REVIEW 3 major objections 6 minor 27 references
Atom probe tomography of hydrated biomacromolecules: preliminary results
T0 review · 3 major / 6 minor · reviewed 2026-08-06 · deepseek-v4-flash
Pith's one-line read The paper demonstrates a cryo-atom-probe pipeline for bulk frozen aqueous samples and shows that plunge freezing destroys ferritin, while high-pressure-frozen water passes through the same workflow as a route to intact biomacromolecules.
desk verdict Honest cryo-APT methods report on hydrated ferritin: real data, clear limitations, but the HPF path to intact biomolecules rests on an unconfirmed vitreous state. 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 piece is the cryogenic specimen pipeline: re-deposition cryo-lift-out, in which a chromium lamella is sputtered from a manipulator to weld the frozen lamella to a silicon post, followed by annular FIB milling and transfer in a cryo-UHV suitcase held below roughly $-180\,^\circ\mathrm{C}$ (90 K). Into this pipeline the paper inserts two freezing routes: manual plunge freezing into liquid nitrogen, which the results show is too slow to vitrify, and high-pressure freezing (HPF) on an aluminum planchette, which is intended to produce amorphous ice. The cluster ion ratio (CIR), the abundance ratio of larger to smaller protonated water clusters, is used as a field-sensitive fingerprint to compare datasets and to argue that longer clusters from the HPF specimen signal vitreous rather than crystalline ice.
What would settle it
Run a high-pressure-frozen ferritin solution through the same cryo-lift-out and transfer workflow, then before atom probe analysis record electron diffraction in a cryo-TEM: sharp Debye-Scherrer rings from crystalline ice, or an APT mass spectrum with only small protonated water clusters ($n\le3$) and iron segregated at the ice/metal interface, would show that the vitreous path to intact biomacromolecules is not yet established.
Extended reading notes
Core claim
The central claim is that a complete cryogenic specimen-preparation-to-atom-probe pipeline—plunge freezing on a metal substrate, cryogenic lift-out with chromium redeposition welding, transfer in a cryo-vacuum suitcase, and laser-pulsed field evaporation—can be applied to bulk hydrated samples, but that the freezing step, not the analysis, is currently the bottleneck. In the ferritin datasets, Fe-rich particles appear near the ice/metal interface with the iron isotope ratio of ferritin cores, surrounded by carbon-enriched shells but with sodium concentrated at roughly 30 at% in the surrounding brine; the authors interpret this as crystallization-driven damage, with solutes rejected into the last liquid and the protein shell unfolded and dispersed. In high-pressure-frozen pure water, the same workflow produced protonated water clusters up to $(\mathrm{H_2O})_{12}\mathrm{H}^+$, in contrast to the small clusters ($n=1$–$3$) from crystalline ice, which the authors take as evidence that vitreous ice can survive cryo-transfer and FIB preparation and that intact biomacromolecules could be analyzed this way in the future.
Load-bearing premise
The load-bearing premise is that the high-pressure-frozen water stayed amorphous from freezing through transfer and FIB milling, since the paper could not confirm the vitreous state of its HPF samples.
Editorial extensions
If this is right
- Bulk aqueous solutions, not just thin ice layers on pre-sharpened tips, can be prepared into atom probe needles by cryo-lift-out and analyzed at near-atomic scale.
- Plunge freezing in liquid nitrogen is not an acceptable preservation route for hydrated biomacromolecules: the crystallization it produces destroys protein shells and redistributes their elements, so future work must use vitrification.
- High-pressure-frozen water is compatible with the existing cryo-lift-out and cryo-transfer workflow, so vitrified biological solutions can be targeted in the same setup.
- Even when ferritin is destroyed, its iron core leaves a detectable isotopic signature, giving a built-in marker for assessing particle preservation in future experiments.
Reading between the lines
- If the HPF ice had devitrified during transfer or FIB milling, the long water clusters would not prove vitreous compatibility; a cryo-TEM electron diffraction check on a specimen from the same planchette would settle that.
- The paper attributes damage to freezing, but Ga implantation (up to nearly 100 at% in some clusters) and FIB heating are documented in the HPF specimen; separating these damage channels will require a vitrified protein sample that is FIB-milled and compared with one that is not.
- Because ferritin cores contain roughly 80 at% Fe and a defined isotope ratio, counting intact cores per analyzed volume could become a quantitative preservation metric once vitrification is confirmed.
- The observed laser-side/shadow-side asymmetry in cluster fragmentation suggests that local field variations across a frozen biological specimen can be read off the mass spectrum, which could be used as an internal field calibration during reconstruction.
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. This manuscript reports cryo-atom probe tomography (cryo-APT) experiments on plunge-frozen ferritin solution and on high-pressure-frozen deionized water. Using cryo-lift-out from nanoporous brass substrates and cryo-transfer under UHV, the authors acquired APT datasets of up to 69 million ions, detected protonated water clusters and various elemental and molecular ions, and identified Fe-rich particles near the ice/metal interface along with C-, Na-, and Cl-containing signals. They conclude that the workflow is feasible for bulk hydrated biological samples, but that LN2 plunge freezing is too slow to vitrify the solution and causes irrecoverable damage to ferritin. A single HPF water dataset is presented as a proof-of-principle that the same workflow could eventually analyze intact biomacromolecules, while explicitly noting that the vitreous state could not be confirmed.
Significance. The paper's main value is a candid, experimentally grounded demonstration of an end-to-end cryo-APT pipeline for frozen biological solutions, including a 69-million-ion dataset and repeatable lift-out from dealloyed brass. It consolidates earlier frozen-water APT work and documents many substrate and preparation failures, which is practically useful for the community. If the ferritin-specific peak assignments and the amorphous state of the HPF specimen were independently confirmed, the forward-looking claim about near-native-state biomacromolecule analysis would be considerably strengthened; as it stands, the significance is incremental but real. The manuscript is honest about background, peak-overlap, and reconstruction limitations, and its central observations are independent experimental evidence rather than circular derivations. The strengths are the actual datasets, the explicit reporting of negative results, and the use of a comparison metric (CIR) consistent with prior published water-APT data.
major comments (3)
- [Section 3.2, Figure 4 and Suppl. Table 1] The attribution of organic peaks to ferritin shell fragments is not supported by a ferritin-free control. The authors state that 'normally the solution only contains NaCl' and that most non-water peaks 'could be related to fragments from the polypeptide chains,' but no buffer-only mass spectrum is shown, and Section 4.4 concedes that unambiguous identification is impossible at the achieved mass resolution (e.g., C2H4+, CNH2+, and CO+ overlap near 28 Da). Because the claim that elements from the ferritin shell are detected is central to the biological interpretation, a control spectrum of 0.15 M NaCl (and ideally of the brass substrate leachate) is needed before these peaks are assigned to ferritin. As written, the shell detection is plausible but not demonstrated.
- [Section 3.4 and Section 4.3] The HPF proof-of-principle is conditional on an unverified assumption: the text admits 'the vitreous state of our samples could not be confirmed' (Section 3.4) and later 'we assume that the specimen is maintained... below the temperature that would cause significant sublimation and/or crystallization' (Section 4.3). If devitrification occurred during transfer or FIB preparation, the HPF dataset would be equivalent to the crystalline plunge-frozen data and would not support the conclusion that the workflow preserves biomacromolecules in their native state. Moreover, the HPF specimen is pure deionized water, so it provides no direct evidence of biomolecule preservation. I recommend either confirming the amorphous state by an independent method (cryo-TEM/electron diffraction or cryo-Raman) or explicitly limiting the conclusion to frozen, not vitreous, specimens.
- [Section 3.3, Figure 6] The evidence for ferritin cores rests on one Fe-rich cluster and one region of overlapping 54/56 Da distributions, while the main 56 Da peak in the dataset is admitted to be misidentified (C2O2+ or C2NOH+ at the same mass). The proximity histogram in Figure 6d reports nearly 80 at% Fe, but the manuscript does not give the background/range correction parameters or the spatial cluster statistic used to define the particle. Given the strong local magnification and trajectory aberrations expected at the ice/brass interface (Section 4.1), an alternative explanation as a reconstruction artifact is not excluded. Please report the analysis parameters and a significance test for Fe clustering.
minor comments (6)
- [Abstract and Section 2.3] The word 'vitrified' is used for HPF water before the caveat in Section 3.4; please rephrase as 'frozen by high-pressure freezing' to avoid making a claim the paper later retracts.
- [Section 3.1] The sentence 'even if the these cannot be monitored' contains a typo and should read 'even if these cannot be monitored'.
- [Figure 6 and text] The text says the composition profile is plotted in Figure 6c, but the caption labels the proximity histogram as (d); please check and correct the cross-reference.
- [Section 4.3] The sentence 'our experimental infrastructure does now allow for transferring the specimen into a TEM' appears to mean 'does not allow'; please correct.
- [Suppl. Figure 9 caption] The caption contains a typo: 'the underlying meta substratel' should be 'the underlying metal substrate'.
- [Acknowledgements] Thanking 'the two reviewers' in the acknowledgements is unusual in a submitted manuscript and may need to be removed before review.
Circularity Check
No significant circularity: the experimental observations are self-contained; self-citations are methodological and non-load-bearing.
full rationale
The paper's central claims are that plunge-frozen ferritin solution can be prepared and analyzed by cryo-APT although the ferritin is damaged by ice-crystal formation, and that the same cryogenic workflow can in principle be applied to high-pressure-frozen (HPF) samples. Both claims rest on measured mass spectra, 3D reconstructions, and composition profiles reported in this paper, not on quantities defined by the cited prior work. The cluster-ion ratio (CIR) from Woods et al. (2025) is used only as a convenient comparison metric for electrostatic-field conditions during analysis; the ferritin observations and the HPF proof-of-principle do not reduce to that metric. The cryo-lift-out protocol and substrate choices are cited from earlier same-group papers, but those citations concern specimen-preparation methods, which are inputs to the experiment rather than the claimed conclusions. The forward-looking HPF claim is explicitly conditional: the paper states in Section 3.4 that 'the vitreous state of our samples could not be confirmed' and that the amorphous state 'should be maintained' if the temperature remained below 90 K. This is an acknowledged unverified assumption and a correctness risk, but it is not circularity—nothing in the paper defines the HPF result in terms of the assumption or derives the assumption from the result. No fitted parameter is renamed as a prediction, no uniqueness theorem is imported from the authors' prior work to force a choice, and no known result is repackaged under new coordinates. The derivation chain is therefore self-contained with respect to the paper's actual experimental evidence, so the circularity score is 0.
Assumptions & free parameters
assumptions (3)
- domain assumption Cryogenic transfer and FIB preparation maintained HPF ice below the devitrification temperature so that the amorphous state was preserved.
- domain assumption No significant sublimation of ice occurred during the cryogenic transfer and APT alignment.
- domain assumption The organic and iron-containing peaks observed in the mass spectra originate from ferritin rather than from contamination or the substrate.
Cite this review
Pith. "Pith review of Atom probe tomography of hydrated biomacromolecules: preliminary results." pith.science (2026). https://pith.science/paper/T6QKP7WW
@misc{pith2026250720970,
author = {Pith},
title = {Pith review of: Atom probe tomography of hydrated biomacromolecules: preliminary results},
year = {2026},
howpublished = {\url{https://pith.science/paper/T6QKP7WW}},
note = {Machine review of arXiv:2507.20970}
}
read the original abstract
The folding and structure of biomacromolecules depend on the 3D distributions of their constituents, which ultimately controls their functionalities and interactions with other biomacromolecules. Atom probe tomography (APT) with its unparalleled compositional sensitivity at nanoscale spatial resolution, could provide complementary information to cryo-electron microscopy, yet routine APT analysis of biomacromolecules in their native state remains challenging. Here, a ferritin solution was used as a model system. Following plunge freezing in liquid nitrogen, cryogenic lift-out and cryo-APT analysis were performed. Elements from the ferritin core and shell are detected yet particles seem destroyed. We hence demonstrate the feasibility of preparing and analyzing bulk hydrated biological samples using APT, however, the cooling was too slow to vitrify the solution. This caused irrecoverable damage to the protein shell surrounding the ferritin particles due to ice crystal formation. We report on preliminary data from high-pressure frozen (HPF) deionized (DI) water, demonstrating a proof-ofprinciple experiments that intact biomacromolecules could be analyzed through a similar workflow in the future. We report on many trials (and errors) on the use of different materials for substrates and different substrate geometries, and provide a perspective on the challenges we faced to facilitate future studies across the community.
Reference graph
Works this paper leans on
-
[1]
t the specimen’s apex in the range of 1010 V.m-1, using moderate high voltages in the range of 2 – 12 kV (Lefebvre-Ulrikson et al., 2016). Such fields enable the removal 2 of atoms from the specimen’s surface in the form of ions, a process termed field evaporation. The ions then fly toward a position-sensitive detector, which collects each ion’s impact po...
work page 2016
-
[3]
that can indicate micro-fractures, possibly due to the brittle and maybe porous nature of the specimen, and uneven surface. However, these are not uncommonly strong, with peaks of only up to a factor of 5 times the set detection rate. Reconstructed 3D maps of different APT specimens vary in length, and some include the ice/metal interfaces or only the fro...
work page 2003
-
[4]
as readily visible in Figure 1d. Therefore, a series of line patterns (30 kV, 40 pA, depth (Si): 0.7 µm) was firstly milled on the interface between the support metal and Cr, to attach the Cr/supporting metal interface via re-deposition (Schreiber et al., 2018; Eric V Woods et al., 2023). To enhance the strength of the junction a rectangular pattern (30 k...
work page 2018
-
[5]
Overall, the background decreases very slightly with increasing laser pulse energy from 80 to 120 pJ, and reaches a maximum at 60pJ. The increase in laser pulse frequency from 60 – 200 kHz leads to only a moderate increase of the level of background from approx. 8 to approx. 11 ppm.ns-1. These trends are consistent with previous reports (Schwarz et al., 2...
work page 2020
-
[6]
At a constant laser pulse energy, the voltage increase is relatively steady, as expected during the APT analysis, to compensate for the progressive blunting of the specimen. The level of background computed by the commercial software comes from the baseline level of the time-of-flight (ToF) spectrum and is hence expressed in relative counts per ns. This i...
work page 2013
-
[8]
Details pertaining to these amino acids are reported in Suppl
When considering the identification of the peaks , colored vertical lines were added in Figure 4b at masses of some of the identified amino acids found in the protein shell around the ferritin core (Hempstead et al., 1997). Details pertaining to these amino acids are reported in Suppl. Table
work page 1997
-
[9]
There is a possibility that series of peaks in the range 115–121 Da could be related to cysteine or threonine or valine that have lost one or more hydrogen atoms. There are similar sets of peaks up to 250 Da and hence above the mass of the heaviest individual amino acid identified, suggesting that combinations of individual amino acids are coming off stil...
work page 2025
-
[10]
Specimen preparation on the half-grid is more time-consuming because the surrounding material needs to be milled away to provide the necessary clearance to enable APT analysis, of approx. 100 µm. This necessitates the removal of a large amount of ice, not unlike what had been proposed by Schwarz et al. (Schwarz et al., 2020), or what was discussed extensi...
work page 2020
Show all 27 references
-
[12]
In addition, the data contains peaks of Al1–3+, Mg1–2+, and Si2+
A similar behavior was previously reported for thin layers of amorphous ice deposited on a pre-sharpened tip (Stintz & Panitz, 1993; Dirks et al., 1992). In addition, the data contains peaks of Al1–3+, Mg1–2+, and Si2+. These detected elements belong to the Al planchette, made...
1993
-
[14]
Here, we reported results on dealloyed brass, but other metal substrates have been tried, including nanoporous gold (NPG), magnesium and aluminum
Milling rate: Considering the specimen preparation time during LO, the milling rate of the metal should be taken into consideration, to reduce the sample preparation time. Here, we reported results on dealloyed brass, but other metal substrates have been tried, including nanop...
2015
-
[15]
Compared to brass, Mg is mechanically softer and has hence a higher FIB-milling rate, closer to that of ice, making is potentially simpler to obtain specimens with a smooth surface
A recent study has demonstrated the formation of a porous hydroxide at the interface during aqueous corrosion of Mg (Schwarz et al., 2024). Compared to brass, Mg is mechanically softer and has hence a higher FIB-milling rate, closer to that of ice, making is potentially simple...
2024
-
[19]
standard
have demonstrated that even simple carbon chains can dissociate into numerous fragments under the extreme electrostatic field conditions of an APT experiment. Several authors recently reported using a “standard” e.g. when analyzing a glucose solution, Schwarz et al. also repor...
2021
-
[20]
Woods et al
to support the identification of peaks as originating from collagen. Woods et al. tried to perform analysis of individual amino acids in solution, but even then the changes in the amplitude of the mass peaks of the different fragments as a function of the electrostatic field c...
2025
-
[21]
on relatively simple molecules already, more will be necessary in the future. There will also be a need to develop 18 data process techniques that offer ways to reconstruct these chains from the incomplete set of information that the APT provides, due to the detector efficienc...
2017
-
[22]
Chang, Y., Lu, W., Guénolé, J., Stephenson, L
https://www.frontiersin.org/articles/10.3389/fmolb.2023.1252529 (Accessed December 4, 2023). Chang, Y., Lu, W., Guénolé, J., Stephenson, L. T., Szczpaniak, A., Kontis, P., Ackerman, A. K., Dear, F., Mouton, I., Zhong, X., Raabe, D., Gault, B., Zhang, S., Dye, D., Liebscher, C....
2019
-
[25]
Parmenter, C. D. & Nizamudeen, Z. A. (2021). Cryo-FIB-lift-out: practically impossible to practical reality. Journal of Microscopy 281, 157–174. Perea, D. E., Gerstl, S. S. A., Chin, J., Hirschi, B. & Evans, James. E. (2017). An environmental transfer hub for multimodal atom p...
2021
-
[26]
E., Liu, J., Bartrand, J., Dicken, Q., Thevuthasan, S
24 Perea, D. E., Liu, J., Bartrand, J., Dicken, Q., Thevuthasan, S. T., Browning, N. D. & Evans, J. E. (2016). Atom probe tomographic mapping directly reveals the atomic distribution of phosphorus in resin embedded ferritin. Scientific Reports 6, 1–9. Prenitzer, B., Urbanik-Sh...
2016
-
[68]
Tsong, T. T. (1978). Field-Ion Image-Formation. Surface Science 70, 211–233. Vurpillot, F., Larson, D. & Cerezo, A. (2004). Improvement of multilayer analyses with a three-dimensional atom probe. In Surface and Interface Analysisvol. 36, pp. 552–558. http://www.scopus.com/inwa...
1978 arXiv
-
[318]
& Sauvage, Xavier
Lefebvre-Ulrikson, Williams., Vurpillot, François. & Sauvage, Xavier. (2016). Atom probe tomography : put theory into practice. Academic Press. 23 Lucas, B. A. & Grigorieff, N. (2023). Quantification of gallium cryo-FIB milling damage in biological lamellae. Proceedings of the...
2016
-
[942]
H., Carpenter, J
Crowe, J. H., Carpenter, J. F. & Crowe, L. M. (1998). THE ROLE OF VITRIFICATION IN ANHYDROBIOSIS. Annual Review of Physiology 60, 73–103. Dietrich, C. A., Schuldt, R., Born, D., Solodenko, H., Schmitz, G. & Kästner, J. (2020). Evaporation and Fragmentation of Organic Molecules...
1998
-
[2000]
The particle contains almost 80 at% Fe, surrounded by a shell enriched in C
is plotted in Figure 6c, including for these two isotopes alongside the main other constituents, namely C, O, H, Cu and Na. The particle contains almost 80 at% Fe, surrounded by a shell enriched in C. Conversely to results from (Perea et al., 2016), no notable increase in P ca...
2016
-
[2006]
and sparse experimental work (Nishikawa & Kato, 1986; Nishikawa et al., 2011, 2009; Gault et al., 2010; Eder et al., 2017; Stoffers et al., 2012; Meng et al.,
1986
-
[2022]
key elements
– and brass – 30-33 V/nm (Tsong, 1978). Considering the complexity of the mass spectrum, which is very similar to Figure 4, not all peaks could be identified and we focus the data analysis here on the distribution of a selected few "key elements": Na+, Fe+, and C+, and the com...
1978
-
[2023]
However, the specimen transfer, data collection and analysis remain extremely challenging, especially for frozen-hydrated specimens
have been reported. However, the specimen transfer, data collection and analysis remain extremely challenging, especially for frozen-hydrated specimens. The development of cryo-transfer devices and gloveboxes (Stephenson et al., 2018; McCarroll et al., 2020; Perea et al., 2017...
2018
-
[2024]
and various solutions (El-Zoka et al., 2020; Schwarz et al., 2021, 2022; Eric V Woods et al., 2023; E. V. Woods et al.,
2020
-
[2025]
or within NPG (E. V. Woods et al., 2023), but this will need to be studied in more details. 16 Figure 10: Mass spectrum from the solution-metal substrate for brass showing a) CuS+ and ZnS+ ions, as well as CuCl+ and ZnCl+ and b) Cu2CN+, Cu2ON+, and Cu2O2H3+. 4.2 Substrate geom...
2023
-
[6082]
The surface contamination with C is also visible, which could come from exposure to the environment or cleaning with alcohol-containing solutions
In Figure 8 the maps of the individual elements and molecular ions are plotted, showing that these elements are primarily located at the bottom of the dataset, confirming that the analysis has reached the interface between the frozen water and the metal, marked with a dashed b...
2003
Reviewed August 6, 2026 · model on record in the stance chip above.
Discussion (0). Continue with ORCID to comment.