Enhanced Thermoelectric Properties in a New Silicon Crystal Si24

Jul 6, 2018 - Thermoelectric device is a promising next-generation energy solution owing to its capability to transform waste heat into useful electri...
0 downloads 0 Views 12MB Size
Subscriber access provided by University of Winnipeg Library

Communication

Enhanced thermoelectric properties in a new silicon crystal Si with intrinsic nanoscale porous structure 24

Kisung Chae, Seoung-Hun Kang, Seon-Myeong Choi, Duck Young Kim, and Young-Woo Son Nano Lett., Just Accepted Manuscript • DOI: 10.1021/acs.nanolett.8b01176 • Publication Date (Web): 06 Jul 2018 Downloaded from http://pubs.acs.org on July 7, 2018

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 40 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

Enhanced Thermoelectric Properties in a New Silicon Crystal Si24 with Intrinsic Nanoscale Porous Structure Kisung Chae,†,§ Seoung-Hun Kang,†,§ Seon-Myeong Choi,†,k Duck Young Kim,∗,‡,¶ and Young-Woo Son∗,† †Korea Institute for Advanced Study, Seoul 02455, South Korea ‡Center for High Pressure Science & Technology Advanced Research, Shanghai 201203, P. R. China ¶Division of Advanced Nuclear Engineering, POSTECH, Pohang 37673, South Korea §Contributed equally to this work kCurrent address: Samsung Advanced Institute of Technology, Suwon 13595, South Korea E-mail: [email protected]; [email protected] Phone: +86-21-8017-7094; +82-2-958-3720. Fax: +82-2-958-3820 Abstract Thermoelectric device is a promising next-generation energy solution owing to its capability to transform waste heat into useful electric energy, which can be realized in materials with high electric conductivities and low thermal conductivities. A recently synthesized silicon allotrope of Si24 features highly anisotropic crystal structure with nanometre-sized regular pores. Here, based on first-principles study without any empirical parameter, we show that the slightly doped Si24 can provide an order-of-magnitude enhanced thermoelectric figure of merit at room temperature, compared with the cubic

1

ACS Paragon Plus Environment

Nano Letters 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 40

diamond phase of silicon. We ascribe the enhancement to the intrinsic nanostructure formed by the nanopore array, which effectively hinders heat conduction while electric conductivity is maintained. This can be a viable option to enhance the thermoelectric figure of merit without further forming an extrinsic nanostructure. In addition, we propose a practical strategy to further diminish the thermal conductivity without affecting electric conductivity by confining rattling guest atoms in the pores.

Keywords first-principles calculations, thermoelectric figure of merit, electron-phonon coupling, intrinsic nanostructure

Thermoelectricity is one of the renewable energy solutions that converts waste heat into viable electric energy without generating greenhouse gases. It creates electric current by thermal motion of charge carriers due to a temperature gradient. Performance of thermoelectric materials can be evaluated through dimensionless figure of merit (ZT),

ZT =

σS 2 T, κl + κe

(1)

which is determined by Seebeck coefficient (S), electric conductivity (σ) and thermal conductivities (κ) at a given temperature (T ). For κ, subscripts l and e refer to lattice and electronic contributions to heat conduction, respectively. To increase the ZT value, a material needs to be a good electrical conductor as well as a good heat insulator as seen in Eq. (1). Such a conflicting condition has recently been realized in a layered crystalline material 1,2 with strong bond anharmonicity, which can be measured by Grüneisen parameter and be a fingerprint for potential thermoelectric materials. However, it is still challenging to achieve a high ZT in three-dimensional (3D) crystals because the materials properties listed above are coupled to each other. For instance, 2

ACS Paragon Plus Environment

Page 3 of 40 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

increasing a charge carrier in 3D semiconductor by doping would increase both σ and κe simultaneously. Also, lifetimes of both charge carriers and phonons in a heavily doped material will be significantly lowered due to increased scattering. In order to overcome this complication, various efforts have been made to reduce the κl by post-processings such as alloying, 3,4 nanostructuring, 5–13 and confining rattling guest atoms in cages of skutterudites and clathrates. 14 The alloy and nanostructure, however, might lower the mobility of charge carriers by enhancing impurity/boundary scattering, and the confinement is only valid for porous materials which can host guest atoms. Aforementioned favorable conditions for a higher ZT can be found in a new silicon allotrope, Si24 , with intriguing electronic properties such as quasi-direct bandgap. 15 Specifically, Si24 shows the intrinsic nanostructure as in the case of a good thermoelectric material, SnSe 1,2 [see Figure 1(a) and (b)]. This is in contrast to the previous studies for Si with postprocessings of extrinsic nanoscale structures. 5–13 Moreover, the Si24 features regular array of nanosized pores as in skutterudites and clathrates, which makes it even more promising as a thermoelectric material. This indicates that an enhanced thermoelectric property can be achieved in a bulk Si material without post-processings, rendering thermal stability in operating conditions. We note here that Si24 is shown to be stable in a wide range of temperature 15 (∼750 K) and pressure 16 (∼8 GPa). We also point out some of the advantages of employing a Si material over other materials containing heavy metal elements (e.g., PbTe) in that Si is cheap, non-toxic, Earth-abundant and free of phase separation. Besides, due to the complex interactions involved in the transport phenomena as mentioned above, describing transport properties to a reasonable degree remains challenging. Some of the previous studies 17,18 used too simplified approximations to the electronic relaxation times. For instance, overestimated values of ZT were reported for black phosphorous obtained with a single-valued relaxation time, adapted from the experiment, 17 from a deformation potential theory 19 or from a constant relaxation time approximation. 20 When the effects of electron-phonon interactions were explicitly taken into account, the ZT values 3

ACS Paragon Plus Environment

Nano Letters 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

decreased by orders of magnitude. 21 In this Letter, we report the enhanced thermoelectric properties of the Si24 over the cubic diamond phase counterpart (dSi) by performing various first-principles calculations without using empirical relaxation times of electron and phonon in Si24 . Specifically, all of the elastic scattering events between electron and phonon with varying energies and momenta were explicitly enumerated for electric conductivity, σ. Similarly, the anharmonic effects of heat transport due to the three-phonon scattering were explicitly considered for the lattice thermal conductivity, κl . We ascribe the order-of-magnitude enhancement in the ZT to a significant anharmonicity in Si24 , which is also confirmed by computing anomalously high Grüneisen parameters. The enhanced ZT of intrinsic Si24 is shown to range from 0.14 to 0.6 at 300 to 700 K. This is still smaller than the best record of SnSe at high temperature (from 0.12 to 2.6 at 300 to 923 K), 1 but is comparable to those of post-processed dSi nanostructure. 5–13 From these, we expect that gentle manipulations of intrinsic atomic structures of Si24 such as rattling effects of guest atoms in the nanopores could enhance the thermoelectric performance further by reducing the lattice thermal conductivities, thus enabling the new silicon crystal be a fruitful playground to engineer thermoelectric properties. To describe the reliable thermoelectric properties for Si24 using first-principles calculations, we performed density functional theory calculations as implemented in Quantum Espresso package. 22 We used norm-conserving pseudopotentials 23 with the plane wave kinetic energy cutoff of 60 Ry (816.34 eV), and exchange-correlation functional of Perdew-BurkeErnzerhof 24 (PBE) within generalized gradient approximation was used. In order to evaluate the σ explicitly from first-principles calculations, the phonon-mediated σ was evaluated by using the transport Eliashberg spectral function 25,26 as implemented in EPW package. 27 The Bloch wave function calculated on a Γ-centered k/q-point mesh of 10×10×6 was interpolated to a much denser grid of 40×40×9 by using the maximally localized Wannier functions e−ph (MLWFs). 28–30 Using the MLWFs, the relaxation time (τn,k ) by electron-phonon coupling

4

ACS Paragon Plus Environment

Page 4 of 40

Page 5 of 40 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

were computed on the denser k/q grids as 1 e−ph τn,k

=

2π X k,λ 2 |g | [δ (εm,k − εn,k + ωλ,q ) (nλ,q + fm,k ) + N m,λ,q m,n δ (εm,k − εn,k − ωλ,q ) (nλ,q − fm,k + 1)] ,

(2)

where ωλ,q is the energy of a phonon with polarization λ and wave vector q in the Brillouin zone (BZ), and nλ,q and fm,k are Bose-Einstein and Fermi-Dirac distribution functions, respectively. N is the total number of k/q grid points in the full BZ. The εn,k is energy of k,λ a Bloch state at nth band at k point in the BZ. 31 The gm,n is an electron-phonon coupling

matrix element, which is computed as hΨm,k+q |∂λ,q V |Ψn,k i. Ψn,k is a Kohn-Sham wave function, and ∂λ,q V is the variation of the Kohn-sham potential for a unit displacement of e−ph e the nuclei along the phonon mode of polarization λ and wave vector q. Using the τn,k ≈ τn,k

as in Eq. (2), electron transport coefficients were calculated by using bolztrap code 32 within   P e−ph ∂fn,k e2 τn,k vα,n,k vβ,n,k − ∂ε , semi-classical Boltzmann transport equation: i.e., σαβ (ε) = N n,k

where α, β=x, y or z, and N is the number of k-points in the full Brillouin zone. The vα,n,k is a group velocity of electrons at εn,k along α. Note that other contributions for σ such as electron-electron and impurity scattering are assumed to be negligible (see SI S1). The Seebeck coefficient, S, shown in Eq. (1) was calculated from the Mott formula at the Fermi level as 2 π 2 kB T d (log (σ)) S= , 3e dε ε=εF where e and kB are the charge of an electron and Boltzmann’s constant, respectively. The electronic contribution to the thermal conductivity (κe ) was evaluated via Wiedemann-Franz law: 33 κe = L0 σT with L0 =2.44×10−8 W·ohm/K2 . The κe computed with Wiedemanne−ph e ) since diffusive Franz law is exact under the assumptions we used above (i.e., τn,k ≈ τn,k

motion of electron gas is described by Boltzmann transport equation with relaxation time approximation.

5

ACS Paragon Plus Environment

Nano Letters 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Similarly, the lattice thermal conductivity (κl ) with anharmonic effects between three phonons was computed by solving Boltzmann transport equation as implemented in phono3py code. 34 The third-order interatomic force constants were evaluated for dynamical matrix by a finite displacement method with a displacement and distance cutoff of 0.03 Å and 15 Å respectively. In a (3×3×2) supercell with 216 Si atoms, a total of 12,108 displaced configurations were generated. Hellmann-Feynman forces in each configuration was computed by first-principles calculations by using Vienna ab initio software package 35–38 with projector augmented wave method 39 and PBE exchange-correlation functional. 24 Plane wave was expanded with the kinetic cutoff energy of 400 eV, and Γ-centered 9×9×3 grid points in the BZ were used for the primitive cell of Si24 . After all the force evaluation, the κl was evaluated by integrating over q points in the full BZ of 14×14×14 grid points. The Si24 crystal features an open-channel structure along a crystallographic axis (x) as shown in Figure 1(a) and (b). Each of the channels, of which the diameter is approximately 0.54 nm, is composed of eight-membered-rings of distorted sp3 Si bonds. When compared to the dSi having the ideal sp3 bonding character with the distance of 2.35 Å and with the angle of 109.5°, those in Si24 range from 2.33 Å to 2.41 Å, and from 97.73° to 135.63°, respectively. 15 In particular, the open-channel structure is attributed to its two-step synthesis process such that a high pressure phase of Na4 Si24 compound is initially synthesized and pressurized at a hydrostatic pressure of 10 GPa, and that only loosely captured Na atoms in the cage are selectively removed from the compound at ambient pressure. 15 As a result, the structure with a regular array of one-dimensional nanopores is retained as seen in Figure 1(b). Note that the structure is highly anisotropic owing to the open channels. For instance, when comparing densities along different orientations of the crystal, Si24 shows denser atomic arrangement along the x-axis unlike the others, which significantly differs from isotropic atomic arrangement for dSi. We investigate thermoelectric properties of Si24 for various temperatures up to 700 K since thermal stability of Si24 in air was experimentally confirmed up to 750 K. 15 The ZT 6

ACS Paragon Plus Environment

Page 6 of 40

Page 7 of 40

(a)

(b)

a3

z x

a3

a2

y

a2

a1

(c)

(d) 300 400 500 600 700

0.5

K K K K K

0.6

0.3

0.3 0.2

0.1

0.1

8 4 dSi

0 −5

0

5

10

0 300

Doping concentration (×1020 cm−3 )

16 12

0.4

0.2

0 −10

Si24

0.5 ZTmax

0.4

20

0.7

Enhancement

0.6 Figure of merit, ZT

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

400

500

600

700

Temperature, T (K)

Figure 1: Ball-and-stick models of the Si24 crystal with (a) a perspective view and (b) orthogonal projection along the x-axis. (a) The primitive cell and the orthorhombic Bravais unitcell are shown as the red and black parallelepipeds, respectively. Twelve atoms in the primitive cell are shown in large blue balls, while the rest of the atoms are shown as small gray balls. (b) Continuous nanopores are shown where a one-dimensional array of Na ions (not shown) were confined for synthesis. (c) Thermoelectric figure of merit (ZT) of Si24 along xaxis are shown in continuous lines as a function of temperature, and the ZT values for dSi are also shown as dashed lines for comparison. Positive and negative doping concentrations (n) refer to that of excess electron and hole, respectively. (d) The maximum values of ZT (ZTmax ) for a varying temperature are shown as squares, of which the optimum doping concentrations (in 1020 cm−3 ) for Si24 are 0.21, 0.35, 0.44, 0.59 and 0.69 for increasing temperature; the values are 0.97, 1.81, 3.12, 3.43 and 4.03 for dSi. The enhancement (ratio of the ZTmax for Si24 to dSi) is also plotted as orange circles.

7

ACS Paragon Plus Environment

Nano Letters 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

obtained from our calculations indicates that Si24 shows superior thermoelectric properties compared to dSi, especially when lightly doped with electrons as shown in Figure 1(c). The overall ZT values are enhanced with an increasing temperature for both Si24 and dSi. On the other hand, enhancement factor, defined as a ratio of the maximum ZT (ZTmax ) between Si24 and dSi, decreases as temperature increases, which reaches ∼17 at 300 K. The optimal doping concentration for the ZTmax increases with temperature as in Figure 1(c). We note that overall values of the optimal doping concentration are much smaller for Si24 than dSi. This is advantageous because impurity scattering becomes significant for heavily doped semiconductors, of which the effects are difficult to be captured directly in the first-principles calculations. It is worth noting here a constant relaxation time approximation can lead to inaccurate ZT value. 18 As is known for other materials, 21 elaborated description of electron-phonon coupling tends to correct the overestimated ZT values obtained from simplified calculations. Likewise, the momentum dependence of relaxation time in Eq. (2) needs to be taken into account in evaluating the electrical conductivity σ to obtain a reliable ZT, specifically for a newly synthesized material. Another point to note is that evaluation of accurate σ values requires remarkably dense k/q grid points, and convergence test should be performed with caution. We confirmed that the data grid used here (40×40×9=14,400) gives reasonably converged ZT values less than 5% relative error compared to that from the most dense grid mesh we tested (50×50×9=22,500). The detailed results are provided in SI S2. To explain the enhanced ZT of Si24 , we will discuss the effects of each of the components in ZT hereafter. Firstly, we consider the effects of electronic contributions to the thermoelectric properties. We find that the electrical conductivity (σ) of Si24 is highly anisotropic (Figure 2(a)), which can be deduced from its anisotropic geometry (Figure 1(a)). Compared with the dSi, it is remarkable that the σxx for Si24 is always higher throughout the doping range considered in this study, while σyy and σzz are largely suppressed. Note that the nanopores run along the x-axis (Figure 1(b)) where the atomic arrangement is most densely 8

ACS Paragon Plus Environment

Page 8 of 40

Page 9 of 40 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

packed. In contrast, along the orientations perpendicular to the nanopore axis, electrons are more likely to be scattered by electron-phonon coupling due to the anisotropic bonding. Similarly, power factors (σS 2 ) also show the high anisotropy as observed in the electric conductivity (Figure 2(b)). In this case, however, the magnitude of the σS 2 also varies significantly due to carrier type and density. When Si24 is lightly doped by electrons, the σxx S 2 reaches the maximum, significantly exceeding that of the dSi (Figure 2(b)). For higher doping concentration, the σxx S 2 drops substantially and becomes smaller than that of dSi. For hole doping, the σS 2 in all directions are significantly diminished compared to the electron-doped cases, and become much smaller than that of dSi, except the peaks at low concentration. As the power factor is linearly proportional to the ZT, this agrees well with the above observation that ZT reaches the maximum at light electron doping as in Figure 1(c). These improved thermoelectric properties of Si24 can be ascribed to its electronic band dispersions (Figure 2(c)). With small electron doping concentration as marked in Figure 2(c), there exist multiple electron pockets (or valleys) in the BZ as can be also seen in the Fermi surface at that doping level; see the inset of Figure 2(c). This multiple valley degeneracy is also known to be responsible for the good thermoelectric performance of PbTe1−x Sex . 40 The highly anisotropic electronic structures are clearly seen in Si24 from the Fermi surface, which is consistent with the anisotropic crystal structure. Note that the relaxation times in Si24 are much longer than those in dSi, especially for low energies, or low doping levels. This agrees with the fact that high ZTmax of Si24 occurs at much lower doping concentration than that of dSi as shown in Figure 1(c). Furthermore, Si24 provides low lattice thermal conductivity (κl ) as seen in Figure 3(a), which renders even better thermoelectric performance compared to dSi. The calculated κl for Si24 is approximately 4 times lower than that of dSi along x-axis throughout the temperature range in this study. The difference becomes even greater to be ∼13 times for the κl along z-axis. If we consider electronic contribution to the thermal conductivity via Wiedemann9

ACS Paragon Plus Environment

Nano Letters

(b) 5

Power Factor, σS 2 (mW/K2 m)

Conductivity, σ (×106 Ohm−1 m−1 )

(a) σxx σyy σzz

4 3 2 1 0 −10

−5

0

5

Doping concentration (×10

(c) Energy, εn,k -εF (eV)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 10 of 40

15

10

5

0 −10

10 20

σxx S 2 σyy S 2 σzz S 2

−3

cm )

−5

0

5

Doping concentration (×10

10 20

−3

cm )

2 1

Si24 dSi

ky

Y

T

S R

0

Z kz

Γ

kx

−1 −2

Γ

Y

S

Γ Z

R

T

Γ 0

250 500 750 1000 e−ph τn,k (fs)

Figure 2: (a) Electrical conductivity tensor components and (b) power factor at 300 K with a varying doping concentration. The continuous and dashed lines are for Si24 and dSi, respectively. (c) Electronic band dispersion of Si24 from first-principles calculations (black lines) and maximally locallized Wannier functions (MLWFs) (red circles). Relaxation times e−ph (τn,k ) as in Eq. (2) of electrons at 300 K for both Si24 and dSi are also provided aside. In the inset, the Fermi surface in the Brillouin zone of Si24 at a doping concentration of 2.5×1020 cm−3 , which is marked as blue dotted line in (c). In (c), the zero energy denoted by a solid horizontal line was set to be middle between conduction and valence band edges.

10

ACS Paragon Plus Environment

Page 11 of 40 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

Franz law, the anisotropy in heat conduction becomes even greater due to higher electrical conductivity along the x-axis compared to the z-axis. We show the Grüneisen parameter for both crystals in Figure 3(b) and (c), which is a measure of anharmonicity of the bonds. The resistivity of the heat conduction through the lattice vibrations due to the Umklapp process increases with the magnitudes of Grüneisen parameters. The highest Grüneisen parameters shown along the Γ-Z path indicate that the anharmonic scattering of phonons occurs mostly along the z-direction, and agrees with the lowest κl along z-direction (Figure 3(a)). For practical aspects of its thermoelectric application, we discuss the effects of conventional defects on transport properties of Si24 . First of all, recent experiment 16 and theoretical calculations 15 of Raman spectra agree well with each other in a wide range of temperature and pressure. From these, we could infer that the local atomic structures of synthesized Si24 samples are quite close to the ideal ones. Moreover, a recent theoretical work on various defects in Si24 41 revealed that population of substitutional dopants are dominant over intrinsic point defects such as vacancy and interstitial thanks to their relatively lower formation energies (< 0.7 eV) compared with the intrinsic ones (2.3–3.7 eV). It is noteworthy that the typical n-type dopant states with very low ionization energy hybridize with conduction bands, 41 realizing good electric conductivity of the doped Si24 . We further discuss the effects of remaining guest atoms in the cages of Si24 . As we mentioned above, the Si24 crystal is synthesized by removing Na atoms confined in the cages of Na4 Si24 , which takes ∼8 days at 400 K. 15 This indicates that, on one hand, the Na atoms are loosely bound to the host Si atoms similar to clathrates, so that the trapped guest atoms can escape from the crystal. On the other hand, the relatively slow degassing process even at an elevated temperature is attributed to the high kinetic energy barrier for migration of Na atoms in the channel from one cage to the adjacent one. Based on our nudged elastic band calculations, the barrier is estimated to be ∼1 eV, which is in a good agreement to the literature. 42 In addition, when comparing the size of the channel windows and Na 11

ACS Paragon Plus Environment

(a)

Page 12 of 40

120

Si24 κxx Si24 κyy Si24 κzz dSi κ

100 80 60 40 20 0 200

300

400

500

600

700

800

Temperature, T (K)

(c) 35

2

Gr¨ uneisen parameter, γi,q

(b) Gr¨ uneisen parameter, γi,q

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Lattice thermal conductivity, κl (W/mK)

Nano Letters

1 0 −1 −2 Γ

X

ΓΓ

30 25 20 15 10 5 0 Γ Y

L

S

ΓZ

R

T Γ

Figure 3: (a) Lattice thermal conductivity (κl ) due to anharmonic phonons. Grüneisen parameters for (b) dSi and (c) Si24 are shown.

12

ACS Paragon Plus Environment

Page 13 of 40 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

atom, the transport behavior of the guest atoms is dominated by a single-file diffusion, of which the degassing rate is limited by diffusive motion of the atoms at the ends of the line. Thus, complete removal of the Na atoms might be difficult, and residual Na atoms might be remaining in the cages. We expect that those residual guest atoms would have the rattling effects as observed in skutterudites and clathrates, 14 enhancing the thermoelectric properties of the crystal. With the guest atoms, their thermal vibrations can interact with the acoustic phonons of the host atoms, which suppresses the lattice thermal conductivity further. Similarly, thermal conductivities of carbon nanotubes (CNTs) filled with water was suppressed by 20–35 % compared to that of empty CNT. 43 Our equilibrium molecular dynamics simulations clearly indicate that the lattice thermal conductivity is reduced by at least ∼50 % due to Na ions in the pores for Na0.37 Si24 as shown in SI S3. In addition, the residual Na atoms in the channel would donate electrons to the host Si24 crystal without affecting the electronic structures. For instance, band structures in Figure 4 are nearly unchanged for a single Na atom out of 288 Si atoms (∼0.35 at.%), and donor level begins to appear when two Na atoms are doped (∼0.7 at.%) as indicated by the arrow. This spontaneous electron doping is advantageous for the Si24 to be used as a TE material as discussed above (see Figure 1(c and d)). Moreover, this spontaneous doping effect would significantly lower the chances of electrons being scattered by impurities (i.e., dopant atoms), which becomes dominant at heavy doping. It is worth to note that large number of guest atoms would donate many electrons to the cage structure and eventually turn the the host materials into a metal. 44 In this case, thermal transport is dominated by electrons, i.e., κ ≈ κe . In conclusion, we demonstrate thermoelectric properties of a new silicon allotrope, Si24 by using various ab initio computational methods without empirical parameters: phononmediated electrical conductivity and lattice thermal conductivity with anharmonic phonon effects. The highly anisotropic structure of the Si24 results in anisotropic electronic and ther13

ACS Paragon Plus Environment

Nano Letters

low 2.0 1.5

Energy (eV)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 14 of 40

high 2.0

(a)

1.5

2.0

(b)

1.5 1.0

1.0

1.0

0.5

0.5

0.5

0.0

0.0

0.0

-0.5

-0.5

-0.5

-1.0

-1.0

-1.0

-1.5

-1.5

-1.5

-2.0

Γ

Y

T

DOS (a.u.)

-2.0

×200

Γ

Y

T

DOS (a.u.)

(c)

-2.0

×200

Γ

Y

T

DOS (a.u.)

Figure 4: Electronic band dispersions and density of states of Na-doped Si24 with (a) 0, (b) 1 and (c) 2 Na ions in a (3×2×2) supercell containing 288 Si atoms. The supercell band structures are unfolded to the irreducible Brillouin zone of the primitive cell of Si24 by using BandUP code. 45,46 Color-coding indicates spectral weight. Total and projected DOS to Na (magnified by a factor of 200) are shown in gray and red shades, respectively. For a single Na ion content, the ionic doping concentration corresponds to 1.58×1020 cm−3 . In (a), the zero energy of the pristine Si24 was set to be the middle between edges of conduction and valence bands and was set to be charge neutral points due to Na atom doping for both (b) and (c). mal transport properties as well. The electron-doped Si24 displays superior thermoelectric behavior to the Si in a cubic diamond phase, which is ascribed to the enhanced power factor and reduced lattice thermal conductivity. We also pointed out that the thermoelectric performance can be further enhanced by guest atoms in the cage, due to their role as electron donator and rattler.

Acknowledgement We thank Korea Institute for Advanced Study for providing computing resources (KIAS Center for Advanced Computation Linux Cluster System) for this work. Y.-W.S. was supported by NRF of Korea (Grant No. 2017R1A5A1014862, SRC program: vdWMRC Center). DYK acknowledges the support by NSFC of China (Grant No. 11774015) and NRF (NRF2017R1D1A1B03031913).

14

ACS Paragon Plus Environment

Page 15 of 40 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

Notes The authors declare no competing financial interest.

Supporting Information Available • Discussion on electric conductivity • Convergence tests for Brillouin zone integration on a discrete grid • Rattling effects of Na ions in pores of Si24

References (1) Zhao, L.-D.; Lo, S.-H.; Zhang, Y.; Sun, H.; Tan, G.; Uher, C.; Wolverton, C.; Dravid, V. P.; Kanatzidis, M. G. Ultralow thermal conductivity and high thermoelectric figure of merit in SnSe crystals. Nature 2014, 508, 373–377. (2) Zhao, L.-D.; Tan, G.; Hao, S.; He, J.; Pei, Y.; Chi, H.; Wang, H.; Gong, S.; Xu, H.; Dravid, V. P.; Uher, C.; Snyder, G. J.; Wolverton, C.; Kanatzidis, M. G. Ultrahigh power factor and thermoelectric performance in hole-doped single-crystal SnSe. Science 2016, 351, 141–144. (3) Dughaish, Z. H. Lead telluride as a thermoelectric material for thermoelectric power generation. Physica B: Condensed Matter 2002, 322, 205–223. (4) Yamashita, O.; Tomiyoshi, S.; Makita, K. Bismuth telluride compounds with high thermoelectric figures of merit. Journal of Applied Physics 2002, 93, 368–374. (5) Lee, J.-H.; Galli, G. A.; Grossman, J. C. Nanoporous Si as an Efficient Thermoelectric Material. Nano Letters 2008, 8, 3750–3754.

15

ACS Paragon Plus Environment

Nano Letters 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

(6) He, Y.; Donadio, D.; Lee, J.-H.; Grossman, J. C.; Galli, G. Thermal Transport in Nanoporous Silicon: Interplay between Disorder at Mesoscopic and Atomic Scales. ACS nano 2011, 5, 1839–1844. (7) Boukai, A. I.; Bunimovich, Y.; Tahir-Kheli, J.; Yu, J.-K.; Goddard III, W. A.; Heath, J. R. Silicon nanowires as efficient thermoelectric materials. Nature 2008, 451, 168–171. (8) Hochbaum, A. I.; Chen, R.; Delgado, R. D.; Liang, W.; Garnett, E. C.; Najarian, M.; Majumdar, A.; Yang, P. Enhanced thermoelectric performance of rough silicon nanowires. Nature 2008, 451, 163–167. (9) Lee, H. J.; Anoop, G.; Lee, H. J.; Kim, C.; Park, J.-W.; Choi, J.; Kim, H.; Kim, Y.J.; Lee, E.; Lee, S.-G.; Kim, Y.-M.; Lee, J.-H.; Jo, J. Y. Enhanced thermoelectric performance of PEDOT:PSS/PANI–CSA polymer multilayer structures. Energy & Environmental Science 2016, 9, 2806–2811. (10) Joshi, G.; Lee, H.; Lan, Y.; Wang, X.; Zhu, G.; Wang, D.; Gould, R. W.; Cuff, D. C.; Tang, M. Y.; Dresselhaus, M. S.; Chen, G.; Ren, Z. Enhanced Thermoelectric Figureof-Merit in Nanostructured p-type Silicon Germanium Bulk Alloys. Nano Letters 2008, 8, 4670–4674. (11) Wang, X. W.; Lee, H.; Lan, Y. C.; Zhu, G. H.; Joshi, G.; Wang, D. Z.; Yang, J.; Muto, A. J.; Tang, M. Y.; Klatsky, J.; Song, S.; Dresselhaus, M. S.; Chen, G.; Ren, Z. F. Enhanced thermoelectric figure of merit in nanostructured n-type silicon germanium bulk alloy. Applied Physics Letters 2008, 93, 193121. (12) Tang, J.; Wang, H.-T.; Lee, D. H.; Fardy, M.; Huo, Z.; Russell, T. P.; Yang, P. Holey Silicon as an Efficient Thermoelectric Material. Nano Letters 2010, 10, 4279–4283. (13) Nakamura, Y.; Isogawa, M.; Ueda, T.; Yamasaka, S.; Matsui, H.; Kikkawa, J.; Ikeuchi, S.; Oyake, T.; Hori, T.; Shiomi, J.; Sakai, A. Anomalous reduction of thermal 16

ACS Paragon Plus Environment

Page 16 of 40

Page 17 of 40 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

conductivity in coherent nanocrystal architecture for silicon thermoelectric material. Nano Energy 2015, 12, 845–851. (14) Dolyniuk, J.-A.; Owens-Baird, B.; Wang, J.; Zaikina, J. V.; Kovnir, K. Clathrate thermoelectrics. Materials Science and Engineering: R: Reports 2016, 108, 1–46. (15) Kim, D. Y.; Stefanoski, S.; Kurakevych, O. O.; Strobel, T. A. Synthesis of an openframework allotrope of silicon. Nature Materials 2015, 14, 169–173. (16) Tong, X.; Xu, X.; Fultz, B.; Zhang, H.; Strobel, T. A.; Kim, D. Y. Phonons in Si24 at simultaneously elevated temperature and pressure. Phys. Rev. B 2017, 95, 094306. (17) Zhang, J.; Liu, H. J.; Cheng, L.; Wei, J.; Liang, J. H.; Fan, D. D.; Jiang, P. H.; Sun, L.; Shi, J. High thermoelectric performance can be achieved in black phosphorus. Journal of Materials Chemistry C 2016, 4, 991–998. (18) Ouyang, T.; Zhang, P.; Xiao, H.; Tang, C.; Li, J.; He, C.; Zhong, J. Potential thermoelectric material open framework Si24 from a first-principles study. Journal of Physics D: Applied Physics 2017, 50, 425501. (19) Qin, G.; Yan, Q.-B.; Qin, Z.; Yue, S.-Y.; Cui, H.-J.; Zheng, Q.-R.; Su, G. Hinge-like structure induced unusual properties of black phosphorus and new strategies to improve the thermoelectric performance. Scientific Reports 2014, 4, 6922. (20) Zhang, J.; Liu, H. J.; Cheng, L.; Wei, J.; Liang, J. H.; Fan, D. D.; Shi, J.; Tang, X. F.; Zhang, Q. J. Phosphorene nanoribbon as a promising candidate for thermoelectric applications. Scientific Reports 2014, 4, 4033. (21) Liao, B.; Zhou, J.; Qiu, B.; Dresselhaus, M. S.; Chen, G. Ab initio study of electronphonon interaction in phosphorene. Phys. Rev. B 2015, 91, 235419. (22) Giannozzi, P.; Baroni, S.; Bonini, N.; Calandra, M.; Car, R.; Cavazzoni, C.; Ceresoli, D.; Chiarotti, G. L.; Cococcioni, M.; Dabo, I.; Corso, A. D.; de Gironcoli, S.; Fabris, S.; 17

ACS Paragon Plus Environment

Nano Letters 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Fratesi, G.; Gebauer, R.; Gerstmann, U.; Gougoussis, C.; Kokalj, A.; Lazzeri, M.; Martin-Samos, L.; Marzari, N.; Mauri, F.; Mazzarello, R.; Paolini, S.; Pasquarello, A.; Paulatto, L.; Sbraccia, C.; Scandolo, S.; Sclauzero, G.; Seitsonen, A. P.; Smogunov, A.; Umari, P.; Wentzcovitch, R. M. QUANTUM ESPRESSO: a modular and open-source software project for quantum simulations of materials. Journal of Physics: Condensed Matter 2009, 21, 395502. (23) Troullier, N.; Martins, J. L. Efficient pseudopotentials for plane-wave calculations. Physical Review B 1991, 43, 1993–2006. (24) Perdew, J. P.; Burke, K.; Ernzerhof, M. Generalized Gradient Approximation Made Simple. Physical Review Letters 1996, 77, 3865–3868. (25) Allen, P. B. New method for solving Boltzmann’s equation for electrons in metals. Physical Review B 1978, 17, 3725–3734. (26) Park, C.-H.; Bonini, N.; Sohier, T.; Samsonidze, G.; Kozinsky, B.; Calandra, M.; Mauri, F.; Marzari, N. Electron–Phonon Interactions and the Intrinsic Electrical Resistivity of Graphene. Nano Letters 2014, 14, 1113–1119. (27) Noffsinger, J.; Giustino, F.; Malone, B. D.; Park, C.-H.; Louie, S. G.; Cohen, M. L. EPW: A program for calculating the electron–phonon coupling using maximally localized Wannier functions. Computer Physics Communications 2010, 181, 2140–2148. (28) Marzari, N.; Mostofi, A. A.; Yates, J. R.; Souza, I.; Vanderbilt, D. Maximally localized Wannier functions: Theory and applications. Reviews of Modern Physics 2012, 84, 1419–1475. (29) Souza, I.; Marzari, N.; Vanderbilt, D. Maximally localized Wannier functions for entangled energy bands. Physical Review B 2001, 65, 035109.

18

ACS Paragon Plus Environment

Page 18 of 40

Page 19 of 40 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

(30) Marzari, N.; Vanderbilt, D. Maximally localized generalized Wannier functions for composite energy bands. Physical Review B 1997, 56, 12847–12865. (31) Giustino, F.; Cohen, M. L.; Louie, S. G. Electron-phonon interaction using Wannier functions. Phys. Rev. B 2007, 76, 165108. (32) Madsen, G. K. H.; Singh, D. J. BoltzTraP. A code for calculating band-structure dependent quantities. Computer Physics Communications 2006, 175, 67–71. (33) Ziman, J. M. Electrons and Phonons; The Theory of Transport Phenomena in Solids; Oxford University Press, 1962. (34) Togo, A.; Chaput, L.; Tanaka, I. Distributions of phonon lifetimes in Brillouin zones. Physical Review B 2015, 91, 094306. (35) Kresse, G.; Hafner, J. Ab initio molecular dynamics for liquid metals. Physical Review B 1993, 47, 558–561. (36) Kresse, G.; Hafner, J. Ab initio molecular-dynamics simulation of the liquidmetal–amorphous-semiconductor transition in germanium. Physical Review B 1994, 49, 14251–14269. (37) Kresse, G.; Furthmüller, J. Efficiency of ab-initio total energy calculations for metals and semiconductors using a plane-wave basis set. Computational Materials Science 1996, 6, 15–50. (38) Kresse, G.; Furthmüller, J. Efficient iterative schemes for ab initio total-energy calculations using a plane-wave basis set. Physical Review B 1996, 54, 11169–11186. (39) Kresse, G.; Joubert, D. From ultrasoft pseudopotentials to the projector augmentedwave method. Physical Review B 1999, 59, 1758–1775. (40) Pei, Y.; Shi, X.; LaLonde, A.; Wang, H.; Chen, L.; Snyder, G. J. Convergence of electronic bands for high performance bulk thermoelectrics. Nature 2011, 473, 66–69. 19

ACS Paragon Plus Environment

Nano Letters 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

(41) Linghu, J.; Shen, L.; Yang, M.; Xu, S.; Feng, Y. P. Si24 : An Efficient Solar Cell Material. The Journal of Physical Chemistry C 2017, 121, 15574–15579. (42) Arrieta, U.; Katcho, N. A.; Arcelus, O.; Carrasco, J. First-Principles Study of Sodium Intercalation in Crystalline Nax Si24 (0≤x≤4) as Anode Material for Na-ion Batteries. Scientific Reports 2017, 7, 510. (43) Thomas, J. A.; Iutzi, R. M.; McGaughey, A. J. H. Thermal conductivity and phonon transport in empty and water-filled carbon nanotubes. Physical Review B 2010, 81, 045413. (44) Stefanoski, S.; Malliakas, C. D.; Kanatzidis, M. G.; Nolas, G. S. Synthesis and Structural Characterization of Nax Si136 (0 < x ≤ 24) Single Crystals and Low-Temperature Transport of Polycrystalline Specimens. Inorganic Chemistry 2012, 51, 8686–8692. (45) Medeiros, P. V. C.; Stafström, S.; Björk, J. Effects of extrinsic and intrinsic perturbations on the electronic structure of graphene: Retaining an effective primitive cell band structure by band unfolding. Phys. Rev. B 2014, 89, 041407. (46) Medeiros, P. V. C.; Tsirkin, S. S.; Stafström, S.; Björk, J. Unfolding spinor wave functions and expectation values of general operators: Introducing the unfolding-density operator. Phys. Rev. B 2015, 91, 041116.

20

ACS Paragon Plus Environment

Page 20 of 40

Page 21 of 40

Graphical TOC Entry 0.7

20

0.6 0.5

Si24

0.4

12

0.3

8

0.2

4

0.1

electric conductivity

σSi24 > σdSi

thermal conductivity

κSi24 < κdSi

dSi

0

0 300

400

500

600

Temperature, T (K)

21

16

ACS Paragon Plus Environment

700

24 ZTSi max ZTdSi max

dSi

Ratio,

Si24

ZTmax

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

Nano Letters 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Disentangling Magnetic Hardening and Molecular Spin Chain Contributions to Exchange Bias in Ferromagnet/Molecule Bilayers Samy Boukari*,†, Hashim Jabbar†, Filip. Schleicher†, Manuel Gruber†,§, Garen Avedissian†, Jacek Arabski†, Victor Da Costa†, Guy Schmerber†, Prashanth Rengasamy†, Bertrand Vileno‡, Wolfgang Weber†, Martin Bowen*,† Eric Beaurepaire†‼ †

Institut de Physique et Chimie des Matériaux de Strasbourg, Université de Strasbourg, CNRS UMR 7504, 23 rue du Loess, BP 43, F-67034 Strasbourg Cedex 2 (France) ‡ Institut de Chimie de Strasbourg, Université de Strasbourg, CNRS UMR7177, 4 rue Blaise Pascal, F-67081 Strasbourg Cedex (France) and French EPR Federation of Research (REseau NAtional de Rpe interDisciplinaire (RENARD), Fédération IR-RPE CNRS 3443), France. AUTHOR INFORMATION Corresponding Author *(S. B) E-mail: [email protected] *(M. B) E-mail: [email protected]

Deceased April 24th, 2018.

Present Addresses § Institut für Experimentelle und Angewandte Physik, Christian-Albrechts-Universität zu Kiel, 24098 Kiel, Germany.

Keywords: organic spintronics, organic-inorganic interfaces, exchange bias, molecular magnetism, ferromagnetic resonance Abstract:

ACS Paragon Plus Environment

Page 22 of 40

Page 23 of 40 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

We performed ferromagnetic resonance and magnetometry experiments to clarify the relationship between two reported magnetic exchange effects arising from interfacial spinpolarized charge transfer in ferromagnetic metal (FM)/molecule bilayers: the magnetic hardening effect, and spinterface-stabilized molecular spin chains. To disentangle these effects we tuned the metal phthalocyanine molecule central site’s magnetic moment to enhance or suppress the formation of spin chains in the molecular film. We find that both effects are distinct, and additive. In the process, we extend the list of FM/molecule candidate pairs that are known to generate magnetic exchange effects, experimentally confirm the predicted increase in anisotropy upon molecular adsorption, and show that spin chains within the molecular film can enhance magnetic exchange. Our results confirm, as an echo to progress regarding inorganic spintronic tunnelling, that spintronic tunnelling across structurally ordered organic barriers has been reached through previous magnetotransport experiments.

A change in our understanding of the spin-polarized solid-state tunneling process (sp-SST) took place in the late 1990s-early 2000s, as the impact of the inorganic barrier’s electronic structure (impact of d sites, structural ordering) on sp-SST was revealed against a backdrop of research using amorphous Al2O3 barriers. The jump in spintronic performance, with an effective spin polarization P that reaches 87% at room temperature (RT), reflects harnessing the conservation of electronic symmetry during SST1, as well as the beneficial impact of oxygen vacancies according to recent research2,3. A similar change is presently underway regarding organic tunnel barriers. Indeed, the initial understanding of sp-SST was established in the late 2000s across amorphous tunnel barriers such as those with Alq3 molecules4. Here again, the barrier was first thought to merely constitute a means of spin transport between decoupled ferromagnetic (FM) electrodes. However, due to sizeable charge transfer, new electronic states with promising spintronic properties can occur at the interface between a ferromagnet and a molecule5. Notably, an ACS Paragon Plus Environment

Nano Letters 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

inversion and high value of spin polarization P of the ferromagnet/molecule interface was detected on a wide variety of interface constituents6–8. However, this high P has thus far been evidenced in a solid-state device only at low temperature9 (up to 99% at 2 K10), while similar measurements of P at RT do not exceed 28%4. Unlocking the RT spintronic potential of the FM/molecule interface requires a better understanding of the magnetic exchange effects within FM/molecular films that involve this interface, which is also called an organic/molecular spinterface. Thus, far, two such effects have been identified: a magnetic hardening of the topmost FM monolayer forming the organic spinterface compared to the underlying FM thin film, and the impact on the organic spinterface’s magnetic properties of molecular spin chains away from the interface and into the organic layer. Indeed, the adsorption of a molecule onto a FM surface promotes an interface with distinct magnetic properties relative to those of the FM substrate, both on the molecule8,11,12 and on the topmost FM layer13. On the metal side of the interface, this interface may have an increased anisotropy and a magnetization parallel or antiparallel to that of the FM substrate. This effect, called magnetic hardening14,15, can in turn lead to magnetoresistance near RT13. However, it remains unclear whether the magnetic hardening effect is more general as it was demonstrated only for the FM Co and the radical molecule zinc methyl phenalenyl. It is speculated that this magnetic hardening effect, which depends on the strength of the interaction with the FM substrate, may lead to other phenomena. In one such phenomenon, the exchange bias (EB), the center of the FM’s hysteresis loop is shifted by a magnetic field Hshift away from H=0. In a scenario involving only magnetic hardening, Hshift should not depend on the thickness of the molecular film deposited atop the FM layer. Since the magnetization of the first molecular monolayer (ML) can be fixed at RT through magnetic interactions with the FM substrate8,11,12,16,17, this can in particular stabilize a molecule’s intrinsic local magnetic moment. Furthermore, due to structural ordering, a ACS Paragon Plus Environment

Page 24 of 40

Page 25 of 40 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

magnetic order can occur between these local magnetic moments so as to form spin chains within the molecular film

18,19

. Combining these two concepts, the spinterface’s magnetism

can stabilize16 at RT an antiferromagnetic (AFM) ordering between the paramagnetic centers M of a metal phthalocyanine (MPc, schematized in Fig. 1(c)) film, thereby stabilizing the magnetic axis of a paramagnetic molecular spin chain.

Figure 1. Magnetic exchange effects within ferromagnet/molecule bilayers. Spinpolarized charge transfer alters the properties of the molecule and the ferromagnetic layer upon adsorption. The magnetism of the resulting interface, also called an organic spinterface, differs from that of its constituent materials. 1) The first molecular layer is ferromagnetically coupled to the topmost FM layer and acquires a strong spin polarization of conduction states at EF8,9. If the molecule carries a local magnetic moment (see the schematic of a phthalocyanine molecule and its central metal atom. Both are depicted by a red dot/arrow. Here, M=Mn with a magnetic moment of 3µB; Fe with 2µB, Co with 1µB ; or Zn with 0µB.19– 22 ), then this magnetic stabilization can be extended to subsequent molecular monolayers away from the interface16 according to structurally imposed magnetic exchange interactions19. 2) The topmost FM layer’s magnetism is altered relative to the FM’s underlying layers. This magnetic hardening effect was first observed in magnetotransport experiments13. We disentangle how magnetic hardening and spinterface-stabilized molecular spin chains can contribute to an effective exchange bias effect on the bulk portion of the FM thin film. Conversely, at low temperature, the magnetization loop of the Co/MnPc bilayer exhibits the phenomenological feature of EB23: an anomalously large coercive field associated with a shift Hshift in the loop center away from H=016. This association is natural since EB, first discovered using inorganic materials but also observed between a submonolayer of TbPc2 molecular magnets and AFM FeMn,24 can involve FM/AFM bilayers25,26. With a FM/molecular film, we expect an increase of EB with the molecular film thickness up to ACS Paragon Plus Environment

Nano Letters 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

several nm, which is typical for inorganic materials27. The EB variation would here be correlated with the length of the spin chains. As a note, Serri et al. found a lower limit of 5 nm for the spin chains length in CoPc films19 . Thus, although this observation of EB is consistent with an explanation in terms of an effective FM/AFM magnetic ordering16 of the Co/MnPc bilayer, it could also be explained in terms of magnetic hardening effects13, which conceptually does not require an AFM ordering of molecular spin chains within the organic layer. Resolving this controversy is of prime importance to understand the origin of magnetotransport results across Pc magnetic tunnel junctions10,28 with Co electrodes: concurrent tunneling magnetoresistance / tunneling anisotropic magnetoresistance (TMR/TAMR); high (up to 2T) bias-dependent coercive fields; concurrent unidirectional/uniaxial magnetotransport; and spectroscopic features that depend on the macroscopic MTJ magnetization state and that were tentatively attributed to spin-flip spectroscopy29 across these solid-state devices. These magnetic effects could underscore a combination of magnetic hardening and spin chain effects. The latter case would amount to transport across a structurally ordered organic tunnel barrier. To address this open question within progress on organic SST toward ordered tunnel barriers, we deploy Superconducting Quantum Interference Device (SQUID) and ferromagnetic resonance (FMR) experiments on FM/MPc (FM=Co, Ni81Fe19 i.e. Py; and M=Mn, Fe, Co, Zn) bilayers. Changing the M site so as to tune the MPc molecule’s intrinsic magnetic moment to be zero (ZnPc) or non-zero (MnPc, FePc, CoPc)19–22, while maintaining a similar adsorption geometry of this planar molecular family onto the FM surface, allows us to selectively suppress spin chains within the MPc film. We find that EB arises here from spinterface-stabilized molecular spin chains and from magnetic hardening13. We reproduce the EB effect within Co/ZnPc, Co/FePc and Co/CoPc bilayers in addition to Co/MnPc. The data on Co/ZnPc are interpreted in terms of the magnetic hardening effect. We find similar effects upon replacing Co with Py. The Hshift ACS Paragon Plus Environment

Page 26 of 40

Page 27 of 40 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

amplitude obtained with magnetic Pcs is larger than with non-magnetic ZnPc, and increases with the CoPc thicknessup to 10 nm. This confirms the impact of molecular spin chains on the EB at organic spinterfaces16, and underscores the interfacial impact of the effect as regards the FM layer. The strength of the EB can be increased by a factor of up to 2-20 when using molecular rather than inorganic materials (see supporting information). We experimentally confirm the change in anisotropy13 that was put forward to explain the magnetic hardening effect by using FMR. We infer from our magnetometry analysis that molecular spin chains play an important role in promoting the EB effect within the MnPc and CoPc films that were used as tunnel barriers in magnetotransport experiments.10,28 This lends strong credence to the impact of magnetic ordering, and therefore structural ordering, on these magnetotransport results, which thus constitute the onset of tunnelling across a structurally ordered organic tunnel barrier. MPc molecules were deposited on Si//SiO2(500 nm)/Co(10 nm) or on Si//SiO2(500 nm)/Cr(15 nm)/Py(10 nm) and protected with a 10 nm gold layer. We emphasize that, to reveal the properties of the FM/molecule interface in solid state devices, special care to be taken to avoid pinholes (for a review on magnetoresistance in organic spin valves and the role of pinholes, see

30,31

and ref. therein). However, gold in direct contact with the bottom Co

electrode was checked to be unable to produce magnetic hardening16 so that even if it penetrates the molecular layer up to the ferromagnet, it cannot be responsible for the observed effects. All materials were deposited by sublimation in ultra-high vacuum (UHV), except Py and Cr which were deposited by sputtering while remaining in UHV. Structural characterization is reported in the Supporting Information (SI). We present in Figure 2 SQUID magnetometry results on Co/MPc (M=Mn, Co, Fe, Zn) bilayers.

ACS Paragon Plus Environment

Nano Letters

a) FC

3

Magnetization (emu/cm )

1500 750 0

MnPc CoPc FePc ZnPc

-750

-1500 -5000

0 Field (Oe)

5000

1500 b)

MnPc/Co CoPc/Co FePc/Co ZnPc/Co

1000

|Hshift| (Oe)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 28 of 40

500

0 0

50

100 150 200 Temperature (K)

250

300

Figure 2. Magnetic exchange effect of Co/MPc bilayers upon suppressing or enhancing the phthalocyanine molecule’s magnetic moment. Magnetization loops at 2 K of Au-capped Co(10 nm)/MPc(10 nm) bilayers after (a) field cooling (FC) at +3T. The shift of the loop center away from H=0 defines Hshift. (b) Temperature dependence of Hshift for Co/MPc bilayers.

Magnetization loops at 2 K after H=+3 T field cooling (FC: panel (a)) reveal, for all Co/MPc bilayers considered, that the hysteresis loop center is shifted away from H=0 by a negative Hshift after FC. We present its temperature dependence in panel (b). For all MPc, we find that Hshift decreases with increasing T, and vanishes at T~100 K. Since we observe a non-zero Hshift for ZnPc, a molecule with neither radical nor 3d magnetism, i.e. with zero nominal magnetic moment, this implies that spin chains within the molecular film are not required to observe the magnetic pinning of the FM substrate. This could, however, reflect the magnetic hardening effect as reported by Raman et al.13, which in turn can pin the magnetization of the underlying FM substrate.

ACS Paragon Plus Environment

Page 29 of 40

Furthermore, we observe that Hshift increases in amplitude upon increasing the magnetic moment on the molecule’s central site M from zero for Zn, to 1 µB for Co and to 3 µB for Mn 19–22

(see Fig. 2c); Hshift is an order of magnitude stronger when Zn is replaced by Mn.

However, Hshift obtained with FePc, , which in the bulk has a moment of 2 µB20, is lower than Hshift obtained with CoPc. A possible explanation is that the FePc moment is reduced, as was observed by Bartolome et al.32 in thin films, leading to a reduction in the hard layer magnetization and a concurrent decrease of Hshift33. Does this increase in Hshift originate from the spin chains or from a variation in magnetic hardening when changing the molecules ? To distinguish between the two scenarios, we examine the impact on Hshift of varying the thickness of the FM and CoPc layers. Referring to Fig. 3a, we find that Hshift decreases roughly as 1/tCo as the thickness of the Co layer tCo is increased. This confirms the interfacial nature of the EB from the standpoint of the FM layer’s magnetism. Apart from experimental errors, the deviation from the 1/tCo law is probably due to an evolution of the interface morphology with tCo23. 0 Co(tCo)/CoPc(10nm)/Au(10nm) a)

|Hshift (Oe)|

100 200 300 400

T=2 K

500 600 700 0.00

0.05 0.10 -1 -1 t Co (nm )

0.15 b)

250 |Hshift(Oe)|

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

200 T=2 K

150 100 50

Co(13.5nm)/CoPc(tCoPc)/Au(10nm)

0 0

5 10 15 CoPc thickness tCoPc (nm)

20

Figure 3. Explicit proof that molecular spin chains contribute to the EB. Dependence of ACS Paragon Plus Environment

Nano Letters 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 30 of 40

Hshift at T=2K upon varying (a) the Co thickness in Au-capped Co(tCo)/CoPc(10nm) bilayers and (b) the CoPc thickness in Au-capped Co(13.5 nm)/CoPc(tCoPc) bilayers. The overall 1/tCo thickness dependence of Hshift (full line) supports the usual interfacial nature of the magnetic exchange effect from the standpoint of the FM layer’s magnetism. The increase in Hshift with increasing CoPc thickness confirms that molecular spin chains can enhance the effect generated by magnetic hardening13 through an EB mechanism16.

Referring to Fig. 3b, we find that Hshift increases upon increasing the CoPc thickness up to ~10 nm, and then decreases slightly up to ~20 nm. Since the Co sites within CoPc exhibit AFM correlations19, this shows that AFM spin chains within the MPc film can enhance Hshift. Since MPc films with M=Co, Fe, Mn can exhibit AFM spin chains19, and Hshift obtained with magnetic Pcs is always larger than with non-magnetic ZnPc, we infer that EB using spinterface-stabilized molecular spin chains16 is a distinct and additive effect to that of magnetic hardening13. Note how this combination of increase and decrease in Hshift with increasing CoPc thickness is typical of an EB effect involving an inorganic AFM material with a varying grain size27. The enduring variation of Hshift for CoPc thicknesses up to 20nm thus implies that spin chains with an effective length of at least 20 nm are involved in the EB effect. Our results thus expand the list of molecules that generate EB from two13,16 to five. To test whether the FM metal Co is required to observe the magnetic exchange effect, we replaced it with

Py,

and

measured

the

Cr(15nm)/Py(10nm)/MPc(10nm)/Au(10nm)

temperature with

dependence

M=Fe,

Cr(15nm)/Py(10nm)/Cr(10nm)/Au(10nm) reference (fig. 4a).

ACS Paragon Plus Environment

Zn

of

alongside

Hshift that

for of

a

Page 31 of 40

a)

|Hshift(Oe)|

0

50

Cr/Py/Cr SQUID Cr/Py/ZnPc SQUID Cr/Py/FePc SQUID Cr/Py/FePc FMR

100 0

1000

50

100 150 200 Temperature(K)

250

300

b)

0 Ha (Oe)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

-1000

Cr/Py/Cr Cr/Py/FePc

-2000 -3000 -4000 0

50

100 150 200 Temperature (K)

250

300

Figure 4. Experimental confirmation of a uniaxial anisotropy within Py/FePc bilayers.(a) Temperature dependence of Hshift extracted from SQUID and FMR measurements for Aucapped Cr(15 nm)/Py(10 nm)/X(10 nm) with X = Cr, ZnPc, FePc. Similar data are found for FePc when using SQUID and FMR techniques. (b) Temperature dependence of the uniaxial magnetic anisotropy field Ha deduced from FMR measurements for Cr/Py/Cr and Cr/Py/FePc films. A positive(negative) Ha favors an out-of-plane(in-plane) easy magnetization axis. The substitution of FePc for Cr generates an additional, large negative contribution to Ha whose temperature dependence mimicks that of Hshift. While the reference sample generates a non-zero Hshift~15 Oe as expected since Cr is AFM34, replacing the top Cr layer with ZnPc(FePc) increases Hshift by a factor of two(six). We thus also witness the magnetic hardening and EB effects when FM=Py. To confirm the prediction13,14 that a strong anisotropy underscores the magnetic hardening effect, we performed FMR measurements on gold-capped Py/FePc bilayers. Since Py films exhibit a FMR resonance field of around 1000 Oe for X-band excitation that is significantly larger than the effective coercive field of the bilayer at all measured temperatures, the FMR technique is well adapted to study the two magnetic exchange effects35 --- magnetic hardening and spinterface-stabilized molecular spin chains --- considered here. We first compare in Fig. 4a the amplitude and temperature dependence of Hshift extracted for Py/FePc bilayers from ACS Paragon Plus Environment

Nano Letters 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

SQUID and FMR measurements. We find that they are similar, as expected since FMR is a perturbation technique. We now extract the uniaxial magnetocrystalline anisotropy field Ha from FMR measurements performed with a magnetic field applied first in the film plane along the direction of field cooling, and then along the sample normal (see SI for details). We compare in Fig. 4b the temperature dependence of Ha for a Cr/Py/FePc stack against that of a Cr/Py/Cr reference stack. As expected, the weak EB contribution from the Cr layers in the reference stack (see previous discussion) generates a low, positive Ha at all temperatures. This is indicative of an out-of-plane uniaxial anisotropy. We find that Ha decreases nearly monotonously with decreasing temperature from 300 K to 2 K. Turning now to the Cr/Py/FePc stack, we find the same sign and amplitude of Ha at 300 K. However, Ha decreases with a rate that increases around the 100 K blocking temperature observed for the Py/FePc bilayer (see Fig. 4(a)), and reaches a large negative value of Ha~-4000 Oe. Our experiments thus explicitly link the magnetic exchange effect to the presence of a large uniaxial anisotropy, in agreement with predictions13,14. The magnetotransport results across Pc magnetic tunnel junctions10,28 are thus due to a conjunction of an increase in the interfacial magnetic anisotropy upon molecular adsorption, i.e. due to the magnetic hardening effect, and of AFM spin chains in the organic tunnel barrier. This confirms that these results represent the onset of sp-SST across a structurally ordered organic tunnel barrier. To conclude, we have performed SQUID and FMR magnetometry experiments to clarify the relationship between two reported magnetic exchange effects within ferromagnetic metal/molecule bilayers: the magnetic hardening effect reported using a custom-tailored molecule13, and spinterface-stabilized molecular spin chains16, both of which can affect the magnetization reversal of the underlying FM thin film. To distinguish between the two effects, we tuned the magnetic moment of the central site of the metal phthalocyanine molecular family to selectively enhance or suppress the formation of spin chains within the molecular film. We find that both effects are distinct, and additive. In the process, we extended the list of ACS Paragon Plus Environment

Page 32 of 40

Page 33 of 40 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

FM/molecule candidate pairs that are known to generate magnetic exchange effects, experimentally confirmed the predicted13 increase in anisotropy upon molecular adsorption; and showed that using molecules with a local magnetic moment enhances magnetic exchange due to the impact on the magnetism of the organic spinterface of molecular spin chains within the organic layer away from the interface. Since the involvement of molecular spin chains implies structural ordering, our work explicitly ascribes the magnetotransport reported across MPc tunnel barriers as the onset in organic solid-state tunnelling from amorphous to ordered organic tunnel barriers.10,28 This paves the way for solid-state devices studies that exploit the quantum physical properties of spin chains. Here, the recent proposal15 and experimental demonstration36 that the organic spinterface can constitute an active component toward multifunctional electronics may be used to electrically alter the molecular spin chain’s ground/excited state, so as to craft spinpolarized transport and thus promote novel device functionalities.

Supporting Information The Supporting Information is available free of charge on the ACS Publications website at DOI:

Acknowledgements The authors thank M. Bailleul, M. Hehn and S. Heutz for discussions as well as A. Derory and Ch. Kieber for technical assistance. We acknowledge the French EPR Federation of Research (REseau NAtional de Rpe interDisciplinaire, RENARD, Fédération IR-RPE CNRS 3443), funding from the Institut Carnot MICA’s ‘Spinterface’ grant, from the Agence Nationale de la Recherche ANR-09-JCJC-0137 and ANR-11-LABX-0058 NIE (Labex NIE), from the International Center for Frontier Research in Chemistry and from the Franco-German

ACS Paragon Plus Environment

Nano Letters 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

University. H. J. acknowledges a grant from the Iraqi Ministry of Higher Education and Campus France. Notes The authors declare no competing financial interest. References (1) (2)

(3)

(4)

(5)

(6)

(7)

(8)

(9)

(10)

(11)

(12)

(13)

(14)

Miao, G.-X.; Münzenberg, M.; Moodera, J. S. Tunneling Path toward Spintronics. Rep. Prog. Phys. 2011, 74 (3), 036501. Schleicher, F.; Halisdemir, U.; Lacour, D.; Gallart, M.; Boukari, S.; Schmerber, G.; Davesne, V.; Panissod, P.; Halley, D.; Majjad, H.; et al. Localized States in Advanced Dielectrics from the Vantage of Spin- and Symmetry-Polarized Tunnelling across MgO. Nat. Commun. 2014, 5, 4547. Taudul, B.; Monteblanco, E. N.; Halisdemir, U.; Lacour, D.; Schleicher, F.; Montaigne, F.; Beaurepaire, E.; Boukari, S.; Hehn, M.; Alouani, M.; et al. Tunneling Spintronics across MgO Driven by Double Oxygen Vacancies. Adv. Electron. Mater. 2017, 1600390. Santos, T.; Lee, J.; Migdal, P.; Lekshmi, I.; Satpati, B.; Moodera, J. Room-Temperature Tunnel Magnetoresistance and Spin-Polarized Tunneling through an Organic Semiconductor Barrier. Phys. Rev. Lett. 2007, 98 (1). Lach, S.; Altenhof, A.; Tarafder, K.; Schmitt, F.; Ali, M. E.; Vogel, M.; Sauther, J.; Oppeneer, P. M.; Ziegler, C. Metal-Organic Hybrid Interface States of A Ferromagnet/Organic Semiconductor Hybrid Junction as Basis For Engineering Spin Injection in Organic Spintronics. Adv. Funct. Mater. 2012, 22 (5), 989–997. Suzuki, T.; Kurahashi, M.; Ju, X.; Yamauchi, Y. Adsorption Structure and Spin Polarization of Pentacene on a Magnetized Fe(100) Substrate: SPMDS and ERDA Study. Surf. Sci. 2004, 549 (2), 97–102. Atodiresei, N.; Brede, J.; Lazić, P.; Caciuc, V.; Hoffmann, G.; Wiesendanger, R.; Blügel, S. Design of the Local Spin Polarization at the Organic-Ferromagnetic Interface. Phys. Rev. Lett. 2010, 105 (6). Djeghloul, F.; Gruber, M.; Urbain, E.; Xenioti, D.; Joly, L.; Boukari, S.; Arabski, J.; Bulou, H.; Scheurer, F.; Bertran, F.; et al. High Spin Polarization at Ferromagnetic Metal-Organic Interfaces: A Generic Property. J. Phys. Chem. Lett. 2016, 7, 2310–2315. Barraud, C.; Seneor, P.; Mattana, R.; Fusil, S.; Bouzehouane, K.; Deranlot, C.; Graziosi, P.; Hueso, L.; Bergenti, I.; Dediu, V.; et al. Unravelling the Role of the Interface for Spin Injection into Organic Semiconductors. Nat. Phys. 2010, 6 (8), 615–620. Barraud, C.; Bouzehouane, K.; Deranlot, C.; Kim, D. J.; Rakshit, R.; Shi, S.; Arabski, J.; Bowen, M.; Beaurepaire, E.; Boukari, S.; et al. Phthalocyanine Based Molecular Spintronic Devices. Dalton Trans 2016, 45, 16694–16699. Scheybal, A.; Ramsvik, T.; Bertschinger, R.; Putero, M.; Nolting, F.; Jung, T. A. Induced Magnetic Ordering in a Molecular Monolayer. Chem. Phys. Lett. 2005, 411 (1– 3), 214–220. Wende, H.; Bernien, M.; Luo, J.; Sorg, C.; Ponpandian, N.; Kurde, J.; Miguel, J.; Piantek, M.; Xu, X.; Eckhold, P.; et al. Substrate-Induced Magnetic Ordering and Switching of Iron Porphyrin Molecules. Nat. Mater. 2007, 6 (7), 516–520. Raman, K. V.; Kamerbeek, A. M.; Mukherjee, A.; Atodiresei, N.; Sen, T. K.; Lazić, P.; Caciuc, V.; Michel, R.; Stalke, D.; Mandal, S. K.; et al. Interface-Engineered Templates for Molecular Spin Memory Devices. Nature 2013, 493 (7433), 509–513. Callsen, M.; Caciuc, V.; Kiselev, N.; Atodiresei, N.; Blügel, S. Magnetic Hardening ACS Paragon Plus Environment

Page 34 of 40

Page 35 of 40 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

Induced by Nonmagnetic Organic Molecules. Phys. Rev. Lett. 2013, 111 (10). (15) Cinchetti, M.; Dediu, V. A.; Hueso, L. E. Activating the Molecular Spinterface. Nat. Mater. 2017, 16 (5), 507–515. (16) Gruber, M.; Ibrahim, F.; Boukari, S.; Isshiki, H.; Joly, L.; Peter, M.; Studniarek, M.; Da Costa, V.; Jabbar, H.; Davesne, V.; et al. Exchange Bias and Room-Temperature Magnetic Order in Molecular Layers. Nat. Mater. 2015, 14, 981–984. (17) Kuch, W.; Bernien, M. Controlling the Magnetism of Adsorbed Metal–Organic Molecules. J. Phys. Condens. Matter 2017, 29 (2), 023001. (18) Heutz, S.; Mitra, C.; Wu, W.; Fisher, A. J.; Kerridge, A.; Stoneham, M.; Harker, A. H.; Gardener, J.; Tseng, H.-H.; Jones, T. S.; et al. Molecular Thin Films: A New Type of Magnetic Switch. Adv. Mater. 2007, 19 (21), 3618–3622. (19) Serri, M.; Wu, W.; Fleet, L. R.; Harrison, N. M.; Hirjibehedin, C. F.; Kay, C. W. M.; Fisher, A. J.; Aeppli, G.; Heutz, S. High-Temperature Antiferromagnetism in Molecular Semiconductor Thin Films and Nanostructures. Nat. Commun. 2014, 5, 3079. (20) Dale, B. W.; Williams, R. J. P.; Johnson, C. E.; Thorp, T. L. S = 1 Spin State of Divalent Iron. I. Magnetic Properties of Phthalocyanine Iron (II). J. Chem. Phys. 1968, 49 (8), 3441–3444. (21) Barraclough, C. G.; Martin, R. L.; Mitra, S. Diamagnetic Anisotropies of Metal‐Free, Nickel(II), and Zinc(II) Phthalocyanines. J. Chem. Phys. 1971, 55 (3), 1426–1429. (22) Miyoshi, H. The Magnetic Properties of Manganese(II) Phthalocyanine. II. Bull. Chem. Soc. Jpn. 1974, 47 (3), 561–565. (23) Nogues, J.; Schuller, I. K. Exchange Bias. J. Magn. Magn. Mater. 1999, 192 (2), 203– 232. (24) Nistor, C.; Krull, C.; Mugarza, A.; Stepanow, S.; Stamm, C.; Soares, M.; Klyatskaya, S.; Ruben, M.; Gambardella, P. Exchange Bias of TbPc 2 Molecular Magnets on Antiferromagnetic FeMn and Ferromagnetic Fe Films. Phys. Rev. B 2015, 92 (18). (25) Klein, T.; Schlage, K.; Buchholz, E.; Marx, U.; Burkel, E.; Röhlsberger, R. Tuning the Remanent Spin Structure of Exchange Coupled Magnetic Films. New J. Phys. 2007, 9 (9), 312–312. (26) Note that if FM/AFM bilayers is by far the most common case for the observation of EB, the effect can also be observed when two ferromagnetic layers with different coercive fields, a hard and a soft Layer, are exchanged coupled in a structure called spring magnet (25). With magnetic molecules, EB could thus not be restricted to AFM intra-chain coupling but exist regardless of the intra-chain coupling sign. (27) O’grady, K.; Fernandez-Outon, L.; Vallejo-Fernandez, G. A New Paradigm for Exchange Bias in Polycrystalline Thin Films. J. Magn. Magn. Mater. 2010, 322 (8), 883–899. (28) Barraud, C.; Bouzehouane, K.; Deranlot, C.; Fusil, S.; Jabbar, H.; Arabski, J.; Rakshit, R.; Kim, D.-J.; Kieber, C.; Boukari, S.; et al. Unidirectional Spin-Dependent MoleculeFerromagnet Hybridized States Anisotropy in Cobalt Phthalocyanine Based Magnetic Tunnel Junctions. Phys. Rev. Lett. 2015, 114 (20), 206603. (29) Ormaza, M.; Bachellier, N.; Faraggi, M. N.; Verlhac, B.; Abufager, P.; Ohresser, P.; Joly, L.; Romeo, M.; Scheurer, F.; Bocquet, M.-L.; et al. Efficient Spin-Flip Excitation of a Nickelocene Molecule. Nano Lett. 2017, 17 (3), 1877–1882. (30) Göckeritz, R.; Homonnay, N.; Müller, A.; Richter, T.; Fuhrmann, B.; Schmidt, G. Nanosized Perpendicular Organic Spin-Valves. Appl. Phys. Lett. 2015, 106 (10), 102403. (31) Geng, R.; Luong, H. M.; Daugherty, T. T.; Hornak, L.; Nguyen, T. D. A Review on Organic Spintronic Materials and Devices: II. Magnetoresistance in Organic Spin Valves and Spin Organic Light Emitting Diodes. J. Sci. Adv. Mater. Devices 2016, 1 (3), 256–272. ACS Paragon Plus Environment

Nano Letters 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

(32) Bartolomé, J.; Bartolomé, F.; García, L. M.; Filoti, G.; Gredig, T.; Colesniuc, C. N.; Schuller, I. K.; Cezar, J. C. Highly Unquenched Orbital Moment in Textured FePhthalocyanine Thin Films. Phys. Rev. B 2010, 81 (19). (33) Alexandrakis, V.; Kechrakos, D.; Moutis, N.; Niarchos, D.; Hadjipanayis, G.; Panagiotopoulos, I. Coercivity and Random Interfacial Exchange Coupling in CoPt/Co Films. J. Appl. Phys. 2016, 119 (12), 123905. (34) Parker, J.; Wang, L.; Steiner, K.; Crowell, P.; Leighton, C. Exchange Bias as a Probe of the Incommensurate Spin-Density Wave in Epitaxial Fe/Cr (001). Phys. Rev. Lett. 2006, 97 (22), 227206. (35) Scott, J. Ferromagnetic Resonance Studies in the Bilayer System Ni0. 80Fe0. 20/Mn0. 50Fe0. 50: Exchange Anisotropy. J. Appl. Phys. 1985, 57 (8), 3681–3683. (36) Studniarek, M.; Cherifi-Hertel, S.; Urbain, E.; Halisdemir, U.; Arras, R.; Taudul, B.; Schleicher, F.; Hervé, M.; Lambert, C.-H.; Hamadeh, A.; et al. Modulating the Ferromagnet/Molecule Spin Hybridization Using an Artificial Magnetoelectric. Adv. Funct. Mater. 2017, 1700259.

ACS Paragon Plus Environment

Page 36 of 40

(a)

(b)

Page 37 of 40

a3

z x

a3

a2

y

a2

a1

(c)

(d) 300 400 500 600 700

0.5

0.6

0.3

0.3 0.2

0.1

0.1

8 4 dSi

0 −5

0

ACS Paragon Plus Environment

5

10

Doping concentration (×1020 cm−3 )

16 12

0.4

0.2

0 −10

Si24

0.5 ZTmax

0.4

20

0.7

K K K K K

300

0 400

500

600

Temperature, T (K)

700

Enhancement

0.6 Figure of merit, ZT

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33

Nano Letters

(b)

Nano Letters

5

Power Factor, σS 2 (mW/K2 m)

Conductivity, σ (×106 Ohm−1 m−1 )

(a) σxx σyy σzz

Page 38 of 40

15

σxx S 2

1 σyy S 2 2 4 σzz S 2 3 10 4 3 5 6 7 2 5 8 9 10 1 11 12 0 0 13 −10 −10 −5 0 5 10 −5 0 5 10 14 20 −3 20 −3 Doping concentration (×10 cm ) Doping concentration (×10 cm ) 15 16 2 17 18 Si24 19 1 dSi k 20 T Y 21 S R 22 Z Γ k k 23 0 24 25 26−1 27 28 29−2 ACS Paragon Plus Environment Γ Y S Γ Z R T Γ 0 250 500 750 1000 30 e−ph 31

Energy, εn,k -εF (eV)

(c)

y

z

x

τn,k

(fs)

Lattice thermal conductivity, κl (W/mK)

(a)

Page 39 of120 40

Si24 κxx Si24 κyy Si24 κzz dSi κ

100 80 60 40 20 0 200

300

400

500

600

700

800

Temperature, T (K)

Gr¨ uneisen parameter, γi,q

(c) 35

Gr¨ uneisen parameter, γi,q

1 2 3 4 5 6 7 8 9 10 11 12 13 14 (b) 15 2 16 17 1 18 19 20 0 21 22−1 23 24 −2 25 26

Nano Letters

30 25 20 15 10

5 ACS Paragon Plus Environment 0 Γ

X

ΓΓ

L

Γ Y

S

ΓZ

R

T Γ

Nano Letters

low

high 2.0

!"#

1.5

2.0

!$#

1.5

1.0

Energy (eV)

1 2 3 4 5 6 7 8 9 10 11 12 13 2.0 14 15 1.5 16 17 18 1.0 19 20 0.5 21 22 23 0.0 24 25 -0.5 26 27 -1.0 28 29 30 -1.5 31 32 -2.0 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47

Page 40 of 40

Γ

Y

T

DOS (a.u.)

!&''

1.0

0.5

0.5

0.0

0.0

-0.5

-0.5

-1.0

-1.0

-1.5

-1.5

-2.0

Γ

Y

T

DOS (a.u.)

ACS Paragon Plus Environment

!%#

-2.0

!&''

Γ

Y

T

DOS (a.u.)