By combining mask-less lithography and chemical vapor deposition (CVD) techniques, a novel two-stage diamond anvil has been fabricated. A nanocrystalline diamond (NCD) micro-anvil 30μm in diameter was grown at the center of a [100]-oriented, diamond anvil by utilizing microwave plasma CVD method. The NCD micro-anvil has a diamond grain size of 115 nm and micro-focused Raman and X-ray Photoelectron spectroscopy analysis indicate sp3-bonded diamond content of 72%. These CVD grown NCD micro-anvils were tested in an opposed anvil configuration and the transition metals osmium and tungsten were compressed to high pressures of 264 GPa in a diamond anvil cell.

Traditional diamond anvils in a diamond anvil cell (DAC) setting have been shown to generate pressures up to 416 GPa.1 Recent advances have made it possible to generate much higher pressures up to 1065 GPa in DACs by utilizing diamond anvils which have ‘second stage’ pressure generators on the culets of primary diamond anvils.2,3 These second stages have been developed by various methods. Dubrovinskaia et. al.2 have taken the approach of synthesizing nanodiamond microballs from glassy carbon under high-pressures and high-temperatures and employing them as the second stage element. Sakai et. al.3 have fabricated their second stage by utilizing focused ion beam technology from both single crystal diamond as well as nano-polycrystalline diamond.

An inherent challenge in these previous approaches is that the micro-anvils have to be secured on the culet of diamond anvil or placed in the sample chamber assembly using micro manipulators. Deviation from precise DAC preparation parameters such as the placements of second stages on the primary anvil, the size of the sample itself could cause the second stages to move and go out of alignment leading to the failure of the experiment. These second stages are not an integral part of the anvils themselves. This could lead to prolonged DAC preparation time, failures during the experiments due to inadvertent movement of the second stage, etc. To eliminate these issues, we have developed novel two-stage diamond micro-anvils where the pressure generating second stage is an integral part of the diamond anvil itself. Our previous study4 showed the feasibility of growth of a homo-epitaxial diamond micro-anvil on a single crystal substrate and static pressure generation of 86 GPa was achieved. However, for generation of ultra-high pressures, a second stage of nanocrystalline diamond (NCD) is preferred due to its higher yield stress under compression as compared to the bulk crystalline diamond.2 It is also well established from extensive literature on NCD coatings on metals that NCD thin-film material has higher fracture toughness and improved mechanical properties compared to their micro-crystalline diamond counterparts.5 We report in this letter on the successful synthesis of a NCD micro-anvil on top of a single crystalline diamond by combining mask-less lithography with microwave plasma CVD and demonstrate the application of CVD grown micro-anvils in ultra-high pressure studies on materials.

The [100]-oriented 0.3 carat single crystal diamond anvils with a central flat size of 300μm in diameter have been selected as substrates for these experiments. The diamond anvils were first covered with a tungsten mask by sputter deposition and a small hole has been set in this tungsten film by a combination of mask-less lithography and wet etching. This step is critical in the fabrication of these two-stage anvils as it determines the location of a second stage pressure generator. Use of mask-less lithography allowed us to place the hole in the tungsten film at the center of the culet with great precision. Chemical vapor deposition of NCD has been carried out on these anvils with a gas phase chemistry of 9% CH4/H2, 10% N2/CH4 at a substrate temperature of 830 °C. The tungsten mask which acted as a deterrent for diamond growth elsewhere on the diamond anvil has been etched away and the resulting diamond anvils showed a distinct second stage on the culet of the primary diamond anvil as shown in Figure 1a. The diameter of the NCD micro-anvil shown at the center of primary anvil in Figure 1a is 30μm. Figure 1b shows the high-resolution scanning electron microscope (SEM) image of the CVD grown second stage. We measured an average grain size to be 115 nm by SEM. Micro-focused X-ray photoelectron spectroscopy (XPS) and Raman spectroscopy have been employed to further characterize the sp3-diamond content in the NCD micro-anvil.

FIG. 1.

(a) Scanning electron microscopy image of two-stage NCD micro-anvil. The diamond central flat size is 300 microns in diameter and the NCD micro-anvil is 30 microns in diameter. (b) A higher magnification image of NCD micro-anvil showing an average diamond grain size of 115 nm.

FIG. 1.

(a) Scanning electron microscopy image of two-stage NCD micro-anvil. The diamond central flat size is 300 microns in diameter and the NCD micro-anvil is 30 microns in diameter. (b) A higher magnification image of NCD micro-anvil showing an average diamond grain size of 115 nm.

Close modal

Diamond anvil cells (DACs) have been prepared with these two-stage diamonds and two different types of sample geometry have been tested. In the first cell, a tungsten metallic gasket of 30μm thickness has been used. This tungsten gasket had no hole drilled in it and was directly compressed between the two NCD micro-anvils. In another cell, a stainless steel gasket which has been pre-indented to a thickness of 40μm was employed. A hole of 75μm in diameter has been drilled in this gasket and was filled with a ground mixture of the transition metals osmium and platinum and compressed between the two NCD micro-anvils. Structural studies were carried out on both DACs at beamline 16-ID-B, HPCAT at the Advanced Photon Source of Argonne National Laboratory. The image plate angle dispersive x-ray diffraction patterns were recorded with a monochromatic beam with a wavelength λ=0.4066Å and the beam was focused to a nominal spot size of 1μm x 2μm.

Raman spectroscopy revealed that the CVD grown second stage diamond is typical of NCD coatings that are deposited on metals.5 We have carried out an extensive study on the growth and characterization of NCD micro-anvils. A total of ten-anvils were fabricated to confirm the reproducibility of the growth process. As evidenced from Figure 1a and 1b, a NCD micro-anvil that is well centered on the single crystalline substrate can be grown and the measured average grain size of the NSD is consistently at 115 nm for the diamond growth chemistry and the substrate temperature conditions described previously.

Figure 2 shows the Raman spectrum from the NCD micro-anvil. It shows a barely detectable characteristic diamond peak at 1332 cm–1, and a Raman band associated with the amorphous carbon centered at 1550 cm–1. There are additional Raman peaks at 1150 cm–1 and 1480 cm–1, which are associated with sp2 bonded carbon in the form of TPA—transpolyacetylene.6 It is difficult to obtain a quantitative estimate of sp2/sp3-bonded carbon content from the Raman spectroscopy alone and hence further characterization has been carried out on the NCD micro-anvils by micro-focused XPS.

FIG. 2.

Raman spectrum from the CVD grown second stage. The characteristic peaks from diamond, sp2 bonded amorphous carbon are present.

FIG. 2.

Raman spectrum from the CVD grown second stage. The characteristic peaks from diamond, sp2 bonded amorphous carbon are present.

Close modal

The XPS analysis was performed on a Phi Electronics, Inc. Versaprobe 5000 with monochromatic X-rays (AlKα1=1486.6eV). These are focused such that the spot size was 10μm with a power of 1.25W. The vacuum level in the instrument prior to ion gun activation was 1x10-8 Pa. The pass energy used for the high resolution scan mode was 23.5 eV. The instrument is equipped with a dual charge neutralization mechanism that sends a flood of electrons (1.3 eV) over the surface and simultaneously bombards the X-ray spot with low energy (15 eV) Ar+ ions. The instrument is regularly calibrated to Ag, Au, and Cu foils. Peak-fitting and background subtraction were performed with the Ph Electronics, Inc. Multipak software using Gaussian-Lorentzian peak shapes and iterated-Shirley background subtraction. For comparison purposes highly oriented pyrolytic graphite—which contains only sp2 bonded carbon—was found to be at 284.25 eV. High resolution (0.1 eV step size) spectra were taken on the second stage and the calculated peaks for sp2 and sp3 bonded carbon were found to be at 283.51 eV and 284.54 eV, respectively as shown in Figure 3. These peaks are shifted from their typically quoted values but with the charge neutralization system in use it is difficult to precisely locate carbon peaks. The separation of the two peaks, however, is known to be approximately 1 eV and the FWHM of the peaks for this system at this pass energy are 1.1 eV for sp2 carbon and 1.45 eV for sp3 carbon. The ratios then of the peaks were determined by holding the FWHM and approximate spacing constant and fitting the height and positions only. Using this method, sp3-bonded diamond content was determined to be 72 % and the remaining sp2-bonded carbon content was 28 %.

FIG. 3.

X-ray photoelectron spectrum of second stage showing the deconvolution of measured spectra in to sp3 and sp2 bonded carbon contributions. The sp3 bonded carbon makes up 72 % of the CVD grown material.

FIG. 3.

X-ray photoelectron spectrum of second stage showing the deconvolution of measured spectra in to sp3 and sp2 bonded carbon contributions. The sp3 bonded carbon makes up 72 % of the CVD grown material.

Close modal

A total of four two-stage diamond anvils were selected for high pressure experiments and two diamond anvil cells were prepared with the gasket geometries as described earlier. In DAC 1, we have compressed a 30μm thick tungsten gasket between the two-stage NCD micro-anvils. The highest pressure achieved in this experiment was 264 GPa, calculated from the equation of state of tungsten obtained from Ref. 7. The measured lattice parameter of tungsten at a pressure of 264 GPa is 2.789 Å and the (111), (200), and (211) diffraction peaks are observed from the body centered cubic phase of tungsten. Figure 4 shows the x-ray transmission profile from this cell indicating the integrity of the NCD micro-anvil as evidenced by the increased x-ray transmission through the thin part of the sample compressed between the two micro-anvils. This clearly demonstrates that the NCD micro-anvil does not show any evidence of plastic deformation up to the highest pressures achieved in this study. The full width at half maximum (FWHM) of the x-transmission scan in Figure 4 corresponds exactly to the diameter of 30 microns of the NCD micro-anvil.

FIG. 4.

X-ray transmission profile of the sample in DAC 1 as a function of displacement in millimeters with a total scan width of 100μm. The highest pressure at this stage in the sample was 230 GPa. The full-width at half maximum (FWHM) of the transmission scans correspond to the size of NCD micro-anvil of 30μm in diameter to the highest pressure thereby demonstrating the structural integrity to extreme pressures.

FIG. 4.

X-ray transmission profile of the sample in DAC 1 as a function of displacement in millimeters with a total scan width of 100μm. The highest pressure at this stage in the sample was 230 GPa. The full-width at half maximum (FWHM) of the transmission scans correspond to the size of NCD micro-anvil of 30μm in diameter to the highest pressure thereby demonstrating the structural integrity to extreme pressures.

Close modal

In DAC 2, osmium and platinum mixture was loaded into a drilled hole in a 40μm thick stainless steel gasket. In this experiment the highest pressure reached was 253 GPa as calculated from the osmium equation of state from Dubrovinsky et. al.8 The measured lattice parameters for osmium at this pressure are a = 2.438 Å, c = 3.924 Å and the (100), (002), (101), (102) and (110) diffraction peaks are observed from the hexagonal close packed phase of osmium. No platinum signal was detected in the highest pressure region; this is most likely due to inhomogeneity in the sample mixture. Figure 5 shows the integrated x-ray diffraction patterns of tungsten and osmium from both experiments at the highest pressures recorded. Figure 5 also clearly demonstrates the existence of large pressure gradients as diffraction peaks from low pressure regions marked by asterisks in Figure 5 were simultaneously recorded with the diffraction from the high pressure region. The pressure difference between the high pressure and low pressure regions was measured to be as large as 150 GPa in each spectrum shown in Figure 5 and our study confirms the ability of NCD micro-anvils to support large pressure gradients or shear stresses.

FIG. 5.

Integrated X-ray diffraction spectra of tungsten (lower-panel) and osmium (upper-panel) at the highest measured pressures from our experiments. The tungsten is in the body centered cubic phase while the osmium is in the hexagonal close packed phase. The peaks marked with asterisk are from the low pressure regions of tungsten and osmium.

FIG. 5.

Integrated X-ray diffraction spectra of tungsten (lower-panel) and osmium (upper-panel) at the highest measured pressures from our experiments. The tungsten is in the body centered cubic phase while the osmium is in the hexagonal close packed phase. The peaks marked with asterisk are from the low pressure regions of tungsten and osmium.

Close modal

Our future experiments will focus on further optimizing NCD micro-anvils. We have utilized the CVD-grown NCD second stage pressure generators without any further modification after deposition. We plan to machine these second stage pressure generators using focused ion beam method to improve their pressure generating capability. This will be achieved by reducing the contact area of the NCD micro-anvils by shaping them into hemispheres. Additional optimization in pressure generating capabilities can be achieved by modifications in the sample chamber designs, changing the geometry of the starting diamond anvil as well as reducing the diamond grain size in the NCD micro-anvils.

In conclusion we have successfully grown NCD micro-anvils on a single crystal diamond substrate using a combination of mask-less lithography and microwave plasma chemical vapor deposition techniques. Very well centered and reproducible NCD micro-anvils of 30μm in diameter were grown on [100]-oriented single crystal diamond anvil with an average diamond grain size of 115 nm. The sp3-bonded diamond content of NCD micro-anvil was measured to be 72%. These NCD micro-anvils were successfully tested in two separate high pressure diamond anvil cell experiments to pressures up to 264 GPa using synchrotron x-ray diffraction studies on osmium and tungsten metals. The measured x-ray pressure profiles indicate that the NCD micro-anvil can support high shear stresses and do not show any evidence of plastic deformation to the highest pressures. Future studies would focus on reducing the diamond grain size as well as focus ion beam machining of NCD micro-anvil stage to a hemi-spherical shape.

We acknowledge support from the National Science Foundation (NSF) under Grant Number DMR-1608682. Samuel Moore is supported by the Carnegie DOE Alliance Center (CDAC) under grant number DE-NA002006. Portions of this work were performed at HPCAT (Sector 16), Advanced Photon Source (APS), Argonne National Laboratory. HPCAT operations are supported by DOE-NNSA under Award No. DE-NA0001974 and DOE-BES under Award No. DE-FG02-99ER45775, with partial instrumentation funding by NSF. APS is supported by DOE-BES, under Contract No. DE-AC02-06CH11357. We acknowledge valuable assistance from Dr. Jesse Smith from HPCAT at APS.

1.
A. L.
Ruoff
,
H.
Xia
,
H.
Luo
, and
Y. K.
Vohra
,
Rev Sci Instrum.
61
,
3830
(
1990
).
2.
N.
Dubrovinskaia
,
L.
Dubrovinsky
,
N. A.
Solopova
,
A.
Abakumov
,
S.
Turner
,
M.
Hanfland
,
E.
Bykova
,
M.
Bykov
,
C.
Prescher
,
V. B.
Prakapenka
,
S.
Petitgirard
,
I.
Chuvashova
,
B.
Gasharova
,
Y.
Mathis
,
P.
Ershov
,
I.
Snigireva
, and
A.
Snigirev
,
Sci. Adv.
2
,
e1600341
(
2016
).
3.
T.
Sakai
,
T.
Yagi
,
H.
Ohfuji
,
T.
Irifune
,
Y.
Ohishi
,
N.
Hirao
,
Y.
Suzuki
,
Y.
Kuroda
,
T.
Asakawa
, and
T.
Kanemura
,
Rev Sci Instrum.
86
,
033905
(
2015
).
4.
Y. K.
Vohra
,
G. K.
Samudrala
,
S. L.
Moore
,
J. M.
Montgomery
,
G. M.
Tsoi
, and
N.
Velisavljevic
,
High Pressure Research
35
,
288
(
2015
).
5.
S. A.
Catledge
,
V.
Thomas
, and
Y. K.
Vohra
, in
Diamond-based Materials for Biomedical Applications
, edited by
R.
Narayan
(
Published by Woodhead Publishing Limited
,
2013
), pp.
105
150
.
6.
A. C.
Ferrari
and
J.
Robertson
,
Phil. Trans. R. Soc. A
362
,
2477
(
2004
).
7.
T. S.
Sokolova
,
P. I.
Dorogokupets
, and
K. D.
Litasov
,
Russ.Geol. Geophys
54
,
181
(
2013
).
8.
L.
Dubrovinsky
,
N.
Dubrovinskaia
,
E.
Bykova
,
M.
Bykov
,
V.
Prakapenka
,
C.
Prescher
,
K.
Glazyrin
,
H.-P.
Liermann
,
M.
Hanfland
,
M.
Ekholm
,
Q.
Feng
,
L. V.
Pourovskii
,
M. I.
Katsnelson
,
J. M.
Wills
, and
I. A.
Abrikosov
,
Nature
525
,
226
(
2015
).