Abstract
In hightransitiontemperature superconducting cuprates and iron arsenides, chemical doping plays an important role in inducing superconductivity. Whereas in the cuprate case, the dominant role of doping is to inject charge carriers, the role for the iron arsenides is complex owing to carrier multiplicity and the diversity of doping. Here, we present a comparative study of the inplane resistivity and the optical spectrum of doped BaFe_{2}As_{2}, which allows for separation of coherent (itinerant) and incoherent (highly dissipative) charge dynamics. The coherence of the system is controlled by doping, and the doping evolution of the charge dynamics exhibits a distinct difference between electron and hole doping. It is found in common with any type of doping that superconductivity with high transition temperature emerges when the normalstate charge dynamics maintains incoherence and when the resistivity associated with the coherent channel exhibits dominant temperaturelinear dependence.
Introduction
Hightransitiontemperature (highT_{c}) superconductivity in iron arsenides is induced by various doping (chemical substitution) processes into metallic parent compounds showing an antiferromagneticorthorhombic (AFO) order. For a representative parent iron arsenide BaFe_{2}As_{2}, the superconducting (SC) phase is attained by doping into all the three different lattice sites^{1,2,3}, and even by isovalent atomic substitution, e.g., P for As^{3} and Ru for Fe^{4}. Notably, both electron and hole doping can generate superconductivity, in stark contrast to the case of highT_{c} cuprates, in which only one type of carrier doping can induce superconductivity for one parent compound, and superconductivity is not achieved by substituting Ni or Zn for Cu in Mott insulating parent compounds.
The mechanism of various doping processes leading to the emergence of superconductivity is a matter of debate. Charge carriers appear to be supplied as revealed by the study of angleresolved photoemission spectroscopy (ARPES), which demonstrates that the electron Fermi surface (FS) expands (the hole FS shrinks) for electron doping (e.g., Co substitution for Fe)^{5,6} and vice versa for hole doping (e.g., K substitution for Ba)^{7}. However, how this contrasting doping evolution of the FS is reflected in charge transport has not been seriously studied. More fundamentally, the question of how a change in electron/hole density in already metallic parent compounds alters the electronic environment to favor superconductivity or whether the change in carrier density is really an important ingredient for superconductivity is nontrivial. In addition to the change in carrier density, a chemicalpressure effect also has a significant influence on the electronic structure^{8,9,10}. To gain a deeper understanding of the physics of doping in ironbased superconductors, it is of fundamental importance to extract universal natures such as the necessary ingredients for superconductivity among different doping processes. Because doping into BaFe_{2}As_{2} covers the entire doping range from the AFO to the SC phase and further to the metallic nonSC end compounds, it is a suitable system for investigating the physical implications of doping.
Optical spectroscopy is a bulksensitive and energyresolved probe and turns out to be a powerful tool to study charge dynamics and its evolution with doping in a multicarrier system, particularly iron pnictides. The optical conductivity spectrum of iron arsenides is characterised by a small zeroenergy peak, corresponding to a relatively coherent carrier dynamics, and a long tail extending to a highenergy region, indicative of the presence of an incoherent charge dynamics with large spectral weight^{11,12,13,14,15}. Such a spectral feature is also seen in the spectrum of the highT_{c} cuprates and characterises the normalstate charge dynamics in highT_{c} superconductors. The combination of resistivity and optical conductivity makes it possible to separate the contribution of multiple carriers or to separate a coherent component from an incoherent one in charge dynamics by disentangling the carrier scattering rate from the carrier density. Indeed, the decomposition of the spectrum for Codoped BaFe_{2}As_{2} explains the temperature and doping dependence of the charge dynamics well^{12}.
Here, we show how the magnitude and the temperature (T) dependence of the inplane resistivity ρ(T) evolve with doping (chemical substitution) into the three different sites in the parent compound BaFe_{2}As_{2}: Ba(Fe_{1−x}Co_{x})_{2}As_{2}, BaFe_{2}(As_{1−x}P_{x})_{2}, and Ba_{1−x}K_{x}Fe_{2}As_{2} (abbreviated as CoBa122, PBa122, and KBa122, respectively). In the present study, we focus on the effect of doping in the hightemperature paramagnetictetragonal (PT) phase. The results of the doping evolution of resistivity and optical conductivity spectrum indicate that a dominant effect is to control coherence/incoherence in charge dynamics in different ways depending on the dopant sites/species. Superconductivity is found to emerge when the normalstate charge dynamics maintains incoherence. It is also found from the analysis of the T dependence of resistivity that, in the region of superconductivity showing fairly high T_{c} values, the normalstate resistivity is dominated by the Tlinear scattering term.
Results
Doping evolution of inplane resistivity
Figure 1 shows the doping evolution of ρ(T) for the three systems, Co, P, and KBa122. The undoped parent compound BaFe_{2}As_{2}, the quality of which is remarkably improved by annealing, shows very low residual resistivity (< 10 µΩcm) at low temperatures in the AFO phase, and the AFOPT transition temperature T_{s} is 143 K, the highest among the temperatures so far reported^{16,17}. Nevertheless, the resistivity at 300 K is fairly high, ~340 µΩcm and is comparable with that for unannealed crystals. The T dependence is weak in the PT phase above T_{s}, suggesting that the charge carriers are highly incoherent. The resistivity in the PT phase rapidly decreases with Co and P doping, in sharp contrast with that in the AFO phase, where the resistivity increases with doping, indicating that doping plays different roles in the two phases. In contrast, the change in magnitude is moderate in the case of K doping.
The evolution of ρ(T) with Co doping is basically the same as that reported by RullierAlbenque et al.^{18}, except for a significantly lower resistivity in the AFO phase in the present result. With Co doping, the temperature dependence changes from T linear to T^{2} in the overdoped nonSC region. Note that a fairly large residual resistivity remains up to x = 0.24, indicating that a doped Co atom works as a relatively stronger scattering center. The substitution of P for As corresponds to chemically isovalent doping. The number of electrons and holes is expected to be balanced over the entire range of doping. Controversially, as in the case of CoBa122, the magnitude of the resistivity in the PT phase monotonically decreases with P doping, and the T dependence gradually changes from Tlinear to T^{2}, in agreement with the result obtained by Kasahara et al.^{3}. A difference is that a doped P atom is a weaker scattering center as evidenced by a smaller residualresistivity component. The doping evolution of ρ(T) in KBa122 is in sharp contrast to the Co and Pdoping cases^{19,20}. Chemically, K doping corresponds to hole doping, but both the magnitude and T dependence of the resistivity only weakly change with doping. At low temperatures and at any doping level, the resistivity rapidly decreases with decreasing temperature, showing the T^{2} dependence typical of a Fermi liquid, whereas it shows a clear tendency for saturation in the hightemperature PT phase. As far as the highT region is concerned, even the end compound KFe_{2}As_{2} with T_{c} ~ 3.5 K is not a good metal. Although a finite residual resistivity appears for the compounds showing the AFO order (x ≤ 0.23), the magnitude of the residual resistivity for the Kdoped compounds is considerably small compared with the Co and Pdoping cases, probably because a doped K atom is located farthest away from the Fe plane.
Doping evolution of optical conductivity spectrum
Figure 2(a) displays the optical conductivity spectrum of BaFe_{2}As_{2}. The spectrum shows a small peak at ω = 0 and a long tail merging into a broad peak in the higherenergy region. Using the Drude–Lorentz model, we decompose the spectrum as shown in Fig. 2(a). Because a single Drude component cannot reproduce the lowenergy spectrum well, we decomposed the spectrum using two Drude components^{12,14,15,21,22,23}. Given that iron pnictides are multiband systems, the presence of multiple Drude components is a valid result. The black and orange lines indicate the experimental and fitting results, respectively. The lowenergy conductivity is dominated by two Drude components with distinct characteristics: one is narrow with a small spectral weight (in blue) and the other is broad with a much larger spectral weight (in gray). A higherenergy component approximated by a Lorentzian function (in green) corresponds to an interband excitation.
The width of the broad Drude component (the scattering rate 1/τ) is ~2000 cm^{−1}, corresponding to τ = 1.7 × 10^{−14} s. From the band dispersions of BaFe_{2}As_{2} obtained by ARPES measurements^{24}, it is seen that the Fermi velocity v_{F} of the band with a heavier effective mass is much less than 1.0 × 10^{6} cm/s. From this, the mean free path l ( = v_{F}τ) for the broad Drude component is estimated to be less than 1.7 Å, which is significantly shorter than the shortest interatomic spacing (2.4 Å). On the other hand, for the narrow Drude component with a width of ~200 cm^{−1}, we estimate the Fermi velocity to be and hence the mean free path to be . It is considerably longer than the lattice parameter (), which indicates relatively coherent charge dynamics. Hereafter, we call the narrow and broad Drude components the coherent and incoherent components (σ_{D} and σ_{in}), respectively. The incoherent component turned out to be nearly T independent as was previously pointed out^{12,21,23}, and the lowenergy conductivity can be written as The evolution of each component with x for CoBa122 is shown in Figs. 2(b)–2(d)^{12}. With Co doping, the weight of the narrow Drude component (coherent Drude weight) increases, whereas that of the broad Drude component shows no appreciable change. Figures 2(e)–2(h) show the evolution of the spectrum with P doping. The substitution of isovalent P for As supplies neither extra electrons nor holes. However, the coherent Drude weight remarkably increases with P doping without changing the total lowenergy spectral weight below 6000 cm^{−1}; the spectral weight is transferred to the coherent component from the higherenergy region^{10}. As a result, the narrow Drude component dominates in the lowenergy spectrum of BaFe_{2}P_{2} [Fig. 2(h)].
The decomposition of the spectrum of KFe_{2}As_{2} [Fig. 2(i)] is nearly the same as that of BaFe_{2}As_{2}, in which the fraction of the coherent Drude weight remains small, or even smaller, and hence the lowenergy conductivity spectrum is also dominated by the incoherent component. In view of the spectrum presented in Ref. 25 for the K content x = 0.4, the highly incoherent spectrum seems to persist over the entire doping range of KBa122 without showing any appreciable spectral change.
Discussion
Here, we discuss the resistivity and optical conductivity results for the three doped BaFe_{2}As_{2} systems in the context of the available results for the Hall coefficient and the FS observed by ARPES. In the hightemperature PT phase of BaFe_{2}As_{2}, there are two electron FS pockets near the Brillouin zone boundary and three hole FS pockets around the zone center. Co doping makes the hole FS shrink and the electron FS expand^{5,6}. The Hall coefficient, indeed, decreases in magnitude with a negative sign as doping proceeds^{18}. Hence, the increase in the coherent Drude weight with Co doping is associated predominantly with the increase in electron carriers supplied by the dopant Co atoms, and introduced electrons are not localised on the Co atom^{26}.
On the other hand, the hole FS expands (the electron FS shrinks) with K doping, and eventually the electron FS disappears in the highly doped region^{7,27}. However, the resistivity does not decrease appreciably with K doping, and the coherent Drude weight remains small over the entire doping range. These observations are in contrast to the electrondoped CoBa122 system. To reconcile these results with the doping evolution of the FS observed by ARPES, it is necessary to assume that quasiparticle states on a substantial part of the hole FS pockets remain incoherent in the holedoped system.
The result for the isovalent P doping gives additional evidence for the electron–hole asymmetry. With P doping, the numbers of the electron and hole FS pockets do not change and thus the FS topology does not significantly change over the entire range of P content, although one of the hole FS pockets is more warped along the caxis direction as x increases^{24}. Nevertheless, the P doping significantly increases the coherent Drude weight and decreases the resistivity, which transforms the bad metallic BaFe_{2}As_{2} into a good metal as in the case of Co doping. The coherent Drude weight is affected by the carrier density and the effective mass (the influence of the effective mass will be discussed later). In view of the negative Hall coefficient, the magnitude of which decreases with increasing P content^{3}, it follows that, as doping proceeds, the coherent region in the momentum space, which is originally restricted to a portion of the electron FS pockets, gradually expands within the electron FS pockets, while the states on the hole FS pockets remain incoherent. This agrees with the conclusion drawn from the contrasting evolution of coherence/incoherence for electron and holedoped BaFe_{2}As_{2}. Earlier, it was reported that holes are subject to stronger scattering than electrons^{18,19,28}. The result of PBa122 is reminiscent of the holedoped highT_{c} cuprates, in which an increase in the hole doping level from the underdoped to the optimally doped region makes the coherent states extend from the nodal to the antinodal region on the FS above the pseudogap temperature.
In Fig. 3(a), we plot the weight of the narrow Drude component N_{D} as a function of the dopant concentration x for the three systems. N_{D} is calculated with where m_{0} and V are the free electron mass and the cell volume containing one Fe atom, respectively. N_{D} is proportional to the coherent carrier density n_{D} and inversely proportional to the carrier effective mass m*: . Open symbols indicate the data compiled from Refs. 10, 12, and 25. Strikingly different evolutions of N_{D} are seen, despite the similar phase diagram for each system. For CoBa122, an initial increase in N_{D} with x is very steep. The N_{D} of PBa122 shows a superlinear increase with x, a gradual increase followed by a rapid increase for x > 0.6. The much faster increase in N_{D} with Co doping indicates that electron carriers are involved in the coherent charge dynamics. By contrast, the N_{D} for KBa122 weakly decreases with increasing x. Compared with the doping evolution of resistivity shown in Fig. 1, it is clear that the magnitude of resistivity follows N_{D} in the highT region in each system. Note that, in the case of K doping, the resistivity only slightly decreases with x, but the hole density steadily increases with x, as evidenced by the increase in the volume of the hole FS observed by ARPES. The measurement of quantum oscillations for KFe_{2}As_{2} revealed an enhancement in m*^{29,30}, indicating that the decrease in N_{D} is partly due to the increase in m*. Probably, the coherent hole carrier density n_{D} would increase with a rate of increase slower than that of m*. Unlike the K doping, a decrease in m* with x (for x > 0.3) is reported for P doping by quantum oscillations^{31}. m* decreases by a factor of 2–3 from x = 0.3 to 1. In this respect, the increase in N_{D} for PBa122 may, in most part, be due to a decrease in m*.
A bad metallic behavior in the iron arsenides manifests in the highT region, where the magnitude of resistivity is fairly high, and the resistivity shows a tendency for saturation. This signals a dominant contribution of the broad (incoherent) Drude component at high temperatures in equation (1). σ_{in} is only weakly dependent on the temperature, which is found to reproduce the lowenergy σ(T, ω) well^{12,23}. The resistivity saturation is understood in terms of conductivity in two channels at ω = 0, where σ_{in} is the ω = 0 value of the incoherent conductivity term. For iron arsenides, σ_{in} is comparable with σ_{D}. The resistivity from the coherent component is expressed by the sum of the contributions from Tindependent elastic and Tdependent inelastic scattering (ρ_{0} and ρ_{ε}, respectively); . With increasing temperature, σ_{D} decreases, and consequently the contribution of σ_{in} becomes dominant.
Equation (2) happens to be similar to a simple parallel resistor formula empirically describing the resistivity saturation, in which σ_{in} is replaced by , ρ_{sat} being the maximum resistivity corresponding to the carrier mean free path equal to the lattice spacing^{32}: . The present σ_{in} is not such a universal quantity but is dependent on the material and the doping level as shown in Fig. 3(b). This is distinct not only from the case of conventional resistivity saturating metals but also from the nonsaturating resistivity of the known bad metals, many of which are in close proximity to a Mott insulating state, such as highT_{c} cuprates^{33,34}. The nonsaturating resistivity is thought to arise from a reduction in lowfrequency conductivity with increasing temperature, the spectral weight of which is transferred to higher frequencies involving transitions to the upper Hubbard band^{34,35}. In iron arsenides, the presence of multiple carriers with distinct characteristics is responsible for the apparent resistivity saturation, and the highly incoherent hole carriers with a mean free path shorter than the lattice spacing^{14} are mainly responsible for the bad metallic behavior.
The result shown in Fig. 3(a) demonstrates that doping controls N_{D} in quite different ways depending on dopant sites and/or species. It is surprising that any of the different doping processes leads to the emergence of superconductivity. The critical doping level above which superconductivity appears more or less depends on the suppression rate of the AFO order, but it is an open question as to what conditions should be fulfilled for superconductivity to appear. On the other hand, the result indicates that the SC phase persists as long as N_{D} is sufficiently small [N_{D} < 0.02–0.03, shaded region in Fig. 3(a)] and that the system cannot sustain superconductivity when N_{D} becomes larger or the charge dynamics becomes more coherent. N_{D} rapidly increases with Co doping, and soon reaches ~0.02 at x ~ 0.10. In the case of P doping, N_{D} increases slowly and exceeds 0.03 at approximately x = 0.6. This seems to corroborate the fact that the SC phase terminates at x ~ 0.15 and ~ 0.7 for Co and P doping, respectively. Unlike these two systems, the N_{D} of KBa122 maintains small values up to x = 1, consistent with the persistent superconductivity to KFe_{2}As_{2}. The incoherent normalstate charge dynamics is therefore a prerequisite for superconductivity.
Finally, we analyse the T dependence of the resistivity in the PT phase predominantly arising from the coherent component and search for a correlation with superconductivity. Assuming the two parallel (inelastic) scattering channels, we adopt a dualcomponent analysis for the Tdependent part of resistivity, ρ_{ε}(T) = α_{1}T + α_{2}T^{2}. The T^{2} term is typical of a conventional Fermi liquid, and the Tlinear term usually results from electron–boson interactions such as antiferromagnetic spin fluctuations. This dualcomponent analysis was applied to fit the resistivity of cuprates^{36} and iron pnictides^{37}. However, this formula does not reproduce the trend for saturation observed for most iron arsenides at high temperatures. Therefore, we have to incorporate the incoherent channel σ_{in} explicitly and use the formula where ρ_{0} represents the elastic scattering component adding to the inelastic scattering processes. σ_{in} can be estimated from the ω → 0 value of the broad (incoherent) Drude component in the present optical conductivity spectrum. For KBa122, larger values of σ_{in} are required to fit the resistivity curves than the experimentally extracted values. This might indicate that the decomposition of optical conductivity underestimates the contribution from the incoherent component, pointing towards the more incoherent charge dynamics inherent in KBa122 at high temperatures. Figure 4 shows the fitting results of the resistivity curves in the PT phase for the three systems under consideration. For undoped and underdoped compounds showing the AFO order, inplane resistivity anisotropy is observed even in the PT phase above T_{s}, but the anisotropy is strongly suppressed by annealing^{16,38}, which guarantees the present fitting down to near T_{s}. Equation (3) with σ_{in} reproduces well the resistivity curves over a wide temperature and doping range.
Figures 5(a)–5(c) show the doping dependences of the coefficients of the Tlinear and square terms multiplied by N_{D}, α_{1}·N_{D} and α_{2}·N_{D}, respectively, for the three doping systems obtained using the bestfit parameters, α_{1} and α_{2}, to reproduce the resistivity curves (Fig. 4). For the compositions we did not measure the optical spectrum, N_{D} is estimated by assuming smooth variation with x, shown as the dashed curves in Fig. 3(a). In the framework of the present twofluid and dualchannel analysis, α_{1}·N_{D} and α_{2}·N_{D} are more directly related to the scattering rate of the coherent carriers 1/τ_{D} (σ_{D} = N_{D}e^{2}τ_{D}/m_{0}). In common with the three systems, α_{1}·N_{D} is largest near the AFOSC phase boundary where T_{c} is highest, whereas it becomes vanishingly small in the overdoped regime where superconductivity disappears in Co and PBa122. Superconductivity in KBa122 persists up to x = 1, but α_{1}·N_{D} remarkably decreases on going from x = 0.69 to 0.88. This probably corresponds to the disappearance of the electron FS pockets^{7} and/or the change in the SC gap function on FS from a nodeless^{39} to a nodal one^{40,41}. Unlike α_{1}·N_{D}, α_{2}·N_{D} remains finite over a wide range of doping regardless of the T_{c} value and does not show any noteworthy feature at the phase boundary. As shown in Fig. 3(a), the weight of the coherent Drude component N_{D} varies monotonically with x. Therefore, the nonmonotonic change in α_{1}·N_{D} certainly arises from that of α_{1}, the Tlinear channel of the carrier scattering rate.
The predominance of the Tlinear term implies stronger inelastic scattering of carriers, and the observed correlation between α_{1}·N_{D} and T_{c} strongly suggest that the scattering mechanism giving rise to the Tlinear resistivity has an intimate connection with the SCpair formation. Cooper et al. attempted to fit the temperature dependence of the inplane resistivity of highly doped La_{2−x}Sr_{x}CuO_{4} (x > 0.15) using a similar formula^{36}. They found that α_{1} is largest at x = 0.19, which does not coincide with the composition showing the highest T_{c}, and decreases with an increase or a decrease in the doping level, whereas α_{2} is only weakly dependent on the hole concentration down to x = 0.19. It should be noted that the Tlinear resistivity in the holedoped cuprates is different in nature. Tlinearity in the cuprates persists up to very high temperatures, ~1000 K or even higher without showing resistivity saturation, and originates from highly incoherent quasiparticle states. The quasiparticle scattering rate (1/τ) in the cuprates is strongly momentum (and frequency) dependent and is extremely high at momenta near the antinodes, , which is near the socalled Planckian dissipation limit^{36,42}. This is in contrast to the present iron pnictide case, in which the Tlinear term arises from the coherent channel in parallel to the incoherent one, and the resistivity exhibits a tendency for saturation due to the nearly Tindependent incoherent term. The width of the narrow Drude component is ~200 cm^{−1} at T = 300 K for every system near the doping level that produces the highest T_{c} value. This corresponds to , which is by a factor of ~6 lower than that in La_{2–x}Sr_{x}CuO_{4}.
The Tlinear resistivity is often invoked in relation to enhanced scattering from critical fluctuations near a quantum critical point (QCP). For PBa122, the presence of a magnetic QCP at x ~ 0.3 was suggested^{31,43}. The Tlinear resistivity coefficient α_{1} peaks at x ~ 0.3, and the effective mass m* estimated from de Haas oscillations increases from the overdoped region towards x ~ 0.3. Unfortunately, it is not clear whether the QCP scenario is applicable to Co and K doping because there have been no reports of the evolution of m* for Co and KBa122. For every doped system, the SC phase is not confined to a region where α_{1} (or α_{1}·N_{D}) shows a peak and persists to x = 1 in the case of KBa122. Therefore, the present results support the scenario that superconductivity in the iron arsenides arises, irrespective of the presence or otherwise of a QCP, from strongly xdependent coupling with some bosonic excitations, which is a source of the Tlinear carrier scattering in the coherent component. For the incoherent component, the temperature dependence is very weak. This component experiences extremely strong scattering, which is presumably related to strong electronic correlations working orbitalselectively, already at low temperatures, and additional scattering channels, such as electron–boson scattering, would be no more visible at elevated temperatures.
The facts that α_{1} and T_{c} attain large values near the AFOSC phase boundary and that they steeply decrease farther away from the boundary suggest that the bosonic excitations originate from the fluctuations of the AFO order. As we discussed in the preceding study of the anomalous response of the AFO phase to dopant impurities^{38,44}, the AFO phase is a unique spinchargeorbital complex. Hence, inelastic scattering from shortrange dynamical spinchargeorbital correlations is a candidate of the Tlinear resistivity and hence pairing interactions.
Methods
Single crystals of Co, P, and Kdoped BaFe_{2}As_{2} were grown by the selfflux method^{12,45,46} and were postannealed to improve the sample quality. Resistivity measurements were performed on the ab plane using the fourterminal method. Optical reflectivity was measured at T = 300 K with the incident light almost normal to the ab plane in the frequency range 50–40000 cm^{−1} using a Fourier transform infrared spectrometer (Bruker IFS113v) and a grating monochromator (JASCO CT25C). The optical conductivity was derived from the Kramers–Kronig transformation of the reflectivity spectrum. The Hagen–Rubens or Drude–Lorentz formula was used for the lowenergy extrapolation in order to smoothly connect to the spectrum in the measured region and to fit the measured resistivity value at ω = 0.
References
 1.
Rotter, M., Pangerl, M., Tegel, M. & Johrendt, D. Superconductivity and crystal structures of (Ba_{1–x}K_{x})Fe_{2}As_{2} (x = 0–1). Angew. Chem. Int. Ed. 47, 7949–7952 (2008).
 2.
Nandi, S. et al. Anomalous suppression of the orthorhombic lattice distortion in superconducting Ba(Fe_{1–x}Co_{x})_{2}As_{2} single crystals. Phys. Rev. Lett. 104, 057006 (2010).
 3.
Kasahara, S. et al. Evolution from nonFermi to Fermiliquid transport via isovalent doping in BaFe_{2}(As_{1–x}P_{x})_{2} superconductors. Phys. Rev. B 81, 184519 (2010).
 4.
Thaler, A. et al. Physical and magnetic properties of Ba(Fe_{1–x}Ru_{x})_{2}As_{2} single crystals. Phys. Rev. B 82, 014534 (2010).
 5.
Liu, C. et al. Evidence for a Lifshitz transition in electrondoped iron arsenic superconductors at the onset of superconductivity. Nature Phys. 6, 419–423 (2010).
 6.
Ideta, S. et al. Carrier doping versus impurity potential effect in transition metalsubstituted ironbased superconductors. Phys. Rev. Lett. 110, 107007 (2013).
 7.
Malaeb, W. et al. Abrupt change in the energy gap of superconducting Ba_{1–x}K_{x}Fe_{2}As_{2} single crystals with hole doping. Phys. Rev. B 86, 165117 (2012).
 8.
Miyake, T., Kosugi, T., Ishibashi, S. & Terakura, K. Electronic structure of novel superconductor Ca_{4}Al_{2}O_{6}Fe_{2}As_{2}. J. Phys. Soc. Jpn. 79, 123713 (2010).
 9.
Usui, H. & Kuroki, K. Maximizing the Fermisurface multiplicity optimizes the superconducting state of iron pnictide compounds. Phys. Rev. B 84, 024505 (2011).
 10.
Nakajima, M. et al. Crossover from bad to good metal in BaFe_{2}(As_{1–x}P_{x})_{2} induced by isovalent P substitution. Phys. Rev. B 88, 094501 (2013).
 11.
Hu, W. Z. et al. Origin of the spin density wave instability in AFe_{2}As_{2} (A = Ba, Sr) as revealed by optical spectroscopy. Phys. Rev. Lett. 101, 257005 (2008).
 12.
Nakajima, M. et al. Evolution of the optical spectrum with doping in Ba(Fe_{1–x}Co_{x})_{2}As_{2}. Phys. Rev. B 81, 104528 (2010).
 13.
Hancock, J. N. et al. Strong coupling to magnetic fluctuations in the charge dynamics of ironbased superconductors. Phys. Rev. B 82, 014523 (2010).
 14.
Tu, J. J. et al. Optical properties of the iron arsenic superconductor BaFe_{1.85}Co_{0.15}As_{2}. Phys. Rev. B 82, 174509 (2010).
 15.
Barišić, N. et al. Electrodynamics of electrondoped iron pnictide superconductors: Normalstate properties. Phys. Rev. B 82, 054518 (2010).
 16.
Nakajima, M. et al. Unprecedented anisotropic metallic state in undoped iron arsenide BaFe_{2}As_{2} revealed by optical spectroscopy. Proc. Natl. Acad. Sci. U.S.A. 108, 12238–12242 (2011).
 17.
Ishida, S. et al. Manifestations of multiplecarrier charge transport in the magnetostructurally ordered phase of BaFe_{2}As_{2}. Phys. Rev. B 84, 184514 (2011).
 18.
RullierAlbenque, F., Colson, D., Forget, A. & Alloul, H. Hall effect and resistivity study of the magnetic transition, carrier content, and Fermiliquid behavior in Ba(Fe_{1–x}Co_{x})_{2}As_{2}. Phys. Rev. Lett. 103, 057001 (2009).
 19.
Shen, B. et al. Transport properties and asymmetric scattering in Ba_{1–x}K_{x}Fe_{2}As_{2} single crystals. Phys. Rev. B 84, 184512 (2011).
 20.
Ohgushi, K. & Kiuchi, Y. Doping dependence of Hall coefficient and evolution of coherent electronic state in the normal state of the Febased superconductor Ba_{1–x}K_{x}Fe_{2}As_{2}. Phys. Rev. B 85, 064522 (2012).
 21.
Wu, D. et al. Optical investigations of the normal and superconducting states reveal two electronic subsystems in iron pnictides. Phys. Rev. B 81, 100512(R) (2010).
 22.
Lucarelli, A. et al. Charge dynamics of Codoped BaFe_{2}As_{2}. New J. Phys. 12, 073036 (2010).
 23.
Dai, Y. M. et al. Hidden Tlinear scattering rate in Ba_{0.6}K_{0.4}Fe_{2}As_{2} revealed by optical spectroscopy. Phys. Rev. Lett. 111, 117001 (2013).
 24.
Ye, Z. R. et al. Doping dependence of the electronic structure in phosphorusdoped ferropnictide superconductor BaFe_{2}(As_{1–x}P_{x})_{2} studied by angleresolved photoemission spectroscopy. Phys. Rev. B 86, 035136 (2012).
 25.
Li, G. et al. Probing the superconducting energy gap from infrared spectroscopy on a Ba_{0.6}K_{0.4}Fe_{2}As_{2} single crystal with T_{c} = 37 K. Phys. Rev. Lett. 101, 107004 (2008).
 26.
Wadati, H., Elfimov, I. & Sawatzky, G. A. Where are the extra d electrons in transitionmetalsubstituted iron pnictides? Phys. Rev. Lett. 105, 157004 (2010).
 27.
Nakayama, K. et al. Universality of superconducting gaps in overdoped Ba_{0.3}K_{0.7}Fe_{2}As_{2} observed by angleresolved photoemission spectroscopy. Phys. Rev. B 83, 020501(R) (2011).
 28.
Muschler, B. et al. Band and momentumdependent electron dynamics in superconducting Ba(Fe_{1–x}Co_{x})_{2}As_{2} as seen via electronic Raman scattering. Phys. Rev. B 80, 180510(R) (2009).
 29.
Terashima, T. et al. Fermi surface and mass enhancement in KFe_{2}As_{2} from de Haasvan Alphen effect measurements. J. Phys. Soc. Jpn. 79, 053702 (2010).
 30.
Terashima, T. et al. Complete Fermi surface in BaFe_{2}As_{2} observed via Shubnikovde Haas oscillation measurements on detwinned single crystals. Phys. Rev. Lett. 107, 176402 (2011).
 31.
Shishido, H. et al. Evolution of the Fermi surface of BaFe_{2}(As_{1–x}P_{x})_{2} on entering the superconducting dome. Phys. Rev. Lett. 104, 057008 (2010).
 32.
Wiesmann, H. et al. Simple model for characterizing the electrical resistivity in A15 superconductors. Phys. Rev. Lett. 38, 782 (1977).
 33.
Gunnarsson, O., Calandra, M. & Han, J. E. Colloquium: Saturation of electrical resistivity. Rev. Mod. Phys. 75, 1085–1099 (2003).
 34.
Hussey, N. E., Takenaka, K. & Takagi, H. Universality of the MottIoffeRegel limit in metals. Phil. Mag. 84, 2847–2864 (2004).
 35.
Deng, X. et al. How bad metals turn good: Spectroscopic signatures of resilient quasiparticles. Phys. Rev. Lett. 110, 086401 (2013).
 36.
Cooper, R. A. et al. Anomalous criticality in the electrical resistivity of La_{2–x}Sr_{x}CuO_{4}. Science 323, 603–607 (2009).
 37.
DoironLeyraud, N. et al. Correlation between linear resistivity and T_{c} in the Bechgaard salts and the pnictide superconductor Ba(Fe_{1–x}Co_{x})_{2}As_{2}. Phys. Rev. B 80, 214531 (2009).
 38.
Ishida, S. et al. Anisotropy of the inplane resistivity of underdoped Ba(Fe_{1–x}Co_{x})_{2}As_{2} superconductors induced by impurity scattering in the antiferromagnetic orthorhombic phase. Phys. Rev. Lett. 110, 207001 (2013).
 39.
Ding, H. et al. Observation of Fermisurfacedependent nodeless superconducting gaps in Ba_{0.6}K_{0.4}Fe_{2}As_{2}. Europhys. Lett. 83, 47001 (2008).
 40.
Fukazawa, H. et al. Possible multiple gap superconductivity with line nodes in heavily holedoped superconductor KFe_{2}As_{2} studied by ^{75}As nuclear quadrupole resonance and specific heat. J. Phys. Soc. Jpn. 78, 083712 (2009).
 41.
Okazaki, K. et al. Octetline node structure of superconducting order parameter in KFe_{2}As_{2}. Science 337, 1314–1317 (2012).
 42.
Zaanen, J. Why the temperature is high. Nature 430, 512–513 (2004).
 43.
Hashimoto, K. et al. A sharp peak of the zerotemperature penetration depth at optimal composition in BaFe_{2}(As_{1–x}P_{x})_{2}. Science 336, 1554–1557 (2012).
 44.
Nakajima, M. et al. Effect of Co doping on the inplane anisotropy in the optical spectrum of underdoped Ba(Fe_{1–x}Co_{x})_{2}As_{2}. Phys. Rev. Lett. 109, 217003 (2012).
 45.
Nakajima, M. et al. Growth of BaFe_{2}(As_{1–x}P_{x})_{2} single crystals (0 ≤ x ≤ 1) by Ba_{2}As_{3}/Ba_{2}P_{3}flux method. J. Phys. Soc. Jpn. 81, 104710 (2012).
 46.
Kihou, K. et al. Single crystal growth and characterization of the ironbased superconductor KFe_{2}As_{2} synthesized by KAs flux method. J. Phys. Soc. Jpn. 79, 124713 (2010).
Acknowledgements
M.N. and S.I. thank the Japan Society for the Promotion of Science (JSPS) for the financial support. Discussions with T. Misawa, M. Imada, and A. Fujimori were helpful in preparing this manuscript. This work was supported by the JapanChinaKorea A3 Foresight Program and GrantinAid for JSPS Fellows from JSPS, GrantinAid for Scientific Research from JSPS and the Ministry of Education, Culture, Sports, Science, and Technology, Japan, and the Strategic International Collaborative Research Program (SICORP) from Japan Science and Technology Agency.
Author information
Author notes
 M. Nakajima
Current address: Department of Physics, Osaka University, Osaka 5600043, Japan.
Affiliations
Department of Physics, University of Tokyo, Tokyo 1130033, Japan
 M. Nakajima
 , S. Ishida
 , T. Tanaka
 , T. Kakeshita
 & S. Uchida
National Institute of Advanced Industrial Science and Technology, Tsukuba 3058568, Japan
 M. Nakajima
 , S. Ishida
 , K. Kihou
 , Y. Tomioka
 , C. H. Lee
 , A. Iyo
 , T. Ito
 & H. Eisaki
JST, Transformative ResearchProject on Iron Pnictides (TRIP), Tokyo 1020075, Japan
 M. Nakajima
 , S. Ishida
 , T. Tanaka
 , K. Kihou
 , Y. Tomioka
 , C. H. Lee
 , H. Fukazawa
 , Y. Kohori
 , T. Kakeshita
 , A. Iyo
 , T. Ito
 , H. Eisaki
 & S. Uchida
Department of Physics, Chiba University, Chiba 2638522, Japan
 T. Saito
 , H. Fukazawa
 & Y. Kohori
Authors
Search for M. Nakajima in:
Search for S. Ishida in:
Search for T. Tanaka in:
Search for K. Kihou in:
Search for Y. Tomioka in:
Search for T. Saito in:
Search for C. H. Lee in:
Search for H. Fukazawa in:
Search for Y. Kohori in:
Search for T. Kakeshita in:
Search for A. Iyo in:
Search for T. Ito in:
Search for H. Eisaki in:
Search for S. Uchida in:
Contributions
M.N. and T.T. conducted the optical spectroscopy measurements and analysed the data. M.N., K.K., T.S., C.H.L., H.F., Y.K. and A.I. synthesised the single crystals. M.N., S.I., Y.T., T.K. and T.I. carried out the resistivity measurements. H.E. and S.U. designed and coordinated the experiment. All authors contributed to and discussed the manuscript.
Competing interests
The authors declare no competing financial interests.
Corresponding author
Correspondence to M. Nakajima.
Rights and permissions
This work is licensed under a Creative Commons AttributionNonCommercialNoDerivs 4.0 International License. The images or other third party material in this article are included in the article's Creative Commons license, unless indicated otherwise in the credit line; if the material is not included under the Creative Commons license, users will need to obtain permission from the license holder in order to reproduce the material. To view a copy of this license, visit http://creativecommons.org/licenses/byncnd/4.0/
To obtain permission to reuse content from this article visit RightsLink.
About this article
Further reading

Effect of Cr substitution for V in Sr2VFeAsO3
Superconductor Science and Technology (2019)

Pressureinduced enhancement of superconductivity and quantum criticality in the 12442type hybridstructure superconductor KCa2Fe4As4F2
Physical Review B (2019)

Comprehensive study of outofplane transport properties in BaFe2As2 : Threedimensional electronic state and effect of chemical substitution
Physical Review B (2018)

Thermal annealing and pressure effects on BaFe2−x Co x As2 single crystals
Journal of Physics: Condensed Matter (2018)

Optical properties of Ba 0.6 K 0.4 Fe 2 As 2 thin film prepared by pulsed laser deposition and subsequent postannealing process
Current Applied Physics (2017)
Comments
By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.