Diamond-like carbon doped with highly π-conjugated molecules by plasma-assisted CVD

The development of plasma processes for carbon-containing molecules are promising new routes for the production of new carbon materials, since solid-state carbon materials generally have very high melting temperature and do not dissolve in solvents. It is expected that this research direction will lead to a new field of chemistry for synthesizing carbon-based materials with novel and useful physical properties. Examples can be found in the plasma synthesis of diamondoid,^1–3) the doping of carbon solid with various elements,^4–7) and the facile synthesis of single wall carbon nanotubes.^8,9)

Diamond-like carbon (DLC) is widely used in various applications such as hard coatings, gas barriers, and anticorrosion coatings.^10,11) DLC is produced by the plasma deposition of carbon and hydrogen using hydrocarbon gases as the source materials. Because of its simple coating process and wide application, DLC is now produced on a large industrial scale. However, the microscopic structures of DLC have been characterized only recently, and precise control of the nanostructures in DLC has not yet been established. The nanostructure control of carbon in DLC will lead to various useful properties such as tribology,^12–14) controlled refractive index^10,15,16) and other optical properties,^10,15,16) electronic and electrochemical properties,^5,6,17–19) catalytic activities,^20–22) and biological functions.^23,24) Although there are various distinctions among sp³-rich carbon–hydrogen systems,¹¹⁾ we use the term "DLC" in this paper for simplicity.

We are studying the fabrication of nanostructured carbon materials using plasma processing. Our interest is in the reaction of extended π-electron systems of organic semiconductor molecules with plasma, which is a kind of interaction between plasma and nanointerfaces.²⁵⁾ In this paper, we report the codeposition of large organic molecules during the fabrication of DLC by plasma CVD. The molecules used were copper phthalocyanine (CuPc, C₃₂H₁₆N₈Cu) and perylenetetracarboxylicdianhydride (PTCDA, C₂₄H₈O₆), both of which are known as robust dye molecules and have recently attracted attention as organic semiconductors. Optical properties and Raman spectra were examined for the characterization of the resulting materials.

We used a rather unique arrangement for the plasma deposition of DLC to confine the plasma and allow for large organic molecules to be incorporated in the DLC without excessive decomposition. A schematic drawing of the deposition chamber is shown in Fig. 1 . The chamber is equipped with the plasma source (an RF-sputter gun), a molecular source (a Knudsen-cell), a gas inlet, and a pumping line with a bypass valve. The sample holder is made of stainless steel and is floated at −400 V from the ground level in order to deposit sp³-carbon-rich DLC. After the introduction of the substrate [Al₂O₃(0001) with dimensions of 5 × 5 × 0.43 mm³] and molecular source, the chamber was pumped down to 10⁻⁴ Pa by a turbo molecular pump (TMP), and then the pumping rate was reduced by closing the main valve. Methane gas, the carbon and hydrogen source, was supplied from a needle valve. The pumping rate was controlled to adjust the pressure of CH₄ to 10 Pa. The sample bias (−400 V) was then applied and the plasma was ignited. The plasma was produced using a 1 inch RF sputtering gun with a graphitic carbon target. The size of the white glow of the plasma is about 5 cm, and it did not directly touch the substrate or the molecular source, as observed visually. After the stabilization of the RF plasma, the molecular source was heated to a certain temperature (∼250 °C for CuPc and ∼200 °C for PTCDA, which is intended to dope these molecules by ∼10% into DLC). The temperature of the substrate was not controlled. It was found that the temperature during the deposition did not exceed 60 °C.

Zoom In Zoom Out Reset image size

The optical emission spectra during the DLC deposition were measured using a fiber coupled spectrometer (Ocean Optics USB2000). The optical absorption spectra and Raman spectra of the films were obtained using Lambda 900 (Lambda Physics) and InVia (Renishaw, 20×–50× objective lens and 532 or 785 nm excitation), respectively.

Figure 2 shows the optical emission spectra (OES) profiles during the deposition. The main emissive components are H_α (657 nm), H_β (486 nm), and CH (A²Δ → X²Π at 430.6 and 430.7 nm,$\text{B}^{2}\Sigma$ → X²Π at 390.3 nm),^26–28) which consistute the typical C–H plasma for the deposition of DLC. Because the emission from C₂ (516 nm) is very weak, the chemistry in the plasma is governed by hydrogen-containing species. The OES profiles did not change noticeably by the codeposition of the organic semiconductor molecules, which means that the concentration of the organic semiconductor molecules was low. This suggests that the organic semiconductor molecules did not react with each other. The C- and H-containing radicals in the plasma mainly interacted with them. This is not always the case, however, owing to the high backpressure of CH₄ (10 Pa), which might cause clustering during the deposition. This issue will be discussed later.

Zoom In Zoom Out Reset image size

In the following, we describe the results of CuPc first, and then those of PTCDA. Figure 3 shows the optical absorption spectra of (a) a DLC film, (b) a DLC with a codeposited CuPc film, and (c) a pristine CuPc film. The DLC (a) shows a very smooth increase in the optical absorption starting from 500 nm. This reflects the amorphous nature of DLC. DLC "doped- with" CuPc (b) shows absorption at 350 nm and a broad peak centered at 630 nm. In the case of the pristine CuPc film (c), a Soret band (∼350 nm) and a Q-band (550–750 nm) with substantial Davydov splitting are observed as reported.²⁹⁾ The resemblance in the optical absorption between (b) and (c) suggests that the CuPc molecules were not decomposed completely. Because the Davydov splitting comes from the interaction between transition dipoles with fixed geometry in crystalline solids, the CuPc molecule separately embedded in the matrix (DLC in the present case) will show a single peak in the optical absorption. The less pronounced splitting of Q-band absorption in (b) might come from this isolation effect or the destruction of the molecule.

Zoom In Zoom Out Reset image size

To examine the above possibilities, we measured the Raman spectra of the samples. Figure 4 shows the Raman spectra of (a) the DLC codeposited with CuPc, (b) the CuPc single crystal, and (c) the pristine CuPc film fabricated by vacuum deposition. A broad peak from 1000 to 1800 cm⁻¹ is typically observed in DLC. The center of this peak is at approximately 1390 cm⁻¹, which indicates the formation of almost full sp³ DLC (below 1400 cm⁻¹). Note that the Raman spectra depend on the crystal structure of CuPc, that is of the β-type in (b) and mainly of the α-type in (c).³⁰⁾ By comparing the spectra shown in Fig. 4(a) and with those in Figs. 4(b) and 4(c), we can observe that the molecular fingerprint of the Raman spectra is surprisingly maintained as marked by the arrows, although the peaks broadened and somewhat shifted. We consider that the vibrational structure of the extended π-conjugate molecule is almost intact, which means that only minor changes occur in the structure, such as the replacement of peripheral hydrogen by CH₃ carbon. However, the CuPc molecules definitely reacted with plasma to form some new chemical bonds, because the broadening and shift of the Raman peaks were observed and many Raman peaks disappeared. The small splitting of the optical absorption Q-band [Fig. 3(b)] may also be explained by this covalent bond formation, without considering the clustering of CuPc in the vapor phase.

Zoom In Zoom Out Reset image size

Figure 5 shows the optical absorption spectra of (a) DLC [same as in Fig. 3(a)], (b) the DLC codeposited with PTCDA, and (c) the pristine PTCDA film. Because the spectrum change between (b) and (c) is substantial, we consider that the π-electron system of PTCDA was substantially altered, which led us to infer that the PTCDA molecules have been decomposed with the C–H plasma. Figure 6 shows the Raman spectra of (a) the DLC co-deposited with PTCDA and (b) the pristine PTCDA film. The shape of the DLC region from 1000 to 1800 cm⁻¹ shows that the DLC is again mainly of the sp³-type, but that the sp²/sp³ ratio is increased from that in the case of the CuPc-doped DLC, because the peak shifted to higher wave numbers. The striking difference from CuPc is that only one distinct peak different from DLC is found in (a), at approximately 1080 nm. There is a strong peak near this wave number in the pristine PTCDA. This peak has been assigned by quantum chemical computation to the in-plane C–O–C bending.³¹⁾ Our experimental result strongly suggests that this moiety (C–O–C) was robust and did not decompose during the interaction with C–H plasma. Combined with the upshift of the DLC peak, we consider that the decomposed π-electron systems of PTCDA by C–H plasma contribute to the increased sp²/sp³ ratio.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

We fabricated novel carbon films by codepositing large organic semiconductor molecules during the plasma-assisted chemical vapor deposition of DLC, which can be regarded as an attempt at doping molecules with extended π-systems into DLC. The OES result shows that C–H radical plasma was formed in the process, even with the codeposition of π-electron molecules. The DLC was composed of mostly sp³ as examined by Raman spectroscopy; however, the sp²/sp³ ratio was slightly dependent on the type of dopant molecules. In the case of CuPc, the doped film showed a similar optical absorption profile to the pristine molecular film, but substantially different Raman spectra were observed. PTCDA showed a substantial destruction of the π-electron system from the optical absorption and Raman spectroscopy results. The Raman spectrum of PTCDA-doped DLC indicated that the carboxylic dianhydride (C–O–C) moiety remains intact during the plasma process. The sp²/sp³ ratio slightly increased in the case of PTCDA, which might have been caused by the decomposed π-electron species.

This work was supported in part by a Grant-in-Aid for Scientific Research on Innovative Areas "Frontier Science of Interactions between Plasmas and Nano-Interfaces" (Grant No. 21110002) from the Ministry of Education, Culture, Sports, Science and Technology of Japan (MEXT). We would like to thank Professor T. Hasegawa (The University of Tokyo) for valuable discussions and support through the JST-CREST program. The experimental work was conducted at Hokkaido University, supported by the "Nanotechnology Platform" Program of MEXT.