Inverted fireball deposition of carbon films with extremely low surface roughness

Abstract

The surface of carbon films deposited with inverted plasma fireballs is analysed in this paper. Measurements were conducted with Raman spectroscopy, atomic force microscopy and nanoindentation. The latter was used to obtain Young's modulus as well as Martens and Vickers hardness. The roughness of the film was measured by atomic force microscopy and its thickness was measured. It was shown with Raman spectroscopy that the films are homogeneous in terms of atomic composition and layer thickness over an area of about 125 × 125 mm. Furthermore, it was demonstrated that inverted plasma fireballs are a viable tool for obtaining homogeneous, large-area carbon films with rapid growth and very little energy consumption. The obtained films show very low roughness.

Full text

Vol.:(0123456789)
1 3
Carbon Letters
https://doi.org/10.1007/s42823-022-00424-9
ORIGINAL ARTICLE
Inverted fireball deposition ofcarbon films withextremely low surface
roughness
J.Gruenwald1 · G.Eichenhofer2· G.Filipič3· Ž.Federl3· W.Feuchtenberger5· K.Panos5· G.HernándezRodríguez4·
A.M.Coclite4
Received: 1 August 2022 / Revised: 26 September 2022 / Accepted: 10 October 2022
© The Author(s), under exclusive licence to Korean Carbon Society 2022
Abstract
The surface of carbon films deposited with inverted plasma fireballs is analysed in this paper. Measurements were conducted
with Raman spectroscopy, atomic force microscopy and nanoindentation. The latter was used to obtain Young’s modulus as
well as Martens and Vickers hardness. The roughness of the film was measured by atomic force microscopy and its thickness
was measured. It was shown with Raman spectroscopy that the films are homogeneous in terms of atomic composition and
layer thickness over an area of about 125 × 125mm. Furthermore, it was demonstrated that inverted plasma fireballs are a
viable tool for obtaining homogeneous, large area carbon films with rapid growth and very little energy consumption. The
obtained films show very low roughness.
Keywords PECVD· Carbon· Inverted fireballs· Deposition
1 Introduction
Various carbon containing films, ranging from graphene over
amorphous carbon and nanostructured surfaces to diamond
like carbon layers have gained more and more momentum in
recent years. This is mostly due to their vast number of prop-
erties that can be tailored by fine tuning the kind of carbon
allotrope deposited onto different substrates. Amongst many
possible technologies for carbon film deposition, plasma
enhanced chemical vapour deposition (PECVD) is a stand-
ard procedure. It has successfully been used to obtain carbon
nano structures [1, 2], graphene [3, 4], diamond like carbon
(DLC) [5, 6], diamond [7, 8] or amorphous carbon films [9,
10]. In recent years, a novel PECVD process, which is based
on so-called inverted fireballs (IFBs) has gained increased
attention, since it allows the creation of very homogeneous
plasmas with enhanced plasma densities [11–15]. There have
also been attempts to incorporate an IFB setup into a sput-
ter chamber but so far with mixed results [16]. The favour-
able combination of high plasma density and homogeneity
is obtained by a highly transparent grid that is put onto a
suitable positive potential with respect to ground. If the bias
becomes larger than the ionisation potential of the working
gas, electrons are attracted from the surrounding background
plasma and accelerated towards the gridded anode. Most of
these highly energetic electrons will pass through the grid
and induce additional ionisation processes in the closed wire
cage. At the same time, the Debye sheath around the mesh
wires overlaps with the sheath of the neighbouring wires
and, thus, forms an equipotential surface around the entire
gridded anode. This leads to a formation of a Faraday cage,
in which the denser plasma is trapped and forms the IFB.
While the ions are trapped in the potential well that forms
inside the mesh cage, electrons can escape through the grid
or passing between two opposite sides of the cage. If this
happens, the escaping electrons are influenced by the electric
field of the mesh as soon as they leave the IFB. This field
accelerates them back into the IFB, artificially prolonging
Online ISSN 2233-4998
Print ISSN 1976-4251
* J. Gruenwald
jgruenwald@g-labs.eu
1 Gruenwald Laboratories GmbH, Taxberg 50,
5660Taxenbach, Austria
2 4A-PLASMA, Aichtalstraße 66, 71088Holzgerlingen,
Germany
3 Jozef Stefan Institute, Jamova Cesta 39, 1000Ljubljana,
Slovenia
4 Institut Für Festkörperphysik, Graz University
ofTechnology, NAWI Graz, Petersgasse 16, 8010Graz,
Austria
5 Institut Für Elektronik Und Messtechnik, Helmut Fischer
GmbH, Industriestraße 21, 71069Sindelfingen, Germany

Carbon Letters
1 3
their collision mean free path and enhancing their kinetic
energy. Therefore, those electrons trigger additional ionising
collisions within the IFB, enhancing the plasma density by
up to an order of magnitude. Although the plasma potential
is essentially flat throughout an IFB, it was experimentally
demonstrated in reference [15] that the potential profile can
be tailored to a certain extent, if one or more sides of the
cage are electrically insulated against the others and biassed
slightly different than the main cage anode [15]. This may
be used in setups where a gradient in the plasma potential
and, hence, in the difference of ion and electron densities
is desirable. IFBs are a novel technology, which offers the
possibility of homogeneous depositions over large areas.
Furthermore, the growth rates are quite high even with low
additional power consumption by the IFB electrode. Nev-
ertheless, the focus of this paper lies on a standard IFB in
a mixture of argon and n-hexane and the films that were
deposited with this method. After the successful deposition
of the carbon films, the physical properties of the layers were
studied with different methods, which will be discussed in
detail in the next section.
2 Materials andmethods
The carbon films were deposited in a vacuum chamber,
type PA 100 from Plasma Technology GmbH at a working
pressure of 3Pa. The chamber was evacuated by a turbo
pump. The background plasma was maintained at 40kHz,
50W. For applying the DC Voltage to the grid, a power
supply from EDF electronics was used. The IFB anode was
manufactured from a stainless steel grid with (wire diameter
120µm, grid constant 250µm). It had a height of 8.5cm and
an edge length of 18.5cm and consumed 20W of electrical
power at a voltage of 174V. The substrate was not addition-
ally heated and remained at room temperature due to the lit-
tle overall power. The IFB setup and the plasma parameters
are shown in detail in reference [17]. A schematic diagram
of the electrical setup of the IFB cage is shown in Fig.1.
The deposition was conducted on standard labora-
tory soda-lime glass slides (as used for microscopy) with
25 × 75 mm2 area. 7 of these glass plates were covering
the substrate holder. An additional slide was put on top of
the others to cover the glass substrate and have a reference
of uncovered glass samples after the deposition. Before
the coating process an IFB was ignited in a mixture of
100sccm H2 and 7.5sccm Ar to clean the substrates. This
cleaning process was upheld for 5min. After the initial
cleaning, a second IFB plasma was ignited in a mixture
of 50sccm Ar and 50sccm tetramethylsilane (TMS) for
1min to obtain an adhesive mediation layer between the
glass substrate and the carbon film. After this, the coating
process was conducted in an Ar/n-hexane IFB (with 100
sccm each) for 15min. The result of this coating experi-
ment is depicted in Fig.2.
It is evident that a smooth carbon containing film was
obtained within the IFB grid. After the coating procedure,
Fig. 1 Schematic diagram of the IFB setup with electrical connections (left) and a photograph of the anode grid on top of the substrate holder
(right, taken from reference [17].)
Fig. 2 Samples after 15 min of IFB deposition. The substrates 1–8
were glass slides with 25 × 75 mm2 area. The slight dislocation of
slide one was due to mechanical vibrations that occurred when open-
ing the chamber

Carbon Letters
1 3
the samples were analysed with different methods that are
outlined in the next section.
3 Results
3.1 Raman measurements
Raman spectroscopy was performed on the two different
areas, shown in Fig.3. On the uncoated glass sample (red
dot) and the carbon film itself (pink dot). The transition
region (green dot) was also investigated with AFM. These
results will be presented below. The samples were ana-
lysed with a Witec Raman spectroscope using a green laser
with 532nm and 4mW output power on the glass substrate
to get Raman peaks with reasonable intensity. However,
the power of the laser was reduced to 1.6mW on the car-
bon films to prevent possible damages. This is still enough
due to the increased Raman sensitivity of carbon films.
Each sample was measured at least at two points in each
of the relevant regions—glass and carbon. The obtained
data are displayed in Fig.4. Since the most interesting
region is the carbon film, Fig.4b depicts the collected data
from random samples (2, 3, 4 and 8 in this case) to illus-
trate that the Raman peaks are indeed identical, regardless
of the substrate’s position within the IFB.
The peaks in Fig.4a can be identified as the symme-
try stretching mode of Si–O bonds from the glass slide
(570 cm−1) [18], a high frequency Si–O-Si binding mode
(794 cm−1) [19] and the so-called Q3 Si–O stretching
vibration mode (1098 cm−1) [20]. With increasing thick-
ness, the Raman peaks from the glass substrate are van-
ishing and only one broad peak remains. This peak was
deconvoluted into two separate peaks at 1383 cm−1 and
1546 cm−1, respectively. The former is as assumed to be
the stretching mode of the CH3 alkyl group while the lat-
ter is attributed to the aromatic Si–C vibration mode [21].
Since the peak position and intensities are identical on
all the measured samples, it can be concluded that the
chemical composition of the films is very homogeneous
throughout the entire substrate and, thus, throughout the
entire plasma inside the IFB. This assumption is also cor-
roborated by the measurements in reference [17].
Fig. 3 The points of measurements for the Raman and AFM data
acquisition
Fig. 4 Raman spectra of the glass substrate a and the carbon film
(b). Image c shows a magnified view of the deconvoluted peak at
1546 cm−1

Carbon Letters
1 3
3.2 AFM measurements
To study the surface morphology, AFM measurements
were performed on the glass and on the carbon coating.
The results are shown in Fig.5.
The uncoated glass as well as the transition region was
analysed on a 3 × 3µm2 area while the carbon coating was
investigated on a 7 × 7µm2 area. It can be seen that the high-
est elevation on the surface is about 22nm with rms rough-
ness of 2.349nm, while the rms roughness on the transi-
tion region decreases down to 0.553nm with the highest
elevations at around 11nm. However, on the carbon coating
the roughness decreases further below the resolution of the
AFM (< 0.25nm). Approaching the resolution limit of the
AFM also causes the artefact at the end of the measure-
ment in Fig.5c. The same results were obtained on all the
samples, indicating again that the morphology of the carbon
layer produced with IFB shows a very smooth and homo-
geneous surface throughout the whole IFB plasma region.
3.3 Nanoindentation measurements
Nanoindention or instrumented indentation testing was per-
formed according to DIN EN ISO 14577, which provides
Fig. 5 AFM images and surface map of the glass substrate (a), the transition region (b) and the carbon film (c)

Carbon Letters
1 3
access to information about the mechanical properties of the
obtained films. The used instrument in these experiments
was a PICODENTOR HM 500 from the Helmut Fischer
GmbH with a Vickers indenter. A maximum force of 100µN
was used as parameter for the standard measurement proce-
dure. The loading and unloading of the force occurred within
3s. In addition, an enhanced stiffness procedure (ESP) with
a maximum force of 100mN applied during 912 s was
utilised to gain further depth-dependent information. The
measured indentation depth h as a function of the applied
force F is depicted in Fig.6.
The Martens hardness (HM) was calculated from the load
curve and is shown as a function of depth in Fig.7.
While the HIT value can be converted into the Vickers
hardness (HV), the elastic properties have been calculated
from the unloading of the lever in accordance with the
method proposed by Oliver and Pharr [22]. This method
yields the so-called indentation modulus EIT.
The force-penetration depth curves (Fig.6) point out
that the deposited layer exhibits a considerable difference in
mechanical properties compared to the glass substrate. The
increase in the curve fraction of the loading demonstrates
a clear difference. Furthermore, the maximum penetration
depth in the coating is higher than in the glass substrate. It
is therefore evident that the coating must have a lower hard-
ness. This was confirmed by the hardness parameter HM
in Fig.7. The graph shows that the hardness of the coating
is not only lower, but also shows a stable or homogeneous
behaviour in depth without any hardness gradient.
Other commonly used hardness properties determined
for the coating, that cannot be represented as a function
of depth in a standard measuring procedure, are HIT of
8.8GPa ± 0.07GPa and the Vickers hardness value of 827
HV ± 7 HV.
In addition to the hardness properties, the elastic proper-
ties were also determined using the characteristic value EIT/
(1-νs2). The mentioned parameter approximately character-
ises the elastic modulus of the coating, without considering
the Poisson’s ratio (νs). Based on these measurements, an
EIT/(1-νs2) of 70.7GPa ± 2.6GPa was determined for the
glass and 47.9GPa ± 0.3GPa for the coating.
When comparing the measurements from the standard
measurement procedures with those from the ESP in Fig.8,
it is clearly evident that this allows an estimation of the layer
thickness. The previously determined characteristic values
from the standard measurement procedure can be seen again
within the measurement uncertainty. Only the characteristic
values EIT/(1-νs2) and HM were shown for clarity.
For both considered characteristic values, a comparable
measured value from the standard procedure can be deter-
mined from the diagram in Fig.8. The special feature here
is that a significantly greater penetration depth was run. This
means that, according to the 10% rule, the properties of the
coating are represented by only one point at approx. 28nm.
Fig. 6 Averaged indentation depth h as a function of the applied force
F
Fig. 7 Depth dependent Martens hardness (HM) as a function of the
penetration depth
Fig. 8 Averaged depth-dependent characteristic value of HM and
EIT/(1-νs2), measured with ESP

Carbon Letters
1 3
The hardness value is influenced by the substrate material
at greater depths. This can also be seen in the elastic behav-
iour, so that two curve components with different slopes
can be distinguished. At the turning point of the slopes, a
layer thickness can be estimated, which can be assumed to
be approx. 200–300nm.
For the analysis of the layer homogeneity, an array of
measuring points was taken on a sample with the nanoin-
dentation instrument. A matrix with a measurement area of
12.65mm by 14.2mm was applied to the sample to obtain
an impression of the homogeneity.
The 341 measurement curves obtained in this area were
used to generate a colour-coded plot of various properties of
the measurement matrix, which can be seen in Fig.9.
The thin layer exhibits a satisfactory homogeneity over
the measured sample area in terms of hardness and Young's
modulus, which can be seen from HM and EIT/(1- νs2). No
gradients of the considered properties over a large area of
the sample can be seen. The results are statistically normal-
distributed. This allows the conclusion that a homogeneous
deposition result was obtained in the process.
3.4 Measurements ofthefilm thickness
The evaluation of the coating thickness was also performed
by Helmut Fischer GmbH. These measurements were con-
ducted with a BETASCOPE using a radioactive source C14
at a temperature of 22°C with a measuring time of 100s.
We measured a mean value of 263nm and a standard devia-
tion of 50nm for the coating thickness by assuming 1.395g/
cm3 for the density of the coating.
4 Conclusions
In this paper, the deposition of large area amorphous carbon
coatings with excellent homogeneity by applying IFBs is
reported. The measurements of the mechanical and surface
properties showed that films with low surface roughness
and homogeneous chemical composition were obtained
with growth rates of about 15–20nm/min at very low over-
all input power (approx. 70W). The coating experiments
demonstrate that IFBs are a versatile tool, especially for the
fabrication of large area carbon films, which demand a small
surface roughness and high level of homogeneity. An over-
view of the results from the mechanical analysis and the
relevant statistical parameters from these measurements are
listed in Table1.
The main purpose of this paper is to show that IFB can
produce carbon layers with relevant properties. However,
there will be more work necessary in the future regarding
suitable precursor gases or gas mixtures and the fine tuning
of other physical parameters such substrate temperature, gas
pressure, et cetera which are expected to enhance the number
of varieties for other carbon allotropes on different substrate
materials.
Acknowlegdements The authors want to thank J. Eisenlohr from
Plasma Technology GmbH, 71083 Herrenberg-Gültstein, Germany for
the making the PA 100 plasma system available for the experiments
and his technical support. J. Gruenwald: Member of the GET-Plasma
Consortium, G. Eichenhofer: Member of the GET-Plasma Consortium.
Declarations
The authors declare that they have no conflicts of interest.
Fig. 9 Colour-coded plot of EIT/(1-νs2) (a) and HM (b) for testing
the layer homogeneity
Table 1 Results and statistical parameters of the mechanical proper-
ties of the carbon films and maximum indentation depth hmax
HM EIT/(1-νs2) HIT HV hmax
N/mm2GPa N/mm2nm
X 2966 47.9 8752 827 27
U (X.) 14 0.25 65 7 0.1
S 137 2.4 608 58 0.7
V/% 4.6 4.9 7 7 2,6
Min 2658 42.3 7148 675 25
Max 3451 55.2 11060 1045 29

Carbon Letters
1 3
References
1. Neyts EC (2012) PECVD growth of carbon nanotubes: from
experiment to simulation. J Vac Sci Technol, B: Nanotechnol
Microelectron: Mater, Process, Meas, Phenom 30(3):030803
2. Santhosh NM etal (2020) Oriented carbon nanostructures from
plasma reformed resorcinol-formaldehyde polymer gels for gas
sensor applications. Nanomaterials 10(9):1704
3. Guo L etal (2018) Direct formation of wafer-scale single-layer
graphene films on the rough surface substrate by PECVD. Carbon
129:456–461
4. Hussain S etal (2020) Low-temperature low-power PECVD
synthesis of vertically aligned graphene. Nanotechnology
31(39):395604
5. Wei X etal (2021) Deposition of DLC films on the inner wall
of U-type pipes by hollow cathode PECVD. Diam Relat Mater
114:108308
6. Mayer M etal (2016) Diamond like carbon deposition by inverted
fireballs. Materials Today: Proceedings 3:S184–S189
7. Mesbahi A, Silva F (2017) Study of the influence of gas flow on
PECVD diamond growth: influence of the separate injection of
gases. J Phys D Appl Phys 50(47):475203
8. Taylor A etal (2022) Growth and comparison of high-quality
MW-PECVD grown B doped diamond layers on 118}, {115
and 113 single crystal diamond substrates. Diam Relat Mater
123:108815
9. Sheng R etal (2018) Effect of unbonded hydrogen on amorphous
carbon film deposited by PECVD with annealing treatment. Diam
Relat Mater 81:146–153
10. Cho Y-S etal (2019) Effect of substrate bias on biocompatibility
of amorphous carbon coatings deposited on Ti6Al4V by PECVD.
Surf Coat Technol 357:212–217
11. Gruenwald, J., etal. (2011) “Further experiments on inverted fire-
balls.” Proc. 38th EPS Conf. Plasma Phys, 2011.
12. Gruenwald J, Reynvaan J, Knoll P (2014) Creation and char-
acterization of inverted fireballs in H2 plasma. Phys Scr
2014(T161):014006
13. Knoll P etal (2016) PECVD of carbon by inverted fireballs: from
sputtering, bias enhanced nucleation to deposition. Diam Relat
Mater 65:96–104
14. Gruenwald J, Reynvaan J, Geistlinger P (2018) Basic plasma
parameters and physical properties of inverted He fireballs.
Plasma Sources Sci Technol 27(1):015008
15. Gruenwald J, Reynvaan J, Geistlinger P (2020) Influence of inho-
mogeneous electrode biasing on the plasma parameters of inverted
H2 fireballs. J Technol Sp Plasmas 1(1):1–4
16. Gruenwald J etal (2021) Application and limitations of inverted
fireballs in a magnetron sputter device. Surf Coat Technol
422:127510
17. Eichenhofer G etal (2022) Measurement of inverted n-hexane
fireball properties with a Multipole Resonance Probe. J Technol
Sp Plasmas 3(1):109–117
18. Duan R-G, Kai-Ming L, Shou-Ren G (1998) Effect of Na+ ion
on the structure of 27CaO· 12Al2O3· 61SiO2 glass. Mater Chem
Phys 57(1):92–94
19. Henderson GS etal (1985) Raman spectra of gallium and germa-
nium substituted silicate glasses: variations in intermediate range
order. Am Miner 70(9–10):946–960
20. O’Shaughnessy C etal (2020) The influence of modifier cations
on the Raman stretching modes of Q n species in alkali silicate
glasses. J Am Ceram Soc 103(7):3991–4001
21. González-Henríquez CM, Terraza Claudio A, Sarabia M (2014)
Theoretical and experimental vibrational spectroscopic inves-
tigation of two R1R2-diphenylsilyl-containing monomers
and their optically active derivative polymer. J Phys Chem A
118(7):1175–1184
22. Oliver WC, Pharr GM (1992) An improved technique for deter-
mining hardness and elastic modulus using load and displacement
sensing indentation experiments. J Mater Res 7(6):1564–1583
Publisher's Note Springer Nature remains neutral with regard to
jurisdictional claims in published maps and institutional affiliations.
Springer Nature or its licensor (e.g. a society or other partner) holds
exclusive rights to this article under a publishing agreement with the
author(s) or other rightsholder(s); author self-archiving of the accepted
manuscript version of this article is solely governed by the terms of
such publishing agreement and applicable law.