SnO2-Catalyzed Oxidation in High-Efficiency CdTe Solar Cells - ACS

Mar 8, 2019 - National Renewable Energy Laboratory , 15013 Denver West Parkway, ... Interfaces at the front of superstrate CdTe-based solar cells are ...
0 downloads 0 Views 1MB Size
Subscriber access provided by McMaster University Library

Surfaces, Interfaces, and Applications 2

SnO-catalyzed oxidation in high efficiency CdTe solar cells Craig L Perkins, Deborah McGott, Matthew O. Reese, and Wyatt K. Metzger ACS Appl. Mater. Interfaces, Just Accepted Manuscript • DOI: 10.1021/acsami.9b00835 • Publication Date (Web): 08 Mar 2019 Downloaded from http://pubs.acs.org on March 16, 2019

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

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

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

ACS Applied Materials & Interfaces

SnO2-catalyzed oxidation in high efficiency CdTe solar cells Craig L. Perkins*, Deborah L. McGott, Matthew O. Reese, and Wyatt K. Metzger National Renewable Energy Laboratory, 15013 Denver West Parkway, Golden CO 80401 USA

KEYWORDS. Cadmium telluride, tin oxide, photovoltaics, X-ray photoelectron spectroscopy, interfaces, passivation, tin oxide, solar cells

Abstract Interfaces at the front of superstrate CdTe-based solar cells are critical to carrier transport, recombination, and device performance, yet determination of the chemical structure of these nanoscale regions has remained elusive. This is partly due to changes that occur at front interfaces during high temperature growth and substantive changes occurring during post-deposition processing. In addition, these buried interfaces are extremely difficult to access in a way that preserves chemical information. In this work we use a recently developed thermomechanical cleaving technique paired with X-ray photoelectron spectroscopy to probe oxidation states at the SnO2 interface of CdTe solar cells. We show that the tin oxide front electrode promotes the formation of nanometer-scale oxides of tellurium and sulfur. Most oxidation occurs during CdCl2/O2 activation. Surprisingly we show that relatively low temperature anneals (180-260 °C)

ACS Paragon Plus Environment

1

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

Page 2 of 29

used to diffuse and activate copper acceptors in a doping/back contact process also cause significant changes in oxidation at the front of the cell, providing a heretofore missing aspect of how back contact processes can modify device transport, recombination, and performance. Device performance is shown to correlate with the extent of tellurium and sulfur oxidation within this nanometer scale region. Mechanisms responsible for these beneficial effects are proposed. 1. Introduction Recent advances in CdTe-based solar cells have put them on the vanguard of photovoltaic technology, competing even with single crystal silicon, a material that has benefited from decades more development time.1,2 A standard superstrate CdTe cell architecture is usually described simply as a stack of the following discrete layers: glass/SnO2:F/SnO2/CdS:O/CdTe/metal, but in fact numerous studies have shown that the processing steps used to create such a device result in an extremely complicated structure that continues to yield surprises in its structures and function.3– 5

Processing begins typically with the production of SnO2:F coated glass, then a high resistivity

transparent (HRT) layer, followed by growth of an emitter layer, typically 100-200 nm of CdS or CdS:O. The thickness and uniformity of the emitter layer is important: light absorbed by CdS at wavelengths shorter than ~520 nm does not contribute to current in the cells, thus providing a motivation to keep this layer as thin as possible. If the emitter layer is too thin however, interdiffusion of CdS and CdTe during high temperature processing can cause direct contact between the SnO2:F front electrode and the CdTe absorber, leading to regions acting as weak diodes or even shorts, with the result being low open circuit voltage.6 A HRT layer such as intrinsic tin oxide that blocks direct contact between the SnO2:F and CdTe layers can mitigate these types of defects. After growth of the CdTe absorber, typically performed in an oxygen-containing ambient, the film stack undergoes another high temperature process in the presence of a chlorine

ACS Paragon Plus Environment

2

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

ACS Applied Materials & Interfaces

compound such as CdCl2 or MgCl2. In the context of this paper, it is important to note that the chloride treatment is also typically performed in an oxygen-containing ambient. Oxygen addition to CdTe growth and CdCl2 treatment process gases was shown to reduce recombination near the CdS-CdTe junction, although the exact mechanism by which this occurs was not delineated.7 In addition to facilitating the growth of larger, less defective grains of the absorber, CdCl 2/O2 treatment can cause interdiffusion of CdS and CdTe, passivation of defects, and modifications in doping.8 Standard superstrate devices are finished with a back contact that supplies copper to act as a p-type dopant. Prior to metallization, copper applied initially as an electronically inactive compound or metal is diffused into the CdTe absorber and thermally converted to CuCd acceptors at temperatures ranging from 150 °C to 300 °C. In a previous publication we demonstrated that the CdCl2 treatment, despite being performed on the back surface of the cell, causes the formation of Cl-Cd-Cl nanosheets at the front of the cell, specifically at the SnO2-emitter interface.4 This example of a late-stage processing step causing changes to the front of the cell prompted us to extend our studies towards understanding the evolution of chalcogen chemical states in this region of CdTe solar cells. Although there are numerous studies8–13 on chemical transformations at free or exposed surfaces of CdTe and related materials, there currently are very few studies on chalcogen chemical states at the front of superstrate cells due to the difficulty of accessing and studying this nanoscale region. Exceptions include two earlier studies that used lift-off methods to expose interfaces near the SnO2 front electrode. Pookpanratana and coworkers used a purely mechanical liftoff process that appeared to cause cleavage between the CdTe and CdS layers.14 Albin et al. also used a purely mechanical process to separate and analyze CdTe interfaces, finding laterally non-uniform cleavages occurring both at the SnO2-CdS interface and at the CdS-CdTe interface.15 In this work however we make

ACS Paragon Plus Environment

3

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

Page 4 of 29

use of a previously described4,16,17 thermomechanical cleaving process that cleanly exposes only the transparent conducting oxide (TCO)-emitter interface, and unlike previous lift-off/analysis experiments, we demonstrate conditions that preserve the exposed surfaces.

This

thermomechanical cleaving sample preparation method is important because the halide and chalcogenide materials in this part of the cell are difficult to probe as they exist within a device: exposure to atmosphere, ion beams, and mechanical polishing usually causes loss of the chemical information that is initially present.

In our current work, thermomechanical cleaving in

conjunction with X-ray photoelectron spectroscopy (XPS) and electrical measurements on completed CdTe solar cells were used to reveal several interesting and previously unknown aspects about redox reactions that occur at the SnO2-emitter interface. Among these is the key finding that three different processing steps, use of tin oxide as the TCO, CdCl2 treatment, and back contact annealing, act together to maximize the extent of tellurium and sulfur oxidation in the region adjacent to the tin oxide front electrode. We demonstrate that relatively low temperature back contact thermal processing used to activate copper can cause chemical changes in the front of the cell. Finally, we show that device performance correlates with the extent of tellurium and sulfur oxidation, with higher performing devices having a greater proportion of oxidized species at the SnO2 interface. 2. Experimental CdTe devices were grown in the superstrate configuration on Schott AF45 glass. The glass was coated in a SnO2:F/SnO2 TCO bilayer via chemical vapor deposition. CdS and CdS:O emitters 100 nm thick were deposited using radio frequency magnetron sputtering in either pure argon or an oxygen/argon ambient.16 CdTe (4 µm) was deposited via close-spaced sublimation (CSS) with the substrate held at 600 ˚C in an oxygen/helium ambient. CSS was also used to perform a vapor-

ACS Paragon Plus Environment

4

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

ACS Applied Materials & Interfaces

phase CdCl2 anneal in an oxygen/helium ambient with the substrate held at 415 ˚C. Copper dopant was then introduced by soaking the device briefly at room temperature in a 0.1 mM CuCl 2 solution.18 This dopant was then activated by annealing in air at temperatures between 180 ˚C and 260 ˚C. For completed devices, a gold back contact (100 nm) was then thermally evaporated onto the device stack. Device stacks were exposed to air between each processing step. Device performance was determined using current-voltage (J-V) behavior measured under standard test conditions (1000 W m-2, 25 ˚C, AM1.5G). Completed and electrically characterized solar cells as well as test film stacks were cleaved using a liquid nitrogen-based thermomechanical technique described previously.4,17 Cleaving was performed in an argon-filled glove-box that is physically connected to a suite of surface analytical tools which allows examination of both sides of cleaved devices without air exposure. XPS measurements were made using monochromatic Al Kα radiation and a pass energy of 29 eV. The spectrometer binding energy scale was calibrated at high and low energy using clean gold and copper foils and known transition energies.19 Because of difficulty in getting good electrical contacts to the CdS-CdTe side of cleaved samples, charging was observed in some cases. For that reason, all spectra were aligned using the Te 3d5/2 peak of the Te2- chemical component, which was measured to be at 572.4 eV on conducting samples that did not exhibit charging. Conductivity was qualitatively determined by comparing core level spectra taken at high and low power on the X-ray source (350 W and 10 W respectively). Data analysis and peak fitting were performed using a combination of Igor and PHI MultiPak.20,21 In some Te 3d5/2 spectra, there appeared to be a minor third chemical component with a binding energy intermediate between the two main chemical components that are the subject of this paper, Te2- and Te4+. For the sake of simplicity

ACS Paragon Plus Environment

5

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

Page 6 of 29

Figure 1. Examples of several different LN2-cleaved devices showing typical mirror smooth surfaces. Sample (a) is a ~17 mm x 5 mm piece of the CdTe side of a cleaved device, with areas 1-4 corresponding to the XPS spectra in Figure 2. Sample (a), an early effort, utilizes epoxy and a soda glass “handle”. Sample (b) shows both sides of a cleaved device and utilizes a polymer for both the adhesive and “handle”. Sample (c) also shows both sides of a cleaved device and utilizes epoxy and an aluminum foil “handle”. Sample (d) shows intact cells. and consistency with other spectra, peak fitting did not include this third chemical state. Sputter depth profiling was performed using a 3 keV argon ion beam. 3. Results 3.1 Lateral uniformity of cleaved devices and method validation Figure 1 consists of photos of different superstrate devices that represent our baseline cell process using a glass/SnO2:F/SnO2/CdS:O/CdTe/gold film stack.

As demonstrated previously,

thermomechanical cleaving of these devices occurs within a couple nanometers of the tin oxideemitter interface, with the interface being defined by CdCl2 nanosheets that are present on both sides of the cleave location.4 Cleaved surfaces were generally mirror smooth to the eye. Figure 1

ACS Paragon Plus Environment

6

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

ACS Applied Materials & Interfaces

Figure 2. XPS Te 3d5/2 spectra (a) and S 2s spectra (b) taken from areas 1-4 on sample of Figure 1a, showing typical good reproducibility of the relative amounts of oxidized chalcogens. Inset tables show percentage of oxide peak relative to total peak area: Te4+/(Te4+ + Te2-) for tellurium and S6+/(S6+ + S2-) for sulfur. Table 1. XPS transitions and binding energies for different oxidation states/chemical environments transition

Oxidation state

Cd 3d5/2 Te 3d5/2 Te 3d5/2 S 2s S 2s O 1s O 1s

2+ 24+ 26+ 2-, O-Te 2-, O-S

Binding Energy (eV) 405.0 572.4 576.3 225.9 233.0 530.7 532.2

also portrays the evolution of our cleaving methodology which in earlier iterations used glass as the “handle” for the side that carries the epoxy and 4 µm thick CdS/CdTe layers (Figure 1a). Occasional factures occurring within the soda glass itself led to the use of foil handles (Figure 1c).

ACS Paragon Plus Environment

7

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

Page 8 of 29

A separate effort to develop the ability to lift off very large areas (up to 3 in x 3 in) has led to the development of a mathematical framework for the delamination process and the use of polymeric handles, an example of which is shown in Figure 1b.17 The numbered regions denoted by red circles in Figure 1a were examined by XPS, the results for which are shown in Figure 2. Figure 2 illustrates a critical aspect of this work and of the cleaving process in general: the relative abundance of oxidized chalcogens across any given cleaved device or separate cleaves of the same devices was remarkably constant. Figure 2 is comprised of high resolution XPS spectra of the Te 3d5/2 and S 2s regions for four different areas taken along the long axis of the sample in Figure 1a, the CdS:O/CdTe side of a cleaved device. In Figure 2a the more intense peaks centered on 572.4 eV stem from Te2- and the less intense peaks centered on 576.3 eV stem from oxidized tellurium, Te4+. The inset table indicates the percentage of oxidized tellurium, Te4+, from each spectrum. Panel (b) shows similar information for sulfur, with the low binding energy peak at 225.8 eV arising from S2- (sulfide) and a peak 7.2 eV higher at 233.0 eV. Based on a ~ 7 eV chemical shift previously observed in S 2p spectra, we assign the 233.0 eV 2s peak to S6+.22 Table 1 lists binding energies for the various transitions used in this work. As in our previous work,4 the 2s transition rather than the 2p was used to probe the oxidation state of sulfur because the 2p transition of S 6+ overlaps almost perfectly with the 4s transition of Te2-. This overlap is apparent in spectra obtained from a sputter depth profile of the CdS/CdTe side of a cleaved sample (Figure S1) in which the sulfide 2p doublet at 161.5 eV decreases in intensity as the peak at 169 eV increases in parallel with the Te 3d5/2 intensity. Use of the S 2s transition to determine sulfur speciation avoids the issue of overestimating the percentage of oxidized sulfur in materials that contain both sulfur and tellurium as may have happened in a previous publication.16

ACS Paragon Plus Environment

8

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

ACS Applied Materials & Interfaces

Although the thermomechanical cleaving process was conducted in an argon-filled glovebox that is connected to our XPS and AES analysis systems, an early concern in this work was whether or not the cleaving and handling process itself was causing oxidation of the newly exposed interfaces. Control experiments were conducted to test for this possibility. A CdTe sample was cleaned in the XPS system using argon ion beam sputtering and thermal annealing. Mimicking a real analysis, the sample was then transported in UHV to the glove-box portion of our cluster tool, immersed in liquid nitrogen, withdrawn into flowing dry argon, then transported back into the XPS system for analysis. The results from this set of measurements are shown in Figure S2 and demonstrate that our device cleaving and sample introduction processes do not significantly affect either the adsorbed oxygen concentration or the oxidation state of Te2-. Only a small increase in carbon concentration is observed in Figure S2. 3.2. Oxidation state dependence on substrate and on CdCl2 processing Figure 3 is comprised of representative Te 3d5/2 and S 2s XPS spectra of four different types of CdTe-containing film stacks, three of which were CdTe devices with the fourth one being a CdTe film grown directly on glass. The topmost (black dotted) traces in Figure 3a and 3b are XPS data from the SnO2 side of a cleaved CdTe solar cell that was made using a CdS emitter. Spectra from the opposite CdS side of these same cleaved devices are the solid black traces second from the top in both graphs. Note that the absolute amounts of tellurium and sulfur on the SnO2 side of the interface are smaller than on the emitter side and that these spectra are normalized in intensity such that a comparison of oxidation states is possible. Comparison of the relative amounts of oxidized chalcogens on each side of the cleave (black dotted versus black solid traces) shows a strikingly large difference in the extent of oxidation: the thin (< 1 nm) layer of material remaining on the SnO2 side is heavily oxidized, while material only a few angstroms away is much less so, with the

ACS Paragon Plus Environment

9

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

Page 10 of 29

Figure 3. In panel (a), Te 3d5/2 spectra of cleaved devices and test structures. Panel (b) shows S 2s spectra from the same samples. At the top of both panels (black dotted trace), XPS spectra from the SnO2 side of a cleaved device based on a CdS emitter that has undergone CdCl2/O2 treatment. Data from the opposite side (CdS) of the same cleaved devices is shown as the solid black traces second from the top. Blue traces in both panels are data taken from the emitter side of a device based on oxygenated CdS (CdS:O), a device also CdCl2-treated. The dark yellow trace fourth from the top corresponds to data taken from the emitter side of a device made using a CdS:O emitter but which was not CdCl2-treated. The bottom green trace in panel (a) is from the CdTe side of a cleaved test structure that consisted of CdTe grown directly on glass and was treated with CdCl2/O2 under standard conditions. tellurium showing a greater extent of oxidation than sulfur. Although varying somewhat with processing, this general trend of heavy oxidation of tellurium and sulfur right at the SnO2 interface was consistent for all the devices in this study. A second set of devices was made using an oxygenated CdS:O emitter. Use of oxygenated CdS as an emitter in CdTe solar cells has been shown to yield higher short-circuit current density and to reduce the interdiffusion of the emitter and absorber layers.16,23 XPS data from the CdS:O/CdTe side of this cleaved device is shown as the blue traces in Figure 3a-b. Interestingly, the amount of

ACS Paragon Plus Environment

10

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

ACS Applied Materials & Interfaces

Figure 4. J-V results for devices that were nominally identical except for the temperature used to active copper. oxidized tellurium found in this CdS:O-based device was not significantly different from that in devices made with pure CdS. As expected, the amount of oxidized sulfur found in the same devices was substantially increased relative to the plain CdS emitter (blue versus black solid trace in Figure 3b). The dark yellow traces in Figure 3, fourth from the top in both graphs, are from another set of devices having CdS:O emitters and made without a standard CdCl2-O2 treatment. Because devices made without this CdCl2 treatment are non-functional, back contacts to these film stacks were not made and J-V measurements were not performed. Absence of this CdCl2 treatment also made

ACS Paragon Plus Environment

11

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

Page 12 of 29

Figure 5. Te 3d5/2 spectra (a) and S 2s spectra (b) from both sides of cleaved devices processed identically except for temperature used to active copper. Panel (c) incorporates data from both sides of cleaved devices and shows the overall fraction of oxidized species. liquid nitrogen-based cleaving of these devices much more difficult: rather than spontaneously ACS Paragon Plus Environment

12

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

ACS Applied Materials & Interfaces

delaminating when immersed in liquid nitrogen, these samples required mechanical assistance with a razor blade to cleave. The percentage of oxidized tellurium decreased substantially in devices without standard CdCl2 treatment (yellow versus blue traces in Figure 3a). The percentage of oxidized sulfur remained about the same as devices made with CdS:O emitters and with the standard CdCl2 treatment (yellow versus blue traces in Figure 3b). The bottom trace (green) of Figure 3b shows a Te 3d5/2 spectrum taken from a film of CdTe grown directly on AF45 glass. The as-grown film was subjected to a standard oxygenated CdCl2 treatment and delaminated from the glass substrate using liquid nitrogen as was done with standard devices. Despite oxygen being present during both CdTe growth and post-growth CdCl2 treatment, the tellurium at the glass-CdTe interface in this sample has an almost undetectable amount of oxidized tellurium. 3.3. Oxidation state dependence on back contact processing Doping and construction of a good ohmic back contact to CdTe using copper usually require a heat treatment to diffuse and convert electronically inactive copper into CuCd acceptors.24 It was observed that the temperature used during this step greatly affected measured device parameters (Figure 4). To test whether this processing step also had an effect on the front SnO2-CdSCdTe interface, the devices measured in Figure 4 were thermomechanically cleaved and examined with XPS. Figure 5 is comprised of Te 3d5/2 spectra (panel a) and of S 2s spectra (panel b) taken from both sides of cleaved devices. Surprisingly, the spectra show systematic variations in the extent of oxidation of both tellurium and sulfur. This is perhaps best seen in panel (c) which represents the sum of the Te4+ and S6+ signals measured on both sides of the cleave and expresses them as a fraction of the total signal for each element: (Te4+/(Te4++Te2-) and S6+/(S6++S2-).

ACS Paragon Plus Environment

13

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

Page 14 of 29

Figure 6. Oxygen 1s spectra (a) for the set of devices there were nominally identical except for the temperature used to activate copper. In (b) are the fraction of the O 1s low binding energy component O 1slow (triangles, left axis) and the fraction of oxidized tellurium in the total of the oxidized tellurium and sulfur. Oxidized chalcogen fractions are calculated from the data in Figure 5. Oxygen chemical states on these same samples were also examined. Panel (a) of Figure 6 shows O 1s spectra taken from the CdS:O/CdTe side of the cleaved devices. Not shown are data from the SnO2 side which are dominated by oxygen signals originating from the tin oxide rather than from oxides of tellurium and sulfur. Two chemical states are observed in all the spectra, with one low binding energy (O 1slow) component centered on 530.7 eV and a higher binding energy component (O 1shigh) centered on 532.2 eV. Panel (b) of the same figure shows that the fraction of O 1slow tracks the fraction of tellurium in the sum of the oxidized tellurium and sulfur species, Te4+/(Te4++S6+). Thus, we assign the O 1slow component to oxygen coordinated with tellurium and the O 1shigh component to oxygen coordinated with sulfur. This assignment matches that expected from the lower electronegativity of tellurium relative to sulfur.25

ACS Paragon Plus Environment

14

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

ACS Applied Materials & Interfaces

4. Discussion It is clear from the comparison of XPS spectra taken on both sides of cleaved devices (Figures 3, 5) that there is an atomically sharp gradient in the oxidation of sulfur and tellurium at the SnO 2 interface. A reasonable question to ask is: what causes oxidation of these species? The mere fact that oxidation appears to take place adjacent to the SnO2 suggests one answer: the tin oxide itself acts as the oxidant or as an oxidation catalyst. We note that tin oxide is used commercially as an oxidation catalyst that can either give up its oxygen atoms directly or act as a Lewis acid and oxidize materials by electron transfer.26 Generally however, oxidation by SnO2 is believed to occur by oxygen atom transfer, also known as the Mars–van Krevelen mechanism.26 In this case of direct oxygen atom transfer, atoms lost from SnO2 and transferred to tellurium and sulfur would be replaced by molecular oxygen used in the CdTe growth and CdCl2 treatment process gases. This overall mechanism is consistent with the fact that under our processing conditions we do not observe any signs of tin reduction in the Sn 3d XPS spectra (data not shown). Figure 3 contains further clues as to the origin of oxidized species at the front of CdTe solar cells. Because there is little difference in the extent of oxidation of tellurium between the devices that use CdS and CdS:O emitter layers (black and blue solid traces, respectively in Figure 3a), we conclude that tellurium oxidation is controlled by something other than the presence of oxygen in the emitter layer through which tellurium diffuses during CdCl2 treatment. Te 3d5/2 spectra from cleaved devices that did not go through the CdCl2/O2 activation step (yellow trace, Figure 3a) show significantly less oxidation than those that did (blue trace), indicating two things: 1) some oxidation of tellurium takes place during CSS growth of CdTe but that 2) most tellurium oxidation occurs during the CdCl2/O2 activation step. Sulfur oxidation follows a somewhat different but predictable track. Comparison of the black and blue traces in Figure 3b show unsurprisingly that

ACS Paragon Plus Environment

15

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

Page 16 of 29

CdS:O emitter layers have much a much higher percentage of S6+ at the TCO interface than is found for CdS emitters. Oxidation of sulfur that takes place during reactive sputtering of CdS:O is such that the oxidizing conditions during CdTe growth and CdCl2/O2 processing do not significantly change the proportion of S6+ as measured on the emitter side of cleaved devices. As discussed below however, a later processing step used for the back contact affects the oxidation of both sulfur and tellurium. To test our hypothesis that SnO2 acts to catalyze oxidation of chalcogens during CdCl2/O2 processing, films of CdTe were grown directly on AF45 glass and subjected to our standard CdCl2/O2 treatment. Using the LN2-cleave method to remove the film, the glass-facing side of the CdTe was examined with XPS. Data for this experiment are shown in the bottom green trace in Figure 3(a), where it can be seen that insignificant Te4+ has formed despite the high temperatures and high oxygen partial pressures involved. This result, in conjunction with the other results in Figure 3, demonstrates conclusively that SnO2 catalyzes oxidation of chalcogens at the front of CdTe solar cells, thereby confirming one of the speculations of the earlier work of Poopanratana et al.14 One might wonder if oxidation of chalcogens at the front of CdTe devices matters to overall device efficiency. Empirically it has been found that for standard copper-doped devices, the CdCl2 activation step works best when performed in an oxygen containing environment.28 The CdTe community has also settled on tin oxide as the standard material for a front-of-cell transparent conducting oxide. As we have shown, both of these aspects of cell construction contribute to oxidation of tellurium and sulfur at the TCO interface. In some of our earlier experiments,29 we found indications that another processing step appeared to be affecting the extent of tellurium and sulfur oxidation: back contact annealing.

ACS Paragon Plus Environment

16

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

ACS Applied Materials & Interfaces

Back contact schemes for CdTe that utilize copper to produce a p+ layer adjacent to the metallization typically rely on a thermal treatment to convert electronically inactive copper into CuCd acceptors.24 This step is typically the last thermal treatment performed on superstrate CdTe solar cells. Because it is done at significantly lower temperatures (200 °C) than those used in CdTe deposition (600 °C) and in CdCl2 treatment (400 °C), one might think this last step would have only minor effects if any on the chemical makeup of front interfaces. Figures 4-6 represent a more in-depth investigation of our earlier experiments29 that indicated that back contact processing was causing changes at the SnO2-emitter interface. Figure 4 shows clearly that such low temperature processes have large effects on the overall cell electrical parameters. In the past, such changes were attributed solely to the optimization of copper-related defects. Although some changes in copper-related defects likely are present here as well, Figure 5 indicates another surprising effect: the low temperatures typically used to activate copper also cause significant changes in the extent of sulfur and tellurium oxidation at the front of the cell. Using data from each side of a cleaved device, the total fractions of oxidized species S 6+/(S6++S2-) and Te4+/(Te4++Te2-) observed in the core level spectra of Figure 5a and 5b are expressed as a function of temperature in Figure 5c. Intriguingly, Figure 5c shows that the fraction of oxidized tellurium and sulfur also goes through a maximum, peaking at roughly the same temperature as cell efficiencies, 230 °C (Figure 4). The data presented in Figure 5c would not be meaningful if devices cleaved inconsistently such that the thickness of emitter/absorber remaining on the SnO2 side varied. This is due to the strong gradient in oxidation in the vicinity of the SnO2 layer. However, the atomic concentrations on each side of the cleaved devices are remarkably constant for different cleaves even while varying the back-contact annealing temperature, indicating that cleaving occurs in the same location close

ACS Paragon Plus Environment

17

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

Page 18 of 29

to the SnO2 each time (Figure S3). This is consistent with our previous Auger electron microscopy results that showed laterally uniform composition in these types of samples.4 Back contact annealing also appears to affect speciation of oxygen atoms at the SnO2 interface. Figure 6b, in addition to providing assignments for the two chemical components observed in the O 1s spectra of Figure 6a, shows that the ratio of tellurium to sulfur among the oxidized constituents changes as a function of back contact annealing temperature. There appears to be a correlation between these data and the open circuit voltages in Figure 4, with Voc maxima at 220 °C and 240 °C corresponding to maxima in the amount of oxidized tellurium relative to oxidized sulfur at the same temperatures. Although it is counterintuitive that the relatively low temperature heat treatments used for back contact activation would affect interfaces that had already been subjected to significantly higher temperatures, there are at least a couple reasons why this might happen. Exposure of the film stacks to air after CdCl2/O2 treatment would allow the in-diffusion of atmospheric constituents such as water, possibly enabling other oxidation pathways at the later processing step. The back contact annealing step is also unique in that it is the first time that the TCO-emitter interface is heated in the presence of copper. Copper is a known oxidation promoter on SnO2,26 and as exemplified in Figure S4, copper was frequently be detected in these experiments at the TCOemitter interface despite the fairly high limit of detection for XPS (~0.1%). The exact mechanism by which in-situ grown chalcogenide oxides increase device efficiency is not clear, however previous work suggests several possible mechanisms. It is tempting to consider the tellurium and sulfur oxides as a type of high resistivity transparent (HRT) layer. Such a layer decreases regions of direct electrical contact between CdTe and SnO2, areas that might otherwise

ACS Paragon Plus Environment

18

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

ACS Applied Materials & Interfaces

act as weak diodes or shorts in the device. We note that in previous work, SnO2 HRTs were observed to be effective down to the thinnest that were measured, 13 nm.30 Based on standard substrate attenuation relations,31 using a 14 Å effective attenuation length32 for 915 eV photoelectrons emanating from the Te 3d5/2 transition, passing through an assumed oxide of CdTeO3, and taking into account the oxides found on both sides of LN2-cleaved devices, we arrive at approximate thicknesses of 1.1 nm to 1.7 nm for the oxide layers present in the six devices corresponding to Figures 4-6.

The assumed oxide species, CdTeO3, is the most

thermodynamically stable CdTe oxide and with its large bandgap (3.9 eV) could plausibly function as an HRT layer.33 It should be noted however that CdSO4 is probably also present, that the exact structure of the oxide species that we have detected in this work is unknown, and that it may not have a bulk analog. Based on our current measurements, the CdS-CdTe oxides consist of a complex mixture of cadmium, tellurium, sulfur, and oxygen. The device schematic shown in Figure 7 attempts to capture the important aspects of the front interface of CdTe-based solar cells derived from this work. Based on earlier Auger electron microscopy results4 that showed laterally uniform composition, the structure shown in panel (a) is more consistent with present results, however it should be noted that if the patches of CdCl2 and CdS/CdTe-oxides depicted in panel (b) were extremely small, and if devices delaminated such that both oxide and chloride were left on both sides of each cleaved device, this also could be consistent with our results. Both panels of Figure 7 underscore one key observation related to the possibility that the in-situ formed chalcogenide oxides may function as a HRT layer: in all the devices examined in the course of this investigation, measurable tellurium (~5%, see e.g. Figure S3 and Reference 4) was always found directly adjacent to the front SnO2 electrode. This contrasts with the standard model of

ACS Paragon Plus Environment

19

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

Page 20 of 29

Figure 7. Schematic of the SnO2 interface showing two possible structures incorporating the oxide layers revealed in this work. In (a), a laterally homogeneous structure with oxide layers on both sides of the CdCl2 cleave location and in (b) a laterally nonuniform structure. CdTe/CdS/TCO superstrate solar cells in which contact between a Te2- compound and the TCO occurs only in low efficiency devices and as a result of the CdS being too thin for the processing conditions.6,8,34 On the topic of whether oxygen alone can chemically passivate CdTe, theory and experiment are somewhat at odds. A recent transient reflectivity study of single crystal CdTe indicates that surface oxidation can reduce tellurium-related defects and increase carrier lifetimes.35 Several older works also appeared to show that CdTe oxides could passivate CdTe surfaces.11,12,36 A first principles investigation of oxygen adsorption on CdTe(110) indicated however that oxygen increased the number of mid-gap surface states.37 In conjunction with other recent work38, our current results suggest that both chlorine and oxygen are necessary to properly passivate the SnO2-emitter interface in CdTe-based solar cells. An analogous situation was reported by Bakulin et al., who showed theoretically that coadsorption of

ACS Paragon Plus Environment

20

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

ACS Applied Materials & Interfaces

oxygen and fluorine on GaAs and InAs caused the elimination of surface states that would otherwise pin the Fermi level near mid gap.39 Experimentally, Ahrenkiel et al. demonstrated that adding fluorine to oxide passivation layers on GaAs reduced the interfacial state density by about a factor of 50.40 In the case of InAs, an oxyfluoride passivation layer about 2 nm in thickness was shown to greatly reduce interfacial density of states and allow the construction of metal-oxidesemiconductor structures with excellent capacitance-voltage characteristics.41 One motivation cited for the theoretical work of Bakulin et al. was the difficulties associated with experimentally probing complex multinary surface passivation layers. The same difficulties hold for the buried chalcogenide oxide layers that are the subject of this paper. We suspect that future progress on making high performing CdTe solar cells will hinge in part on a better understanding of oxygen and halogen induced electronic changes at the surfaces of CdTe and related compounds. 5. Conclusions XPS analysis of thermomechanically cleaved CdTe devices and test structures has revealed that the tin oxide layer commonly used as a front electrode in CdTe-based solar cells catalyzes the formation of ultrathin CdS and CdTe oxide layers during device processing. These oxide layers have most likely been present in CdTe superstrate solar cells since their invention, escaping detection for decades through a combination of being deeply buried at an interface, having a thickness on the order of one nanometer, and being labile during common sample preparation techniques such as ion beam milling. We have shown that oxidation of sulfur and tellurium occurs mainly during post-growth CdCl2/O2 activation, but also develops further during a lower temperature step used to activate copper dopants in the back contact. This latter point is a crucial aspect of this work: late-stage processes used solely for back surface modification also have strong effects on the extent of chalcogen oxidation on the opposite, front side of the cell. We demonstrate

ACS Paragon Plus Environment

21

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

Page 22 of 29

that the best cell parameters coincide with the maximum in chalcogen oxidation. Our work indicates that both oxygen and chlorine are required for well-passivated front interfaces in CdTe solar cells, thereby providing motivation for an increased effort towards understanding defects involving chalcogen-halogen complexes in CdTe and related materials.

ACS Paragon Plus Environment

22

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

ACS Applied Materials & Interfaces

Corresponding Author *Craig L. Perkins

Acknowledgments This work was authored by Alliance for Sustainable Energy, LLC, the manager and operator of the National Renewable Energy Laboratory for the U.S. Department of Energy (DOE) under Contract No. DE-AC36-08GO28308. Funding provided by U.S. Department of Energy Office of Energy Efficiency and Renewable Energy Solar Energy Technologies Office under agreement 30307 and by the U.S. Department of Defense Office of Naval Research. The views expressed in the article do not necessarily represent the views of the DOE or the U.S. Government. The U.S. Government retains and the publisher, by accepting the article for publication, acknowledges that the U.S. Government retains a nonexclusive, paid-up, irrevocable, worldwide license to publish or reproduce the published form of this work, or allow others to do so, for U.S. Government purposes.

ACS Paragon Plus Environment

23

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

Page 24 of 29

Supporting Information Figure S-1, XPS spectra of a sputter depth profile of the CdS/CdTe side of a cleaved device. Panel (a) is Te 3d5/2 spectra. Panel (b), data from the Te 4s/S 2p region. Panel (c) shows the O 1s region and panel (d) shows Cl 2p spectra. Figure S-2, XPS spectra from a control experiment demonstrating that our sample handling procedure does not cause oxidation of Te2-. Black traces in all panels correspond to a sample cleaned by sputtering within the XPS system. Green traces correspond to this sample after transfer into LN2 and back into the XPS. In panel (a) are Te 3d5/2 spectra. In panel (b) are O 1s spectra. Panel (c) shows C 1s spectra. Figure S-3, XPS-derived atomic concentrations from the CdS:O/CdTe side (a) and SnO2-side (b) of cleaved devices corresponding to Figure 4-6. Figure S-4, a representative XPS spectrum of the Cu 2p3/2 region taken from the SnO2 side of a LN2-cleaved, copper-doped device.

ACS Paragon Plus Environment

24

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

ACS Applied Materials & Interfaces

References (1) Feldman, D. J.; Hoskins, J.; Margolis, R. M. Q2/Q3 2017 Solar Industry Update; NREL/PR--6A20-70406, 1409302; 2017. https://doi.org/10.2172/1409302. (2) Margolis, R.; Feldman, D.; Boff, D. Q4 2016/Q1 2017 Solar Industry Update; NREL/PR6A20-68425; National Renewable Energy Lab. (NREL), Golden, CO (United States), 2017. (3) Klein, A. Energy Band Alignment in Chalcogenide Thin Film Solar Cells from Photoelectron Spectroscopy. J. Phys. Condens. Matter 2015, 27 (13), 134201. https://doi.org/10.1088/0953-8984/27/13/134201. (4) Perkins, C. L.; Beall, C.; Reese, M. O.; Barnes, T. M. Two-Dimensional Cadmium Chloride Nanosheets in Cadmium Telluride Solar Cells. ACS Appl. Mater. Interfaces 2017, 9 (24), 20561–20565. https://doi.org/10.1021/acsami.7b03671. (5) Luria, J.; Kutes, Y.; Moore, A.; Zhang, L.; Stach, E. A.; Huey, B. D. Charge Transport in CdTe Solar Cells Revealed by Conductive Tomographic Atomic Force Microscopy. Nat. Energy 2016, 1, 16150. https://doi.org/10.1038/nenergy.2016.150. (6) Romeo, A.; Bätzner, D. .; Zogg, H.; Tiwari, A. . Recrystallization in CdTe/CdS. Thin Solid Films 2000, 361–362, 420–425. https://doi.org/10.1016/S0040-6090(99)00753-1. (7) Metzger, W. K.; Albin, D.; Levi, D.; Sheldon, P.; Li, X.; Keyes, B. M.; Ahrenkiel, R. K. Time-Resolved Photoluminescence Studies of CdTe Solar Cells. J. Appl. Phys. 2003, 94 (5), 3549– 3555. https://doi.org/10.1063/1.1597974. (8) Kumar, S. G.; Rao, K. S. R. K. Physics and Chemistry of CdTe/CdS Thin Film Heterojunction Photovoltaic Devices: Fundamental and Critical Aspects. Energy Environ. Sci. 2013, 7 (1), 45–102. https://doi.org/10.1039/C3EE41981A. (9) AbuEl-Rub, K. M.; Hahn, S.-R.; Tari, S.; Dissanayake, M. A. K. L. Effects of CdCl2 Heat Treatment on the Morphological and Chemical Properties of CdTe/CdS Thin Films Solar Cells. Appl. Surf. Sci. 2012, 258 (16), 6142–6147. https://doi.org/10.1016/j.apsusc.2012.03.020. (10) Niles, D. W.; Waters, D.; Rose, D. Chemical Reactivity of CdCl2 Wet-Deposited on CdTe Films Studied by X-Ray Photoelectron Spectroscopy. Appl. Surf. Sci. 1998, 136 (3), 221–229. https://doi.org/10.1016/S0169-4332(98)00340-7. (11) Rugen-Hankey, S. L.; Clayton, A. J.; Barrioz, V.; Kartopu, G.; Irvine, S. J. C.; McGettrick, J. D.; Hammond, D. Improvement to Thin Film CdTe Solar Cells with Controlled Back Surface Oxidation. Sol. Energy Mater. Sol. Cells 2015, 136, 213–217. https://doi.org/10.1016/j.solmat.2014.10.044. (12) Werthen, J. G.; Häring, J. ‐ P.; Bube, R. H. Correlation between Cadmium Telluride Surface Oxidation and Metal Junctions. J. Appl. Phys. 1983, 54 (2), 1159–1161. https://doi.org/10.1063/1.332091.

ACS Paragon Plus Environment

25

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

Page 26 of 29

(13) Solzbach, U.; Richter, H. Sputter Cleaning and Dry Oxidation of Cdte, Hgte, and Hg0.8cd0.2te Surfaces. Surf. Sci. 1980, 97 (1), 191–205. https://doi.org/10.1016/00396028(80)90113-2. (14) Pookpanratana, S.; Khan, F.; Zhang, Y.; Heske, C.; Weinhardt, L.; Bär, M.; Liu, X.; Paudel, N.; Compaan, A. Chemical Structure of Buried Interfaces in CdTe Thin Film Solar Cells. In 2010 35th IEEE Photovoltaic Specialists Conference; 2010; pp 000024–000027. https://doi.org/10.1109/PVSC.2010.5614076. (15) Albin, D. S.; Yan, Y.; Al-Jassim, M. M. The Effect of Oxygen on Interface Microstructure Evolution in CdS/CdTe Solar Cells. Prog. Photovolt. Res. Appl. 2002, 10 (5), 309–322. https://doi.org/10.1002/pip.426. (16) Meysing, D. M.; Reese, M. O.; Warren, C. W.; Abbas, A.; Burst, J. M.; Mahabaduge, H. P.; Metzger, W. K.; Walls, J. M.; Lonergan, M. C.; Barnes, T. M.; Wolden, C. A. Evolution of Oxygenated Cadmium Sulfide (CdS:O) during High-Temperature CdTe Solar Cell Fabrication. Sol. Energy Mater. Sol. Cells 2016, 157, 276–285. https://doi.org/10.1016/j.solmat.2016.05.038. (17) McGott, D. L.; Kempe, M. D.; Glynn, S.; Bosco, N.; Barnes, T. M.; Haegel, N. M.; Wolden, C. A.; Reese, M. O. Thermomechanical Lift-Off and Recontacting of CdTe Solar Cells. ACS Appl. Mater. Interfaces 2018. https://doi.org/10.1021/acsami.8b16476. (18) Mao, D.; Blatz, G.; Wickersham, C. E.; Gloeckler, M. Correlative Impurity Distribution Analysis in Cadmium Telluride (CdTe) Thin-Film Solar Cells by ToF-SIMS 2D Imaging. Sol. Energy Mater. Sol. Cells 2016, 157, 65–73. https://doi.org/10.1016/j.solmat.2016.05.019. (19) Handbook of X-Ray Photoelectron Spectroscopy: A Reference Book of Standard Spectra for Identification and Interpretation of XPS Data, Update.; Moulder, J. F., Chastain, J., Eds.; Perkin-Elmer Corporation: Eden Prairie, Minn, 1992. (20)

Igor Pro 7; Wavemetrics.

(21)

MultiPak v9.6.1.7; ULVAC-PHI, Inc.

(22) Duncan, D. A.; Kephart, J. M.; Horsley, K.; Blum, M.; Mezher, M.; Weinhardt, L.; Häming, M.; Wilks, R. G.; Hofmann, T.; Yang, W.; Bär, M.; Sampath, W. S.; Heske, C. Characterization of Sulfur Bonding in CdS:O Buffer Layers for CdTe-Based Thin-Film Solar Cells. ACS Appl. Mater. Interfaces 2015, 7 (30), 16382–16386. https://doi.org/10.1021/acsami.5b03503. (23) Wu, X.; Dhere, R. G.; Yan, Y.; Romero, I. J.; Zhang, Y.; Zhou, J.; DeHart, C.; Duda, A.; Perkins, C.; To, B. High-Efficiency Polycrystalline CdTe Thin-Film Solar Cells with an Oxygenated Amorphous Cds (a-CdS:O) Window Layer. In Conference Record of the TwentyNinth IEEE Photovoltaic Specialists Conference, 2002.; 2002; pp 531–534. https://doi.org/10.1109/PVSC.2002.1190619. (24) Wei, S.-H.; Zhang, S. B.; Zunger, A. First-Principles Calculation of Band Offsets, Optical Bowings, and Defects in CdS, CdSe, CdTe, and Their Alloys. J. Appl. Phys. 2000, 87 (3), 1304– 1311. https://doi.org/10.1063/1.372014.

ACS Paragon Plus Environment

26

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

ACS Applied Materials & Interfaces

(25) Gordy, W.; Thomas, W. J. O. Electronegativities of the Elements. J. Chem. Phys. 1956, 24 (2), 439–444. https://doi.org/10.1063/1.1742493. (26) Batzill, M.; Diebold, U. The Surface and Materials Science of Tin Oxide. Prog. Surf. Sci. 2005, 79 (2–4), 47–154. https://doi.org/10.1016/j.progsurf.2005.09.002. (27) CRC Handbook of Chemistry and Physics, 64. ed., 3. printing.; Weast, R. C., Ed.; CRC Press: Boca Raton, Fla, 1984. (28) Romeo, A.; Terheggen, M.; Abou‐ Ras, D.; Bätzner, D. L.; Haug, F.-J.; Kälin, M.; Rudmann, D.; Tiwari, A. N. Development of Thin-Film Cu(In,Ga)Se2 and CdTe Solar Cells. Prog. Photovolt. Res. Appl. 2004, 12 (2–3), 93–111. https://doi.org/10.1002/pip.527. (29) Perkins, C. L.; Beall, C.; Burst, J. M.; Kanevce, A.; Reese, M. O.; Barnes, T. M. New Insights on the Structure and Chemistry of the Tin Oxide-Emitter Interface in CdTe Solar Cells as Revealed by Thermomechanical Cleavage and Electron Spectroscopy; Contributed oral presentation, AVS 63rd International Symposium, Nashville, TN (2016), 2016. (30) Ferekides, C. S.; Mamazza, R.; Balasubramanian, U.; Morel, D. L. Transparent Conductors and Buffer Layers for CdTe Solar Cells. Thin Solid Films 2005, 480–481, 224–229. https://doi.org/10.1016/j.tsf.2004.11.094. (31) Cumpson, P. J.; Seah, M. P. Elastic Scattering Corrections in AES and XPS. II. Estimating Attenuation Lengths and Conditions Required for Their Valid Use in Overlayer/Substrate Experiments. Surf. Interface Anal. 1997, 25 (6), 430–446. https://doi.org/10.1002/(SICI)10969918(199706)25:63.0.CO;2-7. (32)

NIST Standard Reference Database 82.

(33) Menéndez-Proupin, E.; Gutiérrez, G.; Palmero, E.; Peña, J. L. Electronic Structure of Crystalline Binary and Ternary Cd-Te-O Compounds. Phys. Rev. B 2004, 70 (3), 035112. https://doi.org/10.1103/PhysRevB.70.035112. (34) McCandless, B. E.; Moulton, L. V.; Birkmire, R. W. Recrystallization and Sulfur Diffusion in CdCl2-Treated CdTe/CdS Thin Films. Prog. Photovolt. Res. Appl. 1997, 5 (4), 249–260. https://doi.org/10.1002/(SICI)1099-159X(199707/08)5:43.0.CO;2-S. (35) He, X.; Punpongjareorn, N.; Wu, C.; Davydov, I. A.; Yang, D.-S. Ultrafast Carrier Dynamics of CdTe: Surface Effects. J. Phys. Chem. C 2016, 120 (17), 9350–9356. https://doi.org/10.1021/acs.jpcc.6b02771. (36) Suita, M.; Taguchi, T. Thermal Oxidation of CdTe Surfaces and the Properties of MOS Diodes. Nucl. Instrum. Methods Phys. Res. Sect. Accel. Spectrometers Detect. Assoc. Equip. 1989, 283 (2), 268–273. https://doi.org/10.1016/0168-9002(89)91370-3. (37) Miotto, R.; Kiss, F. D.; Ferraz, A. C. Oxygen Adsorption on CdTe(110). Surf. Sci. 2003, 525 (1), 24–32. https://doi.org/10.1016/S0039-6028(02)02569-4.

ACS Paragon Plus Environment

27

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

Page 28 of 29

(38) Perkins, C. L.; Ablekim, T.; Barnes, T. M.; Kuciauskas, D.; Lynn, K. G.; Nemeth, W.; Reese, M. O.; Swain, S. K.; Metzger, W. K. Interfaces Between CdTe and ALD Al2O3. IEEE J. Photovolt. 2018, 1–4. https://doi.org/10.1109/JPHOTOV.2018.2870139. (39) Bakulin, A. V.; Kulkova, S. E.; Aksenov, M. S.; Valisheva, N. A. Fluorine and Oxygen Adsorption and Their Coadsorption on the (111) Surface of InAs and GaAs. J. Phys. Chem. C 2016, 120 (31), 17491–17500. https://doi.org/10.1021/acs.jpcc.6b05308. (40) Ahrenkiel, R. K.; Kazmerski, L. L.; Ireland, P. J.; Jamjoum, O.; Russell, P. E.; Dunlavy, D.; Wagner, R. S.; Pattillo, S.; Jervis, T. Reduction of Surface States on GaAs by the Plasma Growth of Oxyfluorides. J. Vac. Sci. Technol. 1982, 21 (2), 434–437. https://doi.org/10.1116/1.571672. (41) Aksenov, M. S.; Kokhanovskii, A. Y.; Polovodov, P. A.; Devyatova, S. F.; Golyashov, V. A.; Kozhukhov, A. S.; Prosvirin, I. P.; Khandarkhaeva, S. E.; Gutakovskii, A. K.; Valisheva, N. A.; Tereshchenko, O. E. InAs-Based Metal-Oxide-Semiconductor Structure Formation in LowEnergy Townsend Discharge. Appl. Phys. Lett. 2015, 107 (17), 173501. https://doi.org/10.1063/1.4934745.

ACS Paragon Plus Environment

28

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

ACS Applied Materials & Interfaces

TOC graphic

ACS Paragon Plus Environment

29