The growth of atomic layer deposited (ALD) Al2O3 on planar ZnSe substrates is studied using in situ spectroscopic ellipsometry. An untreated ZnSe surface requires an incubation period of 27 cycles of ALD Al2O3 before film growth is observed. Pretreating the surface with an ultraviolet generated ozone lowers the incubation to 17 cycles, whereas a plasma-enhanced ALD Al2O3 process can further lower the incubation period to 13 cycles. The use of ozone or plasma-activated oxygen species on ZnSe is found to create ZnO and SeO2, which are responsible for converting ZnSe from a hydrophobic to a hydrophilic surface. The interfacial layer between Al2O3 and ZnSe is mapped using high-resolution transmission electron microscopy and scanning transmission electron microscopy/energy dispersive spectroscopy. SeO2 is volatile and leaves a zinc-rich interface, which is 4.3 nm thick for the ultraviolet generated ozone pretreated sample and 2.5 nm for the plasma-enhanced ALD process.

ZnSe is a II–VI wide bandgap semiconductor with a direct band gap1 of 2.67 eV and an exciton binding energy of 20 meV.2 It is intrinsically n-type with a doping density of 1015 cm−3 and can be extrinsically doped n-type using Al or Cl or p-type using nitrogen plasma.3,4 ZnSe also offers low absorption in the IR spectral region, can be scalably manufactured, and is mechanically robust.5,6 These properties make ZnSe a versatile material in bulk, thin film, or fiber form for active and passive use in optoelectronic devices, such as UV photodetectors,3 light emitting diodes,7 laser diodes,8 lenses, IR transparent windows, and optical fibers.3–10 

Development of ZnSe-based optoelectronic devices depends on the ability to control ZnSe surfaces and interfaces. For example, devices, such as photodetectors and laser emitting diodes, require high-quality epitaxially matched ZnSe thin films on GaAs substrates.9–11 Recent efforts to dope ZnSe crystals with transition metals, such as Cr2+ or Fe2+, have yielded active single crystal laser host materials, which emit in the spectrally important midinfrared (MIR) region.12 In contrast to ZnSe thin films, recent developments in mid-IR optical fiber laser systems have explored the potential to capture the desirable attributes of difficult to grow single crystals of Fe or Cr-doped ZnSe in host matrices; hence, here, the use of optically active ZnSe particles embedded in a parent glass matrix results in the formation of a highly functional, complex optical composite.13–18 The synthesis for such composites imparts stringent requirements on the stability of the ZnSe particle in the parent matrix during high temperature melting, fabrication, and drawing processes of fibers. Here, any degradation of the dopant crystal or processing protocols that impart undesirable impurities into the matrix that hinders the absorption or emission behavior of the active species are to be avoided.19 To this end, the possibility of using the robust conformal nature of atomic layer deposition to passivate the dopant surface to enhance its stability in a composite is highly desirable.

The varied requirements for ZnSe-based optoelectronic devices create the need to understand conditions that lead to favorable growth of materials on ZnSe surfaces. While many techniques exist to deposit films on planar substrates, atomic layer deposition (ALD) provides the only avenue to deposit functional films with monolayer fidelity on planar and powder substrates, alike. Previously, ZnSe crystalline substrates have been used as a transmission medium to study in situ infrared spectroscopy of Al2O3 ALD.20 However, little is known about initiating ALD reactions on the surface of ZnSe itself. This study evaluates the fundamental mechanism of Al2O3 ALD growth on planar ZnSe.

We present a surface chemistry approach to understand the nature of ALD film growth on a ZnSe surface. On a planar ZnSe substrate, an Al2O3 layer may represent a gate dielectric layer used in a transistor, while a conformal Al2O3 on a ZnSe particle may provide protection against dissolution during high temperature synthesis of mid-IR optical materials.21 The underpinning of these varied requirements requires first the manipulation and control of the surface chemistry of ZnSe. It is shown that a predeposition oxidation step or the use of O2 plasma as a coreactant can change the surface chemistry of ZnSe from hydrophobic to hydrophilic due to the formation of SeO2. The formation of a selenium oxide layer leads to shorter incubation cycles and growth of ALD Al2O3 on a planar substrate.

A Fiji Gen2 PEALD system from Veeco® with an Ebara® multistage dry vacuum pump A30W (pumping speed 3600 l/min) was used to deposit Al2O3 films on ZnSe wafers. The base pressure of the ALD chamber was maintained at 5 × 10−7 Torr with a working pressure of ∼120 mTorr. The ALD process consisted of alternating pulses of trimethylaluminum (TMA) (Millipore Sigma) = 0.06 s and de-ionized water (DIW) pulse = 0.06 s, each separated with an Ar purge for 15 s. The purge is longer than necessary as it allows for recording of the in situ data via ellipsometry with sufficient frequency and fidelity. The deposition of Al2O3 was conducted at 175 °C. As an alternate, a PEALD Al2O3 process at 175 °C consisting of 0.06 s of TMA and O2 plasma (300 W for 4 s) was used as well.

The deposition rate of the film was monitored by in situ spectroscopic ellipsometry using a J. A. Woollam® M-2000, with a wavelength range from 273 to 1690 nm. The thickness was modeled using the completeease® software, consisting of a Cauchy layer optimized with Al2O3 optical constants on a ZnSe substrate layer (ZnSe model provided in the completeease substrate library).

The ZnSe substrate (10 mm in diameter) was supplied by EKSMA Optics USA DBA Altos Photonics, Inc. A piece of black carbon paper was placed under the ZnSe substrate to prevent any backside reflection from ZnSe (see Fig. 1, inset). The ZnSe substrates were ultrasonically cleaned with isopropyl alcohol (IPA) and DIW. For the case where surface oxidation of ZnSe was required, a 15-min exposure to UV-O3 was conducted on a Model E511 made by Ossila®. The dominant wavelengths of the UV light are 185 and 254 nm. The lamp power is 20 μW/cm2. To optically model this substrate, a Cauchy layer (i.e., insulator) was used on a ZnSe substrate. The resulting model consistently predicted a ∼4 nm film, identified as SeO2 using XPS. This interfacial layer was kept fixed during the modeling of the subsequent ALD Al2O3. Prior to ALD, static water contact angle (WCA) measurements were taken on an Ossila E511 and analyzed using Ossila contact angle software (Version 1.1.02). DIW droplets of a volume of 10 μl were dropped using pipette on untreated and UV-O3 treated ZnSe film substrates.

FIG. 1.

In situ growth curves obtained via spectroscopic ellipsometry for 100 cycles of ALD Al2O3 deposited at 175 °C on ZnSe substrates. The untreated ZnSe shows the longest incubation period and results in a final Al2O3 thickness of 6.7 nm. The UV-O3 pretreated ZnSe results in an Al2O3 thickness of 7.8 nm. The PEALD process yields the shortest incubation and results in a final Al2O3 thickness of 11.3 nm. Inset: ZnSe coupons are transparent (right) and must be kept on carbon paper (left) to prevent backside reflection during in situ ellipsometry measurements.

FIG. 1.

In situ growth curves obtained via spectroscopic ellipsometry for 100 cycles of ALD Al2O3 deposited at 175 °C on ZnSe substrates. The untreated ZnSe shows the longest incubation period and results in a final Al2O3 thickness of 6.7 nm. The UV-O3 pretreated ZnSe results in an Al2O3 thickness of 7.8 nm. The PEALD process yields the shortest incubation and results in a final Al2O3 thickness of 11.3 nm. Inset: ZnSe coupons are transparent (right) and must be kept on carbon paper (left) to prevent backside reflection during in situ ellipsometry measurements.

Close modal

The changes to the ZnSe surface chemistry were probed by x-ray photoelectron spectroscopy (XPS) using an ESCALAB-250Xi XPS system from ThermoFisher Scientific®. Experiments were performed at a pressure below 7 × 10−9 mbar using Al-Kα monochromatic radiation and an operating power of 300 W (15 kV, 20 mA). XPS data analysis and peak deconvolutions were analyzed using the open source XPS Peak 41 software. Binding energies were calibrated using a C 1s peak at 284.5 eV.

An FEI® Themis Z scanning/transmission electron microscopy (S/TEM) with Cs probe correction was used to analyze the Al2O3/ZnSe interfaces; the operating voltage was 200 kV, and the system is equipped with a monochromator, a bottom-mounted FEI Ceta 16 M CMOS camera, a Fischione Instruments Model 3000 high angle annular dark-field (HAADF) STEM detector, and a SuperX windowless SDD energy dispersive spectroscopy (EDS) system (solid angle of collection = 0.67 sr). Conventional high-resolution (HR-TEM) imaging, HAADF-STEM imaging, and STEM-EDS mapping were performed; for HAADF-STEM work, the probe semiangle of convergence was 22 mrad and the probe current was set to ∼20 pA for high-resolution imaging and ∼400 pA for STEM-EDS mapping; the inner collection angle of the HAADF detector for HAADF-STEM imaging was ∼70 mrad. An in situ lift-out method using an FEI Helios Nanolab 600i dual focused ion beam/scanning electron microscope was used to extract and prepare lamellas for S/TEM analysis from the specimens as described elsewhere;22–24 prior to preparation of the lamellas, the specimens were ex situ coated with ∼50 nm of Cr using electron beam evaporation to protect/preserve the Al2O3 film and the film/substrate interface.25 

Image J® was used for analysis of TEM images. Thickness measurements were made on cross-sectional images. A set of 20 measurements were usually made to obtain statistically rigorous averages and standard deviations obtained.

The growth behavior of 100 cycles of ALD Al2O3 is captured via in situ ellipsometry in Fig. 1 for three planar samples: (1) untreated ZnSe, (2) ZnSe treated with 15 min of ultraviolet-ozone (UV-O3), and (3) untreated ZnSe with a plasma-enhanced ALD (PEALD) Al2O3 process with O2 plasma as an oxidant. The Al2O3 thickness variation is plotted as a function of deposition time (in minutes). The untreated ZnSe exhibits the longest incubation period of 27 ALD cycles until a linear growth rate (0.093 nm/cy) of Al2O3 is achieved. Here, the incubation cycle, c, is calculated as c=100tg, where t is the final thickness (nm) at 100 cycles and g is the growth rate (nm/cy). The delay in nucleation results in the thinnest Al2O3 film of 6.8 nm after 100 cycles. On the other hand, pretreating with UV-O3 alters the surface chemistry of ZnSe and shortens the incubation delay of Al2O3 to 17 cycles. As a result, a 7.8 nm Al2O3 film is obtained after 100 ALD cycles with a steady-state growth rate in the linear regime of 0.094 nm/cy. Finally, the PEALD process shows the shortest incubation period of 13 cycles on an untreated ZnSe surface. The growth at the end of 100 cycles is 11.3 nm, and the steady-state growth rate in the linear regime is 0.130 nm/cycle, which is in good agreement with an O2 plasma ALD of Al2O3 on ZnSe and silicon substrates.20,26 These data are collectively summarized in Table I. The data suggest that a pretreatment of the ZnSe surface with UV-O3 or the use of a strong oxidant, such as an O2 plasma during ALD, reduces the incubation delays and improves growth characteristics of Al2O3 on a ZnSe surface.

TABLE I.

ALD growth parameters of Al2O3 extracted from in situ spectroscopic ellipsometry on untreated ZnSe (first row) and ZnSe treated with 15 min of UV-O3 prior to deposition (second row). Both the samples were subjected to thermal ALD using TMA and H2O. The third row shows PEALD of Al2O3 using TMA and O2 plasma on untreated ZnSe. The temperature of deposition was maintained at 175 °C across all three samples. Additionally, the thickness measurements from cross-sectional TEM are listed and show a good correlation with ellipsometry.

Thickness via ellipsometry
(nm)
Interfacial thickness via ellipsometry
(nm)
Thickness via TEM
(nm)
Interfacial thickness via TEM
(nm)
Incubation period
(cy)
Steady-state growth state
(nm/cy)
Untreated 6.8 ± 0.026 — 5.9 ± 1.4 — 27 0.093 
UV-O3 7.8 ± 0.029 4.0 7.1 ± 1.2 4.3 ± 0.6 17 0.094 
PEALD 11.3 ± 0.026 — 11.2 ± 0.6 2.5 ± 0.3 13 0.130 
Thickness via ellipsometry
(nm)
Interfacial thickness via ellipsometry
(nm)
Thickness via TEM
(nm)
Interfacial thickness via TEM
(nm)
Incubation period
(cy)
Steady-state growth state
(nm/cy)
Untreated 6.8 ± 0.026 — 5.9 ± 1.4 — 27 0.093 
UV-O3 7.8 ± 0.029 4.0 7.1 ± 1.2 4.3 ± 0.6 17 0.094 
PEALD 11.3 ± 0.026 — 11.2 ± 0.6 2.5 ± 0.3 13 0.130 

Additionally, final film thickness is also determined via HRTEM of film cross sections and is presented in Table I. A good correlation is observed between ellipsometry and TEM. The TEM results are discussed in greater detail below.

Static water contact angle (WCA) measurements were performed to understand the surface chemistry of ZnSe planar substrates and the effect of surface oxidation (Fig. 2). The untreated ZnSe surface is hydrophobic with a contact angle of 112°, which is consistent with previous reports for bulk ZnSe.27 Upon UV-O3 treatment, the WCA reduces to 32°, indicating the hydrophilic nature of the oxidized surface. The WCA results from planar substrates help to infer that the UV-O3 treatment can improve the wettability of ZnSe, resulting in enhanced nucleation and growth of Al2O3 ALD. This result indicates that if functional ALD coatings are required on ZnSe, surface oxidation is necessary to enhance coating adherence and conformal coverage. While other modifications may lead to similar results, surface oxidation through UV-O3 or oxygen plasma is a straightforward process that results in enhanced nucleation and growth on ZnSe.

FIG. 2.

Static water contact angle measurements on untreated ZnSe and UV-O3 ZnSe substrates. The untreated sample is hydrophobic (contact angle ∼112°), whereas exposing to UV-O3 converts the substrate surface to hydrophilic (contact angle ∼32°).

FIG. 2.

Static water contact angle measurements on untreated ZnSe and UV-O3 ZnSe substrates. The untreated sample is hydrophobic (contact angle ∼112°), whereas exposing to UV-O3 converts the substrate surface to hydrophilic (contact angle ∼32°).

Close modal

In order to confirm the oxidation state of the ZnSe surface after pretreatment, the ZnSe substrates were probed by XPS. The Zn 2p spectra of the as-received Zn-Se is shown in Fig. 3(a). The spectrum can be deconvoluted into two peaks. The Zn 2p3/2 binding energy (BE) peak at 1021.98 eV corresponds to the Zn-Se bonding, whereas the peak at 1022.57 eV corresponds to Zn-O.28,29 Even with ZnO detected via XPS, the surface properties of ZnSe remain hydrophobic, suggesting limited oxidation of the surface. In Fig. 3(b), the Se 3d high-resolution spectrum shows peaks at 54.15 and 55.01 eV corresponding to Se 3d5/2 and Se 3d3/2, respectively. These peaks are related to the Se-Zn bond. We note that the peaks are closely separated with a spin–orbit splitting of 0.86 eV only.30 

FIG. 3.

(a) XPS high-resolution spectrum of Zn 2p from the untreated, as-received ZnSe showing primarily Zn-Se bonding with a little Zn-O bonding. (b) XPS high-resolution spectrum of Se 3d from untreated, as-received ZnSe showing Se-Zn bonding. (c) XPS high-resolution spectrum of Zn 2p from the UV-O3 treated ZnSe showing primarily Zn-Se bonding with a little Zn-O bonding, similar to the untreated ZnSe sample. (d) XPS high-resolution spectrum of Se 3d from the UV-O3 treated ZnSe showing Se-Zn bonding but importantly, Se-O formation as well. All peaks were normalized to carbon 1 s = 284.5 eV.

FIG. 3.

(a) XPS high-resolution spectrum of Zn 2p from the untreated, as-received ZnSe showing primarily Zn-Se bonding with a little Zn-O bonding. (b) XPS high-resolution spectrum of Se 3d from untreated, as-received ZnSe showing Se-Zn bonding. (c) XPS high-resolution spectrum of Zn 2p from the UV-O3 treated ZnSe showing primarily Zn-Se bonding with a little Zn-O bonding, similar to the untreated ZnSe sample. (d) XPS high-resolution spectrum of Se 3d from the UV-O3 treated ZnSe showing Se-Zn bonding but importantly, Se-O formation as well. All peaks were normalized to carbon 1 s = 284.5 eV.

Close modal

The Zn high-resolution spectra of the ZnSe substrate pretreated with UV-O3 are shown in Fig. 3(c). Similar to the untreated ZnSe surface, the Zn 2p3/2 peak is present at higher BE—1021.8 eV for Zn-Se and 1022.45 eV for Zn-O. Increased elemental ZnO or ZnOH after UV-O3 treatment is observed.29 The effect of UV-O3 pretreatment on Se is shown in Fig. 3(d). After UV-O3 pretreatment, the primary BE of the Se 3d5/2 can be seen to split into two. The lower BE pair with Se 3d5/2 at 53.69 eV corresponds to ZnSe, whereas the higher BE Se 3d5/2 at 54.64 eV corresponds to elemental (Se°) or can also be assigned to Se-OH.29,31 Additionally, BE peaks are observed at 58.64 eV (3d5/2) and 59.5 eV (3d3/2) corresponding to the Se-O bond formation.28,29,31 The oxidation of the ZnSe is considered to cleave the Zn-Se bonding to form SeO2.32 It is possible that elemental Se may exist at the surface, but the BE is convoluted with Se(OH)4 species.29 Thus, the XPS analysis reveals that pretreating ZnSe with UV-O3 generates oxides of Se.27–29,31,32 While ZnO is present on the surface of untreated ZnSe as well, the presence of SeO2 is observed upon UV-O3 pretreatment. In light of these findings, it can be concluded that the presence of an Se oxide helps the conversion of the ZnSe surface from hydrophobic to hydrophilic.

We next focus on the surface chemical state of the Al2O3 ALD on ZnSe. In Figs. 4(a) and 4(b), the Al 2p and O 1s high-resolution spectra are shown for the 7.8 nm Al2O3 grown on UV-O3 pretreated ZnSe, respectively. The average escape depth of electrons from amorphous Al2O3 is ∼3 nm,33 and thus, it is unlikely that Zn or Se related to the ZnSe substrate can be detected. The Al 2p is centered at 74.26 eV, whereas the O 1s can be deconvoluted into two species. The peak at 530.69 eV is attributed to the O bonding with Al, whereas the peak at 531.44 eV is associated with surface bound OH groups.34 In this respect, the Al and O high-resolution spectra are unremarkable.

FIG. 4.

(a) Al 2p high-resolution spectra of 7.8 nm ALD Al2O3 on UV-O3 treated ZnSe. (b) O 1s high-resolution spectra of 7.8 nm ALD Al2O3 on UV-O3 treated ZnSe. (c) Se 3d high-resolution spectra observed on 7.8 nm ALD Al2O3 on UV-O3 treated ZnSe. This was the only sample on which Se was observed.

FIG. 4.

(a) Al 2p high-resolution spectra of 7.8 nm ALD Al2O3 on UV-O3 treated ZnSe. (b) O 1s high-resolution spectra of 7.8 nm ALD Al2O3 on UV-O3 treated ZnSe. (c) Se 3d high-resolution spectra observed on 7.8 nm ALD Al2O3 on UV-O3 treated ZnSe. This was the only sample on which Se was observed.

Close modal

Surprisingly, in addition to Al and O signals, a signal from Se is also obtained. This is shown in Fig. 4(c). The Se 3d5/2 is positioned at 53.61 eV. This peak cannot be related to ZnSe, as no Zn was detected. The peak position is also not related to Se0/Se-OH as shown previously in Fig. 3(d), as that peak lies at 54.64 eV, i.e., almost 1 eV higher. We speculate that Se discovered on the ALD Al2O3 surface is likely bonded to Al as AlxSey. A previous study by Chen et al. report this behavior where Al evaporated on ZnSe surfaces yields a Se core spectrum similar to ZnSe.35 The complimentary Al high-resolution spectra are unable to resolve the bonding as the Al-O and Al-Se peaks overlap.

Furthermore, it has been verified through XPS (not shown) of the ALD Al2O3 on untreated ZnSe (thinner and at thickness 6.8 nm) and PEALD Al2O3 (thicker and at thickness 11.3 nm) that the Se signal is absent on these surfaces. While the partial diffusion of Se cannot be ruled out for the PEALD sample, it is likely that the origin of the Se detected on the UV-O3 pretreated sample is due to the outdiffusion of Se from the Al2O3/ZnSe interface.

Both Se and SeO2 are volatile species and known to sublimate above 100 °C.36–38 The XPS data suggest that Se in its elemental or oxide form is mobile within the amorphous Al2O3 matrix. The detection of Se on the surface of Al2O3 for a UV-O3 pretreated sample suggests that Se migration to the surface followed by volatilization may occur during the deposition process.

HAADF-STEM cross-sectional images are shown in Fig. 5 for all three samples. We note that ZnSe is a polycrystalline substrate where a majority of the grains have an out-of-plane crystallographic orientation along the [110] direction. In Fig. 5(a), untreated ZnSe with amorphous Al2O3 coating is shown. The Al2O3 ALD film is measured to be 5.9 ± 1.4 nm. In Fig. 5(b), UV-O3 pretreated ZnSe with Al2O3 coating is shown. An Al2O3 film of 7.1 ± 1.2 nm can be measured. In addition, an interfacial region of approximately 4.3 ± 0.6 nm is measured corresponding to SeO2. The interfacial region appears amorphous, albeit lighter in contrast than Al2O3. This is consistent with an ∼4 nm SeO2 film predicted by spectroscopic ellipsometry and confirmed by XPS. In Fig. 4(c), as-received ZnSe with Al2O3 PE-ALD coating is shown. An Al2O3 film of approximately 11.2 ± 0.6 nm can be measured. However, the thickness of the SeO2 layer is 2.5 ± 0.3 nm due to the short exposure time of the ZnSe to O2 plasma before Al2O3 starts to conformally grow and protect the underlying ZnSe substrate. These results are presented collectively in Table I and show a good correlation with in situ ellipsometry data.

FIG. 5.

HAADF-STEM images of (a) untreated ZnSe coated with Al2O3 ALD. (b) UV-O3 treated ZnSe coated with Al2O3 ALD and (c) untreated ZnSe coated with Al2O3 PE-ALD. Scale bar = 5 nm.

FIG. 5.

HAADF-STEM images of (a) untreated ZnSe coated with Al2O3 ALD. (b) UV-O3 treated ZnSe coated with Al2O3 ALD and (c) untreated ZnSe coated with Al2O3 PE-ALD. Scale bar = 5 nm.

Close modal

STEM-EDS elemental mapping of the cross sections of Al2O3 ALD on the three samples are shown in Fig. 6. The Al is shown in blue, Zn in red, O in teal, and Se in green. The Al2O3 deposited on untreated ZnSe shows a uniform distribution of Al and O with a sharp transition to the Zn and Se signals. The Zn and Se signals overlap, indicating the presence of ZnSe. For Al2O3 deposited on the pretreated UV-O3 sample, the O spread is much wider than the Al, extending partially into the substrate. This shows the presence of an O-rich interface layer, in line with the XPS and HAADF-TEM data. Furthermore, Se appears to extend upward into the Al2O3 film, leaving behind a Se-depleted layer, while Zn is rather uniform. Finally, the scan from the PEALD sample reveals the O signal extending deeper into ZnSe than Al. The Zn and Se signals overlap.

FIG. 6.

Elemental maps of untreated ZnSe coated with Al2O3 ALD (first column), UV-O3 treated ZnSe coated with Al2O3 ALD (second column), and untreated ZnSe coated with Al2O3 PE-ALD (third column), indicating position of Al (blue), O (teal), Se (green), and Zn (red). Scale bar = 10 nm.

FIG. 6.

Elemental maps of untreated ZnSe coated with Al2O3 ALD (first column), UV-O3 treated ZnSe coated with Al2O3 ALD (second column), and untreated ZnSe coated with Al2O3 PE-ALD (third column), indicating position of Al (blue), O (teal), Se (green), and Zn (red). Scale bar = 10 nm.

Close modal

To quantify these observations further, the STEM-EDS elemental line scans are shown in Figs. 7(a)7(c) for the untreated, UV-O3, and PEALD samples, respectively. The y axis are the normalized intensity counts. To demarcate the various regions probed, the following procedure is adopted. The region next to the origin on the x axis (i.e., distance = 0 nm) consists of the Cr layer, which is deposited on top of the Al2O3 film during sample preparation. The O signal rises in this region due to the formation of chromium oxide. Thus, to conduct the analysis at the Al2O3 film/ZnSe substrate interface, the data are screened first. The distance at which the O counts first decrease and then stabilize is noted to be the top interface between Cr and Al2O3. The analyses from this point onward until the measured film thickness (determined via HAADF images) encompass the Al2O3 layer and beyond that the ZnSe substrate. As expected, the Zn:Se ratio in the substrate is nearly 1:1. These regions are demarcated on the top x axis for all the three samples.

FIG. 7.

STEM-EDS line scans of (a) untreated, (b) UV-O3 pretreated, and (c) PEALD Al2O3 films. The line scans correspond to normalized counts for O (black), Se (blue), Al (red), and Zn (green). The approximate location of the film and the ZnSe substrate are shown on the top, x axis, whereas the bottom x axis is distance in nm.

FIG. 7.

STEM-EDS line scans of (a) untreated, (b) UV-O3 pretreated, and (c) PEALD Al2O3 films. The line scans correspond to normalized counts for O (black), Se (blue), Al (red), and Zn (green). The approximate location of the film and the ZnSe substrate are shown on the top, x axis, whereas the bottom x axis is distance in nm.

Close modal

For the untreated sample shown in Fig. 7(a), Zn and Se decrease monotonically with Zn extending slightly deeper into Al2O3. The Zn signal extends 3 nm inside the oxide (8 nm ≤ x ≤ 11 nm). For the UV-O3 treated sample shown in Fig. 7(b), the interface is different. The Zn and Se signals extend inside the film, while the interface (12 nm ≤ x ≤ 16 nm) appears to have a distinct layer that is Se depleted, i.e., Zn rich. It is also noted that the O signal rises briefly in this region. For the PEALD sample shown in Fig. 7(c), the interface (16 nm ≤ x ≤ 18 nm) appears to be Zn-rich. The Zn region extends 4 nm inside the film (12 nm ≤ x ≤ 16 nm).

The trend for Zn to extend deeper into the Al2O3 film compared to Se can be explained by the high volatility of SeO2. For the case of the untreated sample, the amount of Se lost depends on the duration of the incubation cycles in which H2O is oxidizing the ZnSe surface, converting Zn to ZnO and Se to SeO2. SeO2 can be subsequently volatilized. Here, we note that SeO2 has a room temperature vapor pressure of 2 × 10−6 Torr.39 Since the deposition temperature for ALD is 175 °C and the working pressure in the ALD chamber is 120 × 10−3 Torr, the SeO2 vapor pressure is expected to be much higher than the room temperature reported value and comparable to the ALD working pressure. For the case of the UV-O3 pretreated sample, there is significant SeO2 formation on the surface prior to the start of the ALD process. This SeO2 will volatilize leaving a Zn-rich, oxide region that is distinctly seen in Fig. 7(b) for 12 nm ≤ x ≤ 16 nm. Additionally, the Zn and Se signal between 8 nm ≤ x ≤ 12 nm attenuates in the same way as shown in Fig. 7(a) from 8 nm ≤ x ≤ 11 nm since the H2O based ALD chemistry remains unchanged. Thus, we hypothesize that the ALD process causes the Zn and trailing Se tails to be present in the oxide as evidence of the Al2O3 nucleation process. For the PEALD process, the SeO2 formation is more aggressive in the presence of the O2 plasma. Thus, the loss of Se in Figs. 7(c), 16 nm ≤ x ≤ 18 nm, is more severe. This leads to faster nucleation of Al2O3. The Al2O3 protective layer formed above limits subsequent oxidation of ZnSe. As a result, the interface is only 2 nm thick.

In conclusion, we have demonstrated that ALD Al2O3 on ZnSe depends on the state and degree of oxidation of the ZnSe surface. An untreated ZnSe substrate requires an incubation of 27 cycles before ALD Al2O3 can be grown on the surface. Pretreating the ZnSe substrate with a UV-O3 exposure reduces the incubation to 17 ALD cycles, while switching to a PEALD process lowers the incubation period to 13 ALD cycles. Extensive characterization of the ZnSe surface and the Al2O3/ZnSe interface reveals oxidation of the Zn, but more importantly, Se upon UV-O3 pretreatment. This interface layer is measured to be ∼4.3 nm in width. Furthermore, Se in the UV-O3 pretreated sample appears to be highly diffusive in nature and migrates to the top of the Al2O3 surface. On the other hand, the PEALD process results in a thinner interface of ∼2.5 nm. In both cases, the interface is left Zn-rich in nature, revealing the volatile nature of the Se when oxidized. These findings will aid in the understanding of developing reliable and robust coatings on ZnSe substrates for optoelectronics devices.

C.F. was funded through the NSF (Award No. 1908167). The authors acknowledge the NSF MRI: ECCS: 1726636 for XPS and the MCF-AMPAC facility at the University of Central Florida. The authors (K.A.R., J.S., A.K., and M.C.) acknowledge the financial support of AFOSR through Grant No. FA9550-19-1-0127. The work of M.C. was partially supported by the University of Central Florida under the Pre-eminent Postdoctoral Program (P3).

The authors have no conflicts to disclose.

Corbin Feit: Data curation (lead); Formal analysis (lead); Investigation (equal); Methodology (equal); Writing – original draft (lead); Writing – review & editing (equal). Jaynlynn Sosa: Investigation (supporting); Writing – review & editing (supporting). Alexandros Kostogiannes: Investigation (supporting); Writing – review & editing (supporting). Matthieu Chazot: Formal analysis (supporting); Writing – review & editing (supporting). Nicholas G. Rudawski: Investigation (supporting); Writing – review & editing (supporting). Titel Jurca: Conceptualization (supporting); Funding acquisition (equal); Resources (equal); Supervision (supporting); Visualization (supporting); Writing – review & editing (equal). Kathleen A. Richardson: Conceptualization (supporting); Funding acquisition (equal); Resources (equal); Supervision (supporting); Validation (equal); Visualization (supporting); Writing – review & editing (supporting). Parag Banerjee: Conceptualization (equal); Funding acquisition (equal); Resources (equal); Supervision (equal); Validation (equal); Visualization (supporting); Writing – review & editing (equal).

The data that support the findings of this study are available from the corresponding author upon reasonable request.

1.
H.
Morkoç
,
S.
Strite
,
G. B.
Gao
,
M. E.
Lin
,
B.
Sverdlov
, and
M.
Burns
,
J. Appl. Phys.
76
,
1363
(
1994
).
2.
G. V.
Astakhov
 et al,
Phys. Rev. B
65
,
165335
(
2002
).
3.
E.
Monroy
,
F.
Omnès
, and
F.
Calle
,
Semicond. Sci. Technol.
18
,
R33
(
2003
).
4.
R. M.
Park
,
M. B.
Troffer
,
C. M.
Rouleau
,
J. M.
Depuydt
, and
M. A.
Haase
,
Appl. Phys. Lett.
57
,
2127
(
1990
).
5.
I.
Yonenaga
,
Phys. B: Condens. Matter
308–310
,
1150
(
2001
).
6.
I.
Yonenaga
,
K.
Watanabe
,
S.
Itoh
, and
S.
Fujiwara
,
J. Mater. Sci.
41
,
2601
(
2006
).
7.
C.
Walker
,
J.
DePuydt
,
M.
Haase
,
J.
Qiu
, and
H.
Cheng
,
Phys. B: Condens. Matter
185
,
27
(
1993
).
8.
A.
Nurmikko
,
presented at the 1992 International Conference on Solid State Devices and Materials, Tsukuba, Japan, 26–28 August 1992
(unpublished).
9.
M.
Godlewski
,
M.
Skrobot
,
E.
Guziewicz
, and
M. R.
Phillips
,
J. Lumin.
125
,
85
(
2007
).
10.
L. A.
Kolodziejski
,
R. L.
Gunshor
, and
A. V.
Nurmikko
,
Annu. Rev. Mater. Sci.
25
,
711
(
1995
).
11.
H.
Jeon
,
J.
Ding
,
W.
Patterson
,
A. V.
Nurmikko
,
W.
Xie
,
D. C.
Grillo
,
M.
Kobayashi
, and
R. L.
Gunshor
,
Appl. Phys. Lett.
59
,
3619
(
1991
).
12.
S. B.
Mirov
,
V. V.
Fedorov
,
D.
Martyshkin
,
I. S.
Moskalev
,
M.
Mirov
, and
S.
Vasilyev
,
IEEE J. Sel. Top. Quantum Electron.
21
,
292
(
2015
).
13.
M.
Chazot
 et al,
J. Non-Cryst. Solids
555
,
120619
(
2021
).
14.
A.
Yang
,
J.
Qiu
,
J.
Ren
,
R.
Wang
,
H.
Guo
,
Y.
Wang
,
H.
Ren
,
J.
Zhang
, and
Z.
Yang
,
J. Non-Cryst. Solids
508
,
21
(
2019
).
15.
E. V.
Karaksina
,
L. A.
Ketkova
,
M. F.
Churbanov
, and
E. M.
Dianov
,
Inorg. Mater.
49
,
223
(
2013
).
16.
R. A.
Mironov
,
E. V.
Karaksina
,
A. O.
Zabezhailov
,
R. M.
Shapashnikov
,
M. F.
Churbanov
, and
E. M.
Dianov
,
Quantum Electron.
40
,
828
(
2010
).
17.
D.
Martyshkin
,
J.
Goldstein
,
V.
Fedorov
, and
S.
Mirov
,
Opt. Lett.
36
,
1530
(
2011
).
18.
E. V.
Karaksina
,
V. S.
Shiryaev
, and
L. A.
Ketkova
,
J. Non-Cryst. Solids
377
,
220
(
2013
).
19.
J.
Cook
 et al,
Opt. Mater. Express
12
,
1555
(
2022
).
20.
V. R.
Rai
,
V.
Vandalon
, and
S.
Agarwal
,
Langmuir
28
,
350
(
2012
).
21.
M.
Chazot
 et al,
J. Non-Cryst. Solids
576,
121259
(
2021
).
22.
L. A.
Giannuzzi
and
F. A.
Stevie
,
Micron
30
,
197
(
1999
).
23.
L. A.
Giannuzzi
,
J. L.
Drown
,
S. R.
Brown
,
R. B.
Irwin
, and
F. A.
Stevie
,
Microsc. Res. Tech.
41
,
285
(
1998
).
24.
M.
Schaffer
,
B.
Schaffer
, and
Q.
Ramasse
,
Ultramicroscopy
114
,
62
(
2012
).
25.
B.
Kempshall
,
L.
Giannuzzi
,
B.
Prenitzer
,
F.
Stevie
, and
S.
Da
,
J. Vac. Sci. Technol. B
20
,
286
(
2002
).
26.
J. L.
Van Hemmen
,
S. B. S.
Heil
,
J. H.
Klootwijk
,
F.
Roozeboom
,
C. J.
Hodson
,
M. C. M.
Van De Sanden
, and
W. M. M.
Kessels
,
J. Electrochem. Soc.
154
,
G165
(
2007
).
27.
V.
Mittal
,
N. P.
Sessions
,
J. S.
Wilkinson
, and
G. S.
Murugan
,
Opt. Mater. Express
7
,
712
(
2017
).
28.
J. R.
Shallenberger
and
N.
Hellgren
,
Surf. Sci. Spectra
27
,
014020
(
2020
).
29.
N.
Hellgren
,
M. A.
Steves
,
J.
Shallenberger
,
S. K.
O’Boyle
,
E.
Mellott
, and
A. R.
Noble
,
Appl. Surf. Sci.
528
,
146604
(
2020
).
30.
J. F.
Moulder
,
W. F.
Stickle
,
P. E.
Sobol
, and
K. D.
Bomben
,
Phys. Electron
464230
(
1995
).
31.
A. M.
Chaparro
,
C.
Maffiotte
,
M. T.
Gutiérrez
, and
J.
Herrero
,
Thin Solid Films
358
,
22
(
2000
).
32.
J. D.
McCullough
,
J. Am. Chem. Soc.
59
,
789
(
1937
).
33.
H.
Nohira
 et al,
J. Non-Cryst. Solids
303
,
83
(
2002
).
34.
I.
Iatsunskyi
,
M.
Kempiński
,
M.
Jancelewicz
,
K.
Załęski
,
S.
Jurga
, and
V.
Smyntyna
,
Vacuum
113
,
52
(
2015
).
35.
W.
Chen
,
A.
Kahn
,
P.
Soukiassian
,
P.
Mangat
,
J.
Gaines
,
C.
Ponzoni
, and
D.
Olego
,
Phys. Rev. B
51
,
14265
(
1995
).
36.
A.
Olin
,
G.
Nolang
,
E.
Osadchii
,
L.
Ohman
, and
E.
Rosen
, Elsevier Sci. (
2005
).
37.
A.
Chaparro
,
C.
Maffiotte
,
J.
Herrero
, and
M.
Gutiérrez
,
Surf. Interface Anal.
30
,
522
(
2000
).
38.
R. F.
Brebrick
,
J. Phase Equilib.
21
,
235
(
2000
).
39.
O.
Kubaschewski
,
Int. Ser. Mater. Sci. Technol.
24
,
478
(
1977
).