Tunable high-temperature itinerant antiferromagnetism in a van der Waals magnet

Discovery of two dimensional (2D) magnets, showing intrinsic ferromagnetic (FM) or antiferromagnetic (AFM) orders, has accelerated development of novel 2D spintronics, in which all the key components are made of van der Waals (vdW) materials and their heterostructures. High-performing and energy-efficient spin functionalities have been proposed, often relying on current-driven manipulation and detection of the spin states. In this regard, metallic vdW magnets are expected to have several advantages over the widely-studied insulating counterparts, but have not been much explored due to the lack of suitable materials. Here, we report tunable itinerant ferro- and antiferromagnetism in Co-doped Fe4GeTe2 utilizing the vdW interlayer coupling, extremely sensitive to the material composition. This leads to high TN antiferromagnetism of TN ~ 226 K in a bulk and ~210 K in 8 nm-thick nanoflakes, together with tunable magnetic anisotropy. The resulting spin configurations and orientations are sensitively controlled by doping, magnetic field, and thickness, which are effectively read out by electrical conduction. These findings manifest strong merits of metallic vdW magnets as an active component of vdW spintronic applications.

2) I think Fig 2 should be modified. Fig. 2a does not really show the reader that the phase diagram in Figure 2g is correct. This is better seen with the data from Fig. S2, which I think might be better placed to the main text. In Fig 2a, it appears as if there is a smoot change in magnetization with doping and one cannot see any clean transition from FM to AFM. It took me a while to realize that the magnetic phase diagram is correct. This is also because of the lack of reference to the SI when starting to describing the trends as discussed above. I think in the current stage, the reader will need a long time to understand the results. 3) For the higher doped samples, it might be nice to take M vs T curves at more applied fields to better understand the appearance of the unknown magnetic phase. 4) The authors motivate their study with 2D magnets and say there is no 2D metallic AFM with a high Neel temperature yet. That is true. However, the authors also did not introduce a 2D one, since their result suggest that the samples become FM when thin. I think they should be more honest about that fact in the abstract and intro, by inserting the word "bulk" before they discuss their results of a metallic vdW AFM. 5) Can the authors comment on the upturn in resistivity at lower T? Why is this? Why is it opposite to the trend in FM samples? 6) I think the authors should discuss the role of disorder in their samples and how this might affect the properties. This might also help for the discussion in respect to point 5. 7) For the calculations, the authors say they used VCA to model the doping, but they do not say how the lattice constants were obtained. Did they use experimental ones? And if yes in accordance with change in doping? And if not did they optimize the structures? And if yes, did they correct for vdW forces? A few more details here would be helpful.
Reviewer #3: Remarks to the Author: This work presents the evolution of magnetic properties in Fe4GeTe2 with Co substitution on Fe site. I will start with some technical comments. 1. Authors have synthesized single crystals, apparently. Therefore, I am puzzled why anisotropy in magnetic properties has not been presented? 2. Supplement section Fig. S1a: crystals have small impurity phase, between the strongest peak around 2theta=27 and (221) there is another reflection in all crystals. Then, x=0.39, 0.3, 0.17 also contain another visible peak between (221) and (444). Since many results in this work rest on bulk magnetic characterization it is of interest to index those peaks and, if due to impurity phases, ensure that bulk magnetic properties are not influenced by impurities. 3. Authors should show saturated moment in mu_B/F.U. where F.U. is formula unit in Fig. 2a in the main text as well as in Fig. S3(a-e). There seems to be a misprint in Fig. 2a, 0.7 is in fact 0.07. 4. I would also show ZFC-FC chi(T) curves. Can authors rule out spin glass state in all investigated crystals? 5. I find Fig. 1b rather interesting…I wonder what is the Ioffe-Regel limit for AFM crystals? Note the absence of metallic resistivity for high x at low temperatures. Throughout the text authors refer to investigated crystals as "metals". Yet, rho(T) looks like a very bad metal for high x. I would advise the use of more precise language. 6. In Fig S5 authors report change in DOS with Co doping. Therefore, it is reasonable to assume some change in electronic structure. Does that have any effect on electrical detection of the spin state. For example, could changes in AMR be influenced by electronic structure changes, in addition to Neel vector changes? I do not think that this work brings high impact or innovation. I agree that AFM transition is high but the last sentence in the introduction is greatly exaggerated. Promising material in spintronics can not operate at 230 K. Authors also say in the conclusion that (Fe,Co)4GeTe2 will enrich the material candidates and spin functionalities for…spintronics. I do not see how. Thicknessdependent magnetic transition falls down with size reduction therefore it is difficult to think of future devices based on nanofabricated crystals investigated in this work. On the other hand, since this paper does report new incremental knowledge, it might be suitable for different, more specialized journal if technical comments above are properly addressed. ----------------------------------------------------------------------------------------------------------------Answers to reviewer #1's questions We sincerely appreciate the reviewer's helpful comments and suggestions. We did our best to answer to each single question and we hope our answers satisfy the reviewer. Figure 1c, The magnetic susceptibility peak near 220 K cannot guarantee a AFM state. Special spin cluster behaviour may also generate a similar peak.

A1-1.
As the reviewer #1 pointed out, the transition to the spin cluster or spin-glass phase can produce a magnetic susceptibility peak, which can be misinterpreted as a signature of the AFM phase transition. In order to rule out the possibility of the spin-cluster state, we measured the temperature dependent magnetic susceptibility, (T), at different cooling procedures, zero-field cooling (ZFC) and field cooling (FC). It has been well known that for the spin-cluster state, the peak of (T) is accompanied by a pronounced bifurcation between ZFC and FC curves. In the (Fe1-xCox) 4GeTe2 single crystal with x = 0.33, however, we observed negligible difference between the ZFC-and FC-(T) curves for both H // c and H // ab as shown in Fig. 2 of the revised manuscript. These results indicate that the peak in (T) for the crystal with x = 0.33 reflects the transition, not to the spin-cluster phase, but to the antiferromagnetic (AFM) phase.
Consistently we also observed no signature of the relaxation behavior, one of the hallmark of the spin cluster state. In the spin cluster phase the change of magnetization requires a sizable time, leading to the time-dependent magnetization phenomena. For example, in the ferromagnetic phase in (Fe1-xCox)4GeTe2 crystals with low x, we found a relaxation behavior with a characteristic time scale of  ~ 600 sec ( Supplementary Fig. S4 with the spin cluster phases in, e.g. PrRhSn3 and Sr2Mn0.7Fe0.3MoO6 [Anand, V. K. et al. Phys. Rev. B 85, 014418 (2012), Wang, X. et al. J. Appl. Phys. 109, 07C322 (2011)]. For x = 0.33, however, we found no signature of relaxation behavior as shown in Supplementary Fig. S4 in the revised manuscript.
Finally, in order to present a more direct experimental evidence of the AFM ordering, we performed resonant soft X-ray scattering experiments for the crystal with x = 0.33, as shown in the supplementary Fig S7. Similar to neutron diffraction, resonant X-ray scattering also offer FIG. S4. Thermoremanent magnetization relaxation in (Fe1−xCox)4GeTe2. a, Magnetic field dependent magnetization for x = 0.07 at T= 10 K. Clear hysteresis behavior with coercive field Hc ~ 6 Oe is observed. b,c, Thermoremanent magnetization relaxation for x = 0.07 at T =10 K with waiting time tw = 193 s (b) and 1350 s (c). The red solid line is the best fit of the relaxation model. d, Magnetic field dependent magnetization for x = 0.33 at T = 10 K. In contrast to x = 0.07 case, no hysteresis behavior is observed. e,f, Thermoremanent magnetization relaxation for x = 0.33 at T = 10 K with waiting time tw = 176 s (e) and 1010 s (f). No signature of relaxation behavior is observed for x = 0.33.
In addition to Bragg peak atq= (0 0 3), additional peak at q= (0 0 3/2) is developed below TN. b, Crystal structure of (Fe0.67Co0.33)4GeTe2with the interlayer AFM structure, indicated by red and blue shades. The period of 19.5 Å, corresponding to q = (0,0,3/2), is consistent with the interlayer AFM phase a probe for AFM phase. In addition to the Bragg peaks at (003), a clear additional peak develops below TN at q = (0,0,3/2), which is absent above TN. The corresponding periodicity is ~ 19.5 Å along the c-axis, in excellent agreement with the interlayer AFM (A-type) structure, predicted by our first principle calculations. Although further studies, using neutron diffraction, are helpful to determine the size and orientation of magnetic moments in the AFM phase of (Fe1-xCox)GeTe2, our resonant X-ray scattering results provide a direct evidence of the AFM phase in (Fe1-xCox)4GeTe2 with x = 0.33.
As the reviewer #1 requested, we added several sentences and additional Figures (Fig. 2) in the main text and Supplementary Information (Figs. S4, and S7), in order to provide experimental evidence of the long-range AFM phase. Figure 2a, The evolution of magnetic loops cannot guarantee a AFM state when the doping is 0.39. It may due to the formation of a large in-plane magnetic anisotropy.

A1-2.
We agree with the reviewer that the magnetization loop alone cannot guarantee the AFM state for (Fe1-xCox)4GeTe2 crystal. As explained in A1-1, we provide several experimental evidence of the AFM transition for x = 0.33, including the temperature dependent magnetic susceptibility (T) with ZFC and FC. As explained for the crystal with x = 0.33 in A1-1, we also performed the magnetic susceptibility measurements on the crystal with x = 0.39 under ZFC and FC procedures for both in H // c and H // ab ( Below T0 = 75 K, however, we found a small bifurcation between ZFC-and FC-(T) curves. This confirms that magnetic structure at low temperatures for x = 0.39 is distinct from the interlayer AFM phase and indicates the competing FM interaction may play a role. In order to identify the low-temperature magnetic structure of the crystal with x = 0.39, the detailed study is needed, which however is beyond the scope of this work. Fig 3a, We can easily explain the data using a ferromagnetic state with large in-plane magnetic anisotropy at low temperature. With increasing temperature, the magnetic anisotropy decreases. Hence, the magnetization can be saturated at lower field at high temperature. Fig  3b, 3c, Using in-plane ferromagnetic state, you can also explain the transport results.

A1-3.
As explained in A1-1, our resonant X-ray scattering provide an experimental evidence for the interlayer AFM phase in (Fe1-xCox)4GeTe2 with x = 0.33. Considering the in-plane magnetic anisotropy of the system, as pointed out by the reviewer, together with the interlayer AFM structure, one can explain the Hall effect and anisotropic magnetoresistance (AMR) results in terms of the canting of moments, as shown illustrated in Fig. 4 of the revised main text. Particularly, under the in-plane magnetic field, the two characteristic kinks at H1 and H2 are consistent with this AFM structure. At low magnetic fields, the Neel vector of each domain is fully aligned at H1, producing a kink in the AMR. With further increasing magnetic fields above H1, canting of magnetic moments along the external field occurs, until all the moments are fully aligned at H2. In the case of the in-plane ferromagnets, however, we expect a single anomaly at the saturation field without a low-field kink. Together with resonant X-ray scattering results, our Hall resistivity and AMR data demonstrate that the spin orientation and configuration can be read out by electrical conduction. Fig 4c and 4d, thicker sample has lower coercivity can be explain quite easily. The reason is that the domain can expand more easily when sample becomes thicker.

A1-4.
As explained in A1-1, the magnetic susceptibility curves for the bulk crystal with x= 0.33 taken during zero-field cooling (ZFC) and field-cooling (FC), show nearly identical temperature dependence. These results rule out the possible spin cluster or spin glass state and are consistent with the resonant X-ray scattering results shown in Supplementary Fig. S7. Similarly, in the nanometer-thick crystals, we observed a clear peak and no thermal hysteresis Anisotropic magnetoresistance ∆ρAMR as a function of temperature and in-plane magnetic field. The low-field and high-field AMR, determined by the relative orientation of Neel vector and the saturated magnetization against the current direction, respectively, which results in a sign change.
in the magnetic susceptibility (T), extracted from the Hall resistivity of the nanoflakes with thickness of 11 L and 16 L. These results indicate the AFM phase is stable in these nanoflakes, while the FM phase appears in thinner crystal with 7 L, as shown in Fig. 5 and Supplementary  Fig. S12.
Furthermore, in the revised manuscript, we present new results on the thickness-dependence of the high Co-doped crystal with x = 0.39. In this case, we found that the AFM phase is stable in nanoflakes with thickness down to ~ 9 L as shown in Fig Supplementary Fig. S12). This spin-flop transition is consistent with the bulk case and is a characteristic feature of the AFM phase with the out-of-plane magnetic anisotropy. These results confirm that the AFM phase is stabilized in the nanoflakes with x = 0.39, which is more robust than the case of with x = 0.33.
In order to show the stable AFM phase in the nanoflakes for x = 0.39, we revised Fig. 5 and added a paragraph for describing new results in the main text. Also a new figure  (Supplementary Fig. S13) and two additional paragraphs are added in the Supplementary information. In order to emphasize the high-TN AFM phase in nanoflakes of vdW antiferromagnets, we included the reported TN of the vdW AFM nanoflakes the Supplementary  Table 1.

Q1-5. The R(T) curves can have many different explanations. I understand the authors did
calculations to show the AFM state. However, it is quite often that the calculations results are different from experimental results. From the magnetic and electric transport results, we cannot rule out the material is still in FM state (a FM state with strong in-plane anisotropy). I think that the authors need to do one more experiment, a neutron diffraction on the bulk crystals. If the neutron diffraction confirms the AFM structure, I will fully support the publication of this paper in Nature Communications.

A1-5.
Due to the current difficult situation, related to the COVID pandemic, we were not able to get an access to the facilities for neutron diffraction experiments. Instead, we carried out resonant soft X-ray scattering experiments on the bulk crystal with x = 0.33, which can also offer a direct probe for the superstructure due to the AFM transition. We observed a clear peak with q = (0 0 3/2), consistent with the calculated interlayer-AFM phase ( Supplementary Fig. S7) as explained in A1-1. This provides a direct evidence of the AFM phase in (Fe1-xCox)4GeTe2 crystals. Further study using neutron diffraction is desirable to determine the size or orientation of magnetic moments of the observed AFM phase, which, we think, is beyond the scope of this work.
In the revised manuscript, new results on resonant soft X-ray scattering are presented ( Supplementary Fig. S7). We sincerely appreciate the reviewer's helpful comments and suggestions. We did our best to answer to each single question and we hope our answers satisfy the reviewer.
Q2-1. The text does not always refer to the SI when necessary. For many statements in the I had to search for a while to find the data that corroborate that statement. This made reviewing this paper very challenging. I recommend that the authors read their manuscript carefully and make sure they refer the reader to the relevant figure whenever it is described. On a similar note, figure captions can be improved. For example, nowhere in Fig. S2 is the applied field giving, In the main text I found that it is supposedly 1kOe, but in this sentence there is no reference to Figure S2. This is very confusing and hard to follow.
A2-1. We really appreciate the reviewer's careful reading and useful suggestions. Following the reviewer's suggestion, we carefully revised the manuscript and clearly stated the location of the relevant information (Figures and Supplementary Notes). In addition, we improved the figure captions so that sufficient information is provided to the readers. Fig 2 should be modified. Fig. 2a does not really show the reader that the phase diagram in Figure 2g is correct. This is better seen with the data from Fig. S2, which I think might be better placed to the main text. In Fig 2a, it appears as if there is a smoot change in magnetization with doping and one cannot see any clean transition from FM to AFM. It took me a while to realize that the magnetic phase diagram is correct. This is also because of the lack of reference to the SI when starting to describing the trends as discussed above. I think in the current stage, the reader will need a long time to understand the results.

A2-2.
As the reviewer suggested, we rearrange graphs in Fig. 2 and Fig. S2 of the main text and the Supplementary information. In the revised manuscript, all the magnetization and susceptibility data as a function of temperature and magnetic field for (Fe1-xCox)4GeTe2 crystals are presented in Fig. 2, in order to emphasize the evolution of the magnetic phase with Co doping. Additional data with different measurement procedures, zero-field-cooling (ZFC) and field-cooling (FC) are also included in Fig. 2 of the revised manuscript. In Fig. 3, we compare the resulting doping-dependent phase diagram with the theoretical calculations. With this arrangement in the revised manuscript, we believe that the readers can understand the results more easily and clearly.

Q2-3.
For the higher doped samples, it might be nice to take M vs T curves at more applied fields to better understand the appearance of the unknown magnetic phase.

A2-3.
As the reviewer suggested, we performed the temperature dependent magnetization measurements at different magnetic fields (H = 100 Oe and 1kOe) below the spin-flop transition field, as shown in Figs. S2 and S3 of the revised Supplementary information. The low-temperature upturn below T0 = 75 K is found to be robust, and a small bifurcation between (T) curves with ZFC and FC was observed. This confirms the low-temperature magnetic structure is distinct from the high-temperature interlayer AFM phase and indicates that the competing FM interaction may play a role. In order to identify the low-temperature magnetic structure of the crystal with x = 0.39, the detailed study is needed, which however is beyond the scope of this work.
Following the reviewer #2's suggestion, we added new results of magnetic susceptibility data taken different applied magnetic fields 100 Oe and 1kOe in Supplementary Figs. S2 and S3.

Q2-4. The authors motivate their study with 2D magnets and say there is no 2D metallic AFM with a high
Neel temperature yet. That is true. However, the authors also did not introduce a 2D one, since their result suggest that the samples become FM when thin. I think they should be more honest about that fact in the abstract and intro, by inserting the word "bulk" before they discuss their results of a metallic vdW AFM.

A2-4.
As we discussed in the previous manuscript, the AFM phase of (Fe1-xCox)4GeTe2 nanoflakes with x = 0.33 is rapidly suppressed with reducing thickness and is eventually changed to be ferromagnetic for thickness of 7L. This is because the system with x = 0.33 is located near to the boundary between FM and AFM phases in the doping-dependent phase diagram, as shown in Fig. 3a of the revised manuscript. Thus we expect that the system with higher Co doping, located deep inside the AFM phase of the phase diagram, would be more stable than the case of x = 0.33. In order to confirm this idea, we performed additional measurement on the crystal with x = 0.39 and present its thickness-dependence in Fig. 5 of the revised manuscript. As expected, we found that the AFM phase is stable down to ~ 9 L sample with a slightly reduced TN by ~ 7 % than the bulk value. These additional results indicate that the AFM phase can be stabilized in (Fe1-xCox)4GeTe2 nanoflakes with higher Co doping. Following the reviewer suggestion, we clearly distinguish the properties of bulk and nanoflakes in abstracts and introduction in order to be more precise on our claim in the revised manuscript.

Q2-5.
Can the authors comment on the upturn in resistivity at lower T? Why is this? Why is it opposite to the trend in FM samples? I think the authors should discuss the role of disorder in their samples and how this might affect the properties. This might also help for the discussion in respect to point 5.

A2-5.
We really appreciate the reviewer for the valuable comments. As the reviewer pointed out, all the crystals exhibit a metallic behavior at high temperatures, whereas they show the upturn of the resistivity at low temperatures. We found that the resistivity at low temperature exhibits the -ln T dependence, followed by deviation at lower temperatures for (Fe 1-xCox)4GeTe2 crystals with x = 0.17, 0.23. 0.26, and 0.33, as presented in the supplementary Fig.  S8. This behavior is a characteristic feature of the Kondo scattering, which has been observed in various AFM thin films with substitutional impurities with magnetic moments [for example, D. Khadka et al. Sci. Adv. 6, eabc1977 (2020]. Recently a metastable phase of Fe5-xGeTe2 has been reported, which contains excess of Fe atoms at the interstitial sites above or below the Ge atoms in the unit cell [May, A. F. et al. ACS Nano, 13, 4436-4442 (2019), Stahl, J. et al. Z. Anorg. Allg. Chem. 644, 1923-1929(2018]. Considering a smaller size of Co than Fe, a small amount of Co atoms can occupy these interstitial sites in our crystals, which behave as magnetic impurities. These magnetic impurities are coupled the conduction band through exchange interaction, resulting in the Kondo scattering. The excess of the resistivity (T) = (T) -min(TK) for (Fe1-xCox)4GeTe2 crystals are nicely scaled as a function of the normalized temperature by the characteristic temperature TK with the resistivity minimum, which is consistent with the prediction of Kondo scattering model as shown in Supplementary Fig. S8.
As the reviewer #2 suggested, we added Fig. S8 and a paragraph in the Supplementary information in order to present detailed analysis on the resistivity upturn at low temperatures,

Q2-6. For the calculations, the authors say they used VCA to model the doping, but they do not say how the lattice constants were obtained. Did they use experimental ones? And if yes in accordance with change in doping? And if not did they optimize the structures? And if yes, did they correct for vdW forces? A few more details here would be helpful.
A2-6. For the calculation, we used experimental lattice constants as mentioned in Methods section briefly, and they are obtained from the X-ray diffraction results shown in Supplementary Fig. S1c. It can be found that lattice constants change linearly with doping, and thus in the case of higher Co doping than experiments, we set the lattice constants by linear extrapolation. Given lattice constants, we did internal atomic position relaxation for each doping. During internal relaxation, vdW correction is not considered, because internal relaxation is hardly affected by vdW correction term for the thick slab of Fe4GeTe2 structure.
In the Methods section of the revised manuscript, we included additional information for calculations, as mentioned above.

Q3-1.
Authors have synthesized single crystals, apparently. Therefore, I am puzzled why anisotropy in magnetic properties has not been presented? A3-1. We presented the magnetization curves at different magnetic fields H // c and H // ab in Fig. 2. As emphasized in the manuscript, depending on Co doping, (Fe1-xCox)4GeTe2 exhibits four different magnetic phases, FM with the out-of-plane magnetic anisotropy (x = 0), FM with in-plane anisotropy (0.07 ≤ x < 0.26), AFM with in-plane magnetic anisotropy (x = 0.3 or 0.33), AFM with perpendicular magnetic anisotropy (x = 0.39). Particularly the magnetic anisotropy of the AFM phases, for example, the in-plane (x = 0.33) and the out-of-plane (x = 0.39) anisotropy remain the same in nanoflakes as shown in Fig. 5 of the revised manuscript. These results demonstrate that (Fe1-xCox)4GeTe2 can provide vdW magnetic layers with various spin configuration and orientations, which may be suitable for vdW-material-based spintronic devices . Fig. S1a: crystals have small impurity phase, between the strongest peak around 2theta=27 and (221) there is another reflection in all crystals. Then,x=0.39,0.3,0.17 also contain another visible peak between (221)  rest on bulk magnetic characterization it is of interest to index those peaks and, if due to impurity phases, ensure that bulk magnetic properties are not influenced by impurities.

A3-2.
We appreciate the reviewer for careful reading on our manuscript. All the peaks, that the reviewer 3 mentioned, are well indexed by the single phase of (Fe1-xCox)4GeTe2. We added the corresponding Bragg indices in Supplementary Fig. S1a of the revised manuscript. Fig.  2a in the main text as well as in Fig. S3(a-e). There seems to be a misprint in Fig. 2a, 0.7 is in fact 0.07.

A3-3.
We appreciate the careful reading of our manuscript. Following the reviewer suggestion, we corrected the typos and also present the saturation magnetization in terms of B/f.u. in Fig. 2 and Fig. S6 of the revised manuscript.

A3-4.
As the reviewer #3 pointed out, the transition to the spin cluster or spin-glass phase can produce a magnetic susceptibility peak, which can be misinterpreted as a signature of the AFM phase transition. In order to rule out the possibility of the spin-cluster state, we measured the temperature dependent magnetic susceptibility, (T), at different cooling procedures, zero-field cooling (ZFC) and field cooling (FC). It has been well known that for the spin-cluster state, the peak of (T) is accompanied by a pronounced bifurcation between ZFC and FC curves. In the (Fe1-xCox) 4GeTe2 single crystal with x = 0.33, however, we observed negligible difference between the ZFC-and FC-(T) curves for both H // c and H // ab as shown in Fig. 2 of the revised manuscript. These results indicate that the peak in (T) for the crystal with x = 0.33 reflects the transition, not to the spin-cluster phase, but to the antiferromagnetic (AFM) phase.
Consistently we also observed no signature of the relaxation behavior, one of the hallmark of the spin cluster state. In the spin cluster phase the change of magnetization requires a sizable time, leading to the time-dependent magnetization phenomena. For example, in the ferromagnetic phase in (Fe1-xCox)4GeTe2 crystals with low x, we found a relaxation behavior with a characteristic time scale of  ~ 600 sec ( Supplementary Fig. S4), which is comparable with the spin cluster phases in, e.g. PrRhSn3 and Sr2Mn0.7Fe0.3MoO6 [Anand, V. K. et al. Phys. Rev. B 85, 014418 (2012), Wang, X. et al. J. Appl. Phys. 109, 07C322 (2011)]. For x = 0.33, however, we found no signature of relaxation behavior as shown in Supplementary Fig. S4 in the revised manuscript. Finally, in order to present a more direct experimental evidence of the AFM ordering, we performed resonant soft X-ray scattering experiments for the crystal with x = 0.33, as shown in the supplementary Fig S7. Similar to neutron diffraction, resonant X-ray scattering also offer a probe for AFM phase. In addition to the Bragg peaks at (003), a clear additional peak develops below TN at q = (0,0,3/2), which is absent above TN. The corresponding periodicity is ~ 19.5 Å along the c-axis, in excellent agreement with the interlayer AFM (A-type) structure, predicted by our first principle calculations. Although further studies, using neutron diffraction, are In addition to Bragg peak at q = (0 0 3), additional peak at q = (0,0,3/2) is developed below TN. b, Crystal structure of (Fe0.67Co0.33)4GeTe2 with the interlayer AFM structure, indicated by red and blue shades. The period of 19.5 Å, corresponding to q = (0,0,3/2), is consistent with the interlayer AFM phase. helpful to determine the size and orientation of magnetic moments in the AFM phase of (Fe 1-xCox)GeTe2, our resonant X-ray scattering results provide a direct evidence of the AFM phase in (Fe1-xCox)4GeTe2 with x = 0.33.
We added several sentences and additional Figures (Fig. 2) in the main text and Supplementary  Information (Figs. S4, and S7), in order to provide experimental evidence of the long-range AFM phase. Fig. 1b rather interesting…I wonder what is the Ioffe-Regel limit for AFM crystals? Note the absence of metallic resistivity for high x at low temperatures. Throughout the text authors refer to investigated crystals as "metals". Yet, rho(T) looks like a very bad metal for high x. I would advise the use of more precise language.

A3-5.
We agree with the reviewer about the definition of "metallicity". A rigorous distinction of metals from insulators is based on the finite resistivity extrapolated at zero temperature. Thus, in principle, metallicity can be defined, even though the resistivity increases with lowering temperatures, which is one of the signatures of 'bad metal'. As the reviewer #3 suggested, we estimated the mean free path of our crystal with x = 0.33. Using the carrier density n = ~ 6  10 27 m -3 , extracted from the contribution of the normal Hall effect and the measured resistivity we obtained the mean free path of ~ 2 nm, which is rather close to the in-plane lattice constant a ~ 0.4 nm. Also from the quantum conductance of the 2D layers and the interlayer distance of ~ 1 nm in (Fe1-xCox)4GeTe2 crystals, we obtained a critical resistivity  ~ 200 cm, which is comparable with the measured resistivity. These results show that (Fe1-xCox)4GeTe2 crystals are in the bad metal regime.
In the revised manuscript, we clearly mentioned that our system is in the bad metal regime. Fig S5 authors report change in DOS with Co doping. Therefore, it is reasonable to assume some change in electronic structure. Does that have any effect on electrical detection of the spin state. For example, could changes in AMR be influenced by electronic structure changes, in addition to Neel vector changes? A3-6. As the reviewer #3 pointed out, there is electronic structure change due to Co doping, which is expected to affect the magnetotransport properties such as the anisotropic magnetoresistance (AMR). Although the AMR itself is defined by the relative orientation between Neel vector and the current direction, its magnitude is material specific and thus can be significantly modulated by Co doping. For the crystal with x = 0.33 the maximum AMR is found to be ~ 0.3 %, comparable with those of other AFM metals. We believe that the size of the AMR can be further optimized by control of Co doping level. Considering the narrow doping window for the in-plane AFM phase in the phase diagram shown in Fig. 3a, however, the detailed study requires precise control of Co doping level, which remains as a future work. Q3-7. I do not think that this work brings high impact or innovation. I agree that AFM transition is high but the last sentence in the introduction is greatly exaggerated. Promising material in spintronics can not operate at 230 K. Authors also say in the conclusion that (Fe,Co)4GeTe2 will enrich the material candidates and spin functionalities for…spintronics. I do not see how. Thickness-dependent magnetic transition falls down with size reduction therefore it is difficult to think of future devices based on nanofabricated crystals investigated in this work.

A3-7.
Realization of stable room temperature antiferromagnetism in nanoflakes is one of the challenge in the current research field of vdW magnets. We note that most of vdW antiferromagnets so far investigated, show various spintronic functionalities below TN much lower than our system. For example, CrI3, one of the most widely studied vdW magnets, has a TN ~ 30 K, far below room temperature. This contrasts to FM vdW materials, some of which exhibit ferromagnetism nearly at room temperature. Although our (Fe1-xCox)4GeTe2 shows TN ~ 226 K, below room temperature, this is significant increase as compared to the previously studied vdW antiferromagnets. Unlike the previously studied vdW antiferromagnets with intralayer AFM structures, our approach for high-TN vdW antiferromagnetism is to use high-Tc FM vdW metals as a parent material and convert them to antiferromagnet by changing their interlayer interaction. We believe that our approach can be applied to other high-Tc FM vdW materials, which will eventually result in higher-TN antiferromagnetism in the vdW magnets.
We agree with the reviewer that the stability issue of the AFM phase in nanoflakes is even more important than the issue of high-T N in bulk materials. As the reviewer pointed out, the AFM phase of our crystal with x = 0.33 is rapidly suppressed with lowering thickness and eventually changes to be ferromagnetic for thickness of 7 layers. However as explained in the manuscript, this is because the system with x = 0.33 is located close to the boundary between FM and AFM phases in the doping-dependent phase diagram, as shown in Fig. 3a of the revised manuscript. Thus we expect that the system with higher Co doping, located deep inside the AFM phase of the phase diagram, would be more stable than the case of x = 0.33. In order to confirm this idea, we performed new measurements on the crystal with x = 0.39 and presented its thicknessdependence in the revised manuscript. As expected, we found that the AFM phase is stable down to ~ 9 L sample with a slightly reduced TN, by ~7% than the bulk value. To best our knowledge, this is the highest TN in the AFM vdW nanoflakes (Supplementary Table. S1).
We envision that our finding of the stable AFM phase in ultrathin (Fe1-xCox)4GeTe2 nanoflakes leads to more researches on spintronic properties using metallic vdW AFM layers. As demonstrated in one of the FM vdW metal Fe3GeTe2 [Deng, Y. et al. Nature, 563, 94-99 (2018)], electrical tuning of the Neel temperature can be possible using (Fe1-xCox)4GeTe2 nanoflakes. Also heterostructures, consisting of AFM layers of (Fe,Co) 4 GeTe 2 and FM layers of Fe 3 GeTe 2 or Fe4GeTe2, may induce the exchange bias effect, which has been demonstrated by 25-75 layer-thick CrCl3 and Fe3GeTe2 [Zhu, R. et al. Nano Lett. 20, 5030-5035 (2020)]. Thus, we believe that our work will bring attention to AFM vdW nanoflakes as one of the active component in vdW-material-based spintronics.
In order to show the stable AFM phase in the nanoflakes for x = 0.39, we revised Fig. 5 and added a paragraph for describing new results in the main text. Also a new figure  (Supplementary Fig. S13) and two additional paragraphs are added in the Supplementary information. In order to emphasize the high-TN AFM phase in nanoflakes of vdW antiferromagnets, we included the reported TN of the vdW AFM nanoflakes the Supplementary  Table 1.