Magnetoresistivity and filamentary superconductivity in nickel-doped BaFe2As2
Zhang Wei1, 2, Dai Yao-Min2, Xu Bing2, Yang Run2, Liu Jin-Yun2, Sui Qiang-Tao2, Luo Hui-Qian2, Zhang Rui2, Lu Xing-Ye2, Yang Hao3, †, , Qiu Xiang-Gang2, ‡,
College of Physics, Optoelectronics and Energy and Collaborative Innovation Center of Suzhou Nano Science and Technology, Soochow University, Suzhou 215006, China
Beijing National Laboratory for Condensed Matter Physics, Institute of Physics, Chinese Academy of Sciences, Beijing 100190, China
College of Science, Nanjing University of Aeronautics and Astronautics, Nanjing 211106, China

 

† Corresponding author. E-mail: yanghao@nuaa.edu.cn

‡ Corresponding author. E-mail: xgqiu@iphy.ac.cn

Project supported by the National Basic Research Program of China (Grant Nos. 2012CB821400, 2012CB921302, and 2015CB921303) and the National Natural Science Foundation of China (Grant Nos. 11274237, 91121004, 51228201, 11004238, and 11374011).

Abstract
Abstract

We present magnetotransport studies on a series of BaFe2−xNixAs2 (0.03 ≤ x ≤ 0.10) single crystals. In the underdoped (x = 0.03) non-superconducting sample, the temperature-dependent resistivity exhibits a peak at 22 K, which is associated with the onset of filamentary superconductivity (FLSC). FLSC is suppressed by an external magnetic field in a manner similar to the suppression of bulk superconductivity in an optimally-doped (x = 0.10) compound, suggesting the same possible origin as the bulk superconductivity. Our magnetoresistivity measurements reveal that FLSC persists up to the optimal doping and disappears in the overdoped regime where the long-range antiferromagnetic order is completely suppressed, pointing to a close relation between FLSC and the magnetic order.

1. Introduction

The discovery of unconventional superconductivity in iron pnictides[1] has aroused a tremendous amount of research into this class of materials in an effort to understand the mechanism of high-Tc superconductivity, as well as to explore materials with higher Tc. The parent compounds of these iron-based superconductors (FeSCs) are poor Pauli-paramagnetic metals in their high-temperature phase, and undergo structural and antiferromagnetic (AFM) phase transitions upon cooling down to low temperature.[24] Superconductivity emerges with the suppression of the AFM order via chemical substitution[57] or the application of pressure.[811] Since optimal superconductivity is accompanied by the disappearance of the AFM phase, AFM spin fluctuations have been proposed to mediate the electronic pairing in FeSCs.[12] On the other hand, previous spectroscopic studies have also revealed strong competition between the AFM order and superconductivity in the Ba122 system.[13,14] Therefore, the relation between magnetism and superconductivity is still a matter of debate in FeSCs.

In this article, we report our results of magnetotransport measurements on BaFe2−xNixAs2 single crystals with different Ni concentration. An abrupt drop of the resistivity at ∼22 K in underdoped x = 0.03 has been observed in the temperature-dependent resistivity as evidence for the onset of filamentary superconductivity (FLSC). An external magnetic field suppresses FLSC in a manner similar to the suppression of the bulk Tc in an optimally-doped (x = 0.10) sample, suggesting a possible connection between FLSC and bulk SC. In addition, FLSC is robust and exhibits little doping dependence in the underdoped regime, but vanishes abruptly in the overdoped region where long-range AFM order is absent. These results indicate that the emergence of FLSC is intimately related to the magnetic order in FeSCs.

2. Experiments

High-quality single crystals of BaFe2−xNixAs2 with a series of Ni doping were grown using a self-flux method.[15] Typical dimensions of single crystals for measurements are approximately 1 mm × 0.25 mm × 0.05 mm. In-plane resistivity was measured as a function of temperature in a physical property measurements system (PPMS-9, Quantum Design) using a standard four-electrode method. The current used for measurements is parallel to the basel plane and I = 1 mA.

3. Results and discussion

Figure 1 shows the temperature dependence of normalized resistivity ρ(T)/ρ(300 K) for x = 0.03 underdoped and x = 0.10 optimally doped BaFe2−xNixAs2 samples. For x = 0.10, the curve is characterized by a steep superconducting transition at Tc = 20.3 K. The x = 0.03 sample features a metallic behavior followed by an upturn resistivity at the spin-density-wave (SDW) transition temperature TN = 109 K, corresponding to a sharp dip in the derivative of the resistivity dρ/dT as a function of temperature. Meanwhile, as shown in the upper inset in Fig. 1, an abrupt drop of resistivity can be clearly identified at low temperature Tfl ≈ 22 K for the x = 0.03 sample. Our recent studies on CaFe2As2[16] and BaFe2−xCoxAs2[17] has demonstrated that this step in the resistivity is due to the presence of weakly pinned superconducting filaments. Furthermore, we have also provided evidence that filamentary superconductivity nucleated at antiphase domain walls in antiferromagnetic CaFe2As2. Therefore, we define the temperature of this step in resistivity as the onset of FLSC (Tfl), similar to the high-Tc cuprates[18,19] and heavy fermions superconductors.[20,21]

Fig. 1. The temperature-dependent normalized resistivity ρ(T)/ρ(300 K) curves for BaFe2−xNixAs2 single crystals with x = 0.03 and x = 0.10. Top inset: the enlarged view of ρ(T)/ρ(300 K) curve for underdoped x = 0.03 sample; Bottom inset: the derivative curve dρ/dT for x = 0.03 sample.

The top panel of Fig. 2 shows ρ(T) curves from 2 K to 110 K at different magnetic fields for underdoped BaFe2−xNixAs2 (x = 0.03). While no magnetoresistance is observed above the magnetic transition temperature TN, a peak feature related to FLSC is observed at temperature Tfl in all curves. It is noticed that Tfl decreases with increasing magnetic fields. The HT phase diagram obtained from the field dependence of Tfl for the x = 0.03 sample and the field dependence of bulk Tc for the x = 0.10 sample are shown in the bottom panel of Fig. 2. A striking similarity of the suppression by external magnetic field between Tfl and Tc is found through the comparison of these HT phase diagrams, which suggests a close connection between FLSC and bulk SC and rules out the possible contribution of an impurity phase to this phenomena.

Fig. 2. (a) Temperature-dependent reduced resistivity ρ(T)/ρ(300 K) curves for BaFe1.97Ni0.03As2 single crystal with different applied magnetic fields (H = 1 T–9 T). (b) HT phase diagrams of BaFe2−xNixAs2 single crystal (x = 0.03 and x = 0.10). The black solid lines are fits of the data by using the GL expression, the red dotted lines are fits with the Werthemer–Helfand–Hohenberg (WHH) model.

In addition, the upper critical fields in these two samples are also fitted by the Ginzburg–Landau (GL) model and the Werthamer–Helfand–Hohenberg (WHH) model. As shown in the bottom panel of Fig. 2, the solid curve is a fit with GL expression,

which gives Hc2(0) = 23.7 T and Tc = 20.6 K for the underdoped x = 0.03 sample, and Hc2(0) = 40.8 T and Tc = 20.7 K for the optimally-doped x = 0.10 sample. The dashed curve is a fit with the WHH relation, Hc2(0) = −0.7Tc (dHc2/dTc), which gives a slope of −1.34 T/K and Hc2(0) = 19.4 T with Tc = 20.6 K for the x = 0.03 sample, and a slope of −2.14 T/K and Hc2(0) = 31.1 T with Tc = 20.7 K for the x = 0.10 sample. The value of Hc2(0) obtained from WHH is lower than GL, which is similar to that reported in other 122-type iron pnictide superconductors.[22] Meanwhile, it is noted that recently Wang et al. reported that the of x = 0.10 sample with high field data can be described by a two-band model and it gave a higher Hc2 = 47.5 T.[23]

While we observed an obvious drop of resistivity in ρ(T) curves for the x = 0.03 sample, some studies showed that the step in the low-temperature resistivity is not always observed in 122-type parent compound in iron pnictide superconductors. For example, Xiao et al. reported that this step was observed only for small and thin single crystals in undoped BaFe2As2.[17] Tanatar et al. studied different samples of parent compounds CaFe2As2 and BaFe2As2, and found that two samples showed the partial superconducting transition in three undoped CaFe2As2, and even only two samples showed the partial superconducting transition in five different BaFe2As2 samples.[24] For this reason, we performed magnetoresistivity measurements, which is believed to be more sensitive to the small decrease in resistivity due to filamentary superconductivity.

Figure 3 shows the temperature dependence of magnetoresistivity Δρ/ρ(0) = [ρ(9 T) − ρ(0 T)]/ρ(0 T) for different Ni concentrations. As the temperature decreases, the magnetoresistivity Δρ/ρ(0) for the underdoped x = 0.03 sample increases very gently at first, but then an abrupt increase occurs close to 22 K, which is exactly the FLSC onset temperature Tfl in the zero-field ρ(T) curve. Therefore, we attribute this abrupt change in magnetoresistivity to the occurrence of filamentary superconductivity and identify the temperature as Tfl. We also performed magnetoresistivity measurements on several other pieces of crystal of underdoped x = 0.03 samples and confirmed that the abrupt change in magnetoresistivity always exists at a temperature Tfl even in some samples where the step in ρ(T) curve has not been detected. Furthermore, the abrupt change in magnetoresistivity is observed in other Ni-doping BaFe2−xNixAs2 single crystals in Fig. 3.

Fig. 3. (a) Temperature-dependent magnetoresistivity [ρ(9 T) − ρ(0 T)]/ρ(0 T) curves for BaFe2−xNixAs2 single crystals with x = 0.03, 0.05, 0.065, 0.085, 0.092, and 0.096. Data were offset vertically to separate the overlapped curves between different doping.

In Fig. 4, we make a comprehensive plot of the phase diagram of the doping dependence of structural transition temperature Ts, magnetic transition temperature TN, superconducting transition temperature Tc, and FLSC transition temperature Tfl. The defination of Ts, TN, and Tc is shown in Fig. 1. From the phase diagram we find that the parent compound BaFe2As2 exhibits simultaneous structural and magnetic phase transitions at ∼140 K, from high-temperature tetragonal paramagnetic to low-temperature orthorhombic antiferromagnetic phase. With increasing Ni doping, the structural and magnetic transitions separate from each other with TN slightly lower than Ts. The antiferromagnetic order is suppressed by replacing Fe by Ni in BaFe2As2. Superconductivity emerges at x ≈ 0.05 and reaches a maximum at x = 0.10 where long-range AFM order vanishes. Most importantly, we find that Tfl is less doping dependent compared with the bulk superconductivity Tc and it exists in underdoped samples and persists up to the edge of the optimally-doped x = 0.10 sample, where the AFM order vanishes. This suggests a close relationship between FLSC and AFM order.

Fig. 4. Doping-x dependence of Ts, TN, Tc, and Tfl phase diagram. Tfl in parent compound BaFe2As2 is obtained from Ref. [17].

However, some recent experimental results revealed that SC and AFM order parameters are spatially modulated on a microscopic scale in iron pnictides.[2426] For example, many studies showed the evidence for the existence of the structural domain like twin boundary and antiphase domain wall in the iron pictides.[24,26] Meanwhile, a recent SQUID microscopy study in underdoped Ba(Fe1−xCox)2As2 showed an enhanced superfluid density on twin boundaries[25] and 75As nuclear magnetic resonance (NMR) measurements in antiferromagnetic CaFe2As2 suggests the presence of FLSC nucleated at the AFM domain walls.[16] All these results indicate that AFM order is vital to the emergence of SC in iron pnictides.

In combination with our previous studies in CaFe2As2 and Ba(Fe1−xCox)2As2,[16,17] we find that FLSC exists in electron-doped 122-type iron-based superconductors and it has a close connection with AFM order. Besides, Chu et al. also revealed an obvious FLSC with a high Tc = 49 K in single crystalline CaFe2As2 via electron-doping by partial replacement of Ca by rare-earth.[27,28] Thus our results suggest that the FLSC may be a common feature in the AFM phase of iron pnictides due to strong coupling between AFM and superconductivity, further exploration is needed to verify it.

4. Conclusion

In conclusion, magnetotransport measurements have been carried out on a series of Ni doping BaFe2As2 single crystals. A similar suppression of FLSC Tfl and bulk SC Tc by external magnetic field is depicted in the HT phase diagram. The doping-temperature phase diagram reveals the following results: (i) FLSC is less doping-dependent compared with bulk SC; (ii) FLSC coexists with AFM order and vanishes in the edge of optimal doping where long-range AFM order is absent. Based on our results, we conclude that FLSC is closely linked with magnetic order and the bulk SC in iron pnictides.

Reference
1Kamihara YWatanabe THirano MHosono H 2008 J. Am. Chem. Soc. 130 3296
2Ni NNandi SKreyssig AGoldman A IMun E DBud’ko S LCanfield P C 2008 Phys. Rev. 78 014523
3Huang QQiu YBao WGreen M ALynn J WGasparovic Y CWu TWu GChen X H 2008 Phys. Rev. Lett. 101 257003
4Rotter MTegel MJohrendt DSchellenberg IHermes WPöttgen R 2008 Phys. Rev. 78 020503
5Rotter MTegel MJohrendt D 2008 Phys. Rev. Lett. 101 107006
6Ni NThaler AYan J QKracher AColombier EBud’ko S LCanfield P CHannahs S T 2010 Phys. Rev. 82 024519
7Jiang SXing HXuan GWang CRen ZFeng CDai JXu ZGao G 2009 J. Phys.: Condens. Matter 21 382203
8Torikachvili M SBud’ko S LNi NCanfield P C 2008 Phys. Rev. Lett. 101 057006
9Park TPark ELee HKlimczuk TBauer E DRonning FThompson J D 2008 J. Phys.: Condens. Matter 20 322204
10Alireza P LKo Y T CGillett JPetrone C MCole J MLonzarich G GSebastian S E 2009 J. Phys.: Condens. Matter 21 012208
11Kitagawa KKatayama NGotou HYagi TOhgushi KMatsumoto TUwatoko YTakigawa 2009 Phys. Rev. Lett. 103 257002
12Mazin I ISingh D JJohannes M DTakigawa M 2008 Phys. Rev. Lett. 101 057003
13Chia E E MTalbayev DZhu J XYuan H QPark TThompson J DPanagopoulos CChen G FLuo J LWang N LTaylor A J 2010 Phys. Rev. Lett. 104 027003
14Dai Y MXu BShen BWen H HHu J PQiu X GLobo R O S M 2012 Phys. Rev. 86 100501
15Chen YLu XWang MLuo HLi S 2011 Supercond. Sci. Technol. 24 065004
16Xiao HHu TDioguardi A PRoberts-Warren NShockley A CCrocker JNisson D MViskadourakis ZTee X YRadulov IAlmasan C CCurro N JPanagopoulos C 2012 Phys. Rev. 85 024530
17Xiao HHu THe S KShen BZhang W JXu BHe K FHan JSingh Y PWen H HQiu X GPanagopoulos CAlmasan C C 2012 Phys. Rev. 86 064521
18Gomes K KPasupathy A NPushp AOno SAndo YYazdani A 2007 Nature 447 569
19Li QHücker MGu G DTsvelik A MTranquada J M 2007 Phys. Rev. Lett. 99 067001
20Bianchi AMovshovich RJaime MThompson J DPagliuso Sarrao J L 2001 Phys. Rev. 64 220504
21Chen YJiang W BGuo C YYuan H QFisk ZThompson J DLu X 2015 Phys. Rev. Lett. 114 146403
22Kumar NNagalakshmi RKulkarni RPaulose P LNigam A KDhar S KThamizhavel A2009Phys. Rev. B79012504
23Wang Z SXie TKampert EFörster TLu X YZhang RGong D LLi S LHerrmannsdörfer TWosnitza JLuo H Q 2015 Phys. Rev. 92 174509
24Tanatar M AKreyssig ANandiNi NBud’ko S LCanfield P CGoldman A IProzorov 2009 Phys. Rev. 79 180508
25Kalisky BKirtley J RAnalytis J GChu J WVailionis AFisher I RMoler K A 2010 Phys. Rev. 81 184513
26Li GHe XZhang JJin RSefat A SMcGuire M AMandrus D GSales B CPlummer E W 2012 Phys. Rev. 86 060512
27Lv BDeng L ZGooch MWei F YSun Y YMeen J KXue Y YLorenz BChu C W 2011 Proc. Natl. Acad. Sci 108 15705
28Gofryk KPan M HCantoni CSaparov BMitchell JSefat S 2014 Phys. Rev. Lett. 112 047005