Pith. sign in

REVIEW 4 major objections 6 minor 14 references

Rate-independent hysteretic energy dissipation in collagen fibrils

T0 review · 4 major / 6 minor · reviewed 2026-08-06 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read Hydrated collagen fibrils show rate-independent hysteresis with return point memory, and a three-function hysteresis operator fitted to one indentation cycle predicts the force and dissipated energy for arbitrary tip trajectories.

desk verdict The core rate-independence observation is solid and the model does real out-of-sample work; the 'arbitrary trajectories' claim is broader than the single tested trajectory supports. read the letter →

arxiv 2507.10841 v1 pith:UP5ZYOJM submitted 2025-07-14 cond-mat.soft physics.bio-ph

Robert Magerle , Paul Zech , Martin Dehnert , Alexandra Bendixen , Andreas Otto This is my paper · ORCID
classification cond-mat.softphysics.bio-ph
keywords rate-independenthysteresisreturnpointmemorycollagenfibrilsatomicforcemicroscopynanoindentationenergydissipationoperatorelastoplasticindentation
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper reports that when an atomic force microscope tip indents a hydrated collagen fibril slowly, at tip speeds between 0.01 and 1 µm/s, the measured force-distance curves do not change shape with speed: the dissipated energy per cycle is rate-independent and the system shows return point memory. It proposes a generic hysteresis model in which the entire force response is determined by three time-independent functions—the approach curve, a straight unloading branch with a fixed contact stiffness, and a constant attractive force limit—combined through the operator Γ̂[δ] = max{fR, min{fA, fT}}. The claim is that this operator, calibrated on one large approach-retract cycle, predicts force and dissipated energy for arbitrary indentation trajectories, unifying elastoplastic indentation, capillary adhesion, and surface leveling at these slow rates. A sympathetic reader would care because it identifies a previously unrecognized dissipation mechanism in a load-bearing connective-tissue component and provides a parameter-light predictive tool for nanoindentation.

What carries the argument

The central object is the hysteresis operator Γ̂ defined in eqn (3): for indentation δ and last return point δ*i, the output force is max{fR(δ), min{fA(δ), fT(δ, δ*i)}}. Here fA is a power-law approach curve of contact-model form, fT is a straight elastic unloading line with slope κT, and fR is the constant attractive (capillary and adhesion) force. The operator is rate-independent by construction because it depends on the sign of δ − δ*i rather than on time, and it implements return point memory by storing only the last return point. The same form is a generalized stop operator, with fA and fR serving as the upper and lower output bounds; the predicted forces, dissipated energies, and complex-trajectory behaviors all follow from plugging calibrated parameters from one measured cycle into this operator.

What would settle it

Measure cyclic force-distance curves with small retract-approach loops starting from several different return-point depths on the same fibril and compare the measured unloading slopes to the κT value from one deep cycle; if the initial slope of a small loop at shallow depth is systematically different, the straight-line unloading assumption fails. Alternatively, ramp the tip velocity above 1 µm/s, where inner hysteresis loops appear due to viscous phase lag, to test the stated rate-independent regime boundary.

Watch

Extended reading notes

Core claim

On its own terms, the central discovery is that hydrated type I collagen fibrils, when indented perpendicular to their axis at velocities at or below 1 µm/s, dissipate energy through a rate-independent hysteretic process with return point memory rather than through viscous friction. The shape of force-distance curves is independent of tip velocity in this range, and the dissipated energy Edis depends only on the maximal force setpoint FM, with an analytical expression provided by the model. Return point memory manifests in trajectories with intermediate retract-approach cycles: at the end of each small cycle the force returns to the value at the cycle's start, so the system state depends only on the last return point. The model's operator reproduces the measured force and dissipated energy for complex trajectories, with deviations only at the first small internal loop and the final retraction.

Load-bearing premise

The model assumes that unloading from any return point follows a straight line with a single, depth-independent contact stiffness κT taken from one large indentation cycle; if κT actually varies with indentation depth or with the position of the return point, the predicted force and dissipated energy for complex trajectories will drift away from the data.

Editorial extensions

If this is right

  • Given one calibrated large indentation cycle, the operator predicts the force output for any other indentation trajectory without fitting time-dependent viscoelastic parameters.
  • The analytical expression for Edis(FM) gives a velocity-free prediction of how much energy each indentation cycle dissipates as a function of indentation amplitude.
  • The model unifies three previously distinct phenomena—elastoplastic indentation, capillary adhesion and bridge formation, and surface leveling—into one rate-independent description.
  • Because native hydrated collagen fibrils are in the glassy state, the authors anticipate that rate-independent hysteretic energy dissipation will also be found in other soft condensed matter and biological materials at slow deformation rates.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • If κT turns out to vary with indentation depth or with the position of the return point, a natural extension would be to replace the straight unloading line fT with a curved family parameterized by the return point; the max/min operator form could survive but would then encode memory beyond the last return point.
  • The operator's resemblance to a generalized stop operator suggests that superposition-based hysteresis descriptions in the Preisach family could be adapted to nanoindentation data if nested-loop measurements at multiple amplitudes were available.
  • A testable extension is to fit the EPICAL operator on one fibril and then check whether it transfers to other positions along the same fibril or to other hydrated biopolymer fibrils; if the mechanism is generic, the calibrated parameters should predict new force curves without refitting.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

4 major / 6 minor

Summary. The manuscript reports AFM nanoindentation measurements on hydrated collagen fibrils showing that the force-distance hysteresis and dissipated energy are independent of tip velocity in the range 0.01–1 um/s, and that the response exhibits return point memory. The authors construct a phenomenological hysteresis operator (eqns 1–3) whose parameters are extracted from one measured approach-retract cycle, and show that it predicts the dissipated energy as a function of maximum force (eqn 6) and reproduces the main features of a complex indentation trajectory with multiple intermediate retract-approach cycles (Fig. 3). The model combines elastoplastic indentation, a linear elastic unloading branch with constant stiffness kT, and an attractive capillary force floor, with an additional elementary hysteresis in the capillary region.

Significance. If the rate-independence result is robust, it is a significant finding for AFM-based nanomechanics of collagen and soft materials, because low-speed hysteresis is usually attributed to viscoelasticity. The closed-form Edis(FM) relation and the out-of-sample prediction of a complex trajectory are clear strengths, and the model is parameterized from a single cycle, which aids reproducibility. The main weaknesses are the limited validation of the 'arbitrary trajectories' claim and the reliance on an amplitude-independent kT. The paper is otherwise careful in distinguishing rate-independent from rate-dependent regimes and provides direct stress-relaxation and creep controls.

major comments (4)
  1. [§2.9, Fig. 3] The claim that the hysteresis operator predicts the force and dissipated energy for 'arbitrary indentation trajectories' rests on a single complex trajectory measured at one velocity (vtip = 0.04 um/s) on one isolated fibril. The authors themselves note deviations on the first small retract-approach loop and the last retraction (Section 2.9). Please add a quantitative error analysis (e.g., RMS force error, relative Edis error) for the Fig. 3 prediction and, ideally, test additional trajectories with different return-point depths and loop amplitudes. If such data are not available, the wording in the abstract and Section 2.8 should be tempered to 'the tested trajectory class.'
  2. [§2.4, eqn (5)] The model's unloading branch fT uses a single constant kT determined from the retraction slope of one large indentation, yet Fig. 2c shows that kT increases for FM > 5 nN and Section 2.6 reports only that intermediate excursions have stiffness 'similar to kT.' Because this parameter controls the entire unloading behavior, a depth- or amplitude-dependent kT would propagate directly into the predicted force and Edis for arbitrary trajectories. Please report kT for return points at different indentation depths and amplitudes (e.g., the intermediate loops of Fig. 3) and quantify how the observed variation affects the predictions, or state the range of return-point amplitudes over which kT is effectively constant.
  3. [§2.8, eqns (1)-(3)] The quantitative model specification is incomplete for the attractive region 0 < delta < delta_off, where the text invokes an 'elementary hysteresis model' (ref. 43) without giving its equations. Since the full-trajectory prediction in Fig. 3 includes the attractive parts of the force signal, the model as written is not fully reproducible from the equations provided. Please include the explicit capillary-bridge operator or clearly restrict the quantitative predictions to delta <= 0.
  4. [§2.10, eqn (6)] The closed-form Edis(fM) derivation assumes that the linear unloading from the maximum-force point reaches the attractive limit fc before the tip returns to delta_s (i.e., delta_c = delta_min + (fM - fc)/kT <= delta_s). Please state this condition explicitly and confirm that all FM data points in Fig. 2b satisfy it; for large fM or small kT the loop-closure geometry changes and eqn (6) would need modification.
minor comments (6)
  1. [Section 2 heading] The heading 'Results and disscusion' contains a typo; it should read 'Results and discussion.'
  2. [Fig. 2b] The caption does not mention error bars or the scatter reported as 'typically below 20%.' Please add representative error bars to the main figure or explicitly refer to Fig. S3 in the caption.
  3. [Section 2.3] State the number of measurement positions and cycles per velocity used for the Edis statistics in the main text, rather than only in Fig. S3.
  4. [Eqn (6)] In the compiled manuscript the formula appears with split exponents; please ensure the typeset equation is unambiguous, for example (f_M - f_c)^(1+1/gamma) / ((1+gamma) kappa_A^(1/gamma)).
  5. [ESI footnote] The ESI footnote uses a placeholder DOI (10.1039/cXsm00000x/); update it to the actual supplementary information DOI.
  6. [Section 2.6] The return-point-memory discussion would benefit from an explicit demonstration of wiping out (a larger loop erasing the memory of a smaller one), since that property is central to the operator's definition.

Circularity Check

1 steps flagged · score 2.0 of 10

Central predictions are out-of-sample and non-circular; the only flagged item is an in-sample 'prediction' in Fig. 5b, with self-citations confined to methodology.

  1. fitted input called prediction [Section 2.9 (Comparison with experiments), Fig. 5b]
    "For a given FD data set, measured with large indentation amplitude (as in Fig. 1b), we determine the parameters that describe the EPICAL hysteresis model by fitting functions corresponding to eqns (1a), (1b), and (3) to the measured data. ... An example for the force predicted with the hysteresis operator for the entire indentation cycle is shown in Fig. 5b."

    The parameters of eqns (1a), (1b), and (3) are fitted to the same FD data set shown in Fig. 5a, and Fig. 5b then applies the resulting hysteresis operator to that same entire indentation cycle. The plotted force is therefore the fitted model evaluated on its training data, so calling it 'predicted' is in-sample reproduction rather than an independent forecast. This is a minor wording issue: the paper's actual out-of-sample tests are the complex trajectory in Fig. 3 (parameters fitted to one data set, applied to a different trajectory) and eqn (6), which predicts the Edis(FM) curve without using Edis data in the fit. Those central comparisons are not forced by the fitting step.

full rationale

The paper's central empirical claim is rate-independent hysteresis with return point memory, supported directly by cyclic AFM force-distance measurements at different tip velocities (Fig. 2) and by multi-loop trajectories (Fig. 3). No model is needed for that observation, and no fitted parameter is renamed as the observed rate-independence. The hysteresis operator in eqn (3) is parameterized from one large approach-retract cycle: fA is fitted to the large approach curve, kT is obtained from the large retraction slope via eqn (5), and the capillary force fc is read off from the data. The model is then applied to a different complex indentation trajectory in Fig. 3 and used to derive the closed-form Edis(fM) relation in eqn (6). These are out-of-sample predictions in the sense that the predicted quantities (inner-loop force excursions and dissipated energy as a function of FM) were not used to determine the parameters. The acknowledged deviations at the first small loop and the final retraction show the comparison is genuine rather than tautological. Self-citations (refs 20-22, 25, 42) are confined to measurement methodology, data-analysis procedures, and background; none is used as a load-bearing uniqueness theorem or as a substitute for the present data. The only in-sample presentation is Fig. 5b, which shows the model's force for the same cycle used for parameter fitting under the label 'predicted'; this is a fitted input called prediction, but it is not the basis of the paper's central claim. Overall, the central derivation chain is self-contained and the predictions have independent content, so the circularity score is low.

Assumptions & free parameters 6 free parameters · 5 assumptions · 0 invented entities

The central predictions rest on fitted material-specific parameters (fc, δs, κA, γ, κT, δoff) and on phenomenological assumptions about the hysteresis operator shape, linear elastic unloading, and capillary bistability; no new physical entities are introduced.

free parameters (6)
  • fc = -1.08 (dimensionless; Table 1, Fig 1b dataset)
    Attractive/capillary force, set to the minimal force in the FD approach data; describes adhesion and capillary interaction.
  • δs = -10.0 (dimensionless; Table 1, Fig 1b dataset)
    Surface position parameter used in the approach power law; fitted together with κA and γ and noted to be sensitive.
  • κA = 0.192 (dimensionless; Table 1, Fig 1b dataset)
    Stiffness-like coefficient for the elastoplastic approach curve fA(δ); exhibits large scatter across positions due to scale invariance.
  • γ = 1.26 (dimensionless; Table 1, Fig 1b dataset)
    Power-law exponent of the approach curve; fitted to FD approach data.
  • κT = 3.58 (dimensionless; Table 1, Fig 1b dataset)
    Dimensionless contact stiffness obtained from the retraction slope via eqn (5); used for all linear elastic unloading lines.
  • δoff = not tabulated (determined from retraction data)
    Distance at which the capillary bridge breaks; sets the upper bound of the bistable capillary region and the memory-erasing leveling domain.
assumptions (5)
  • domain assumption At vtip <= 1 µm/s, viscous friction and time-dependent relaxation/creep are negligible during FD measurements, so hysteresis is purely rate-independent.
    Supported by overlap of FD curves for vtip = 0.01-1 µm/s (Fig 2a) and by stress relaxation/creep rates below tip velocity (Appendix C); if false, the central rate-independence claim fails.
  • domain assumption The tip-sample interaction for δ <= 0 is captured by a generalized stop operator with boundaries fA and fR and linear elastic unloading fT (eqns 1-3).
    Phenomenological; the operator form is chosen to reproduce measured FD hysteresis and return point memory; not derived from molecular mechanisms.
  • domain assumption Linear elastic response with constant stiffness κT holds for all intermediate retract-approach cycles regardless of amplitude and history (eqn 1b).
    Authors note deviations for first small loop and last retract (Section 2.9); if κT were amplitude-dependent, arbitrary-trajectory predictions would fail.
  • domain assumption Capillary bridge formation/collapse is an elementary bistable hysteresis for 0 < δ < δoff, and memory is erased when δ >= δoff (surface leveling).
    Modeled after ref 43, not derived here; the paper does not provide independent verification of this bistability.
  • domain assumption The repulsive force during approach follows a power-law form resembling Hertz contact even beyond elastic regime (eqn 1a).
    Used phenomenologically; authors acknowledge Hertz models are for elastic materials only (Section 2.5).

how reviews work

0 comments
Cite this review

Pith. "Pith review of Rate-independent hysteretic energy dissipation in collagen fibrils." pith.science (2026). https://pith.science/paper/UP5ZYOJM

@misc{pith2026250710841,
  author       = {Pith},
  title        = {Pith review of: Rate-independent hysteretic energy dissipation in collagen fibrils},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/UP5ZYOJM}},
  note         = {Machine review of arXiv:2507.10841}
}
abstract

Nanoindentation cycles measured with an atomic force microscope on hydrated collagen fibrils exhibit a rate-independent hysteresis with return point memory. This previously unknown energy dissipation mechanism describes in unified form elastoplastic indentation, capillary adhesion, and surface leveling at indentation velocities smaller than 1 $\mu$m s$^{-1}$, where viscous friction is negligible. A generic hysteresis model, based on force-distance data measured during one large approach-retract cycle, predicts the force (output) and the dissipated energy for arbitrary indentation trajectories (input). While both quantities are rate independent, they do depend nonlinearly on indentation history and on indentation amplitude.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

14 extracted references · 14 canonical work pages

  1. [1]

    Accepted Manuscript first published on 20 Feb 2024 inSoft Matter, 2024, 20, 2831–2839; DOI: 10.1039/D3SM01625K Rate-independent hysteretic energy dissipation in colla- gen fibrils† Robert Magerle, a Paul Zech, a Martin Dehnert, a Alexandra Bendixen, a and Andreas Otto a Nanoindentation cycles measured with an atomic force microscope on hydrated collagen f...

    work page Pith review arXiv 2024

  2. [14]

    57 J. Li, B. Lu, Y. Zhang, H. Zhou, G. Hu and R. Xia,Mater . Chem. Phys., 2020, 241, 122391. 58 H. Batzer and U. T. Kreibich, Polym. Bull., 1981, 5, 585–590. 59 A. V. Tobolsky ,J. Appl. Phys. , 1956, 27, 673–685. 60 A. Nicolas, E. E. Ferrero, K. Martens and J.-L. Barrat, Reviews of Modern Physics , 2018, 90, 045006. 61 J. E. Sader, J. W. M. Chon and P. Mu...

    work page 2020

  3. [226]

    Zitzler, S

    38 L. Zitzler, S. Herminghaus and F. Mugele, Phys. Rev. B , 2002, 66, 155436. 39 B. J. Briscoe, L. Fiori and E. Pelillo, J. Phys. D: Appl. Phys. , 1998, 31,

    work page 2002

  4. [344]

    Stolz, R

    14 M. Stolz, R. Raiteri, A. Daniels, M. R. VanLandingham, W. Baschong and U. Aebi, Biophys. J., 2004, 86, 3269–3283. 15 C. A. Grant, D. J. Brockwell, S. E. Radford and N. H. Thomson, Biophys. J., 2009, 97, 2985–2992. 16 S. J. Baldwin, A. S. Quigley , C. Clegg and L. Kreplak, Biophys. J., 2014, 107, 1794–1801. 17 O. G. Andriotis, P. J. Thurner, W. Manuyako...

    work page 2004

  5. [847]

    Krieg, G

    28 M. Krieg, G. Fläschner, D. Alsteens, B. M. Gaub, W. H. Roos, G. J. L. Wuite, H. E. Gaub, C. Gerber, Y. F. Dufrêne and D. J. Müller, Nat. Rev. Phys., 2019, 1, 41–57. 29 J. P. Sethna, K. Dahmen, S. Kartha, J. A. Krumhansl, B. W. Roberts and J. D. Shore, Phys. Rev. Lett. , 1993, 70, 3347–

    work page 2019

  6. [1991]

    Bobbio, G

    44 S. Bobbio, G. Milano, C. Serpico and C. Visone, IEEE T. Mag., 1997, 33, 4417–4426. 45 T. Wang, M. Noori, W. A. Altabey , M. Farrokh and R. Ghiasi, P. I. Mech. Eng. L.: J. Mat. , 2021, 235, 2639–2653. 46 A. Visintin, Differential Models of Hysteresis , Springer, Berlin, Heidelberg,

    work page 1997

  7. [1993]

    24 Y. M. Efremov, T. Okajima and A. Raman, Soft Matter, 2020, 16, 64–81. 25 R. Garcia, C. J. Gómez, N. F. Martinez, S. Patil, C. Dietz and R. Magerle, Phys. Rev. Lett., 2006, 97, 016103. 26 R. Proksch and D. G. Yablon, Appl. Phys. Lett. , 2012, 100, 073106. 27 Y. F. Dufrêne, D. Martínez-Martín, I. Medalsy , D. Alsteens and D. J. Müller, Nat. Methods, 2013, 10,

    work page 2020

  8. [1996]

    Kuhnen, Eur

    48 K. Kuhnen, Eur . J. Control, 2003, 9, 407–418. 49 F. Moisy ,EzyFit 2.44, MATLAB Central File Exchange,

    work page 2003

Show all 14 references
  1. [2008]

    Gautieri, S

    2 A. Gautieri, S. Vesentini, A. Redaelli and M. J. Buehler, Nano Lett., 2011, 11, 757–766. 3 J. P. R. O. Orgel, T. C. Irving, A. Miller and T. J. Wess, Proc. Natl. Acad. Sci. U. S. A. , 2006, 103, 9001–9005. 4 D. W. L. Hukins and J. Woodhead-Galloway , Mol. Cryst. Liq. Cryst.,...

    2011

  2. [2023]

    50 Z. L. Shen, M. R. Dodge, H. Kahn, R. Ballarini and S. J. Eppell, Biophys. J., 2010, 99, 1986–1995. 51 Z. L. Shen, H. Kahn, R. Ballarini and S. Eppell, Biophys. J. , 2011, 100, 3008–3015. 52 J. Liu, D. Das, F. Yang, A. G. Schwartz, G. M. Genin, S. Tho- mopoulos and I. Chasio...

    2010

  3. [2315]

    Stolz, R

    10 M. Stolz, R. Gottardi, R. Raiteri, S. Miot, I. Martin, R. Imer, U. Staufer, A. Raducanu, M. Düggelin, W. Baschong, A. U. Daniels, N. F. Friederich, A. Aszodi and U. Aebi, Nat. Nan- otechnol., 2009, 4, 186–192. 11 A. D. Kemp, C. C. Harding, W. A. Cabral, J. C. Marini and J. ...

    2009

  4. [2395]

    Oliver and G

    40 W. Oliver and G. Pharr, J. Mater . Res., 1992, 7, 1564–1583. 41 B. Cappella and G. Dietler, Surf. Sci. Rep., 1999, 34, 1–104. 42 M. Dehnert and R. Magerle, Nanoscale, 2018, 10, 5695–5707. 43 I. D. Mayergoyz, Mathematical Models of Hysteresis , Springer, New York,

    1992

  5. [3350]

    Bertotti, Hysteresis in Magnetism , Academic Press, San Diego, 1998, pp

    8 | 1–9 + P V S O B M / B N F < Z F B S > < W P M > 30 G. Bertotti, Hysteresis in Magnetism , Academic Press, San Diego, 1998, pp. 433–477. 31 F. Preisach, Z. Phys., 1935, 94, 277–302. 32 L. Delaey , R. V. Krishnan, H. Tas and H. Warlimont, J. Mater . Sci., 1974, 9, 1521–1535....

    1998

  6. [5942]

    8 O. G. Andriotis, S. Desissaire and P. J. Thurner, ACS Nano , 2018, 12, 3671–3680. 9 C. Zapp, A. Obarska-Kosinska, B. Rennekamp, M. Kurth, D. M. Hudson, D. Mercadante, U. Barayeu, T. P. Dick, V. Deny- senkov, T. Prisner, M. Bennati, C. Daday , R. Kappl and F. Gräter, Nat. Com...

    2018

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.