Fabrication of NIS and SIS Nanojunctions with Aluminum Electrodes and Studies of Magnetic Field Influence on IV Curves

Round 1

Reviewer 1 Report

In this work, authors claimed the superiority of Al films to fabricate NIS and SIS devices. It is also demonstrated that the Al-based superconducting properties are affected by the magnetic field in detail. And finally, the theoretical and experimental results of IV characteristics are highly accorded. A clear introduction for background knowledge, together with full discussions enable us to believe that this work is well done and very suitable for the scope of Electronics. I therefore recommend this manuscript is accepted in present form without any change.

Author Response

Dear reviewer, thank you very much for prompt reply and positive estimation of our work

Reviewer 2 Report

As attached.

Comments for author File: Comments.pdf

Author Response

Dear reviewer, thank you very much for careful reading of our manuscript.  Concerning questions: 1) we don’t understand quite well what do you mean exactly by IV curve link to EJ, if it is about the energy gap, it is clear visible for all NIS junctions at about V_D=0.19 mV, corresponds to D=1.76kT_c.  For SIS junctions the gap voltage is twice as much, about 0.38mV.  Concerning Jc(B) for Al SIS junctions this subject was presented in our recent publication (M. Tarasov et al., Development of Josephson travelling-wave parametric amplifier based on aluminum SIS junctions, Physics of Solid State 2021, Vol. 63, No 9, p. 1377 DOI: 10.1134/S1063783421090419) with modulation of dc SQUID critical current from 45 to 10 nA and 0.5-0.3 µA for different junction areas.  2) line 201, now line 231, we added: magnetic flux of applied field 4.65mT for the 1.8 µm² area of junction F=8.37 Weber and F₀=2.07 W that makes F/F₀=4.  3) About cryogenics: we added (line 150) a portable dilution cryostat [15] and description of measurement setup (lines 151-154). 4) changed “cells” not “sells”. 5) Specified permanent magnetic field, see title for section 3 (Influence of permanent magnetic field on conductivity of NIS junctions). 6) added references 15, 16,17 and equations 3a and 3b

Reviewer 3 Report

The authors report on the study of the influence of magnetic field on the transport properties on metal-insulator- superconductor/superconductor-insulator-superconductor nanojunctions. The results deviate from those expected for the systems with perfect aluminum layer with the properties of a type-I superconductor.  So, the authors interpret them considering grainy structure of the Al layers, with imperfections and impurities. Presence of those features is linked with formation of Abrikosov vortices, shunting resistance and local overheating. The problem is interesting both from the point of view of the physics of real superconducting structures and for proper assessment of the properties of devices. However, the manuscript should be made much clearer and consistent, prior the publication.

The title suggests that fabrication of the nanojuctions is the main topic of the paper. A considerable part of the abstract is also devoted to the methods of preparation of the structures. However, this subject is described  in the text superficially. As concerns the preparation methods, the reader obtains just the reference to the previous publications (Ref. 8 seems to be unavailable yet). Then, the profiles and roughness of a few Al layers are shown. It has been promised in the abstract: “We compared Al film quality, grain size, surface roughness, resistivity deposited by thermal evaporation and magnetron sputtering. We studied the dependence of the oxide barrier and quality of tunnel junctions on the thickness, morphology and microcrystal structure of films, determined by the deposition technique.” Nothing of that appeared in the text. If it is a description of the previous studies it should not appear in the abstract. So, the title, the abstract and the main text should be made consistent and focused on the main subject of the paper. I have an impression that the manuscript is focused on the modeling the properties of the structures with imperfect Al layers under the influence of the magnetic field and not on fabrication of the structures, as suggested in the title and the abstract. Nevertheless, more detailed description of the investigated structures should be given, apart from the brief description in the page 4 and Fig. 3a. It should be indicated which system exhibited the properties shown in Fig. 4 and 5 – single NIS structures, SINIS structures, NIS thermometers?

Taking into account the collected experimental results and the model calculations the authors claim (in  “Conclusions”) that “Thin granular films suffer from Abrikosov vortices that convert Al from a type-1 to a type-2 superconductor.” This seems to be an overstatement. I agree that the properties of the studied system deviate from the properties of a type-I superconductor. However, the granular Al layer with grain boundaries most probably decorated with defects and impurities differs substantially from a chemically uniform type-II superconductor with coexisting superconducting and normal phases for some intermediate ranges of magnetic field and temperature. The wording used in “Introduction” (“can demonstrate some features of superconductor type-2”) seems to be more appropriate.

The units for the magnetic field should be taken consistently from the SI or CGS-Gaussian system of units. And the symbol of tesla is T (not Tl).

In Fig. 5, a detailed description of the difference between “G”, G_single” and “G_single calc.” should be given.  The scale labels are too small in Fig. 1,2, and 3.

Author Response

Dear reviewer, thank you very much for careful reading of our manuscript.  Concerning your recommendations: We made corrections both in title and abstract. The details of fabrication methods were presented in our recent publication, Ref.8, is published and free available. Comparison of studied films is added and now presented in Table 1 and paragraph below, lines 113-121.  Investigated structures now are described in capture and presented in Fig.4a (photo of actual sample).  Conclusion is edited, we added “Penetration of a magnetic field normal to the surface in the form of Abrikosov vortices into granular aluminum films and the retention of superconductivity at fields higher than the critical value for bulk aluminum indicate that the films are type II superconductors “.  Units are now consistent, magnetic field in mT. In Fig.5 and capture we clarified difference between curves. Mentioned that curves with label “single” correspond to single-electron tunneling – experimentally measured current with subtracted Andreev current. Label “full” corresponds to experimentally measured current.  Scale labels in Fig.1-3 all enlarged.  

Reviewer 4 Report

Please, read the attached pdf file

Comments for author File: Comments.pdf

Author Response

Dear reviewer, thank you very much for careful reading of our manuscript and detailed comments.  Actually many changes according to recommendations of reviewers 1-3 (similar to your requests) are already made in a revised version from November 9.  Please find below our answers corresponding to the latest version of the manuscript from November 16.  

TITLE, ABSTRACT, INTRODUCTION

The title refers to NIS and SIS junctions, as well as the abstract, but in the manuscript text only NIS and SINIS samples have been deeply analyzed.

Title, abstract, introduction already changed in previous version.  Concerning SIS junctions: we fabricated Al SIS junctions for dc SQUID amplifier with the same layout as in Fig.3a and published results in [M. Tarasov, et al., Development of Josephson travelling-wave parametric amplifier based on aluminum SIS junctions, Phys. of Sol. State, 2021, Vol. 63, No.9, pp.1377-1381. DOI: 10.1134/S1063783421090419].  This reference is added as [18].

SECTION 2:

The author state they used three different technologies for the fabrication. The details about the three different techs are reported in Ref. 8, but this is not sufficient to understand which techniques are used to fabricate the samples described in detail in the section. Indeed, it seems that the morphological analysis has been carried out only on two of the expect three types of samples.

Actually AFM analysis was done for all types of films and we did not present in the article similar pictures.  Numerical results now are presented in Table 1 for all 3 methods of fabrication.

Moreover, data from electrical transport properties characterization are reported for only the samples realized by magnetron sputtering.

Actually results in Fig.4 and Fig.5 are for samples fabricated by e-beam evaporation.  Magnetron sputtering was used for sample in Fig.3.

Thus, a real comparison of “the dependence of the oxide barrier and quality of tunnel junctions on the thickness, morphology and microcrystal structure of films, determined by the deposition technique.” is lacking.

The main conclusion is that regardless the fabrication method for films deposited at room temperature, the best quality of junctions without Abrikosov vortices is obtained for film thicknesses 150 nm and above, as it is stated in conclusion. 

SECTION 3:

First of all, it is not clear which is the link between the sample described in section 2 and the ones analysed in this section.

Samples in Fig.3a,b fabricated by magnetron sputtering, and in Fig.4,5,6 by e-beam evaporation.

Then, which is the “nonlinearity of up to ~30mT..” that was observed?

In figs 5 and 6 we present dynamic conductivity dependencies at 28 mT which are clear nonlinear, and some nonlinearity can still be observed up to 30 mT.  We added (lines 192-195): However, for a higher magnetic field of 28 mT, the conductivity is still slightly below unity, nonlinearity and superconductivity are still present, which means that although the critical field is not achieved but is very close to it. Therefore, to estimate the correlation length with an accuracy of ~10%, we can use the value of 30 mT as a critical field B_C2.    

Moreover, how did the author estimate the upper critical field Bc2, whose value is not reported, and which leads to the estimation of xi = 110 nm and lambda = 300 nm on page 7?

See now clarification in lines 207-209.  These rough estimations are made just to demonstrate that film is a type-2 superconductor and the higher is critical field – the lower is x and higher is l.   The exact value of B_C2 is not fundamental, it can vary from film to film. 

Still, which is the so called “region of the tunnel junction”? Without a clear definition of this region, the reader cannot evaluate which is the junction area and thus which is the fraction of this area that is occupied by vortices. Moreover, if it is true that “Even in the field of 23.3 Gs, if the junction contains only one vortex…”, thus we are dealing with vortices inside a junction, which are probably not Abrikosov vortices, but Josephson vortices. These second type of vortices have a far more complicated core and simple model of a NIN structure is probably wrong. 

To clarify the layout of single NIS junction we added Fig.4b in which separate top blue box (1) of the area 1 1x1.8 µm² is exactly junction area with Al film. The area of superconductor is much larger, but tunneling conductance is determined only by vortices in the area of tunnel junction.

Concerning nature of vortices: in our case it is NIS junction, and Josephson vortices are expected to present in SNS junctions.  In general, the Josephson vortex is a quantum vortex of supercurrent in Josephson junction. Supercurrent circulate around the vortex center which is situated inside the Josephson barrier, unlike Abrikosov vortices in the Type-2 superconductors, which are located in the superconducting condensate.  Abrikosov vortices are characterized by normal cores where superconducting condensate is destroyed on a scale of superconducting length.   

In general, even dealing with Abrikosov vortices, the author should explain why they refer to a NIN model. Indeed, the simplest description of an Abrikosov vortex is a cylindrical normal core hosting a tube of flux with a radius of xi surrounded by the superconducting medium where screening supercurrents circulate around the core to a distance up to lambda.  

Finally, if we are dealing with possible effects from the presence of Abrikosov vortices in the superconducting electrode(s), what about the dissipative effects associated with vortex motion? Indeed, Al is known as a very weak pinning material, with possible very low irreversibility field values.

True, dissipation due to vortex motion is another drawback of Abrikosov vortices in Al films, we did not study in detail such dissipation, but if in thick films we avoid such vortices, such dissipation should not affect NIS and SIS devices. For vortex motion tunneling currents are too low, flux pinning according to our article [2] is clear visible, hysteresis is about 5-10 mT, so we do not expect vortex motion in our experiment in ordinary low magnetic field.  For pinning/depinning, change in number of vortex by one should lead to current change by tens of percent, that was not observed.

SECTION 4

Where is Section 4?

Sorry, we changed number

SECTION 5

More than a discussion, it looks like a repetition of concepts which have been already introduced and discussed in the previous sections.

Now it is 4.  In this section we present our point of view on prospects for improvement of Al-based tunnel junctions, and contrary to present deposition at room-temperature substrate covered with SiO2 or Al2O3 amorphous insulator, two recipes are mentioned: epitaxial growth on hot single-crystal substrate with proper lattice parameter, or amorphous small-grain film deposition on cold substrate.  Both methods can reduce surface roughness and improve junction parameters

SECTION 5 (again)

They are reported as conclusion statements which does not find any reference in the text. For example, where is shown in the manuscript that “Improvement can be achieved by reducing the surface roughness and, at the same time, by increasing film thickness to over 200 nm. Such films can be fabricated by cross-type lithography with thermal evaporation, or magnetron sputtering with double patterning. Grain size can be reduced by evaporation on a cold substrate, obtained even by a Peltier cooler, down to -27^OC, or better, down to 77 K by liquid nitrogen.”?

Together with the previous flaws, that should be addressed in order to allow the reader to evaluate the quality of the presented work, there are some other improvements which can help the readability of the manuscript.

In conclusions (rewritten) we present a brief summary of practical consequences of our research.  First statement is that magnetic field can affect superconductivity of Al films and devices quality.  First simple solution is to increase thickness over 150 nm.  The second is to reduce surface roughness.  Third statement is that conversion to superconductor type-2 is manifested by appearance of Abrikosov vortices in magnetic field that can be detected by shunting resistance and overheating.  If mentioned film parameters cannot be achieved, it is necessary to provide magnetic shielding down to the field below 1 mT. 

There is not an homogeneous use of units in the text. The applied magnetic field could be expressed in gauss or tesla, even if the SI unit is tesla, but the symbols used for the units should be the same in the whole text. Thus, the author should decide if use Gs or G, unless in the first case they are referring to gauss per second. Or they should use T for tesla, since Tl means tesla per litre.

Sure, we already changed units from Gs to T all over the manuscript. 

On row 67, the sentence “In order to avoid magnetic flux penetration, the first obvious decision is increasing film thickness to over ξ0 or even λ, that is, 15-300 nm.” is not clear. Are the given 3 numbers related to film thickness of to the coherence length and penetration depth? In the first case, an Al film with a thickness lower than 100 nm could be easily a type-II superconductor.

Below 150 nm conversion from Type-1 to Type-2 can be related to intergranular barriers and vary with the grain size, so this estimation is approximate, just to show the scale order.  To our experience in our films thickness over 150 nm is always a guarantee that we have Type-1 superconductor.

3. In Figure 3b, it is not clear which data are related to the current axis and which ones to the resistivity axis.

Now marked with arrows.

4. I would prefer to have the sample/junction dimension explicitly given and not to infer them from an undeclared obscure mathematical formula as for example in rows 135, 137 or 253. By the way, some pictures/sketches of the junctions geometry would surely help to understand how the devices are.

Added: Fig.4a photo, 4b schematic view, and in capture: junction size 1x1.8 µm².

5. Talking about schemes, I doubt that the scheme in Figure 7a refer to a NIN structure, as stated in the text at row 209.

In 7a vortex core is marked as V in the center of top Al layer marked as S, see also capture. 

6. The author should check the coherence between main text and captions. For example, at row 90 tit is reported: “The top layer of superconducting aluminum should be over 300 nm (Figure 3a).”, but in the caption of the mentioned figure it is “Al NIS junction with 150 nm-thick aluminum electrode”.

Thanks for mentioning, changed.

Author Response File: Author Response.pdf

Round 2

Reviewer 3 Report

I accept the answers of the authors to my comments and the amendments made in the manuscript. So, I consider the manuscript as suitable for publication in Electronics, in its present form.

Author Response

Dear reviewer, thank you very much for reading manuscript and comments

Reviewer 4 Report

Please, see the attachment.

Comments for author File: Comments.pdf

Author Response

Dear reviewer, thank you very much for valuable comments, see our changes^

First of all, since the title is “Fabrication of NIS and SIS nanojunctions with aluminum electrodes and studies of magnetic field influence on IV curves” and in the abstract is written “Samples of superconductor–insulator–superconductor (SIS) and normal metal–insulator– superconductor (NIS) junctions with superconducting aluminum of different thickness were fabricated and experimentally studied…” I would expect to find in the manuscript data and details about SIS samples, not just a reference to a previous work.

As it is mentioned in line 120, we fabricated also SIS junctions with the similar layout [18].  The difference in SIS fabrication is that only Al is deposited as the first layer, contrary to NIS junctions for which the underlayer of 1-2 nm of Fe was deposited before Al. Influence of magnetic field on SIS junction is more complicated, including oscillations of Josephson critical current of SIS junction and penetration of Abrikosov vortices in lower Al layer, as well as in upper Al film.  Contrary to SIS, in NIS junctions we have only one effect of vortices in upper Al film that makes explanation clear and unambiguous. (lines 120-127)

I still find hard to understand what is what in the different formulas which are presented in the text. As for example, at row 231 it is written “The region of the tunnel junction with area 1.8*1 μm2 ”. The presence of the scheme in Fig4b can help, but still is not clear which size is 1.8 μm long which 1 μm long.

We added in the text:

The region of each tunnel junction with area 1.8*1 μm2   (line 237)

 At row 284 “The total volume of a gold film is 14*100*0.1*100*5=7*104 μm3 .” There is a product of 5 numbers, when one expect a product of three lengths.

We changed

The total volume of all 100*5 gold films is

 At row 290 it is written “we have 7e-9=6.5e-9*7e4*(T4-0.14) and Т=0.104 К”,

Sorry, it is our oversight, changed as follows:

7*10^-9 =6.5*10^-9 *7*10⁴ *(T⁴-0.1⁴) and Т=0.104 К

And also one line above, should be     

 at P=7 nW

where do these numbers come from? Is “e” the electron charge or the exponent of the power of 10? In the second case, why this different notation here which can be confused with “h/2e2=12 kΩ” at row 345. The authors should carefully revise the text in order to make clear to what physical quantity the given numbers refer in any of the mentioned cases and in the similar ones.

Hope that now it is clear

Author Response File: Author Response.pdf