Synthetic CVD Diamond
The present invention relates to a method for synthesizing synthetic CVD diamond
material and high quality synthetic CVD diamond material.
Diamond has a high hardness, good abrasion resistance, a low compressibility and a low
coefficient of thermal expansion. Diamond can also have a very high coefficient of
thermal conductivity and it can be an extremely good electrical insulator. This makes
diamond a desirable material for many applications. For example, by making use of its
10 thermal conductivity, diamond can be an excellent heat spreader material for electronic
devices.
15
In certain electronic devices, the ability to dope diamond with nitrogen is important in
order to pin the Fermi Level.
Synthetic diamond material synthesized using high pressure-high temperature (HPHT)
synthesis techniques typically contains significant concentrations of nitrogen impurities,
particularly single substitutional nitrogen (Ns0
), making it yellow. To avoid this, specific
precautions may be taken to exclude nitrogen from the synthesis environment. In
20 addition, diamond material produced using HPHT synthesis techniques exhibits strongly
differential uptake of nitrogen impurity on the surfaces with different crystallographic
orientation (which are the surfaces corresponding to differing growth sectors) that form
during synthesis. The diamond material therefore tends to show zones with different
colours, resulting from the differing nitrogen impurity concentrations in different growth
25 sectors. In addition, it is hard to control the HPHT synthesis process sufficiently to give
a uniform and desired nitrogen concentration throughout even a single growth sector
within the synthesized diamond material. Furthermore, in HPHT synthesis, it is typical
to see impurities resulting from the synthesis process and the catalysts used- examples
would be inclusions comprising cobalt or nickel - features that can result in localised
30 and inhomogeneous strain that degrade the mechanical, optical and thermal properties.
Using a CVD method (such as the process described in US 7,172,655) it is possible to
synthesize diamond material that contains significant concentrations (up to
5
wo 2011/076643 PCT /EP201 0/069828
2
approximately 10 ppm [parts per million]) of N8°, but such diamond is normally brown
in colour. This brown colour is believed to be due to the presence of defects other than
Ns0 incorporated in the material, thought to be caused by the addition of nitrogen to the
CVD synthesis environment.
Intrinsic diamond material has an indirect bandgap of about 5.5 eV and is transparent in
the visible part of the spectrum. Introducing defects, also referred to as "colour centres",
which have associated energy levels within the band gap, gives the diamond a
characteristic colour which is dependent on the type and concentration of the colour
10 centres. This colour can result from absorption or photoluminescence, or a combination
of the two, but generally absorption is the dominant factor. For example, it is well
known that the Ns0 defect causes absorption at the blue end of the visible spectrum
which, by itself, causes the material to have a yellow colour. Similarly, it is known from
Walker (J. Walker, 'Optical Absorption and Luminescence in Diamond', Rep. Prog.
15 Phys., 42 (1979), 1605-1659) that when such yellow material is irradiated with high
energy electrons to create vacancies (sites in the crystal lattice from which carbon atoms
have been displaced), and annealed to cause the migration and eventual trapping of
vacancies at nitrogen impurity atoms, NV centres are formed.
20 EP 0 671 482, US 5,672,395 and US 5,451,430 describe methods of reducing undesired
defect centres in CVD diamond using a HPHT treatment, and US 7,172,655 applies an
annealing technique to reduce the brownness of a single crystal stone. The most
complete removal of brown colour is achieved at annealing temperatures above about
1600°C and generally requires diamond-stabilising pressures. However, such treatment
25 is an expensive and complicated process in which yields can be seriously affected by
cracking of stones etc. Furthermore, due to defect diffusion, such an annealing strategy
is not necessarily consistent with avoiding nitrogen aggregation or with the fabrication of
high performance electronic devices, where controlling the location of defects may be
important. It is therefore considered desirable to be able to directly synthesize diamond
30 material that is not brown but retains the desired high concentration of N s 0 using CVD
methods.
wo 2011/076643 PCT /EP201 0/069828
3
There are many variations of the CVD method for the deposition of diamond which are
now well established and have been described extensively in the patent and other
literature. The method generally involves providing a source gas which, on dissociation
to form a plasma, can provide reactive gas species such as radicals and other reactive
5 species. Dissociation of the source gas is brought about by an energy source such as
microwaves, RF energy, a flame, a hot filament or a jet based technique, and produces
reactive gas species which are allowed to deposit onto a substrate and form diamond.
Most of the recent art is focused on the use of hydrogen-based (H-based) plasmas,
typically comprising H2 with small additions of methane, typically in the range 1 to 10
10 volume% (see, for example, J. Achard et al, "High quality MPACVD diamond single
crystal growth: high microwave power density regime", J. Phys. D: Appl. Phys., 40
(2007), 6175-6188), and oxygen or oxygen-containing species typically at the level of 0
to 3 volume% (for example Chia-Fu Chen and Tsao-Ming Hong, Surf. Coat. Technol.,
54155 (1992), 368-373). Hereinafter, oxygen and oxygen-containing species will be
15 referred to collectively as "0-containing species" and are formed from the 0-containing
sources in the source gas.
Deliberate nitrogen additions into the synthesis environment are known (e.g. Samlenski
et al, Diamond and Related Materials, 5 (1996), 947-951), typically with the purpose of
20 enhancing the growth rate, or improving the quality of the diamond in some other way
e.g. reducing stress and cracking (W02004/046427). In these processes, whilst the
addition of nitrogen into the synthesis environment does introduce some level of
nitrogen into the solid, this is not the primary intent and the overall method is generally
to minimise the inclusion of the nitrogen and the other associated defects in the diamond
25 material that is synthesised. One exception is where the intent is specifically to create
colour centres in the diamond material in the form of nitrogen defects (e.g.
W02004/046427), however such diamond material is of little use in the applications
envisaged here, because of the high defects concentrations other than single
substitutional nitrogen incorporated into the diamond material.
30
W02004/063430 discloses high pressure and microwave power density (MWPD) to be
important for the growth of synthetic CVD diamond material with a low defect
wo 2011/076643 PCT /EP201 0/069828
4
concentration (i.e. what IS generally termed 'high quality' synthetic CVD diamond
material).
Whilst most of the recent art is focused on H-based plasmas comprising little or no 0-
5 containing species, there are also references to the importance of 0-containing species in
the etching of non-diamond carbon, in particular in the context of the synthesis of
polycrystalline diamond by CVD methods (see, for example, Chen et al., Phys. Rev. B,
vol 62 (2000), pages 7581-7586; Yoon-Kee Kim et al., J. Mater. Sci.: Materials in
Electronics, vol 6 (1995), pages 28-33), and in synthesizing "high colour" diamond
10 which is free from the impurities which arise when using H-based plasmas in synthesis
processes runs at similar pressures and powers. Critically, in this area of prior art,
nitrogen incorporated into the diamond material is considered to be one of the defects
which is being minimised, and the methods taught reduce the nitrogen content along
with other defect types (for example WO 20061127611).
15
20
On the basis of the above, there remains a need for a CVD process for producing single
crystal diamond in which the defect content can be controlled and the synthetic CVD
diamond product thereof. There is also a need for high colour (i.e. high quality)
synthetic CVD diamond material per se.
In particular, there is a need for a CVD diamond material, obtained by direct synthesis,
with a relatively high nitrogen content that is uniformly distributed, and which is free of
other defects, such as inclusions, normally associated with HPHT synthesis processes.
Furthermore, there is a need for such CVD diamond material to have a colour which is
25 not dominated by brown defects that do not contain nitrogen, but is instead dominated by
the yellow hue due to the presence of single substitutional nitrogen. Likewise, there is a
need for CVD diamond material where the electronic properties are dominated by single
substitutional nitrogen, but not degraded by the other defects normally resulting from
nitrogen in the growth process.
30
The present invention provides a chemical vapour deposition (CVD) method for
synthesizing diamond material on a substrate in a synthesis environment, said method
compnsmg:
wo 2011/076643
5
providing the substrate;
providing a source gas;
dissociating the source gas; and
PCT /EP201 0/069828
allowing homoepitaxial diamond synthesis on the substrate;
5 wherein the synthesis environment comprises nitrogen at an atomic concentration of
from about 0.4 ppm to about 200 ppm;
10
15
and wherein the source gas comprises:
a) an atomic fraction of hydrogen, Hf, from about 0.40 to about 0.75;
b) an atomic fraction of carbon, Cf, from about 0.15 to about 0.30;
c) an atomic fraction of oxygen, Of, from about 0.13 to about 0.40;
wherein Hf + Cf +Of= 1;
wherein the ratio of atomic fraction of carbon to the atomic fraction of oxygen,
Cf:Of, satisfies the ratio of about 0.45:1 < Cf:Of < about 1.25:1; wherein the
source gas comprises hydrogen atoms added as hydrogen molecules, Hz, at an
atomic fraction of the total number of hydrogen, oxygen and carbon atoms
present of between about 0.05 and about 0.40; and
wherein the atomic fractions Hf, Cf and Of are fractions of the total number of
hydrogen, oxygen and carbon atoms present in the source gas.
20 The source gas comprises hydrogen, oxygen and carbon in the atomic fractions based on
the total number of hydrogen, oxygen and carbon atoms present in the source gas of Hf,
Of and Cf. The hydrogen may be provided by Hz or other sources such as CH4 etc.
The inventors have surprisingly found that by adding hydrogen as Hz molecules to a
25 source gas comprising 0-containing sources (e.g. a CH4/COz source gas), it is possible to
modify the diamond growth conditions to obtain synthetic CVD diamond material with
both a high concentration of nitrogen in the form of Ns0 and a low concentration of other
defects.
30 CVD processes are advantageous as methods of synthesising single crystal diamond
because they offer a more uniform, controllable method of producing synthetic diamond
material. In contrast, HPHT techniques produce material with poorly controlled levels
wo 2011/076643 PCT /EP201 0/069828
6
of N8°, and therefore great variation in concentration of Ns0 between growth sectors, and
material with other impurities and inclusions.
Without wishing to be bound by theory, it is believed that the usable ranges for the
5 synthesis parameters of temperature, pressure and power density for the synthesis of
CVD diamond material are reduced with the addition of 0-containing sources to the
source gas. Plasmas comprising 0-containing species can be less stable and more prone
to the formation of, for example, mono-polar arcs than H-based plasmas, and the
microwave power density and pressures of processes using 0-containing species in the
10 plasma may therefore be forced to be lower than for H-based plasma processes. One
advantage of the addition of some hydrogen, preferably in the form of Hz, into a process
comprising 0-containing species is that it can allow the use of a higher microwave
power density and pressure. In this way, the operating pressure and microwave power
density can both be increased before the synthesis process is disrupted by the tendency to
15 form mono-polar arcs.
Herein, a "H-based plasma" is defined as a plasma comprising hydrogen, oxygen and
carbon atoms in which the atomic fraction of hydrogen atoms (Hf), expressed as a
fraction of the total number of hydrogen, oxygen and carbon atoms in the source gas that
20 forms the plasma, is about 0.80 or greater, alternatively about 0.85 or greater,
alternatively about 0.90 or greater. For example, a source gas composition of 600 seem
Hz, 30 seem CH4, (where "seem" is "standard cubic centimetres" and these are the flows
into the plasma containing chamber) has a hydrogen atomic fraction of
((600 X 2) + (30 X 4))/((600 X 2) + (30 X 5)) = 0.98
25 and such a plasma would be a H-based plasma.
Herein, an "0-based plasma" is defined as a plasma comprising hydrogen, oxygen and
carbon atoms in which the atomic fraction of oxygen atoms (Of), expressed as a fraction
of the total number of hydrogen, oxygen and carbon atoms in the source gas that forms
30 the plasma, is about 0.10 or greater, alternatively about 0.15 or greater, alternatively
about 0.20 or greater and the atomic fraction of hydrogen atoms (Hf) in the source gas
that forms the plasma is about 0.75 or less, alternatively about 0.70 or less, alternatively
about 0.60 or less, alternatively about 0.50 or less. For example a source gas
wo 2011/076643 PCT /EP201 0/069828
7
composition of 290 seem COz, 250 seem CH4, and 230 seem Hz has an oxygen atomic
fraction (Of) of
(290 X 2)/((290 X 2) + (25 X 5) + (230 X 2)) = 0.22
and a hydrogen atomic fraction (Hf) of
5 ((250 X 4) + (230 X 2))/((290 X 2) + (25 X 5) + (230 X 2)) = 0.56
and such a plasma would be an 0-based plasma.
One reason why so much of the recent art relates to H-based plasmas (e.g. plasmas
formed from a source gas mixture that is predominantly Hz) is thought to be that plasmas
10 comprising large fractions of 0-containing species (e.g. plasmas formed from the
dissociation of COz/CH4-based source gases), are more difficult to control at high
pressures and high microwave power densities. Under some circumstances, the
"quality" of diamond synthesized at a particular pressure and power density may be
better (i.e. the defect content may be lower) for a plasma comprising a large fraction of
15 0-containing species than for a H-based plasma under similar pressure and power
density combination, and models generally assume this is based upon the superior
etching ability of 0-containing species. However, the H-based plasma can be run at
significantly higher pressures and powers, which means that whilst the 0-containing
species may etch non-diamond carbon more efficiently "per atom", there are much lower
20 fluxes of the etchant species at the diamond surface than for a higher power and pressure
H-based plasma. These lower fluxes of the etchant species generally cause the defect
content of the product to be higher for material synthesised in an optimised plasma
comprising 0-containing species compared to an optimised H-based plasma process.
25 Furthermore, hydrogen addition alone to COz/CH4-based source gases has previously
been shown to reduce the quality of the diamond material synthesized (Chia-Fu Chen
and Tsao-Ming Hong, "The role of hydrogen in diamond synthesis from carbon dioxidehydrocarbon
gases by microwave plasma chemical vapor deposition", Surface and
Coatings Technology, vol54/55 (1992), pages 368-373).
30
In contrast, the inventors here have surprisingly found that by adding H-containing
sources (in particular by the addition of hydrogen atoms as Hz) to the source gas in a
carefully controlled amount, the pressure at which the CVD synthesis process can be run
wo 2011/076643 PCT /EP201 0/069828
8
stably may be increased (i.e. the presence of hydrogen is believed to increase the
stability of the plasma). This can allow for an increase in the synthesis pressure and
microwave power density, and thus by combining substantial levels of oxygen, with the
power and pressure regime that may be enabled by the presence of appropriate levels of
5 hydrogen, conditions have been found which can enable significant levels of single
substitutional nitrogen to enter the diamond lattice whilst at the same time minimising
the other defects normally associated with nitrogen present in the synthesis environment.
This can result in higher quality synthetic CVD diamond material.
10 Without wishing to be bound by theory, the inventors believe that under the conditions
proposed in the present invention, the increased microwave power density allows more
rapid etching of non-diamond carbon than is usually possible in plasmas comprising 0-
containing species. This rapid etching can arise from the superior etching ability of the
0-containing species which may be substantially enhanced by the higher process
15 pressures. It is not clear, and certainly was not predictable, that this would give the
substantial selectivity in etching found in practice, giving extremely efficient etching of
non-diamond carbon and those defects believed to cause brown colour, but allowing an
unusually high level of nitrogen incorporation in the CVD diamond material. Thus,
uniquely, the material produced by the process of the present invention shows significant
20 absorption associated with single substitutional nitrogen, but very low absorption
associated with brownness arising from defects other than single substitutional nitrogen
thought to be caused by the addition of nitrogen to the CVD synthesis environment.
The present invention further provides synthetic CVD diamond material produced by the
25 method of the present invention.
The present invention further provides synthetic CVD diamond material compnsmg
single substitutional nitrogen (Ns0
) at a concentration of greater than about 0.5 ppm and
having a total integrated absorption in the visible range from 350 nm to 750 nm such that
30 at least about 35 % of the absorption is attributable to Ns0
Preferably the CVD diamond material of the invention is single crystal CVD diamond
material.
wo 2011/076643 PCT /EP201 0/069828
9
The present invention further provides a gemstone comprising synthetic CVD diamond
material of the present invention. The present invention further provides the use of
synthetic CVD diamond material of the present invention as a gemstone. These
5 gemstones are advantageous because the defects present in the synthetic CVD diamond
material of the present invention provide a yellow colour that is similar to the yellow
colour of a yellow-coloured natural diamond (e.g. a so-called "Cape Yellow" diamond),
making the gemstones more attractive to those who have a preference for natural-looking
colours.
10
The present invention further provides an electronic device comprising synthetic CVD
diamond material of the present invention.
The present invention further provides the use of synthetic CVD diamond material of the
15 present invention in an electronic device. These electronic devices are advantageous
because the high and uniform concentration of Ns0 in the substrate upon which the
device is fabricated efficiently pins the Fermi level whilst providing a substrate upon
which subsequent epilayers of intrinsic or doped diamond with low extended defect
density can be deposited.
20
As used herein, where diamond material of a particular colour grade is described, the
colour grade scale referred to is that of the Gemological Institute of America ("GIA'')
wherein "high colour" is a grade of "L" or higher (i.e. grades D, E, F, G, H, I, J, K or L),
GIA grade "D" is the "highest colour", that is the closest to being completely colourless.
25 The meaning of colour within the art of colour-grading of diamond is well understood
and diamonds are graded by the same methods against the same scale in gem grading
laboratories worldwide. The relationship between the perceived colour and the grade is
given in V. Pagel-Theisen, "Diamond Grading ABC The Manual", 9th Edition (2001),
Rubin & Son, Antwerp, Belgium, page 61 (fold-out table).
30
The quantitative measurement of colour in diamond is well established and is described
using the "CIE L *a*b* Chromaticity Coordinate" and its use in diamond is described in
WO 2004/022821. a* and b* are plotted as x andy axes of a graph and the hue angle is
5
10
wo 2011/076643 PCT /EP201 0/069828
10
measured from the positive a* axis towards the positive b* axis. Thus a hue angle of
greater than 90° and less than 180° lies in the upper left quadrant of the a*b* graph. In
this scheme for describing colour, L * is the lightness and a fourth coordinate C* is the
saturation.
In the context of this invention, the term "quality" is used to mean "fitness for purpose"
and therefore a material when used for a specific application is considered to be of
"higher quality" than another material, if it is providing a better or improved technical
solution to the problem being solved.
It is common in the art to refer to the "synthesis" of diamond material as the "growth" of
diamond material. Such terms as "growth rate" and "growth sector" should therefore be
interpreted with this in mind.
15 The synthesis environment compnses the source gas which itself comprises carbon
atoms, hydrogen atoms and oxygen atoms in the form of molecules, atoms, radicals or
ions. There may be other species deliberately added to the synthesis environment (in
particular by adding these species to the source gas) in larger or smaller quantities, such
as inert gases (e.g. helium, neon, argon, krypton, etc.) or nitrogen. In addition, there may
20 be impurities in the source gas mixture.
The proportion of atoms in the synthesis environment that are C, H and 0 may be about
70% or greater, alternatively about 80% or greater, alternatively about 85% or greater,
alternatively about 90% or greater, alternatively about 95% or greater, alternatively
25 about 98% or greater.
The synthesis environment may further comprise one or more inert gases selected from
helium, neon, argon, krypton and xenon. These inert gases may be added to the
synthesis environment by the source gas, i.e. they are present in the source gas. Inert
30 gases may be present in the source gas at an atomic fraction, Xf, of between about 0 and
about 0.5, alternatively between about 0 and about 0.3, alternatively between about 0 and
about 0.2, alternatively between about 0 and about 0.1, alternatively between about 0 and
about 0.05; where Xf + Hf' + Cf' + Of'= 1. The atomic fractions Hf', Cf' and Of' are
wo 2011/076643 PCT /EP201 0/069828
11
fractions of the total number of hydrogen, oxygen, carbon and inert gas atoms present in
the source gas. Clearly, when Xf = 0; Hf = Hf', Cf = Cf' and Of = Of'. Since the inert
gases are inert from a chemical viewpoint, they do not have a role in the chemical
processes that occur in the plasma and can be ignored in this respect. However, the
5 presence of inert gases can affect the physical characteristics of the plasma, such as its
thermal conductivity, or may act as third bodies that can facilitate chemical reactions
between other species without actually participating chemically in the reaction.
Consequently the inventors, without being bound by any particular theory, believe that
the presence of inert gases, in particular Ar, in small quantities, whilst not necessary for
10 the current invention, may have beneficial effects.
The synthesis environment comprises the source gas, wherein the source gas comprises
hydrogen atoms added as hydrogen molecules, H2, at an atomic fraction expressed as a
fraction of the total number of hydrogen, carbon and oxygen atoms present of from
15 about 0.05 to about 0.40, alternatively from about 0.10 to about 0.35, alternatively from
about 0.15 to about 0.30, alternatively from about 0.05 to about 0.10, alternatively from
about 0.10 to about 0.15, alternatively from about 0.15 to about 0.20, alternatively from
about 0.20 to about 0.25, alternatively from about 0.25 to about 0.30, alternatively from
about 0.30 to about 0.35, alternatively from about 0.35 to about 0.40. The remaining
20 hydrogen atoms (i.e. those not from the added H2 molecules) are from other sources such
as CH4 etc.
The hydrogen in the source gas is in the form of hydrogen (H2) or hydrogen-containing
sources (hereinafter collectively referred to as H-containing sources), e.g. H2, CH4 and
25 other hydrocarbon species including hydrocarbons that also contain oxygen such as
aldehydes, ketones etc.
The carbon in the source gas is in the form of carbon-containing sources, e.g. CO
(carbon monoxide), C02 (carbon dioxide), CH4, other hydrocarbons (alkanes such as
30 ethane, propane, butane etc; alkenes such as ethene, propene etc; alkynes such as ethyne,
propyne etc), oxygen containing hydrocarbons such as alcohols, aldehydes, esters,
carboxylic acids, etc.
5
wo 2011/076643 PCT /EP201 0/069828
12
The oxygen in the source gas is in the form of oxygen or oxygen-containing sources
(hereinafter collectively referred to as 0-containing sources), e.g. Oz, 0 3 (ozone), CO,
COz, oxygen-containing hydrocarbons such as alcohols, aldehydes, esters, ethers,
carboxylic acids, etc.
Since the source gas mixture is essentially broken up into its constituent atoms in the
synthesis process and these components will reform into a mixture of species that is, or
is close to, the thermodynamic equilibrium composition for the particular mixture of
atoms, the choice of the molecular species that make up the source gas mixture is
10 dictated by the requirement to achieve a particular plasma composition and the ability to
strike and maintain a stable plasma in the chosen mixture of molecules. The choice of
gases forming the source gas mixture is to some extent further dictated by cost,
availability, purity and ease of handling; for example, CH4 , CO, COz, Hz and Oz are all
readily available as stable, bulk gases in a range of chemical purities and therefore might
15 be preferred.
On dissociation of the source gas, these sources form hydrogen or hydrogen-containing
species (collectively H-containing species), carbon-containing species and oxygen or
oxygen-containing species (collectively 0-containing species) respectively in the
20 plasma. Typically, the dissociated source gas will comprise hydrogen radicals (H"),
carbon monoxide radicals (CO") and two-carbon radical species (e.g. CzHx. where x less
than 6). The presence of these species may be determined by techniques such as optical
emission spectroscopy ("OES").
25 The source gas may be comprised of molecular species such CO, COz, CH4 and Hz such
that the atomic fractions of carbon, Cf, hydrogen, Hf, and oxygen, Of, in the source gas
are in the following ranges:
30
0.15 < cf < 0.30, alternatively 0.18 < cf < 0.28, alternatively 0.20 < cf < 0.25;
0.40 < Hf < 0.75, alternatively 0.42 < Hf < 0.72, alternatively 0.45 < Hf < 0.70;
and
0.13 directions.
The substrate may also have a growth surface with a normal that is within about 10° of
20 the [001] direction, alternatively within about so of the [001] direction, alternatively
within about 4° of the [001] direction, alternatively within about 3° of the [001]
direction. The edges of the substrate may be within about 10° of the <100> directions,
alternatively within about so of the <100> directions, alternatively within about 3° of the
<100> directions. The edges of the substrate may be within about 10° of the <110>
2S directions, alternatively within about so of the <110> directions, alternatively within
about 3° of the <110> directions. The growth surface of the substrate may be
substantially the {001}, {101}, {113}, {311} or {111} surface, and is generally the
{ 001} surface.
30 The following convention is applied to the crystallography of diamond to enable the
growth surface of a substrate to be distinguished from the other surfaces. As used
herein, for a typical substrate having the shape of a rectangular parallelepiped (for which
all faces are nominally part of the { 100} form) with two opposed major and four smaller
5
wo 2011/076643 PCT /EP201 0/069828
15
faces, the opposed major faces are the (001) and (001) faces (collectively referred to as
{ 001}) and the surface upon which growth occurs is taken to be the (00 1) surface. The
growth direction is thus the [001] direction and the edges of the substrate are parallel to
the [100] and [010] directions.
As used herein, the term "substantially" when referring to a direction, e.g. a
crystallographic direction or a direction with respect to the growth surface of the
substrate, means within about 10° of said direction, alternatively within about 5° of said
direction, alternatively within about 4° of said direction, alternatively within about 3° of
10 said direction.
It is important for the production of high quality CVD diamond material that the growth
surface of the diamond substrate is substantially free of crystal defects. In this context,
crystal defects primarily means dislocations and microcracks, but also includes twin
15 boundaries, stacking faults, point defects, low angle boundaries and any other disruption
to the crystal lattice.
The nature of the defects that are responsible for the brown colour of diamond are not
properly understood at present but are believed by the inventors to be related to the
20 presence of multivacancy clusters that are grown-in under large growth rates,
concomitant with the addition of nitrogen to the plasma via a hydrogen I methane
(H2/CH4) source gas. Such clusters are thermally unstable and may be removed to some
degree, though often not completely, by high-temperature treatment (i.e. annealing). It is
believed that smaller vacancy-related defects, such as NVH (nitrogen-vacancy-
25 hydrogen) defects that are made up of nitrogen and hydrogen and a missing carbon atom,
may be partially responsible for the brown colour and these defects may also be removed
by high-temperature treatment.
The defect density is most easily characterised by optical evaluation after usmg a
30 revealing plasma or chemical etch (referred to collectively as a "revealing etch"). The
revealing etch will be optimised to reveal the defects in the diamond substrate and may
be a brief anisotropic plasma etch of the type described below. In general, two types of
defect can be revealed:
wo 2011/076643 PCT /EP201 0/069828
16
1) Those intrinsic to the diamond substrate material quality. In selected natural
diamond substrates, the density of these defects can be as low as 50 /mm2
, with
more typical values being 102 /mm2
, whilst in other diamond materials, the density
5 of these defects can be 106 /mm2 or greater.
2) Those resulting from polishing, including dislocation structures and microcracks in
the form of "chatter tracks" along polishing lines. The density of these defects can
vary considerably over a sample, with typical values ranging from about 102 /mm2
,
10 up to more than 104 /mm2 in poorly polished regions or substrates.
15
The preferred low density of defects is thus such that the density of surface etch features
related to defects, as described above, is below 5 x 103 /mm2
, and may be below
102 /mm2
.
The defect level at and below the growth surface of a diamond substrate may be
minimised by careful preparation of the substrate. Here preparation includes any process
applied to the substrate material from mine recovery (in the case of natural diamond) or
synthesis (in the case of synthetic diamond material) as each stage can influence the
20 defect density within the substrate material at the plane which will ultimately form the
growth surface of the substrate when processing to form the substrate is complete.
Particular processing steps may include conventional diamond processes such as
mechanical sawing, lapping and polishing conditions, and less conventional techniques
such as laser processing or ion implantation and lift off techniques, chemical/mechanical
25 polishing, and both liquid and plasma chemical processing techniques. In addition, the
roughness of the surface, as described by its Ra value, should be minimised; typical
values prior to any plasma etch are a few nanometres, i.e. 10 nm or less.
The Ra (sometimes referred to as "RA" or "centre line average" or "c.l.a.") refers to the
30 arithmetic mean of the absolute deviation of surface profile from the mean line measured
by stylus profilometer, measured over a length of 0.08 mm and measured according to
British Standard BS 1134 Part 1 and Part 2. The mathematical description of Ra (from
"Tribology", I. M. Hutchings, Pub. Edward Arnold (London), 1992, pages 8-9) is:
wo 2011/076643 PCT /EP201 0/069828
17
Ra = ± fly(x)ldx
(arithmetic mean of the absolute deviation of surface profile measured by stylus
profilometer, generally over 0.08 mm length). Measurement of Ra using a stylus
profilometer is well known in the art and there are many instruments suitable for making
5 such measurements; for example, the inventors have used a "Taylor Hobson
FormTalysurf 50", (Taylor Hobson Ltd, Leicester, UK).
10
15
The surface damage of the growth surface of a diamond substrate may be minimised by
providing an anisotropic etch, for example a plasma etch.
Anisotropic etching involves the removal of material from the growth surface of the
substrate to provide a growth surface which is substantially flat and also free or
substantially-free of residual damage features arising from any prior processing steps, in
particular mechanical processing steps.
The anisotropic plasma etch may be an oxygen etch using an etching gas comprising
hydrogen and oxygen. Alternatively, the plasma etch may be a hydrogen etch. In a
further alternative, the plasma etch may comprise both an oxygen and a hydrogen etch,
wherein in some cases, the oxygen etch is followed by the hydrogen etch. It is
20 advantageous for the oxygen etch to follow the hydrogen etch since the hydrogen etch is
less specific to crystal defects and rounds off any angularities caused by the oxygen etch
(which aggressively attacks such crystal defects) and provides a smoother, better growth
surface on the substrate.
25 The duration and temperature of the anisotropic plasma etch are selected to enable any
damage to the surface of the substrate to be removed, and for any surface contaminants
to be removed. The surface of the substrate may have been damaged by processing steps
undertaken before the plasma etch. Generally, the plasma etch does not form a highly
roughened surface on the substrate and does not etch extensively along extended defects
30 (such as dislocations) which intersect the surface and thus cause deep pits.
The anisotropic etch may be an in situ plasma etch. In principle this etch need not be in
situ, nor immediately prior to the synthesis of the diamond material, but the greatest
wo 2011/076643 PCT /EP201 0/069828
18
benefit is achieved if it is in situ, because it avoids any risk of further physical damage or
chemical contamination of the substrate. An in situ etch is also generally most
convenient when the synthesis of the diamond material is also plasma based. The
plasma etch can use similar conditions to those used for the synthesis of the diamond
5 material, but with the absence of any carbon-containing source gas and generally at a
slightly lower temperature to give better control of the etch rate.
The anisotropic oxygen etch conditions may be a pressure of from about 50 x 102 Pa to
about 450 x 102 Pa, an etching gas containing an oxygen content of from about 1% to
10 about 4%, an argon content of about 30% or less and the balance hydrogen, all
percentages being by volume, a substrate temperature of from about 600°C to about
ll00°C (generally about 800°C), and an etch duration of from about 3 minutes to about
60 minutes.
15 The anisotropic hydrogen etch conditions may be a pressure of from about 50 x 102 Pa to
about 450 x 102 Pa, an etching gas containing hydrogen and about 30% or less by
volume argon, a substrate temperature of from about 600°C to about ll00°C (generally
about 800°C), and an etch duration of from about 3 minutes to about 60 minutes.
20 Alternative methods for the etch not solely based on argon, hydrogen and oxygen may
be used, for example, those utilising halogens, other inert gases or nitrogen.
For the duration of the synthesis process, the substrate may be maintained at a
temperature of from about 750°C to about 1000°C, alternatively from about 750°C to
25 about 950°C, alternatively from about 780°C to about 870°C.
As used herein, the "synthesis environment" is the part of the synthesis apparatus in
which CVD diamond material synthesis takes place. It is generally down stream of a gas
delivery system and up stream of a means of controlling the pressure within the synthesis
30 environment, for example a 'throttle valve' controlling the gas flow impedance and a
vacuum pump. The synthesis environment comprises a source gas and any gases due to
leaks, back-streaming through vacuum pumps and desorption from the synthesis
apparatus. Therefore, changes in the composition of a source gas provided by the gas
wo 2011/076643 PCT /EP201 0/069828
19
delivery system are reflected in the gas composition of the synthesis environment.
However, the gas composition in the synthesis environment is also affected by other
factors such as the synthesis process. In principle the gas delivery system may have
more than one point of entry into the synthesis environment, such that the composition
5 of the source gas, provided by the gas delivery system, does not pre-exist as a mixture
within the gas delivery system, but is determined by the relative flows of the gases being
added into the synthesis environment by the gas delivery system.
All references herein to the gas composition in the synthesis environment are based on
10 the sum of the gases added to the synthesis environment by the gas delivery system and
any gases due to leaks, back-streaming through vacuum pumps and desorption from the
synthesis apparatus. This can be determined by activating the gas delivery system to
deliver the source gas to the synthesis system where the CVD diamond material
synthesis would take place, any gases due to leaks, back-streaming through vacuum
15 pumps and desorption from the synthesis apparatus will also be generated at this point.
The gas composition in the synthesis environment can then be determined in the
synthesis system where the CVD diamond material synthesis would take place without
actually synthesizing any CVD diamond material.
20 The concentrations of the gases in the synthesis environment are controlled by altering
the concentration of the gases in the source gas before inputting the source gas into the
synthesis apparatus. Therefore, the measurements given herein for gas concentrations in
the source gas have been determined by the source gas composition before it is inputted
into the synthesis apparatus and have not been measured in the synthesis environment in
25 situ (i.e. inside the synthesis apparatus). The person skilled in the art would be able to
calculate the required flow of any separate gases that would be needed to provide the
required concentration in the source gas.
The synthesis environment comprises nitrogen at an atomic concentration of from about
30 0.2 ppm to about 100 ppm, alternatively from about 0.5 ppm to about 50 ppm,
alternatively from about 1 ppm to about 30 ppm, alternatively from about 2 ppm to about
20 ppm, alternatively from about 5 ppm to about 20 ppm. Preferably the atomic nitrogen
content of the synthesis environment is from about 5 ppm to about 20 ppm. The
5
wo 2011/076643 PCT /EP201 0/069828
20
nitrogen concentration may be determined as the atomic fraction nitrogen, N, in the total
gas flow of the source gas; for example a gas flow of 1000 seem (standard cubic
centimetres) of Hz and 150 seem of Hz containing 100 ppm of Nz would have an atomic
nitrogen content of:
(150 X 100 X 2)/(1000 X 2 +150 X 2) = 13 ppm
The nitrogen in the synthesis environment may be provided by nitrogen (i.e. Nz) or
nitrogen-containing gases (e.g. ammonia (NH3), hydrazine (NzH4) etc).
The method of the present invention additionally comprises dissociating the source gas.
10 Dissociation of the source gas in the synthesis environment is brought about by an
energy source such as microwave, RF (radio frequency) energy, a flame, a hot filament
or jet based technique and the reactive gas species so produced (also herein termed the
"plasma") are allowed to deposit onto the growth surface of a substrate and form the
synthetic CVD diamond material. In one embodiment, the energy source is a microwave
15 energy source. The frequency of the microwave source may be from about 800 MHz to
about 2500 MHz and is generally one of the industrial heating frequencies of which
2450 MHz, 915 MHz and 896 MHz are examples.
The process of the present invention may be carried out at a pressure of about 8,000 Pa
20 (60 Torr) or greater, alternatively about 10,600 Pa (80 Torr) or greater, alternatively
about 13,300 Pa (100 Torr) or greater, alternatively about 16,000 Pa (120 Torr) or
greater.
The process of the present invention may be carried out at a pressure of from about
25 10,600 Pa (about 80 Torr) to about 40,000 Pa (about 300 Torr), alternatively about
12,000 Pa (about 90 Torr) to about 40,000 Pa (about 300 Torr), alternatively about
13,300 Pa (about 100 Torr) to about 40,000 Pa (about 300 Torr), alternatively about
10,600 Pa (about 80 Torr) to about 26,600 Pa (about 200 Torr), alternatively about
12,000 Pa (about 90 Torr) to about 24,000 Pa (about 180 Torr), alternatively about
30 13,300 Pa (about 100 Torr) to about 20,000 Pa (about 150 Torr). The method of the
present invention surprisingly allows these high pressures to be used in combination
with a source gas comprising 0-containing sources allowing more efficient etching of
wo 2011/076643 PCT /EP201 0/069828
21
non-diamond carbon and defects, as discussed above, thereby allowing the synthesis of
high quality CVD diamond material.
In one embodiment, the process of the present invention may be carried out at a pressure
5 of from 13,300 Pa (100 Torr) to 40,000 Pa (300 Torr).
The inventors have found that there is a pressure, Pare, above which there is a greatly
increased risk of the formation of monopolar arcs in the plasma that may disrupt or stop
the synthesis process, which is related to the atomic fraction of hydrogen, Hf, in the
10 source gas. The inventors have found that Pare is given by:
Pare= 170(Hf + 0.25) +X
where the units of Pare are Torr (1 Torr= 133.3 Pa). Although this equation results in a
15 P arc measured in Torr, the skilled person would know how to convert it to a measurement
in Pascals. X represents the fact that this upper pressure limit may vary slightly
dependent on the given reactor configuration and process conditions other than Hf, but
for any given reactor and process it is a simple exercise to establish where this limit lies
by varying the pressure and observing the pressure of formation of unipolar arcs. The
20 inventors have found that X typically ranges from about 20 to about -50; alternatively
about 20 to about -30; alternatively about 10 to about -30; alternatively about 20 to about
-20; alternatively about 10 to about -20; alternatively about 10 to about -10; alternatively
about 5 to about -10; alternatively about 5 to about -5. Alternatively X may be about 10,
alternatively about 5, alternatively about 0, alternatively about -5, alternatively about-
25 10.
Having determined a value for X, this expression has been found to hold over a range of
Hf values of between about 0.4 and about 0.95 (e.g. between about 0.4 and about 0.75),
and for a wide range of Cf and Of making up the balance of the synthesis environment.
30 In addition, the inventors have found the addition of one or more inert gases in the
atomic fractions previously disclosed, does not have a significant impact on the pressure
limit P arc·
wo 2011/076643 PCT /EP201 0/069828
22
The inventors have found that the preferred pressure of operation in the method of this
invention exceeds Plower where Plower= Pare- Y, where the units of Plower andY are Torr
and the value of Y is about 50 Torr or less, alternatively about 40 Torr or less,
alternatively about 30 Torr or less, alternatively about 20 Torr or less, alternatively about
5 10 Torr.
In one embodiment, the method of the present invention is carried out at a pressure
greater than Plower Torr, where Plower= P arc- Y and P arc = 170(Hf + 0.25) +X, where Y is
about 50 Torr, X is about 0 Torr and P arc is the pressure of the onset of unipolar arcing in
10 the process.
The pressure of operation may be about 120 Torr or more, alternatively about 130 Torr
or more, alternatively about 140 Torr or more, alternatively about 150 Torr or more,
alternatively about 160 Torr or more. These operating pressures may be useful when Hf
15 is 0.75.
20
The preferred pressure of operation lies at or below P arc, and for reasons of process
stability, preferably a small but significant pressure below Pare· This pressure is referred
to as "Pupper" and is measured in Torr.
Thus, the inventors have found that the preferred pressure of operation in the method of
this invention is given by Pupper =Pare - Z, where Z is about 0 Torr, alternatively Z is
about 5 Torr, alternatively Z is about 10 Torr.
25 It will be appreciated by the skilled person that any of the X, Y and Z values provided
above may be combined.
In one embodiment, the method of the present invention is carried out at a pressure less
than or equal to P upper, wherein P upper = P arc - Z and P arc = 170(Hf + 0 .25) + X, where X is
30 from about 20 to about -50, Pare is the pressure of the onset of unipolar arcing in the
process and Z is about 0 Torr.
wo 2011/076643 PCT /EP201 0/069828
23
In one embodiment, the process is carried out at a pressure less than or equal to P upper,
wherein Pupper =Pare- Z and Pare= 170(Hf + 0.25) +X, where X is about 0, Z is about 0
Torr and P arc is the pressure of the onset of unipolar arcing in the process.
5 The pressure of operation, Pupper, may be about 170 Torr or less, alternatively about 165
Torr or less, alternatively about 160 Torr or less. These operating pressures may be
useful when Hds 0.75.
As mentioned previously, the present invention provides synthetic CVD diamond
10 material comprising Ns0 at a concentration of greater than about 0.5 ppm (about
8.8 x 1016 atoms per cm3
) and having a total integrated absorption in the visible range
from 350 nm to 750 nm such that at least about 35 % of the absorption is attributable to
0 Ns.
15 Alternatively, the concentration of Ns0 in the synthetic CVD diamond material may be
greater than about 1.0 ppm, alternatively greater than about 1.5 ppm, alternatively
greater than about 1.8 ppm, alternatively greater than about 2.0 ppm, alternatively
greater than about 2.5 ppm, alternatively greater than about 3.0 ppm, alternatively
greater than about 3.5 ppm, alternatively greater than about 4.0 ppm, alternatively
20 greater than about 5.0 ppm, alternatively greater than about 7.0 ppm, alternatively
greater than about 10.0 ppm, alternatively greater than about 15.0 ppm, alternatively
greater than about 20.0 ppm, alternatively greater than about 25.0 ppm, alternatively
greater than about 50.0 ppm, alternatively greater than about 75.0 ppm, alternatively
greater than about 100.0 ppm.
25
Preferably the concentration of Ns0 is about 200 ppm or less.
Preferably the concentration of Ns0 in the synthetic CVD diamond material is greater
than about 1.0 ppm and less than about 25 ppm, alternatively greater than about 2.0 ppm
30 and less than about 15 ppm.
Since diamond is a wide band gap semiconductor, diamond material, and in particular
diamond material containing defects, does not necessarily have a well defined Fermi
wo 2011/076643 PCT /EP201 0/069828
24
level. At room temperature (i.e. about 300 K), charge that is initially trapped at a defect
having an energy level that is relatively shallow compared with either the valence band
maximum or the conduction band minimum, will reach an equilibrium distribution via
transport after thermal excitation to the valance or conduction bands. However, diamond
5 material may contain defects having energy levels that are relatively deep within the
band gap such that, at room temperature, there is a low probability that electrons will be
thermally excited between the valence band and the defect or between the defect and the
conduction band. When such defects are present, the charge distribution across the
various defects may depend on the thermal and excitation history of the sample. In such
10 cases, to the extent that the optical absorption properties of the material depend on the
charge state of defects within it, they will also depend on the thermal and excitation
history of the sample. For example, the proportions of isolated substitutional nitrogen
defects that exist in the neutral charge state may depend on the prior thermal and
excitation history of the sample and therefore the proportion of the total optical
15 absorption that is attributable to this neutral defect will also depend on the history of the
sample.
For the avoidance of doubt, when the history of the sample is not specified, the
properties of the material described in this invention should be taken to be properties that
20 can be measured at room temperature (i.e. about 300 K) with no additional excitation of
the sample during the measurement other than that required in making the measurement.
Preferably the properties are measured after the sample has been irradiated with light
from a deuterium lamp under the following conditions:
25 (a) distance between the sample and the lamp of about 10 em or less;
(b) lamp operating electrical power of at least 10 Watts; and
(c) a duration of between about 5 minutes and about 60 minutes.
In particular, the properties are measured after the sample has been irradiated with light
30 from a deuterium lamp under the following conditions:
(d)
(e)
(f)
distance between the sample and the lamp of 8 em;
lamp operating electrical power of 10 Watts; and
a duration of 10 minutes.
5
10
wo 2011/076643 PCT /EP201 0/069828
25
Deuterium lamps (also known as "deuterium arc lamps") are widely used in
spectroscopy where a continuous output is required between wavelengths of about 180
nm and about 370 nm.
The concentration of N s
0 present in the synthetic CVD diamond material of the present
invention may be measured using the 270 nm peak using UV -visible absorption
spectroscopy. The technique of UV-visible absorption spectroscopy is well-known in
the art.
The concentration of Ns0 in synthetic CVD diamond material may be found by
measuring infrared absorption peaks at wavenumbers of 1332 cm-1 and 1344 cm-1
.
Using a spectrometer with a resolution of 1 cm-1
, the conversion factors between the
absorption coefficient values in cm-1 for the peaks at 1332 cm-1 and 1344 cm-1 and the
15 concentrations of single nitrogen in the positively-charged and neutral states respectively
are 5.5 (S.C. Lawson et al., J. Phys. Condens. Matter, 10 (1998), 6171-6181) and 44.
However, it must be noted that the value derived from the 1332 cm-1 peak is only an
upper limit.
20 Alternatively, the total concentration of nitrogen may be determined using secondary ion
mass spectroscopy (SIMS). SIMS has a lower detection limit for nitrogen in diamond of
approximately 0.1 ppm and its use is well-known in the art. For synthetic diamond
produced by a CVD method, the vast majority of nitrogen present in the solid is in the
form of neutral single substitutional nitrogen, N8°, and therefore, whilst SIMS
25 measurements of the total nitrogen concentration inevitably provide an upper limit to the
concentration of Ns0
, they typically also provide a reasonable estimate of its actual
concentration.
Alternatively, the concentration of Ns0 may be determined by electron paramagnetic
30 resonance ("EPR"). Whilst the method is well-known in the art, for completeness, it is
summarised here. In measurements conducted using EPR, the abundance of a particular
paramagnetic defect (e.g. the neutral single-substitutional nitrogen defect, N,0
) is
proportional to the integrated intensity of all the EPR absorption resonance lines
wo 2011/076643 PCT /EP201 0/069828
26
originating from that centre. This permits the concentration of the defect to be
detem1ined by compming the integrated intensity to that which is observed from a
reference sarnple, provided care is taken to prevent or correct for the effects of
microwave power saturation. Since continuous wave EPR spectra are recorded using
5 field modulation, double integration is required to determine the EPR intensity and
hence the defect concentration. To minimise the errors associated with double
integration, base line conection, finite limits of integration, etc., especially in cases
where overlapping EPR spectra are present, a spectral fltting method (using a NelderMead
simplex algorithm (J. A. Neider and R. Mead, The Computer Journal, 7 (1965),
10 308)) is employed to detennine the integrated intensity of the EPR centres present in the
sample of interest This entails fitting the experimental spectra with simulated spectra of
the defects present in the sample and determining the integrated intensity of each from
the simulation. Expmimentally it is observed that neither a Lorentzian nor Gaussian line
shape provides a good fit to the experimental EPR spectra, therefore a Tsallis function is
15 used to produce the simulated spectra (D.F. Howarth, J.A. Wei!, Z. Zimpel, J. Magn.
Res., 161 (2003), 215). Furthermore, in the case of lovl nitrogen concentrations, it is
often necessary to use modulation amplitudes approaching or exceeding the line width of
the EPR signals to achieve a good signalinoise ratio (enabling accurate concentration
determination \Vithin a reasonable time frame). Hence pseudo-rnodulation is employed,
20 with the Tsallis line shape in order to produce a good fit to the recorded EPR spectra
(J.S. Fiyde, M. Pasenkie\vicz-CJiemla, A. Jesrnanowicz, W.E. Antholine, Appl. Magn.
Reson., 1 (1990), 483). Using this method the concentration can be determined with a
reproducibility of better than ±5%.
25 The person skilled in the art would be able to determine which Ns0 measurement method
would be appropriate to use in any given situation.
The skilled person is familiar with the methods which can be used to distinguish between
synthetic CVD diamond and synthetic HPHT diamond. The following are a number of
30 non-limiting examples of these methods.
One method of distinguishing synthetic CVD diamond material from a synthetic
diamond material synthesized using HPHT techniques is by the dislocation structure. In
wo 2011/076643 PCT /EP201 0/069828
27
synthetic CVD diamond, dislocations generally thread in a direction that is roughly
perpendicular to the initial growth surface of the substrate, i.e. where the substrate is a
(001) substrate, the dislocations are approximately aligned parallel to the [001] direction.
In synthetic diamond material synthesized using HPHT techniques, dislocations that
5 nucleate on the surface of the seed crystal (often a surface close to { 001}) typically grow
in <110> directions. Thus the two types of material can be distinguished by their
different dislocation structures observed, for example, in an X-ray topograph. However,
obtaining X-ray topographs is an onerous task and clearly an alternative, less onerous
method that enables positive distinction would be desirable.
10
A further method for distinguishing synthetic CVD diamond material from a synthetic
diamond material synthesized using HPHT techniques is by the presence of metallic
inclusions in the HPHT -synthesized material that are incorporated as a result of the
synthesis process. The inclusions are comprised of the metals used as the solvent
15 catalyst metal, e.g. Fe, Co, Ni etc. Inclusions can vary in size from less than 1 f..Lm to
more than 100 f..Lm. Larger inclusions can be observed using a stereo-microscope (e.g. a
Zeiss DV 4 ); whilst smaller inclusion can be observed using transmitted light in a
metallurgical microscope (e.g. a Zeiss "Axiophot").
20 A further method that can be used to provide a positive distinction between synthetic
diamonds produced by CVD and HPHT methods is photoluminescence spectroscopy
(PL). In the case of HPHT-synthesised material, defects containing atoms from the
catalyst metals (typically transition metals) used in the synthesis process (e.g. Ni, Co, Fe
etc.) are frequently present and the detection of such defects by PL positively indicates
25 that the material has been synthesised by an HPHT method.
The absence of defects related to the presence of catalyst metal atoms in the diamond is
an advantage of diamond produced by the current invention over material produced by
HPHT methods as such defects can locally disrupt the Fermi level affecting the
30 suitability of the material for use as a substrate for the fabrication of electronic devices
such as PETs.
wo 2011/076643 PCT /EP201 0/069828
28
Synthetic CVD diamond material of the present invention may be identified by its
unique integrated absorption and its association with N8°. The integrated absorption is
measured using a UV /visible absorption spectrum of the synthetic CVD diamond
material taken at room temperature. All the absorption spectra mentioned herein were
5 collected using a Perkin Elmer Lambda-9 spectrometer. The UV /visible absorption
spectra of diamond material may show characteristic bands at 360 nm and 510 nm.
The data recorded in the spectra ("the measured spectrum") were deconvolved in the
following way to give information on the proportion of the measured absorption that can
10 be attributed to N s
0 and the proportion of the measured absorption that can be attributed
to other defects.
a. A reflection loss spectrum was created using tabulated refractive index data and
standard expressions for the reflection loss for a parallel-sided plate.
15
b. The reflection loss spectrum was subtracted from the measured absorbance data
and an absorption coefficient spectrum is created from the resulting spectrum.
c. In order to determine the component of the measured spectrum that is attributable
20 to N8°, an absorption spectrum for type lb HPHT synthetic diamond (for which the
absorption is attributed solely to Ns0
) was scaled until it is substantially removed the
270 nm peak from the measured spectrum when subtracted from it.
d. Using the visible regwn of the spectrum as stretching from 380 nm (i.e.
25 3.2618 eV) to 750 nm (i.e. 1.6527 eV), the integrated absorption in the visible region
was determined for the measured spectrum (C) and for the component of it attributable
to Ns0 (B). The ratio of the integrated absorption in the visible region for the component
attributable to N s
0 and for the measured spectrum (B/C) can then be calculated. The
contribution to the optical absorption in the visible region not due to Ns0 is given by the
30 value C-B.
e. In practice reflection losses are generally greater than the theoretical values and
this can make it difficult to determine absolute absorption coefficient values. In order to
wo 2011/076643 PCT /EP201 0/069828
29
correct for additional losses not directly related to absorption, the following routine was
used. Towards lower energies it was generally the case that below a particular energy
the measured absorption no longer showed significant variation with energy. Absorption
coefficient data were shifted so that absorption coefficient was zero at the energy at
5 which there was no further significant variation in absorption coefficient.
The residual spectrum can be further deconvolved in to a component proportional to 1111?
and two overlapping bands, one centred on 360 nm and the other centred on 510 nm.
The skilled person would know how to calculate the absorption coefficients of these
10 bands
Synthetic CVD diamond material of the present invention may have a total integrated
absorption in the visible range from 350 nm to 750 nm such that at least about 35 %,
alternatively at least about 40%, alternatively at least about 45%, alternatively at least
15 about 50%, alternatively at least about 55%, alternatively at least about 60%,
alternatively at least about 65%, alternatively at least about 70%, alternatively at least
about 75%, alternatively at least about 80%, alternatively at least about 85%,
alternatively at least about 90%, alternatively at least about 92%, alternatively at least
about 94%, alternatively at least about 96%, alternatively at least about 98%,
20 alternatively at least about 99% of the integrated absorption (in eV.cm-1
) is attributable
0 toNs.
25
Preferably, for synthetic CVD diamond material of the present invention, at least about
85% of the integrated absorption between 350 nm and 750 nm is attributable to N8°.
These values of integrated absorption indicate that synthetic CVD diamond material of
the present invention is similar to intrinsic diamond, having a high colour, and that this is
due to a lower concentration of defects other than Ns0 that absorb in the visible part of
the spectrum. This shows that synthetic CVD diamond material of the present invention
30 is of a high quality.
Synthetic CVD diamond material of the present invention may also be characterised by
its photoluminescence (PL) spectrum. These are typically performed by excitation using
wo 2011/076643 PCT /EP201 0/069828
30
a continuous-wave laser source employing a power between about 10 mW and about
100 mW focussed onto a diamond sample surface using a microscope objective, and
detection via a high-resolution (better than about 0.5 nm) grating spectrometer. At 77 K,
using 488 nm excitation of an argon-ion laser, the PL spectrum of the synthetic CVD
5 diamond material of the present invention shows a peak between about 543.0 nm and
about 543.2 nm, with an intensity ratio of this peak normalized to the 1st order diamond
Raman (at 521.9 nm for this excitation wavelength) that is greater than about 0.005. The
inventors believe that the peak between about 543.0 nm and about 543.2 nm is
associated with the presence of oxygen in the synthesis process at an atomic fraction of
10 at least about 0.05 relative to the atomic fraction of carbon as the peak has only been
observed in such circumstances. At 77 K, using 488 nm excitation of an argon-ion laser,
the PL spectrum of the synthetic CVD diamond material of the present invention also
shows a second peak between about 539.9 nm and 540.1 nm. This peak is typically
between l/8th and l/12th of the strength of the peak between 543.0 nm and 543.2 nm and
15 therefore is believed to be observed only when there is a significant atomic fraction of
oxygen in the synthesis environment. The composition and structure of the defects
responsible for either the peak between about 543.0 nm and about 543.2 nm or the peak
between about 539.9 nm and about 540.1 nm have not yet been elucidated. It is possible
that these two peaks, which are evident in the synthetic CVD diamond material of the
20 present invention, are inter-related.
Synthetic CVD diamond material of the present invention may also be indicated by its
PL spectrum using the 514.5 nm excitation of an argon-ion laser. Using this type of
excitation, the PL spectrum of synthetic CVD diamond material of the present invention
25 shows two photoluminescence peaks, the first at 574.8- 575.1 nm (approximately 575
nm) and the second at 636.9- 637.1 nm (approximately 637 nm) such that the ratio of
the integrated intensity of the 637 nm peak to the 575 nm peak exceeds about 1.0,
alternatively exceeds about 1.2, alternatively exceeds about 1.4. Without wishing to be
bound by theory, the first and second peaks of this PL spectrum correspond to the
30 nitrogen vacancy defect in its neutral and negative charge states, respectively.
Furthermore, the material of the current invention shows a marked reduction in the level
of the 737 nm PL line that is believed to be related to the silicon-vacancy (Si-V) defect.
5
wo 2011/076643 PCT /EP201 0/069828
31
The reduction in the 737 nm PL line intensity is believed to occur in the absence of a
specific reduction in the concentration of silicon in the synthesis environment and
therefore the inventors believe that the reduction is caused by an as yet unidentified
change in the silicon incorporation mechanism.
The use of photoluminescence (PL) spectroscopy to characterise defects in diamond is
well known in the art. In PL the sample is exposed to photons of a particular wavelength
(e.g. 514.5 nm radiation of an argon-ion laser). This excites electrons in the material to
higher energy levels. The excited electrons decay back to their ground states and emit
10 photons having wavelengths that are characteristic of the transition and that can be
characterised using a spectrometer. There are numerous PL spectrometers commercially
available.
Synthetic CVD diamond material of the present invention may embody IR absorption
15 coefficients greater than (i) 0.1 cm-1 at 1332 cm-1 and (ii) 0.05 cm-1 at 1344 cm-1 using an
infrared absorption spectrometer with a resolution of 1 cm-1
. These respectively indicate
that the single nitrogen concentrations are (i) more than 0.55 parts per million in the
positive charge state and (ii) more than 2.2 parts per million in the neutral charge state.
The correlation between the absorption coefficient at 1332 cm-1 and single substitutional
20 nitrogen in the neutral charge state is that an absorption coefficient of 1 cm-1 corresponds
to approximately 5.5 ppm. The correlation between the absorption coefficient at 1344
cm-1 and single substitutional nitrogen in the positive charge state is that an absorption
coefficient of 1 cm-1 corresponds to approximately 44 ppm.
25 Synthetic CVD diamond material of the present invention may have at least about 50%,
alternatively at least about 80%, alternatively at least about 90%, alternatively at least
about 95% of the volume of the synthetic CVD diamond material formed from a single
growth sector. The material of the single growth sector may have Ns0 levels within
±10% of the mean for greater than about 50% of the volume of the growth sector,
30 alternatively greater than about 60% of the volume of the growth sector, alternatively
greater than about 80% of the volume of the growth sector. Forming synthetic CVD
diamond material from a single growth sector is advantageous as the CVD diamond
material will have fewer surfaces with different crystallographic orientations (which are
wo 2011/076643 PCT /EP201 0/069828
32
the surfaces corresponding to differing growth sectors). Surfaces with different
crystallographic orientations exhibit strongly differential uptakes of nitrogen impurity
and the synthetic CVD diamond material therefore tends to show undesirable zones with
different colour, resulting from the different concentrations of Ns0 in different growth
5 sectors. Therefore, using a single growth sector will lead to higher quality synthetic
CVD diamond material.
Synthetic CVD diamond material of the present invention may contain impurities other
than N8°. In one embodiment, the elemental concentration of individual chemical
10 impurities other than nitrogen and hydrogen is less than about 0.1 ppm, alternatively less
than about 0.05 ppm, alternatively less than about 0.02 ppm, alternatively less than about
0.01 ppm. As used herein, "elemental concentration" means the absolute chemical
concentration of the impurities referred to.
15 The concentration of substitutional boron may be about 1 x 1017 atoms per cm3 or less,
alternatively about 5 x 1016 atoms per cm3 or less, alternatively about 1 x 1016 atoms per
cm3 or less. The concentration of hydrogen (including the isotopes of hydrogen) may be
about 1 x 1019 atoms per cm3 or less, alternatively about 1 x 1018 atoms per cm3 or less,
alternatively about 1 x 1017 atoms per cm3 or less.
20
The perceived colour of an object depends on the transmittance/absorbance spectrum of
the object, the spectral power distribution of the illumination source and the response
curves of the observer's eyes. The CIE L*a*b* chromaticity coordinates (and therefore
hue angles) quoted herein have been derived in the way described below. Using a
25 standard D65 illumination spectrum and standard (red, green and blue) response curves
of the eye (G. Wyszecki and W. S. Stiles, John Wiley, New York-London-Sydney,
1967) CIE L*a*b* chromaticity coordinates of a parallel-sided plate of diamond have
been derived from its transmittance spectrum using the relationships below, between
350 nm and 800 nm with a data interval of 1 nm:
30
SA.= transmittance at wavelength A
LA,= spectral power distribution of the illumination
XA, = red response function of the eye
wo 2011/076643
y~e = green response function of the eye
z~e = blue response function of the eye
X= L~e [S~ex~eL~e] I Yo
5 Y = L~e [S~e y~e L~e] I Yo
Z = L~e [S~ez~eL~e] I Yo
Where Y0 =I.~, y~, L~,
L* = 116 (Y/Y0)
113
- 16 =Lightness
10 a*= 500[(X/X0)
113
- (Y/Y0)
113
]
b* = 200[(Y/Y0)
113
- (Z/Z0)
113
]
C* = (a*2 + b*2
)
112 =saturation
hab =arctan (b* I a*)= hue angle
33
PCT /EP201 0/069828
(for Y/Yo > 0.008856)
(for X/Xo > 0.008856, Y/Yo > 0.008856)
(for Z/Z0 > 0.008856)
15 Modified versions of these equations must be used outside the limits of Y /Y0, X/X0 and
Z/Z0• The modified versions are given in a technical report prepared by the Commission
Internationale de L'Eclairage (Colorimetry (1986)).
It is standard to plot a* and b* coordinates on a graph with a* corresponding to the x
20 axis and b* corresponding to the y axis. Positive a* and b* values correspond
respectively to red and yellow components of the hue. Negative a* and b* values
correspond respectively to green and blue components. The positive quadrant of the
graph then covers hues ranging from yellow through orange to red, with saturations (C*)
given by the distance from the origin.
25
It is possible to predict how the a*b* coordinates of diamond with a given absorption
coefficient spectrum will change as the optical path length is varied. In order to do this,
the reflection loss must first be subtracted from the measured absorbance spectrum. The
absorbance is then scaled to allow for a different path length and then the reflection loss
30 is added back on. The absorbance spectrum can then be converted to a transmittance
spectrum which is used to derive the CIE L *a*b* coordinates for the new thickness. In
this way the dependence of the hue, saturation and lightness on optical path length can
wo 2011/076643 PCT /EP201 0/069828
34
be modelled to giVe an understanding of how the colour of diamond with giVen
absorption properties per unit thickness will depend on the optical path length.
L*, the lightness, forms the third dimension of the CIE L*a*b* colour space. It is
5 important to understand the way in which the lightness and saturation vary as the optical
path length is changed for diamond with particular optical absorption properties. This
can be illustrated on a colour tone diagram in which L * is plotted along the y-axis and
C* is plotted along the x-axis. The method described in the preceding paragraph can
also be used to predict how the L *C* coordinates of diamond with a given absorption
10 coefficient spectrum depend on the optical path length.
The C* (saturation) numbers can be divided into saturation ranges of 10 C* units and
assigned descriptive terms as below.
15 0-10 weak
10-20 weak-moderate
20-30 moderate
30-40 moderate-strong
40-50 strong
20 50-60 strong-very strong
60-70 very strong
70-80+ very very strong
Similarly the L * numbers can be divided up into lightness ranges as follows:
25
5-15 very very dark
15-25 very dark
25-35 dark
35-45 medium/dark
30 45-55 medium
55-65 light/medium
65-75 light
75-85 very light
5
wo 2011/076643 PCT /EP201 0/069828
35
85-95 very very light
There are four basic colour tones defined by the following combinations of lightness and
saturation:
Bright: Light and high saturation,
Deep: High saturation and dark,
Pale: Light and low saturation,
Dull: Low saturation and dark.
A hue angle of greater than 80° for a transmission pathlength of 1 mm indicates that the
10 colour of the synthetic CVD diamond material of the present invention is dominated by
the Ns0
, with little contribution from other colour centres in the material. In particular,
synthetic CVD diamond material of the present invention, when in the form of a parallel
sided plate with a thickness of approximately 1 mm, may have the following colour
parameters in the CIE L *a*b* colour space:
15
20
a* between about -20 and about 1, alternatively between about -10 and about 1,
alternatively between about -5 and about 1;
b* greater than about 5 and less than about 20, alternatively greater than about
10 and less than about 20;
C* (saturation) between about 0 and about 30, alternatively between about 1
and about 25, alternatively between about 2 and about 30; and
L * (lightness) greater than about 40 and less than about 100, alternatively
greater than about 50 and less than about 100, alternatively greater than about
60 and less than about 100.
This provides a quantitative measure the quality of synthetic CVD diamond material of
25 the present invention. These colour properties are advantageous because they give the
diamond a pure yellow colour and can be used for ornamental purposes such as
gemstones for jewellery.
The diamond material of the present invention may have a hue angle for a transmission
30 pathlength of 1 mm of about 80° or greater, alternatively about 85° or greater,
alternatively about 90° or greater, alternatively about 95° or greater.
wo 2011/076643 PCT /EP201 0/069828
36
The CVD diamond material of the present invention may have a hue angle of less than
about 180° for a transmission pathlength of 1 mm.
It is possible for synthetic CVD diamond material produced using a H-based plasma
5 (e.g. using H2/CH4 source gas mixtures) to include greater than 0.5 ppm of N8°, but such
synthetic CVD diamond material produced using these H2/CH4-based source gases will
have extremely high levels of other defects, reducing the quality and appearance (or
colour) of the material. Therefore, the synthetic CVD diamond materials of the prior art
will typically have a hue angle significantly less than about 80°, and often less than 70°,
10 for a transmission path length of 1 mm and will consequently be brown coloured.
15
20
In one embodiment, synthetic CVD diamond material of the present invention may be in
the form of a freestanding entity having a thickness of greater than about 0.2mm,
alternatively greater than about 0.5mm, alternatively greater than about 1.0 mm,
alternatively greater than about 1.5 mm, alternatively greater than about 2.0mm,
alternatively greater than about 2.5mm, alternatively greater than about 3.0mm,
alternatively greater than about 3.5mm, alternatively greater than about 4.0mm,
alternatively greater than about 5.0mm, alternatively greater than about 6.0mm,
alternatively greater than about 10 mm, alternatively greater than about 15 mm.
In one embodiment, synthetic CVD diamond material of the present invention may be in
the form of a freestanding entity having a thickness of less than about 50 mm;
alternatively less than about 45 mm; alternatively less than about 40 mm; alternatively
less than about 35 mm; alternatively less than about 30 mm; alternatively less than about
25 25 mm alternatively less than about 20 mm.
Such thicknesses of freestanding entities made from synthetic CVD diamond material of
the present invention are advantageous because they can be used in ornamental
applications such as gemstones for jewellery. Such entities can also be used in diamond
30 electronic devices, such as PETs.
In another embodiment, synthetic CVD diamond material of the present invention may
be in the form of a layer attached to a diamond having different characteristics (e.g.
wo 2011/076643 PCT /EP201 0/069828
37
impurity level, dislocation density, carbon isotope ratio, etc.). The layer of CVD
diamond material of the present invention may have a thickness of about 0.5 mm or less,
alternatively about 0.2 mm or less, alternatively about 0.1 mm or less, alternatively about
10 f..Lm or less, alternatively about 1 f..Lm or less, alternatively about 300 nm or less,
5 alternatively about 100 nm or less.
In another embodiment, synthetic CVD diamond material of the present invention may
be in the form of a layer having a thickness of about 0.1 nm or more; alternatively about
0.2 nm or more; alternatively about 0.3 nm or more; alternatively about 0.4 nm or more;
10 alternatively about 0.5 nm or more.
Such layers made from synthetic CVD diamond material of the present invention are
advantageous because they can be used in small scale electronic devices.
15 Synthetic CVD diamond material of the present invention may be in the form of a
doublet. A doublet is a synthetic CVD diamond material made in layered sections. The
lower, larger portion is made from lower quality synthetic CVD diamond material and
has a smaller layer of higher quality synthetic CVD diamond material attached to the top
of it. These doublets are advantageous because they can be used in gemstone
20 applications.
The present invention further provides, a gemstone comprising synthetic CVD diamond
material of the present invention and/or made by the method of the present invention.
25 The present invention further provides the use of synthetic CVD diamond material
according to the present invention and/or made by the method of the present invention as
a gemstone.
The present invention further provides an electronic device comprising synthetic CVD
30 diamond material of the present invention and/or made by the method of the present
invention.
wo 2011/076643 PCT /EP201 0/069828
38
The present invention further provides a luminescent detector comprising a layer of
diamond material of the invention that emits photoluminescence on irradiation with Xrays.
5 The present invention further provides the use of synthetic CVD diamond material
according to the present invention and/or made by the method of the present invention in
an electronic device.
The present invention provides synthetic CVD diamond material comprising Ns0 at a
10 concentration of greater than about 0.5 ppm and having a total integrated absorption in
the visible range from 350 nm to 750 nm such that at least about 35 % of the absorption
is attributable to N8°, wherein the material optionally has one or more of the following
properties:
15 (a)
(b)
hue angle greater than about 80° for a transmission path length of 1 mm;
a photoluminescence spectrum at 77 K using the 488 nm excitation of an argon -
ion laser which shows a peak at from about 543.0 to about 543.2 nm, with an
intensity ratio of this peak normalized to the 1st order diamond Raman (at 521.9
20 nm for this excitation wavelength) of greater than about 1/50 or preferably about
(c)
25
11100 or more preferably about 1/200;
a photoluminescence spectrum using the 514.5 nm excitation of an argon - ion
laser showing two photoluminescence peaks, the first at 57 4. 8 - 57 5.1 nm
(approximately 575 nm) and the second at 636.9- 637.1 nm (approximately 637
nm) such that the ratio of the integrated intensity of the 637 nm peak to the 575
nm peak exceeds about 1.0, alternatively exceeds about 1.2, alternatively exceeds
about 1.4.
30 Therefore, in one embodiment of the present invention, there is provided synthetic CVD
diamond material comprising Ns0 at a concentration of greater than about 0.5 ppm and
having a total integrated absorption in the visible range from 350 nm to 750 nm such that
wo 2011/076643 PCT /EP201 0/069828
39
at least about 35% of the absorption is attributable to N8°, wherein the synthetic CVD
diamond material has feature (a) above.
In a further embodiment of the present invention, there is provided synthetic CVD
5 diamond material comprising Ns0 at a concentration of greater than about 0.5 ppm and
having a total integrated absorption in the visible range from 350 nm to 750 nm such that
at least about 35% of the absorption is attributable to Ns0
, wherein the synthetic CVD
diamond material has feature (b) above.
10 In a still further embodiment of the present invention, there is provided synthetic CVD
diamond material comprising Ns0 at a concentration of greater than about 0.5 ppm and
having a total integrated absorption in the visible range from 350 nm to 750 nm such that
at least about 35% of the absorption is attributable to N8°, wherein the synthetic CVD
diamond material has feature (c) above.
15
In a still further embodiment of the present invention, there is provided synthetic CVD
diamond material comprising Ns0 at a concentration of greater than about 0.5 ppm and
having a total integrated absorption in the visible range from 350 nm to 750 nm such that
at least about 35% of the absorption is attributable to Ns0
, wherein the synthetic CVD
20 diamond material has features (a) and (b) above.
In a still further embodiment of the present invention, there is provided synthetic CVD
diamond material comprising Ns0 at a concentration of greater than about 0.5 ppm and
having a total integrated absorption in the visible range from 350 nm to 750 nm such that
25 at least about 35% of the absorption is attributable to N8°, wherein the synthetic CVD
diamond material has features (a) and (c) above.
In a further embodiment of the present invention, there is provided synthetic CVD
diamond material comprising Ns0 at a concentration of greater than about 0.5 ppm and
30 having a total integrated absorption in the visible range from 350 nm to 750 nm such that
at least about 35% of the absorption is attributable to Ns0
, wherein the synthetic CVD
diamond material has features (b) and (c) above.
wo 2011/076643 PCT /EP201 0/069828
40
In a still further embodiment of the present invention, there is provided synthetic CVD
diamond material comprising Ns0 at a concentration of greater than about 0.5 ppm and
having a total integrated absorption in the visible range from 350 nm to 750 nm such that
at least about 35% of the absorption is attributable to N8°, wherein the synthetic CVD
5 diamond material has features (a), (b) and (c) above.
The present invention is now described, by way of illustration only, with reference to the
accompanying drawings, in which:
10 Figure 1 shows a ternary diagram for the C-H-0 space;
15
Figure 2 shows a comparison of a UV /visible optical absorption spectrum obtained from
"Sample 1, Layer 2" of Example 1 with a spectrum obtained from an HPHT Type lb
synthetic single crystal diamond ;
Figure 3 shows a UV/visible optical absorption spectrum obtained from "Sample 1,
Layer 2" of Example 1 deconvolved into its components;
Figure 4 shows UV/visible optical absorption spectra obtained from "Sample 2, Layer 1"
20 and "Sample 2, Layer 2" of Example 2;
Figure 5 shows the UV /visible optical absorption spectrum "Sample 2, Layer 2" of
Example 2 deconvolved into its components;
25 Figure 6 shows a photoluminescence (PL) spectrum obtained at 77 K from "Sample 2,
Layer 1" and "Sample 2, Layer 2" of Example 2 by exciting with radiation having a
wavelength of 488.2 nm using a 50 mW Ar-ion laser;
Figure 7 shows a photoluminescence (PL) spectrum obtained at 77 K from "Sample 2,
30 Layer 1" and "Sample 2, Layer 2" of Example 2 by exciting with radiation having a
wavelength of 514.5 nm using a 50 mW Ar-ion laser;
wo 2011/076643 PCT /EP201 0/069828
41
Figure 8 shows a PL image obtained from "Sample 3" of Example 3 showing the clear
distinction between "Sample 3, Layer 1" and "Sample 3, Layer 2", overlaid with the
spatially resolved 737 nm PL trace from the sample;
5 Figure 9 shows UV/visible optical absorption spectra obtained from "Sample 4, Layer 2"
of Example 4;
10
Figure 10 shows a PL spectra obtained from "Sample 5, Layer 2" and "Sample 6, Layer
2" of Example 5 obtained by exciting with radiation having a wavelength of 458 nm; and
Figure 11 shows a projection X-ray topograph obtained from "Sample 6" of Example 5.
Examples
15 The foregoing examples are intended to describe the current invention without limiting
the invention to the content of the examples.
Several of the examples were made using processes in which the synthesis environment
was changed partway through the CVD diamond synthesis so that a C02/CH4/H2 process
20 could be compared directly with a CH4/ Ar/H2.
Example 1
Example 1 describes the preparation of single crystal diamond substrates suitable for the
25 deposition of the diamond material of the invention, the deposition of a layer of diamond
material using a CH4/H2 synthesis process, and the subsequent deposition of a layer of
material made by the method of the invention.
1) The substrate was prepared using the following steps:
30 a) A single crystal diamond was selected from a stock of material (type Ia natural
stones and type lb HPHT stones) on the basis of microscopic investigation and
birefringence imaging to identify a stone which was substantially free of strain
and imperfections.
wo 2011/076643 PCT /EP201 0/069828
42
b) A parallel-surfaced plate with lateral dimensions of approximately 4 mm x 4 mm
and approximately 500 f..Lm thick and surface Ra of less than 1 nm, with all faces
being within 5° of { 100} surfaces, was prepared from the selected diamond using
processes including laser sawing, and mechanical lapping and polishing. The
5 processes used had previously been optimised to minimise subsurface defects
using a method of a revealing plasma etch to determine the defect levels being
introduced by the processing.
A substrate produced by above steps typically has a density of defects measurable after a
10 revealing etch that is dependent primarily on the material quality and is below about 5 x
103 defects/mm2
, and generally below about 102 defects/mm2
•
2) The diamond substrate was mounted on a tungsten carrier using a Au-Ta high
temperature diamond braze. This was introduced into an 896 MHz microwave plasma
15 CVD diamond reactor.
3) The reactor was started and the substrate was subjected to a two stage pre-growth
etching sequence consisting of:
(a) an in situ oxygen plasma etch performed using gas flows of 40/20/3000 seem of
20 0 2/Ar/H2 at a pressure of about 236 x 102 Pa (about 180 Torr) and a substrate
temperature of about 716°C for a duration of about 30 minutes;
25 4)
(b) followed without interruption by a hydrogen etch with the removal of the 0 2
from the gas flow for a duration of about 30 minutes.
A first layer of CVD diamond ("Sample 1, Layer 1") was deposited on the etched
substrate by the introduction of CH4 into the gas flow, giving a gas flow comprising
140/20/3000 seem of CH4/Ar/H2 at a pressure of about 236 x 102 Pa (about 180 Torr).
The source gas additionally comprised an atomic fraction of nitrogen of 1.4 ppm. The
substrate temperature was 840°C. Sample 1, Layer 1 has been produced by a process of
30 a type well known in the art and thoroughly characterised, and therefore provides an in
situ standard against which Sample 1, Layer 2 can be compared. The inventors have
found that the optical properties of this diamond material are highly consistent and
repeatable between synthesis runs.
wo 2011/076643 PCT /EP201 0/069828
43
5) A second layer, "Sample 1, Layer 2", was prepared usmg the following
conditions: gas flows of 290/250/230 seem for COz/CH4/Hz, a Cf:Of:Hf ratio of
0.21:0.22:0.57, a Cf:Of ratio of 0.95:1, a pressure of 184 x lOz Pa (138 Torr), nitrogen
5 present at 18 ppm of atomic nitrogen equivalent, and a substrate temperature of 830°C.
Of the atomic fraction of hydrogen in the source gases, 0.18 is added as Hz and 0.39 is
added from sources other than Hz. The Hf value of 0.57 gives a Pare value of
Pare= 170(Hf + 0.25) = 139 Torr.
10 The surfaces of the sample were processed sufficiently to facilitate optical
characterisation of the diamond material.
UV /visible absorption spectra were recorded by illuminating one of the polished side
surfaces of the sample such that the light path was entirely in either Layer 1 or Layer 2.
15 The pathlength used for obtaining the absorption spectra was roughly the lateral
dimension of the layer.
Prior to the optical characterisation, the sample was exposed to a deuterium lamp (15
Watt electrical power consumption) for 10 minutes with the sample approximately 80
20 mm from the filament of the bulb.
25
The experimental absorption spectrum was obtained as described in the body of the text
Figure 2 shows a comparison of the UV /visible absorption spectrum for Sample 1, Layer
2 with a spectrum from an HPHT type Ib synthetic diamond.
The experimental absorption spectrum was then de-convolved, as described elsewhere in
the specification, to determine the concentration of N8°. Figure 3 shows the measured
spectrum and the Type Ib component for Sample 1, Layer 2.
30 The measured optical absorption spectrum was integrated over the visible wavelength
range (i.e. 380 nm, equivalent to 3.2168 eV, to 750 nm, equivalent to 1.6527 eV) giving
a value (C) with units of eV.cm-1
. The absorption attributable to Ns0 was similarly
determined in eV.cm-1 over the same wavelength range giving a value (B). The ratio
wo 2011/076643 PCT /EP201 0/069828
44
B/C and the difference C-B are then calculated and are used for characterising the
material.
Sample 1 [Ns0
] Integrated Integrated Difference Ratio
(ppm) absorption in absorption in visible (C-B), (B/C)
visible (C), attributable to N s 0 (eV.cm-1
)
(eV.cm-1
) (B), (eV.cm-1
)
Layer 2 1.95 0.5989 0.5202 0.0787 0.87
5 The results of the optical analysis show that for "Sample 1, Layer 2", the proportion of
the optical absorption over the range 350 nm to 750 nm that is due to Ns0 is greater than
0.35 or 35%.
In addition, the absorption coefficients for the 360 nm and 510 nm bands were measured
10 from the deconvolved spectrum at the peaks of the respective bands.
Sample 1 Absorption coefficient (cm-1
)
360 nm band (cm-1
) 510 nm band (cm-1
)
Layer 2 0.1 0.2
The CIE L *a*b* coordinates for an optical pathlength of 1 mm were derived from the
absorption spectrum, in the way described in the detailed description of the invention,
15 and are tabulated below.
Sample 1 CIE L *a*b* Coordinates Hue angle at 1
at 1 mm thickness mm optical path
a* b* C* L* length (degrees)
Layer 2 -0.1 2.1 2.1 85.3 93.6
"Sample 1, Layer 2", made by the method of the invention, has a hue angle for an optical
pathlength of 1 mm of greater than 80° in addition to at least 35% of the optical
20 absorption in the visible spectrum being due to N8°.
5
wo 2011/076643 PCT /EP201 0/069828
45
Photoluminescence (PL) spectra from Sample 1, Layer 1 and Sample 1, Layer 2 were
recorded at 77 K excited using 514.5 nm light from a 50 mW Ar-ion laser. The ratio of
the intensities of the peaks at 637 nm and 575 nm were 0.8 and 1.4 for layers 1 and 2
respectively.
The inventors have found that the higher the value of the ratio of the 637 nm PL line to
the 575 nm PL line, the closer the optical absorption spectrum is to that of a pure Type
Ib component. In a diamond containing only N and NV centres, the ratio of NV-:NV0
(i.e. the ratio of the intensities of the 637 nm and 575 nm lines, 637 nm:575 nm) is
10 thought to be largely governed through the following equation:
N +NV (575 nm) -7 N+ +NV- (637 nm)
However in diamond material which is found to contain significant contributions from
15 absorptions at around 360 nm and 510 nm, another trap(s) competes for this electron
transfer. Using "X" to denote this trap(s), it is found that:
20 where N+ can be characterized through an absorption band with a peak at the onephonon
energy of 1332 cm-1
.
25
This competing electron trap mechanism results in the 637 nm:575 nm intensity ratios
reducing.
Example 2
The method of Example 1 was followed for steps 1) to 4)
30 A second layer of CVD diamond ("Sample 2, Layer 2") was deposited on the first layer
by gradually changing the input gas mixture to 375/430/290 seem of C02/CHJH2 at a
pressure of about 190 x 102 Pa (about 142 Torr) over a period of about 10 minutes. The
source gas additionally comprised 20 ppm of atomic N. The substrate temperature was
wo 2011/076643 PCT /EP201 0/069828
46
840°C. The synthesis conditions for Sample 2, Layer 2 have Cf:Of:Hf atomic fractions
of 0.21:0.19:0.60 and a Cf:Of ratio of 1.1:1. Of the atomic fraction of hydrogen in the
source gases, 0.15 is added as Hz and 0.45 is added from sources other than Hz. The Hf
value of 0.60 gives a Pare value of 144.5 Torr, so the operating pressure is approximately
5 2.5 Torr below P arc·
On completion of the growth period, the substrate was removed from the reactor and the
CVD diamond layer removed from the substrate, the top and bottom surfaces and two
opposed side surfaces of the CVD diamond layer were polished sufficiently for the layer
10 to be optically characterised. The final product was a layer of CVD diamond with a total
thickness of 2.4 mm (approximately equally split between Layer 1 and Layer 2), with
lateral dimensions of approximately 3.8 mm x 3.8 mm.
Prior to the optical characterisation, the sample was exposed to a deuterium lamp (15
15 Watt electrical power consumption) for 10 minutes with the sample approximately 80
mm from the filament of the bulb.
UV /visible absorption spectra were recorded by illuminating one of the polished side
surfaces of the sample such that the light path was entirely in either Layer 1 or Layer 2.
20 Thus the pathlength used for obtaining the absorption spectra was roughly the lateral
dimension of the layer.
25
The optical absorption spectra for Sample 2, Layer 1 and Sample 2, Layer 2 are shown in
Figure 4. The spectrum for Sample 2, Layer 2 was analysed as follows:
The Type Ib component was de-convolved from the measured spectrum (Figure 5).
The measured optical absorption spectrum was integrated over the visible wavelength
range (i.e. 380 nm, equivalent to 3.2168 eV, to 750 nm, equivalent to 1.6527 eV) giving
30 a value with units of eV.cm-1 (C). The absorption attributable to Ns0 was similarly
determined in eV.cm-1 over the same wavelength range (B). The ratio B/C and the
difference C-B are then calculated and are used for characterising the material. The
values for Sample 2, Layer 2 are given below.
5
10
15
wo 2011/076643 PCT /EP201 0/069828
47
Sample 2 [Ns0
] Integrated Integrated Difference Ratio
(ppm) absorption in absorption in visible (C-B), (B/C)
visible (C), attributable to N s 0 (eV.cm-1
)
(eV.cm-1
) (B), (eV.cm-1
)
Layer 2 1.25 0.737 0.3334 0.4037 0.45
For "Sample 2, Layer 2", the proportion of the absorption over the range 380 nm to 750
nm due to Ns0 is 0.45 or 45%.
In addition, the absorption coefficients for the 360 nm and 510 nm bands were
determined by further deconvolution of the spectrum as described elsewhere in the
specification.
Sample 2 Absorption coefficient (cm-1
)
360 nm band (cm-1
) 510 nm band (cm-1
)
Layer 2 0.3 0.3
The CIE L *a*b* coordinates were derived from the absorption spectrum, in the way
described in the detailed description of the invention. The values tabulated below are
those calculated from these values for an optical pathlength of 1.0 mm.
Sample 1 CIE L *a*b* Coordinates Hue angle at 1. 0 mm
a* b* C* L* optical pathlength, degrees
Layer 2 0.4 2.1 2.0 86.2 81.2
PL spectra from Sample 2, Layer 1 and Sample 2, Layer 2, recorded at 77 K using 488.2
nm light from a 50 mW Ar-ion laser, are shown in Figure 6. The spectra have been
normalised by ratioing the integrated area beneath the first-order Raman line at 521.9 nm
in Figure 6. The PL spectrum for Sample 1, Layer 2 shows a peak at 543.1 nm that is
20 absent in Sample 1, Layer 1, and is believed to be related to the use of large fractions of
oxygen in the synthesis process.
wo 2011/076643 PCT /EP201 0/069828
48
PL spectra recorded at 77 K using 514.5 nm light from a 50 mW Ar-ion laser from
Sample 2, Layer 1 and Sample 2, Layer 2, are shown in Figure 7. The ratio of the peaks
at 637 nm and 575 nm are 0.7 for Sample 2, Layer 1, and 1.1 for Sample 2, Layer 2. The
higher value of this ratio for Sample 1, Layer 2 compared with Sample 2, Layer 1
5 indicates that Sample 2, Layer 2 has a higher concentration of Ns0 in the material.
Example 3
Sample 3 was prepared using the same sequence of steps described in Example 1 except
10 that the following conditions were used for step 5 to form "Sample 3, Layer 2": gas
flows of 501/604/500 seem for COz/CH4/Hz, a Cf:Of:Hf ratio of 0.20:0.18:0.62, a Cf:Of
ratio of 1.11:1, a pressure of 190 x 102 Pa (about 143 Torr), nitrogen present as 15 ppm
of atomic nitrogen equivalent, and a substrate temperature of 860°C. The of the
hydrogen atomic fraction of 0.62, 0.18 was added as H2 and 0.44 was added in other
15 forms (in this case as CH4). The Hf value of 0.62 gives a value for Pare of 148 Torr, so
the operating pressure is 5 Torr less than P arc·
For this sample, the amount of silicon incorporated in each of the layers was
characterised using the intensity of the 737 nm PL line that is believed to be associated
20 with the silicon-vacancy defect. Figure 8 shows a PL image (excited with 633 nm
radiation from a He-Ne laser) of the cross-section (substrate to the left-hand-side of the
image) overlaid with the intensity plot for the 737 nm line. This example demonstrates
that the method of the invention suppresses the uptake of Si into the material compared
with H2/CH4 chemistries. The peak intensity of the 737 nm line is near the interface
25 between substrate and the first diamond layer and is typically associated with higher
levels of contaminants during the first stage of diamond growth, for example due to
exposure of sources of silicon within the growth environment.
Analysis ofPL spectra obtained at 77 K using 514.5 nm light from a 50 mW Ar-ion laser
30 for "Sample 3, Layer 1" and "Sample 3, Layer 2" gave ratios of the intensities of the 637
nm and 575 nm peaks of 0.5 and 1.1 respectively.
5
wo 2011/076643 PCT /EP201 0/069828
49
Example 4
This comparative example demonstrates the effect that not providing some of the
hydrogen atoms in the form of Hz molecules has on the synthesis process.
The procedure set out in Example 1 was repeated with the following variation m
conditions for step 5 to form "Sample 4, Layer 2": synthesis conditions of COz/CH4
based plasma growth process without interruption by a ramped change in gas
composition and process (pressure and power) window. The final pressure (limited by
10 control issues due to no Hz addition) was fixed at 130 x lOz Pa (about 97 Torr), with a
Cf:Of ratio of 1.07:1, and Cf:Of:Hf ratios of 0.246:0.229:0.525. The gas flows were
375/430 seem of COz/CH4 . The source gas additionally comprised 20 ppm of atomic N.
Substrate temperature was 810 oc. The proportion of Hf that was added in the form of
Hz molecules was zero. For the Hf fraction in the source gas for the synthesis of Sample
15 4, Layer 2, the value of P arc would be expected to be 132 Torr; this is substantially higher
than the operating pressure of about 97 Torr and is believed by the inventors to be due to
the absence of H added as Hz in the source gas mixture.
The UV/visible optical absorption spectrum for Sample 4, Layer 2 is shown in Figure 9.
20 In contrast to the previous examples grown at the higher pressures made possible by the
addition of hydrogen as Hz to the source gas mixture, the spectrum from Sample 4,
Layer 2 shows considerable absorption in addition to the type lb component.
The optical properties of Sample 4, Layer 2 were deconvolved, as previously described,
25 to determine the concentration of N8°. The hue angle was measured and converted to the
hue angle for an optical pathlength of 1 mm.
Sample 4 CIE L *a*b* Coordinates at 1 Hue angle for 1 mm
mm optical path length
a* b* C* L* (degrees)
Layer 2 1.7 4.8 5.0 81.3 70.5
wo 2011/076643 PCT /EP201 0/069828
50
Sample 4 [Ns0
] Integrated Integrated absorption Difference Ratio
(ppm) absorption in in visible attributable (C-B) (B/C)
visible (C) to Ns0 (B), (eV.cm-1
) (eV.cm-1
)
(eV.cm-1
)
Layer 2 2.8 5.314 0.75 4.564 0.14
The ratio B/C of 0.14 means that the proportion of the optical absorption in the visible
region that is due to Ns0 is only 0.14 or 14%, very much less than for those samples
prepared by the method of the invention, demonstrating the importance of H2 and the
5 desirability of operating at high pressures.
The ratio of the peaks at 637 nm and 575 nm is 0.7 for Sample 4, Layer 1, and 1.0 for
Sample 4, Layer 2.
10 In addition, the absorption coefficients for the 360 nm and 510 nm bands were
determined by further deconvolution of the measured spectrum as described elsewhere in
this specification.
Sample 4 Absorption coefficient (cm-1
)
360 nm band (cm-1
) 510 nm band (cm-1
)
Layer 2 3.2 1.5
15 Example 5
20
This example demonstrates that provided the atomic fractions of 0, C and H in the
source gas are the same, the optical properties of the diamond material that is produced
will be substantially similar.
The methodology used in steps 1) to 3) of Example 1 was repeated to produce two
further samples, referred to as "Sample 5, Layer 2" and "Sample 6, Layer 2". The
starting gas compositions for Sample 5, Layer 2 and Sample 6, Layer 2 are summarized
in the table below (gas flows in seem). The table also shows the fractions of C, H and 0
25 in the gas phase.
wo 2011/076643 PCT /EP201 0/069828
51
Sample Flow, seem Proportion, atomic fraction
COz co Hz CH4 Total c H 0
Sample 5, 500 0 500 590 1590 0.20 0.62 0.18
Layer 2
Sample 6, 0 614 902 75 1591 0.20 0.62 0.18
Layer 2
The starting gases were chosen to result in the same C, H and 0 atomic fractions in the
5 plasma. The Cf:Hf ratio is 1.11:1. For both samples, thick single crystal diamond bodies
were produced (3.7 mm for Sample 5 and 3.6 mm for Sample 6), grown at a gas pressure
of 170 x 102 Pa (approximately 127 Torr), with the source gas also comprising 14 ppm
of equivalent atomic nitrogen. In this case P arc for Hf = 0.62 is 148 Torr.
10 After completion of the synthesis process, Sample 5 and Sample 6 were processed such
that they could be optically characterised. UV -visible absorption spectra were obtained.
The optical absorption spectrum was deconvolved, as described elsewhere in this
specification, to determine the concentration of N8°. The absorption coefficients for the
360 nm and 510 nm bands were measured at the peaks of the respective bands. The key
15 parameters are given in the table below.
Sample Ns0 360 nm band 510 nm band
(ppm) (cm-1) (cm-1)
Sample 5 1.5 0.5 0.45
Sample 6 1.2 0.4 0.35
The deconvolution and absorption coefficients of the optical absorption spectrum show
that different gas mixtures give essentially the same result in terms of the absorption due
20 to each of the components.
wo 2011/076643 PCT /EP201 0/069828
52
PL spectra were obtained from Sample 5 and Sample 6 using 514.5 nm excitation from
an Ar-ion laser. The ratio of the peaks at 637 nm and 575 nm are 0.94 for Sample 5 and
0.84 for Sample 6.
5 PL spectra obtained from Sample 5, Layer 2 and Sample 6, Layer 2 using 458 nm
excitation are shown in Figure 10. Both the characteristics and the relative intensities of
the spectral features are very similar; similar results were found at other PL excitation
wavelengths studied (325 nm, 488 nm, 515 nm and 660 nm).
10 A projection X-ray topograph of Sample 6 is shown in Figure 11. There is little X-ray
contrast indicating high crystalline quality and low dislocation density. This material
property makes the material suitable for some optical and mechanical applications.
15
Example 6
In this example, the variation of the CIELAB parameters as a function of optical path
length is shown. These results are derived from the model described and referred to in
the specification.
20 In this case, the sample was produced according to Example 1, Layer 2 and the
absorption spectrum required for calculating the CIE L *a*b* coordinates was obtained
from a 1.37 mm thick sample.
Thickness (mm) 0.5 1 1.5 2 3 4 5 6 10
L* 86.4 85.3 84.3 83.4 81.4 79.5 77.6 75.8 68.8
a* -0.1 -0.1 -0.2 -0.2 -0.3 -0.3 -0.2 -0.2 0.4
b* 1.1 2.1 3.2 4.2 6.1 7.9 9.6 11.2 16.6
c* 1.1 2.1 3.2 4.2 6.1 7.9 9.6 11.2 16.6
Hue angle, o 93.9 93.6 93.4 93.1 92.5 91.9 91.4 90.8 88.8
25 It can be seen that, in particular, the hue angle is greater than 80° for all the thicknesses
where it has been calculated.

Claims
1. A chemical vapour deposition (CVD) method for synthesizing diamond material
on a substrate in a synthesis environment, said method comprising:
5 providing the substrate;
providing a source gas;
dissociating the source gas; and
allowing homoepitaxial diamond synthesis on the substrate;
wherein the synthesis environment comprises nitrogen at an atomic concentration of
10 from about 0.4 ppm to about 50 ppm; and
15
20
and wherein the source gas comprises:
a) an atomic fraction of hydrogen, Hf, from about 0.40 to about 0.75;
b) an atomic fraction of carbon, Cf, from about 0.15 to about 0.30;
c) an atomic fraction of oxygen, Of, from about 0.13 to about 0.40;
wherein Hf + Cf +Of= 1;
wherein the ratio of atomic fraction of carbon to the atomic fraction of oxygen,
Cf:Of, satisfies the ratio of about 0.45:1 < Cf:Of