# Fileset

[Journal of Applied Physics--Oxygen incorporation effects on the structural and thermoelectric properties of copper(I) iodide.pdf](https://mdr.nims.go.jp/filesets/047328fb-73ae-4bd1-b994-2ccd0d3d5ba2/download)

## Creator

Martin Markwitz, Niall Malone, [Song Yi Back](https://orcid.org/0009-0000-8890-1484), Alexander Gobbi, Jake Hardy, Peter P. Murmu, [Takao Mori](https://orcid.org/0000-0003-2682-1846), Ben J. Ruck, John V. Kennedy

## Rights

[Creative Commons BY Attribution 4.0 International](https://creativecommons.org/licenses/by/4.0/)

## Other metadata

[Oxygen incorporation effects on the structural and thermoelectric properties of copper(I) iodide](https://mdr.nims.go.jp/datasets/69518d78-990a-4807-9754-41b5fbe7cd05)

## Fulltext

Oxygen incorporation effects on the structural and thermoelectric properties of copper(I) iodideViewOnlineExportCitationRESEARCH ARTICLE |  NOVEMBER 22 2024Oxygen incorporation effects on the structural andthermoelectric properties of copper(I) iodideMartin Markwitz   ; Niall Malone  ; Song Yi Back  ; Alexander Gobbi  ; Jake Hardy  ;Peter P. Murmu  ; Takao Mori  ; Ben J. Ruck  ; John V. Kennedy J. Appl. Phys. 136, 205702 (2024)https://doi.org/10.1063/5.0235467 CHORUSArticles You May Be Interested InKoopmanLab: Machine learning for solving complex physics equationsAPL Mach. Learn. (September 2023)Experimental realization of a quantum classification: Bell state measurement via machine learningAPL Mach. Learn. (September 2023) 22 November 2024 12:26:18https://pubs.aip.org/aip/jap/article/136/20/205702/3321820/Oxygen-incorporation-effects-on-the-structural-andhttps://pubs.aip.org/aip/jap/article/136/20/205702/3321820/Oxygen-incorporation-effects-on-the-structural-and?pdfCoverIconEvent=citejavascript:;https://orcid.org/0009-0007-6516-3571javascript:;https://orcid.org/0009-0004-0485-8271javascript:;https://orcid.org/0009-0000-8890-1484javascript:;https://orcid.org/0009-0002-4090-5093javascript:;https://orcid.org/0000-0001-5006-1918javascript:;https://orcid.org/0000-0002-0109-1798javascript:;https://orcid.org/0000-0003-2682-1846javascript:;https://orcid.org/0000-0002-3719-7375javascript:;https://orcid.org/0000-0002-9126-4997https://crossmark.crossref.org/dialog/?doi=10.1063/5.0235467&domain=pdf&date_stamp=2024-11-22https://doi.org/10.1063/5.0235467https://pubs.aip.org/aip/jap/article-pdf/doi/10.1063/5.0235467/20264698/205702_1_5.0235467.am.pdfhttps://pubs.aip.org/aip/aml/article/1/3/036110/2910717/KoopmanLab-Machine-learning-for-solving-complexhttps://pubs.aip.org/aip/aml/article/1/3/036111/2910912/Experimental-realization-of-a-quantumhttps://e-11492.adzerk.net/r?e=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&s=7AfoYSh3C4XR0EDuBl0qcVCXwwsOxygen incorporation effects on the structural andthermoelectric properties of copper(I) iodideCite as: J. Appl. Phys. 136, 205702 (2024); doi: 10.1063/5.0235467View Online Export Citation CrossMarkSubmitted: 27 August 2024 · Accepted: 3 November 2024 ·Published Online: 22 November 2024Martin Markwitz,1,2,3,a) Niall Malone,2,3 Song Yi Back,4 Alexander Gobbi,2 Jake Hardy,1,2,3Peter P. Murmu,2 Takao Mori,4,5 Ben J. Ruck,1,3 and John V. Kennedy2,3AFFILIATIONS1School of Chemical and Physical Sciences, Victoria University of Wellington, PO Box 600, Wellington 6140, New Zealand2National Isotope Centre, GNS Science, PO Box 30368, Lower Hutt 5010, New Zealand3The MacDiarmid Institute for Advanced Materials and Nanotechnology, Victoria University of Wellington, PO Box 600,Wellington 6140, New Zealand4International Center for Materials Nanoarchitectonics (WPI-MANA), National Institute for Materials Science (NIMS), 1-1 Namiki,Tsukuba, Ibaraki 305-0044, Japan5Graduate School of Pure and Applied Science, University of Tsukuba, 1-1-1 Tennodai, Tsukuba, Ibaraki 305-8671, Japana)Author to whom correspondence should be addressed: martin.markwitz@vuw.ac.nzABSTRACTOxygen is a ubiquitous contaminant in thin films grown in high vacuum systems, and it was hypothesized to play an important role in theproperties of the p-type conductivity of transparent copper(I) iodide, CuI. We study the ambient properties of CuI deposited at variouspartial pressures of oxygen gas. Through a variety of experimental techniques, we find that achieving a critical oxygen partial pressure ofbelow p(O2) = 3� 10�5 mbar is essential for depositing stoichiometric and conductive CuI thin films. Notably, we relate the commonlyreported copper excess to the presence of oxygen within the CuI films. Notably, we relate the commonly reported excess of copper in CuIthin films to to the presence of oxygen. Finally, we infer from transport and optical measurements that the hole transporting properties ofsputtered CuI films are dominated by an abundance of VCu defects with an acceptor transition energy of 84+ 3 meV rather than OI defectswith an acceptor transition energy of 175+ 14 meV.© 2024 Author(s). All article content, except where otherwise noted, is licensed under a Creative Commons Attribution (CC BY) license(https://creativecommons.org/licenses/by/4.0/). https://doi.org/10.1063/5.0235467I. INTRODUCTIONTransparent conducting materials offer the potential for thedevelopment of invisible electronics for everyday use. Currently,inorganic n-type transparent conducting oxides dominate sucha commercial market with the leading In2O3:Sn possessing anoptical gap greater than 3 eV and an electrical conductivity nearσ � 104 S cm�1, both of which are crucial parameters.1,2 On theother hand, inorganic p-type transparent conductors are limited tothe order of σ � 102 S cm�1. Of those compounds, CuI possessesthe highest conductivity and optical transparency, derived from alarge bandgap of EG ¼ 3:1 eV, high thin film carrier mobility of upto μ � 20 cm2 V�1 s�1, and high dopability with copper vacancies.3The high electrical performance of the ambient room temperaturezincblende phase γ-CuI is derived from the diffuse hybridized Cu3d and I 5p valence band, with two degenerated bands with low(ml ¼ 0:3m0), and high (mh ¼ 2:4m0) effective mass hole bandswhere m0 is the free electron mass.4,5CuI leads its competitor p-type transparent conductors due tothe localization of the O 2p valence band in delafossites and strongpolaron scattering of the carriers, leading to significantly lowercarrier mobilities in such materials.1,6,7 CuI is p-type doped withcopper vacancies (VCu) acting as shallow acceptor states with athermodynamic activation energy between 78 and 150 meV.8–10The unintentional doping with VCu can push CuI to the point ofdegenerate conductivity with carrier concentrations beyond1020 cm�3 while simultaneously maintaining its high mobility.11–13Due to the dual-band nature with a heavy carrier band, CuI pos-sesses a large Seebeck coefficient (α � 200 μVK�1),14,15 imbuingJournal ofApplied PhysicsARTICLE pubs.aip.org/aip/japJ. Appl. Phys. 136, 205702 (2024); doi: 10.1063/5.0235467 136, 205702-1© Author(s) 2024 22 November 2024 12:26:18https://doi.org/10.1063/5.0235467https://doi.org/10.1063/5.0235467https://pubs.aip.org/action/showCitFormats?type=show&doi=10.1063/5.0235467http://crossmark.crossref.org/dialog/?doi=10.1063/5.0235467&domain=pdf&date_stamp=2024-11-22https://orcid.org/0009-0007-6516-3571https://orcid.org/0009-0004-0485-8271https://orcid.org/0009-0000-8890-1484https://orcid.org/0009-0002-4090-5093https://orcid.org/0000-0001-5006-1918https://orcid.org/0000-0002-0109-1798https://orcid.org/0000-0003-2682-1846https://orcid.org/0000-0002-3719-7375https://orcid.org/0000-0002-9126-4997mailto:martin.markwitz@vuw.ac.nzhttps://creativecommons.org/licenses/by/4.0/https://creativecommons.org/licenses/by/4.0/https://doi.org/10.1063/5.0235467https://pubs.aip.org/aip/japit with a thermoelectric figure of merit of ZT ¼ 0:21.16–18 Highthermoelectric power conversion efficiencies rely on a large ther-moelectric figure of merit (ZT ¼ α2σT=κ), wherein κ ¼ κe þ κ p isthe thermal conductivity separated by the carrier and phonon con-tributions and T is the temperature. Notably, CuI has a low latticethermal conductivity of κ p � 0:5Wm�1 K�1 due to the large massof I.16 Its earth-abundant and non-toxic constituents with facilethin film and powder synthesis methods makes CuI an attractivecandidate for potential low-cost device integration for thermoelec-tric and other applications.19The comparatively limited conductivity of CuI is partially dueto its only recent rediscovery as a high performance optoelectronicmaterial.4,6 Theoretical research found that the lighter chalcogen-ides O, S, and Se could be promising p-type dopants when substi-tuted onto the iodine site.11 The computationally predicted mostpromising extrinsic dopant was S due to its shallow thermody-namic transition energy, ranging from 0:215 to 0:33 eV.11–13 Theother chalcogens (O, Se, and Te) have low solubility limits due tothe greater formation energy of the introduced acceptor state, andtheir incorporation is inhibited by the formation of secondaryCu-chalcogen phases. Experimentally, there have been a plethora ofrecent advancements in the carrier concentration (and thereby con-ductivity) of CuI in recent years by (a) enhancing the intrinsicproperties of CuI by deposition and exposure to I-rich condi-tions,6,20 (b) extrinsic shallow acceptor introduction with S or Sedoping,7,21–23 and (c) extrinsic doping to enhance the intrinsicproperties with Cs, Al, Fe, or Tb.24–28 These improvements havebrought CuI closer to the 103 S cm�1 threshold for potential com-mercial transparent conductor applications, such as in thin filmtransistors,29 thermoelectric generators,18 perovskite-based solarcells,30 and diodes.31 The high electrical conductivity and transpar-ency of p-type CuI is also important for high power conversionefficiency in an inverted perovskite solar cell architecture.32,33Similarly to S and Se, O has also been clearly identified as ap-type dopant in CuI,34–37 but its role in the measured electrical prop-erties is not yet clear. Challenging the conventional view that it is thecopper vacancies, which promote high film carrier concentrations,research by Storm et al.34 suggests that oxygen acceptor defects are thereason for high conductivity in CuI. In addition, it is also understoodthat CuI undergoes a transient electrical transport property changesthrough atmospheric exposure.34,38 To achieve high conductivity andtransparency, sputtering or pulsed laser deposition is known to beuseful to prepare CuI with high carrier mobility (>10 cm2V�1 s�1)and low roughness (<5 nm).6,7 However, in such vacuum depositionprocesses, there is a poor understanding of how residual O2 within thechamber during CuI film deposition (potentially provided by animpure carrier gas) contributes to the measured properties of thematerial compared to Cu2O or In2O3:Sn.39,40Focusing on the thermoelectric application where CuI is theleading p-type candidate, there is to our best knowledge no system-atic report for the role of O doping on the thermoelectric proper-ties of CuI. In this paper, we prepared CuI films through reactiveion beam sputtering with various partial O2 gas pressures. We con-ducted optical, compositional, structural, and transport measure-ments to characterize the effect of the oxygen partial pressure[p(O2)] on the properties of the films, finding a notable changeonce p(O2) during deposition was raised to near 1� 10�4 mbar.We identified the cause for the common [Cu]/[I] excess measuredin composition measurements to be due to O incorporation.II. METHODSThe Si(001), 500 nm thermally oxidized Si(001), microscopeslides, and vitreous carbon substrates were cleaned by sonication inacetone, ethanol, and de-ionized water for 10 min each prior to CuIdeposition. The reactive ion beam sputtering was conducted underhigh vacuum. The base pressure of the sputter system was�3� 10�7 mbar during continuous pumping with a turbopump.41A liquid nitrogen cold trap was used to further reduce the residualgas pressure prior to sputtering. For sputtering, Ar gas was fed intothe Penning ion source region, some of which entered the depositionchamber region while the gate valve separating the two regions wasopen. This results in a necessary operating argon partial pressure ofp(Ar) = 8� 10�6 mbar in the deposition chamber region. For con-trolled oxygen incorporation, 99:99 % O2 gas was leaked directly intothe deposition chamber through a needle valve with a closed gatevalve between the Penning ion source and deposition chamberregions, providing a measurement and control of p(O2). Then, thegate valve separating the Penning ion source and depositionchamber was opened, and sputtering was started. Since the Penningion source region volume was much lower than that of the deposi-tion chamber the p(O2) variation before and after opening the gatevalve was negligible. The chamber pressure was kept constant duringdeposition by constantly flooding the chamber with oxygen toreplenish the oxygen gas extracted through the turbopumps. Thiscontinuous oxygen input provided precise p(O2) control between3� 10�6 and 1� 10�4 mbar. Sputtering was conducted with an Ar+ion beam impinging on a commercial CuI sputter target held at 45�,and substrate holders held at 60� relative to the baseplate. An accel-erating voltage of 16 kV with a sputter target current of 0:15 mA wasused to deposit the CuI thin films on the aforementioned substrates.The film deposition rate depended on the deposition pressure butwas ranged between 0:5 and 1 nmmin�1.Rutherford backscattering spectrometry (RBS) measurementswere conducted with a 2:0MeV 4He+ beam with a beam currentbelow 10 nA by using a surface barrier detector mounted at a back-scattering angle of 165�.42 A 100 nm thin film of Gold deposited ona Si(001) substrate and Si(001) with a 500 nm oxidized layer wereused for the energy calibration. To investigate the crystallinity of thefilms, angle-symmetric x-ray diffraction (XRD) was conducted in theθ=2θ geometry with a Rigaku SmartLabs diffractometer using acopper x-ray source with a principal x-ray wavelength ofλ ¼ 1:54059 Å. A Ni filter was used to remove Cu K β and W Lα-derived diffraction peaks from the diffraction patterns.Transmittance spectrometry was conducted by using a Perkin ElmerLambda 365 spectrophotometer over a wavelength range between300 and 800 nm. Steady-state photoluminescence (PL) measure-ments were performed using a 375 nm Thorlabs L375P70MLD70mW light source with a Kymera 328i Spectrograph. Room tem-perature Hall effect and four-terminal resistance measurements wereconducted with a 10 μA driving current by using the HMS-3000Hall effect measurement system. A 0:55T permanent magnet wasused for the Hall effect measurements. Gold contacts with approxi-mate thicknesses of 100 nm were deposited by ion beam sputteringJournal ofApplied PhysicsARTICLE pubs.aip.org/aip/japJ. Appl. Phys. 136, 205702 (2024); doi: 10.1063/5.0235467 136, 205702-2© Author(s) 2024 22 November 2024 12:26:18https://pubs.aip.org/aip/japon the corners of the CuI samples through a mask. The Hall effectconductivity and Hall coefficient uncertainties are 5 % of the mea-sured values. Near-room temperature Seebeck effect measurementswere conducted in an ULVAC ZEM-3 Seebeck effect measurementsystem. The Seebeck coefficient uncertainty is 6 % of the measuredvalue.III. RESULTS AND DISCUSSIONRBS measurements on vitreous carbon substrates were con-ducted to investigate the degree of oxygen incorporation when CuIis deposited at a variety of oxygen partial pressures (Fig. 1). Fromthe measured data the film compositions were derived from thepeak areas and their respective Rutherford scattering cross sec-tions,43 the results of which are summarized in Table I. The filmsdeposited at and below p(O2) = 7� 10�5 mbar showed no presenceof oxygen and a near-unitary ratio of [Cu]/[I] [Figs. 1(a)–1(c)]while the film deposited at p(O2) = 1� 10�4 mbar [Fig. 1(d)]revealed a [Cu]/[I] concentration ratio of 1:14+ 0:01, and anoxygen concentration of 9:0+ 2:2 %, suggesting a relationshipbetween the incorporated concentration of O and an increase inthe [Cu]/[I] fraction.The RBS data of CuI films deposited on Si substrates areshown in Fig. 2(a). Notably, due to the Si, the in-film O backscat-tering peak is obfuscated for all films. Figure 2(b) shows the[Cu]/[I] ratios for films deposited on Si and C substrates depend-ing on the p(O2). This suggests that there is a threshold pressurebetween p(O2) = 7� 10�5 and p(O2) = 1� 10�4 mbar abovewhich the deposition leads to the incorporation of O in the CuIfilms. This incorporated oxygen in the films will likely substituteiodine in CuI or lead to the formation of a secondary oxide phase,both of which are compatible with an increase in the [Cu]/[I]ratio.XRD measurements were then conducted after the RBS mea-surements to further investigate the O incorporation effects on thecrystallinity and microstrain of the deposited films. The XRD pat-terns of CuI samples deposited on Si(001) substrates are depicted inFig. 3(a) for the angle-symmetric θ=2θ measurements. The diffracto-grams show the presence of strongly textured CuI in the h111i orien-tation.14 The CuI(111), CuI(222), and CuI(333) peaks wereindependently used to derive the out-of-plane lattice constant (a?)with Bragg’s law λ ¼ 2dhk‘ sin θ, wherein dhk‘ ¼ a?=ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffih2 þ k2 þ ‘2p.The derived out-of-plane lattice constant varied slightly based on thediffraction peak order, suggesting imperfect sample alignment withinthe goniometer. This alignment error was corrected by use of theFIG. 1. Rutherford backscattering spectra of films deposited on vitreous carbonat partial pressures of (a) 3� 10�6, (b) 3� 10�5, (c) 7� 10�5, and (d)1� 10�4 mbar. The inset graphs show the absence or presence of oxygen inthe respective films.TABLE I. Copper, iodine, and oxygen concentrations and the copper to iodine ratioof CuI films derived from RBS measurements deposited on vitreous carbon sub-strates deposited at various oxygen partial pressures.p(O2) (mbar) Cu (at. %) I (at. %) O (at. %) [Cu]/[I]3 × 10−6 49.7 ± 0.4 49.2 ± 0.2 1.1 ± 2.0 1.01 ± 0.013 × 10−5 49.7 ± 0.4 49.5 ± 0.2 0.7 ± 1.9 1.00 ± 0.017 × 10−5 49.8 ± 0.4 49.1 ± 0.2 1.1 ± 2.2 1.01 ± 0.011 × 10−4 48.5 ± 0.4 42.5 ± 0.2 9.0 ± 2.2 1.14 ± 0.01Journal ofApplied PhysicsARTICLE pubs.aip.org/aip/japJ. Appl. Phys. 136, 205702 (2024); doi: 10.1063/5.0235467 136, 205702-3© Author(s) 2024 22 November 2024 12:26:18https://pubs.aip.org/aip/japgoniometer error function G(θ) ¼ 12 cot2 (θ)[1þ sin (θ)]3 to derivethe true out-of-plane lattice constants, the p(O2)-dependence thereofdepicted in Fig. 3(b). As the partial oxygen pressure increased, theout-of-plane CuI lattice constant monotonically reduced from6:068+ 0:001 Å for the films deposited at p(O2)� 3� 10�7 mbar to6:055+ 0:001 Å for the films deposited at p(O2) = 1� 10�4 mbar. Itis known that room temperature vacuum-deposited CuI filmspossess larger out-of-plane lattice constants than the bulk powdervalue of 6:054 Å, with in-plane values that are smaller suggestingthat the films are subjected to biaxial tensile strain.3,4,44 The reducingout-of-plane lattice constant suggests some strain relaxation effectsdue to a differing chemical composition of the film.In order to deconvolute macrostrain, microstrain, and broad-ening effects of the out-of-plane peak series, the Williamson–Hallplot approach was employed. The Williamson–Hall plot approachdraws a linear fit β cos (θ) ¼ 4ε sin (θ)þ Kλ=D through the peakpositions and peak widths wherein β is the width (in radians), ε isthe strain, K is a numerical constant, and D is the crystallite size.45This method was conducted after subtracting a constant systematicbroadening approximated by the broadening of the Si(004) sub-strate peak width.46 Based on this approach, the out-of-plane crys-tallite sizes did not significantly vary as the oxygen partial pressurevaried. On the other hand, the out-of-plane strain (ε?) at firstdecreases from 0:94� 10�3 + 0:17� 10�3 for the film depositedat p(O2)� 3� 10�7 mbar to 0:72� 10�3 + 0:07� 10�3 for thefilm deposited at p(O2) = 3� 10�5 mbar. Subsequently ε? increasedto 1:33� 10�3 + 0:09� 10�3 for the film deposited at p(O2)= 1� 10�4 mbar. The derived out-of-plane microstrain relationshipwith p(O2) is shown in Fig. 3(c). The ε? becomes relaxed as thegrowth pressure increases toward p(O2) = 3� 10�5 mbar, whileabove it, ε? sharply increases again. The cause for this onset ofstrain is not yet understood, but due to the low calculated solubil-ity of chalcogenides in CuI,8,12,14,21 it is possible that a secondaryoxide phase becomes incorporated in the films.36,37 In summary,it is reasonable to interpret that the variation of the structuralproperties are related to variations in the chemical composition ofthe films, which itself can be controlled by variation of the p(O2)during deposition.To investigate variations in the transport properties of thesamples, they were subjected to four-point van der Pauw electricalconductivity measurements, the results of which are summarized inTable II. The films deposited at p(O2)� 3� 10�7 mbar exhibitedan electrical conductivity of 65:7+ 2:4 S cm�1, which did notchange significantly over a wide range of oxygen partial pressures.Notably, the films deposited at p(O2) = 1� 10�4 mbar had areduced electrical conductivity of 39:9+ 1:8 S cm�1. The sampleswere subjected to room temperature Hall and Seebeck effect mea-surements, which identified p-type carrier conduction in all films.The carrier concentration was derived from the Hall coefficientthrough pH ¼ qRHð Þ�1, and the carrier mobility was derived fromμ ¼ σRH . The Hall factor was assumed to be unitary due to thehigh carrier concentrations in these films. Hall carrier mobilitiesof 8:7+ 0:8 cm2 V�1 s�1 were found for the samples deposited atp(O2) � 3� 10�7 mbar, which increased to 9:4+ 0:8 cm2 V�1 s�1for the samples deposited at p(O2) = 7� 10�5 mbar, and reducedto 6:5+ 1:8 cm2 V�1 s�1 for the samples deposited atp(O2) = 1� 10�4 mbar. The Hall carrier concentration derivedfrom the Hall coefficient remained approximately constant near4:2 �1019 cm�3 for all films (comparable to previously measuredvalues14,15). The series of films produced similar Seebeck coeffi-cients (�200 μVK�1), except for the film deposited atp(O2)� 3� 10�7 mbar, which provided a greater Seebeck coeffi-cient of 235+ 14 μVK �1.The increased mobility at moderate oxygen partial gas pressuresis similar to the results by Storm et al.,3 who instead used nitrogenas a background gas during deposition of CuI with pulsed laser dep-osition. Depositing CuI at moderate gas pressures results in animproved crystal quality, in this case identified by a reduction in themicrostrain, therefore improving the carrier mobility by a reductionof the concentration of scattering sites. The cause for the reducedcarrier mobility for the film deposited at the highest oxygen partialpressure was presumably due to scattering from the increased con-centration of additional oxygen defects and oxide impurities. Finally,the power factor (α2σ) derived for the films is greatest for thosedeposited at p(O2) � 3� 10�7 mbar, finding a power factor of363+ 45 μWm�1 K�2, reducing to 168+ 22 μWm�1 K�2 for thefilms deposited at p(O2) = 1� 10�4 mbar. This implies that for ther-moelectric applications CuI should be deposited at the lowest possi-ble p(O2), while it does not strongly affect the electrical properties.Overall, oxygen does not appear to have a strong effect on thecarrier concentration and Seebeck coefficient, implying that the OIacceptor energy is deeper within the bandgap than the VCu acceptorFIG. 2. (a) Rutherford backscattering spectra of films deposited on silicon at avariety of oxygen partial pressures [p(O2)]. (b) Copper-to-iodine ratio dependingon p(O2). The dashed line is used to guide the reader.Journal ofApplied PhysicsARTICLE pubs.aip.org/aip/japJ. Appl. Phys. 136, 205702 (2024); doi: 10.1063/5.0235467 136, 205702-4© Author(s) 2024 22 November 2024 12:26:18https://pubs.aip.org/aip/japenergy, and that there is a low concentration thereof. Further, thelow solubility limit of oxygen in CuI limits the concentration of OIacceptors,11 which could instead lead to forming oxide phases,which provide a limited contribution to the conductivity in thefilms.37The samples’ optical properties were measured with transmit-tance spectrometry and PL. The transmittance spectra are depictedin Fig. 4(a), which show that the samples deposited atp(O2)� 3� 10�5 mbar) have high transmittance (.70%) through-out the highlighted visible regime. These samples also clearlyexhibit the characteristic CuI Z1/Z2 (3:10+ 0:01 eV) and the spinorbit split Z3 (3:73+ 0:01 eV) excitonic absorption resonances.3The Z1/Z2 absorption is at �0:05 eV higher than the standardvalue of 3:05 eV,4 possibly due to the high carrier concentrations inthese films which would cause a blueshift in the excitonic absorp-tion energy.25,47 The samples deposited at p(O2) > 3� 10�5 mbarexhibit slightly lower average transmittances with values of65%–67%, and appear to have suppressed excitonic absorptionresonances.The generalized Elliot formula was used in order to investigatethe absorption properties (single particle gap energy EG, spectralbroadening Γ, and exciton binding energy EB,i) of the depositedfilms, a method recently finding use for analyzing absorption dataof perovskite compounds.48,49 This type of absorption model con-siders two types of absorption: excitonic absorption peaks, and thecontinuum band-to-band transition absorption contribution, whichalso include band nonparabolicity and Sommerfeld enhancementeffects, written asFIG. 3. (a) Labeled θ=2θ x-ray diffraction patterns of CuI films deposited on thermally oxidized silicon substrates at various oxygen partial pressures. Calculated (b)out-of-plane lattice constant a? and (c) out-of-plane strain ε? from the diffraction patterns. The dashed lines are used to guide the reader.TABLE II. Room temperature electrical and thermoelectric properties of CuI thin films deposited at different oxygen partial pressures.p(O2) (mbar) σ (S cm−1) μH (cm2 V−1s−1) pH (×1019 cm−3) α (μVK−1) α2σ (μWm−1 K−2)≤3 × 10−7 65.7 ± 2.4 8.7 ± 0.8 4.7 ± 0.3 235 ± 14 363 ± 453 × 10−6 61.3 ± 2.3 9.2 ± 0.8 4.2 ± 0.2 200 ± 12 245 ± 313 × 10−5 57.9 ± 2.1 9.4 ± 0.8 3.9 ± 0.2 198 ± 12 227 ± 287 × 10−5 64.2 ± 2.4 9.3 ± 0.9 4.3 ± 0.3 207 ± 12 275 ± 341 × 10−4 39.9 ± 1.8 6.5 ± 1.8 3.9 ± 0.9 205 ± 12 168 ± 22Journal ofApplied PhysicsARTICLE pubs.aip.org/aip/japJ. Appl. Phys. 136, 205702 (2024); doi: 10.1063/5.0235467 136, 205702-5© Author(s) 2024 22 November 2024 12:26:18https://pubs.aip.org/aip/japαT �hωð Þ/ 1�hωXiXn2EB,in3sech�hω� EB,n,iΓ� �þð1EGsech�hω� EΓ� �11� exp �2πffiffiffiffiffiffiffiffiffiEB,iE�EGq� � 11� 8μ2i b�h4E � EGð ÞdE264375 (1)for a two-valence band system i ¼ l, h considering contributionfrom both light and heavy holes, n is the exciton order,EB,n,i ¼ EG � EB,in�2 is the nth order exciton binding energyanalogous to the electron energy levels of the hydrogen atom,μi ¼ mime mi þmeð Þ�1 is the reduced mass for each absorptionchannel, and b is band nonparabolicity defined throughECB(~k)� EV ,i(~k) ¼ EG þ �h2 ~k��� ���2 2μið Þ�1�b~k��� ���4. To fit Eq. (1) to themeasured data, the absorption coefficient (αT) was derived throughαT ¼ z�1 ln 1=T 0ð Þ, where T 0 is the sample transmittance. Thesample reflectance was assumed much less than one (R0 � 1). Anelectron effective mass me ¼ 0:23m0, a single particle bandgapEG ¼ 3:145 eV, and a band nonparabolicity factor b ¼ 0:55 eVm4were found to provide good absorption profiles that fitted theexperimental results. This allowed the spectral broadening Γ to befitted as the only free variable, a descriptor for instrumental andsample-derived broadening factors. The derived spectral broadeningfactors Γ are depicted in Fig. 4(b) while the Elliot fits themselves areshown in Figs. 4(c)–4(g). The variation in the broadening factorΓ appears to be coupled to the evolution of the microstrain, suggest-ing that there is a relationship between the two. The exciton bind-ing energies were calculated with EB,i ¼ μiq4 2 4πεð Þ2�h2�  �1whereε ¼ εrε0 using εr ¼ 6:5.4 The choice of electron effective massresulted in binding energies of EB,l ¼ 42 meV and EB,h ¼ 68 meVfor the light and heavy hole-derived excitons, respectively. Due tothe greater hole effective mass and, therefore, exciton binding energyand absorption intensity, the excitonic absorption derived from theheavy hole band dominates the excitonic absorption data. Thederived exciton binding energies are similar to those reported byNikitine50 and Inagaki et al.44 of �62 meV. The weighted averageexciton binding energy is 54meV, and the lower binding energy islikely a result of the electron effective mass and the higher freecarrier concentration as expected by the model of excitons byMahan.47 This type of analysis using the Elliot fit is recommendedFIG. 4. (a) Transmittance spectra with highlighted visible region and annotated excitonic absorption energies. The room temperature thermal energy kBT is highlightedwith the dashed line. (b) Spectral broadening factor derived from Elliot fits (red) comprised of excitonic (blue) and continuum (green) components, which are shownin (c)–(g).Journal ofApplied PhysicsARTICLE pubs.aip.org/aip/japJ. Appl. Phys. 136, 205702 (2024); doi: 10.1063/5.0235467 136, 205702-6© Author(s) 2024 22 November 2024 12:26:18https://pubs.aip.org/aip/japanytime the presence of excitonic states influences the absorptionspectra of a compound49,51 and is used in this work for determiningthe bandgap of CuI rather than the commonly used Tauc plot fittingprocedure. The Tauc plot procedure uses the low-energy shoulder ofthe excitonic absorption spectrum rather than the basic assumptionof absorption due to a parabolic band absorption edge, which cate-gorically underestimates the bandgap of CuI as 3:0 eV compared tothe more-realistic value of EG ¼ 3:1 eV.Steady-state PL spectra were conducted to investigate the evo-lution of shallow acceptor states in CuI as a function of depositionpressure which are shown in Fig. 5. Three gaussian functions wereused to resemble PL emissions derived from excitonic recombina-tion (red), VCu defects (green), and OI defects (blue). Beyond thep(O2) = 3� 10�5 mbar threshold pressure, the PL spectra of theCuI films showed an increased high-wavelength shoulder ofthe PL data, attributed the incorporation of OI defects. Notably, thePL intensity of the OI defect relative to the other PL peaks scaledwith the deposition pressure. The emission energies (and peak fullwidth at half maxima) were restricted to 2:990+ 0:001 (0:045+0:001), 2:958+ 0:002 (0:077+ 0:003), and 2:869+ 0:013 eV(0:114+ 0:006 eV), for the excitonic,52 VCu,53 and OI-derivedemission lines, respectively. If the energy difference between theexciton emission energy from PL to the bandgap is assumed to beidentical to the (average) exciton binding energy, the VCu defectstates lay 84+ 3 meV above the valence band maximum, whilethe OI defect states lay 175+ 14 meV above the valence bandmaximum. Storm et al.,8 Bar et al.,9 and Koyasu et al.10 derived aVCu acceptor energy of 78, 95, and 150 meV, with which our resultsare in good agreement. Also, density functional theory using thePerdew–Burke–Ernzerhof exchange correlation functional with aHubbard parameter applied to the Cu 3d states calculations foundthat the OI0 to OI1- acceptor transition is �100 meV deeper than theVCu0 to VCu1- acceptor transition.11 A schematic of the CuI valenceband structure and the acceptor energy levels is visualized in Fig. 6.In our films, the limited effectiveness of the OI acceptor is furthermasked by the abundance of VCu acceptors in the present films, incontrast with the films grown by Storm et al.,34 which possessoverall lower carrier concentrations (�1017 cm�3).IV. CONCLUSIONThe aim of this work was to investigate whether small con-centrations of O have notable effects on the optical and trans-port properties of CuI thin films. This research clarifies how theO partial pressure plays a role in the measured electrical andoptical properties of CuI films, and finds that the properties ofCuI are insensitive to the O partial pressure at or below a thresh-old pressure of p(O2) = 3� 10�5 mbar. The electrical, optical,and structural properties notably vary above this threshold pres-sure, where a O is incorporated in the films, as measured withRBS. The highest mobility of the CuI films is observed whendeposited at a moderate partial pressure, which suggests a poten-tial direction for achieving heavily doped and high mobility CuIfilm deposition.The controlled gas partial pressure environment we used toincorporate O in our films is not compositionally equivalent to airas it is missing other gaseous constituents such as nitrogen. ForCuI, however, O is expected to be the active ingredient in air,which can be incorporated during film deposition. This is impor-tant information for reducing the time needed to pump downvacuum systems from atmosphere before CuI film deposition and,therefore, reducing the cost of such a process. The VCu defect isfound to be a shallower acceptor in CuI than OI by �100 meV, sug-gesting that O is not an effective p-type dopant in already heavilydoped CuI.ACKNOWLEDGMENTSThis research was funded by the Royal Society of NewZealand through the Marsden Fund, New Zealand (Grant No.FIG. 5. (a) Normalized PL spectra fitted with gaussian functions resemblingexcitonic (red), VCu (green), and OI (blue) PL features.FIG. 6. Near-valence band structure schematic with parabolic light (ml ) andheavy (mh) bands, and the positions of VCu and OI acceptors relative to thevalence band.Journal ofApplied PhysicsARTICLE pubs.aip.org/aip/japJ. Appl. Phys. 136, 205702 (2024); doi: 10.1063/5.0235467 136, 205702-7© Author(s) 2024 22 November 2024 12:26:18https://pubs.aip.org/aip/japMFP-GNS2301) and the JST Mirai Program, Japan (Grant No.JPMJMI19A1). The authors wish to acknowledge Mr. Chris Purcell(National Isotope Centre, GNS Science, New Zealand) for carryingout RBS measurements.AUTHOR DECLARATIONSConflict of InterestThe authors have no conflicts to disclose.Author ContributionsMartin Markwitz: Conceptualization (equal); Data curation (lead);Formal analysis (equal); Investigation (lead); Methodology (equal);Visualization (equal); Writing – original draft (lead); Writing –review & editing (equal). Niall Malone: Conceptualization (equal);Formal analysis (equal); Validation (equal); Visualization (equal);Writing – review & editing (equal). Song Yi Back: Data curation(equal); Writing – review & editing (equal). Alexander Gobbi:Data curation (equal); Investigation (equal); Writing – review &editing (equal). Jake Hardy: Data curation (equal); Writing –review & editing (equal). Peter P. Murmu: Funding acquisition(equal); Resources (equal); Supervision (equal); Validation (equal);Writing – review & editing (equal). Takao Mori: Resources(equal); Supervision (equal); Writing – review & editing (equal).Ben J. Ruck: Formal analysis (equal); Funding acquisition (equal);Supervision (equal); Validation (equal); Writing – review & editing(equal). John V. Kennedy: Funding acquisition (equal); Projectadministration (equal); Resources (equal); Supervision (equal);Validation (equal); Writing – review & editing (equal).DATA AVAILABILITYThe data that support the findings of this study are availablefrom the corresponding author upon reasonable request.REFERENCES1J. Willis and D. O. Scanlon, J. Mater. Chem. C 9, 11995 (2021).2M. Markwitz, S. Y. Back, E. X. M. Trewick, P. P. Murmu, T. Mori, B. J. Ruck,and J. V. Kennedy, Phys. Rev. B 109, 115201 (2024).3P. Storm, M. S. Bar, G. Benndorf, S. Selle, C. Yang, H. von Wenckstern,M. Grundmann, and M. Lorenz, APL Mater. 8, 091115 (2020).4M. Grundmann, F. L. Schein, M. Lorenz, T. Böntgen, J. Lenzner, and H. vonWenckstern, Physica Status Solidi A 210, 1671 (2013).5J. Wang, J. Li, and S. S. Li, J. Appl. Phys. 110, 054907 (2011).6C. Yang, M. Kneiß, M. Lorenz, and M. Grundmann, Proc. Natl. Acad. Sci.U.S.A. 113, 12929 (2016).7A. S. Mirza, M. Pols, W. Soltanpoor, S. Tao, G. Brocks, and M. Morales-Masis,Matter 6, 4306 (2023).8P. Storm, M. S. Bar, S. Selle, H. von Wenckstern, M. Grundmann, andM. Lorenz, Physica Status Solidi (RRL) 15, 2100214 (2021).9M. S. Bar, D. Splith, Y. Chen, M. Grundmann, H. von Wenckstern, T. Rauch,S. Blaurock, and H. Krautscheid, Phys. Rev. Appl. 22, 044027 (2024).10S. Koyasu, N. Umezawa, J. D. Baniecki, A. Yamaguchi, and M. Miyauchi,Cryst. Growth Des. 18, 6748 (2018).11D. Huang, Y. J. Zhao, S. Li, C. S. Li, J. J. Nie, X. H. Cai, and C. M. Yao,J. Phys. D: Appl. Phys. 45, 145102 (2012).12M. Graužinytė, S. Botti, M. A. L. Marques, S. Goedecker, andJ. A. Flores-Livas, Phys. Chem. Chem. Phys. 21(35), 18839–18849 (2019).13J. Willis, R. Claes, Q. Zhou, M. Giantomassi, G. Rignanese, G. Hautier, andD. O. Scanlon, Chem. Mater. 35, 8995 (2023).14M. Markwitz, P. P. Murmu, S. Y. Back, T. Mori, B. J. Ruck, and J. V. Kennedy,Surf. Interfaces 41, 103190 (2023).15M. Markwitz, P. P. Murmu, S. Y. Back, T. Mori, J. V. Kennedy, and B. J. Ruck,Mater. Today Phys. 46, 101513 (2024).16C. Yang, D. Souchay, M. Kneiß, M. Bogner, H. M. Wei, M. Lorenz,O. Oeckler, G. Benstetter, Y. Q. Fu, and M. Grundmann, Nat. Commun. 8,16076 (2017).17T. Zhu, Y. Liu, C. Fu, J. P. Heremans, J. G. Snyder, and X. Zhao, Adv. Mater.29, 1605884 (2017).18J. Coroa, B. M. Morais Faustino, A. Marques, C. Bianchi, T. Koskinen,T. Juntunen, I. Tittonen, and I. Ferreira, RSC Adv. 9, 35384 (2019).19A. Liu, H. Zhu, M. G. Kim, J. Kim, and Y. Y. Noh, Adv. Sci. 8, 2100546 (2021).20E. J. Bae, J. Kim, M. Han, and Y. H. Kang, ACS Mater. Lett. 5, 2009 (2023).21P. P. Murmu, J. Kennedy, Z. Liu, and T. Mori, J. Alloys Compd. 921, 166103(2022).22K. H. Ahn, G. H. Kim, S. J. Kim, J. H. Kim, G. S. Ryu, P. Lee, B. K. Ryu,J. Y. Cho, Y. H. Kim, J. H. Kang et al., Chem. Mater. 34, 10517 (2022).23P. P. Murmu, M. Markwitz, S. V. Chong, N. Malone, T. Mori, H. Vyas,L. J. Kennedy, S. Rubanov, C. I. Sathish, J. Yi, and J. V. Kennedy, Mater. TodayEnergy 44, 101639 (2024).24K. Matsuzaki, N. Tsunoda, Y. Kumagai, Y. Tang, K. Nomura, F. Oba, andH. Hosono, J. Am. Chem. Soc. 144, 16572 (2022).25A. S. Mirza, B. Vishal, P. Dally, C. S. Schnohr, S. De Wolf, andM. Morales-Masis, Adv. Funct. Mater. 34, 2316144 (2024).26N. Salah, A. Alshahrie, M. Aljaghtham, A. Saeed, N. Baghdadi, andK. Koumoto, Adv. Electron. Mater. 8, 2101214 (2022).27S. Byeon, K. Pryga, J. H. Kim, J. S. Rhyee, K. Komedera, B. Wiendlocha, andH. Jin, J. Alloys Compd. 1002, 175349 (2024).28N. Salah, A. M. Abusorrah, Y. N. Salah, M. Almasoudi, N. Baghdadi,A. Alshahri, and K. Koumoto, Ceram. Int. 46, 27244 (2020).29A. Liu, H. Zhu, W. T. Park, S. J. Kim, H. Kim, M. G. Kim, and Y. Y. Noh, Nat.Commun. 11, 4309 (2020).30H. Wang, Z. Yu, X. Jiang, J. Li, B. Cai, X. Yang, and L. Sun, Energy Technol.5, 1836 (2017).31C. Yang, M. Kneiß, F. L. Schein, M. Lorenz, and M. Grundmann, Sci. Rep. 6,21937 (2016).32S. Z. Haider, H. Anwar, S. Manzoor, A. G. Ismail, and M. Wang, Curr. Appl.Phys. 20, 1080 (2020).33K. Wang, Z. Xu, Z. Guo, H. Wang, S. M. H. Qaid, K. Yang, and Z. Zang, Adv.Energy Mater. 14, 2402249 (2024).34P. Storm, S. Gierth, S. Selle, M. S. Bar, H. von Wenckstern, M. Grundmann,and M. Lorenz, APL Mater. 9, 051101 (2021).35C. S. Herrick and A. D. Tevebaugh, J. Electrochem. Soc. 110, 119 (1963).36B. L. Zhu and X. Z. Zhao, Physica Status Solidi A 208, 91 (2011).37R. Xue, G. Gao, L. Yang, L. Xu, Y. Zhang, and J. Zhu, J. Mater. Chem. C 11,13681 (2023).38A. Crovetto, H. Hempel, M. Rusu, L. Choubrac, D. Kojda, K. Habicht, andT. Unold, ACS Appl. Mater. Interfaces 12, 48741 (2020).39S. Shibasaki, Y. Honishi, N. Nakagawa, M. Yamazaki, Y. Mizuno, Y. Nishida,K. Sugimoto, and K. Yamamoto, Appl. Phys. Lett. 119, 242102 (2021).40P. P. Murmu, A. Shettigar, S. V. Chong, Z. Liu, D. Goodacre, V. Jovic, T. Mori,K. E. Smith, and J. V. Kennedy, J. Materiom. 7, 612 (2021).41J. Kennedy, P. P. Murmu, J. Leveneur, A. Markwitz, and J. Futter, Appl. Surf.Sci. 367, 52 (2016).42J. V. Kennedy, A. Markwitz, H. J. Trodahl, B. J. Ruck, S. Durbin, and W. Gao,J. Electron. Mater. 36, 472 (2007).43J. W. Mayer and E. Rimini, Ion Beam Handbook for Material Analysis(Elsevier, 2012).44S. Inagaki, M. Nakamura, Y. Okamura, M. Ogino, Y. Takahashi, L. C. Peng,X. Z. Yu, Y. Tokura, and M. Kawasaki, Appl. Phys. Lett. 118, 012103 (2021).45G. F. Harrington and J. Santiso, J. Electroceram. 47, 141 (2021).Journal ofApplied PhysicsARTICLE pubs.aip.org/aip/japJ. Appl. Phys. 136, 205702 (2024); doi: 10.1063/5.0235467 136, 205702-8© Author(s) 2024 22 November 2024 12:26:18https://doi.org/10.1039/D1TC02547Chttps://doi.org/10.1103/PhysRevB.109.115201https://doi.org/10.1063/5.0021781https://doi.org/10.1002/pssa.201329349https://doi.org/10.1063/1.3633220https://doi.org/10.1073/pnas.1613643113https://doi.org/10.1073/pnas.1613643113https://doi.org/10.1016/j.matt.2023.10.003https://doi.org/10.1002/pssr.202100214https://doi.org/10.1103/PhysRevApplied.22.044027https://doi.org/10.1021/acs.cgd.8b01024https://doi.org/10.1088/0022-3727/45/14/145102https://doi.org/10.1039/c9cp02711dhttps://doi.org/10.1021/acs.chemmater.3c01628https://doi.org/10.1016/j.surfin.2023.103190https://doi.org/10.1016/j.mtphys.2024.101513https://doi.org/10.1038/ncomms16076https://doi.org/10.1002/adma.201605884https://doi.org/10.1039/C9RA07309Dhttps://doi.org/10.1002/advs.202100546https://doi.org/10.1021/acsmaterialslett.3c00315https://doi.org/10.1016/j.jallcom.2022.166103https://doi.org/10.1021/acs.chemmater.2c02603https://doi.org/10.1016/j.mtener.2024.101639https://doi.org/10.1016/j.mtener.2024.101639https://doi.org/10.1021/jacs.2c06283https://doi.org/10.1002/adfm.202316144https://doi.org/10.1002/aelm.202101214https://doi.org/10.1016/j.jallcom.2024.175349https://doi.org/10.1016/j.ceramint.2020.07.209https://doi.org/10.1038/s41467-020-18006-6https://doi.org/10.1038/s41467-020-18006-6https://doi.org/10.1002/ente.201700422https://doi.org/10.1038/srep21937https://doi.org/10.1016/j.cap.2020.06.022https://doi.org/10.1016/j.cap.2020.06.022https://doi.org/10.1002/aenm.202402249https://doi.org/10.1002/aenm.202402249https://doi.org/10.1063/5.0047723https://doi.org/10.1149/1.2425687https://doi.org/10.1002/pssa.201026239https://doi.org/10.1039/D3TC02422Ahttps://doi.org/10.1021/acsami.0c11040https://doi.org/10.1063/5.0072310https://doi.org/10.1016/j.jmat.2020.10.015https://doi.org/10.1016/j.apsusc.2016.01.160https://doi.org/10.1016/j.apsusc.2016.01.160https://doi.org/10.1007/s11664-006-0051-yhttps://doi.org/10.1063/5.0036862https://doi.org/10.1007/s10832-021-00263-6https://pubs.aip.org/aip/jap46A. K. Zak, W. H. A. Majid, M. E. Abrishami, and R. Yousefi, Solid State Sci.13(1), 251–256 (2011).47G. D. Mahan, Phys. Rev. 153, 882 (1967).48F. Ruf, M. F. Aygüler, N. Giesbrecht, B. Rendenbach, A. Magin, P. Docampo,H. Kalt, and M. Hetterich, APL Mater. 7, 031113 (2019).49T. Palmieri, E. Baldini, A. Steinhoff, A. Akrap, M. Kollár, E. Horváth, L. Forró,F. Jahnke, and M. Chergui, Nat. Commun. 11, 850 (2020).50S. Nikitine, Philos. Mag. 4, 1 (1959).51C. Malerba, F. Biccari, C. L. A. Ricardo, M. D’Incau, P. Scardi, and A. Mittiga,Solar Energy Mater. Solar Cells 95, 2848 (2011).52K. X. Zhang, S. W. Wang, L. Y. Bai, Y. Wang, K. Ou, Y. W. Zhang, andL. X. Yi, J. Lumin. 214, 116522 (2019).53S. Koyasu, N. Umezawa, A. Yamaguchi, and M. Miyauchi, J. Appl. Phys. 125,115101 (2019).Journal ofApplied PhysicsARTICLE pubs.aip.org/aip/japJ. Appl. Phys. 136, 205702 (2024); doi: 10.1063/5.0235467 136, 205702-9© Author(s) 2024 22 November 2024 12:26:18https://doi.org/10.1016/j.solidstatesciences.2010.11.024https://doi.org/10.1103/PhysRev.153.882https://doi.org/10.1063/1.5083792https://doi.org/10.1038/s41467-020-14683-5https://doi.org/10.1080/14786435908238225https://doi.org/10.1016/j.solmat.2011.05.047https://doi.org/10.1016/j.jlumin.2019.116522https://doi.org/10.1063/1.5082865https://pubs.aip.org/aip/jap