Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2019 Oct 15.
Published in final edited form as: Appl Surf Sci. 2018 May 24;455:438–445. doi: 10.1016/j.apsusc.2018.05.182

Surface Chemistry of Thermal Dry Etching of Cobalt Thin Films Using Hexafluoroacetylacetone (hfacH)

Jing Zhao 1,&, Mahsa Konh 1,&, Andrew Teplyakov 1,*
PMCID: PMC6013264  NIHMSID: NIHMS973590  PMID: 29937610

Abstract

Amechanism of thermal dry etching process of cobalt thin films by using 1,1,1,5,5,5-hexafluoro-2,4-pentanedione (hexafluoroacetylacetone, hfacH) was investigated. This process, relevant to atomic layer etching (ALE) technology directed towards oxidized cobalt films, requires adsorption of molecular organic precursor, such as hfacH, at moderate temperatures and is often thought of as releasing water and Co(hfac)2 at elevated temperatures. The reaction was analyzed in situ by temperature-programmed desorption (TPD) and the resulting surface was investigated ex situ by X-ray photoelectron spectroscopy (XPS). The changes in surface morphology during the process were monitored by atomic force microscopy (AFM) and scanning electron microscopy (SEM). The removal of Co(hfac)2 from the surface was observed above 650 K, a temperature well above commercially desired etching conditions, suggesting that the thermal etching process is more complex than originally envisioned. In addition, the upper limit of thermal treatment is established at 800 K, as the microscopic techniques clearly indicated surface morphology changes above this temperature. In addition, the structure of the surface at the nanoscale is observed to be affected by the presence of surface bound organic ligands even at room temperature. Thus, further mechanistic studies should address the kinetic regime and surface morphology to make inroads into mechanistic understanding of the dry etching process.

Keywords: Dry etching, cobalt thin films, reaction mechanism

1. Introduction

Magnetic random access memory (MRAM) devices are the foundation of the emerging energy-efficient technologies requiring further miniaturization of the desired microelectronic components [1]. MRAMs are generally made of two magnetic metal layers separated by a magnetic tunnel junction [2–4]. The magnetic metal layers are commonly made of Fe, Co, Ni. In the recent years, an increased interest in Co-based thin films started to develop not only because of the magnetic properties of this metal [5], but also because of its high magnetoresistance [6]. With the development of microelectronic devices with smaller and smaller dimensions over the past decades, to retain the high magnetoresistance, the thickness of magnetic metal layers, for example cobalt thin films, is required to be controlled properly as well. Traditionally, cobalt thin films are deposited by atomic layer deposition (ALD) approach [7–12]. Although some ALD approaches to cobalt deposition can be quite successful, often they are hampered by slow nucleation rate and increased roughness of the resulting surface [13]. Therefore, more and more studies began to focus on controlling the thickness of cobalt films starting with a thick layer by etching methods. Typically, as the first step, halogens (such as Cl2) or oxygen are used to produce metal halides or metal oxides directly or with a help of plasma-based methods. In the following step, organic molecules are introduced as co-reactants to form organometallic products, which can be removed by proper additional stimuli, such as sputtering and thermal processing. The development of this etching approach led to establishing the processing for highly-controlled metal removal by atomic layer etching (ALE), where typically several atomic layers are removed in every treatment cycle [14,15]. More importantly, ALE can provide smooth and uniform surfaces compared, for example, with plasma etching methodologies [16].

Previously, an efficient ALE method has been demonstrated by applying halogen and hydrogen plasma to etch CoFe alloy films [17]. It was discovered that the combination of Cl2 and H2 plasma can etch CoFe alloy at a reasonable rate (up to 4.2 nm/min) and can preserve its magnetic properties. However, this method required plasma instrumentation and high-price pure gas chemicals. More recently, it was reported that instead of using exclusively plasma methods, the combination of Cl2 or O2 plasma and organic vapor deposition are also efficient in etching Co and Fe thin films [18,19]. Halogen and oxygen can initially modify the metal surfaces. Once the oxidized metal surfaces are exposed to organic vapor, organometallic bonds can be formed between oxidized metal atoms and organic chemicals. Then, the removal of the produced organometallic entities from surfaces may occur spontaneously or with additional thermal treatment or light sputtering.

A β-diketone is commonly used as the organic vapor to be exposed onto oxidized metal surfaces because of its high reactivity and affinity to metals [20,21]. Among various types of β-diketones, acetylacetone (ACAC) and hexafluoroacetylacetone (hfacH) are the most widely used molecules because of high volatility and reactivity and because of the formation of stable metal diketonates. For instance, it has been reported that Co, Fe, Ni can be etched by hfacH in supercritical CO2 [22]. In our study, we also chose hfacH as the organic vapor etchant for Co thin films. In addition to the attractive physical properties and chemical reactivity of hfacH, the presence of fluorine in the ligand formed on a surface can be used as a spectroscopic probe of surface chemical reactions to help us identify the mechanism of the etching process.

Since the target of the current investigation is the half-cycle of the ALE process, instead of utilizing plasma methods for surface oxidation, as reported previously [17,19], this study simply utilized cobalt films whose surface are covered by oxide, as the starting point. The overall mechanisms of ALE can be very complicated and they can certainly depend on the surface morphology, cleanliness, and etching conditions. It is indeed possible that plasma-treatments or sputtering steps can alter surface morphology differently compared to thermally induced metal removal demonstrated in the present work. Nevertheless, similar starting point in all those methods is at least a partially oxidized metal surface. The goal of this work is to place the initial thermal boundaries for the ALE processing of cobalt films, to confirm the role of surface oxide in etching mechanism, and to follow changes in surface morphology during initial etching steps.

The cobalt films were investigated in vacuum by exposing them to hfacH vapor and then following the temperature-dependent changes spectroscopically and microscopically. This approach is sufficient to assess the chemistry of a half-cycle for the dry etching of cobalt films. It also allows us to carefully investigate the effects of dosing conditions and thermal treatment onto the film, specifically targeting the reactions of organic ligands with oxidized surfaces to assess the chemical and morphological changes during ALE processing.

2. Experimental and Computational Methods

The cobalt film samples utilized in this work are cut from a standard 300 mm single-side polished silicon wafers (Advantiv). The samples were prepared by standard physical vapor deposition of cobalt (300 nm) onto SiOx covered silicon wafer (500 nm) using tantalum as an adhesion layer (50 nm). For qualitative infrared spectroscopy studies, the thickness of the cobalt layer was 50 nm. The thickness of the films was verified by scanning electron microscopy (not shown). A separate set of control studies used sputtering depth profiling combined with X-ray photoelectron spectroscopy to demonstrate that the thickness of cobalt oxide layer on top of the samples is less than 3 nm (not shown). Once mounted on a sample holder in vacuum chambers, the samples were prepared by annealing to 440 K for 40 min in order to remove physisorbed surface contaminants. 1,1,1,5,5,5-hexafluoro-2,4-pentanedione (hfacH) (98+%, Alfa Aesar) was introduced into the chambers using leak valves, following a freeze-pump-thaw procedure. The purity of the compound was confirmed in situ by mass spectroscopy. The ultra-high vacuum (UHV) chamber used for TPD studies has a 10−9 Torr base pressure. It is equipped with a differentially pumped mass spectrometer (Hiden Analytical) with the detection range from 0-510 amu that was used for thermal desorption investigation. After the desired dose of hfacH was introduced, the temperature was increased linearly at a rate of approximately 2 K/s, controlled by a temperature controller (Eurotherm, Model 818). Only one sample was mounted and investigated with TPD at a time, as the highest temperature reached in the temperature sweep was 800 K. Lower temperatures were set as the upper limits for selected samples, as indicated below. Following the single set of TPD experiments, the sample was removed from the chamber and the new sample was mounted for further studies. The high vacuum chamber utilized to prepare samples for ex-situ spectroscopy and microscopy investigations has a 10−5 Torr base pressure. It is also equipped with an infrared spectrometer (Nicolet Magna 560 with an MCT (mercury-cadmium-telluride) detector) set up in a transmission mode that was used to confirm the dosing of hfacH. For the experiments performed in this chamber, all the dosing of the hfacH was performed at the same conditions. The dosing pressure was 2 Torr, and a dosing time of 5 min. This dose is sufficient to produce a fully saturated surface as confirmed by the infrared signal recorded as a function of exposure. Following the exposure, the sample was briefly annealed to the desired temperatures and cooled down to room temperature. It should be noted that even though only 50 nm cobalt layer samples were used for the infrared studies, the use of infrared spectroscopy to reliably and quantitatively investigate surface species on Co-covered side is limited by the size and type (single-side polished) of the samples utilized here. The obtained signal-to-noise ratio was not sufficient to obtain quantitatively reproducible data. The vibrational spectroscopy studies of thinner films deposited on double-side polished wafers will be the subject of further work.

XPS studies were performed on a K-Alpha+ XPS System from Thermo Scientific in the surface facility at the University of Delaware using Al K-Alpha X-ray source (hv = 1486.6 eV) at a 35.3° takeoff angle with respect to the analyzer. The high-resolution spectra were collected over the range of 20 at 0.1 eV/ step with the pass energy of 20 eV. The survey spectra were collected over the energy range of 0–1000 eV. CasaXPS (version 2.3.16) software was utilized to analyze all the raw data [23]. The carbon peak at 284.6 eV [24–26] was used to calibrate the XPS scale.

The AFM images shown in the studies were collected with a Nanoscope V controller (Veeco Multimode SPM) in the Interdisciplinary Science and Engineering Laboratory (ISE) Facility lab at the University of Delaware with tapping mode at 512 lines per scan and Aluminum-coated silicon nitride tips resonating at 300 kHz (Budget Sensors).

A Zeiss Auriga 60 scanning electron microscope (SEM) with an accelerating voltage of 3 kV was used in electron microscopy investigations. All images were collected with a secondary electron (in-lens) detector at a working distance of 5.0 mm.

The DFT investigations presented in this work were performed using Gaussian 09 suite of programs [27] and the graphical interface GaussView 5. The proposed structures of selected molecular models (Co(hfac)2 and CoF2) were predicted by B3LYP/LANL2DZ [28–32] with an intent of identifying representative F 1s binding energies in the experimental spectra by applying calibrating procedures based on Koopmans’ theorem, similarly to the procedures for identifying N- and C-containing species described previously [33–35].

3. Results and Discussion

3.1. Insight into the mechanism of cobalt etching based on thermal desorption investigations

Scheme 1 presents a generalized mechanism proposed for dry etching of cobalt starting with oxidized surface. Figure 1 shows the summary of the results of thermal desorption investigation of the cobalt thin films following hfacH adsorption at room temperature. Even though there are a total of 22 radioisotopes of Co that can be characterized, the only stable cobalt isotope is 59Co. According to the proposed thermal etching process for cobalt thin films shown in Scheme 1, we would expect to observe the desorption of Co(hfac)2, as a product of thermal dry etching. Therefore, we monitored the representative fragments of a cracking pattern for Co(hfac)2: m/z=59, 207, 266 and 473. As shown in Figure 1, Co(hfac)2 desorption represented in this figure by m/z=473 occurs at approximately 625 K, supporting the hypothesis of Co(hfac)2 formation as a result of a reaction with hfacH. If the formation of the hfacH as a result of temperature-programmed reaction is investigated, in addition to the sharp peak at 625 K (from recombination processes involving Co(hfac)2), a broad feature centered at approximately 720 K is observed, as shown in Figure 1 by a representative m/z = 207 trace corresponding to the desorption of hfacH. This assignment is supported by other traces in the hfacH cracking pattern recorded for m/z=91, 119, and 208. It should be pointed out that at temperatures above 750 K, the desorption features observed for hfacH could potentially be influenced by the contribution from the background desorption of hfacH (from heater and manipulator parts); however, as will be demonstrated further, this is not the case.

Scheme 1.

Scheme 1

Open in a new tab

Schematic illustration of cobalt thermal dry etching after saturated with hfacH at room temperature.

Figure 1.

Figure 1

Open in a new tab

Summary of temperature-programmed desorption studies of hfacH adsorption on cobalt thin films at room temperature. The representative traces illustrating evolution processes for (a) H2O, (c) hfacH, (e) Co(hfac)2 from the cobalt film surfaces prepared by annealing in vacuum (oxidized surface) are compared to the (d) trace for Co(hfac)2 and (b) trace for hfacH evolution from the cobalt sample sputtered with Ar+ to remove surface impurities (carbon and oxygen) and annealed at 440 K for 40 min.

The overall schematic illustration of cobalt etching process following hfacH adsorption and thermal treatment shown in Scheme 1 suggests the formation of water during thermal etching, and the TPD results in Figure 1 tracking the corresponding mass-to-charge ratio (m/z=18) do show the evolution of water from the surface that starts nearly concurrently with hfacH and extends at higher temperatures. Based on previous studies of hfacH reacting with metal oxides, some of the remaining organic surface species can be very stable and could be described as ketene-like structures, as was shown by infrared investigations and computational studies of ZnO chemistry [36]. Thus, as a result of thermal treatment, Co(hfac)2 is formed and desorbs around 625 K, accompanied by the formation and removal of H2O. This observation is consistent with etching mechanism reported previously for ZnO etching [37] and nickel etching [38], where the formation of partially oxidized metal is a pre-requisite for the removal of metal-containing surface products. In order to examine if the cobalt surface without oxygen could react with hfacH in a similar fashion to the process shown in Scheme 1, the evolution of Co(hfac)2 was investigated on a cobalt film surface pre-cleaned by argon sputtering and brief annealing at 440 K for 40 min to remove surface oxygen as confirmed in situ by Auger electron spectroscopy. In this case, no desorption of Co(hfac)2 is recorded, as shown in Figure 1. This observation confirms that oxidized metal surface is required as a starting point for dry etching process summarized in Scheme 1, and that pristine metal surface exposed to hfacH will only lead to the decomposition of surface-bound organic species when heated. The last statement was also supported by the observation of the −CF3 containing products of surface decomposition of adsorbed hfacH at temperatures above 600 K in thermal desorption spectra (not shown), although the precise nature of these species was not determined. Based on the absence of m/z = 207 and m/z = 473 signals in the thermal desorption spectra following hfacH reaction with a clean (pre-sputtered) cobalt surface, it can be concluded that the background desorption (from heater or manipulator parts) of any species with these mass-to-charge ratios does not affect the results of thermal desorption studies of the oxidized cobalt surfaces reacting with the same exposures of hfacH.

3.2. Surface chemical changes and morphology transformations following hfacH precursor exposure and elevated temperature

The changes in cobalt chemical state, based on Co 2p spectral range, as well as F 1s and C 1s spectra are shown in Figures 2–4. According to the Co 2p spectrum shown in Figure 2a, there is only Co(II) oxidation state, 780.7 eV for Co 2p3/2 [39] present on the surface of cobalt thin films before chemical treatment. However, when hfacH is dosed at room temperature, a lower binding energy peak at 777.9 eV increases in intensity. At the same time, the Co(II) peak does not decrease substantially. This implies that some of the surface Co(II) may be reduced to Co(0) by hfacH. After the surface is annealed to 800 K, a dramatic oxidation state change of Co was observed, as demonstrated in Co 2p spectrum in Figure 2c. Co(0) peak is dominant after the hfacH-exposed cobalt surface is heated to 800 K. According to the proposed mechanism discussed above, hfacH is expected to saturate the top layer of partially oxidized cobalt and react with cobalt (II) to form Co-bound surface species. Above 650 K, some of these species are removed from the surface as Co(hfac)2. This fully explains why following high temperature treatment, more Co(0) is observed by XPS and Co(II) peak almost disappears. Interestingly, the X-ray photoelectron spectrum of the oxidized cobalt film exposed to hfacH and annealed to 800 K is very similar to that obtained following hfacH exposure onto a cobalt film pre-cleaned by Ar+ sputtering and annealing at 440 K for 40 min. In both cases, the dominant spectral feature indicates the presence of Co(0) within few nanometers of the surface. In other words, clean cobalt surface is stabilized by exposure to hfacH with respect to oxidation and a very similar surface is obtained by exposure of oxidized cobalt film to hfacH followed by annealing to 800 K. The summary of F 1s spectral region shown in Figure 3 indicates that the peak assigned to C-F is dominant at room temperature after hfacH is dosed onto an oxidized cobalt film and decreases following surface annealing to 800 K. Another F 1s peak at approximately 685 eV appears at elevated temperature. The position of this peak suggests that the decomposition of hfacH or Co(hfac)2 will cause the formation of CoFx species [40]. The assignment of the two observed features to −CFx and CoFx species is supported by the DFT calculations performed for Co(hfac)2 and CoF2 structures. The computationally predicted positions of these features based on Koopmans theorem aligned with the experimentally observed −CF3 peak are shown in Figure 3 as solid bars underneath experimental data. The intensity of the overall F 1s signal decreases only by about 10% following annealing, suggesting that the TPD studies followed the desorption of Co(hfac)2 only as a minor channel for surface transformations at elevated temperatures. It is important to note that the formation of CoFx species confirms that the high temperature required for thermal desorption of Co(hfac)2 yields partial decomposition of the surface-bound hfac ligands. One more point that has to be noted based on the XPS analysis is that if the cobalt surface is pre-cleaned by sputtering and annealing to remove surface oxygen in a UHV chamber and then exposed to hfacH at room temperature, the second feature corresponding to the formation at room of CoFx immediately species is observed at approximately 685 eV temperature, without additional thermal treatments. This observation is consistent with the hypothesis that adsorption of hfacH on oxidized cobalt surface is required for the formation of Co(hfac)2 at elevated temperature. hfacH adsorption onto a clean cobalt surface does result chemisorption; however, the only surface reaction pathway for hfac on a clean cobalt surface is decomposition. In C 1s spectra summarized in Figure 4, when hfacH adsorbs on cobalt surfaces at room temperature, the binding energies corresponding to CF3 (292.6 eV), C=O (288.5 eV) and C-H (284.6 eV) [41] were observed. Unfortunately, since the XPS studies are conducted following by a brief exposure of the samples to ambient conditions upon transfer, quantification or correct assignment of all the observed features is difficult because of possible contamination. Nevertheless, as expected, after the after surface was annealed to 800K, C 1s the intensities of the features at binding energies corresponding to CF3 and C=O decreased.

Figure 2.

Figure 2

Open in a new tab

Summary of high-resolution XPS investigations of Co 2p spectral region for (a) pristine (oxidized) cobalt sample; (b) the cobalt sample in (a) exposed to hfacH at room temperature; (c) the sample in (b) briefly heated to 800 K; (d) cobalt sample sputtered with Ar+ to remove surface impurities (carbon and oxygen) and annealed at 440 K for 40 min, then exposed to a saturation exposure of hfacH and briefly heated to 800 K.

Figure 4.

Figure 4

Open in a new tab

Summary of high-resolution XPS investigations of C 1s spectral region for (a) pristine (oxidized) cobalt sample; (b) the cobalt sample in (a) exposed to hfacH at room temperature; (c) the sample in (b) briefly heated to 800 K; (d) cobalt sample sputtered with Ar+ to remove surface impurities (carbon and oxygen) and annealed at 440 K for 40 min, then exposed to a saturation exposure of hfacH.

Figure 3.

Figure 3

Open in a new tab

Summary of high-resolution XPS investigations of F 1s spectral region for (a) pristine (oxidized) cobalt sample; (b) the cobalt sample in (a) exposed to hfacH at room temperature; (c) the sample in (b) briefly heated to 800 K; (d) cobalt sample sputtered with Ar+ to remove surface impurities (carbon and oxygen) and annealed at 440 K for 40 min, then exposed to a saturation exposure of hfacH. The computationally predicted results for models indicated are shown as vertical bars underneath experimental data.

Besides the chemical state changes observed by XPS, we also investigated the surface morphology transformations upon hfacH precursor exposure and thermal treatment by examining the corresponding SEM and AFM images. The initial oxidized cobalt surface is shown in Figure 5(a). When this surface is exposed to hfacH at room temperature, it appears rougher, as demonstrated in Figure 5(b); however, quantitative assessment of this change is difficult based solely on the SEM spectra. If the temperature of the sample if increased to 650 K, no additional morphological changes are observed (as shown in Figure S-2 in the Supporting Information section). However, after the surface is annealed to 800 K, even following room-temperature hfacH dose, some changes in surface morphology are observed, as shown in Figure 5(c). Specifically, large portions of the surface appear to be much smoother than that following hfacH dose. Based on these observations, it can be concluded that the exposure of the oxidized cobalt surface to hfacH changes surface morphology even at room temperature by increasing roughness and that annealing the surface to 800 K results in substantially smoother surface. In order to test whether the surface morphological changes resulted only from the high temperature annealing or from the combination of hfacH exposure and thermal treatment, we annealed the original cobalt surface (without exposure to hfacH) to the same temperature of 800 K. The resulting image is shown in Figure 5(d) for reference. Annealing of the cobalt surface to 800 K turned out to cause minor surface morphological changes as well, which implies that at the temperature of 800 K, the morphology of the surface starts to change and heating the samples above this temperature should be avoided. However, the surface in image 5(c) appears much smoother than that in image 5(d).

Figure 5.

Figure 5

Open in a new tab

SEM images of cobalt thin films: (a) starting cobalt thin films; (b) dosing hfacH onto the cobalt surfaces at room temperature; (c) annealing the cobalt sample exposed to hfacH at room temperature to 800 K; (d) annealing the clean (no hfacH exposure) cobalt sample to 800K.

In order to assess the morphological changes more quantitatively, AFM images were also recorded to be compared to SEM experiments, as summarized in Figure 6. This set of studies demonstrates a change of surface structure following a saturation exposure of the cobalt surface to hfacH at room temperature, consistent with the observations based on SEM studies. The image in Figure 6(b) has an RMS (root mean square) roughness of 3.98 nm, which is substantially larger than the RMS roughness of the starting cobalt sample, 1.56 nm. More importantly, we observed the evidence of surface nanostructuring, which is consistent with the hypothesis that the formation of surface-bound organic species affects the positions of Co atoms, as the Co(hfac)2 is ultimately produced. With the surface annealed to 800 K, a noticeably smoother surface, compared to the surface exposed to hfacH at room temperature, is observed, which is again consistent with the SEM investigation described above. Much less of nanostructuring is recorded in image 6(c), recorded following brief annealing to 800 K, although a number of pits are starting to form at this temperature. Overall, this observation is consistent with the proposed mechanism of dry etching based on Co(hfac)2 removal at elevated temperatures. The hfacH adsorption roughens the surface up but then Co(hfac)2 removal, likely targeting predominantly adatoms, defect sites, and other thermodynamically less stable Co positions, makes the surface smooth again.

Figure 6.

Figure 6

Open in a new tab

AFM images of cobalt thin films: (a) pristine (oxidized) cobalt thin films; (b) dosing hfacH onto a cobalt film surface in (a) at room temperature; (c) annealing the cobalt sample exposed to hfacH at room temperature in (b) to 800 K.

3.3. The observation of Co (II) reduction to Co (0) based on XPS investigation at different hfacH exposure temperatures

According to the results discussed in section 3.1, the actual etching will not occur until approximately 650 K after the exposure of cobalt films to hfacH at room temperature. In addition, we observed a Co oxidation state change and surface morphology change after the room temperature hfacH exposure. These changes imply that even though there is no species leaving the surface at room temperature, a reaction has indeed occurred and caused substantial alteration of surface properties. It is reasonable to assume that some hydrogen atoms coming from hfacH may react with oxygen on cobalt at room temperature (and even produce water directly upon adsorption), however, the temperature is not sufficiently high to produce the bidentate hfac ligands on selected cobalt atoms or to yield sufficient number of hydrogen atoms to reduce all the oxidized cobalt. In other words, there is an increase in surface concentration of Co (0) but Co (II) is still dominant. Assuming that the reaction may be driven kinetically, we exposed the starting cobalt surface to hfacH at different substrate temperatures, staying below the onset of decomposition of fluorine-containing surface species. As shown in Figure 7 (spectra a-d), with higher reaction temperature, more Co (II) can be reduced to metallic Co. F 1s position does not change and a single feature corresponding to −CFx species remains within the temperature range tested. Thus, at 360 K and even at 440 K, no hfac is desorbed as a part of hfacH or any other hfac-containing species and no ligand decomposition is observed. It does appear that the coverage of hfac increases at higher temperature, which is signified by the increased intensity of the F 1s feature and also by the clear increase of the corresponding O 1s feature, as shown in Figure 7. Within the O 1s spectral range in Figure 7, the peak at approximately 531.5 eV can be assigned the oxygen in oxidized Co samples, as the same peak is observed in a starting cobalt films, before any chemical treatment is conducted, as confirmed by spectrum (a). Although the intensity analysis for the other feature observed at approximately 529.5 eV may be complicated by the fact that the surfaces are exposed to ambient upon transfer to the XPS setup, the intensity of this feature clearly increases substantially upon temperature increase, consistent with the hypothesis that the overall process may be driven kinetically. Finally, if the hfacH exposure is conducted at 650 K, as summarized in spectra (e) of Figure 7, a large number of different processes may occur. Co 2p spectral region recorded following hfacH exposure at 650 K shows much more pronounced surface oxidation than at any other temperature investigated. O 1s spectral region confirms the formation of different O-containing surface species indicative of hfac entity decomposition or desorption. The F 1s and C 1s spectral regions confirm the decomposition of hfacH observed if the dosing temperature is that high.

Figure 7.

Figure 7

Open in a new tab

XPS spectra of Co 2p, F 1s, C 1s and O 1s for: (a) pristine (oxidized) cobalt thin films; (b) hfacH adsorption at room temperature; (c) hfacH adsorption at 360 K; (d) hfacH adsorption at 440 K, and (e) hfacH adsorption at 650 K.

4. Conclusions

Based on the investigation of the cobalt thermal dry etching with spectroscopic and microscopic analytical techniques supplemented with DFT calculations, the reaction mechanism is proposed to consist of the following steps: (1) hfacH will adsorb on oxidized cobalt surfaces at mild deposition temperatures and donate hydrogen to reduce surface cobalt and to produce water as a side product; (2) once surface-bound hfac species are formed, selected surface cobalt atoms become less thermodynamically stable and are removed at elevated temperatures as Co(hfac)2, thus constituting the actual etching of cobalt surfaces. Cobalt removal is a minor surface reaction pathway at the conditions studied, and it requires the temperatures substantially higher than those required for commercial processing. The cobalt surface has to be oxidized for the Co(hfac)2 to be formed. Otherwise, the main surface reaction pathway for the adsorbed hfac is decomposition. Microscopy studies suggest that hfacH adsorption causes increased surface roughening; however, Co(hfac)2 removal above 650 K does result in a smoother surface. Above 800 K, irreversible morphological changes occur, suggesting that this temperature is the upper limit for any chemical treatment resulting in cobalt dry etching. In fact, the upper limit has to be placed at 650 K, since at this temperature very substantial decomposition of surface hfac ligands is observed.

Further work is necessary to determine the species desorbing from the surface in realistic etching conditions used in industry. This will likely require overcoming kinetic barriers at substantially higher pressures than used in vacuum studies.

Supplementary Material

supplement

Acknowledgments

This work was partially supported by the National Science Foundation (DMR1609973 (GOALI)). AVT acknowledges the support of NSF (9724307; 1428149) and the NIH NIGMS COBRE program (P30-GM110758) for research activities in the University of Delaware Surface Analysis Facility and W. M. Keck Center for Advanced Microscopy and Microanalysis. We want to thank Ms. Mackenzie Williams (University of Delaware) for her help with AFM and SEM investigations used in this work.

Footnotes

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

References

  • 1.Senni S, Torres L, Sassatelli G, Gamatie A, Mussard B. Exploring MRAM Technologies for Energy Efficient Systems-On-Chip. IEEE J Emerg Sel Top Circuits Syst. 2016;6:279–292. doi: 10.1109/JETCAS.2016.2547680. [DOI] [Google Scholar]
  • 2.Tehrani S, Slaughter JM, Deherrera M, Engel BN, Rizzo ND, Salter J, Durlam M, Dave RW, Janesky J, Butcher B, Smith K, Grynkewich G. Magnetoresistive random access memory using magnetic tunnel junctions. Proc IEEE. 2003;91:703–714. doi: 10.1109/JPROC.2003.811804. [DOI] [Google Scholar]
  • 3.Pushp A, Phung T, Rettner C, Hughes BP, Yang SH, Parkin SSP. Giant thermal spin-torque–assisted magnetic tunnel junction switching. Proc Natl Acad Sci. 2015;112:6585–6590. doi: 10.1073/pnas.1507084112. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4.Khvalkovskiy AV, Apalkov D, Watts S, Chepulskii R, Beach RS, Ong A, Tang X, Driskill-Smith A, Butler WH, Visscher PB, Lottis D, Chen E, Nikitin V, Krounbi M. Erratum: Basic principles of STT-MRAM cell operation in memory arrays (Journal of Physics D: Applied Physics 2013 46 (074001)) J Phys D Appl Phys. 2013;46 doi: 10.1088/0022-3727/46/13/139601. [DOI] [Google Scholar]
  • 5.Daosheng D, Peng C, Fang R. Magnetic anisotropy and interlayer exchange coupling of evaporated Au/Co multilayers. 1992;46:12022–12025. doi: 10.1103/physrevb.46.12022. [DOI] [PubMed] [Google Scholar]
  • 6.Luciński T, Reiss G, Brückl H. In situ monitoring of GMR and M(H) evolution during Co/Cu multilayers growth. J Magn Magn Mater. 1999;193:484–487. doi: 10.1016/S0304-8853(98)00479-X. [DOI] [Google Scholar]
  • 7.Lim BS, Rahtu A, Gordon RG. Atomic layer deposition of transition metals. Nat Mater. 2003;2:749–754. doi: 10.1038/nmat1000. [DOI] [PubMed] [Google Scholar]
  • 8.Kerrigan MM, Klesko JP, Rupich SM, Dezelah CL, Kanjolia RK, Chabal YJ, Winter CH. Substrate selectivity in the low temperature atomic layer deposition of cobalt metal films from bis(1,4-di- tert -butyl-1,3-diazadienyl)cobalt and formic acid. J Chem Phys. 2017;146 doi: 10.1063/1.4968848. [DOI] [PubMed] [Google Scholar]
  • 9.Kerrigan MM, Klesko JP, Winter CH. Low Temperature, Selective Atomic Layer Deposition of Cobalt Metal Films Using Bis(1,4-di-tert-butyl-1,3-diazadienyl)cobalt and Alkylamine Precursors. Chem Mater. 2017;29:7458–7466. doi: 10.1021/acs.chemmater.7b02456. [DOI] [PubMed] [Google Scholar]
  • 10.Kwon J, Saly M, Halls MD, Kanjolia RK, Chabal YJ. Substrate Selectivity of (tBu-Allyl)Co(CO)3 during Thermal Atomic Layer Deposition of Cobalt. Chem Mater. 2012;24:1025–1030. doi: 10.1021/cm2029189. [DOI] [Google Scholar]
  • 11.Kalutarage LC, Martin PD, Heeg MJ, Winter CH. Volatile and thermally stable mid to late transition metal complexes containing α-imino alkoxide ligands, a new strongly reducing coreagent, and thermal atomic layer deposition of Ni, Co, Fe, and Cr metal films. J Am Chem Soc. 2013;135:12588–12591. doi: 10.1021/ja407014w. [DOI] [PubMed] [Google Scholar]
  • 12.Klesko JP, Kerrigan MM, Winter CH. Low Temperature Thermal Atomic Layer Deposition of Cobalt Metal Films. Chem Mater. 2016;28:700–703. doi: 10.1021/acs.chemmater.5b03504. [DOI] [Google Scholar]
  • 13.Johnson RW, Hultqvist A, Bent SF. A brief review of atomic layer deposition: From fundamentals to applications. Mater Today. 2014;17:236–246. doi: 10.1016/j.mattod.2014.04.026. [DOI] [Google Scholar]
  • 14.Kanarik KJ, Lill T, Hudson EA, Sriraman S, Tan S, Marks J, Vahedi V, Gottscho RA. Overview of atomic layer etching in the semiconductor industry. J Vac Sci Technol A. 2015;33:20802. doi: 10.1116/1.4913379. [DOI] [Google Scholar]
  • 15.Lee Y, George SM. Prospects for Thermal Atomic Layer Etching Using Sequential, Self-Limiting Fluorination and Ligand-Exchange Reactions. ACS Nano. 2016;10:4889–4894. doi: 10.1021/acs.chemmater.6b02543. [DOI] [PubMed] [Google Scholar]
  • 16.Manos D, Flamm D. Plasma Etching: An Introduction. Elsevier; 1989. [Google Scholar]
  • 17.Kim T, Kim Y, Chen JKC, Chang JP. Viable chemical approach for patterning nanoscale magnetoresistive random access memory. J Vac Sci Technol A. 2015;33:21308. doi: 10.1116/1.4904215. [DOI] [Google Scholar]
  • 18.Chen JKC, Kim T, Altieri ND, Chen E, Chang JP. Ion beam assisted organic chemical vapor etch of magnetic thin films. J Vac Sci Technol A. 2017;35:31304. doi: 10.1116/1.4978553. [DOI] [Google Scholar]
  • 19.Chen JKC, Altieri ND, Kim T, Chen E, Lill T, Shen M, Chang JP. Directional etch of magnetic and noble metals. II. Organic chemical vapor etch. J Vac Sci Technol A. 2017;35:05C305. doi: 10.1116/1.4983830. [DOI] [Google Scholar]
  • 20.Kung H, Teplyakov AV. Formation of copper nanoparticles on ZnO powder by a surface-limited reaction. J Phys Chem C. 2014;118:1990–1998. doi: 10.1021/jp409902c. [DOI] [Google Scholar]
  • 21.Yichen D, Gao F, Teplyakov AV. Role of the Deposition Precursor Molecules in Defining Oxidation State of Deposited Copper in Surface Reduction Reactions on H-Terminated Si(111) Surface. J Phys Chem C. 2015;119:27018–27027. doi: 10.1126/science.1249098.Sleep. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22.Rasadujjaman M, Nakamura Y, Watanabe M, Kondoh E, Baklanov MR. Supercritical carbon dioxide etching of transition metal (Cu, Ni, Co, Fe) thin films. Microelectron Eng. 2016;153:5–10. doi: 10.1016/j.mee.2015.12.018. [DOI] [Google Scholar]
  • 23.Casa XPS Processing Software. Casa Software Ltd; 2010. [Google Scholar]
  • 24.Gammon WJ, Kraft O, Reilly AC, Holloway BC. Experimental comparison of N(1s) X-ray photoelectron spectroscopy binding energies of hard and elastic amorphous carbon nitride films with reference organic compounds. Carbon. 2003;41:1917–1923. doi: 10.1016/S0008-6223(03)00170-2. [DOI] [Google Scholar]
  • 25.Tan X, Fan Q. Eu (III) Sorption to TiO2 (Anatase and Rutile): Batch, XPS, and EXAFS Studies. Environ Sci Technol. 2009;43:3115–3121. doi: 10.1021/es803431c. [DOI] [PubMed] [Google Scholar]
  • 26.Duan Y, Teplyakov AV. Deposition of copper from Cu(I) and Cu(II) precursors onto HOPG surface: Role of surface defects and choice of a precursor. J Chem Phys. 2017;146:52814. doi: 10.1063/1.4971287. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27.Frisch MJ, Trucks GW, Schlegel HB, Scuseria GE, Robb MA, Cheeseman JR, Scalmani G, Barone V, Petersson GA, Nakatsuji H, Li X, Caricato M, Marenich A, Bloino J, Janesko BG, Gomperts R, Mennucci B, Hratchian HP, Ortiz JV, Izmaylov AF, Sonnenberg JL, Williams-Young D, Ding F, Lipparini F, Egidi F, Goings J, Peng B, Petrone A, Henderson T, Ranasinghe D, Zakrzewski VG, Gao J, Rega N, Zheng G, Liang W, Hada M, Ehara M, Toyota K, Fukuda R, Hasegawa J, Ishida M, Nakajima T, Honda Y, Kitao O, Nakai H, Vreven T, Throssell K, Montgomery JA, Jr, Peralta JE, Ogliaro F, Bearpark M, Heyd JJ, Brothers E, Kudin KN, Staroverov VN, Keith T, Kobayashi R, Normand J, Raghavachari K, Rendell A, Burant JC, Iyengar SS, Tomasi J, Cossi M, Millam JM, Klene M, Adamo C, Cammi R, Ochterski JW, Martin RL, Morokuma K, Farkas O, Foresman JB, Fox DJ. Gaussian 09, Revision A.02. Gaussian, Inc; Wallingford CT: 2016. [Google Scholar]
  • 28.Lee C, Yang W, Parr RG. Development of the Colle-Salvetti correlation-energy formula into a functional of the electron density. Phys Rev B. 1988;37:785–789. doi: 10.1103/PhysRevB.37.785. [DOI] [PubMed] [Google Scholar]
  • 29.Becke AD. A new mixing of Hartee-Fock and local density functional theories. J Chem Phys. 1993;98:1372. [Google Scholar]
  • 30.Wadt WR, Hay PJ, Wadt WA, Hay PJ. Ab initio effective core potentials for molecular calculations. Potentials for main group elements Na to Bi. J Chem Phys. 1997;82:284–298. doi: 10.1063/1.448800. [DOI] [Google Scholar]
  • 31.Krishnan R, Binkley JS, Seeger R, Pople JA, Krishnan R, Binkley JS, Seeger R, Pople JA. Self ‐ consistent molecular orbital methods. XX. A basis set for correlated wave functions. J Chem Phys. 1984;72:650–654. doi: 10.1063/1.438955. [DOI] [Google Scholar]
  • 32.Hay PJ, Wadt WR. Ab initio effective core potentials for molecular calculations. Potentials for K to Au including the outermost core orbitals. J Chem Phys. 1985;82:299–310. doi: 10.1063/1.448975. [DOI] [Google Scholar]
  • 33.Leftwich TR, Teplyakov AV. Calibration of computationally predicted N 1s binding energies by comparison with X-ray photoelectron spectroscopy measurements. J Electron Spectrosc Rel Phenom. 2009;175:31–40. doi: 10.1016/j.elspec.2009.07.002. [DOI] [Google Scholar]
  • 34.Giesbers M, Marcelis ATM, Zuilhof H. Simulation of XPS C1s spectra of organic monolayers by quantum chemical methods. Langmuir. 2013;29:4782–4788. doi: 10.1021/la400445s. [DOI] [PubMed] [Google Scholar]
  • 35.Zhao J, Gao F, Pujari SP, Zuilhof H, Teplyakov AV. Universal Calibration of Computationally Predicted N 1s Binding Energies for Interpretation of XPS Experimental Measurements. Langmuir. 2017;33:10792–10799. doi: 10.1021/acs.langmuir.7b02301. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 36.Teplyakov A, Kung H. Selectivity and mechanism of thermal decomposition of β-diketones on ZnO powder. J Catal. 2015;330:145–153. doi: 10.1016/j.jcat.2015.07.021. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 37.Droes SR, Kodas TT, Hampden-Smith MJ. Etching of ZnO Films with Hexafluoroacetylacetone. Adv Mater. 1998;10:1129–1133. doi: 10.1002/(SICI)1521-4095(199810)10:14<1129∷AID-ADMA1129>3.0.CO;2-I. [DOI] [Google Scholar]
  • 38.Umezaki T, Gunji I, Takeda Y, Mori I. Thermal dry-etching of nickel using oxygen and 1,1,1,5,5,5-hexafluoro-2,4-pentanedione (Hhfac) 16th Int Conf Nanotechnol - IEEE NANO 2016. 2016:135–138. doi: 10.1109/NANO.2016.7751359. [DOI] [Google Scholar]
  • 39.Biesinger MC, Payne BP, Grosvenor AP, Lau LWM, Gerson AR, Smart RSC. Resolving surface chemical states in XPS analysis of first row transition metals, oxides and hydroxides: Cr, Mn, Fe, Co and Ni. Appl Surf Sci. 2011;257:2717–2730. doi: 10.1016/j.apsusc.2010.10.051. [DOI] [Google Scholar]
  • 40.Teng YT, Pramana SS, Ding J, Wu T, Yazami R. Investigation of the conversion mechanism of nanosized CoF2. Electrochim Acta. 2013;107:301–312. doi: 10.1016/j.electacta.2013.05.107. [DOI] [Google Scholar]
  • 41.Popovici D, Czeremuzkin G, Meunier M, Sacher E. Laser-induced metal-organic chemical vapor deposition (MOCVD) of Cu(hfac)(TMVS) on amorphous Teflon AF1600 : an XPS study of the interface. Appl Surf Sci. 1998;126:198–204. [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

supplement
NIHMS973590-supplement.docx (1.1MB, docx)

RESOURCES