Antiferromagnetic Kondo lattice compound CePt3P

A new ternary platinum phosphide CePt3P was synthesized and characterized by means of magnetic, thermodynamic and transport measurements. The compound crystallizes in an antiperovskite tetragonal structure similar to that in the canonical family of platinum-based superconductors APt3P (A = Sr, Ca, La) and closely related to the noncentrosymmetric heavy fermion superconductor CePt3Si. In contrast to all the superconducting counterparts, however, no superconductivity is observed in CePt3P down to 0.5 K. Instead, CePt3P displays a coexistence of antiferromagnetic ordering, Kondo effect and crystalline electric field effect. A field-induced spin-flop transition is observed below the magnetic ordering temperature TN1 of 3.0 K while the Kondo temperature is of similar magnitude as TN1. The obtained Sommerfeld coefficient of electronic specific heat is γCe = 86 mJ/mol·K2 indicating that CePt3P is a moderately correlated antiferromagnetic Kondo lattice compound.

The interplay among spin, charge and orbital degrees of freedom in transition metal compounds has triggered enormous research interests in condensed matter physics and material science. For a large family of layered 3d electron superconductors (SCs) such as the copper oxides 1 and iron pnictides 2 , the spin fluctuations caused by strong 3d electron correlations play a vital role in the unconventional superconductivity. Besides these 3d transition metal systems, several platinum-based SCs exhibit remarkably rich physical properties and therefore have also attracted considerable attention, partly owing to the moderately strong spin-orbit coupling of the platinum 5d electrons. The most prominent example is the heavy fermion noncentrosymmetric (NCS) SC CePt 3 Si, in which exotic superconductivity is observed below T c = 0.75 K 3 : an admixture of spin-singlet and spin-triplet pairing symmetry, nodal gap structure and huge upper critical field (B c2 ≈ 4 T) 4 . The delicate interplay between the cerium 4f and the platinum 5d electrons places this material on the border of the magnetic quantum critical point (QCP) but still in the antiferromagnetic (AFM) gound state, rendering the role of inversion symmetry unclear 5 . Among a series of filled skutterudite MT 4 X 12 (M = rare-earth or alkaline-earth metals, T = transition metals and X = P, As, Sb and Ge) with the cubic space group Im3 (No. 204), PrPt 4 Ge 12 was reported to exhibit time-reversal symmetry breaking from zero-field μSR measurements 6 . As a result of the unexpectedly high transition temperature T c = 7.9 K and the moderately enhanced Sommerfeld coefficient γ = 76 mJ/mol ⋅ K 2 , PrPt 4 Ge 12 has been extensively studied and multiband superconductivity has been proposed based on the analysis of the photoemission spectroscopy 7 as well as the magnetic penetration depth 8 . Moreover, SrPtAs is recently reported to crystallize in a hexagonal structure (P6 3 /mmc, No. 194) with weakly coupled PtAs layers forming a honeycomb lattice 9 . The peculiar locally NCS structure within PtAs layer together with a strong spin-orbit coupling demonstrates SrPtAs as an attractive material to explore superconductivity with a spontaneous static magnetic field B s 10 .
It is interesting that among the platinum-based superconductors, the newly reported family of APt 3 P (A = Ca, Sr and La) shares the structural similarity with that of iron pnictides 11 . These compounds crystallize in a tetragonal structure with space group P4/nmm (No. 129) with stacking in the order of A-Pt 6 P-A along the c-axis. The distorted antiperovskite Pt 6 P octahedral unit alternates within the ab plane, forming an antipolar pattern. The z → − z inversion operation is thus preserved. Due to the structural distortion, the platinum atoms take two different sites as Pt(I) and Pt(II) so that the Pt(II) and P atoms form a Pt 2 P layer resembling the FeAs layer in the iron-based superconductors. Of course, the structure of APt 3 P is also somewhat similar to that of CePt 3 Si, but the latter is actually isotypic to the NCS compound CePt 3 B with the space group P4mm (No. 99) 3 . The corresponding Pt 6 Si unit has the polar structure under this space group leading to the absence of inversion symmetry, different from the antipolar structure in APt 3 P. Noticeably, the APt 3 P family shows a significant variation of T c , i.e., T c = 8.4 K, 6.6 K and 1.5 K for A = Sr, Ca and La, respectively. It was reported theoretically that spin-orbit coupling (SOC) effect is significant in LaPt 3 P but negligible in CaPt 3 P and SrPt 3 P 12-14 . The origin of significantly enhanced T c in SrPt 3 P is still debatable. It was suggested to be due to a possible dynamic charge-density-wave (CDW) 12 . However, a theoretical work by Zocco et al. indicated SOC could strongly renormalize the electron-phonon coupling of SrPt 3 P and thus enhance the electronic density of states near the Fermi level 15 . Moreover, several theoretical works claimed that the CDW instability could not be reproduced in SrPt 3 P 13,14 . The centrosymmetric (CS) compounds APt 3 P reported so far do not involve the 4f electrons. The interplay between strong 4f electron correlation and superconductivity of 5d electrons in the APt 3 P family remains an open issue.
In this paper, we report our successful synthesis of such a candidate compound CePt 3 P in the platinum-based phosphides APt 3 P family. We performed systematic measurements of the physical properties including the magnetic susceptibility, magnetization, specific heat and electrical resistivity. However, no evidence of superconductivity is observed down to 0.5 K in CePt 3 P, in contrast to other APt 3 P compounds. Instead, the compound displays the rich physics involving the coexistence of magnetic ordering, Kondo coherence as well as crystalline electric field (CEF) effect. We shall discuss these properties and highlight the delicate 4f-5d interplay in this system. Figure 1 shows the Rietveld refinement of the XRD pattern of polycrystalline CePt 3 P samples. Almost all peaks can be well indexed with the tetragonal structure with the space group P4/nmm (No. 129), except for a tiny peak of an impurity phase around 31.4° which might be PtP 2 . The result of the Rietveld refinement 16 shows a good convergence: R wp = 13.4%, S = 3.3. The refined lattice parameters of CePt 3 P are a = 5.7123(7) Å and c = 5.4679(6) Å as listed in Table 1. The room temperature XRD patterns of LaPt 3 P are also refined with R wp = 14.9%, S = 2.7  (data not shown). The refined lattice parameters of LaPt 3 P are a = 5.7597(3) Å and c = 5.4736(3) Å. For comparison, the lattice parameters of the other APt 3 P compounds are also provided in Table 1. One can see obviously that a of CePt 3 P is smaller, while c is larger, compared with the lattice parameters of SrPt 3 P. Due to the lanthanide contraction, both of a and c of CePt 3 P are smaller than those of LaPt 3 P. From the EDS measurements, the molar ratio is Ce:Pt:P = (1.0 ± 0.1):(3.2 ± 0.2):(0.7 ± 0.2) for CePt 3 P and La:Pt:P = (1.0 ± 0.1):(2.6 ± 0.1): (0.8 ± 0.1) for LaPt 3 P. The actual chemical compositions are close to the nominal ones, while there seems a deficiency on the P site for both CePt 3 P and LaPt 3 P. The temperature-dependent molar magnetic susceptibility χ(T) = M/H and inverse magnetic susceptibility 1/χ(T) of CePt 3 P measured at H = 1000 Oe are presented in Fig. 2(a). χ(T) obeys a modified Curie-Weiss law above 200 K, χ = χ 0 + /(T − θ). χ 0 is a temperature independent susceptibility from the core diamagnetism, the van Vleck and Pauli paramagnetism,  is the Curie constant and θ is the Weiss temperature. The relatively large absolute value of θ = − 28.3 K may be attributed to the hybridization of the 4f electronic states with the conduction band 17 . The derived effective moment μ eff = 2.52μ B is almost equal to that of a free Ce 3+ ion, indicating the trivalent Ce ion and well localized moment of Ce-4f 1 electrons at high temperature. χ 0 is in the magnitude order of 10 −3 . For T < 100 K, a change of the slope of 1/χ(T) can be clearly observed and the fitting parameters are μ eff = 2.11μ B , and θ = − 15.3 K. Here the change of the slope and the decreased value of μ eff can be ascribed to the CEF effect. With decreasing temperature, χ(T) increases and shows a round peak around 3.0 K. Upon further cooling, another anomaly is observed near our base temperature. Two magnetic transition temperatures are determined from the peaks of derivative susceptibility Tdχ/dT as T N1 = 3.0 K and T N2 = 1.9 K (seen from Fig. 2(b)). Considering the negative Weiss temperature, the first anomaly marks the AFM ordering below T N1 which is compatible with the magnetization measurement (discussed below). While the second anomaly is attributed to a spin-reorientation. A similar phenomenon was observed in CeNiAsO 18 . Further experimental studies, especially neutron diffraction measurement on single crystals of CePt 3 P, are necessary to clarify the magnetic structure at low temperature.

Results and Discussion
The isothermal magnetization M(B) of CePt 3 P, measured in the B-sweep mode containing both field-up and down loops, is displayed in Fig. 3(a). In the AFM ordering state, M(B) displays a linear field dependence when B < 2.0 T, but undergoes a weak step-like increase around 3.0 T. This anomaly, which is ascribed to a field-induced metamagnetic transition (MMT), can be independently determined to be B m = 3.0 T by the peak in dM/dB curve (inset to Fig. 3(a)) and the hump in ρ(B) curve ( Fig. 3(b)) measured at T = 2 K. The expected hysteresis around B m is not observed and such absence of hysteresis around MMT was also reported in the single-crystalline samples CeAuSb 2 19 and YbNiSi 3 20 . No hysteresis in resistivity is observed for CePt 3 P in this magnetic field range either. Note that the M(B) curve does not show a saturation trend in the highest field limit and the value M ~ 0.6μ B at B = 5 T is much lower than the theoretical value of 2.14μ B for the saturated moment of free Ce 3+ ions which is probably due to the CEF effect. Figure 3(b) shows the isothermal resistivity versus the applied field. ρ decreases monotonously with increasing magnetic field at T = 6 K > T N1 . Whereas at T = 2 K < T N1 , a hump around B m = 3.0 T is added to the decreasing trend. This feature is compatible with the MMT observed in the magnetization measurement.
The specific heats of CePt 3 P and LaPt 3 P divided by T, C(T)/T, are plotted in the main panel of Fig. 4(a) in a semi-logarithm scale. At room temperature, C(T) saturates to about 135 and 140 J/mol ⋅ K for La and Ce compound, respectively, which are, within an acceptable error range, compatible with the classical Dulong-Petit law 3NR with N = 5 and R = 8.31 J/mol ⋅ K, where R is the universal gas constant. The specific heat C(T) of LaPt 3 P is typical for nonmagnetic metals since no typical anomaly can be observed at high temperature. At low temperature, the specific heat of LaPt 3 P is dominated by the electronic and phonon contributions for T < Θ D /10, therefore, it can be fitted to a power law C/T = γ La + β La T 2 over 10-20 K (data not shown). Here Θ D is the Debye temperature, and γ La and β La denote the coefficients of the electronic and phonon contributions, respectively. It should be noted that there is a small jump around 1 K in the specific heat of LaPt 3 P which should correspond to a superconducting transition though it is too small to observe in Fig. 4.
In the paramagnetic region above the magnetic transition, the specific heat of CePt 3 P can be expressed as where the coefficients γ Ce and β Ce are of electronic and phonon contributions of CePt 3 P, respectively, while C Sch describes the Schottky anomaly item. A linear T 2 -dependence is clearly seen in C/T vs T 2 plot for temperature below 20 K (see inset to Fig. 4(a)). The derived Sommerfeld coefficient is γ Ce = (86 ± 1) mJ/mol ⋅ K 2 . The value is moderately enhanced by a factor of 57 compared with that of LaPt 3 P where γ La = (1.5 ± 0.1) mJ/mol ⋅ K 2 , manifesting the correlation effect contributed from the Ce-4f electrons. Therefore, CePt 3 P is a Kondo lattice compound due to the strong 4f electron correlation and moderate effective 4f− 5d hybridization. Note that γ La for LaPt 3 P derived here is slightly smaller but still in the same magnitude order with that obtained in ref. 11. The reported phonon coefficients are in reasonable agreement with each other: β Ce = 0.98(1) mJ/mol ⋅ K 4 for CePt 3 P and β La = 0.94(1) mJ/mol ⋅ K 4 for LaPt 3 P, indicating similar phonon contributions. The Debye temperature Θ D estimated by using Θ D = (12π 4 NR/5β) 1/3 is (215 ± 1) K for CePt 3 P and (218 ± 1) K for LaPt 3 P, implying that the above analysis is quite self-consistent.
The Ce-4f contribution to the specific heat of CePt 3 P is then deduced by subtracting the measured specific heat of the nonmagnetic isostructural reference sample LaPt 3 P from the total specific heat of CePt 3 P, i.e., C 4f = C Ce − C La . The result is shown in the main panel of Fig. 4(b), plotted as C 4f /T vs T in a logarithmic scale. The Schottky anomaly, which is visible as a broad peak centered around 90 K in C 4f /T curve, should be caused by the excitations between different CEF levels. The Schottky anomaly with three Kramers doublets (one doublet ground state and two excited doublets) for Ce 3+ ion with j = 5/2 experiencing a tetragonal crystal-field potential can be expressed by refs 21,22  Here g i = 2 is the degeneracy of the ith doublet state and Δ i is the energy difference between the ground state and the i-th excited state (see the schematic sketch drawn in the inset of Fig. 4(b)). Eq. 2 is applied to C 4f /T of CePt 3 P over a temperature range of 50-130 K. The derived CEF energy differences are Δ 1 = (20.9 ± 0.1) meV (~(240 ± 1) K) and Δ 2 = (60.9 ± 0.3) meV (~(700 ± 3) K). This result may explain the slope change in 1/χ(T) curve as well as the broad hump in both ρ mag and S. Furthermore, the large value of Δ 1 is consistent with the reduced effective Ce moment below 100 K. The magnetic entropy gain S m is calculated by integrating C 4f /T over T and plotted on the right axis in Fig. 4(b). One can see that S m reaches about 0.51Rln2 at T N1 and Rln2 is recovered at ~50 K, indicating that the ground state with the AFM ordering of Ce 3+ moments is Kramers two-fold degenerate. The plateau over the temperature range of T = 10-30 K indicates that the first excited CEF level is far above T N1 . S m reaches Rln4 at ~150 K and increases substantially above the Schottky anomaly. For a Kondo lattice, the Kondo temperature can be estimated by the magnetic entropy at T N via ref. 23 m where ξ = T K /T N . The yielded T K is about (6.1 ± 0.1) K for CePt 3 P. At low temperature, C 4f /T shows a pronounced λ-shape peak at T N1 = 3.0 K, implying a second-order phase transition. The expected jump in specific heat is δ ~6 J/mol ⋅ K. A slight slope change in C 4f /T is also observed around T N2 = 1.9 K, consistent with the low-temperature anomaly observed in aforepresented χ(T) curve. Based on the mean-field theory of Besnus et al. 24 and Bredl et al. 25 , the specific heat jump δ = C T T N is related to the Kondo temperature T K by the following formula Here ζ = (T K /T N )/2π, ψ denotes the digamma function and ψ′ , ψ′ ′ and ψ′ ′ ′ are the first three derivatives of ψ.
Then the Kondo temperature can be also estimated by applying Eq. 4, obtaining a ratio of T K /T N1 = 0.88, or T K ~ (2.7 ± 0.1) K. Therefore, based on both magnetic entropy and specific heat jump, it is reasonable to estimate T K ~ 2-6 K in this compound. In the magnetically ordered state, the AFM spin-wave spectrum follows a dispersion relation of ε k = ∆ + Dk 2 2 . Here ε k is the excitation energy, Δ is the gap in the spin-wave spectrum, and D is the spin-wave stiffness. The phonon contribution, β Ce T 3 item, can be subtracted from the total specific heat C as Δ C = C − β Ce T 3 . At low temperature, Δ C is described by the following expression 26,27 : Fitting the specific heat below T N2 (solid line in Fig. 4(a)) gives the fitting parameters γ 0 = 247 mJ/mol ⋅ K 2 , Δ = 2.6 K, and A C = 67.5 mJ/mol ⋅ K 4 . The considerably enhanced zero-temperature Sommerfeld coefficient γ 0 is about 3 times of γ Ce obtained in the paramagnetic state, indicating the formation of moderate-heavy quasiparticles in the antiferromagnetically ordered state. It is worthwile noting that the obtained spin-wave gap Δ is of the order of magnitude often found in cerium intermetallics with AFM ground states 17 . The temperature variation of the electrical resistivity of CePt 3 P, ρ(T), is plotted in Fig. 5(a). The resistivity at room temperature is ρ 300K = 1140 μΩ ⋅ cm, a value rather typical for the Ce-based Kondo compounds with narrow f-band 28 . The resistivity decreases with decreasing temperature and exhibits two features. A broad hump around 110 K reflects the 4f-electron contribution via Kondo scattering from different CEF levels 21,22 . At low temperature, a pronounced peak in ρ(T) around 3 K is directly visible, indicating the AFM ordering phase below T N1 = 3.0 K. Above T N1 , ρ increases in a minus logarithmic temperature manner over T = 5-20 K, reflecting the Kondo-type scattering. Further evaluation of ρ(T) requires information of the phonon contribution which could be taken from the homologous and isostructural analog, LaPt 3 P. The ρ(T) of LaPt 3 P, which is also presented in Fig. 5(a), can be well described by a Bloch-Grüeneisen-Mott (BGM) relation: where ρ 0 is the residual resistivity due to lattice defects, the second term denotes electron-phonon scattering, and the third one accounts for the contribution due to Mott's s-d interband electron scattering. A least square fitting of the BGM formula to the experimental data over the temperature range 2-300 K leads to the following parameters: ρ 0 = 32 μΩ ⋅ m, Θ R = 160 K, R = 1.25 μΩ ⋅ cm/K, and K = 4.1 × 10 −8 μΩ ⋅ cm/K 3 . Note that the residual resistivity ρ 0 is smaller than that in ref. 11. The parameter Θ R is usually considered as an approximation of the Debye temperature Θ D in spite of some contribution due to electron-electron correlations in Θ R

29
. Θ D yielded from the specific heat data is 218 K which is in accordance with Θ R from the resistivity data. LaPt 3 P exhibits simple metallic behavior as we expected, without the characteristic features due to the interplay of Kondo and CEF effects in CePt 3 P mentioned above.
In order to analyze the magnetic contribution to the electrical resistivity of CePt 3 P, it is reasonable to assume that the phonon contribution in this compound can be properly approximated by that in LaPt 3 P, ρ ph = ρ(La) − ρ 0 (La), so we have The temperature dependence of ρ mag + ρ 0 derived in this way is presented in Fig. 5(b) in a semilogarithmic scale. As a distinct feature in a Kondo lattice system, a pronounced broad hump centered at T * = 110 K become obvious in ρ mag curve, which could be ascribed to the Kondo scattering from different CEF levels. According to Cornut and Coqblin 21 , this maximum provides an estimate of the CEF splitting energy scale ~200 K of Ce-4f 1 state with j = 5/2. On the other hand, as temperature is decreased, ρ mag increases in a logarithmic scale, as shown as the dotted lines in Fig. 5(b) above T > 200 K and between 5-20 K, respectively. Following the theoretical predictions of Cornut and Coqblin 21 , the logarithmic slopes c K LT and c K HT in the low-temperature and high-temperature regions, respectively, are proportional to the squared effective degeneracy λ of the thermally populated levels: c K ∝ λ 2 − 1. For cerium compouns with Ce 3+ ion placed in a noncubic crystalline environment the ground multiplet splits into three doublets, thus the expected ratio is c K LT : c K HT = 3:35. In the case of CePt 3 P, with the coefficients c K LT = − 0.063 and c K HT = − 0.57 yielded from linear fitting of ρ mag vs logT (see the dashed lines in Fig. 5(b)), the ratio is about 3:27, reasonably close to the theoretical prediction.
From the inset of Fig. 5(a), ρ drops rapidly below about 3.0 K owing to the reduction of spin-flip scattering upon entering the AFM ordered state. This magnetic transition temperature is determined from a slope change of dρ/dT in Fig. 2(b). Upon further cooling, a second slope change in ρ is observed around 1.9 K, corresponding to the pronounced kink in dρ/dT. Therefore, two magnetic transitions in CePt 3 P are apparent from the analysis of magnetic susceptibility χ(T), specific heat C(T) and electrical resistivity ρ(T), as shown in Fig. 2(b): the first transition T N1 corresponds to the AFM ordering temperature, while the second one T N2 is presumably associated with the spin reorientation. The values of T N1 and T N2 derived from different measurements agree well with each other. It is noted that while LaPt 3 P shows superconductivity around T c = 1.0 K (from specific heat), no superconductivity is observed in CePt 3 P down to 0.5 K.
Considering the relativistic dispersion relation for the AFM magnon spectrum, the electrical resistivity ρ(T) for T < Δ can be well described by the following equations 26,27 : where ρ 0 is the temperature-independent residual resistivity, the constant coefficient B ρ is related to the spin-wave stiffness D by the proportionality D −3/2 and Δ is the same gap in the spin-wave spectrum as in Eq. (5). AT 2 stems from the electron-electron scattering following the Fermi liquid theory, while the third term describes the electron-magnon scattering. This formula is applied to the electrical resistivity of CePt 3 P (dotted line in the inset of Fig. 5(a)) and a very good fit is obtained with the fitting parameters: ρ 0 = 688 μΩ ⋅ cm, Δ = 4.0 K, A = 9.0 μΩ ⋅ cm/K 2 and B ρ = 25 μΩ ⋅ cm/K 2 . Considering the relatively short fitting range of temperature, the derived Δ value for the measured polycrystalline sample is still reasonably compared with that obtained from the specific heat data. Based on the above analyses, CePt 3 P displays the coexistence of three important characteristics: AFM ordering of the cerium local moments due to the Ruderman-Kittel-Kasuya-Yosida exchange interaction, the Kondo effect due to the strong 4f electron correlation and moderate effective 4f− 5d hybridization, and the CEF interactions. The AFM ordering at T N1 = 3.0 K is clearly identified by the pronounced anomalies in the temperature-dependent magnetic, thermodynamic and electrical measurements. In addition, another anomaly at T N2 = 1.9 K is also visible from the physical properties, and is probably due to a change in the magnetic configuration within the AFM ordered phase. The behavior of ρ(T) and C(T) in the ordered region is well describable in terms of AFM spin-wave spectrum. The field-dependent behavior of the magnetization and electrical resistivity also indicates a MMT from the magnetic ordering to a spin-polarized state around B m = 3.0 T. The magnetic structure of CePt 3 P is still unclear and the neutron diffraction or Mössbauer spectroscopy experiments are helpful to clarify the details of the magnetic structure.
Scientific RepoRts | 7:41853 | DOI: 10.1038/srep41853 The Kondo effect displays itself by the large value of Weiss temperature θ (compared with the ordering temperature), the reduced magnetic entropy and the specific heat jump at T N , as well as the enhanced Sommerfeld coefficient γ Ce . From the analysis of the specific heat data, the Kondo temperature T K is estimated to be in the range of 2-6 K. Its value can be also estimated from the magnetic susceptibility as T K ~ |θ|/4  7.1 K 30 , in reasonable agreement with other estimates. Also, the Kondo effect is well manifested in the electrical resistivity for Kondo systems with strong CEF interactions which follows the negative logarithmic-temperature dependence as ρ(T) = ρ 0 + c T ln k , with Kondo coefficient c k < 0 21 . The inverse susceptibility (1/χ(T)) curve shows a slope change between T = 100-200 K which is also attributed to the CEF effect. This temperature region is in accordance with the energy scale Δ 1 = 240 K of the multiplet Ce 3+ ion estimated from the Schottky contributions of the specific heat 21,22 .
Finally, it is very interesting to compare this CS compound CePt 3 P with the extensively studied NCS heavy fermion SC CePt 3 Si (T c = 0.75 K) 3 . The crystal structure of CePt 3 P consists of alternative stacking of layers of Ce atoms and layers of distorted antiperovskite Pt 6 P octahedral units along the c-axis. The Pt 6 P octahedra is asymmetrically distorted perpendicular to the ab-plane but alternatively distributed in the ab-plane, resulting in a symmetric antipolar analogue of CePt 3 Si. CePt 3 Si shows antisymmetric spin-orbit coupling of the platinum 5d electrons due to the absence of z → − z symmetry as well as mixing spin-singlet and spin-triplet pairing states. The parity mixing alone can hardly account for the heavy fermion phenomena unless the strong electron-electron correlation effects which are ensured by the presence of Ce 3+ ions are taken into consideration together 31 . Correspondingly, the suppression of superconductivity in CePt 3 P may be attributed to the enhanced AFM ordering. CePt 3 P is, therefore, probably placed further away from the magnetic QCP compared with CePt 3 Si (T N = 2.2 K). With an external control parameter δ, such as doping or positive pressure, the system may be shifted towards T N = 0, namely the QCP 32,33 . It is thus of great interest to investigate whether superconductivity exists in CePt 3 P at even lower temperature than 0.5 K; if superconductivity does exist, it will provide strong evidence for the proximity to a magnetic QCP in CePt 3 P. Comparing with CePt 3 P, the occurrence of superconductivity at T c = 0.75 K in CePt 3 Si implies that the NCS crystal structure may favor unconventional superconductivity within the AFM ground state.

Conclusion
In summary, we report the successful synthesis of a new compound CePt 3 P. From the collected experimental data of magnetization, specific heat and transport measurements, this compound is characterized as an antiferromagnetic Kondo lattice with crystal electric field effect. Two successive magnetic transitions of Ce 4f moments are observed: the magnetic ordering at T N1 = 3.0 K and the spin reorientation at T N2 = 1.9 K. Considering the moderately enhanced Sommerfeld coefficient of γ Ce = 86 mJ/mol ⋅ K 2 in the paramagnetic region and large value of γ 0 = 247 mJ/mol ⋅ K 2 in the the AFM region, the Kondo effect and the AFM order should coexist in the ground state. Thus a relatively large Fermi surface formed by the heavy quasiparticles is expected in CePt 3 P with a Kondo temperature T K ~ 2-6 K. The ab initio crystal-field and electronic band structure calculations are necessary to further complement the present results. Further experiments such as chemical doping are presently underway in order to tune the ground state from the AFM ordering to strongly-correlated paramagnetic region.

Experimental Methods
The polycrystalline sample of CePt 3 P was synthesized by solid state reaction. Ce piece (99.8%), Pt powder (99.9%) and P lump (99.999%) of high purity from Alfa Aesar were used as starting materials. Firstly, CeP was pre-synthesized by reacting Ce and P at 1173 K for 72 h. Secondly, powders of CeP and Pt were weighed according to the stoichiometric ratio, thoroughly ground and pressed into pellets. The pellets were then packed in Al 2 O 3 crucibles and sealed in an evacuated quartz tube which were slowly heated to 1273 K and kept at that temperature for 7 days. Finally, the samples were thoroughly ground, cold pressed and annealed in vacuum to improve the sample homogeneity. For comparison, the polycrystalline sample LaPt 3 P was also synthesized in the similar process. All the preparation procedures except heating were carried out in an argon protected glove box with the water and oxygen content below 0.1 ppm. The obtained CePt 3 P sample is less compact than LaPt 3 P and both of them are quite stable in the air. Powder x-ray diffraction (XRD) measurements at room temperature were carried out on a PANalytical x-ray diffractometer (Model EMPYREAN) with a monochromatic Cu K α1 radiation and a graphite monochromator. Lattice parameters were derived by Rietveld refinement using the program RIETAN 2000 16 . The energy dispersion x-ray spectroscopy (EDS) analysis was performed on a EDS spectrometer affiliated to a field emission scanning electron microscope (FEI Model SIRION). The electron beam was focused on a crystalline grain and the chemical compositions were averaged on at least 4 EDS spectra from different grains. The electrical resistivity ρ(T) was measured by the standard four-probe method in a Quantum Design physical property measurement system (PPMS-9). The dc magnetization was measured in a Quantum Design magnetic property measurement system (MPMS-5) with the temperature range of T = 2-400 K. The specific heat measurements were performed in the PPMS-9 down to about 0.5 K.