REVIEW 3 major objections 5 minor 33 references
Anomalous Shear Stress Growth During Relaxation of a Soft Glass
T0 review · 3 major / 5 minor · reviewed 2026-08-07 · deepseek-v4-flash
Pith's one-line read Soft glassy fluids remember past shear direction and show a growing stress while resting.
desk verdict A solid new experimental protocol and observation of directional memory in soft glasses, with a model that explains the effect but whose headline claim about post-yield persistence outruns the data. read the letter →
The pith
A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.
The reading
What carries the argument
The load-bearing object is the elastoplastic model with power-law yielding (EP-PLY), built on a time-varying distribution $P(\ell,t)$ of local strain elements $\ell$. Each element is an elastic mesoscopic region that yields at a rate $r(\ell)$ that is zero inside a yield window $-\ell_y<\ell<\ell_y$ and grows as a power law of the distance outside it; upon yielding, elements are reset to a Gaussian post-hop distribution. The macroscopic stress is the first moment $\sigma(t)=\int \ell P(\ell,t)\,d\ell$. The mechanism that carries the argument is the residual anisotropy imprinted during training: a long tail on one side and a small hump on the other, which yield at different rates and therefore first increase and later decrease the first moment. A background fluidity $\varepsilon$ controls the terminal relaxation.
What would settle it
Directly image the local strain distribution $P(\ell)$ during the reading step—for example by tracking individual bubbles in a quasi-2D foam or by X-ray photon correlation spectroscopy—and check whether it is asymmetric, with a long tail on one side and a small hump on the other, and whether the tail relaxes faster than the hump. Alternatively, train a sample at zero stress for long enough to erase the tail (or test a freshly made material that has never been pre-sheared) and verify that the stress rise disappears, as the model predicts.
Extended reading notes
Core claim
The paper claims that several soft glassy fluids store a directional mechanical memory of prior shear even after bulk stress has equilibrated to zero. When a small step strain is applied and the sample is held at fixed strain, the shear stress can first grow for minutes to hours and only then relax away, rather than monotonically decaying. This nonmonotonic response is reproduced by an elastoplastic model with power-law yielding (EP-PLY) that tracks a distribution of local strain elements; the model shows the stress rise arises from an asymmetric strain distribution whose long tail yields quickly while a small hump on the opposite side yields slowly. Reversing the direction of the pre-shear flips the sign of the residual response, and the memory survives probe strains well beyond the yield strain, superimposing on elastic responses. The author argues that high-shear pre-conditioning does not fully erase prior deformation history, so past processing may influence rheological measurements more than previously assumed.
Load-bearing premise
The mechanism rests on a residual asymmetric distribution of local strain states that is never measured directly; the model parameters that produce it are fitted to the same relaxation data the model explains, so if real microstructures relax symmetrically the central claim collapses.
Editorial extensions
If this is right
- Standard pre-shear protocols do not fully reset soft glasses, so yield-stress and creep measurements may carry hidden contributions from earlier deformation.
- Nonmonotonic stress relaxation after a strain step can arise without thixotropy or shear banding; a purely elastoplastic mechanism suffices.
- The directional memory persists beyond the yield strain, so large deformation steps do not erase the conditioning; the residual stress can be about one tenth of the yield stress.
- A nonzero background fluidity implies a finite zero-shear viscosity and thus no true yield stress for the tested materials, with estimates given in the paper.
- Low-amplitude oscillations after conditioning may erode long tails of $P(\ell)$ and offer a way to mechanically anneal soft materials toward a reproducible initial state.
Reading between the lines
- The asymmetry mechanism, if confirmed, would generalize to other athermal amorphous solids, so granular piles under cyclic loading might show the same nonmonotonic relaxation after unloading.
- A testable prediction the author does not make: the height of the stress rise should scale with the length of the long tail in $P(\ell)$, so deliberately broadening the distribution (e.g., by a brief reverse shear) should amplify the effect.
- If the memory is a superposition of residual and elastic responses, then protocols could be designed to write prescribed small stress responses into a material, which may matter for additive manufacturing or soft actuators whose rheological history is never fully reset.
- The model parameter $\ell_0/\ell_y$ falls with foam age, and older foams show smaller stress rises; this suggests an aging protocol could tune the memory strength, though the author does not explore this.
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. The manuscript reports a new experimental protocol (conditioning, training, reading) applied to two Carbopol microgels and a foam at two ages. It shows that after a small step strain, the shear stress during relaxation can increase before terminal decay, and that the response depends on the direction of prior conditioning. The authors propose an elastoplastic model (EP-PLY) with a distribution of local strain states and power-law yielding, which reproduces the observed stress rise and the directional dependence. The abstract further claims that these directional memory effects persist even when the material is stressed beyond the yield stress, and that the mechanism is residual anisotropy in strain states relaxing at different rates.
Significance. If the experimental observation is robust, it is significant: it suggests soft glassy fluids retain directional mechanical memory beyond the yield stress, challenging the common assumption that pre-shear fully resets the material. The study includes direction-reversal and no-conditioning controls, and uses four sample variants. The EP-PLY model is transparent and analytically tractable in limiting cases (EM-1 A, B), yielding closed-form flow curves, which is a useful contribution. However, the model parameters are fitted to the same relaxation data, and the beyond-yield persistence claim is not directly demonstrated experimentally; the paper's value currently rests on the experimental phenomenology and the model as an illustrative mechanism rather than as an independently confirmed prediction.
major comments (3)
- [Abstract and Directional Memory (Fig 4)] The abstract states, "We demonstrate that these effects persist even after material is stressed beyond the yield stress," but this claim is not supported by the experimental results shown in Fig 4. For the largest experimental step (γstep = 24%), the text says the directional difference is "nearly invisible on a log scale." The quantitative support for persistence over 0.1–100% strain comes from Fig 4c, whose caption reads "Results from a range of experiments" while the main text says "A larger set of strain steps were simulated." This mismatch matters: the simulated persistence is obtained with parameters (ℓ0/ℓy, ε) fitted to the same relaxation data, so the post-yield persistence is a model extrapolation, not an experimental finding. The claim should be softened to a model prediction unless direct experimental evidence with a visible, replicable directional difference at post-yield strains is provided.
- [Model Development and Distribution of Local Strain (Eqs. 1–2, Table I)] The model-experiment agreement is not an independent test: the values of ℓ0/ℓy and ε were selected "to correspond to observed experimental stress relaxation results" (Model Development), and the proposed mechanism is then read off from the simulated P(ℓ, t) generated with those fitted parameters. No out-of-sample prediction is made, and P(ℓ) is never measured. The theoretical explanation is therefore a consistent interpolation rather than a confirmed mechanism. The manuscript should state this limitation explicitly and, ideally, test EP-PLY on a protocol not used for fitting (e.g., different training times or a different material) to establish predictive power.
- [SI-1 D and Figs. 1c, 4a] No replicate measurements or error bars are reported, and the low-stress, short-time data are post-processed with an averaging filter whose parameters are only described qualitatively. The main nonmonotonic rise in Fig 1c occurs at short times and low stresses, and the directional differences in Fig 4a are small and, at large steps, nearly invisible. Without repeated measurements, standard errors, and a precise statement of the filter kernel and window, it is difficult to rule out that the reported rise and small directional differences are smoothing or noise artifacts. Reporting these details is necessary to support the central experimental claim.
minor comments (5)
- [Fig 4c caption] The caption reads "Results from a range of experiments" but the main text says these results are simulated; the caption should be corrected to distinguish simulation data from experimental data, and the symbols should be defined.
- [Directional Memory paragraph] The description of Fig 4c ("absolute stress after about 10^3 s in (b) for both step directions") is ambiguous; please specify which quantity is plotted on each axis and what the symbols represent.
- [Distribution of Local Strain] The claim "In simulation, a fully relaxed state prior to reading presents a response that decays monotonically to zero stress (Fig SI-2)" does not clearly match Fig SI-2, which shows a (0,+) simulation for Carbopol rather than a dedicated fully relaxed initial state; please clarify the simulation protocol for this statement.
- [Throughout] There are several typographical errors: "createas" should be "creates" (Directional Memory), "viscoelatic" should be "viscoelastic" (final discussion), and "fluiditiy" should be "fluidity" (caption of Fig SI-2).
- [EM-1 C] The no-conditioning control is based on a single experiment (Fig EM-2); the text should state explicitly that this is a single measurement so readers can gauge its weight.
Circularity Check
Partial circularity: the post-yield persistence headline and the P(ℓ) anisotropy mechanism 'prediction' reduce to a model fitted on the same relaxation data, and the Fig 4c caption labels simulated results as 'experiments.'
-
fitted input called prediction
[The mechanism claim appears in the 'Model Development' and 'Distribution of Local Strain' sections (Fig 3), and is reflected in the Abstract's 'Using insight from the model, we suggest a…]
"To complete our model-experiment link, values for the post-hop distribution width ℓ0/ℓy and background fluidity ε were selected to correspond to observed experimental stress relaxation results. ... EP-PLY makes a prediction of the internal mechanisms that may lead to an increase in stress over long times when strain is held fixed at zero."
The model parameters (ℓ0/ℓy and ε) are explicitly fitted so that the simulated σ(t) matches the very nonmonotonic relaxation the paper then explains. Because stress is defined as the first moment of P(ℓ), σ(t) = ∫ ℓ P(ℓ,t) dℓ, the 'prediction' that the stress rise is caused by a fast-yielding negative tail and a slow-yielding positive hump is a re-description of the fitted model's internal state, not an out-of-sample test: no direct measurement of P(ℓ) is reported anywhere. The paper's hedge ('we suggest') and the reversed-conditioning and no-conditioning controls give the anisotropy mechanism partial independent behavioral support, but framing the fitted model's internal state as a 'prediction' overstates the evidence.
-
fitted input called prediction
[The post-yield persistence claim appears in the Abstract ('We demonstrate that these effects persist even after material is stressed beyond the yield stress'), in the 'Directional Memory' section…]
"the difference in (+ , +) and (−, +) steps ... remaining perceptible over the full range of tested strains, and retaining a value slightly less than one tenth of the yield stress σy, even when the initial stress response exceeded the yield stress (Fig 4a,c). For large steps the apparent difference was small compared to the overall stress response, and nearly invisible on a log scale. Fig 4c caption: 'Results from a range of experiments with γstep = 0.1-100% show the initial stress ... and the difference between stress maximum in (+ , +) and minimum in (−, +).'"
The abstract's headline claim that the memory effects persist beyond the yield stress is supported only by Fig 4c, whose caption labels the data 'Results from a range of experiments' while the main text states that 'a larger set of strain steps were simulated, spanning 0.1%-100%.' That simulation uses parameters (ℓ0/ℓy = 0.35, ε = 0.0005) fitted to the same foam's relaxation data, and the paper itself admits that at the largest experimental step (24%, Fig 4a) the directional difference is 'nearly invisible on a log scale.' The post-yield persistence claim therefore reduces to an extrapolation of the fitted model presented as a demonstrated experimental result.
full rationale
The central experimental observation — directional, nonmonotonic stress relaxation after small strain steps — is independently grounded, and the paper includes genuine controls: an in-situ (no-conditioning) foam shows an order-of-magnitude smaller stress rise, and reversing the conditioning direction reverses the sign of the response. These give the residual-anisotropy mechanism non-circular behavioral support, and the model also matches steady flow curves from an independently chosen exponent ν. However, two load-bearing claims reduce to the fitted model. First, the 'prediction of the internal mechanisms' is read from EP-PLY after ℓ0/ℓy and ε were 'selected to correspond to observed experimental stress relaxation results'; since σ(t) ≡ ∫ ℓ P(ℓ,t) dℓ by definition, the P(ℓ) asymmetry explanation re-describes the fitted model's own output without any measured microstructure. Second, the abstract's 'We demonstrate that these effects persist even after material is stressed beyond the yield stress' is carried by Fig 4c, whose large-step portion is simulated (with the caption mislabeling it as 'experiments') and whose largest experimental step (24%) is admitted to be 'nearly invisible on a log scale.' No load-bearing self-citation chain is present: the author's own references [13,14] support only vane-rheometry methods, and the η0 estimates (Section EM-1 F) are cross-checked by two independent routes. The sub-yield directional memory phenomenon stands on its own, but the headline post-yield persistence claim and the mechanism 'prediction' are partially circular, giving a score of 6.
Assumptions & free parameters
free parameters (4)
- ℓ0/ℓy (normalized post-hop distribution width) =
0.80 (young foam), 0.35 (aged foam), 0.65 (standard Carbopol), 0.50 (mixed Carbopol)
- ε (background fluidity) =
0.007 (young foam), 0.0005 (aged foam), 0.01 (standard Carbopol), 0.0034 (Carbopol in SI-2)
- ν (power-law yielding exponent) =
2 for most simulations; n = 1/(1+ν) maps to the flow-curve exponent
- τ0 (simulation time scale) =
22 s (foam), 44 s (Carbopol)
assumptions (6)
- domain assumption The stress of the material is the first moment of the strain-element distribution, σ(t) = ∫ℓ P(ℓ,t)dℓ
- ad hoc to paper Local yielding rate r(ℓ) has a flat well with power-law growth outside the well (Eq. 2)
- ad hoc to paper After yielding, elements reset into a Gaussian post-hop distribution P0(ℓ) of width ℓ0
- domain assumption The materials have negligible thixotropy and negligible aging over the protocol
- ad hoc to paper Residual asymmetric strain distribution survives the zero-stress training step because full relaxation is slower than the experimental window
- domain assumption Training and reading steps can be mapped to model time via ttrain and τ0
Cite this review
Pith. "Pith review of Anomalous Shear Stress Growth During Relaxation of a Soft Glass." pith.science (2026). https://pith.science/paper/OSZQST42
@misc{pith2026250606393,
author = {Pith},
title = {Pith review of: Anomalous Shear Stress Growth During Relaxation of a Soft Glass},
year = {2026},
howpublished = {\url{https://pith.science/paper/OSZQST42}},
note = {Machine review of arXiv:2506.06393}
}
read the original abstract
We show experimentally that multiple soft glassy fluids are capable of storing directional rheological signatures from past shear history, evidenced during stress growth and overall nonmonotonic stress relaxation after small steps in strain. We illustrate theoretically that these responses can be reproduced without requiring thixotropy or shear-banding, which are typically implicated in time-dependent rheological complexities, but by using a simple elastoplastic rheological model with power-law yielding (EP-PLY) that incorporates a distribution of local strain states. Using insight from the model, we suggest a mechanism for the experimentally observed stress increase to be driven by residual anisotropy in strain states that are relaxed at different rates. We demonstrate that these effects persist even after material is stressed beyond the yield stress, indicating that past deformation may have more influence than previously thought.
Figures
Reference graph
Works this paper leans on
-
[1]
J. Song, Q. Zhang, F. de Quesada, M. H. Rizvi, J. B. Tracy, J. Ilavsky, S. Narayanan, E. Del Gado, R. L. Leheny, N. Holten-Andersen, et al., Microscopic dynam- ics underlying the stress relaxation of arrested soft ma- terials, Proceedings of the National Academy of Sciences 119, e2201566119 (2022)
work page 2022
-
[2]
K. A. Murphy, J. W. Kruppe, and H. M. Jaeger, Mem- ory in Nonmonotonic Stress Relaxation of a Granular System, Physical Review Letters 124, 168002 (2020), arXiv:1911.01560
work page Pith review arXiv 2020
-
[3]
B. F. Di Dio, F. Khabaz, R. T. Bonnecaze, and M. Cloitre, Transient dynamics of soft particle glasses in startup shear flow. part ii: Memory and aging, Journal of Rheology 66, 717 (2022)
work page 2022
-
[4]
H. J. Barlow, J. O. Cochran, and S. M. Fielding, Ductile and brittle yielding in thermal and athermal amorphous materials, Phys. Rev. Lett. 125, 168003 (2020)
work page 2020
-
[5]
Koumakis and G
N. Koumakis and G. Petekidis, Two step yielding in attractive colloids: transition from gels to attractive glasses, Soft Matter 7, 2456 (2011)
2011
-
[6]
P. Coussot, H. Tabuteau, X. Chateau, L. Tocquer, and G. Ovarlez, Aging and solid or liquid behavior in pastes, Journal of Rheology 50, 975 (2006)
work page 2006
-
[7]
M. Cloitre, R. Borrega, and L. Leibler, Rheological Aging and Rejuvenation in Microgel Pastes, Phys. Rev. Lett. , 4819 (2000)
work page 2000
- [8]
Show all 33 references
-
[9]
Choi and S
J. Choi and S. A. Rogers, Optimal conditions for pre- shearing thixotropic or aging soft materials, Rheologica Acta , 1 (2020)
2020
-
[10]
Finally, our results probe at whether yield stress mate- rials ultimately relax if given unreasonably long time
by sharpening P (ℓ) below the width set by the ma- terial’s native ℓ0. Finally, our results probe at whether yield stress mate- rials ultimately relax if given unreasonably long time. We remark that a non-zero background fluidity ε is necessary to recreate our experimentally o...
-
[11]
Sudreau, M
I. Sudreau, M. Auxois, M. Servel, E. L´ ecolier, S. Man- neville, and T. Divoux, Residual stresses and shear- induced overaging in boehmite gels, Phys. Rev. Mater. 6, L042601 (2022)
2022
-
[12]
Edera, M
P. Edera, M. Bantawa, S. Aime, R. T. Bonnecaze, and M. Cloitre, Mechanical tuning of residual stress, memory, and aging in soft glassy materials, Physical Review X 15, 011043 (2025)
2025
-
[13]
C. E. Owens, A. J. Hart, and G. H. McKinley, Improved rheometry of yield stress fluids using bespoke fractal 3d printed vanes, Journal of Rheology 64, 643 (2020)
2020
-
[14]
Hendricks, A
J. Hendricks, A. Louhichi, V. Metri, R. Fournier, N. Reddy, L. Bouteiller, M. Cloitre, C. Clasen, D. Vlas- sopoulos, and W. J. Briels, Nonmonotonic stress relax- ation after cessation of steady shear flow in supramolec- ular assemblies, Phys. Rev. Lett. 123, 218003 (2019)
2019
-
[15]
Vinutha, M
H. Vinutha, M. Marchand, M. Caggioni, V. V. Vasisht, E. Del Gado, and V. Trappe, Memory of shear flow in soft jammed materials, arXiv preprint arXiv:2312.00251 (2023)
2023 arXiv
-
[16]
I. M. Carraretto, C. E. Owens, and G. H. McKin- ley, Time-resolved rheometry of coarsening foams using three-dimensionally printed fractal vanes, Physics of Flu- ids 34 (2022)
2022
-
[17]
Barik and S
S. Barik and S. Majumdar, Origin of two distinct stress relaxation regimes in shear jammed dense suspensions, Physical Review Letters 128, 258002 (2022)
2022
-
[18]
Khabaz, B
F. Khabaz, B. F. Di Dio, M. Cloitre, and R. T. Bon- necaze, Transient dynamics of soft particle glasses in startup shear flow. part i: Microstructure and time scales, Journal of Rheology 65, 241 (2021)
2021
-
[19]
Nicolas, E
A. Nicolas, E. E. Ferrero, K. Martens, and J. L. Bar- rat, Deformation and flow of amorphous solids: Insights from elastoplastic models, Reviews of Modern Physics 90, 45006 (2018)
2018
-
[20]
Kamani, G
K. Kamani, G. J. Donley, and S. A. Rogers, Unification of the rheological physics of yield stress fluids, Physical review letters 126, 218002 (2021)
2021
-
[21]
Full test parameters are detailed in Sec SI-1 E with additional exploration of model parameters in Sec SI-1 B
and the lower values for older foam align with results showing that soft glasses may embrittle with age [5]. Full test parameters are detailed in Sec SI-1 E with additional exploration of model parameters in Sec SI-1 B. Distribution of Local Strain - EP-PLY makes a pre- dictio...
-
[22]
S. M. Fielding, P. Sollich, and M. E. Cates, Aging and rheology in soft materials, Journal of Rheology 44, 323 (2000)
2000
-
[23]
H. J. Barlow, J. O. Cochran, and S. M. Fielding, Ductile and Brittle Yielding in Thermal and Athermal Amor- phous Materials, Physical Review Letters 125, 168003 (2020), arXiv:1912.07457
2020 arXiv
-
[24]
Caggioni, V
M. Caggioni, V. Trappe, and P. T. Spicer, Variations of the herschel–bulkley exponent reflecting contributions of the viscous continuous phase to the shear rate-dependent stress of soft glassy materials, Journal of Rheology 64, 413 (2020)
2020
-
[25]
For our use it is interchangeable with a time-strain plot
We use ‘Kymograph’ to denote a plot showing the change in a function over time, although it is not a ‘space’ plot as the name would suggest, but rather is a function of strain element. For our use it is interchangeable with a time-strain plot
-
[26]
F. A. Lavergne, P. Sollich, and V. Trappe, Delayed elastic contributions to the viscoelastic response of foams, Jour- nal of Chemical Physics 156, 10.1063/5.0085773 (2022)
2022 doi
-
[27]
Nikoumanesh, C
E. Nikoumanesh, C. J. M. Jouaneh, and R. Poling- Skutvik, Elucidating the role of physicochemical inter- 6 actions on gel rheology, Soft Matter (2024)
2024
-
[28]
V. K. Ward and S. M. Fielding, Shear banding as a cause of non-monotonic stress relaxation after flow cessation, arXiv preprint arXiv:2408.11223 (2024)
2024 arXiv
-
[29]
N. C. Keim, J. D. Paulsen, Z. Zeravcic, S. Sastry, and S. R. Nagel, Memory formation in matter, Rev. Mod. Phys. 91, 035002 (2019)
2019
-
[30]
K¨ ursten, V
R. K¨ ursten, V. Sushkov, and T. Ihle, Giant kovacs-like memory effect for active particles, Physical Review Let- ters 119, 188001 (2017)
2017
-
[31]
Barnes and K
H. Barnes and K. Walters, The yield stress myth?, Rhe- ologica acta 24, 323 (1985)
1985
-
[32]
H. A. Barnes, The ‘yield stress myth?’paper–21 years on, Applied Rheology 17, 43110 (2007)
2007
-
[33]
high enough
E. Younes, M. Himl, Z. Stary, and T. Burghelea, In- situ visualisation of the micro-structure of a carbopol gel during a confined microscopic flow, Journal of Non- Newtonian Fluid Mechanics 296, 104630 (2021). 7 EM-1. END MA TTER A. Analytical yield stress from model Equation ...
2021
Reviewed August 7, 2026 · model on record in the stance chip above.
Discussion (0). Sign in to comment.