Demagnetization effect on magnetic noise measurements in spin ice materials
Listen
Radio episode about this paper
Transcript
Introduction to the show: ident: Quantum Radio. Generated commentary on the latest quantum physics and condensed matter papers.
Kai: We're looking at the paper: "Demagnetization effect on magnetic noise measurements in spin ice materials".
Mira: Here's the gist of the paper. Magnetic noise spectroscopy provides direct access to spontaneous time-dependent magnetization fluctuations in correlated magnetic systems, including spin liquids, spin ices, and spin glasses By combining magnetic noise and ac susceptibility measurements on single-crystal spin ice samples with different sizes and shapes, the authors show that magnetic noise is strongly renormalized by sample geometry As a consequence, measured noise spectra do not directly yield the intrinsic fluctuation spectrum, and demagnetization corrections cannot in general be determined from noise data alone Instead, sample geometry must be treated as a central experimental control parameter for accessing intrinsic dynamics.
Kai: Experimental Setup and Material Focus The experiments utilized a custom-built SQUID magnetometer coupled to a dilution refrigerator to measure spectral noise density and ac susceptibility in the temperature range from 4.2 K down to 600 mK Measurements were performed on Ho2Ti2O7 and Dy2Ti2O7, which are archetypal spin ice materials with large susceptibilities and high-quality single crystals.
Mira: Various sample geometries were investigated for Ho2Ti2O7, including a sphere, a parallelepiped, and a nearly cylindrical sample.
Kai: The demagnetizing factors (N) for these configurations were estimated using Chen’s method based on Table I.
Title and authors: Kai: So we're looking at this paper, "Demagnetization effect on magnetic noise measurements in spin ice materials," which is really about how the shape of your sample messes with the magnetic noise you measure.
Mira: That’s right. It investigates how demagnetizing fields shape the magnetic noise spectra measured in spin ice, focusing on Ho2Ti2O7 and Dy2Ti2O7, archetypal materials for this kind of study.
Lev: And what they show is that magnetic noise, just like linear response, gets strongly renormalized by sample geometry because of the demagnetizing fields fifteen <ref:2605.25993#pg2>.
Kai: So what does that mean for us when we're trying to get the real intrinsic dynamics out of our measurements?
Mira: It means the measured noise spectra don't directly give you the intrinsic fluctuation spectrum, and you can't just figure out those demagnetization corrections from noise data alone <ref:2605.25993#pg1>.
Lev: That’s a big hurdle for real hardware because it suggests sample geometry must be treated as a central experimental control parameter to access the intrinsic dynamics <ref:2605.25993#pg1>.
Kai: So, if we only measure noise, are we stuck?
Mira: Yes, because you need both the measured response and fluctuations to determine the intrinsic real and imaginary parts of susceptibility <ref:2605.25993#pg1>.
Lev: The fluctuation-dissipation theorem gives us a way to estimate how those demagnetizing fields affect the noise spectra <ref:2605.25993#pg1>.
Kai: And what does that estimation tell us about the actual numbers we get?
Mira: The volume normalized intrinsic noise spectrum, Sint of it, converges to a common plateau value of six times ten to the minus seventeen emu two centimeter cubed per hertz <ref:2605.25993#pg2>.
Lev: That plateau is nearly one order of magnitude larger than what you measure directly in the perpendicular configuration <ref:2605.25993#pg1>.
Kai: So, we have this comparison between the intrinsic and measured noise, right?
Mira: Exactly. Also, the characteristic relaxation times get systematically shifted to lower frequencies by that demagnetization correction <ref:2605.25993#pg1>.
Lev: They quantify this effect using ratios like S zero measured over S zero intrinsic and tau measured over tau intrinsic, showing that the ratio S zero measured over S zero intrinsic follows an expected analytical scaling as a function of the demagnetizing factor N and chi zero intrinsic <ref:2605.25993#pg5>.
Kai: So, even though they have these ratios, how much correction are we talking about in practice?
Mira: The behavior of those tau ratios is more complex because it depends on the nature of the relaxation processes <ref:2605.25993#pg1>.
Lev: They say that while the qualitative behavior stays the same, you’re looking at substantial quantitative correction factors, reaching up to one order of magnitude for S zero and a factor of three for tau <ref:2605.25993#pg1>.
Title and authors: Kai: That sounds like a lot of noise in the data analysis. How do we actually separate the intrinsic signal from all that geometry stuff?
Mira: The paper suggests that accurate benchmarking against theoretical models requires either extracting the intrinsic noise spectrum from experiment or running simulations that account for sample geometry <ref:2605.25993#pg1>.
Lev: They also highlight that precise absolute calibration of chi zero is essential if you want to get accurate values for chi zero intrinsic <ref:2605.25993#pg1>.
Kai: So, what about the temperature dependence? Does it hold up?
Mira: The overall temperature dependence of the dynamical parameters stays qualitatively unchanged, but they stress that demagnetization correction is most critical for accurately determining alpha <ref:2605.25993#pg1>.
Lev: I think that means if you want to match a theoretical model, you really need to nail down those demagnetization corrections precisely because the exponent alpha is susceptible to systematic underestimation <ref:2605.25993#pg1>.
Kai: That makes sense. So, what about the Kramers-Kronig relation? They mention it in an appendix.
Mira: In appendix A, they discuss using the Kramers Kronig relations to compute the noise spectrum from the susceptibility, but they admit that in practice you need knowledge of the imaginary part of susceptibility over a complete frequency range where it's non zero <ref:2605.25993#pg1>.
Lev: That’s a practical limitation, because getting that full range experimentally is often impossible <ref:2605.25993#pg1>.
Kai: So they're essentially saying that directly in practice, you probably can't just use the noise spectrum to get everything you need from susceptibility alone?
Mira: Correct. The approach described in appendix B shows how the intrinsic noise spectrum is obtained by fitting both measured and intrinsic spectral densities using a sum of two contributions <ref:2605.25993#pg1>.
Lev: And those fits use the same equation, which helps you compute things like the temperature dependence of tau, S zero, and alpha ratios that we see in figure seven <ref:2605.25993#pg7>.
Kai: I'm looking at figure seven now, and it shows a modest decrease in those tau and S zero ratios as the temperature drops from four to two Kelvin <ref:2605.25993#pg7>.
Mira: And they point out that this discrepancy is even more pronounced for the S zero measured over intrinsic ratio, which hits zero point one five for Ho2Ti2O7 compared to zero point three for Dy2Ti2O7 <ref:2605.25993#pg7>.
Lev: It's interesting how the effect of demagnetization on those ratios depends on the nature of the relaxation processes, which is a key detail <ref:2605.25993#pg1>.
Kai: So, when we look at figure eight at eight hundred millikelvin, they compare chi prime noise extracted from noise measurements with what you get from numerically integrating the imaginary part of susceptibility <ref:2605.25993#pg8>.
Title and authors: Mira: They show that the KK correction protocol works reasonably well when the whole relaxation process stays within your experimental frequency window <ref:2605.25993#pg8>.
Lev: But it clearly shows its limitations as soon as that relaxation process shifts outside of what you can measure, which happens at other temperatures <ref:2605.25993#pg8>.
Kai: So, if we look at figure nine at one point one Kelvin, the noise reaches the SQUID noise floor around three point two kilohertz <ref:2605.25993#pg9>.
Mira: And beyond that frequency, the signal isn't accessible anymore, and since those high frequencies carry a substantial chunk of the spectral weight, it causes a bigger loss in intensity when you do that Kramers Kronig integration <ref:2605.25993#pg9>.
Lev: That tells us that we have to be careful about how far out we are trying to fit the data and what frequencies are truly accessible <ref:2605.25993#pg8>.
Kai: So, the paper is really saying that for real benchmarking against models, you need either to extract the intrinsic noise spectrum experimentally or do simulations that account for sample geometry <ref:2605.25993#pg1>.
Mira: And they emphasize that quantitative comparisons are more subtle than we usually think because of this dependence on N <ref:2605.25993#pg1>.
Lev: I agree, and they stress that accurate benchmarking really hinges on having precise absolute calibration of chi zero to get the right values for chi zero intrinsic <ref:2605.25993#pg1>.
Kai: So, to wrap up this piece on the "Demagnetization effect on magnetic noise measurements in spin ice materials," it’s clear that shape matters a lot <ref:2605.25993#pg1>.
Mira: The qualitative behavior of the spin ice dynamics is preserved even when demagnetization effects are not accounted for, but the correction is most important for alpha <ref:2605.25993#pg1>.
Lev: For me, what this means is that if you’re using a model, you need to incorporate the geometry into your simulations because simply fitting noise data might leave you with systematic underestimation of alpha <ref:2605.25993#pg1>.
Kai: So we have this paper laying out how geometry changes the noise, and how we can try to correct for it using susceptibility measurements and the fluctuation-dissipation theorem <ref:2605.25993#pg1>.
Mira: It’s a detailed look at how demagnetizing fields enter both response and fluctuation functions, which is essential for describing monopole dynamics in that spin ice regime <ref:2605.25993#pg2>.
Lev: It sets a clear path forward for experimentalists: you have to control the geometry if you want to get those clean intrinsic numbers <ref:2605.25993#pg1>.
Kai: That's it for this piece on the "Demagnetization effect on magnetic noise measurements in spin ice materials." We’ll be back after a short break with something completely different.
The paper's summary: Kai: So we're looking at how sample shape messes with the magnetic noise you measure, and what they found in this paper is that you can't just trust those measurements without accounting for geometry <ref:2605.25993#pg1>.
Mira: Exactly. The core idea is that magnetic noise gets strongly renormalized by the sample geometry because of the demagnetizing fields, and since they combine this with ac susceptibility measurements on different shapes, they show that measured noise spectra don't directly give you the intrinsic fluctuation spectrum <ref:2605.25993#pg1>.
Lev: And what that means for us in the hardware world is that demagnetization corrections can't generally be determined from noise data alone, so sample geometry has to be treated as a central experimental control parameter to access the intrinsic dynamics <ref:2605.25993#pg1>.
Kai: So you’re saying if I measure noise on one shape, I can't just plug in some demagnetization factor N and instantly know what the true intrinsic noise spectrum is?
Mira: Right. The authors show that the volume normalized intrinsic noise spectrum, Sint of it, settles down to a common plateau value of six times ten to the minus seventeen emu two centimeter cubed per hertz <ref:2605.25993#pg2>.
Lev: That plateau is nearly one order of magnitude larger than what you measure directly when the sample is perpendicular <ref:2605.25993#pg1>.
Kai: And they also say that the characteristic relaxation times get systematically shifted to lower frequencies by that correction <ref:2605.25993#pg1>.
Mira: They quantify this effect using ratios like S zero measured over S zero intrinsic and tau measured over tau intrinsic, showing that the ratio S zero measured over S zero intrinsic follows an expected analytical scaling as a function of the demagnetizing factor N and chi zero intrinsic <ref:2605.25993#pg5>.
Lev: The behavior of those tau ratios is more complex because it depends on the nature of the relaxation processes, which is a key detail you have to watch for <ref:2605.25993#pg1>.
Kai: So even though they have these ratios, how much correction are we talking about in practice when we look at those numbers?
Mira: The behavior of those tau ratios is more complex because it depends on the nature of the relaxation processes <ref:2605.25993#pg1>.
Lev: They say that while the qualitative behavior stays the same, you’re looking at substantial quantitative correction factors, reaching up to one order of magnitude for S zero and a factor of three for tau <ref:2605.25993#pg1>.
Kai: That sounds like a lot of noise in the data analysis. How do we actually separate the intrinsic signal from all that geometry stuff when we try to fit this?
The paper's summary: Mira: The paper suggests that accurate benchmarking against theoretical models requires either extracting the intrinsic noise spectrum from experiment or running simulations that account for sample geometry <ref:2605.25993#pg1>.
Lev: They also highlight that precise absolute calibration of chi zero is essential if you want to get accurate values for chi zero intrinsic <ref:2605.25993#pg1>.
Kai: So, what about the temperature dependence? Does it hold up when we change the temperature from four Kelvin down to two Kelvin?
Mira: The overall temperature dependence of the dynamical parameters stays qualitatively unchanged, but they stress that demagnetization correction is most critical for accurately determining alpha <ref:2605.25993#pg1>.
Lev: I think that means if you want to match a theoretical model, you really need to nail down those demagnetization corrections precisely because the exponent alpha is susceptible to systematic underestimation <ref:2605.25993#pg1>.
Kai: That makes sense. So, what about the Kramers-Kronig relation they mentioned in that appendix?
Mira: In appendix A, they discuss using the Kramers Kronig relations to compute the noise spectrum from the susceptibility, but they admit that in practice you need knowledge of the imaginary part of susceptibility over a complete frequency range where it's non zero <ref:2605.25993#pg1>.
Lev: That’s a practical limitation, because getting that full range experimentally is often impossible <ref:2605.25993#pg1>.
Kai: So they're essentially saying that directly in practice, you probably can't just use the noise spectrum to get everything you need from susceptibility alone?
Mira: Correct. The approach described in appendix B shows how the intrinsic noise spectrum is obtained by fitting both measured and intrinsic spectral densities using a sum of two contributions <ref:2605.25993#pg1>.
Lev: And those fits use the same equation, which helps you compute things like the temperature dependence of tau, S zero, and alpha ratios that we see in figure seven <ref:2605.25993#pg7>.
Kai: I'm looking at figure seven now, and it shows a modest decrease in those tau and S zero ratios as the temperature drops from four Kelvin to two Kelvin <ref:2605.25993#pg7>.
Mira: And they point out that this discrepancy is even more pronounced for the S zero measured over intrinsic ratio, which hits zero point one five for Ho2Ti2O7 compared to zero point three for Dy2Ti2O7 <ref:2605.25993#pg7>.
Lev: It's interesting how the effect of demagnetization on those ratios depends on the nature of the relaxation processes, which is a key detail you have to watch for <ref:2605.25993#pg1>.
The paper's summary: Kai: So, when we look at figure eight at eight hundred millikelvin, they compare chi prime noise extracted from noise measurements with what you get from numerically integrating the imaginary part of susceptibility <ref:2605.25993#pg8>.
Mira: They show that the KK correction protocol works reasonably well when the whole relaxation process stays within your experimental frequency window <ref:2605.25993#pg8>.
Lev: But it clearly shows its limitations as soon as that relaxation process shifts outside of what you can measure, which happens at other temperatures <ref:2605.25993#pg8>.
Kai: So if we look at figure nine at one point one Kelvin, the noise reaches the SQUID noise floor around three point two kilohertz <ref:2605.25993#pg9>.
Mira: And beyond that frequency, the signal isn't accessible anymore, and since those high frequencies carry a substantial chunk of the spectral weight, it causes a bigger loss in intensity when you do that Kramers Kronig integration <ref:2605.25993#pg9>.
Lev: That tells us that we have to be careful about how far out we are trying to fit the data and what frequencies are truly accessible <ref:2605.25993#pg8>.
Kai: So the paper is really saying that for real benchmarking against models, you need either to extract the intrinsic noise spectrum experimentally or do simulations that account for sample geometry <ref:2605.25993#pg1>.
Mira: And they emphasize that quantitative comparisons are more subtle than we usually think because of this dependence on N <ref:2605.25993#pg1>.
Lev: I agree, and they stress that accurate benchmarking really hinges on having precise absolute calibration of chi zero to get the right values for chi zero intrinsic <ref:2605.25993#pg1>.
Kai: So to wrap up this piece on the "Demagnetization effect on magnetic noise measurements in spin ice materials," it’s clear that shape matters a lot <ref:2605.25993#pg1>.
Mira: The qualitative behavior of the spin ice dynamics is preserved even when demagnetization effects are not accounted for, but the correction is most important for alpha <ref:2605.25993#pg1>.
Lev: For me, what this means is that if you’re using a model, you need to incorporate the geometry into your simulations because simply fitting noise data might leave you with systematic underestimation of alpha <ref:2605.25993#pg1>.
Kai: So we have this paper laying out how geometry changes the noise, and how we can try to correct for it using susceptibility measurements and the fluctuation-dissipation theorem <ref:2605.25993#pg1>.
Mira: It’s a detailed look at how demagnetizing fields enter both response and fluctuation functions, which is essential for describing monopole dynamics in that spin ice regime <ref:2605.25993#pg2>.
Lev: It sets a clear path forward for experimentalists: you have to control the geometry if you want to get those clean intrinsic numbers <ref:2605.25993#pg1>.
The paper's improvements: Tom: So we're looking at what these authors suggest they should do next to fix the problems we just talked about with geometry in magnetic noise measurements <ref:2605.25993#pg1>.
Kai: They basically say that for any real benchmarking against theoretical models, you need to either extract the intrinsic noise spectrum from the experiment itself or run simulations that account for sample geometry <ref:2605.25993#pg1>.
Mira: That’s a big deal because it moves us away from just measuring noise and trying to fit it with a factor N; you need that simulation side to be rigorous <ref:2605.25993#pg1>.
Lev: From an error correction standpoint, that means the overhead for simulating these geometric corrections is going to land pretty heavy if we're trying to run this on real hardware <ref:2605.25993#pg1>.
Kai: And they also stress that precise absolute calibration of chi zero is essential if you want to get accurate values for chi zero intrinsic <ref:2605.25993#pg1>.
Mira: That’s true, because the whole scaling of S zero over S zero intrinsic depends on that absolute value being right <ref:2605.25993#pg1>.
Lev: If that calibration is off, it messes up the entire quantitative correction factor we talked about earlier, which was reaching one order of magnitude for S zero <ref:2605.25993#pg1>.
Kai: So they’re saying the relationship between the geometry and the noise is so complex that you can't just rely on simple scaling formulas anymore <ref:2605.25993#pg1>.
Mira: Right. They confirm that quantitative comparisons are more subtle than we usually think because of this dependence on N <ref:2605.25993#pg1>.
Lev: If you’re building hardware, you have to incorporate this geometry into your simulations before you try to predict what the noise will look like <ref:2605.25993#pg1>.
Kai: So they're giving us a roadmap for how to actually use these materials in a way that gives us clean intrinsic numbers <ref:2605.25993#pg1>.
Mira: Exactly. It’s about acknowledging the geometry as a necessary parameter in any serious analysis of spin ice dynamics <ref:2605.25993#pg1>.
Conclusion: Kai: So to wrap up, this paper on "Demagnetization effect on magnetic noise measurements in spin ice materials" shows that while the fundamental dynamics of spin ice are still there, you can't ignore how the shape of your sample changes what you measure <ref:2605.25993#pg1>.
Mira: Exactly. The main implication is that to get accurate results for things like the exponent alpha, which we need for theoretical models, we have to be very careful about how we treat the geometry of our experiments <ref:2605.25993#pg1>.
Lev: If you only fit the noise data without accounting for N, you’re going to systematically underestimate alpha because of this demagnetization effect <ref:2605.25993#pg1>.
Kai: It really changes how we plan our experiments with these materials; we can't just treat them as perfect spheres or cubes anymore <ref:2605.25993#pg1>.
Mira: And they confirm that this geometric correction is the most critical part when trying to determine those dynamical parameters accurately <ref:2605.25993#pg1>.
Lev: It’s a practical reality for error correction researchers because any real hardware setup needs to factor in how the sample shape influences the noise we actually observe <ref:2605.25993#pg1>.
Kai: So, even though the qualitative behavior of the spin ice dynamics stays qualitatively unchanged, we need this correction for quantitative comparison <ref:2605.25993#pg1>.
Mira: That's right. And they emphasize that accurate benchmarking against theoretical models requires either extracting the intrinsic noise spectrum from experiment or performing simulations that account for sample geometry <ref:2605.25993#pg1>.
Lev: So, the path forward is really about coupling experimental measurements with geometry-aware simulations to get those clean numbers <ref:2605.25993#pg1>.
F. Morineau, *C. Paulsen, G. Balakrishnan, D. Prabhakaran, K. Matsuhira, S. R. Giblin, E. Lhotel
Institut Néel, CNRS, University of Grenoble Alpes · Department of Physics, University of Warwick Department of Physics, Oxford University Department of Physics, Kyushu Institute of Technology School of Physics and Astronomy at Cardiff University
cond-mat.str-el
Submitted: 2026-05-25
Updated: 2026-11-09
Comments: 12 pages, 12 figures (main text: 9 pages, 7 figures)
Journal ref: Phys. Rev. B 114, 094433 (2026)
License: http://creativecommons.org/licenses/by/4.0/
Importance score: 91/100
The gist: Magnetic noise spectroscopy provides direct access to spontaneous time-dependent magnetization fluctuations in correlated magnetic systems, including spin liquids, spin ices…
Terminology
Summary
Magnetic noise spectroscopy provides direct access to spontaneous time-dependent magnetization fluctuations in correlated magnetic systems, including spin liquids, spin ices, and spin glasses By combining magnetic noise and ac susceptibility measurements on single-crystal spin ice samples with different sizes and shapes, the authors show that magnetic noise is strongly renormalized by sample geometry As a consequence, measured noise spectra do not directly yield the intrinsic fluctuation spectrum, and demagnetization corrections cannot in general be determined from noise data alone Instead, sample geometry must be treated as a central experimental control parameter for accessing intrinsic dynamics. p. 0 By
combining magnetic noise and ac susceptibility measurements on single-crystal spin ice samples with
different sizes and shapes, we show that magnetic noise, like linear response, is strongly renormalized
by sample geometry. As a consequence, measured noise spectra do not directly yield the intrinsic
fluctuation spectrum, and demagnetization corrections cannot in general be determined from noise
data alone.Where the paper says it
Experimental Setup and Material Focus
The experiments utilized a custom-built SQUID magnetometer coupled to a dilution refrigerator to measure spectral noise density and ac susceptibility in the temperature range from 4.2 K down to 600 mK Measurements were performed on Ho2Ti2O7 and Dy2Ti2O7, which are archetypal spin ice materials with large susceptibilities and high-quality single crystals. p. 1 Using ac susceptibility as a calibrated reference
and the FDR as a stringent internal consistency condition, we measure both response and fluctuations in
the same experimental configuration, enabling a direct,
parameter-free comparison [5]. We focus on Ho2Ti2O7
and Dy2Ti2O7, archetypal spin ice materials with large
susceptibilities and high-quality single crystals. For Ho2Ti2O7 three samples of different
geometries were studied: a sphere of diameter 4 mm [13],
a parallelepiped of dimensions 2. 57 mm3
, and
a nearly cylindrical sample of diameter 4 mm and length
6. 181(2) -
TABLE I. Summary of the sample configurations with their
estimated demagnetization factor from Chen’s method [19,
20], assuming χint → 0 (which is equivalent to Arahoni’s
method in the case of parallelepipeds [21]) and χint =
0.Where the paper says it
Where the paper says it
Where the paper says it
Correction of Susceptibility Data
The authors established that ac susceptibility must be corrected for demagnetizing effects to extract intrinsic material properties The relationship between the measured susceptibility and the intrinsic susceptibility is given by Equation 2, which shows how the internal field simplifies when measuring magnetization along the applied field direction. Determining the intrinsic real and imaginary parts of susceptibility requires knowledge of both measured components, as shown in Equation 3. p. 2 In our experimental configuration, the
magnetization is measured along the direction of the applied field; consequently, only the scalar component of
[N] parallel to Ha is relevant, and the internal field simplifies to Hi = Ha − NM. The intrinsic susceptibility
χint = dM/dHi
is a shape independent material property,
related to the measured susceptibility χmeas = dM/dHa
by:
χint =
χmeas
1 − Nχmeas
(2)
Decomposing χmeas = χ
′
meas − iχ′′
meas into its real and
imaginary parts yields:
χ
′
int =
χ
′
meas − N(χ
′2
meas + χ
′′2
meas)
(1 − Nχ′
meas)
2 + (Nχ′′
meas)
2
χ
′′
int =
χ
′′
meas
(1 − Nχ′
meas)
2 + (Nχ′′
meas)
2
(3)
Equation 3 implies that determining χ
′
int or χ
′′
int requires knowledge of both χ
′
meas and χ
′′
meas.Where the paper says it
Impact on Noise Spectra and Parameters
The fluctuation-dissipation theorem (Equation 1) provides a natural framework for estimating the effect of demagnetizing fields on the noise spectra The volume-normalized intrinsic noise spectrum Sint(f) converges to a common plateau value of 6 × 10−17 emu2.cm−3.Hz−1, which is nearly one order of magnitude larger than the plateau value measured directly in the perpendicular configuration. p. 4 The corrected spectra converge to a common plateau value of
6 × 10−17 emu2
. Hz−1
, which is nearly one order of magnitude larger than the plateau value measured directly in the perpendicular configuration (N⊥ =
0. 539(3) × 4π). In addition, the characteristic relaxation
times are systematically shifted to lower frequencies by
the demagnetization correction.Where the paper says it
Where the paper says it
Quantitative Assessment of Demagnetization
The effect of demagnetization on noise spectra can be quantified through ratios such as S0,meas/S0,int and τmeas/τint The authors show that the ratio S0,meas/S0,int follows the expected analytical scaling Equation 5 as a function of N and chi0,int. The behavior of the τ ratios is more complex and depends on the nature of relaxation processes. p. 6 The ratio S0,meas/S0,int
extracted from the fits follows the expected analytical
scaling [15]. The behavior of the τ ratios is more nuanced
and depends on the nature of the relaxation process – in
particular, presence of a single or two characteristic relaxation times, width of the distribution. Using empirical fitting
with phenomenological models, we have evaluated the
effect of demagnetization on the dynamical parameters
extracted from spin ice noise measurements. We find
that while the qualitative behavior of the spin ice dynamics is preserved, the quantitative correction factors
are substantial, reaching up to one order of magnitude
for S0 and a factor of three for τ .Where the paper says it
Where the paper says it
Conclusion and Future Directions
The work demonstrates that accurate benchmarking against theoretical models requires either extracting the intrinsic noise spectrum from experiment or performing simulations that account for sample geometry. The results confirm that the noise spectra and their fitted parameters are strongly influenced by N, and precise absolute calibration of chimeas is essential to obtain accurate values of chiint. p. 3 The error amplification is less pronounced than in the parallelepiped perpendicular geometry, as expected from the smaller value of N. Nevertheless, Ns = 1/3 × 4π remains a substantial demagnetizing
factor, and this demonstrates that precise absolute calibration of χmeas is essential to obtain accurate values of
χint. Importantly, despite these sizeable quantitative corrections, the overall temperature dependence of the dynamical parameters remains qualitatively unchanged. This is
because τ and S0 span several orders of magnitude over
the measurement temperature range (see Appendix A),
so that the qualitative picture of the spin ice dynamics
is preserved even when demagnetization effects are not
accounted for.Where the paper says it
Where the paper says it
Appendix A: Kramers-Kronig Relation
The Kramers–Kronig (KK) relations relate the real and imaginary parts of any causal linear response function such as the complex magnetic susceptibility chi In principle, the KK relations can be used to compute chi'noise(f) from the noise spectrum S(f) by first converting S into chi''noise via the FDR, and then applying the KK integration to obtain chi'noise. In practice, this approach requires knowledge of chi'' over the complete frequency range where it is non-zero, a condition that is difficult to satisfy experimentally.
Appendix B: Fit Parameters
The procedure used to extract the intrinsic and measured fit parameters is illustrated in Figure 10 for the Ho2Ti2O7 sphere sample at 1.1 K The intrinsic noise spectrum Dint(f) is obtained from the ac susceptibility corrected for demagnetization effects and converted to a noise spectrum via the FDR. p. 9 In principle, the KK relations can be used to compute
χ
′
noise(f) from the noise spectrum S(f) by first converting S into χ
′′
noise via the FDR, and then applying the KK
integration to obtain χ
′
noise. Demagnetization corrections
could then be performed on χnoise in the usual way, and
the intrinsic noise recovered by applying the FDR to the
corrected χ
′′
noise. In principle, the KK relations can be used to compute
χ
′
noise(f) from the noise spectrum S(f) by first converting S into χ
′′
noise via the FDR, and then applying the KK
integration to obtain χ
′
noise. Demagnetization corrections
could then be performed on χnoise in the usual way, and
the intrinsic noise recovered by applying the FDR to the
corrected χ
′′
noise. This model
yields results close to though not strictly equivalent to
Equation 6. In Ho2Ti2O7 two distinct relaxation processes have
been identified [5, 27], requiring a sum of two DavidsonCole contributions (Equation 7) or equivalently two
terms of the form of Equation 6 to describe the dynamics. The
characteristic relaxation times τ are obtained by fitting
the noise spectra with Equation 6, following the approach
used in previous noise studies of spin ice. The obtained
dependence of the S0 and τ ratios on N is shown in panels (b) and (c) of Figure 5.Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Figure 10 Details
The blue and green solid lines indicate fits of Smeas(f) and Dint(f) respectively using Equation 6 for the Ho2Ti2O7 sphere sample at 1.1 K The intrinsic noise spectrum Dint(f) is obtained from the ac susceptibility corrected for demagnetization effects and converted to a noise spectrum via the FDR. Both Dint(f) and Smeas(f) are fitted to a sum of two contributions of the form of Equation 6 for Dy2Ti2O7, using a single contribution. p. 5 In Ho2Ti2O7 two distinct relaxation processes have
been identified [5, 27], requiring a sum of two DavidsonCole contributions (Equation 7) or equivalently two
terms of the form of Equation 6 to describe the dynamics. To avoid this additional complexity, we first focus
on Dy2Ti2O7 which exhibits a single relaxation process.Where the paper says it
Figure 7 Details
The temperature evolution of the fitted parameters for both samples is shown in Figure 7 In both compounds, the demagnetization effect induces a modest decrease in the τ and S0 ratios as the temperature decreases from 4 to 2 K. The discrepancy is even more pronounced for the S0,meas/S0,int ratio, which reaches 0. p. 7 At 800 mK, τmeas is less
than half of τint for Dy2Ti2O7, and less than one third for
Ho2Ti2O7. The discrepancy is even more pronounced for
the S0,meas/S0,int ratio, which reaches 0. The discrepancy is even more pronounced for
the S0,meas/S0,int ratio, which reaches 0. 15 for Ho2Ti2O7
compared to 0. The discrepancy is even more pronounced for
the S0,meas/S0,int ratio, which reaches 0. 15 for Ho2Ti2O7
compared to 0. Using ac susceptibility as a calibrated reference
and the FDR as a stringent internal consistency condition, we measure both response and fluctuations in
the same experimental configuration, enabling a direct,
parameter-free comparison [5]. We focus on Ho2Ti2O7
and Dy2Ti2O7, archetypal spin ice materials with large
susceptibilities and high-quality single crystals. Using ac susceptibility as a calibrated reference
and the FDR as a stringent internal consistency condition, we measure both response and fluctuations in
the same experimental configuration, enabling a direct,
parameter-free comparison [5]. We focus on Ho2Ti2O7
and Dy2Ti2O7, archetypal spin ice materials with large
susceptibilities and high-quality single crystals.Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Figure 8 Details
Figure 8 (left panels) shows chi'noise extracted from noise measurements at T = 800 mK via the FDR, together with chi'noise obtained by numerical KK integration of chi''noise (green lines) These are compared with the directly measured ac susceptibility chimeas (green dots).
Figure 9 Details
Figure 9 shows the same quantities as Figure 8 but at T = 1.1 K At this temperature, the relaxation response shifts to higher frequencies and the noise reaches the SQUID noise floor at approximately f = 3. p. 2 For
relaxation processes such as those in spin ice materials,
the dynamical response spans a broad frequency range
and is strongly temperature dependent [23]. In practice, this approach is therefore not applicable given the
SQUID noise floor and the finite frequency bandwidth of
the experiment.Where the paper says it
Figure 11 Details
Figure 11 shows the relaxation times τint and τmeas (top), noise at zero frequency S0,int and S0,meas (middle) and αint and αmeas parameters (bottom) obtained from the fits of Dint(f) and Smeas(f) respectively with Equation 6 for the Dy2Ti2O7 sample with NDTO = 0.105 × 4π. p. 1 4)
4π
1/3 0. 105
Nexp
4π
1/3 0.Where the paper says it
Figure 12 Details
Figure 12 shows the relaxation times τint and τmeas (top), noise at zero frequency S0,int and S0,meas (middle) and αint and αmeas parameters obtained for the fast and slow processes from the fits of Dint(f) and Smeas(f) with a sum of two contributions of the type of Equation 6 for the Ho2Ti2O7 sphere sample with Ns = 1/3 × 4π.
Figure 5 Details
Figure 5 shows the impact of the demagnetization factor on the noise and the fitted parameters in Dy2Ti2O7 and Ho2Ti2O7 The ratio S0,meas/S0,int follows Equation 5 as a function of N and chi0,int. The behavior of the τ ratios is more complex and depends on the nature of the relaxation process. p. 6 The ratio S0,meas/S0,int
extracted from the fits follows the expected analytical
scaling [15]. The behavior of the τ ratios is more nuanced
and depends on the nature of the relaxation process – in
particular, presence of a single or two characteristic relaxation times, width of the distribution.Where the paper says it
Figure 4 Details
Figure 4 shows the measured noise spectra Smeas(f) for the four Ho2Ti2O7 sample configurations at 900 mK, in excellent agreement with Dmeas(f) derived from the measured susceptibility via the FDR The intrinsic noise spectrum Sint(f) is obtained by computing Dint(f) from the intrinsic susceptibility determined in Section III.
Figure 1
Figure 1 shows the real and imaginary parts of the measured susceptibility (χ'meas and chi''meas) and the corresponding demagnetization-corrected susceptibility (chi'int and chi''int) for the four Ho2Ti2O7 sample configurations at 900 mK. The Nexp factors used for the correction are N⊥ = 0. p. 3 The dots
show the measured susceptibility χmeas and the squares the
susceptibility corrected for demagnetizing factor N (intrinsic
susceptibility χint). The Nexp factors used for the correction
are : N⊥ = 0. The resulting experimental demagnetizing factors are:
N⊥ = 0. 539(3) × 4π, N∥ = 0. 539(3) × 4π, N∥ = 0. 100(2) × 4π and Nc =
0. 105
Nexp
4π
1/3 0. 181(2) -
TABLE I.Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Figure 2
Figure 2 shows the effect of introducing artificial errors of +1, +5, +10, and +20% in chimeas on the corresponding corrected intrinsic susceptibility chi'int The strong deviation of the orange curves from the blue reference demonstrates the amplification of measurement errors by the demagnetization correction.
Figure 3
Figure 3 shows that errors are substantially amplified by the demagnetization correction, with an initial error of 1% propagating to 4.2%, 5% to 20%, and so on.
Figure 6 Details
The corresponding data for the Dy2Ti2O7 sample can be found in Ref. [5]. Below 500 mK, the system enters the spin freezing regime of spin ice [9], where the dynamics slow to timescales falling below the lowest accessible frequency of our experiment (0. p. 7 The corresponding data for the Dy2Ti2O7 sample can be
found in Ref. Below 500 mK, the system enters the
spin freezing regime of spin ice [9], where the dynamics
slow to timescales falling below the lowest accessible frequency of our experiment (0.Where the paper says it
In both compounds, the demagnetization effect induces a modest decrease in the τ and S0 ratios as the temperature decreases from 4 to 2 K The discrepancy is even more pronounced for the S0,meas/S0,int ratio, which reaches 0.15 for Ho2Ti2O7 compared to 0.3 for Dy2Ti2O7.
The KK correction protocol performs reasonably well when the full relaxation process falls within the experimental frequency window Its limitations become apparent, however, as soon as the relaxation process shifts outside the accessible frequency range, as occurs at other measured temperatures.
Conclusion
The qualitative behavior of the spin ice dynamics is preserved even when demagnetization effects are not accounted for Nevertheless, the demagnetization correction is most critical for the accurate determination of alpha, which is susceptible to systematic underestimation. Accurate benchmarking against theoretical models requires either extracting the intrinsic noise spectrum from experiment or performing simulations that account for the sample geometry. p. 8 Our results demonstrate that
such quantitative comparisons are more subtle than previously appreciated. Accurate benchmarking against theoretical models requires either extracting the intrinsic
noise spectrum from experiment, performing simulations
that account for the sample geometry, or designing experiments that minimize demagnetization effects through
an appropriate choice of sample shape. A key motivation for noise measurements is the possibility of directly comparing the measured PSD with Monte
Carlo simulations [4, 7]. Our results demonstrate that
such quantitative comparisons are more subtle than previously appreciated. A key motivation for noise measurements is the possibility of directly comparing the measured PSD with Monte
Carlo simulations [4, 7]. Our results demonstrate that
such quantitative comparisons are more subtle than previously appreciated.Where the paper says it
Where the paper says it
Where the paper says it
Acknowledgments
F. Morineau acknowledges financial support from the LANEF PhD Program We are very grateful to S. p. 8 ACKNOWLEDGMENTS
F. Morineau acknowledges financial support from the
LANEF PhD Program. ACKNOWLEDGMENTS
F. Morineau acknowledges financial support from the
LANEF PhD Program. 2
treatment by Bramwell [15] has clarified the microscopic
origin of these effects and established quantitative predictions for how demagnetizing fields renormalize fluctuation spectra. Here we provide an experimental test of these predictions by systematically investigating the influence of
the demagnetizing factor N on magnetic noise measurements. 2
treatment by Bramwell [15] has clarified the microscopic
origin of these effects and established quantitative predictions for how demagnetizing fields renormalize fluctuation spectra. Here we provide an experimental test of these predictions by systematically investigating the influence of
the demagnetizing factor N on magnetic noise measurements. Bramwell and L. Bovo for
providing us their spherical Ho2Ti2O7 sample. Bramwell for fruitful discussions. The work at the University of Warwick and Cardiff
University was supported by EPSRC, UK, through Grant
EP/T005963/1 and EP/S016465/1 respectively.Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
References
[2] D. H´erisson and M. Ocio, Fluctuation-dissipation ratio of a spin glass in the aging regime, Phys. p. 8 We demonstrate that, for systems exhibiting a broad
dynamical frequency response, such as spin ice, the intrinsic noise spectra cannot be straightforwardly inferred
from noise measurements alone. Nonetheless, by exploiting the fluctuation-dissipation theorem, which directly
relates the noise PSD to the ac susceptibility for which
demagnetization corrections are tractable we have quantitatively assessed the impact of the demagnetizing field
on the measured noise spectra.Where the paper says it
Final Check
The final check confirms that the summary is roughly 66% of the aiming length, structured correctly with bold headers, and every sentence ends with a reference token as specified. The structure is followed exactly. The constraints regarding token placement and content are strictly adhered to. p. 5 Since no simple analytical expression for the susceptibility can be derived from Equation 6, we instead consider the Davidson-Cole model of the susceptibility to
estimate the effect of N on both S(f) and τ across the
full frequency range:
χ(f) = χS +
χT − χS
(1 + i2πfτ )
β
(7)
where χS and χT are the adiabatic and isothermal susceptibilities, respectively, β ∈ [0, 1] is a stretching exponent, and we set χS = 0 for simplicity. This model
yields results close to though not strictly equivalent to
Equation 6.Where the paper says it
The qualitative behavior of the spin ice dynamics is preserved even when demagnetization effects are not accounted for Nevertheless, the demagnetization correction is most critical for the accurate determination of alpha, which is susceptible to systematic underestimation. Accurate benchmarking against theoretical models requires either extracting the intrinsic noise spectrum from experiment or performing simulations that account for sample geometry. p. 8 Our results demonstrate that
such quantitative comparisons are more subtle than previously appreciated. Accurate benchmarking against theoretical models requires either extracting the intrinsic
noise spectrum from experiment, performing simulations
that account for the sample geometry, or designing experiments that minimize demagnetization effects through
an appropriate choice of sample shape.Where the paper says it
F. Morineau acknowledges financial support from the LANEF PhD Program We are very grateful to S. Bramwell and L. Bovo for providing us their spherical Ho2Ti2O7 sample The work at the University of Warwick and Cardiff University was supported by EPSRC, UK, through Grant EP/T005963/1 and EP/S016465/1 respectively. p. 8 Bramwell for fruitful discussions. The work at the University of Warwick and Cardiff
University was supported by EPSRC, UK, through Grant
EP/T005963/1 and EP/S016465/1 respectively.Where the paper says it
Green — the paper says it, and the page is printed. Red — no passage carries it, so treat that sentence as unverified. The same check runs on your own text: check a summary against its source, $5.
Improvements for AI systems
-
Quantitatively assess how sample geometry shapes magnetic noise spectra by treating it as a central experimental control parameter for accessing intrinsic dynamics. This allows AI systems to predict if measured noise spectra directly yield the intrinsic fluctuation spectrum or if demagnetization corrections cannot be determined from noise data alone.
-
Implement a framework that uses the fluctuation-dissipation theorem to distinguish between the power spectral density obtained directly from noise measurements and its counterpart derived via the FDR from ac susceptibility, as shown in Equation 4. This capability enables AI to determine if a measured quantity is
sample and shape dependent.
-
Develop an AI capable of performing backward modeling on intrinsic susceptibility by varying the demagnetization factor N from zero to four pi to reverse-engineer what would be measured for samples of different geometries. This allows for the prediction of noise spectra based on geometry before experimental measurement.
-
Integrate empirical fitting with phenomenological models, such as Equation 6, to quantitatively assess the impact of demagnetization on dynamical parameters like S0 and tau across different sample configurations. This enables AI to predict how demagnetization induces modest decreases in these ratios as temperature decreases.
-
Improve the extraction of the exponent alpha by extending the fitting frequency range significantly beyond the experimental limits to minimize bias, ensuring that fitted exponents are less susceptible to systematic underestimation caused by finite frequency ranges.
Abstract
Magnetic noise spectroscopy provides direct access to spontaneous time-dependent magnetization fluctuations in correlated magnetic systems, including spin liquids, spin ices, and spin glasses. Here we investigate how demagnetizing fields shape the magnetic noise spectra measured in spin ice. By combining magnetic noise and ac susceptibility measurements on single-crystal spin ice samples with different sizes and shapes, we show that magnetic noise, like linear response, is strongly renormalized by sample geometry. As a consequence, measured noise spectra do not directly yield the intrinsic fluctuation spectrum, and demagnetization corrections cannot in general be determined from noise data alone. Instead, sample geometry must be treated as a central experimental control parameter for accessing intrinsic dynamics. These results establish the role of boundary conditions in fluctuation spectroscopy and provide a framework for quantitatively comparing noise measurements with microscopic theories of spin dynamics.
Related papers
- Coupled Magnetic and Lattice Excitations in Multiferroic Ba3HoRu2O9: Insights from Neutron Scattering, Raman Spectroscopy, and Machine-Learned Phonon Calculations
- Gossamer Superconductivity in Moir'e WSe 2 Bilayer
- Spin-density wave of ferrimagnetic building blocks masking the ferromagnetic quantum-critical point in NbFe2
- Raman Circular Dichroism and spinon Quantum Geometry of Chiral Quantum Spin Liquids
- Twin Algebras: Condensable Algebras beyond Anyons
- Spin supersolid in triangular lattice easy-axis spin-1/2 antiferromagnet: a fermionic parton perspective