Beruflich Dokumente
Kultur Dokumente
https://doi.org/10.1007/s00894-017-3521-7
ORIGINAL PAPER
Abstract
Effects of size, shape, and pyrene doping on electronic properties of graphene nanoflakes (GNFs) were theoretically investigated
using density functional theory method with PBE, B3PW91, and M06-2X functionals and cc-pVDZ basis set. Two shapes of
zigzag GNFs, hexagonal (HGN) and rhomboidal (RGN), were considered. The energy band gap of GNF depends on shape and
decreases with size. The HGN has larger band gap energy (1.23–3.96 eV) than the RGN (0.13–2.12 eV). The doping of pyrene
and pyrene derivatives on both HGN and RGN was also studied. The adsorption energy of pyrene and pyrene derivatives on GNF
does not depend on the shape of GNFs with energies between 21 and 27 kcal mol−1. The substituent on pyrene enhances the
binding to GNF but the strength does not depend on electron withdrawing or donating capability. The doping by pyrene and
pyrene derivatives also shifts the HOMO and LUMO energies of GNFs. Both positive (destabilizing) and negative (stabilizing)
shifts on HOMO and LUMO of GNFs were seen. The direction and magnitude of the shift do not follow the electron withdrawing
and donating capability of pyrene substituents. However, only a slight shift was observed for doped RGN. A shift of 0.19 eV was
noticed for HOMO of HGN doped with 1-aminopyrene (pyNH2) and of 0.04 eV for LUMO of HGN doped with 1-
pyrenecarboxylic acid (pyCOOH).
Keywords Pristine graphene . Pyrene doping . Diode . Graphene nanoflake . DFT calculations
interactions with aromatic molecules through π-π interaction. The knowledge could provide a new insight into the de-
Because pyrene could stably bind to the graphene through sign of a complete circuit of donor- and acceptor-type
strong π-π interactions [31–34] and electron withdrawing or graphene diodes.
donating substituents can be attached to pyrene with simple
reactions and facile purification [35, 36]. They then doped
rGO with pyrene and pyrene derivatives and successfully de- Computational models and methods
vised a bottom-gated field-effect transistor. Several theoretical
studies have been carried out regarding the properties of Two different shapes of zigzag-edge GNFs, HGN and
physisorbed graphene. Deng et al. [37] using density function- RGN, were selected for the study. The hexagonal carbon
al theory (DFT) suggested that the computed band gap energy ring is the core structure for the construction of HGN,
of fully H-terminated armchair- and zigzag-edge graphene which expands along the C2 symmetry axis as C6 exten-
decreases with increasing size of nanosheet. Zhou and sion within the D6h point group. Thus, it expands along
Zhang [38] theoretically investigated the adsorption behavior the C2’ axis within D6h symmetry as zigzag circles around
of a series of benzene derivatives physisorbed on graphene. the nanoflake. The total number of C atoms in HGN is
They suggested that adsorption energies were insensitive to 6n2, where n is the peripheral zigzag circle of HGN. Our
the adsorption sites (hollow, bridging, on-top), since energy HGN models were generated for n = 2–6, which contains
differences between these sites are within 1 kcal mol−1. C24H12, C54H18, C96H24, C150H30, and C216H36, respec-
Moreover, the adsorption energies for all the derivatives are tively. For RGN, the m x m model was used to generate
larger than that of benzene. This can be ascribed to the increas- GNFs, where m is the number of benzene rings. Our RGN
ing orbital overlap of substituents with graphene. models were produced for m = 3–8, which are comprised
A graphene nanoflakes (GNF) is a finite sized of C 3 0 H 1 4 C 4 8 H 1 8 , C 7 0 H 2 2 , C 9 6 H 2 6 , C 1 2 6 H 3 0 , and
graphene. It can be prepared by both bottom up [39] and C160H34, respectively. All molecular structures were fully
top down [40–48] approaches. Whilst it has properties optimized by DFT method [52] using Perdew Burke
similar to graphene, it can be made into many sizes and Ernzerhof (PBE) functional [53] and cc-pVDZ basis set
shapes. Nagai et al. [49] have studied electrical properties [54]. Geometry optimizations were performed in gas
of functionalized GNFs with various shapes and sizes and phase using the Gaussian 09 software [55]. Illustrations
found that electrical properties of GNFs can be tuned by of the HGNs and RGNs being investigated in this study
making them the proper size and shape. Shakourian-Fard are given in Fig. 1.
et al. [50] used M06-2X/cc-pVDZ level of theory to study To address the dispersion-type interaction for the study
the effect of defect types on electronic and optical prop- of physisorption, we chose the Minnesota M06-2X func-
erties of GNFs physisorbed by ionic liquids (ILs). They tional [56] with cc-pVDZ basis set in cooperation with
reported that the increasing size of the defective GNF BSSE correction since from our previous study we found
affects the geometry of the surface and enhances the bind- that the method could yield a reasonable description for
ing affinity of ILs by about 10%. Chutia et al. [51] inves- the dispersion interaction (π-π stacking) [57]. Two differ-
tigated the properties of GNFs by modifying their surface ent adsorption sites, on-top and hollow, were investigated
chemistry and stacking conformation. They have shown for both doped HGN and RGN (Fig. 2). We chose the C96
that there is a remarkable relationship between the chem- model to study the physisorption because in this model
ical hardness with electrical conductivity of GNFs. They both HGN and RGN have the same number of carbon
also revealed that HOMO-LUMO energy gap can directly atoms. Doped GNF structures at the two positions were
be related with the chemical hardness and hence the elec- obtained by fully optimizing the structures in Fig. 2 at
trical conductivity. Like graphene, electronic properties of BSSE corrected M06-2X/cm3-pVDZ level of theory.
GNF can also be modified by doping with π-conjugated Energies of HOMO and LUMO and band gap energies (Eg)
compound. of HGNs, RGNs, and their physisorbed complexes were de-
In this work, geometries and electronic properties of termined using the DFT method with B3PW91 hybrid func-
two shapes of graphene nanoflakes, rhomboidal (RGN) tional [58, 59] and cc-pVDZ basis set. The B3PW91 hybrid
and hexagonal (HGN), at various sizes were investigated functional has been shown to provide excellent prediction for
using DFT calculations. Moreover, the physisorption on band gap energies of 70 insulating compounds [60]. Band gap
HGN and RGN by pyrene and its derivatives with various energies of both HGNs (C24H12, C54H18, C96H24, C150H30,
electron withdrawing/donating substituents and their ef- C216H36) and RGNs (C30H14, C48H18, C70H22, C96H26,
fects on electronic properties of GNF were discussed. C126H30, C160H34) were estimated by:
This research aimed to understand the fundamental that
Eg ¼ ELUMO EHOMO ð2Þ
governs electronic properties of GNF and graphene re-
garding shape and size and especially pyrene doping. where E LUMO is the energy of the lowest unoccupied
J Mol Model (2017) 23:355 Page 3 of 9 355
Fig. 1 Hexagonal graphene nanoflakes (a) and rhomboidal graphene nanoflakes (b)
molecular orbital and EHOMO is the energy of the highest All HGNs being studied have large band gap. They could be
occupied molecular orbital. classified as semiconductors. From the relation between the
number of carbon atoms on HGN and its band gap energy,
HGN will have zero band gap when the number of carbon
atoms is beyond 480 (see Fig. S1). Thus, very large HGN will
Results and discussion have electronic properties similar to graphene. However, mod-
erately large RGNs (C > 200) have negligible band gap energy
Size and shape dependent and they show the properties of metals. Though, for small
RGNs there remains a sizable band gap and they can be made
HOMO and LUMO energies of five HGNs, C24H12, C54H18,
into semiconductors.
C96H 24, C150H30, and C216H 36 and six RGNs, C30H 14,
To understand why HOMO and LUMO energies of GNFs
C48H18, C70H22, C96H26, C126H30, and C160H34 are displayed
depend on shape, we compared HOMO and LUMO orbitals
in Fig. 3. The relation between band gap energies and numbers
of C96H24 (HGN) and C96H26 (RGN), Fig. S2 and S3. HOMO
of carbon atoms are given in Fig. S1. Corresponding band gap
and LUMO orbitals of HGN look very distinct; however,
energies for HGNs are 3.96, 2.76, 2.06, 1.58, and 1.23 eV,
HOMO and LUMO orbitals of RGN appear to be similar.
respectively and for RGNs are 2.12, 1.16, 0.62, 0.34, 0.20,
Thus, while HGN has large band gap, HOMO and LUMO
and 0.13 eV, respectively. It can be seen that HOMO and
of RGN have similar energy and hence negligible band gap.
LUMO energies and hence band gap energies depend on
shape and size of GNFs. The decreasing of band gap energy
with the increasing of size was observed for both HGN and Physisorption
RGN consistent with the result of Deng et al. [37]. Thus, the
observation could be explained by the increment of free π Adsorption energies (Ead1) and BSSE corrected adsorption
electrons in occupied orbitals perpendicular to the GNF plane energies (Ead2) of pyrene and seven pyrene derivatives
(pz orbital). RGN generally has a band gap smaller than HGN. adsorbed on HGN and RGN were given in Table 1. The
Fig. 2 Pyrene doped HGN (a) and RGN (b) at on-top and hollow adsorbed positions
355 Page 4 of 9 J Mol Model (2017) 23:355
Fig. 4 Optimized structures for adsorption complexes of pyrene/seven pyrene derivatives and HGN (top view)
substituents on pyrene and the HOMO/LUMO energy shift. as their magnitude do not have any relation with the electron
For RGNs, only a small shift around 0.01–0.08 eV were withdrawing/donating capability of the substituent. Thus, the
noted. electronic effect on pyrene ring does not have any direct in-
Thus, it could be concluded that the doping has a minimal fluence on the electronic properties of GNF. This conclusion
effect on the electronic structure of RGN. However, it should surprised us. To see whether the electronic effect of the sub-
be noticed that the doping with pyNH2 (the strongest electron stituent on pyrene affects its electronic structure, we carried
donating substituents) provides the largest positive shift. For out B3PW91/cc-pVDZ calculations for pyrene and seven
HGN, the more pronounced positive shift was observed on pyrene dervatives. More pronouced shifts on HOMO (0.01–
HOMO energy when doping with pyNH2 (0.19 eV), while 0.55 eV) and LUMO (0.03–1.03 eV) as compared to those of
HGN with pyCOOH doping provides the smallest postive physisorbed complexes were found, see Table S3. The nega-
shift (0.04 eV) on LUMO energy. Likewise, the positive and tive shift was noticed to be associated with pyrene with elec-
the negative shifts on HOMO/LUMO energy of HGN as well tron withdrawing substituents, while the positive shift was
Fig. 5 Optimized structures for adsorption complexes of pyrene/seven derivatives and RGN (top view)
355 Page 6 of 9 J Mol Model (2017) 23:355
Table 4 Sum of atomic net charge of pyrene/pyrene derivatives in HGN 3.96 eV). Thus, large RGNs have electronic properties closer
and RGN complexes
to graphene. Very large HGNs with twice the number of car-
Adsorbates Net charge bon atoms as large RGNs will also have a close-to-zero band
gap. Pyrene and all pyrene derivatives have large binding
HGN complex RGN complex affinity with GNFs (21–27 kcal mol−1, BSSE corrected).
PyH −0.0191 −0.01548
Adsorption energies with pyrene and pyrene derivatives are
similar for RGNs and HGNs. The strenght of adsorption does
PyNO2 −0.01755 −0.01643
not depend on the type of substituent (whether it is electron
PySO3H −0.01253 −0.01423
withdrawing or donating). All pyrene derivatives adsorbed to
PyCOOH −0.01559 −0.01348
GNF stronger than pyrene due to the orbital overlap of the
PyNHCOCH3 −0.01459 −0.01030
substituent. Doping can shift HOMO and LUMO of GNF
PyCH3 −0.01643 −0.01355
but the magnitude and the direction of the shift does not follow
PyOH −0.01045 −0.00800
the electron withdrawing/donating capability of the substitu-
PyNH2 −0.00699 −0.01085
ent. Thus, the shift of HOMO and LUMO in GNFs cannot be
controlled by the physisorption since all pyrene dervatives act
as electron withdrawing groups. Therefore, the substituent on
pyrene does not provide a substantially different electronic
effect on GNF. A slight shift in HOMO and LUMO was ob-
served for RGN, while a more significant shift is noticed for
HGN doped with pyNH 2 (0.19 eV). HGN doped with
pyCOOH provides the least positive shift on LUMO
(0.04 eV). Therefore, HGN-pyNH2 and HGN-pyCOOH can
be made into a diode.
6. Ghosh S, An X, Shah R, Rawat D, Dave B, Kar S, Talapatra S by noncovalent functionalization: effects of organic donor, acceptor
(2012) Effect of 1-pyrene carboxylic-acid functionalization of and metal atoms. Nanotechnology 21:065201–065207
graphene on its capacitive energy storage. J Phys Chem C 116: 26. Zhang ZX, Huang HL, Yang XM, Zang L (2011) Tailoring elec-
20688–20693 tronic properties of graphene by pi-pi stacking with aromatic mol-
7. Mohammed M, Li Z, Cui J, Chen TP (2012) Junction investigation ecules. J Phys Chem Lett 2:2897–2905
of graphene/silicon Schottky diodes, nanoscale. Res Lett 7:302– 27. Wang WZ, Sun T, Zhang Y, Wang YB (2014) Substituent effects in
307 the pi...pi interaction between graphene and benzene: an indication
8. Gowtham S, Scheicher RH, Ahuja R, Pandey R, Karna SP (2007) for the noncovalent functionalization of graphene. Comput Theor
Physisorption of nucleobases on graphene: density-functional cal- Chem 1046:64–69
culations. Phys Rev B 76:033401–033404 28. Wang WZ, Sun T, Zhang Y, Wang YB (2015) Benchmark calcula-
9. Lee J-H, Choi Y-K, Kim H-J, Scheicher RH, Cho JH (2013) tions of the adsorption of aromatic molecules on graphene. J
Physisorption of DNA nucleobases on H-BN and graphene: Comput Chem 36:1763–1771
Vdw-corrected DFT calculations. J Phys Chem C 117:13435– 29. Yin YF, Cervenka J, Medhekar NV (2015) Tunable hybridization
13441 between electronic states of graphene and physisorbed hexacene. J
10. Cortés-Arriagada D, Sanhueza L, Santander-Nelli M (2013) Phys Chem C 119:19526–19534
Modeling the physisorption of bisphenol a on graphene and 30. Lee J, Hwang E, Lee E, Seo S, Lee H (2012) Tuning of n- and p-
graphene oxide. J Mol Model 19:3569–3580 type reduced graphene oxide transistors with the same molecular
11. Duy L, Abdelkader K, Elsebeth S, Per H, Talat SR (2012) backbone. Chem Eur J 18:5155–5159
Physisorption of nucleobases on graphene: a comparative van der 31. Chen RJ, Zhang YG, Wang DW, Dai HJ (2001) Noncovalent side-
Waals study. J Phys Condens Matter 24:424210 wall functionalization of single-walled carbon nanotubes for protein
12. Martins TB, Miwa RH, da Silva AJR, Fazzio A (2007) Electronic immobilization. J Am Chem Soc 123:3838–3839
and transport properties of boron-doped graphene nanoribbons. 32. Besteman K, Lee JO, Wiertz FGM, Heering HA, Dekker C (2003)
Phys Rev Lett 98:196803–196806 Enzyme-coated carbon nanotubes as single-molecule biosensors.
13. Wang XR, Li XL, Zhang L, Yoon Y, Weber PK, Wang HL, Guo J, Nano Lett 3:727–730
Dai HJ (2009) N-doping of graphene through electrothermal reac- 33. Xu YX, Bai H, Lu GW, Li C, Shi GQ (2008) Flexible graphene
tions with ammonia. Science 324:768–771 films via the filtration of water-soluble noncovalent functionalized
14. Wei DC, Liu YQ, Wang Y, Zhang HL, Huang LP, Yu G (2009) graphene sheets. J Am Chem Soc 130:5856–5857
Synthesis of N-doped graphene by chemical vapor deposition and 34. An XH, Simmons TJ, Shah R, Wolfe C, Lewis KM, Washington M,
its electrical properties. Nano Lett 9:1752–1758 Nayak SK, Talapatra S, Kar S (2010) Stable aqueous dispersions of
15. Panchokarla LS, Subrahmanyam KS, Saha SK, Govindaraj A, noncovalently functionalized graphene from graphite and their mul-
Krishnamurthy HR, Waghmare UV, Rao CNR (2009) Synthesis, tifunctional high-performance applications. Nano Lett 10:4295–
structure, and properties of boron- and nitrogen-doped graphene. 4301
Adv Mater 21:4726–4730 35. Babu P, Sangeetha NM, Vijaykumar P, Maitra U, Rissanen K, Raju
16. Li X, Wang H, Robinson JT, Sanchez H, Diankov G, Dai H (2009) AR (2003) Pyrene-derived novel one- and two-component
Simultaneous nitrogen doping and reduction of graphene oxide. J organogelators. Chem Eur J 9:1922–1932
Am Chem Soc 131:15939–15944 36. Soustek P, Michl M, Almonasy N, Machahcky O, Dvorak M,
17. Choi J, Kim KJ, Kim B, Lee H, Kim S (2009) Covalent Lycka A (2008) The synthesis and fluorescence of N-substituted
functionalization of epitaxial graphene by azidotrimethylsilane. J 1- and 2-aminopyrenes. Dyes Pigments 78:139–147
Phys Chem C 113:9433–9435 37. Deng JP, Chen WH, Chiu SP, Lin CH, Wang BC (2014) Edge-
18. Elias DC, Nair RR, Mohiuddin TMG, Morozov SV, Blake P, termination and core-modification effects of hexagonal Nanosheet
Halsall MP, Ferrari AC, Boukhvalov DW, Katsnelson MI, Geim graphene. Molecules 19:2361–2373
AK, Novoselov KS (2009) Control of graphene’s properties by 38. Zhou PP, Zhang RQ (2015) Physisorption of benzene derivatives
reversible hydrogenation: evidence for Graphane. Science 323: on graphene: critical roles of steric and stereoelectronic effects of
610–613 the substituent. Phys Chem Chem Phys 17:12185–12193
19. Balog R, Jorgensen B, Nilsson L, Andersen M, Rienks E, Bianchi 39. Li H, Zou L, Pan L, Sun Z (2010) Using graphene nano-flakes as
M, Fanetti M, Laegsgaard E, Baraldi A, Lizzit S, Sljivancanin Z, electrodes to remove ferric ions by capacitive deionization. Sep
Besenbacher F, Hammer B, Pedersen TG, Hofmann P, Hornekaer L Purif Technol 75:8–14
(2010) Bandgap opening in graphene induced by patterned hydro- 40. Berger C, Song Z, Li X, Wu X, Brown N, Naud C, Mayou D, Li T,
gen adsorption. Nat Mater 9:315–319 Hass J, Marchenkov AN, Conrad EH, First PN, de Heer WA (2006)
20. Choi J, Lee H, Kim KJ, Kim B, Kim S (2010) Chemical doping of Electronic confinement and coherence in patterned epitaxial
epitaxial graphene by organic free radicals. J Phys Chem Lett 1: graphene. Science 312:1191–1196
505–509 41. Schedin F, Geim AK, Morozov SV, Hill EW, Blake P, Katsnelson
21. Chen W, Chen S, Qi DC, Gao XY, Wee ATS (2007) Surface transfer MI, Novoselov KS (2007) Detection of individual gas molecules
p-type doping of epitaxial graphene. J Am Chem Soc 129:10418– adsorbed on graphene. Nat Mater 6:652–655
10422 42. Molitor F, Güttinger J, Stampfer C, Dröscher S, Jacobsen A, Ihn T,
22. Voggu R, Das B, Rout CS, Rao CNR (2008) Effects of charge Ensslin K (2011) Electronic properties of graphene nanostructures.
transfer interaction of graphene with electron donor and acceptor J Phys Condens Matter 23:243201–243215
molecules examined using Raman spectroscopy and cognate tech- 43. Tapas S, Das I, Banerjee S (2009) Contribution of energy-gap in the
niques. J Phys Condens Matter 20:472204–472208 ferromagnetic spin–wave spectrum on magnetocaloric parameters
23. Grimme S (2008) Do special noncovalent pi-pi stacking interac- of Ceru 2 Ge 2. J Phys Condens Matter 21:026010
tions really exist? Angew Chem 47:3430–3434 44. Neubeck S, Ponomarenko LA, Freitag F, Giesbers AJM, Zeitler U,
24. Dong XC, Fu DL, Fang WJ, Shi YM, Chen P, Li LJ (2009) Doping Morozov SV, Blake P, Geim AK, Novoselov KS (2010) From one
single-layer graphene with aromatic molecules. Small 5:1422–1426 electron to one hole: quasiparticle counting in graphene quantum
25. Zhang YH, Zhou KG, Xie KF, Zeng J, Zhang HL, Peng Y (2010) dots determined by electrochemical and plasma etching. Small 6:
Tuning the electronic structure and transport properties of graphene 1469–1473
J Mol Model (2017) 23:355 Page 9 of 9 355
45. Datta SS, Strachan DR, Khamis SM, Johnson ATC (2008) 55. Frisch M, Trucks G, Schlegel HB, Scuseria G, Robb M, Cheeseman
Crystallographic etching of few-layer graphene. Nano Lett 8: J, Scalmani G, Barone V, Mennucci B, Petersson G et al. (2009)
1912–1915 Gaussian 09, revision a.02. Gaussian Inc., Wallingford, p 200
46. Ci L, Xu Z, Wang L, Gao W, Ding F, Kelly KF, Yakobson BI, 56. Zhao Y, Truhlar DG (2008) The M06 suite of density Functionals for
Ajayan PM (2008) Controlled nanocutting of graphene. Nano main group thermochemistry, thermochemical kinetics, noncovalent
Res. 1:116–122 interactions, excited states, and transition elements: two new functionals
47. Ci L, Song L, Jariwala D, Elías AL, Gao W, Terrones M, Ajayan and systematic testing of four M06-class functionals and 12 other func-
PM (2009) Graphene shape control by multistage cutting and trans- tionals. Theor Chem Accounts 120:215–241
fer. Adv Mater 21:4487–4491 57. Rungnim C, Chanajaree R, Rungrotmongkol T, Hannongbua S,
48. Campos LC, Manfrinato VR, Sanchez-Yamagishi JD, Kong J, Kungwan N, Wolschann P, Karpfen A, Parasuk V (2016) How
Jarillo-Herrero P (2009) Anisotropic etching and nanoribbon for- strong is the edge effect in the adsorption of anticancer drugs on a
mation in single-layer graphene. Nano Lett 9:2600–2604 graphene cluster? J Mol Model 22:85–93
49. Nagai H, Nakano M, Yoneda K, Fukui H, Minami T, Bonness S, 58. Becke AD (1993) Becke’s three parameter hybrid method using the
Kishi R, Takahashi H, Kubo T, Kamada K, Ohta K, Champagne B, LYP correlation functional. J Chem Phys 98:5648–5652
Botek E (2009) Theoretical study on third-order nonlinear optical 59. Perdew JP (1991) In: Ziesche P, Eschrig H (eds) Electronic structure
properties in hexagonal graphene nanoflakes: edge shape effect. of solids. Akademie, Berlin
Chem Phys Lett 477:355–359
60. Crowley JM, Tahir-Kheli J, Goddard WA (2016) Resolution of the
50. Shakourian-Fard M, Kamath G (2017) The effect of defect types on
band gap prediction problem for materials design. J Phys Chem Lett
the electronic and optical properties of graphene nanoflakes
7:1198–1203
physisorbed by ionic liquids. Phys Chem Chem Phys 19:4383–
4395 61. Colherinhas G, Fileti EE, Chaban VV (2015) Can inorganic salts
51. Chutia A, Sahnoun R, Deka RC, Zhu Z, Tsuboi H, Takaba H, tune electronic properties of graphene quantum dots? Phys Chem
Miyamoto A (2011) Local electronic and electrical properties of Chem Phys 17:17413–17420
functionalized graphene nanoflakes. PhyB 406:1665–1672 62. Hunter CA (1993) Arene—arene interactions: electrostatic or
52. Parr RG (1982) Density functional theory. In: Electron distributions charge transfer? Angew Chem 32:1584–1586
and the chemical bond. Springer, Boston, pp 95–100 63. Hunter CA, Sanders JKM (1990) The nature of pi-pi interactions. J
53. Paier J, Hirschl R, Marsman M, Kresse G (2005) The Perdew– Am Chem Soc 112:5525–5534
Burke–Ernzerhof exchange-correlation functional applied to the 64. Lüssem B, Riede M, Leo K (2013) Doping of organic semiconduc-
G2-1 test set using a plane-wave basis set. J Chem Phys 122: tors. Phys Status Solidi A 210:9–43
234102–234114 65. Salzmann I, Heimel G (2015) Toward a comprehensive understand-
54. Woon DE, Dunning Jr TH (1993) Gaussian basis sets for use in ing of molecular doping organic semiconductors (review). J
correlated molecular calculations. III. The atoms aluminum through Electron Spectrosc 204:208–222
argon. J Chem Phys 98:1358–1371