PLASMA-ENHANCED CHEMICAL VAPOR DEPOSITION METHODS FOR GRAPHENE DEPOSITION

Abstract

A graphene deposition process. The process includes the steps of placing a substrate into a deposition chamber and heating the chamber, generating radio frequency plasma at a location proximate to the substrate while flowing a precursor gas containing carbon through the plasma and over the substrate.

Claims

1. A graphene deposition process, comprising: placing a substrate into a deposition chamber; heating the chamber; and generating radio frequency plasma at a location proximate to the substrate while flowing a precursor gas containing carbon through the plasma and over the substrate.

2. The process of claim 1, wherein the chamber is heated to a temperature of between 200 and 1000 C.

3. The process of claim 2, wherein the chamber is heated to a temperature of between 550 and 650 C.

4. The process of claim 1, the substrate is a metal selected from the group consisting of Copper, Aluminum, Nickel, Titanium, Titanium Nitride, Cobalt, Ruthenium, Platinum, Iridium, and Silver.

5. The process of claim 1, the substrate is an insulator selected from the group consisting of SiO.sub.2, Sapphire, MgO, and Quartz.

6. The process of claim 1, the substrate is a semiconductor selected from the group consisting of Silicon, Silicon Carbide, and GaAs.

7. The process of claim 1, the precursor gas is selected from the group consisting of CH.sub.4, C.sub.2H.sub.2, C.sub.2H.sub.4, and C.sub.2H.sub.6.

8. The process of claim 1, prior to placing the substrate into the deposition chamber, comprising: washing the substrate with i) solvents; or ii) an etching solution.

9. The process of claim 8, the solvent is selected from the group consisting of Toluene, Acetone, Methanol and Isopropyl alcohol; and the etching solution is piranha etch.

10. The process of claim 1, the precursor gas is flown over the substrate at a partial pressure of between 10.sup.04 to 1 Torr.

11. The process of claim 10, the precursor gas is flown over the substrate at a partial pressure of about 1 mTorr.

12. The process of claim 1, the precursor gas is flown over the substrate at a flow rate of between 1 sccm to 500 sccm.

13. The process of claim 12, the flow rate is about 6 sccm.

14. The process of claim 1, the radio frequency plasma is generated at about 13.56 MHz.

15. The process of claim 1, the plasma is generated upstream from the substrate.

16. The process of claim 1, the plasma is generated adjacent to the substrate.

17. The process of claim 1, precursor gas is flown through the plasma and over the substrate for a predetermined period of time, ranging between about 1 to about 30 minutes.

18. The process of claim 17, the predetermined period of time is between about 1 to about 15 minutes, when the substrate is copper.

19. The process of claim 17, the predetermined period of time is between about 5 to about 15 minutes, when the substrate is Quartz.

20. The process of claim 17, the predetermined period of time is between about 10 to about 15 minutes, when the substrate is SiO.sub.2.

Description

BRIEF DESCRIPTION OF DRAWINGS

[0011] FIG. 1a is a process flowchart which depicts steps involved in generating graphene, according to the present disclosure.

[0012] FIG. 1b is a Raman spectroscopy output of intensity vs. Raman shift measured in cm.sup.1 which depicts different growth periods of graphene on SiO.sub.2 for 1 min, 5 min, 10 min, and 15 min.

[0013] FIG. 1c is a top view scanning electron microscope (SEM) output depicting graphene growth.

[0014] FIG. 1d is an atomic force microscopy (AFM) scan of a substrate with graphene grown thereon.

[0015] FIG. 1e is a graph of height of graphene grown on a substrate as a function of a linear scanning.

[0016] FIGS. 1f, 1g, and 1h are graphs that statistically (i.e., frequency of occurrence) express area of graphene islands in nm.sup.2 (1f), distance to the nearest island in nm (1g), and height of the island in nm (1h).

[0017] FIG. 2a is an SEM image of further growth from 10 minutes to 15 minutes on an SiO.sub.2 substrate.

[0018] FIG. 2b is a TEM image of a 15-minute growth trial on an SiO.sub.2 substrate.

[0019] FIG. 2c is a graph of a device constructed by deposited graphene according to the present disclosure depicting source/drain current as a function of the Si back gate voltage with a drain voltage of 10 mV.

[0020] FIG. 3a is a Raman spectroscopy output of intensity vs. Raman shift measured in cm.sup.1 for different growth periods of graphene on quartz for 1 min, 5 min, 10 min, and 15 min trials.

[0021] FIG. 3b is an SEM output for graphene grown on quartz for a 10-minute trial.

[0022] FIG. 3c is an SEM output for graphene grown on quartz for a 15-minute trial.

[0023] FIG. 4a are graphs of measured reflectance (associated with left axis) of graphene grown on an SiO.sub.2 substrate and extracted contrast (dash lines associated with right axis) vs. wavelength measured in nm for various growth periods.

[0024] FIG. 4b are graphs of measured transmittance (associated with left axis) of graphene grown on a quartz substrate and extracted absorbance of graphene (associated with right axisthe bottom four traces) vs. wavelength measured in nm for various growth periods.

[0025] FIG. 5a is a Raman spectroscopy output of intensity vs. Raman shift measured in cm.sup.1 for different growth periods of graphene on Cu for 1 min, 5 min, 10 min, and 15 min trials.

[0026] FIG. 5b is a TEM image of the 15-minute trial for graphene grown on Cu.

[0027] FIG. 5c is an XRD output comparing intensity vs. angle measured in degrees for Cu on Si and Si.

DETAILED DESCRIPTION

[0028] For the purposes of promoting an understanding of the principles of the present disclosure, reference will now be made to the embodiments illustrated in the drawings, and specific language will be used to describe the same. It will nevertheless be understood that no limitation of the scope of this disclosure is thereby intended.

[0029] In the present disclosure, the term about can allow for a degree of variability in a value or range, for example, within 10%, within 5%, or within 1% of a stated value or of a stated limit of a range.

[0030] In the present disclosure, the term substantially can allow for a degree of variability in a value or range, for example, within 90%, within 95%, or within 99% of a stated value or of a stated limit of a range.

[0031] A novel scalable approach for generating graphene that overcomes the above-discussed shortcomings is presented. A plasma-enhanced chemical vapor deposition (PECVD) process is described herein that can generate graphene on a substrate at relatively low temperatures without the use of a catalytic layer. According to the present disclosure, a process for growth of few-layer graphene (FLGless than 10 layers) and multilayer graphene (MLGmore than 10 layers) is described at deposition temperatures ranging between 200-1000 C. using a plasma-enhanced CVD on both non-catalytic and traditional catalytic substrates within a few minutes, e.g., 1-15 minutes depending on the particular substrate chosen. This approach provides a one-step method towards catalyst-free, rapid growth of multilayer graphene at reduced temperatures. Raman spectroscopy, Scanning Electron Microscopy (SEM) and optical measurements were used to analyze the progression of graphene growth with time, providing a striking contrast with the self-limiting scheme observed in thermal CVD.

[0032] Referring to FIG. 1a, a process flowchart 100 is provided which depicts steps involved in generating graphene, according to the present disclosure. The process starts in the block 102. As a first step, in block 104, the substrate upon which the graphene is to be deposited is initially washed with a i) solvent, e.g., using a triple sonication wash in toluene, acetone and methanol or isopropyl alcohol for 5 min each; or ii) with an etching solution, e.g., piranha etch, or other solutions known to a person having ordinary skill in the art. The substrate is then loaded into a heating-plasma chamber for further processing.

[0033] Next, in block 106, the chamber within which graphene is to be deposited on to the loaded substrate is heated. An example of such a chamber is EASYTUBE 3000 FIRST NANO system, utilizing a 6-inch diameter quartz furnace with a radio frequency (RF) plasma generator, e.g., at about 13.56 MHz, or other suitable and industrial heating chambers as known to a person having ordinary skill in the art. The chamber is then heated to about 650 C. in about 20 min and evacuated in a flow of 1 slpm of Ar at 300 mTorr. Other gases known to a person having ordinary skill in the art can also be used.

[0034] The temperature to which the chamber is heated determines the grain size of the graphene that will be deposited. Higher temperatures result in higher grain sizes which in turn result in a smaller number of grain boundaries, thereby resulting in a more efficient current density once the graphene is used in an electronic device. However, high temperatures may be unsuitable for some substrates. Therefore, a balance between grain size and choice of substrate is achieved by determining the optimum heating temperature. It should be appreciated that temperature ranging between 200 and 1000 C. are within the scope of the present disclosure for substrates ranging from 1) metal-based material, 2) insulators, and 3) semiconductor-based material. These three categories include: 1) Copper, Aluminum, Nickel, Titanium, Titanium Nitride, Cobalt, Ruthenium, Platinum, Iridium, and Silver; 2) SiO.sub.2, Sapphire, MgO, and Quartz; and 3) Silicon, Silicon Carbide, III-V semiconductors such as GaAs, and GaN, respectively. Another factor that affects the operational temperature and growth of graphene is the choice of the precursor gas. For example, in accordance with another embodiment of the present disclosure, MLG films were directly deposited onto SiO.sub.2 at a lower temperature of 550 C., using C.sub.2H.sub.2 as the precursor gas.

[0035] When the temperature reaches a predetermined set point, Ar flow is shut off. To begin the deposition process of graphene, a precursor gas containing carbon is injected into the chamber while RF plasma is generated, as shown in Block 110. The plasma can be generated remotely from the substrate upstream, where at plasma species are generated away from the substrate and any graphene already deposited to avoid ion bombardment damage further providing independent control of the substrate temperature; or near the substrate. The reflected RF power can be minimized using a matching network for maximum plasma efficiency. The precursor gas, is injected into the chamber at a predetermined partial pressure and flow rate. In one exemplary embodiment, CH.sub.4 is used as the precursor gas at about 1 mTorr (other pressures from 10.sup.04 to 1 Torr are also within the scope of this disclosure) and at about a flow rate of about 6 sccm. However, other precursor gases such as C.sub.2H.sub.2, C.sub.2H.sub.4, C.sub.2H.sub.6, or other suitable gases known to a person having ordinary skill in the art can also be used. At a predetermined growth period, the precursor gas, e.g., CH.sub.4, flow is shut off and the chamber is evacuated using, e.g., Ar gas. Subsequently, the chamber is cooled down at a predetermined rate, e.g., about 15 C./min with a flow of about 100 sccm of Ar at a pressure of about 1 Torr, as provided in Block 112. Substrates with deposited graphene are then removed from the CVD chamber at temperatures below about 150 C., as provided in block 114.

[0036] One important aspect of the process described according to FIG. 1a is the ability to grow graphene on a substrate of choice without a catalytic interface. This process facilitates dismissal of myriad issues associated with use of a catalytic interface and thereby disassociation of graphene from the catalytic interface such as introduction of unwanted contamination and defects. Furthermore, the process according to the present disclosure can be accomplished at relatively low temperatures as compared to prior art CVD processes, e.g., as low as about 200 C. This low temperature provides the flexibility of certain substrate choices not available at the traditional temperatures seen in traditional CVD processes, e.g., 1500 C. Furthermore, deposition (and hence growth time) are reduced to minutes, e.g., according to one embodiment of about 1 to about 15 minutes, vs. hours making this process scalable for high volume fabrication processes.

[0037] According to one exemplary embodiment, using 550 W of plasma power both FLG and MLG growth on non-catalytic substrates, e.g., SiO.sub.2 (90 nm thick), and quartz (1 mm thick), are achieved. The plasma power can range from between 10-1000 W.

[0038] In addition to non-catalytic deposition, the process described herein is also capable of a depositing graphene on substrates typically used as a catalyst in a catalytic deposition. The process was reduced to practice with growth on 500 nm thick physical-vapor-deposited copper (Cu) films on SiO.sub.2Si substrates. A Ta underlayer of about 5 nm thick was deposited to promote Cu adhesion. Results of these and non-catalytic depositions are provided below using Raman spectroscopy, Scanning Electron Microscopy (SEM) and optical measurements.

[0039] Raman spectroscopy was carried out in ambient environment at room temperature to determine the effectiveness of the deposition process of the present disclosure. A 532 nm Diode pumped solid state (DPSS) green laser beam was focused on substrates before and after graphene was deposited thereon by an objective, e.g., OLYMPUS 50 objective (NA=0.75, WD=0.38 mm). The excited Raman scattering was collected by a Raman spectrometer, e.g., HORIBA LABRAM HR800 Raman spectrometer with an 1800 mm.sup.1 grating (spectral resolution=0.27 cm.sup.1). Transmission measurements were performed using a commercially available spectrophotometer, e.g., PERKIN ELMER Lambda 950 spectrophotometer having an integrating sphere. Transmission electron microscopy (TEM) sample was made by FIB system, e.g., FIB-SEM NOVA 200 and SiO.sub.2 and Pt layers were deposited to protect the TEM sample area during ion-beam milling. High resolution microscopy was performed with a TEM, e.g., a JEOL ARM200F operated at 200 kV. For X-ray photoelectron spectroscopy (XPS) studies, the photoelectrons were excited using an Mg-Ka (energy=1253.6 eV) X-ray radiation source (SPECS XRC-1000) and analyzed using an electron analyzer, e.g., a OMICRON ARGUS hemispherical electron analyzer with a round aperture of 6.3 mm. X-Ray Diffraction (XRD) measurements were carried out in a commercially available XRD analyzer, e.g., a BRUKER D8 XRD set-up using mixed Cu-K line (wavelength=0.151838 nm).

[0040] Referring to FIG. 1b, a Raman spectroscopy output of intensity vs. Raman shift measured in cm.sup.1 is provided for different growth periods of graphene on SiO.sub.2 for 1 min, 5 min, 10 min, and 15 min. The Raman spectrum of graphene has three dominant peaks, denoted by D, G and 2D at around 1350 cm.sup.1, 1600 cm.sup.1 and 2690 cm.sup.1, respectively. The Raman spectra provided no Raman signals from the substrate for about 1-min and about 5-min growths on SiO.sub.2, indicating no graphene growth. For longer growth time (greater or equal to about 10 min), clear Raman signals are observed, providing concrete evidence that graphene is deposited. It should be appreciated that graphene in this case was deposited without a catalytic interface. It should also be appreciated that the deposition time (i.e., the time period for which the precursor gas is flown over the substrate) is a predetermined time period based on the choice of the substrate as will be made apparent below. However, time periods of about 1 minute to about 30 minutes are within the scope of the present disclosure.

[0041] Referring to FIG. 1c, a top view SEM output is provided. This visual evidence further substantiates formation of graphene layers starting from the initial stages of growth for the 10-minute growth trial as small graphene islands begin to appear on SiO.sub.2 after 10 min of growth. Islands are disjointed clusters of graphene each containing multiple graphene grains of various sizes, e.g., 5-10, as further discussed below.

[0042] Referring to FIG. 1d, an atomic force microscopy (AFM) scan of the same substrate with graphene grown thereon is provided. FIG. 1e depicts a graph of height of graphene grown on the substrate as a function of a linear scanning. FIGS. 1d and 1e show that these graphene islands have a thickness of about 1-3 nm, indicating few-layer structures. Referring back to FIG. 1b, due to low coverage in the 10-minute sample, the corresponding Raman spectrum is rather noisy. The D-to-G intensity ratio of the Raman peaks of the 10-min sample can be approximately used to find the grain size of graphene. For the 532 nm Raman wavelength excitation, the grain size is calculated to be about 5-10 nm. These numbers are smaller than the size of the graphene islands observed under the SEM. Hence, these islands are not necessarily individual graphene grains but several grains merged together to form arbitrarily sized islands. As growth progresses, the island sizes increase.

[0043] A statistical analysis of the island size, spacing between islands and island height has been provided in FIGS. 1f, 1g, and 1h, which statistically (i.e., frequency of occurrence) express area of graphene islands in nm.sup.2, distance to the nearest island in nm, and height of the island in nm, respectively. The majority of the islands formed after about 10-min growth on the SiO.sub.2 substrate are few tens of nanometers in sizes, with sub-100 nm to 100 nm spacing.

[0044] Referring to FIG. 2a, an SEM image of further growth from 10 minutes to 15 minutes on SiO.sub.2, shows the already formed graphene islands coalesce to entirely cover the SiO.sub.2 surface (yet with visible boundaries between islands, achieving a complete coverage and yielding sharp Raman peaks, as shown in D, G, and 2D peaks of FIG. 1b for the 15-minute trial. A TEM image of the 15-minute trial is depicted in FIG. 2b. The layered structure of directly grown graphene is seen in the cross-sectional TEM of FIG. 2b, which shows that the 15-min SiO.sub.2 sample has multiple graphene layers (12 layers).

[0045] In order to investigate the efficacy of the graphene deposited on the SiO.sub.2 substrate, the transport properties of the directly grown MLG in a back-gated field-effect transistor configuration was studied by patterning graphene channels formed from the 15-min SiO.sub.2 substrate and depositing Ti/Au source-drain electrodes in a 2-probe arrangement. Referring to FIG. 2c the source/drain current as a function of the Si back gate voltage with a drain voltage of 10 mV is shown. The observed ambipolar nature of transport is a defining characteristic of a graphene-constructed device.

[0046] Referring to FIG. 3a, a Raman spectroscopy output of intensity vs. Raman shift measured in cm.sup.1 is provided for different growth periods of graphene on quartz for 1 min, 5 min, 10 min, and 15 min. Substantially identical growth recipes were used on quartz substrate and, in contrast to SiO.sub.2, quartz as a substrate begins to show Raman peaks at a shorter growth time of 5 min compared to 10 min for SiO.sub.2. Referring to FIG. 3b, an SEM output for graphene grown on quartz for a 10-minute trial are provided. As can be seen, 10 minutes allows well-connected graphene islands covering a large fraction of the substrate, in agreement with the distinct Raman peaks observed in FIG. 3a. Referring to FIG. 3c, an SEM output for graphene grown on quartz for a 15-minute trial are provided. Increasing the growth time to 15 min leads to a complete coverage of graphene on quartz. Based on the SEM and Raman results shown in FIGS. 3a, 3b, and 3c, quartz assists graphene growth faster than SiO.sub.2.

[0047] The growth of graphene as a function of growth time was further analyzed using optical measurements. Referring to FIG. 4a, graphs of measured reflectance of graphene grown on SiO.sub.2 (associated with left axis) and extracted contrast

[00001]$\left(C=\frac{R_{\mathrm{Substrate}}-R_{\mathrm{Gr}+\mathrm{Substrate}}}{R_{\mathrm{Substrate}}}\right)$

(dash lines associated with right axis) vs. wavelength measured in nm is shown for various growth periods. 1-10 min of growth on SiO.sub.2 gives no appreciable change in contrast with respect to bare substrate in spite of the formation of small islands seen under the SEM. During the next 5 min, while the coverage becomes complete, a large change in contrast is observed, suggestive of MLG grown at that stage as confirmed by TEM (FIG. 2b).

[0048] As a comparison and referring to FIG. 4b, measured transmittance and extracted absorbance

[00002]$\left(T_{\mathrm{Gr}}=\frac{T_{\mathrm{Gr}+\mathrm{Quartz}}}{T_{\mathrm{Quartz}}},A_{\mathrm{Gr}}=1-T_{\mathrm{Gr}},R_{\mathrm{Gr}}~0\right)$

of graphene (the bottom four traces) grown on a quartz substrate is provided. From the extracted absorbance, it should be noted that the 1-min sample shows no graphene growth as further from its Raman spectrum (see FIG. 3a). Monolayer graphene absorbs about 2.3% of light in near-IR regime, hence, an absorption value of 9% shown by the 10-min sample would then correspond to 4 layers of graphene. However, since the growth is not yet complete at 10 min, it should be noted that the PECVD process yields few layer graphene islands during the initial stage of the growth, which have transmission equivalent to 4 uniform layers of graphene. This observation suggests that absorption cannot be used to quantitatively analyze the number of layers unless growth is confirmed to be uniform across the surface. As the time increases to 15 min, these islands grow in size and eventually merge together to give a complete coverage. Absorption in near-IR range also increases from 9% to 27%, suggesting a large increase in layer count. Moreover, the 15-min sample shows a sheet resistivity of 1.3 kohm sq.sup.1, which makes the directly grown PECVD graphene an ideal candidate for transparent conductive electrodes. Furthermore, in spite of the large number of layers, both SiO.sub.2 and quartz samples show very high 2D-to-G peak intensity ratios and small 2D peak-widths in their Raman spectra. The multi-layer graphene grown on non-catalytic substrates is turbostratic (random angle of orientation between layers), thereby resulting in the high 2D-to-G ratio.

[0049] In order to test the efficacy of the process according to the present disclosure with catalytic interfaces, substantially identical growth processes were simultaneously carried out on Cu films to study the impact of traditional metal catalysts. In order to verify complete coverage of graphene, Cu samples were heated at 180 C. in ambient conditions for 30 min. Graphene, which acts as a passivation barrier, prevents the oxidation of Cu. The heating test was performed on all Cu samples and it was found that even 5 min of growth prevents Cu oxidation, indicating complete coverage of graphene on Cu. This result serves as a clear evidence of faster graphene growth on Cu than on quartz and, by extension, on SiO.sub.2 as well. Referring to FIG. 5a, a Raman spectroscopy output of intensity vs. Raman shift measured in cm.sup.1 is provided for different growth periods of graphene on Cu for 1 min, 5 min, 10 min, and 15 min. Measurements were acquired directly on the substrate to avoid any defects induced by graphene transfer. It can be seen that even 1 min of growth on Cu shows D and G peaks, albeit with low intensity because of the incomplete growth. Referring to FIG. 5b, a TEM image of the 15-minute trial for graphene grown on Cu is provided. As seen in FIG. 5b, provides more than 20 layers after about 15-min PECVD growth, according to the present disclosure to be contrasted with the case of Cu in traditional CVD, which is generally self-limiting for thermal CVD. Consequently, the results according to the present disclosure prove that PECVD growth is not self-limiting and the number of layers can be tuned by simply changing the duration of growth, an advantage of the process according the present disclosure over the thermal CVD provided in the prior art.

[0050] To further understand the improved graphene deposition on Cu, XRD measurements carried out on the Cu sample just after about 1-min growth revealed that Cu's crystal orientation is mostly (111) even during the initial stages. Referring to FIG. 5c, an XRD output comparing intensity vs angle measured in degrees for Cu on Si and Si is provided. As a result, enhanced adsorption and high diffusion of carbon species on Cu (111) is a major factor for the high growth rate of graphene when compared to arbitrary substrates, e.g. Si. The high D-peak observed in the Raman spectra of the samples discussed above signifies the nano-crystalline nature of PECVD graphene which is a direct consequence of the high carbon flux in CH.sub.4 plasma and the low growth temperature, thus leading to high nucleation density and small grain sizes. This is followed by further adsorption of carbon species on the host substrate that can attach to the edges of the islands or direct adsorption on top of graphene islands to induce multi-layer growth. Once the substrate is entirely covered with graphene, interaction of incoming carbon species with substrate vanishes which leads to these incoming carbon species being directly adsorbed on existing graphene layers and facilitating growth of top layers.

[0051] Those having ordinary skill in the art will recognize that numerous modifications can be made to the specific implementations described above. The implementations should not be limited to the particular limitations described. Other implementations may be possible.