Atomic Layer Deposition of Nickel Carbide from a ... - ACS Publications

Subscriber access provided by UNIV OF NEW ENGLAND ARMIDALE

Article

Atomic Layer Deposition of Nickel Carbide from a Nickel Amidinate Precursor and Hydrogen Plasma Qun Guo, Zheng Guo, Jianmin Shi, Wei Xiong, Haibao Zhang, Qiang Chen, Zhongwei Liu, and Xinwei Wang ACS Appl. Mater. Interfaces, Just Accepted Manuscript • DOI: 10.1021/acsami.8b00388 • Publication Date (Web): 14 Feb 2018 Downloaded from http://pubs.acs.org on February 15, 2018

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.

ACS Applied Materials & Interfaces 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 23 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

Atomic Layer Deposition of Nickel Carbide from a Nickel Amidinate Precursor and Hydrogen Plasma

Qun Guo,1,a Zheng Guo,2,a Jianmin Shi,3 Wei Xiong,2 Haibao Zhang,1 Qiang Chen,1 Zhongwei Liu,1,* and Xinwei Wang2,*

1

Laboratory of Plasma Physics and Materials, Beijing Institute of Graphic Communication,

Beijing 102600, China

2

School of Advanced Materials, Shenzhen Graduate School, Peking University, Shenzhen

518055, China

3

Institute of Nuclear Physics and Chemistry, China Academy of Engineering Physics, Mianyang

621000, China

* Email: [email protected] (Z.L.); [email protected] (X.W.)

a

Q.G. and Z.G. contributed equally to this work.

1

ACS Paragon Plus Environment

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 23

ABSTRACT

A

new

ALD

process

for

depositing

Ni3Cx

thin

films

is

reported,

using

bis(N,N′-di-tert-butylacetamidinato)nickel(II) and H2 plasma. The process shows a good layer-by-layer film growth behavior with a saturated film growth rate of 0.039 nm/cycle for a fairly wide process temperature window from 75 to 250 °C. Comprehensive material characterizations are performed on the Ni3Cx films deposited at 95 °C with various H2 plasma pulse lengths from 5 to 12 s, and no appreciable difference is found with the change of the plasma pulse length. The deposited Ni3Cx films are fairly pure, smooth, and conductive, and the x in the nominal formula of Ni3Cx is approximately 0.7. The ALD Ni3Cx films are polycrystalline with a rhombohedral Ni3C crystal structure, and the films are free of nanocrystalline graphite or amorphous carbon. Lastly, we demonstrate that, by using this ALD process, highly uniform Ni3Cx films can be conformally deposited into deep narrow trenches with aspect ratio as high as 20:1, which thereby highlights the broad and promising applicability of this process for conformal Ni3Cx film coatings on complex high-aspect-ratio 3D architectures in general.

Keywords: atomic layer deposition, nickel carbide, nickel amidinate precursor, hydrogen plasma, conformal coating

2


Page 3 of 23 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


INTRODUCTION

Nickel and its compounds have broad and important applications in many areas such as microelectronics,1 catalysis,2 and biochemistry.3 In particular, nickel carbide (Ni3Cx) is an important material for its promising applications in electrocatalysis,4-8 electrochemical supercapacitors,9 fabrication of carbon nanotubes10 and graphene,11,12 and photothermal therapy.13 Among these applications, many need the Ni3Cx to be synthesized into various nanoscale structures. For instance, nanoparticular Ni3C was found as a superior electrocatalyst for sodium borohydride electrooxidation.4 However, prior reports on the synthesis of nanostructured Ni3Cx were relatively few as compared to Ni metal, NiOx, and NiSx, which was perhaps due to the fact that Ni3Cx is metastable and subject to thermal decomposition into Ni metal and carbon at a fairly low temperature around 300 °C.14 The reported Ni3Cx synthesis methods are mainly vacuum4 or solution phase thermolysis15-18 and mechanical ball milling.19

Recently, atomic layer deposition (ALD) has aroused great attention for synthesizing nanomaterials.20-22 ALD is a powerful thin-film deposition technique that employs alternating self-limiting surface reactions to enable film growth in a layer-by-layer fashion.23 ALD is featured for its atomic-precision control over film thickness and composition, and it enables for highly conformal film coatings on complex 3D nanoscale architectures, which can hardly be achieved by other methods.24 With these merits, ALD has been widely adopted for nanoscale thin-film coatings and surface engineering for a variety of applications in microelectronics,24-26 catalysis,27-29 energy technology,20-22,30-33 and so forth. 3



Page 4 of 23

Despite the numerous benefits of ALD, the process development for the ALD of Ni3Cx is considerably lagged.34 To the best of our knowledge, there were only two recent reports that specifically targeted for the ALD of Ni3Cx. Sarr et al.35 reported the first ALD process for nickel carbide by using nickel acetylacetonate and a primary alcohol, but, perhaps due to its relatively high deposition temperature (220~300 °C), the deposited films contained a high amount of excess carbon, which essentially afforded mixed carbon-Ni3C films. The other ALD report of nickel carbide

was

from

our

own

group,9

where

a

nickel

diazadienyl

precursor

(i.e.,

bis(1.4-di-tert-butyl-1,3-diazabutadienyl)nickel(II)) was used along with hydrogen (H2) plasma. In that ALD process, crystalline stoichiometric Ni3C films were deposited at a relatively low temperature around 95 °C, but, possibly due to the relatively low thermal stability of the nickel diazadienyl precursor, the deposition temperature was limited to be only ≤105 °C.9 In addition to the above two reports, there were also several other works that reported the formation of Ni3C as a minor impurity phase in their ALD processes targeting for pure Ni metal.36-38 These works may also provide some additional chemistry insights to the precursor choice and process engineering for the ALD of Ni3Cx. Nevertheless, considerable effort is still largely in need for the process development of ALD Ni3Cx, especially in order to match up the rapidly increasing demands from the application end.

Herein, we report a new ALD process for Ni3Cx from a nickel amidinate precursor of bis(N,N′-di-tert-butylacetamidinato)nickel(II) (Ni(amd)2) with the use of H2 plasma as the coreactant. The Ni(amd)2 precursor is a fairly volatile compound, and it only needs mild heating 4


Page 5 of 23 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


to ~70 °C to afford sufficient vapor pressure for ALD.39 Ni(amd)2 is also believed to be of high reactivity on metal surface, since the metal amidinate family have been widely used for depositing various metals.40 On the other hand, H2 plasma is a highly reactive coreactant, as it contains many reactive species, such as electrons, photons, ions, radicals, and excited-state molecules. In particular, H2 plasma is much more reactive than molecular H2 for ALD, and, in fact, our trial experiments showed that using molecular H2 instead did not afford film deposition at temperature up to 250 °C. Generally speaking, using plasma in ALD is beneficial for reducing the deposition temperature and broadening the precursor choice.41

EXPERIMENTAL SECTION

ALD of Ni3Cx films. Nickel carbide (Ni3Cx) thin films were deposited in a home-made tubular ALD reactor, using Ni(amd)2 as the metal precursor and hydrogen (H2) plasma as the coreactant. The molecular structure of Ni(amd)2 is schematically shown in the inset of Figure 1(a). Our ALD chamber consisted of a 60 cm long fused silica tube (diameter 40 mm) and an aluminum heatable sample stage. Additional information regarding the ALD reactor structure can be found in our previous publication.42 The nickel precursor, Ni(amd)2, is a solid at room temperature, and it was kept in a glass container and heated to 75 °C during deposition to afford sufficient vapor pressure for film growth. The Ni(amd)2 vapor was delivered into the deposition chamber with the assist of purified N2 gas (50 sccm) as the carrier gas. H2 plasma was generated from discharging molecular H2 gas (99.999%, 50 sccm) using a 13.56 MHz radio frequency (RF) power source (output power 0~500 W). As will be shown in the following, Ni3Cx films could be 5



deposited from Ni(amd)2 and hydrogen plasma, but Ni(amd)2 and molecular H2 gas do not directly react at low deposition temperatures (at least up to 250 °C). Thus, in this case, we can safely use H2 gas (50 sccm) as the purging gas to remove excess nickel precursor. Using H2 gas for purging can provide additional convenience for deposition, because in the following plasma steps, one only needs to turn on the RF power in order to generate the hydrogen plasma, but no need to wait for the gas flow to stabilize as required for switching gas types. To investigate the deposition behavior of Ni3Cx films, the lengths of the Ni(amd)2 pulse, purge, and plasma pulse as well as the plasma RF power were systematically varied, and the deposition temperature was also varied from 75 to 300 °C. Si wafers and glass slides were used as the deposition substrates in this study, and no appreciable difference was observed for the films deposited on these substrates. All the substrates were sequentially cleaned by using acetone, methanol, and isopropanol and then pretreated by 1 min of H2 plasma prior to the film deposition.

Ni3Cx film characterizations. The film thickness was measured by first scratching the deposited Ni3Cx film and then measuring the step profile (Veeco, Dektak 150) across the scratched line. The obtained thickness values were also verified by cross-sectional scanning electron microscopy (SEM) (Hitachi, SU8020) for thick films. The surface morphology of the films was examined by SEM and atomic force microscopy (AFM) (Veeco, diInnova). The film microstructure was examined by transmission electron microscopy (TEM) (Jeol, JEM-3200F), and the film crystal structure was examined by X-ray diffraction (XRD) (Rigaku, D/max-2200 PC). The composition of the films was analyzed by X-ray photoelectron spectroscopy (XPS) 6


Page 6 of 23

Page 7 of 23 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


(Thermo Scientific, Escalab 250Xi). The Raman spectra were obtained by a Raman spectrometer (Titan, RTS-2) using an excitation wavelength of 532 nm. The sheet resistance of the films was measured by four-point probes method (RTS-8).

RESULTS AND DISCUSSION

FIGURE 1. Data in panels (a) and (b) show that the growth rate of Ni3Cx approaches a constant value as the pulse lengths of (a) Ni(amd)2 and (b) H2 plasma increase, respectively. Inset of (a) shows the molecular structure of Ni(amd)2. (c) Growth rate as a function of the purge length after the Ni(amd)2 pulse. (d) Growth rate as a function of the plasma RF input power. (e) Film thickness as a function of the total ALD cycles. (f) Growth rate as a function of deposition temperature. ALD of Ni3Cx thin films from Ni(amd)2 and H2 plasma was carefully investigated by systematically varying the deposition parameters. We started from a representative deposition recipe of using 5 s, 15 s, 10 s, and 10 s for the lengths of the Ni(amd)2 pulse, purge after the Ni(amd)2 pulse, plasma pulse, and purge after the plasma pulse, respectively, along with a plasma 7



RF power of 60 W at a deposition temperature of 95 °C, and sequentially varied each of the above parameters while keeping fixed the others to study the film growth behavior. The results are shown in Figure 1, in which the presented film growth rates were calculated from the thicknesses of 300-cycle ALD films. Figure 1(a) displays the change of the film growth rate with the pulse length of the nickel precursor, Ni(amd)2. As the Ni(amd)2 pulse length increased from 1 to 9 s, the film growth rate first appreciably increased from 0.029 to 0.039 nm/cycle and then remained constant when the Ni(amd)2 pulse length exceeded 3 s. These results suggested that the surface chemistry reaction of this precursor was self-limiting and the film growth saturated when the Ni(amd)2 pulse was longer than 3 s. Similarly, Figure 1(b) shows the change of the growth rate with the pulse length of H2 plasma. As the plasma pulse was elongated from 3 to 10 s, the film growth rate first increased and then reached constant when the plasma pulse length exceeded 4 s, which indicated that the film growth saturated when the plasma pulse was longer than 4 s. The saturated film growth behavior is of high importance for an ALD process; a good saturation behavior, as in this case, indicates that the film growth can be well controlled in a layer-by-layer growth fashion. The effect of the H2 gas purging time was also investigated. As shown in Figure 1(c), by elongating the purge length after Ni(amd)2 from 1 to 15 s, the film growth rate first considerably decreased and then reached constant when the purge length exceeded 7 s. These results indicated that >7 s of purging was sufficient to remove all the excess Ni(amd)2. Figure 1(d) shows the film growth rate as a function of the plasma RF input power. As the RF power increased from 40 to 80 W, the growth rate first increased from 0.033 to 0.039 nm/cycle and then remained almost constant at 0.039 nm/cycle when the power exceeded 50 W. In a RF plasma, the 8


Page 8 of 23

Page 9 of 23 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


input power is in positive correlation with the electron density and electron temperature, and therefore a higher input power is expected to generate more energetic electrons which would produce a higher density of reactive hydrogen radicals in the gas phase. Accordingly, increasing the RF power would have a similar effect as elongating the plasma pulse. Nevertheless, all the characterizations so far have shown that the deposition process of Ni3Cx films followed a typical ALD behavior, where the film growth was self-limiting and saturated in thickness in each ALD cycle, and the deposition parameters where we initially started (i.e., 5 s, 15 s, 10 s, and 10 s for the sequential Ni(amd)2 pulse, purge, plasma pulse, and purge lengths, respectively, and 60 W for the plasma power) were actually in the regime for saturated film growth. Therefore, unless otherwise specified, we chose this set of parameters for the following depositions.

Linear film growth behavior was examined by varying the number of total ALD cycles. As shown in Figure 1(e), a good linear relation was indeed found for the film thickness with respect to the total ALD cycles. The linear growth behavior is an important feature for an ALD process, since it implies that the thickness of the ALD films could be precisely controlled by digitally varying the total ALD cycles. In addition, the linear fit shown in Figure 1(e) gave out a negligible intercept, which indicated that no nucleation delay occurred during the initial film growth. Also, the fitted slope gave out a growth rate of 0.039 nm/cycle, which was the same as previously obtained using single-point thickness data. The temperature dependence of the film growth is shown in Figure 1(f). The growth rate was found to be roughly a constant of 0.039 nm/cycle for the deposition temperature up to 250 °C. However, when the temperature was elevated beyond 9



280 °C, the growth rate drastically increased, possibly owing to the onset of the thermal decomposition of the Ni(amd)2 precursor.39 Nevertheless, the ALD process of Ni3Cx herein showed a fairly wide temperature window for processing (i.e., at least 75~250 °C).

In the following, we focused our film characterizations on the ALD Ni3Cx films deposited at 95 °C. The surface morphology of the films was examined by SEM and AFM. Figure 2 shows typical SEM and AFM images for a ~12 nm film deposited at 95 °C with 300 ALD cycles. As shown in Figure 2, the ALD Ni3Cx film was continuous and smooth, and the rms roughness value was 1.72 nm, which was only 14% of the film thickness. Also, both the SEM and AFM images show granular features, which suggests that the film was polycrystalline.

FIGURE 2. Representative (a) SEM and (b) AFM images for a ~12 nm Ni3Cx film deposited at 95 °C with 300 ALD cycles. Both scale bars represent 200 nm. XPS was employed to analyze the chemical composition of the ALD films. The XPS measurements were taken on ~45 nm Ni3Cx films deposited at 95 °C using various pulse lengths 10


Page 10 of 23

Page 11 of 23 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


of the H2 plasma from 5 to 12 s. Prior to the collection of the spectra, all the samples were sputtered using 2 kV Ar+ ions for 60 s to remove the adventitious carbon and surface oxide. The obtained XPS spectra are displayed in Figure 3, and the spectra for different H2 plasma pulse lengths exhibit generally the same features. The survey spectra (Figure 3(a)) show only the peaks corresponding to Ni and C, and the possible impurity peaks of N and O are all within the noise level. High-resolution XPS spectra were taken for Ni 2p, C 1s, N 1s, and O 1s. The Ni 2p spectra (Figure 3(b)) show pairs of well-defined spin-orbit split peaks at 852.9 eV (Ni 2p3/2) and 870.2 eV (Ni 2p1/2) and the C 1s spectra (Figure 3(c)) show single peaks at 283.5 eV, which are all consistent with the formation of Ni-C bonds as in previous reports of nickel carbide.16 On the other hand, the N 1s and O 1s emission intensities were merely above the noise levels, as shown in Figure 3(d-e), which indicates that the ALD Ni3Cx films were quite pure. We further extracted the atomic percentages of these elements from their peak areas and listed the obtained numbers in Table 1, and the impurity levels of N and O were found to be only ~0.5 at.% and ~0.1 at.%, respectively. The x as in Ni3Cx was also extracted to be around 0.7 (the error in determining x is about ±10%). This number shows that the films contained fairly appreciable amount of carbon but less than that of stoichiometric Ni3C (vide infra). In addition, prior works35,43 suggested that the position of the satellite of Ni 2p3/2 is correlated with the presence of C atoms in Ni lattice. For pure Ni metal, the satellite peak is located at +6 eV higher in bind energy than the main Ni 2p3/2 peak; whereas for Ni3C, the satellite should be located at +7 eV due to the modification of the core-hole screening via less-localized 4sp electrons by the C atoms.43 In our case, we found that the satellite position of the ALD Ni3Cx was at +6.4±0.1 eV (Figure 2(b)), which fell in between of 11



Page 12 of 23

the above two numbers. The result is reasonable and consistent, since the carbon content of the ALD Ni3Cx was also in between of those for pure Ni and Ni3C.

FIGURE 3. XPS results for ~45 nm Ni3Cx films deposited at 95 °C using various H2 plasma pulse lengths of 5, 7, 10, and 12 s. (a) Survey spectra and high-resolution spectra for (b) Ni 2p, (c) C 1s, (d) N 1s, and (e) O 1s core-level emissions. Symbol ∆ in (b) represents the binding energy difference from the satellite to the main peak of Ni 2p3/2 emission. TABLE 1. Elemental composition of the ALD Ni3Cx films deposited at 95 °C using various H2 plasma pulse lengths. Data were extracted from XPS results, and the error in determining x is about ±10%. Pulse length of H2 plasma (s)

Ni (at.%)

C (at.%)

N (at.%)

O (at.%)

x in Ni3Cx

5 7 10 12

79.9 80.7 82.1

19.4 18.9 17.5 18.3

0.5 0.3 0.3 0.8

0.2 0.1 0.1 0.1

0.73 0.70 0.64 0.68

80.8

12


Page 13 of 23 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


FIGURE 4. XRD results for ~80 nm Ni3Cx films deposited at 95 °C using various H2 plasma pulse lengths of 5, 7, 10, and 12 s. Vertical bars (green) represent the data of rhombohedral Ni3C from PDF#06-0697. The crystal structure of the films was analyzed by XRD. Figure 4 shows the XRD results for ~80 nm Ni3Cx films deposited at 95 °C using various H2 plasma pulse lengths. The diffraction peaks shown in Figure 4 can be assigned to the (110), (006), (113), and (116) planes of a rhombohedral Ni3C crystal structure (PDF#06-0697), and the crystallite size was estimated to be ~10 nm by using Scherrer equation from the diffraction peak widths. It should be noted that, while bulk elemental Ni adopts a face centered cubic (fcc) structure, it also has an allotrope of a hexagonal close packed (hcp) structure; the rhombohedral Ni3C is in fact a superstructure of hcp-Ni with ordered distribution of interstitial carbon, and Ni3Cx (x = 0~1) is a solid solution that bridges the hcp-Ni and rhombohedral Ni3C.15 The lattice expansion induced by the interstitial carbon is very small (

Atomic Layer Deposition of Nickel Carbide from a ... - ACS Publications

Recommend Documents