Radiation induced electronic trap states and local structural disorder in van~der~Waals bonded semiconductor crystals
RRadiation induced electronic trap states and local structural disorder invan der Waals bonded semiconductor crystals
Tobias Morf, ∗ Tino Zimmerling, † Simon Haas, ‡ and Bertram Batlogg § Laboratory for Solid State Physics, ETH Zurich, 8093 Zurich, Switzerland (Dated: October 16, 2018)In controlled X-ray irradiation experiments, the formation of trap states in the prototyp-ical van der Waals bonded semiconductor Rubrene is studied quantitatively for doses up to82 Gray (Gy = J kg − ). About 100 electronic trap states, located around 0 . PACS numbers: 71.20.Rv 72.80.LeKeywords: CHARGE-LIMITED CURRENT; SINGLE-CRYSTAL; TRANSPORT; MOBILITY; ENERGY;X-RAY; TRAP STATE; ORGANIC SEMICONDUCTOR
The understanding of charge traps in van der Waalsbonded semiconductors has reached new levels in recentyears and organic electronic devices’ performance hasmuch improved with promising perspectives for applica-tions. With the emerging understanding and spectralanalysis of the trap density of states (DOS) [1, 2] it ishighly desirable to quantify the interaction of environ-mental influences with van der Waals bonded semicon-ductors in terms of density and spectral distribution ofthe induced trap states. Such environmental influencesmay occur during fabrication, storage or operation of anorganic electronic device. They cover a broad range ofchemical [3], mechanical [4] and also radiative [5–7] phe-nomena.The commonly known adverse influence of traps oncharge transport [2] as well as spin diffusion length [8] iscontrasted by an unexpected positive influence of X-rayradiation in electron-beam evaporation processes on themagnetotransport in organic materials [8]. These obser-vations are a strong motivation to further investigate thedefects arising from X-ray radiation.Furthermore, after intense scintillation studies in the1970s [9–12], organic materials are considered anew fordirect X-ray detection through the photoconductivity ef-fect [6, 13]. For applications in low-cost, large-area inte-grated X-ray imaging panels it will be of central impor-tance to assess the radiation damage and whether thosedefects can be healed by thermal annealing.In this study we address the formation of electronictrap states upon X-ray irradiation in Rubrene sin-gle crystals, a grain boundary free model material forvan der Waals bonded semiconductors. The spectral den-sity of trap states (DOS) is determined by measuring ∗ [email protected] † [email protected] ‡ [email protected] § [email protected] current voltage characteristics at different temperaturesand applying temperature-dependent space-charge lim-ited current spectroscopy (TD-SCLC). The basic conceptof SCLC is electrical transport by charge carriers ther-mally excited above a certain energy separating extendedfrom localised states. No further a priori assumptions —in particular no specific transport model — are required.Due to the Fermi-Dirac statistics this excitation from lo-calised traps to delocalised conducting states takes placein a small energy window. With increasing voltage, morespace charge is injected, hence the Fermi energy E F isshifted towards the delocalised states. The trap DOS iscalculated from this differential increment. The energyscale is given by the thermal activation energy at a givenvoltage corrected by the statistical shift, which accountsfor the asymmetry of the DOS around E F . The fullprocedure is formally discussed in references 14–19 andnumerically implemented using cubic smoothing splines[20].Rubrene crystals were grown by physical vapour trans-port in high purity argon flow. The platelet-like crystalswere then laminated onto prefabricated gold electrodes,similar to the ‘flip-crystal’ technique [21]. For the vac-uum deposition of the top electrode, the samples werecooled in order to minimise the thermal load on the crys-tals. The sample layout is schematically shown in Fig. 1.Current flows along the 26 .
86 ˚A long crystallographic a axis [22–24]. The simultaneous measurement of up tofour sites on the same crystal — called channels — pro-vides verifiable results and the shielding of some of thesechannels during irradiation provides the necessary refer-ence. Furthermore, this sample arrangement (c.f. Fig. 1)enables checking of reproducibility and device stability.The SCLC measurements were performed in darknessin a cryostat’s helium atmosphere. Charge carrier injec-tion from the laminated bottom electrodes turned out tobe more efficient than from the deposited top electrodeand thus the polarity for all measurements was chosenaccordingly. Current and power limits prevent crystaldamage [25, 26] or local heating. a r X i v : . [ c ond - m a t . m t r l - s c i ] M a y For the quantitative study of the radiation damage, acrystal diffractometer served as a well-defined monochro-matic CuK α (8 keV) radiation source. The lateral inten-sity distribution (beam profile) was measured with thediffractometer’s image-plate detector and suitable Zirco-nium attenuators. For the measurement of the inten-sity in absolute units, a suitably calibrated instrumentwas kindly provided by the University Hospital Zurich.The beam’s peak intensity was 3 µ W cm − correspond-ing to a photon flux of 2 × s − cm − . Approximat-ing Rubrene as a 42:28 mixture of carbon and hydrogenthe crystal at the center of the beam absorbed a doseof (43 ±
9) Gy (= J kg − ) [27] during one hour of ex-posure. The 30 nm thick gold top contact absorbs onlyabout 1 % of the incident intensity. For comparison,a single computed tomography (CT) scan accounts forup to 10 mGy [28], typical radiotherapy doses are some10 Gy [29] and the accumulated lifetime dose of X-rayimaging sensors is a few 100 Gy [13].After first measuring the trap DOS of the pristine crys-tals, the samples were transfered to an Argon filled glasstube with a Kapton window and aligned in the X-raybeam with fluorescent marks. During irradiation, twoout of four channels were shielded by a 0 . gold contactsPb shield X−rays crystalRubrene FIG. 1: (Colour online) Schematic of the experiment. TheRubrene crystal is laminated onto prefabricated bottom elec-trodes, then the top electrode is evaporated. The lead shieldscreens part of the crystal from X-rays, thus allowing directcomparison of irradiated to unirradiated crystal sites. s h i e l d e dp r i s t i n e4 1 G y X - r a y
C r y s t a l B p r i s t i n ea n n e a l e d8 2 G y X - r a y4 1 G y X - r a y
C r y s t a l A trap DOS (cm-3eV-1)
C r y s t a l A
DOS change (1016 cm-3eV-1) e n e r g y a b o v e m o b i l i t y e d g e ( e V )
FIG. 2: (Colour online) Trap density of states and its changeupon X-ray irradiation in two different Rubrene single crys-tals. The mobility edge is chosen as the energy referencepoint. The distinct increase peaked at 0 . are clearly identified. Furthermore, the unchanged DOSin unirradiated channels reflects the stability and repro-ducibility of sample handling and measurement.In the exposed channels, the trap DOS increases byup to 8 × cm − eV − (Fig. 2, crystal B) in a nar-row energy range peaked around 0 . ∼ × cm − traps generated during one hour of X-ray irradiation. Af-ter the second hour of irradiation, the induced trap den-sity has doubled within experimental uncertainty. TheX-ray absorption length λ ≈ ∼ µ m (only 0 .
05 % ofthe photons are absorbed), and with an hourly dose of43 Gy, approximately 4 × photons are abosrbed percm creating a uniform defect density.To quantitatively compare X-ray and ion irradiationin Fig. 3 it is appropriate to consider the microscopicinteraction mechanism as sketched in the insets. An ab-sorbed X-ray photon will deposit its full energy of 8 keVin a single primary event causing a cascade of secondaryevents which in turn create numerous microscopic de-fects. In contrast, every proton of 1 MeV experiences ap-proximately 14 primary interactions on its way througha 1 µ m thick crystal, each time transfering ≈ . ∼ . ∼
100 traps. A central result of this study is: per pri-mary interaction event, X-rays are found to be 100 to1000 times more effective than ions in trap generation.This is attributed to the shower of secondary events fol-lowing every photon absorption. These showers spatiallydistribute the energy in the crystal as opposed to thepoint-like event of an ion interaction. These secondaryevents apparently carry enough energy to create defectsand their number would account for the 100- to 1000-folddefect creation rate.The spectral distribution of the additional traps sug-gests their common microscopic origin: they are peaked ≈ . ≈ . covered surfacefree surfaceCrystal B X - r a y s8 k e V induced traps (cm-3) p r i m a r y e v e n t s ( c m - 3 ) p r o t o n s1 M e V Crystal A
Rubrene crystal
FIG. 3: (Colour online) Trap density in Rubrene crystals ir-radiated by X-rays (red) or protons (blue, [7]). For each pri-mary interaction event, protons create ∼ . . ∼ traps. The 100 to 1000 timeshigher trap creation efficiency of X-rays is attributed to sec-ondary events. Trap creation by ions in covered crystals satu-rates due to re-attachement of hydrogen. The insets schemat-ically show the energy deposition processes. An X-ray photonwill either pass the crystal undisturbed or deposit its full en-ergy in a single event creating numerous secondary events.On the other hand, every ion will experience several interac-tions with (mainly) target electrons every time depositing afraction of its initial energy. ne w s t a t e s ( e V - c m - ) energy above mobility edge (eV)0 0.1 0.2 0.3 0.4 0.5 Rubrene crystal +He ionsprotonsX-rays (B)X-rays (A)
FIG. 4: (Colour online) Spectral distribution of disorder in-duced states in Rubrene single crystals after Helium ion ir-radiation (blue, [7]), proton irradiation (magenta, [7]) andX-ray exposure (green and red, this study). The inset showsan exagerated sketch of local disorder as the possible originof these traps.
A motivation to consider structural defects as a possi-ble cause is the partial recovery after thermal annealingat moderate temperatures. Annealing of structural de-fects even at room temperature has been shown to takeplace in Pentacene thin films always kept in high vacuum[40]. Similarly, intercalation with inert gases (N andAr) induces trap states at a similar energy [34]. Alsoin organic polymer solar cells, radiation induced dam-age recovering by annealing has been observed and inter-preted in terms of hydrogen detachment and rearrange-ment [35, 37, 41]. Particularly interesting is the bend-ing of a pentacene molecule when two hydrogen atomsare attached to the initially flat entity, creating localisedelectronic states [36]. Similar detailed calculations forRubrene [42] with attached oxygen of hydrogen in variousconfigurations give no evidence for new electronic stateswithin the few tenths of eV above the HOMO band acces-sible in the experiment. However, they reveal a slight ro-tation of a phenyl side-group. While the authors are notaware of a corresponding calculation involving a missinghydrogen in Rubrene, the previous experimental obser-vations [7] clearly suggest hydrogen detachment to be akey step in trap state formation.The discussion about the microscopic nature of inten-tionally induced defect states in organic semiconductorcrystals is now stimulated by (a) the formation of elec-tronically active states ≈ . π electron system (e.g. H detachment, O orOH attachement), and on modifications of the molecule’s structure and its environment in the crystal. Combininglocal probes and macroscopic transport measurementswill produce such new insights.In conclusion, we have quantitatively studied the for-mation of bulk trap states in van der Waals bonded sin-gle crystals by X-ray irradiation. For each absorbed8 keV photon, approximately 100 trap states are cre-ated, while proton irradiation [7] generates up to 1 trapstate per primary interaction. The spectral trap distri-bution is peaked near 0 . ≈ .
15 eV wide. Very similar trap distribu-tions which can be partially annealed are produced byhydrogen- and oxygen-related chemical defects but alsowhen van der Waals bonded semiconductors are locallydisturbed by proton or Helium ion radiation [7] or bypenetration of oxygen or chemically neutral nitrogen orargon [2, 3, 33, 34]. This formation of energetically verysimilar trap states by a wide range of treatments and theobservation of partial annealing of these states set theframework for future studies focussing in the respectivecontributions of chemical and structural defects.
Acknowledgments
The authors thank Carlo Bernasconi from the Lab-oratory of Crystallography, ETH Zurich for access tothe X-ray diffractometer. Stephan Kl¨ock and J´erˆomeKrayenb¨uhl from University Hospital Zurich are great-fully acknowledged for their help in measuring the X-rayintensity and Kurt Mattenberger for support with vari-ous technical issues. [1] A. Salleo,
Electronic Traps in Organic Semiconductors (Wiley-VCH Verlag GmbH & Co. KGaA, 2013), chap. 14,pp. 341–380, ISBN 9783527650965.[2] W. L. Kalb, S. Haas, C. Krellner, T. Mathis, and B. Bat-logg, Phys. Rev. B , 155315 (2010).[3] C. Krellner, S. Haas, C. Goldmann, K. Pernstich,D. Gundlach, and B. Batlogg, Phys. Rev. B: Condens.Matter Mater. Phys. , 245115 (2007).[4] T. Sekitani, Y. Kato, S. Iba, H. Shinaoka, T. Someya,T. Sakurai, and S. Takagi, Appl. Phys. Lett. , 073511(2005).[5] A. Quaranta, A. Vomiero, S. Carturan, G. Maggioni, andG. Della Mae, Synth. Met. , 275 (2003).[6] C. R. Newman, H. Sirringhaus, J. C. Blakesley, andR. Speller, Appl. Phys. Lett. , 142105 (2007).[7] T. Zimmerling, K. Mattenberger, M. D¨obeli, M. J. Si-mon, and B. Batlogg, Phys. Rev. B , 134101 (2012).[8] J. Rybicki, R. Lin, F. Wang, M. Wohlgenannt, C. He,T. Sanders, and Y. Suzuki, Phys. Rev. Le , 076603(2012).[9] S. Weisz, A. Cobas, P. E. Richardson, H. H. Szmant, andS. Trester, J. Chem. Phys. , 1364 (1966).[10] J. Birks, Proc. Phys. Soc. A , 874 (1951).[11] K. Yokoi and Y. Ohba, Chem. Phys. Lett. , 560 (1978).[12] N. Shiomi, J. Phys. Soc. Jpn. , 1177 (1967).[13] J. Blakesley, P. Keivanidis, M. Campoy-Quiles, C. New-man, Y. Jin, R. Speller, H. Sirringhaus, N. Greenham,J. Nelson, and P. Stavrinou, Nucl. Instrum. MethodsPhys. Res., Sect. A , 774 (2007).[14] O. Zmeˇskal, F. Schauer, and S. Neˇsp˚urek, J. Phys. C:Solid State Phys. , 1873 (1985).[15] F. Schauer, S. Nˇesp˚urek, and O. Zmeˇskal, J. Phys. C:Solid State Phys. , 7231 (1986).[16] F. Schauer, S. Nˇesp˚urek, and H. Valeri´an, J. Appl. Phys. , 880 (1996).[17] F. Schauer, R. Novotny, and S. Neˇsp˚urek, J. Appl. Phys. , 1244 (1997).[18] S. Neˇsp˚urek, O. Zmeˇskal, and F. Schauer, Phys. StatusSolidi A , 619 (1984).[19] D. Braga, N. Battaglini, A. Yassar, G. Horowitz,M. Campione, A. Sassella, and A. Borghesi, Phys. Rev. B , 115205 (2008).[20] C. de Boor, A practical guide to splines , vol. 27 of
Appliedmathematical science (Springer Verlag, New York, 1987).[21] J. Takeya, C. Goldmann, S. Haas, K. Pernstich, B. Ket-terer, and B. Batlogg, J. Appl. Phys. , 5800 (2003).[22] O. D. Jurchescu, A. Meetsma, and T. T. Palstra, Acta Crystallogr., Sect. B: Struct. Sci. , 330 (2006).[23] E. Menard, A. Marchenko, V. Podzorov, M. E. Gershen-son, D. Fichou, and J. A. Rogers, Adv. Mater. , 1552(2006).[24] T. Minato, H. Aoki, H. Fukidome, T. Wagner, andK. Itaya, Appl. Phys. Lett. , 093302 (2009).[25] J. Srour, G. Vendura, D. Lo, C. Toporow, M. Dooley,R. Nakano, and E. King, IEEE T. Nucl. Sci. , 2624(1998).[26] J. Srour, J. Palko, D. Lo, S. Liu, R. Mueller, and J. No-cerino, IEEE T. Nucl. Sci. , 3300 (2009).[27] J. Hubbell and S. Seltzer, Tables of x-ray mass at-tenuation coefficients and mass energy-absorption coef-ficients , URL http://physics.nist.gov/PhysRefData/XrayMassCoef/cover.html .[28] D. J. Brenner and E. J. Hall, N. Engl. J. Med. , 2277(2007).[29] J. Krayenb¨uhl, private communications.[30] M. Pope and C. E. Swenberg,
Electronic processes in or-ganic crystals and polymers (Oxford University Press,1999).[31] J. Ziegler, J. Biersack, and U. Littmark,
SRIM The Stop-ping and Range of Ions in Solids (Pergamon Press, New York, 1985), URL .[32] R. Barillon and T. Yamauchi, Nucl. Instrum. MethodsPhys. Res., Sect. B , 336 (2003).[33] W. L. Kalb, K. Mattenberger, and B. Batlogg, Phys. Rev.B , 035334 (2008).[34] F. Bussolotti, S. Kera, K. Kudo, A. Kahn, and N. Ueno,Phys. Rev. Lett. , 267602 (2013).[35] R. Street, J. Northrup, and B. Krusor, Phys. Rev. B:Condens. Matter Mater. Phys. , 205211 (2012).[36] J. E. Northrup and M. L. Chabinyc, Phys. Rev. B ,041202(R) (2003).[37] J. E. Northrup, Applied Physics Express , 121601(2013).[38] G. Heppell and R. Hardwick, Trans. Faraday Soc. ,2651 (1967).[39] C. Zorn, Nucl. Phys. B, Proc. Suppl. , 377 (1993).[40] W. L. Kalb, F. Meier, K. Mattenberger, and B. Batlogg,Phys. Rev. B: Condens. Matter Mater. Phys. , 184112(2007).[41] K. Nakagawa and N. Itoh, Chem. Phys. , 461 (1976).[42] L. Tsetseris and S. Pantelides, Phys. Rev. B: Condens.Matter Mater. Phys.78