Enhancement of lower critical field by reducing the thickness of epitaxial and polycrystalline mgb2 thin films

Enhancement of lower critical field by reducing the thickness of epitaxial and polycrystalline MgB2 thin films Enhancement of lower critical field by reducing the thickness of epitaxial and polycrysta[.]

Enhancement of lower critical field by reducing the thickness of epitaxial and

polycrystalline MgB2 thin films

Teng Tan, M A Wolak, Narendra Acharya, Alex Krick, Andrew C Lang, Jennifer Sloppy, Mitra L Taheri, L Civale, Ke Chen, and X X Xi

Citation: APL Mater 3, 041101 (2015); doi: 10.1063/1.4916696

View online: http://dx.doi.org/10.1063/1.4916696

View Table of Contents: http://aip.scitation.org/toc/apm/3/4

Published by the American Institute of Physics

Enhancement of lower critical field by reducing

the thickness of epitaxial and polycrystalline MgB2

thin films

Teng Tan,1, aM A Wolak,1Narendra Acharya,1Alex Krick,1,2

Andrew C Lang,2Jennifer Sloppy,2Mitra L Taheri,2L Civale,3Ke Chen,1

and X X Xi1

1Department of Physics, Temple University, Philadelphia, Pennsylvania 19122, USA

2Department of Materials Science and Engineering, Drexel University, Philadelphia,

Pennsylvania 19104, USA

3Los Alamos National Laboratory, Los Alamos, New Mexico 87545, USA

(Received 11 November 2014; accepted 16 March 2015; published online 1 April 2015)

For potential applications in superconducting RF cavities, we have investigated the properties of polycrystalline MgB2films, including the thickness dependence of the lower critical field Hc1 MgB2thin films were fabricated by hybrid physical-chemical vapor deposition on (0001) SiC substrate either directly (for epitaxial films) or with

a MgO buffer layer (for polycrystalline films) When the film thickness decreased from 300 nm to 100 nm, Hc1at 5 K increased from around 600 Oe to 1880 Oe in epitaxial films and to 1520 Oe in polycrystalline films The result is promising for using MgB2/MgO multilayers to enhance the vortex penetration field C2015 Author(s) All article content, except where otherwise noted, is licensed under a Creative Commons Attribution 3.0 Unported License.[http://dx.doi.org/10.1063/1.4916696]

The acceleration gradient of superconducting radio frequency (SRF) cavities, in which the RF magnetic field is parallel to the superconductor surface, is limited by the thermodynamic critical field Hc of the superconductor.1SRF cavities made of niobium (transition temperature Tc= 9.2 K and Hc∼ 1900 Oe) have been perfected such that their accelerating gradient is close to the mate-rial’s theoretical limit.2,3Magnesium diboride (MgB2) is a promising alternative to Nb for future SRF cavities.4MgB2has a higher Tcof 39 K, a lower residual resistivity ( ρ0) of ∼0.1 µΩ cm,5and higher Hcvalues than Nb (Hc > 3000 Oe have been reported6 , 7) The larger energy gaps and the low residual resistivity, which is higher than8but closer to an optimal value,9promise a smaller surface resistivity, thus a higher quality factor, and the higher Hc could lead to a higher RF breakdown field.10 , 11Meanwhile, the grain boundaries in MgB2do not show substantial weak link behavior,12

making MgB2coating suitable for non-epitaxial substrates, such as the inner wall of SRF cavities Earlier studies showed that the lower critical field Hc1of a type II superconductor (Hc1of Nb

∼ 1700 Oe) also played an important role in limiting the RF breakdown field.13 , 14Recent exper-imental results supported this hypothesis and showed significant vortex dissipation loss occurs at

Hc1.3 Gurevich proposed2 a multilayer coating of alternating insulator and superconductor layers with a thickness less than its London penetration depth λ to impede the vortex penetration into the cavity bulk Such an enhancement of Hc1has been shown in MgO/NbN multilayers,15although the benefit of the multilayer structure is being actively debated.16,17For MgB2, Hc1(0) values from 300

Oe to 1100 Oe have been reported for bulk samples,18–20and Hc1(0) = 6500 Oe has been reported

in 60 nm-thick epitaxial thin films.21 In this work, we investigate whether the enhancement of

Hc1with decreasing film thickness is attainable in both epitaxial and polycrystalline MgB2films Because MgB2coatings on cavity walls and the top MgB2layers in S-I-S multilayer structures are usually polycrystalline,22,23an investigation on polycrystalline MgB2films is necessary to evaluate

a Author to whom correspondence should be addressed Electronic mail: phys.tan@temple.edu

041101-2 Tan et al. APL Mater 3, 041101 (2015)

the suitability of MgB2 for SRF cavity applications Our results show that the Hc1 of MgB2thin films is enhanced with decreasing thickness in both epitaxial and polycrystalline films with Hc1 (5 K)= 1880 Oe for a 100 nm epitaxial film and Hc1(5 K)= 1520 Oe for a 100 nm polycrystalline film

The MgB2 films were grown by hybrid physical-chemical vapor deposition (HPCVD)24 at

a substrate temperature of 730◦C, hydrogen carrier gas pressure of 40 Torr, hydrogen flow rate

of 400 sccm, and diborane mixture (5% B2H6 in H2) flow rate of 20 sccm The corresponding deposition rate was 55 nm/min The thickness of the MgB2films ranged from 100 nm to 300 nm and was controlled by the deposition time Epitaxial MgB2 films were deposited directly on SiC (0001) substrate For polycrystalline MgB2films, a MgO layer was deposited on the same substrate prior to depositing MgB2 by reactive DC magnetron sputtering from a 2 in.-diameter Mg metal target at room temperature in an Ar/O2(25:8) mixture of 4.0 mTorr The sputtering voltage was

160 V and the sputter power was 50 W, yielding a deposition rate of 0.7 nm/min After the MgB2 deposition, the sides of the substrate were cleaned with 10% hydrogen chloride acid to remove the MgB2 deposit and prevent magnetization signals from the edge The structure of the MgB2 thin films was characterized by x-ray diffraction (XRD) and cross-sectional transmission electron microscopy (TEM) using a JEOL 2100 LaB6 operated at 200 kV The DC transport properties were measured using the four-probe van der Pauw method The lower critical field was determined

by measuring the magnetic moment m versus the applied magnetic field (H) of zero-field-cooled samples using either an AC measurement system (ACMS),25a SQUID magnetometer, or a vibrat-ing sample magnetometer (VSM) manufactured by Quantum Design In the experiment, field was applied to both sides of the sample and parallel to the film’s surface Thin films were attached to

a vertical sample rod with a specification of less than 0.5◦off axis parallel to the field To confirm the sample alignment, we have measured some samples using a rotating sample stage, where the misalignment can be determined The results from the rotating sample stage with misalignment

∼0.2◦are consistent with those from the fixed sample stage for similar samples

XRD θ-2θ scans for an epitaxial MgB2 film on a (0001) SiC substrate and a polycrystalline MgB2film on a MgO buffer layer are shown in Figs.1(a)and1(b), respectively Both figures show only the (000l) MgB2diffractions besides the substrate and sample stage peaks, indicating that the c

FIG 1 (a) XRD θ-2θ scan of a 300 nm MgB 2 film on SiC substrate (b) XRD θ-2θ scan of a 200 nm MgB 2 film on 15 nm MgO bu ffer layer on SiC substrate (c) TEM image of a MgB 2 /MgO multilayer structure on SiC substrate (d) SAD pattern across the SiC /MgB interface (e) SAD pattern across the MgB /MgO multilayer.

FIG 2 Resistivity versus temperature curves for epitaxial MgB 2 films on SiC substrates and polycrystalline MgB 2 films on

15 nm MgO buffer layer The film thickness ranges from 100 nm to 300 nm.

axis of the film is perpendicular to the substrate surface in both films No diffraction peak associated with the MgO layer was observed, as MgO films deposited at room temperature are polycrystalline with grains too small to observe in an XRD measurement.26 The MgB2 film peaks on MgO are broader and weaker than those from the film on (0001) SiC, a consequence of the polycrystalline nature of the film on MgO While the φ scans of the (10¯11) MgB2peak for epitaxial films show a sixfold symmetry indicative of in-plane alignment with the substrate lattice, no such φ dependence could be observed in the MgB2 film on MgO buffer layer In order to illustrate that the MgB2 films grown on MgO layer are polycrystalline, we performed cross-sectional TEM analysis of a MgB2/MgO S-I-S multilayer on (0001) SiC substrate in which the MgB2and MgO layers were grown alternately under the same conditions as described above As shown by the TEM image in Fig.1(c), except for the bottom MgB2layer which grows epitaxially on the (0001) SiC substrate, the subsequent MgO and MgB2layers are all polycrystalline, consistent with the result of the MgB2 film on MgO buffer layer The crystal structure was confirmed by selected area diffractions (SAD) [Figs.1(d)and1(e)]

The resistivity versus temperature curves for several MgB2films with thicknesses ranging from

100 nm to 300 nm on either (0001) SiC or MgO buffer layer are shown in Fig.2 All the samples showed a zero-resistance Tcabove 38 K The epitaxial films on bare SiC substrates have lower residual resistivity ρ0than films on MgO buffer layers of the same thickness The values are consistent with previous results for epitaxial27and polycrystalline28MgB2thin films The Tcvalues in the polycrys-talline films (around 39 K) are lower than in the epitaxial films (around 41 K) in part because the biaxial tensile strain that raises Tcin epitaxial MgB2films27is absent in the polycrystalline films Figures3(a)and3(b)are schematics of the fixed sample holder and the rotation sample holder used in this work, respectively In both cases, a vertical magnetic field was applied to both sides of the sample and parallel to the film’s surface; however the rotation sample holder can rotate about the vertical axis of the sample rod, marked by angle φ, while the fixed sample holder cannot The horizontal pickup coil measures the longitudinal moment mL, and the vertical pickup coil measures the transverse moment mT Most results in this work were obtained using the fixed sample holder, which has a specification of misalignment angle θ off the field direction less than 0.5◦ Some samples were measured using the rotation sample stage, whose misalignment angle θ can be calculated from the maximum mTwhen φ changes Figure3(c)shows the m−H curves of a zero-field-cooled 300-nm epitaxial MgB2film on SiC measured from 5 K to 30 K using the fixed sample holder When the applied field is below Hc1of a superconductor thin film with a thickness of d, there is no vortex in the film and

041101-4 Tan et al. APL Mater 3, 041101 (2015)

FIG 3 (a) Diagram of fixed sample stage measurement (b) Diagram of rotation sample stage measurement (c) Typical m−H curves measured at di fferent temperatures for a 300 nm epitaxial MgB 2 film on SiC substrate (d) Derivative dm /dH versus H calculated from the curve at T = 5 K The value of H c1 is determined when dm /dH deviates from its original value by more than 10% (e) Rotation test of a 250 nm epitaxial MgB 2 film on SiC substrate under 400 Oe applied field (f) Comparison of m−H curves of 250 nm epitaxial films measured with fixed and rotation sample holders at 5 K Arrows indicate the H c1 of each sample.

4πm=

d/2



−d/2

H · A ·[1 − cosh(

x

λ) cosh(d 2λ)]dx = H · A · d[1 −

2λ

d tanh( d 2λ)], (1)

where λ is the penetration depth and A is the film area.14Thus, a linear dependence of m on H should

be observed below Hc1 Above Hc1, vortices enter the superconductor and m deviates from the linear dependence on H Figure3(d)shows dm/dH vs H for the result of T = 5 K The derivative fluctuates around −4.75 × 10−7(erg/G Oe) at low magnetic field (corresponding to a λ value of 32 nm) and rises significantly above Hc1 We define the field at which the absolute value of dm/dH deviates

by more than 10% from the original value as Hc1, which yields a value Hc1(5 K)= 600 Oe for the

300 nm epitaxial film The magnetic moment signal became weaker when the MgB2film thickness decreased Our measurements on samples thicker than 100 nm were well reproducible, while the results for thinner samples were noisier and the Hc1values were more difficult to determine precisely The sample alignment was confirmed by measuring samples’ transverse (mT) magnetic moments during horizontal rotation Figure3(e)shows mLand mTas a function of the axial angle φ for a 250

nm epitaxial film under 400 Oe applied field The maximum mT∼(V/4π)(L/2d)Hθ allows us to estimate the misalignment, and the largest misalignment in Fig.3(c)(for φ ∼ 300◦) corresponds to

θ ∼ 0.2◦ Figure3(f)shows the m−H curves for two 250 nm epitaxial films, one mounted on the fixed sample holder (open squares) and one on the rotation sample holder (solid triangles) The curves overlap on each other, indicating similar misalignment angles for both sample holders The arrows indicate Hc1of 920 Oe for the fixed sample holder and 1240 Oe for the rotation sample holder The

difference could partly arise from the different film qualities, but may also reflect the different sample alignments

In Fig.4, we show the temperature dependence of Hc1for both the epitaxial and polycrystalline films For the epitaxial films, shown in Fig.4(a), the lowest Hc1was observed in the 300 nm thick film, about 600 Oe at 5 K with a linear Hc1− T dependence These properties are similar to bulk MgB2 samples reported in the literature.19 Hc1increases with decreasing MgB2film thickness, reaching

1880 Oe when the MgB2film thickness is 100 nm The Hc1values of 80 nm samples determined from the m−H measurements (not shown) contain large uncertainties, ranging from 1400 Oe to 4500 Oe

FIG 4 Temperature dependence of H c1 for di fferent film thicknesses for (a) epitaxial and (b) polycrystalline MgB 2 films.

FIG 5 Thickness dependence of H c1 (5 K) of (a) epitaxial and (b) polycrystal MgB 2 films The solid lines are the result of fitting to Eq ( 2 ).

The Hc1− T curves of the polycrystalline MgB2films grown on MgO buffer layer, plotted in Fig.4(b), show the same thickness dependence trend as the epitaxial films However, the Hc1values are slightly lower than the epitaxial films of the same thickness, being 520 Oe at 5 K for the 300 nm film and

1520 Oe for the 100 nm film From these results, we can conclude that the thin polycrystalline MgB2 films in the S-I-S structures should have similar abilities to prevent vortex entry as in clean, epitaxial MgB2thin films The temperature dependence of Hc1is predominately determined by 1/λ2(T),14and

a decreasing λ when the temperature decreases leads to an increasing Hc1 The two-band nature of MgB2directly impacts the shape of the 1/λ2

(T) curve, depending on the relative contributions from the π and σ bands and the strength of scattering.29The largely linear temperature dependence in most films indicates predominant π band contribution to λ while the curvature for the thinnest epitaxial films suggests enhanced influence from the σ band, presumably due to the reduced c-axis dimension

In Fig.5, we plot Hc1(5 K) as a function of MgB2film thickness from 100 nm to 300 nm for (a) epitaxial and (b) polycrystalline MgB2films In both cases, Hc1increases with decreasing film thickness to values similar to those of bulk Nb The solid lines in Fig.5is the fitting to the theoretical thickness dependence of Hc1applicable to film thickness greater than λ, based on the free energy of

a single vortex in a thin film,30 , 31

Hc1

Hc1b = {1 +

∞

 0

(tanhk2+ (d/λ)2− 1)dk (ln κ + 0.5)k2+ (d/λ)2 }/(1 − sechd

where κ ≡ λ/ξ is the Ginzburg-Landau parameter (ξ is the coherence length), k is the integrating variable, and Hc1bis the lower critical field for bulk materials at this temperature The Hc/Hc1bratio depends exclusively on the ratio of the film thickness to the penetration depth d/λ Taking both bulk

Hc1band λ as fitting parameters, very good fittings are obtained with Hc1b= 610 Oe and λ = 51 nm for the epitaxial films and Hc1b= 514 Oe and λ = 54 nm for the polycrystalline films The Hc1bobtained are within the range of values reported in the literature.17–19The λ values are slightly higher than the previous measurements on our clean film samples.32

041101-6 Tan et al. APL Mater 3, 041101 (2015)

The lower critical field of thin superconducting films is difficult to measure First, aligning the film surface parallel to the magnetic field is critical as vortices penetrate the sample much more easily in the perpendicular direction (Hc1//∼600 Oe and Hc1⊥∼10 Oe at 5 K for a 300 nm epitaxial MgB2film) Second, the diamagnetic signal is small in thin superconducting films A 100 nm film with λ= 40 nm will have a dm/dH slope ∼0.1 (µerg/G Oe), approaching the limit of commercial magnetometers thus increasing the uncertainty of Hc1determination Third, in our pure MgB2thin films, vortices penetrate the sample gradually above Hc1rather than avalanching into it.33This slow penetration significantly reduces the sensitivity of the AC susceptibility method34and other methods that depend on the m−H hysteresis.35Therefore, the magnetization method was chosen as the technique for this study despite various shortcomings and uncertainties in the absolute values of Hc1 We believe that the behavior

of the film thickness dependence of Hc1, which is of critical importance for SRF applications, was accurately revealed by our results

In conclusion, the lower critical field increases as the film thickness decreases for both the epitax-ial and polycrystalline MgB2 films Even though the Hc of bulk MgB2, around 500–600 Oe at

5 K determined from the thickness dependence of Hc of our films, is lower than Nb, it increases to values near that of Nb at the film thickness of 100 nm and is even higher in thinner films The result demonstrates that the approach of preventing vortex entry, thus enhancing the acceleration gradient,

by thin superconducting films can be realized using polycrystalline MgB2coatings on SRF cavities and in S-I-S multilayer structures

This work was supported by the U.S Department of Energy, Office of Science, High Energy Physics, under Award No DE-SC0011616 The authors are grateful to Dr Tan Yuen and Dr S J May for the assistance in the measurements The authors also would like to thanks Dr Alex Gurevich for helpful discussion

1 H S Padamsee, IEEE Trans Appl Supercond 15, 2432 (2005).

2 A Gurevich, Appl Phys Lett 88, 012511 (2006).

3 G Ciovati, in Proceedings of the IPAC2013, Shanghai, China, 16 May 2013, p.THYB201.

4 J Nagamatsu, N Nakagawa, T Muranaka, Y Zenitani, and J Akimitsu, Nature 410, 63 (2001).

5 X X Xi, Supercond Sci Technol 22, 043001 (2009).

6 M Zehetmayer, M Eisterer, J Jun, S M Kazakov, J Karpinski, A Wisniewski, and H W Weber, Phys Rev B 66, 052505 (2002).

7 F Bouquet, R A Fisher, N E Phillips, D G Hinks, and J D Jorgensen, Phys Rev Lett 87, 047001 (2001).

8 B B Jin, T Dahm, C Iniotakis, A I Gubin, E M Choi, H J Kim, S I Lee, W N Kang, S F Wang, Y L Zhou, A V Pogrebnyakov, J M Redwing, X X Xi, and N Klein, Supercond Sci Technol 18, L1 (2005).

9 J P Turneaure, J Halbritter, and H A Schwettman, J Supercond 4, 341 (1991).

10 H Padamsee, J Knobloch, and T Hays, RF Superconductivity for Particle Accelerators (Wiley, New York, 1998).

11 E W Collings, M D Sumption, and T Tajima, Supercond Sci Technol 17, S595 (2004).

12 D C Larbalestier, L D Cooley, M O Rikel, A A Polyanskii, J Jiang, S Patnaik, X Y Cai, D M Feldmann, A Gurevich,

A A Squitieri, M T Naus, C B Eom, E E Hellstrom, R J Cava, K A Regan, N Rogado, M A Hayward, T He, J S Slusky, P Khalifah, K Inumaru, and M Haas, Nature 410, 186 (2001).

13 B Piosczyk, P Kneisel, O Stolz, and J Halbritter, IEEE Trans Nucl Sci 20, 108 (1973).

14 M Tinkham, Introduction to Superconductivity, 2nd ed (McGraw-Hill, New York, 1996).

15 C Z Antoine, S Berry, S Bouat, J F Jacquot, J C Villegier, G Lamura, and A Gurevich, Phys Rev Spec Top.–Accel Beams 13, 121001 (2010).

16 S Posen, M U Liepe, G Catelani, J P Sethna, and M K Transtrum, “Response to comment on theoretical RF field limits of multilayer coating structures of superconducting resonator cavities,” eprint arXiv:1310.4479 (2013).

17 A Gurevich, AIP Adv 5, 017112 (2015).

18 L Lyard, T Klein, J Marcus, R Brusetti, C Marcenat, M Konczykowski, V Mosser, K Kim, B Kang, H Lee, and S Lee,

Phys Rev B 70, 180504 (2004).

19 S L Li, H H Wen, Z W Zhao, Y M Ni, Z A Ren, G C Che, H P Yang, Z Y Liu, and Z X Zhao, Phys Rev B 64,

094522 (2001).

20 Y Feng, G Yan, Y Zhao, A K Pradhan, C F Liu, P X Zhang, and L Zhou, J Phys.: Condens Matter 15, 6395 (2003).

21 D B Beringer, C Clavero, T Tan, X X Xi, W M Roach, and R A Lukaszew, IEEE Trans Appl Supercond 23, 7500604 (2013).

22 C Zhuang, T Tan, A Krick, Q Lei, K Chen, and X X Xi, J Supercond Novel Magn 26, 1563 (2013).

23 M A Wolak, T Tan, A Krick, E Johnson, M Hambe, K Chen, and X X Xi, Phys Rev STAB 17, 012001 (2014).

24 X H Zeng, A V Pogrebnyakov, A Kotcharov, J E Jones, X X Xi, E M Lysczek, J M Redwing, S Y Xu, J Lettieri,

D G Schlom, W Tian, X Q Pan, and Z K Liu, Nat Mater 1, 35 (2002).

25 ACMS is a magnetometer manufactured by Quantum Design which can conduct magnetic momentum measurements using extraction method, see http: //www.qdusa.com/sitedocs/productBrochures/1070-002.pdf

26 Y Matsuda, Y Koyama, K Tashiro, and H Fujiyama, Thin Solid Films 435, 154 (2003).

27 A V Pogrebnyakov, J M Redwing, S Raghavan, V Vaithyanathan, D G Schlom, S Y Xu, Q Li, D A Tenne, A Souki-assian, X X Xi, M D Johannes, D Kasinathan, W E Pickett, J S Wu, and J C H Spence, Phys Rev Lett 93, 147006 (2004).

28 C Zhuang, T Tan, Y Wang, S Bai, X Ma, H Yang, G Zhang, Y He, H Wen, X X Xi, Q Feng, and Z Gan, Supercond Sci Tech 22, 025002 (2009).

29 A A Golubov, A Brinkman, O V Dolgov, J Kortus, and O Jepsen, Phys Rev B 66, 054524 (2002).

30 G Stejic, A Gurevich, E Kadyrov, D Christen, R Joynt, and D C Larbalestier, Phys Rev B 49, 1274 (1994).

31 A A Abrikosov, Sov Phys JETP 19, 988 (1964), available at http://www.jetp.ac.ru/cgi-bin/e/index/e/19/4/p988?a=list

32 D Cunnane, C Zhuang, K Chen, X X Xi, J Yong, and T R Lemberger, Appl Phys Lett 102, 072603 (2013).

33 Z X Ye, Q Li, Y F Hu, A V Pogrebnyakov, Y Cui, X X Xi, and J M Redwing, Appl Phys Lett 85, 5284 (2004).

34 D Saint-James and P G Gennes, Phys Lett 7, 306 (1963).

35 C Böhmer, G Brandstätter, and H W Weber, Supercond Sci Technol 10, A1 (1997).