Polarizability, Susceptibility, and Dielectric Constant of Nanometer‐Scale Molecular Films: A Microscopic View
Source abstract
Abstract
Abstract The size‐dependence of the polarizability, susceptibility, and dielectric constant of nanometer‐scale molecular layers is explored theoretically. First‐principles calculations based on density functional theory are compared to phenomenological modeling based on polarizable dipolar arrays for a model system of organized monolayers composed of oligophenyl chains. Size trends for all three quantities are primarily governed by a competition between out‐of‐plane polarization enhancement and in‐plane polarization suppression. Molecular packing density is the single most important factor controlling this competition and it strongly affects the bulk limit of the dielectric constant as well as the rate at which it is approached. Finally, the polarization does not reach its “bulk” limit, as determined from the Clausius–Mossotti model, but the susceptibility and dielectric constant do converge to the correct bulk limit. However, whereas the Clausius–Mossotti model describes the dielectric constant well at low lateral densities, finite size effects of the monomer units cause it to be increasingly inaccurate at high lateral densities.
Source: crossref · Quality: baseline
Research context
OpenAlex topics: Molecular Junctions and Nanostructures · Semiconductor materials and devices
SciVX discovery tags are terminology-based. OpenAlex topics are source-provided classifications.
References (89)
Show source-provided references
1. 10.1002/(SICI)1521-4095(199906)11:8<605::AID-ADMA605>3.0.CO;2-Q10.1002/(SICI)1521-4095(199906)11:8<605::AID-ADMA605>3.0.CO;2-Q ↗
2. 10.1146/annurev.physchem.50.1.41310.1146/annurev.physchem.50.1.413 ↗
3. 10.1021/ar990047t10.1021/ar990047t ↗
4. 10.1002/adma.20050051710.1002/adma.200500517 ↗
5. 10.1002/cphc.20070017710.1002/cphc.200700177 ↗
6. 10.1021/jo070543s10.1021/jo070543s ↗
7. 10.1088/0953-8984/20/01/01300110.1088/0953-8984/20/01/013001 ↗
8. 10.1021/ar800010210.1021/ar8000102 ↗
9. 10.1016/j.susc.2008.10.06310.1016/j.susc.2008.10.063 ↗
10. 10.1002/adma.20090183410.1002/adma.200901834 ↗
11. 10.1002/adfm.20050018710.1002/adfm.200500187 ↗
12. 10.1021/nl048372j10.1021/nl048372j ↗
13. 10.1038/nature0527010.1038/nature05270 ↗
14. 10.1021/ja068417d10.1021/ja068417d ↗
15. 10.1021/ar700284q10.1021/ar700284q ↗
16. 10.1166/jctn.2008.249610.1166/jctn.2008.2496 ↗
17. 10.1021/jp900301f10.1021/jp900301f ↗
18. 10.1038/nmat211910.1038/nmat2119 ↗
19. 10.1002/adfm.20080087610.1002/adfm.200800876 ↗
20. 10.1021/jp062096a10.1021/jp062096a ↗
21. 10.1063/1.172985310.1063/1.1729853 ↗
22. 10.1063/1.12193010.1063/1.121930 ↗
23. 10.1073/pnas.050102710210.1073/pnas.0501027102 ↗
24. 10.1002/adma.20050051710.1002/adma.200500517 ↗
25. 10.1002/adma.20080326710.1002/adma.200803267 ↗
26. Jackson J. D. · 1975
27. Peck E. R. · 1953
28. 10.1098/rspa.1927.002510.1098/rspa.1927.0025 ↗
29. 10.1103/PhysRevE.49.143910.1103/PhysRevE.49.1439 ↗
30. 10.1063/1.173426310.1063/1.1734263 ↗
31. 10.1103/PhysRevB.73.19331010.1103/PhysRevB.73.193310 ↗
32. 10.1002/adfm.20060111610.1002/adfm.200601116 ↗
33. 10.1103/PhysRevB.73.04130210.1103/PhysRevB.73.041302 ↗
34. 10.1021/jp962581d10.1021/jp962581d ↗
35. 10.1016/S0022-0728(02)00764-710.1016/S0022-0728(02)00764-7 ↗
36. 10.1103/PhysRevLett.72.124010.1103/PhysRevLett.72.1240 ↗
37. 10.1103/PhysRevB.50.1135510.1103/PhysRevB.50.11355 ↗
38. 10.1002/pssb.20054146310.1002/pssb.200541463 ↗
39. 10.1063/1.180277110.1063/1.1802771 ↗
40. 10.1103/PhysRevB.43.199310.1103/PhysRevB.43.1993 ↗
41. 10.1103/PhysRevB.78.07510910.1103/PhysRevB.78.075109 ↗
42. Most calculations were carried out using a grid spacing of 0.45 a.u which was found to be sufficient for the converging the polarizability values. Some cases were also calculated using a grid spacing of 0.35 or 0.3 a.u in order to verify convergence. Denser lateral coverage monolayers (above 0.25 ML) were calculated using a 3 × 3 × 1 k‐point sampling grid For sparser lateral coverage monolayers (0.25 ML and below) only the gamma point was sampled. A vacuum separation of at least 33 a.u. between the molecular slab and the isolatedz‐axis boundaries was used throughout.
43. 10.1103/PhysRevLett.78.480510.1103/PhysRevLett.78.4805 ↗
44. 10.1103/PhysRevB.46.1606710.1103/PhysRevB.46.16067 ↗
45. 10.1103/PhysRevB.51.401410.1103/PhysRevB.51.4014 ↗
46. 10.1103/PhysRevB.59.1230110.1103/PhysRevB.59.12301 ↗
47. 10.1103/PhysRevB.60.155110.1103/PhysRevB.60.1551 ↗
48. 10.1016/j.apsusc.2006.03.05210.1016/j.apsusc.2006.03.052 ↗
49. 10.1021/ja00764a01010.1021/ja00764a010 ↗
50. 10.1021/jp075643g10.1021/jp075643g ↗
51. 10.1039/tf928240009210.1039/tf9282400092 ↗
52. 10.1002/adma.20070168110.1002/adma.200701681 ↗
53. 10.1103/PhysRevLett.83.69410.1103/PhysRevLett.83.694 ↗
54. 10.1103/PhysRevLett.93.21300210.1103/PhysRevLett.93.213002 ↗
55. 10.1103/PhysRevB.80.15310110.1103/PhysRevB.80.153101 ↗
56. 10.1080/1478644010863561810.1080/14786440108635618 ↗
57. 10.1080/1478644020863563110.1080/14786440208635631 ↗
58. 10.1080/1478644060863566610.1080/14786440608635666 ↗
59. 10.1063/1.171232510.1063/1.1712325 ↗
60. 10.1103/PhysRevLett.96.19680610.1103/PhysRevLett.96.196806 ↗
61. 10.1103/PhysRevB.77.04511310.1103/PhysRevB.77.045113 ↗
62. 10.1021/jp064284710.1021/jp0642847 ↗
63. 10.1103/PhysRevB.74.07341410.1103/PhysRevB.74.073414 ↗
64. 10.1021/jp068855710.1021/jp0688557 ↗
65. 10.1002/adfm.20070130510.1002/adfm.200701305 ↗
66. 10.1021/jz900051710.1021/jz9000517 ↗
67. Throughout our analysis we assume perfect order of the monolayer and a uniform field. If the monolayer posesses domains holes or islands (the latter being particularly likely at low coverage) a careful analysis of finite domain and edge effects should be performed in the spirit of the ideas presented in Ref. [43].
68. In this work we defineLas the inter‐monomer distance multiplied by the number of momomers. One could have used alternative definitions (e.g the distance between the upper‐most and the lower‐most hydrogens in the oligomer chain). In all but the shortest chains this would only have a slight effect on the precise numerical values of the results. The same also holds for the volume of the molecule as it is directly dependent onL.
69. 10.1103/PhysRev.183.83410.1103/PhysRev.183.834 ↗
70. 10.1016/0021-9797(74)90002-210.1016/0021-9797(74)90002-2 ↗
71. 10.1016/S0009-2614(01)00175-010.1016/S0009-2614(01)00175-0 ↗
72. 10.1016/0040-6090(92)90172-810.1016/0040-6090(92)90172-8 ↗
73. For a long oligomer in the dense layer the per‐monomer polarizability deduced from Figure 1b is only 7.0Å3 which is somewhat higher than the per‐ring polarizability of 6.17Å3obtained for the dense benzene monolayer but still notably smaller than the 12.45Å3obtained for the isolated benzene ring.
74. 10.1103/PhysRev.99.112810.1103/PhysRev.99.1128 ↗
75. 10.1016/S0031-8914(58)95803-810.1016/S0031-8914(58)95803-8 ↗
76. 10.1103/PhysRevB.74.03340710.1103/PhysRevB.74.033407 ↗
77. If one is interested in the response to the external electrical field—thenis the relevant quantity if one is interested the dielectric constant thenχis the relevant quantity.
78. 10.1063/1.163002910.1063/1.1630029 ↗
79. Recall that just like the original CM expression this result only applies to a slab with a cubic crystalline geometry. If the unit cell is different than a cubic one a modified expression should be used.
80. We note that if one similarly defines a modified polarizability as dipole changes due toEint the CM limit would be restored for the polarizability as well. We choose not do so here because we wish to emphasize that polarizability definitions aside there is no enhancement of actual polarization with respect to that of an isolated unit no matter how thick the molecular film is.
81. This expression forεis in a sense an average value even for the case of a dense monolayer. This is because the top‐most and bottom‐most layers possess a polarizability value different from that of other layers. Still this average value reflects the total capacitance of the system.
82. 10.1103/PhysRevB.75.20512110.1103/PhysRevB.75.205121 ↗
83. 10.1038/nphys118510.1038/nphys1185 ↗
84. Because biphenyl is a relatively short molecule the results of Figure 3a are somewhat sensitive to the exact choice ofL(see Ref. [52]). TakingL = 2*4.34Å (i.e twice the inter‐phenyl distance) as in the figure yields a different result than that obtained by takingL = 9.35Å which is the total distance from the bottom H to the top H in the molecule. The formerLyieldsε = 4.64 from the DFT calculation andε = 3.91 from the electrostatic model. The latterLyieldsε = 3.68 from the the DFT calculation andε = 3.25 from the electrostatic model.
85. A possible correction of the depolarization term [first term on the right hand side of Eq. (10)] due to the finite size of the benzene ring is presented in Section E of the Supporting Information. Its application shows that indeed most of the polarizability underestimation of the model at high densities can be attributed to this effect.
86. The negative value arises because the denominator in Equation (3) becomes negative. This is known as the “dipole catastrophe” and is a direct result of the point dipole approximation being an inadequate description for the actual physical system.
87. 10.1103/PhysRevB.74.16532310.1103/PhysRevB.74.165323 ↗
88. 10.1073/pnas.070147210410.1073/pnas.0701472104 ↗
89. 10.1002/pol.1958.120281181310.1002/pol.1958.1202811813 ↗