Overhauser distribution of local fields, \(p(B) = \pi^{-1}(B_1^2 - B^2)^{-1/2}\) for \(|B| < B_1\); the resulting polarisation is the zeroth-order Bessel function with \(f = \gamma_\mu B_1/2\pi\) set by the field-distribution edge. At late times
a damped cosine with a characteristic \(-45^\circ\) phase — so a
free-phase Oscillatory fit that insists on a phase near
\(-45^\circ\) is the classic sign that this component is needed.
Compose with a relaxation envelope for additional damping; for commensurate
order use Oscillatory or OscillatoryField. For a powder sample,
where the ⅓ non-precessing tail is also visible, use OverhauserPowder
instead.
Name |
Symbol |
Unit |
Description |
|---|---|---|---|
|
\(A\) |
% |
Component asymmetry amplitude. |
|
\(f\) |
MHz |
Field-distribution edge, γμB₁/2π. |
|
\(\phi\) |
rad |
Phase offset. |
References
Le et al., Phys. Rev. B 48, 7284 (1993).
OverhauserPowder
The polarisation of a powder (polycrystalline) sample of the same
incommensurate, single-site structure as Bessel — a spin-density wave with
one muon site, sampled over every crystallite orientation. Each crystallite
still sees the Overhauser field distribution
\(p(B) = \pi^{-1}(B_{\max}^2 - B^2)^{-1/2}\) for \(|B| < B_{\max}\),
with \(f = \gamma_\mu B_{\max}/2\pi\) the distribution edge and the
order parameter to trend versus temperature, but the powder average splits
the polarisation into two exact fractions: a non-precessing \(\tfrac{1}{3}\)
of the muon spins lie along the local field at their site and relax at
\(\lambda_L\), while the precessing \(\tfrac{2}{3}\) follow the
\(J_0\) line shape of Bessel and relax at \(\lambda_T\).
Name |
Symbol |
Unit |
Description |
|---|---|---|---|
|
\(A\) |
% |
Component asymmetry amplitude. |
|
\(f\) |
MHz |
Field-distribution edge, γμBmax/2π. |
|
\(\lambda_T\) |
µs⁻¹ |
Relaxation of the precessing ⅔ fraction. |
|
\(\lambda_L\) |
µs⁻¹ |
Relaxation of the non-precessing ⅓ fraction. |
frequency, lambda_T and lambda_L are bounded non-negative. A
Bessel-like line does not on its own prove incommensurate order — a
distribution of fields from a commensurate structure with several sites, or
from disorder, can mimic it — so corroborate with the ordering wavevector
from diffraction where possible.
References
Savici et al., Phys. Rev. B 66, 014524 (2002).
Le et al., Phys. Rev. B 48, 7284 (1993).
Dalmas de Réotier, A. Yaouanc, and A. Maisuradze, arXiv:1410.2767 (2014).
OverhauserPowderCutoff and OverhauserPowderCentre
The polarisation of a powder sample in which the local field at the muon
keeps one direction while its size is modulated sinusoidally between two
non-zero cut-offs, \(B = B_{\mathrm{av}} + \Delta B\cos(\mathbf{q}\cdot\mathbf{r})\)
— a shifted Overhauser distribution,
\(p(B) = \pi^{-1}[(B - B_{\min})(B_{\max} - B)]^{-1/2}\), rather than one
running down to zero as in OverhauserPowder. As with
OverhauserPowder the polarisation splits into a non-precessing
\(\tfrac{1}{3}\) tail relaxing at \(\lambda_L\) and a precessing
\(\tfrac{2}{3}\) fraction relaxing at \(\lambda_T\); \(\phi\) is
a phase offset.
The two entries fit the same lineshape in two parametrisations, sharing one implementation:
OverhauserPowderCutoff —
frequencyis the upper cut-off \(f_{\max}\) andratiois \(r = B_{\min}/B_{\max}\). Shareratioacross a series, since it is a structural constant of the magnetic structure, and trendfrequencyas the order parameter.OverhauserPowderCentre —
frequencyis the centre \(f_{\mathrm{av}}\) anddelta_frequencyis the half-width \(\Delta f\), the form of Amato et al. (2014). Use it to fit the two edges as independent numbers, or to compare directly against literature written in \((f_{\mathrm{av}}, \Delta f)\).
The two parameter sets convert into each other directly:
Name |
Symbol |
Unit |
Description |
|---|---|---|---|
|
\(A\) |
% |
Component asymmetry amplitude. |
|
\(f_{\max}\) |
MHz |
Upper cut-off frequency (OverhauserPowderCutoff). |
|
\(r\) |
— |
\(B_{\min}/B_{\max}\) (OverhauserPowderCutoff). |
|
\(\phi\) |
rad |
Phase offset. |
|
\(\lambda_T\) |
µs⁻¹ |
Relaxation of the precessing ⅔ fraction. |
|
\(\lambda_L\) |
µs⁻¹ |
Relaxation of the non-precessing ⅓ fraction. |
Name |
Symbol |
Unit |
Description |
|---|---|---|---|
|
\(A\) |
% |
Component asymmetry amplitude. |
|
\(f_{\mathrm{av}}\) |
MHz |
Centre frequency (OverhauserPowderCentre). |
|
\(\Delta f\) |
MHz |
Half-width between the two cut-offs. |
|
\(\phi\) |
rad |
Phase offset. |
|
\(\lambda_T\) |
µs⁻¹ |
Relaxation of the precessing ⅔ fraction. |
|
\(\lambda_L\) |
µs⁻¹ |
Relaxation of the non-precessing ⅓ fraction. |
In the Fit Wizard, ratio and \(\phi\) are bounded to \([0, 1]\)
and \([-\pi, \pi]\) respectively; in a manual fit only the lower bound of
0 is set on ratio, delta_frequency, lambda_T, lambda_L and
frequency. delta_frequency exceeding \(f_{\mathrm{av}}\) is not
clamped: it is telling you \(B_{\min}\) has reached zero, and you should
switch to OverhauserPowder rather than read a negative \(f_{\min}\).
In the cut-off form \(r > 1\) is not the same line: it makes
frequency the lower cut-off and \(r f\) the upper one, so set an
upper bound of 1 on ratio in the fit table (the Fit Wizard does), or
read a fit that crosses it with the two edges swapped.
The closed form above is exact for that fixed-direction field. For a helical
or cycloidal structure, where the field vector rotates on an ellipse, the
distribution is instead
\(p(B) = (2/\pi)\,B\,[(B^2 - B_{\min}^2)(B_{\max}^2 - B^2)]^{-1/2}\),
which carries an extra factor of \(2B/\sqrt{(B+B_{\min})(B_{\max}+B)}\),
and the closed form is Amato et al.’s narrow-distribution approximation to
it: exact as \(r \to 1\) (a pure cosine at \(f_{\max}\)) and worse as
\(r\) falls, with a maximum deviation of the precessing part of about
0.03 at \(r = 0.8\), 0.07 at \(r = 0.6\), 0.12 at \(r = 0.4\) and
0.31 at \(r = 0\) (in units of the precessing amplitude). Use
HelicalPowder for that geometry. At \(r = 0\) the closed form does
not reduce to OverhauserPowder’s \(J_0(2\pi f_{\max} t)\): a
fixed-direction field swinging between zero and \(B_{\max}\) is not a
spin-density wave, whose field reverses sign. A fit that drives ratio (or
\(f_{\min}\)) to zero is worth comparing with OverhauserPowder and
with HelicalPowder, which reduces to it there. As with Bessel, a
Bessel-like line does not on its own prove incommensurate or single-q
order — several commensurate sites, or disorder, can mimic it — so
corroborate with the ordering wavevector from diffraction where possible.
The Fit Parameters panel’s Create Composite Parameter (Parameter trending) recovers the other parametrisation’s quantities as derived, uncertainty-propagated trends. From the centre form:
frequency + delta_frequencyfor \(f_{\max}\)frequency - delta_frequencyfor \(f_{\min}\)(frequency - delta_frequency)/(frequency + delta_frequency)for \(r\)
From the cut-off form:
frequency*(1 + ratio)/2for \(f_{\mathrm{av}}\)frequency*(1 - ratio)/2for \(\Delta f\)frequency*ratiofor \(f_{\min}\)
In a composite model, a name shared by several components carries the index
shown in the fit table (e.g. frequency_1).
The Fit Wizard offers OverhauserPowder and OverhauserPowderCutoff,
each with an optional extra Exponential, in its Precession family for
the single-run and Global Fit Wizards. OverhauserPowderCentre is not
offered as a wizard template, because it is numerically identical to the
cut-off form.
References
Amato et al., Phys. Rev. B 89, 184425 (2014).
Dalmas de Réotier et al., Phys. Rev. B 93, 144419 (2016).
Dalmas de Réotier, A. Yaouanc, and A. Maisuradze, arXiv:1410.2767 (2014).
HelicalPowder
The polarisation of a powder sample of a helical, cycloidal or other
single-q magnetic structure in which the local field at the muon site
rotates on an ellipse centred on zero, with semi-axes \(B_{\max}\) and
\(B_{\min}\). As the helix phase runs uniformly over the sites, the field
magnitude follows \(p(B)\), which diverges at both cut-offs. frequency
is the upper cut-off \(f = \gamma_\mu B_{\max}/2\pi\), the order parameter
to trend versus temperature, and ratio is \(r = B_{\min}/B_{\max}\), a
structural constant to share across a series. The powder average splits the
polarisation into a non-precessing \(\tfrac{1}{3}\) tail relaxing at
\(\lambda_L\) and a precessing \(\tfrac{2}{3}\) fraction relaxing at
\(\lambda_T\).
The transform of \(p(B)\) has no closed form, so it is evaluated
numerically to \(10^{-12}\) of the line amplitude. \(p(B)\) is the
shifted Overhauser density of OverhauserPowderCutoff times the smooth
factor \(h(B) = 2B/\sqrt{(B+B_{\min})(B+B_{\max})}\), and expanding
\(h\) in Chebyshev polynomials turns the integral into a series of Bessel
functions,
with \(f_{\mathrm{av}} = f(1+r)/2\), \(\Delta f = f(1-r)/2\), and
coefficients \(c_n\) that depend only on \(r\). The first term,
\(c_0 = 1\), is the OverhauserPowderCutoff line; the rest are the
correction. The series needs more terms as \(r\) falls, so a fit near
\(r = 0\) evaluates more slowly; at \(r = 0\) the line is exactly
OverhauserPowder’s \(J_0(2\pi f t)\), and at \(r = 1\) a single
cosine at \(f\).
Name |
Symbol |
Unit |
Description |
|---|---|---|---|
|
\(A\) |
% |
Component asymmetry amplitude. |
|
\(f\) |
MHz |
Upper cut-off, γμBmax/2π. |
|
\(r\) |
— |
\(B_{\min}/B_{\max}\). |
|
\(\phi\) |
rad |
Phase offset; starts fixed at 0. |
|
\(\lambda_T\) |
µs⁻¹ |
Relaxation of the precessing ⅔ fraction. |
|
\(\lambda_L\) |
µs⁻¹ |
Relaxation of the non-precessing ⅓ fraction. |
phase starts fixed at 0. In zero field every muon starts precessing from
the direction of its initial polarisation, and every field \(\mathbf{B}\)
on the ellipse is paired with \(-\mathbf{B}\), so the true phase is zero.
A fit that improves when the phase is freed is usually absorbing a time-zero
offset, which shifts the phase in proportion to frequency, or a missing
component, rather than describing the lineshape. ratio above 1 describes
the same ellipse with its axes relabelled, so frequency then names the
lower cut-off; bound ratio to \([0, 1]\) to keep the usual meaning.
Compose with an additional relaxation component when the line is broader than
the ideal distribution.
References
Amato et al., Phys. Rev. B 89, 184425 (2014).
A. Yaouanc and P. Dalmas de Réotier, Muon Spin Rotation, Relaxation, and Resonance (Oxford University Press, Oxford, 2011).
HelicalCrystal
The same elliptical field distribution as HelicalPowder in a single
crystal, where the non-precessing weight of each field is fixed by its
direction relative to the initial muon polarisation \(\hat{P}_0\)
rather than averaged over orientations. With \(a\) and \(c\) the
components of \(\hat{P}_0\) along the \(B_{\max}\) and
\(B_{\min}\) axes of the ellipse,
the squared component of the field direction along \(\hat{P}_0\) depends on the field magnitude only,
\(\theta_h\) is the angle between \(\hat{P}_0\) and the normal to the
plane the field rotates in, and \(\phi_h\) the angle, within that plane,
between the projection of \(\hat{P}_0\) and the \(B_{\max}\) axis.
Fields parallel to the polarisation do not precess, so the orientation
reshapes the line as well as setting \(W_0\): with \(\hat{P}_0\)
along the \(B_{\max}\) axis (\(\theta_h = 90^\circ\),
\(\phi_h = 0\)) the upper cut-off drops out of the oscillation, and along
the \(B_{\min}\) axis the lower one does. The magic orientation
\(\theta_h = 54.7356^\circ\), \(\phi_h = 45^\circ\) gives
\(a^2 = c^2 = \tfrac{1}{3}\) and reproduces HelicalPowder exactly. The
line is evaluated by the same series as HelicalPowder.
Name |
Symbol |
Unit |
Description |
|---|---|---|---|
|
\(A\) |
% |
Component asymmetry amplitude. |
|
\(f\) |
MHz |
Upper cut-off, γμBmax/2π. |
|
\(r\) |
— |
\(B_{\min}/B_{\max}\). |
|
\(\theta_h\) |
° |
Angle between the polarisation and the helix-plane normal. |
|
\(\phi_h\) |
° |
In-plane angle of the polarisation from the Bmax axis. |
|
\(\phi\) |
rad |
Phase offset; starts fixed at 0. |
|
\(\lambda_T\) |
µs⁻¹ |
Relaxation of the precessing fraction. |
|
\(\lambda_L\) |
µs⁻¹ |
Relaxation of the non-precessing fraction. |
Take theta_h and phi_h from the known magnetic structure and the
crystal mounting and hold them fixed where possible; freeing them alongside
A and ratio leaves strong correlations. Only \(a^2\) and
\(c^2\) enter, so both angles are defined on \([0^\circ, 90^\circ]\).
Symmetry-inequivalent muon sites, or magnetic domains, generally see
differently oriented ellipses: model them with one HelicalCrystal per
inequivalent set, sharing frequency and ratio where the structure
requires it. The weights assume the detector axis lies along
\(\hat{P}_0\); with ratio above 1 the roles of the two axes, and so of
phi_h, are exchanged.
References
Amato et al., Phys. Rev. B 89, 184425 (2014).
A. Yaouanc and P. Dalmas de Réotier, Muon Spin Rotation, Relaxation, and Resonance (Oxford University Press, Oxford, 2011).
VortexLattice / VortexLatticePowder
Transverse-field precession in the mixed state of a type-II superconductor.
Below \(T_c\) the muon samples the inhomogeneous field of the flux-line
lattice, whose distribution \(p(B)\) is strongly non-Gaussian — a sharp
low-field cutoff at the saddle point, a most-probable field below the mean, and
a long tail to high field near the vortex cores (a positively skewed line). The
relaxation \(R(t)=\langle e^{i 2\pi\gamma_\mu(B-\bar B)t}\rangle\) is the
characteristic function of the modified-London field distribution of an ideal
triangular lattice. Fitting the lineshape directly — rather than a single
Gaussian proxy, whose returned rate depends on the fit window and binning —
gives a window-independent penetration depth \(\lambda\) and upper critical
field \(B_{c2}\).
VortexLatticePowder applies the \(3^{1/4}\lambda_{ab}\) polycrystalline
average and returns the ab-plane depth \(\lambda_{ab}\). The line’s second
moment is calibrated to the Brandt result (see Superconducting penetration depth models),
so the depth read from this lineshape matches the field-domain
SC_Brandt_VortexLattice trend models. Multiply by a Gaussian for the
nuclear dipolar background and add Oscillatory + Constant for the
(weakly relaxing) sample-holder signal:
VortexLatticePowder * Gaussian + Oscillatory + Constant
Name |
Symbol |
Unit |
Description |
|---|---|---|---|
|
\(A\) |
% |
Component asymmetry amplitude. |
|
\(B\) |
G |
Applied transverse field (usually fixed). |
|
\(\phi\) |
rad |
Phase offset. |
|
\(\lambda\) |
nm |
Penetration depth (ab-plane for powder). |
|
\(B_{c2}\) |
T |
Upper critical field (core-size cutoff). |
field starts fixed at the applied value. \(B_{c2}\) is weakly
constrained by a single low-field run (where \(b=B/B_{c2}\to 0\)); fix it
from the literature or fit the field dependence to pin it. lambda_ab is
strongly correlated with the nuclear Gaussian rate, so constrain the latter
from a normal-state (above \(T_c\)) measurement at the same field.
References
Brandt, Phys. Rev. B 68, 054506 (2003).
Sonier, J. H. Brewer, and R. F. Kiefl, Rev. Mod. Phys. 72, 769 (2000).
Pratt et al., Phys. Rev. B 79, 052508 (2009).