WO1994012680A1

WO1994012680A1 - Doping of highly tetrahedral diamond-like amorphous carbon

Info

Publication number: WO1994012680A1
Application number: PCT/GB1993/002424
Authority: WO
Inventors: Gehan Anil Joseph Amaratunga; David Robert Mckenzie
Original assignee: Gehan Anil Joseph Amaratunga; David Robert Mckenzie
Priority date: 1992-11-25
Filing date: 1993-11-25
Publication date: 1994-06-09
Also published as: GB9224697D0; AU5532194A

Abstract

A method of forming doped highly tetrahedral diamond-like amorphous carbon is disclosed in which a cathode (1) made of carbon is ionised in an evacuated chamber (2) with a dopant present in the chamber. A dopant is ionised and the carbon ions and dopant ions are collected on a substrate (4). Suitable dopants include B, P, Fe, Al, Au, Ag, N, and MoS2.

Description

DOPING OF HIGHLY TETRAHEDRAL DIAMOND-LIKE AMORPHOUS CARBON

The present invention relates to doping of highly tetrahedral diamond-like amorphous carbon.

A filtered cathodic arc is an efficient source of hydrogen free carbon plasma which can be used to deposit extremely hard and resistive thin films on a variety of substrates. These films have been denoted tetrahedral amorphous carbon (ta-C) or "amorphous diamond" due to the high proportion (generally greater than 50% and less than 90%, and more particularly greater than 65% and less than 85%, ±10%) of tetrahedral (sp ) bonds which give a structure analogous to amorphous silicon and properties similar to those of diamond, see, for example, McKenzie, D.R., Muller, D.A. & Pailthorpe, B.A. Phys . Rev. Lett . 67, 773 (1991) and Robertson, J. Prog. Solid St . Chem . 21, 199 (1991) . Work to date on the electronic properties of ta-C has been limited to studies of the intrinsic material, which is weakly p-type.

Hydrogenated amorphous silicon (a-Si:H) has become an important semiconductor because of its high sensitivity to light and the ability to deposit it at low temperatures and over large areas. Solar cells and large transistor arrays for switching display elements in portable computers and television sets are examples of where it is widely used. An amorphous semiconductor with a wider band gap than that of Si, which could operate at higher electric fields and temperatures, would be useful in its own right as well as in combination with Si for a number of electronic devices. The use of another group IV material such as carbon to form a-C:H would at first sight seem to be a viable alternative. However, attempts to form a semiconductor which can be doped using a-C:H have been unsuccessful because, unlike a-Si:H, a-C:H contains a high proportion of non-tetrahedral (sp ) bonding which prevents the formation of a semiconductor band gap. The equivalent tetrahedral form of amorphous carbon (ta-C) is produced by cathodic arc deposition in the absence of hydrogen.

Attempts have previously been made to dope ta-C. See for example Amaratunga, G.A.J. , Segal, D.E. & McKenzie, D.R. Appl .Phys . Lett .59, 69 (1991) in which attempts to dope ta-C using a dopant in the gaseous phase introduced during vacuum deposition of the carbon are reported as being unsuccessful.

According to a first aspect of the present invention, there is provided a method of forming doped highly tetrahedral diamond-like amorphous carbon, the method comprising the steps of: ionising a cathode made of carbon in an evacuated chamber with a dopant present in the chamber; is characterised by: ionising the dopant; and, collecting the carbon ions and dopant ions on a substrate.

Generally, the present invention allows a ta-C film to be produced which has been doped in a controlled way to control its electronic, optical, magnetic and/or tribological properties. The film generally has a high proportion (generally greater than 50% and less than 90%, and more particularly greater than 65% and less than 85%, ±10%) of tetrahedral (sp^z) bonds.

The dopant may be incorporated into the cathode or provided as a secondary source in the chamber.

Alternatively, the dopant may be introduced as a gas into the chamber; the gas is preferably ionised prior to its introduction into the chamber.

The dopant may be such as to produce n-type or p-type doped highly tetrahedral diamond-like amorphous carbon.

The dopant may be phosphorous. The cathode may consist of up to substantially 5% by mass of phosphorous and preferably up to substantially 1% by mass of phosphorous to 99% carbon. The dopant may be boron. The boron may be introduced into the chamber as boron ions by ionising a compound including boron, such as B₂H₆. Boron powder may alternatively be incorporated in a carbon cathode. The dopant may alternatively be aluminium or an aluminium- containing compound.

The dopant may be aluminium or an aluminium-containing compound, or iron or an iron-containing compound (eg Fe₂0₃) , or at least one of Au and Ag, or nitrogen, or MoS₂. According to a second aspect of the present invention, there is provided apparatus for producing doped highly tetrahedral diamond-like amorphous carbon, the apparatus comprising: a vacuum chamber; a source of carbon; means for ionising the carbon; a substrate; means for directing the carbon ions into the chamber to the substrate; and, a source of dopant; characterised by: means for ionising the dopant.

The carbon source may be a solid. The dopant source may be incorporated into the solid carbon source. Alternatively, the dopant may be incorporated into a separate, secondary source within the vacuum chamber.

Alternatively, the dopant ions may be generated from a gaseous source of the dopant outside of the chamber. The ions are introduced into the chamber as the carbon ions are produced from the carbon source.

An example of the present invention will now be described with reference to the accompanying drawings, in which:

Fig. 1 is a composite diagram of alternative examples of apparatus for use in the present invention;

Fig. 2 is a current-voltage curve for a phosphorous- containing ta-C film deposited on an insulating quartz substrate;

Fig. 3 shows Arrhenius plots for a phosphorous- containing ta-C film; Fig. 4 shows current-voltage curves of heterojunctions between ta-C and 1.5 Ωcm n-type silicon for (i) undoped ta-C and (ii) phosphorous-containing ta-C;

Fig. 5 shows the band levels for a rectifying junction between n-type silicon and phosphorous-containing ta-C;

Fig. 6 is a graph of nitrogen content in a nitrogen- doped film;

Fig. 7(a) to (d) are graphs of activation energy, electrical resistivity, Tauc optical gap, and compressive stress for varying N₂ partial pressure respectively;

Fig. 8 is a graph showing the C K-edge spectra for a range of films having different level of nitrogen doping; and,

Fig. 9 is a graph showing a comparison between C K-edge and N K-edge spectra for a 10% N doped ta-C film.

In Figure 1, variations of the apparatus suitable for the present invention are shown. A cathode 1, described in more detail below, is used as a source of at least carbon ions which are directed into a vacuum chamber 2 through a curved magnetic plasma guide 3 as suggested by Veerasamy,

V.S. et al (presented at a conference "Diamond Films '92" in Heidelberg, Germany on 31 August to 4 September 1992 and published in "IEEE Transactions on Plasma Science",

Vol. 21, no.3, pp322-328) . The ions are directed towards a substrate 4 within the vacuum chamber 2.

The cathode was fabricated from a mixture of 1% by mass of 99.999% pure red phosphorous powder in 99.999% pure graphite powder. The powders were ultrasonically mixed and compressed into a 50mm diameter disc under 20-30MPa pressure.

A cathodic arc was struck with the phosphorous- containing carbon disc as the cathode 1 and using a high voltage source 5 to an electrode 6 in close proximity to the cathode 2. The vacuum base pressure in the chamber 2 was 10^" Pa and care was taken to eliminate water vapour and oxygen from the system. The arc voltage of 32-35V at a current of 60A was significantly larger than the 19-24V observed for a pure graphite cathode. Once the arc is struck, it is self-sustaining and the power supply can be removed.

In order to eliminate micron-sized solid particles also emitted from the cathode 1, the plasma passes through a 90° curved solenoid as the plasma guide 3 with an axial magnetic field of 30mT. The substrate 4 is held at a negative voltage of approximately -30V.

Films of 30-35nm thickness were deposited at a rate of l-2nms^' onto (100) silicon and quartz as the substrate. The species P⁺, P and P₂ were observed by a mass spectrometer in the deposition chamber 2.

The advantage of introducing the phosphorous at the cathode 1 is that it enables phosphorous-containing carbon films to be grown from ionic species. The phosphorous and carbon ions are emitted from the cathode 1 with 20-50eV kinetic energy and are further accelerated by the bias voltage of the substrate 4. Under these conditions a significant fraction of the phosphorous ions will be implanted a few monolayers beneath the surface of the growing film.

A typical current-voltage curve measured across a phosphorous containing film deposited on an insulating quartz substrate is shown in Figure 2. Electrical contact to the film was made through 25nm thick gold coatings which were thermally evaporated under a vacuum of 10^' Pa. A shadow mask was used to restrict the gold contacts to a 5mm wide strip containing the 0.5mm gaps across which the measurements were made. Leakage current through the quartz substrate was less than 10^" A. The linearity of this curve clearly demonstrates that there is an ohmic contact between the phosphorous-containing carbon film and the gold contact. The resistivity of the film was calculated to be 5Ω.cm at room temperature. This is more than 6 orders of magnitude lower than the resistivity of undoped ta-C on the same substrates, which is typically 10 Ω.cm. Undoped ta-C also shows non-linear space charge limited current flow characteristics below |2V|.

The conductivity of the phosphorous containing film of Figure 2 was measured as a function of temperature between 13OK and 33OK. The Arrhenius plot (a) in Figure 3 shows conductivity in the dark and shows that the dark conduction in the phosphorous containing film is thermally activated with an activation energy of 0.13eV in the temperature range 20OK to 33OK. At temperatures below 16OK, the conductivity had an additional component which gives the logarithm of conductivity a T -1/4 dependency which is typical of hopping conduction in amorphous materials.

Curve (b) in Figure 3 shows the Arrhenius plot which is obtained when the sample is exposed to AM-1 light. The photoconductive component of the conductivity is larger than the thermally activated conduction component only for temperatures below 200K, which is typical of a doped semiconductor. Preliminary Hall effect measurements on the same sample gave a Hall voltage consistent with conduction by electrons. The carrier density at room temperature estimated from the Hall voltage is of order 10 cm^" . The carrier density when combined with the conductivity gives an electron mobility in the extended states of around 10cm V^" s^" . Together with the dramatic change in conductivity, the low activation energy suggests that phosphorous has been incorporated into the film and is electronically active. The sample was not thermally annealed.

Heterojunction diodes which are formed by depositing ta-C on silicon can give useful information about the band structure and defects in the material, as mentioned in Amaratunga, G.A.J., Segal, D.E. & McKenzie, D.R. Appl .Phys . Lett .59, 69 (1991) . The current-voltage curve of an n-type silicon to undoped ta-C heterojunction diode is shown as curve (i) in Figure 4. At positive voltages, the junction is forward biased and there is majority carrier (electron) injection from the silicon into the ta-C. In contrast, curve (ii) in Figure 4 shows that, when the film contains phosphorous, the polarity of the device is reversed. Injection of electrons now occurs from the phosphorous-containing film into the n-type silicon, as would be expected if incorporation of phosphorous has moved the Fermi level in the carbon film close to the conduction band (see Oldham, W.G. & Milnes, A.G. Solid State Electronics 7, 153 (1964) ) . This shows that an n-n heterojunction has been formed as represented by the band diagram in Figure 5. Note that for forward voltages greater than 2V, the current injection from the phosphorous containing film into the silicon is limited by resistance at the gold-carbon interface. The energy difference between the conduction band and the Fermi level is taken to be O.leV in the ta-C and 0.3V in the silicon. The Fermi level in the ta-C was estimated on the basis of a 10 c carrier density and a density of states of I0¹⁸cm^*3 at the conduction band edge calculated from space charge limited current data. Contact to the film was made through 1mm diameter vacuum evaporated gold contacts to form Au/carbon film/Si sandwich structures. The silicon substrates were pressure contacted onto a large metal base which was held at earth potential. The bias voltage was applied to the gold contact and the current-voltage characteristic measured using a Hewlett Packard 4140B pA meter/DC voltage source.

A free-standing film was analysed by transmission electron microscopy and parallel electron energy loss spectroscopy (PEELS) . The diffraction pattern showed diffuse rings characteristic of an amorphous material. The plasmon energy, which is related to the valence electron density, was found to be 30.9eV. Detailed structural investigation of ta-C has shown that a plasmon energy of 30.5eV corresponds to about or more than 90% tetrahedral bonding in the film. Energy dispersive spectroscopy (EDS) in a scanning electro microscope operated at 2KeV confirmed the pressure of phosphorous in a ta-C film on the quartz substrate. No other impurities were detected. These results indicate that the material retains essentially the same structure as ta-C. The fact that there was sufficient phosphorous in the films for detection by EDS, but an insufficient amount for detection by PEELS, suggests that the atomic fraction of phosphorous in the films is between 0.1% and 1%. Using a carrier density of 10 cm^" from the Hall measurements and a density of 3 gem^" for the ta-C, for a phosphorous content between 0.1% and 1%, a doping efficiency of the order of 10 -3 to 10-2 is estimated at room temperature.

The dramatic increase in the conductivity of ta-C films containing phosphorous, together with the fact that the highly tetrahedral amorphous structure of intrinsic ta-C is retained, shows that ta-C films have been successfully doped. The properties of heterojunctions with silicon, together with Hall effect and activation energy measurements, are consistent with shallow n-type doping of ta-C incorporation of phosphorous during growth. Where the dopant is a solid, it can easily be incorporated into the cathode. Accordingly, materials such as aluminium can be used instead of the phosphorous described above.

Another dopant which may be incorporated into the solid cathode is MoS₂, which may be incorporated as a powder. This would give a ta-C film mixed with MoS₂ to enhance the lubrication and wear properties of the films.

Further dopants which may be incorporated into the solid cathode include metals such as Ag and Au, which would allow a mixed ta-C metal film, the optical transmission properties of which can be controlled. These films would have application in optical coatings (e.g. for sunglasses, windows) where the ta-C matrix provides a hard mechanical coating and the metal particles allow for optical absorption. Furthermore, Au and Ag are known to have surface plasmon modes which can emit light. By confining Ag, Au or a similar material in small particle form, it would be possible to enhance the photoemission properties of the ta-C/Au(Ag) films. This may have application in electroluminescent cells.

Fe or Fe-containing compounds (such as Fe₂0₃) can also be used as a dopant in order to control the magnetic properties of the ta-C film. This may provide a mechanically hard magnetic film, which may have use for fabricating magnetic disks and tapes.

Where the dopant is a gas, this can be introduced into the system as a gas in the vacuum chamber in which deposition takes place. Apparatus suitable for this method is shown in Figure 1. In particular, the cathode 1 is substantially pure carbon. A Kaufmann ion source 7 is connected to the vacuum chamber 2 to form dopant ions from gaseous sources supplied to the Kaufmann source 7, and the Kaufmann ion source 7 directs those ions into the vacuum chamber 2. As the carbon cathode 1 is ionised and the carbon ions are accelerated towards the substrate 4, the dopant ions from the Kaufmann source 7 are similarly accelerated towards the substrate 4 and a doped film of ta-C builds up on the surface of the substrate.

For example, B₂H₆ can be used as the gaseous source to introduce boron ions in this way. Other suitable gaseous dopant sources include PH₃, AsH₃, N₂. As a further alternative, a gaseous dopant may be introduced into the apparatus via a leak valve (not shown) in the bend region 10 of the curved magnetic plasma guide 3, used to remove macroparticles and neutrals from the plasma stream. For example, N₂ (99.9995% pure) was introduced into the vacuum chamber in this way. The N₂ partial pressure was monitored using an ion gauge and mass spectrometer assembly fitted to the deposition chamber

-₄ unit. The pressure during deposition varied between 10 to 10^" bar (10^" to 10^" Pa) and hence only the initial N₂ partial pressure and flow rate prior to arcing was used as the controllable parameter. The arc current and voltage used were 60A and 20V respectively. The N₂ background partial pressures were varied from below 10^" bar to 10^"2 mbar (lθ^"5 to 1 Pa) . Dissociation and ionisation of the background gas (N₂ in this case) is effected by the highly energetic ions and electrons from the plasma stream. Evidence for the presence of N ions in the plasma was obtained from optical spectroscopic analysis which showed the most likely N²P→⁴S transition at 346.6 nm. Highly reactive N⁺ gas ions were thus incorporated into the growing films. The substrates were left at floating potential and their temperature never exceeded 80°C during deposition.

Films with thicknesses ranging from 50 to 100 nm and doped with nitrogen in this manner were grown on quartz and silicon (100) substrates. Using a photo-resist patterning and lift-off technique, each substrate was partly coated from the plasma stream so as to allow step measurements of film thickness using a DekTak profilometer. Film thickness on silicon substrates was also estimated by ellipso etry. The N content in the films was determined using X-ray Photoelectron Spectroscopy (XPS) carried out on a Perkin Elmer ESCA-system, model 5500, coupled with a spherical capacitor analyser (SCA) with a pass energy set to 117.4 eV, giving an energy resolution of the order of ΔE=2.3 eV. The X-ray source was operated at 250 W (15kV) using a Mg anode. The source was positioned at an angle of 54° with respect to the import lens axis of the analysing system kept at a distance of 6 mm from the sample. The sample positioned was at an angle of 45° with respect to the lens axis and the area analysed at a time was 0.8-1.2 mm 2. The data evaluation was carried out on the basis of empirical sensitivity factors taken to be 0.477 for N in the present system. The detection limit of this technique is 0.2 at% N, which is sensitive enough for the detection of low levels of N(<1%) . Analysis of ta-C samples with no intentional leakage of N₂ gas yields concentrations of 99.9% C with traces of O and Ar. The traces of Ar are attributed to the sputter-cleaning of the ♦ sample surface by Ar bombardment prior to the XPS measurements. Within the partial pressure range of 10^" - 10^* mbar (10^*5 to lθ^" Pa) , the lowest observable N concentration was 0.25% and the highest concentration of N was found to be about 1% (Figure 6) . The N content was below the level of detection in the ta-C film with no intentional introduction of N. The resistivities of the films were found to vary as a function of N₂ partial pressure and hence N content. Films doped with nitrogen as described above and deposited on quartz were used for both optical and electrical measurements. The variation of film resistivity with N₂ partial pressure is shown in Figure 7(b). Changes in resistivity as well as the thermopower were also measured as a function of temperature over the range 180- 500 K. The variation of electrical conductivity with temperature provides important information on the conduction mechanism in the bulk material. All films containing below and above 0.45 % of N show a linear dependence in log(conductivity) with reciprocal of thermodynamic temperature, indicative of thermally activated conduction. The gradient of the log(σ) vs (1/T) lines change with the differing N contents in the film. Ascribing these gradients over the range of temperature measured to the corresponding activation energies E_a, an increase in activation energy is first observed followed by a gradual decrease with increasing N partial pressures, as shown in Figure 7(a) . The film with the highest resistivity, containing 0.45 N%, displayed a linear relationship between log(σ) and T -1/4 over a wide temperature range, characteristic of variable range hopping.

In amorphous semiconductors, thermopower measurements are essential to determine whether doping has been achieved and, if it has, whether the carriers are electrons or holes. From our results, there is a clear change in the sign of the thermopower from positive to negative in the case of undoped films compared to films containing 2% N. In the latter films, the offer of magnitude of thermopower values (in the mV/K range) suggests a conduction mechanism involving electrons thermally activated into the conduction band, corresponding to regular n-type doping. In the case of undoped films, the thermopower is positive and of the order of 0.1 mV/K indicating that conduction takes place via valence band extended states. In the film with the highest resistivity (0.45 %N; N₂ partial pressure 10^" mbar (10 -4 Pa)), the thermopower i.s in the μV/K regime which suggests a conduction mechanism around the Fermi level and correlates with the variable range hopping type mechanism suggested by the T 1/4 dependence discussed above.

The optical bandgap E_g was determined for nitrogen- doped films grown under similar conditions on quartz substrates. The corresponding absorbtion coefficients of the thin films in the range of 30-50 nm were measured in the wavelength range 190-750 nm. E_g was determined using a Tauc plot, the usual procedure for amorphous materials. The variation of E_g with N content is shown in Figure 7(c) . Thus, within the limits of experimental error, the optical bandgap decreases slightly to 1.8 eV with up to 1.5% N. With 10% incorporation, the optical bandgap reduces to 1.5 eV. Using the reflectance and transmission of the ta-C thin films, the real and the imaginary parts of the refractive index have also been extracted. This allows for the subsequent calculation of the imaginary permittivity e₂₀ of ta-C. The optical data is compared with the EELS data below. in order further to establish clearly whether chemical doping with nitrogen has taken place, a comparative study of ta-C films with and without doping was made using EELS. Thin films grown on silicon were lifted off using a mixture of HF, HN0₃ and acetic acid. Each set of films was placed on electron microscope support grids. The typical resolution of the spectra was 0.5 eV. The energy loss spectra of the valence excitation as well as the C K-edges were collected with a focused probe scanning over a large area so as to average over any possible composition fluctuations. The convergence angle of the electron probe was 5 mrad and the spectrometer collector subtends a semi- angle of 7 mrad. With a total beam current of the order of 10^" A and a defocused probe, no sign of radiation-induced change in the energy-loss spectra was observed. Thus, we are confident that the results obtained are intrinsic to the samples under investigation. Four samples were examined, the first (a) undoped, the rest containing (b) 0.45%, (c) 1% and (d) 10% N respectively. The first three samples cover the entire range of N concentrations where controlled doping has been achieved.

The C K-edge spectra of these ta-C films are shown in Figure 8 where they have been deconvoluted within the valence electron spectra obtained from the same area to remove both the multiple scattering contribution as well as the effect of instrumental response profile. A comparative study was made between the undoped films and N-doped films. For the 0.5% N sample, both the pre-edge as well as the shape of the broad peak is very similar to that of the undoped film. In the case of the 1% N-doped sample, the shape of the pre-edge remains unchanged within the experimental error, but the main absorption peak centred around 292 eV becomes broadened. After taking into account the latter observation, the area under the ls→2p(π*) peak is constant, indicating that the effect of the N doping on the film at this level is minimal and local as far as the sp²/sp³ fraction is concerned. The broadening of the peak at 292 eV with N doping up to 1% N could therefore be attributed to an increased range of disorder in the σ- bonded structure, instead of a growth of the fraction of sp² bonds. The C K-edge spectrum in the case of the sample containing 10% N shows a very different fine structure from the others. The pre-edge peak is not only increased in intensity, but its position is also shifted to a lower energy by 0.5 eV and becomes broadened. This suggests that 2 excessive N doping promotes a disordered sp structure. Another noteworthy feature is that the N K-edge shows a structure similar to that of the C K-edge (Figure 9) , indicating that the N atoms do not appear exclusively as either sp 2 or sp3 bonded si•tes, but rather on average sample the same local environment as that of the C atoms. Small differences do appear however in the fine structure of the C and N K-edge, but this probably reflects atomic differences between N and C. The net role of N is consistent with the fact that the increase in the fraction of sp 2-bonded C i.s much higher than the i.ncrease in N concentration.

The compressive stress in these films has been identified to be an important parameter which is strongly correlated with the sp bonding in the material and the stress with differing N₂ partial pressures and hence flow rate has therefore been determined. Measurements have been made for three independent runs over the same range of N₂ partial pressures. The stress in the films has been calculated using a method in which the curvature of the silicon {100} substrate is measured prior to and after deposition using a Dektak 3030 stylus profilometer. The subsequent use of Stoney's equation allows calculation of the film stress. The stress is maintained and interestingly rises initially with increasing gas flow rate, and then at higher flow rate starts to decrease (Figure 7(d)). We conclude that the sp fraction in the films is preserved with low N incorporation. The reason for a decrease in stress at high pressures of N is two- fold: a reduction in the energy of the incident C ions in collision with N₂ gas or the incorporation of a high level of N causes stress relaxation. A control experiment with argon gas was also performed and showed a monotonic decrease in stress with increasing N₂ gas pressure, implying that the initial increase in stress in N-doped films is due to the incorporation of N. As an alternative to introducing the dopant as a solid incorporated into the cathode 1, the cathode 1 may be a substantially pure carbon source and the solid dopant may be present in a secondary target 8. An RF supply 9 generates an RF field which is followed by the lighter mass electrons travelling with the plasma beam, causing the secondary source 8 to develop a DC bias relative to the vacuum arc plasma beam. Ions from the main beam are thereby attracted to the secondary source 8, sputtering that source 8 to produce dopant ions in the vicinity of the substrate 4. The dopant ions and carbon ions then deposit on the substrate, itself held at a DC bias voltage, to form doped ta-C. Almost any solid dopant can be introduced in this manner, including Fe, Au, Ag and MoS₂ mentioned above, either instead of or in addition to incorporation of the dopant into the cathode 1.

Claims

1. A method of forming doped highly tetrahedral diamond¬ like amorphous carbon, the method comprising the steps of: ionising a cathode (1) made of carbon in an evacuated chamber (2) with a dopant present in the chamber; characterised by: ionising the dopant; and, collecting the carbon ions and dopant ions on a substrate (4) .

2. A method according to claim 1, wherein the dopant is incorporated into the cathode (1) .

3. A method according to claim 1 , wherein the dopant is provided as a secondary source (8) in the chamber (2) .

4. A method according to claim 1, wherein the dopant is introduced as a gas into the chamber (2) .

5. A method according to claim 4, wherein the gas is ionised prior to its introduction into the chamber (2) .

6. A method according to any of claims 1 to 5, wherein the dopant is phosphorous.

7. A method according to claim 2, wherein the cathode consists of up to substantially 5% by mass of phosphorous.

8. A method according to claim 1 , wherein the cathode consists of up to substantially 1% by mass of phosphorous to 99% carbon.

9. A method according to any of claims 1 to 5, wherein the dopant is boron.

10. A method according to claim 1, wherein the dopant is boron which is introduced into the chamber as boron ions by ionising a compound including boron.

11. A method according to claim 10, wherein said compound is B₂H₆.

12. A method according to claim 2, wherein boron powder is incorporated in a carbon cathode.

13. A method according to any of claims 1 to 5, wherein the dopant is aluminium or an aluminium-containing compound.

14. A method according to any of claims 1 to 5, wherein the dopant is iron or an iron-containing compound.

15. A method according to claim 1 or claim 2, wherein the dopant is at least one of Au and Ag.

16. A method according to claim 4 or claim 5, wherein the dopant is nitrogen.

17. A method according to claim 1 or claim 2, wherein the dopant is MoS₂.

18. Apparatus for producing doped highly tetrahedral diamond-like amorphous carbon, the apparatus comprising: a vacuum chamber (2); a source (1) of carbon; means (5,6) for ionising the carbon; a substrate (4); means (3) for directing the carbon ions into the chamber to the substrate; and, a source of dopant (1;7) ; characterised by: means (5,6;7) for ionising the dopant.

19. Apparatus according to claim 18, wherein the carbon source is a solid.

20. Apparatus according to claim 19, wherein the dopant source is incorporated into the solid carbon source (1) .

21. Apparatus according to claim 18, wherein the dopant is incorporated into a separate, secondary source (8) within the vacuum chamber (2) .

22. Apparatus according to claim 18, comprising a gaseous source (7) of the dopant outside of the vacuum chamber (2) .