Atomic layer deposition for device integration of graphene
– A just published review paper

A new paper from our group has just appeared in Advanced Material Interfaces. In this paper I – together with my colleagues, have reviewed the methods explored to achieve uniform atomic layer deposition (ALD) on graphene. The work is part of my PhD project entitled: ‘A physical-analytical approach to atomic layer deposition of ultra-thin high-κ dielectrics on graphene’, which is funded by the Netherlands Foundation for Fundamental Research on Matter (FOM). The project aims at the fabrication of ultrathin (< 10 nm), uniform, pinhole free metals and metal-oxides on graphene. In this blog I would like discuss the importance ALD for the device integration of graphene and related 2D materials.

The promise of Graphene

Let me start with first reiterating the usual story. Graphene is a single atomic layer of sp² carbon atoms arranged in a hexagonal pattern. The unique properties of graphene have attracted interest throughout the research community and elevated graphene’s status to that of a true wonder material. Amongst others, it is the thinnest material discovered to date (only 3.35Å thick), it has a record high charge carrier mobility (above of 200.000 cm²/Vs), it is extremely flexible and stretchable and it is transparent over a wide spectral range (30µm – 280 nm) with an absorption coefficient of only 2.3 %. Applications of graphene are envisioned in many areas ranging from energy storage applications to flexible electronics and from chemical sensors to spintronic and photonic devices [1–3].

How will we integrate graphene in devices?

No matter what will ultimately be the application of graphene, it is important to realize that in the end we will always need to connect it to its 3D surroundings. Graphene therefore needs to be made compatible with existing semiconductor processing technology. Of particular importance for semiconductor device integration are the creation of dielectric interfaces with the graphene for electrostatic control and the formation of metal-graphene contacts for the injection of charge carriers into the graphene [2].  It is important that the intrinsic properties of graphene are not adversely affected by the creation of these interfaces and contacts. One therefore needs deposition techniques that are able to deposit high quality materials, with good thickness control on graphene, while at the same time damage to the graphene is avoided. Not unexpectedly, considering the background of our group, we think that ALD is ideally suited to meet these requirements.

Why ALD?

So why would we want to use ALD for the deposition of materials on graphene? The ALD process is a cyclic process that typically consists of two self-limiting half reactions separated by pump and purge steps. The alternative pulsing and purging of reactants and co-reactants results in a self-limiting process in which film growth is independent of the precursor flux (see Figure 1 below). This makes is possible to deposit high quality materials uniformly over large areas with a precise control of the layer thickness [4]. Compared to other deposition techniques such as e-beam evaporation, sputtering, and pulsed laser deposition, thermal ALD also has the advantage that no highly reactive radicals, ions and photons are present during deposition. Such reactive species can damage the graphene, deteriorating its electrical properties [5]. For the deposition of thin dielectric layers on graphene ALD is therefore the method of choice.

Figure 1: Schematic representation of the ALD process of Al₂O₃ consisting of two self-limiting half reactions. The reactants in the first half-cycle (Al(CH₃)₃ precursor exposure) and second half-cycle (H₂O co-reactant exposure) are self-limiting, i.e. the process stops once all available surface sites have reacted. The precursor and co-reactant dose are separated by pump or purge steps to prevent unwanted reactions between the precursor and co-reactant. At the end of the second half-cycle a surface is obtained identical to the starting surface of the first half cycle. This makes it possible to obtain the desired film thickness by repeating the half-cycles in an ABAB fashion. The resulting coverage, or growth per cycle (GPC), as a function of the exposure and purge time is indicated as well. Care should be taken that the exposure steps and purge steps are sufficiently long such that saturated growth is obtained and the reaction between the precursor and the co-reactant in the gas phase is prevented (CVD-like growth). The schematic shows the atoms oxygen (in red), hydrogen (in white), aluminium (in blue) and carbon (in gray).

The Challenge

The cyclic nature of the ALD process makes it the ideal candidate for the deposition of thin dielectric and metal layers on graphene. However, ALD relies on the presence of reactive surface sites for the precursor molecules to adsorb on a surface. Graphene lacks these reactive sites due to its sp² carbon configuration. All bonds lie in plane and only a weak van der Waals interaction between the graphene and its surroundings exist. As we and other people have found [6–9], the ALD precursor molecules consequently have difficulty adsorbing on the graphene plane, which is illustrated in the figure 2 below. Only on reactive surface sites such as defects, grain boundaries and wrinkles ALD growth is obtained. The lack of reactive surface sites on the graphene plane makes the initiation of ALD growth on graphene challenging. Although this “area-selective ALD” might offer opportunities for catalytic applications [10], most applications require the deposition of uniform films on graphene.

Figure 2: a) Scanning electron microscopy image of graphene on a SiO₂ substrate. The bi-layers spots and wrinkles, which are a result of the graphene growth process are indicated as well. b) Graphene after 100 cycles of Pt ALD. No uniform growth of Pt on the graphene plane is obtained, due to the absence of reactive surface sites. Only on reactive sites such as defects, grain boundaries, and wrinkles Pt growth is obtained.

The solution

To overcome this issue different surface preparation techniques to initialize ALD growth on graphene have been investigated in the literature[11–14]. As outlined in our review paper, I have divided these techniques into three categories as also indicated in Figure 3. Briefly: 1) Tuning the ALD process conditions to facilitate optimum coverage of precursor molecules on the graphene by lowering the deposition temperature, choosing a precursor that interacts more strongly with the graphene and optimizing the dose and purge times; 2) The use of seed-layers, such as self-assembled monolayers, polymers, evaporated metals (which are oxidized in air before doing ALD) and layers deposited by chemical vapor deposition (CVD); 3) The creation of functional groups on the graphene surface by, for example, ozone and plasma treatments.

The overview given in Figure 3 indicates that there are many possibilities to initiate uniform ALD growth on graphene. Each method having its advantages and disadvantages as I have elaborately discussed in the review. In the ideal case one would like to have a method that enables direct uniform ALD growth on graphene without damage to the graphene at a sufficiently high temperature to obtain the best material quality.
Figure 3: Overview of the different approaches to achieve uniform ALD on graphene. These can be divided into three categories: 1) The adjustment of the ALD process, such as lowering the deposition temperature to prevent precursor molecules form desorbing, choosing a precursor with more affinity towards the graphene surface and changing the dose and purge times; 2) The deposition of seed-layers on the graphene such as spin-coating polymers, evaporating metals and metal-oxides; 3) The creation of reactive groups on the graphene surface by plasma treatments, ozone and NO₂ exposure or wet chemical treatments [12].

What can we improve further?

Despite the significant advances in the growth of materials on graphene by ALD, further improvements are desired. Currently, to avoid damage to the graphene, many of the ALD processes used to achieve growth on graphene use a relatively low deposition temperature. This results in the deposition of material with less than optimal properties. Furthermore the surface functionalization approaches tend to disrupt the sp² carbon bonding of the graphene, by converting it to sp³ carbon, deteriorating the graphene’s electrical properties. The use of reversible functionalization approaches where the graphene is restored to its pristine sp² carbon bonded form after ALD would circumvent this problem. An example of such an approach, that we recently published in Chemistry of Materials, uses a H₂ plasma to functionalize the graphene and is shown in Figure 4. Other approaches such as choosing a precursor specifically tailored for the adsorption on graphene or the adjustment of the ALD process parameters also have demonstrated their potential [15,16].

Figure 4:  New approach to deposit Al₂O₃ uniformly on graphene without damage by using a H₂ plasma treatment. The H₂ plasma treatment creates functional groups on the graphene (indicated by the appearance of a Defect-peak (D) in the Raman spectra), which aid in the adsorption of precursor molecules on graphene. Because H₂ is released during the precursor adsorption process the electrical properties of graphene remain unaffected, indicated by the recovery of the mobility of the graphene after ALD. Any hydrogen that remains after the ALD process can be removed by heating the graphene to 400°C [17].

When can we expect the introduction of graphene in devices?

This brief overview shows that the combination graphene with ALD processing makes it possible to deposit thin dielectric and metallic layers on graphene as needed for the integration of graphene in devices. So when can we expect devices with graphene to become available? The short answer is “This might take a while” [1]. As for other new promising technologies it is very difficult to replace existing process technologies. It is therefore much more likely that new application areas where no existing technology exist will be the first to introduce graphene. Some of these applications might include the combination of graphene with other 2D materials, such as MoS₂, which in contrast to graphene do have a band gap. By the way, the challenges faced to make the integration of other 2D materials possible are very similar to those of graphene [18]. Therefore, even if graphene won’t be used in the end, I consider the knowledge obtained for graphene will be useful elsewhere. Irrespective of which application will be the first one making use of graphene or other 2D materials, ALD will likely play an important role in making the application possible.

References

1. Ferrari, A.C., Bonaccorso, F., Fal’ko, V., et al. (2015) Science and technology roadmap for graphene, related two-dimensional crystals, and hybrid systems. Nanoscale, 7 (11), 4598–4810.
2. Lemme, M.C., Li, L., Palacios, T., and Schwierz, F. (2014) Two-dimensional materials for electronic applications. MRS Bull., 39 (8), 711–718.
3. Schwierz, F. (2010) Graphene transistors. Nat. Nanotechnol., 5 (7), 487–496.
4. Pakkala, A., and Putkonen, M. (2010) Atomic Layer Deposition, in Handbook of Deposition Technologies for Films and Coatings, Elsevier, pp. 364–391.
5. Ni, Z.H., Wang, H.M., Ma, Y., et al. (2008) Tunable Stress and Controlled Thickness Modification in Graphene by Annealing. ACS Nano, 2 (5), 1033–1039.
6. Karasulu, B., Vervuurt, R.H.J., Kessels, W.M.M., and Bol, A.A. (2016) Continuous and ultrathin platinum films on graphene using atomic layer deposition: a combined computational and experimental study. Nanoscale, 8 (47), 19829–19845.
7. Wang, X., Tabakman, S.M., and Dai, H. (2008) Atomic Layer Deposition of Metal Oxides on Pristine and Functionalized Graphene. J. Am. Chem. Soc., 130 (26), 8152–8153.
8. Xuan, Y., Wu, Y.Q., Shen, T., et al. (2008) Atomic-layer-deposited nanostructures for graphene-based nanoelectronics. Appl. Phys. Lett., 92 (1), 13101.
9. Dlubak, B., Kidambi, P.R., Weatherup, R.S., et al. (2012) Substrate-assisted nucleation of ultra-thin dielectric layers on graphene by atomic layer deposition. Appl. Phys. Lett., 100 (17), 173113.
10. Kim, K., Lee, H.-B.-R., Johnson, R.W., et al. (2014) Selective metal deposition at graphene line defects by atomic layer deposition. Nat. Commun., 5, 4781.
11. Garces, N.Y., Wheeler, V.D., and Gaskill, D.K. (2012) Graphene functionalization and seeding for dielectric deposition and device integration. J. Vac. Sci. Technol. B Microelectron. Nanom. Struct., 30 (3), 30801.
12. Vervuurt, R.H.J., Kessels, W.M.M.E., and Bol, A.A. (2017) Atomic Layer Deposition for Graphene Device Integration. Adv. Mater. Interfaces, 1700232, 1700232.
13. Kim, J., and Jandhyala, S. (2013) Atomic layer deposition of dielectrics for carbon-based electronics. Thin Solid Films, 546, 85–93.
14. Liao, L., and Duan, X. (2010) Graphene–dielectric integration for graphene transistors. Mater. Sci. Eng. R Reports, 70 (3–6), 354–370.
15. Aria, A.I., Nakanishi, K., Xiao, L., et al. (2016) Parameter Space of Atomic Layer Deposition of Ultrathin Oxides on Graphene. ACS Appl. Mater. Interfaces, 8 (44), 30564–30575.
16. Oh, I.-K., Tanskanen, J., Jung, H., et al. (2015) Nucleation and Growth of the HfO 2 Dielectric Layer for Graphene-Based Devices. Chem. Mater., 27 (17), 5868–5877.
17. Vervuurt, R.H.J., Karasulu, B., Verheijen, M.A., et al. (2017) Uniform Atomic Layer Deposition of Al2O3 on Graphene by Reversible Hydrogen Plasma Functionalization. Chem. Mater., 29 (5), 2090–2100.
18. Kim, H.G., and Lee, H.-B.-R. (2017) Atomic Layer Deposition on 2D Materials. Chem. Mater., 29 (9), 3809–3826.