Method for laser machining inside materials

Abstract

The invention provides a method for laser modification of a sample to form a modified region at a target location within the sample. The method comprises positioning a sample in a laser system for modification by a laser; measuring tilt of a surface of the sample through which the laser focusses; using at least the measured tilt to determine a correction to be applied to an active optical element of the laser system; applying the correction to the active optical element to modify wavefront properties of the laser to counteract an effect of coma on laser focus; and laser modifying the sample at the target location using the laser with the corrected wavefront properties to produce the modified region.

Claims

1. A method for laser modification of a sample comprising diamond or gemstone to form a modified region at a target location within the sample, comprising: positioning a sample comprising diamond or gemstone in a laser system for modification by a laser; measuring tilt of a surface of the sample through which the laser focusses; using at least the measured tilt to determine a correction to be applied to an active optical element of the laser system; applying the correction to the active optical element to modify wavefront properties of the laser to counteract an effect of coma on laser focus; and laser modifying the sample at the target location using the laser with the corrected wavefront properties to produce the modified region.

2. A method as claimed in claim 1, further comprising measuring the position of the sample within the laser system; using also the measured position to determine a fabrication depth and the correction to be applied to the active optical element of the laser system.

3. A method as claimed in claim 1, further comprising applying the correction to the active optical element to modify wavefront properties of the laser to counteract an effect of spherical aberration on laser focus.

4. A method as claimed in claim 1, further comprising measuring the sample after forming the modified region, and modifying the correction applied to the active optical element based on the further measurement.

5. A method as claimed in claim 1, wherein laser focus has a Strehl ratio of greater than 0.5.

6. A method as claimed in claim 1, further comprising measuring a focus of the laser within the sample, and modifying the correction applied to the active optical element based on the measured focus.

7. A method as claimed in claim 1, wherein determining a correction is based at least in part on a pulse energy of the laser.

8. A method as claimed in claim 1, further comprising laser modifying the sample using a single laser pulse; and/or comprising laser modifying the sample using a plurality of laser beams.

9. A method as claimed in claim 1, wherein forming a modified region includes using non-linear optical interactions to cause modification of the material.

10. A method as claimed in claim 1, wherein forming a modified region comprises modifying only material within the bulk of the sample.

11. A method as claimed in claim 1, further comprising using a sample comprising diamond.

12. A method as claimed in claim 1, wherein determining a correction comprises calculating coefficients of a Zernike expansion.

13. A method as claimed in claim 1, wherein the correction counteracts the effects on the focus of tilt aberration caused by the sample.

14. A method as claimed in claim 1, further comprising using a pulsed laser with pulse energies of between 10 nJ and 300 nJ.

15. A method as claimed in claim 1, further comprising modifying a region less than 1 micrometre in the propagation direction.

16. A laser system for laser modification of a sample to form a modified region at a target depth within the sample, comprising: a measurement device arranged to measure tilt of a sample, wherein the sample comprises diamond or gemstone.

17. A laser system as claimed in claim 16, further comprising a processor and an active optical element, wherein the processor is configured to determine a correction to be applied to the active optical element using measured tilt of the laser system, and to communicate the correction to the active optical element.

18. A laser system as claimed in claim 17, wherein the active optical element is configured to modify wavefront properties of the laser to counteract an effect of coma on laser focus; and/or wherein the active optical element is configured to modify wavefront properties of the laser to counteract an effect of spherical aberration on laser focus.

19. A laser system as claimed in claim 16, arranged to perform a method for laser modification of the sample to form a modified region at a target location within the sample, further comprising: positioning a sample in a laser system for modification by a laser; measuring tilt of a surface of the sample through which the laser focusses; using at least the measured tilt to determine a correction to be applied to an active optical element of the laser system; applying the correction to the active optical element to modify wavefront properties of the laser to counteract an effect of coma on laser focus; and laser modifying the sample at the target location using the laser with the corrected wavefront properties to produce the modified region.

20. A method for laser modification of a gem stone to form a modified region at a target location within the gem stone, comprising: positioning a gem stone in a laser system for modification by a laser; measuring a focus of the laser within the gem stone; using at least the measurement of the focus to determine a correction to be applied to an active optical element of the laser system; applying the correction to the active optical element to modify wavefront properties of the laser to counteract an effect of aberration on laser focus; and laser modifying the gem stone at the target location using the laser with the corrected wavefront properties to produce the modified region.

21. A method as claimed in claim 20, wherein applying the correction to the active optical element results in a laser focus with a Strehl ratio of greater than 0.5.

22. A method as claimed in claim 20, wherein the correction counteracts an effect of coma aberration on laser focus; and/or wherein the correction counteracts an effect of spherical aberration on laser focus.

Description

LIST OF FIGURES

(1) Embodiments of the invention are described below by way of example only and with reference to the accompanying drawings in which:

(2) FIGS. 1A, 1B, 1C, 1D show a graphitic track written inside a diamond substrate from different perspectives;

(3) FIG. 2A shows a schematic example of a distorted focus within a substrate without using aberration correction;

(4) FIG. 2B shows a schematic example of a focus within a substrate achieved using aberration correction;

(5) FIGS. 3A and 3B shows features written in a diamond substrate, the left-hand set of features were written using an aberration correction technique, while the right-hand set of features were written without using an aberration correction technique; and

(6) FIG. 4 shows an exemplary laser system for laser modification of a sample.

DETAILED DESCRIPTION OF EMBODIMENTS

(7) In the following there is described a system and method for laser processing at high resolution inside transparent materials incorporating aberration correction to compensate for effects of refraction at the material surface. The system uses feedback from measurements of the material to optimise the aberration correction and focal intensity to obtain the necessary level of material modification. A particular application is for the fabrication of features inside diamond.

(8) Applications of the disclosed method include the generation of light guides inside glasses through local increase of the refractive index. Similar structures may be created in crystals such as KDP or lithium niobate using localised increases in refractive index caused by stress fields around damaged tracks in the medium. Microfluidic devices may be created through exposure of glass followed by chemical etching. Nonlinear photopolymerisation may be used to create complex 3D polymer structures from appropriate solutions.

(9) The disclosed method may be used to create modified regions within a material which are approximately the same size as a diffraction-limited laser focus, which may be less than 1 micrometre in all dimensions.

(10) Fabrication in Diamond

(11) When femtosecond pulsed lasers are focused tightly inside diamond, the non-linear optical interactions cause modification of the crystal lattice in various ways, dependent upon the energy density at the focus. At low energies, there is minor disruption of the crystal lattice that can be used to generate colour centres following annealing. At higher energies, there is significant lattice disruption to the extent that there is significant conversion of the carbon from the sp3 phase (diamond crystal structure) to the sp2 phase (graphite). Typically, the laser modified regions take the form of amorphous carbon, which is a combination of the sp2 and sp3 phases.

(12) Fabrication of fine optical features in diamond uses short-pulsed lasers and high numerical aperture (NA) objective lenses. This ensures that features are well confined in three-dimensions within the material and there is no damage to the surface of the material. A single laser exposure can create a point-like feature of modified material. Complex structures, which may be two or three dimensional, can be built up using collections of point-like features. Alternatively, linear structures can be composed of closely spaced features.

(13) There are two regimes for sub-surface laser processing inside diamond: (i) at very low pulse energy the highly non-linear interaction generates an ensemble of lattice vacancies at the laser focus, while (ii) at higher pulse energies, there is break-down of the diamond lattice leaving a conductive graphitic phase. Modifications of Regime (i) are invisible by transmission microscopy and can only be seen in a fluorescence microscope. They are an important precursor for the formation of coherent NV (nitrogen vacancy) centers for quantum applications. Regime (ii) modifications comprise small (<several 100 nm) clusters of sp2 bonded carbon accompanied by micro-cracks in the diamond lattice. Tracing the diamond through the laser focus in Regime (ii) enables the writing of continuous tracks of sp2 bonded carbon which can be used as electrically conductive wires.

(14) The features formed without translation of the diamond during exposure take the form of an ellipsoid. The extent of the modification is longer along the direction of propagation for the fabrication laser and depends upon the NA (numerical aperture) of the objective lens used. The size of the features is also dependent on the pulse energy used and the dosage (number of pulses in the exposure). If the wavefront is well corrected as described here, highly regular modifications can be made from exposure to exposure. If the wavefront is not well corrected as described here, highly irregular modifications (in size and shape) can result from different exposures with the same conditions.

(15) High NA optics (NA>˜0.8) are used for fabrication both of features which are not axially extended (<2 μm along the optical axis) but also for reliable fabrication from point to point. Fabrication at lower NA (˜0.5) may be possible but is heavily position dependent and inconsistent. At higher NA, the fabrication is highly repeatable over a large volume with exactly the same pulse energy and laser dose. There is no position dependence to the fabrication. This is important to industrial application.

(16) Other demonstrations have consisted of graphitic point-like features in an array. In other applications, continuous graphitic structures have been generated that may be used as electrical conductors.

(17) Tracing the diamond through the laser focus (or scanning the laser relative to a fixed diamond sample) enables the creation of continuous tracks of laser modified material. Working in fabrication Regime (ii), these tracks contain sp2 bonded carbon and are electrically conductive. They may be used to form conductive wires that run in 3 dimensions through the diamond. For example, FIG. 1 shows various continuous graphitic tracks 110 following 2D and 3D paths beneath the surface 120 of a diamond sample. FIG. 1A shows a top view of the graphitic track 110. FIGS. 1B and 1C show side views of the graphitic track 110, and FIG. 1D shows an imaged perspective view of the written graphitic track 110. Scale 125 shows 5 μm. The dashed ellipse in FIG. 1C shows a portion of the graphitic track which is fabricated with increasing depth.

(18) Such graphitic tracks 110 may serve as conductive wires and are useful in the manufacture of diamond based sensors for radiation or chemical sensing. In one implementation, wires running through the diamond may have a voltage applied across them to set up a local electric field within the diamond. If ionising radiation is incident on the diamond, it may create free charge on passing through the diamond, which is collected by the electrodes. In another implementation, the embedded wires may be used for electrochemistry, taking advantage of the wide potential window of diamond. The embedded wires can be used to create an electric field near the surface of the diamond, which is then immersed in a solution. Such embedded laser-written wires can be connected to external electronics by bringing them up to the diamond surface, which is subsequently coated in metal for efficient electrical connection.

(19) Aberration Correction

(20) In the ideal case, the size of the laser focus should be at the diffraction limit i.e. the smallest spot size possible for a given wavelength, refractive index and numerical aperture of the objective lens. However, frequently this limit is not achieved due to the effects of aberrations. Aberrations are deviations of the optical system from its ideal focusing performance. In terms of ray optics, aberrations cause the rays in the focusing cone to no longer meet at the same point. In terms of wave optics, the wavefronts converging on the focus no longer take the form of the spherical cap required for focusing to a diffraction limited spot. In this wave optics case, the aberrations are often quantified in terms of the optical phase error between the ideal and distorted wave front, and different types of distortions are characterised by different phase errors. The effects of aberrations on the focus are to spread out or blur the focus while reducing its peak intensity. By the nature of focusing, the spreading takes place predominantly along the optical axis.

(21) FIG. 2A shows an example of a type of particular aberration mode: spherical aberration. In laser fabrication, aberrations frequently occur due to the refraction of rays at the surface of the transparent material inside which one intends to fabricate. This could occur at the interface between, for example, the immersion medium of the objective lens (typically air, oil or other media) and the fabrication material. The light rays 210 which enter the sample 220 are refracted by the sample surface 222 and the resulting focus 230 is distorted and elongated. The flat wavefront 240 means parallel rays 210 enter the high NA lens 250, which would result in ideal focusing if the sample 220 were absent.

(22) FIG. 2B shows the case where the wavefront 240 has already been corrected by an active optical element such as an SLM, which has modified its phase to counteract the refraction of the sample surface 222. As a consequence, in the absence of a sample 220 the rays 210 leaving the lens would not result in an ideal focus. However, given the presence of the sample, the rays 210 refract at the surface 222 and result in an improved focus 232. Therefore, by shaping the incident wavefront using adaptive optics the aberration is cancelled, allowing accurate focusing and reliable fabrication of sub micrometre features.

(23) FIG. 3A shows fabricated features 310 (left) created by focusing inside diamond at a depth of 50 μm using adaptive optics aberration correction. Features 320 (right) are the same structure but made without aberration correction. The laser pulse energy had to be substantially increased to see any fabrication at all. The fabrication laser was incident along the z-axis. It will be see that the fabrication of features 310 is far better controlled than that of features 320. FIG. 3B shows the same features from the side. It will be appreciated that the features created using the aberration correction technique of the present invention are significantly better controlled in the propagation direction (i.e. the z-direction).

(24) If the interface of the sample is normal to the optical axis, then the aberration consists of a refocusing effect (a focal shift along to optical axis) in addition to spherical aberration. The amplitude of these effects is proportional to the focusing depth. An expression for spherical aberration of this type is:

(25) ${\phi }_{\mathrm{SA}}\left(\rho \right)=\frac{-2\pi d_{\mathrm{nom}}}{\lambda }\left(\sqrt{n_2^2-{\left(\mathrm{NA}\rho \right)}^2}-\sqrt{n_1^2-{\left(\mathrm{NA}\rho \right)}^2}\right)$

(26) This equation is an analytic description of the spherical aberration phase φ.sub.SA for light of wavelength λ when focusing to a depth d.sub.nom inside a material of refractive index n.sub.2, using an objective lens with numerical aperture NA and immersion medium n.sub.1. The coordinate ρ is the normalised radius in the pupil of the objective lens.

(27) If the surface normal is tilted with respect to the optical axis then other aberrations such as coma (which includes wavefront tilt, causing a lateral focal shift) are introduced. These effects are proportional to both the focusing depth and the angle of surface tilt. The additional aberration due to a small surface tilt is given by:

(28) ${\phi }_{\mathrm{tilt}}\left(\rho ,\theta \right)=\frac{t{d}_{\mathrm{nom}}}{\lambda }\left(a\rho +b{\rho }^3\right)\cos \left(\theta -\xi \right)$

(29) This equation is an analytic description of the aberration component due to surface tilt, where the tilt is at a small angle t. θ is the azimuthal coordinate in the pupil and ξ represents the orientation of the tilt. Factors a and b are scalar coefficients whose values depend upon the refractive indices and NA.

(30) The combination of spherical aberration and coma causes reductions in the focal intensity and distortions of the intensity distribution that affect the fabrication efficiency and precision.

(31) The effects of refraction at the diamond surface are strong, due to the high refractive index of diamond (refractive index of 2.4, compared to 1.0 for air and around 1.5 for immersion oil). This means that aberration correction will allow production of fine features at depths which would otherwise not be possible. Aberrations correction is readily implemented using a liquid crystal spatial light modulator (SLM), but may also be implemented using a deformable mirror.

(32) While static correction methods are possible, there are indications from various trials that static corrections are not effective at correcting aberrations between nominally similar samples due to slight changes in composition and positioning of the material, and that instead fine-tuned adaptive aberration correction is needed. Therefore, a one-size-fits-all approach cannot achieve the same degree of control as the present method.

(33) The aberration correction demands are increased if using a dry objective lens to focus deep into diamond. The aberration correction requirements are less severe using an oil immersion lens to focus the laser, as the refractive index contrast is lower than with a dry lens. However, correction is still required in this case to obtain viable results.

(34) A phase pattern calculated from the above equations can be imparted upon the aberration correction device (i.e. the active optical element) such as a spatial light modulator (SLM) in order to correct for the aberration induced by the sample. As SLMs typically have a phase modulation range limited to one wavelength (or a small number of wavelengths) the phase is usually wrapped so that it lies within the accessible range. For example, if only a single wavelength is accessible, then the phase function applied will be φ.sub.SA modulo 2π, as 2π radians of phase corresponds to one wavelength.

(35) The phase pattern applied to the SLM can be simplified by noting that the spherical aberration term contains defocus, which is another aberration mode or component that shifts the focus (i.e. the peak optical intensity) along the optical axis, but does not change its shape. By removing the defocus component from the correction, the size of the phase correction can be reduced, thus more effectively using the SLM for aberration correction. The defocusing component of the spherical aberration may be compensated by translation of the sample by an appropriate amount. Similarly, the coma aberration for a tilted sample includes a “tilt” aberration, a constant phase gradient that causes lateral shift of the focus. Again, this tilt can be removed from the phase pattern before it is imparted on the SLM. The lateral shift can be compensated by translation of the sample.

(36) Instead of using the analytic expressions directly, the aberrations can be considered as a series of basis functions. Commonly, the Zernike polynomials are used for this purpose. Hence, an aberration may be described a sum of aberration modes. For example, spherical aberration may be expressed as an expansion in terms of Zernike polynomials. Using functions such as these aids in the design of feedback systems for the measurement and correction of unknown aberrations.

(37) The methods described herein are further relevant for focusing through non-planar surfaces. This might include through curved surfaces or near/across edges. Again, an accurate pre-measurement of the surface topography can be used to predict a starting phase pattern that is close to optimum and can be used as a good starting point for subsequent optimisation using focal feedback. Fabricating across or near an edge involves pupil segmentation. The methods described here can be adopted for setting the phase for correcting spherical and coma aberrations in the segmented pupil. Focusing through a curved surface will require correction of a combination of spherical aberration, astigmatism and coma.

(38) Adaptive Control of the Focus

(39) In order to maintain consistent quality of fabrication between different positions (particularly depths) within a sample and between different samples, it is necessary to implement an adaptive control system that can maintain an appropriate combination of aberration correction and pulse energy. This may use a method of feedback from the focal region to the devices controlling the wavefront and energy.

(40) A first measurement of surface position and tilt based upon reflection from the material surface as described above provides a prediction of the correction needed to counteract spherical aberration and coma. In one implementation, a measurement of the position for best optical focus at three points which are not all on the same axis can provide information on the relative sample tilt. This can allow predictive aberration correction for coma and spherical aberration at a particular depth in a sample, e.g. diamond.

(41) Finer compensation may be carried out by observation through a microscope of the fabrication process at the focus. A combination of measurements is possible. Using a transmission microscope it is possible to observe changes in absorption at the focus or a change in optical phase due to refractive index modification. This indicates the degree to which the material is modified by the laser pulse and can provide a feedback signal for optimisation of the aberration correction and the pulse energy. Alternative feedback signals could be provided by photoluminescence or plasma emission from the focus.

(42) In order to reduce the number of measurements required (and hence the time taken) for the optimisation process, algorithms can be used in which coefficients of the aberration modes (particularly spherical aberration and coma) and the pulse energy are considered as unknown coordinates in a search space.

(43) The optimisation process can be expressed mathematically as the minimisation of a cost function f (or alternatively, g may be a merit function that should be maximised) that is related to the aberration components and the pulse energy, collectively represented by the symbol P. The optimal value of P is given by either

(44) $P_{\mathrm{opt}}=\arg \underset{P}{\min }\left[f\left(P\right)\right]$

(45) where the cost function should be minimised or

(46) $P_{\mathrm{opt}}=\arg \underset{P}{\min }\left[g\left(P\right)\right]$

(47) where the merit function should be maximised. The function for g can be defined as a combination of measurements. For example, the intensity of the focal plasma generated during laser fabrication is dependent upon the total aberration content, so that a corrected system shows a maximum in focal plasma emission. Alternatively, the fabrication laser may be used below the threshold for structural modification to non-linearly excite photoluminescence (PL) from intrinsic defects contained in the sample. The detected PL is maximised when the aberrations are minimised. Similarly, luminescence or fluorescence emission in a confocal microscope is maximised when the aberrations are corrected. These signals can therefore be used as a feedback mechanism to allow optimisation of the cost/merit function.

(48) Various methods are possible for the implementation of optimisation processes. The minimum number of unknown parameters that need to be optimised for laser fabrication through a tilted flat surface, where the refractive indices and NA are known, is three: one coefficient of spherical aberration, and two for coma (i.e. the two orthogonal coma components). The process may therefore be considered as a three-dimensional optimisation problem. Another variable in the form of pulse energy can be considered, which then extends the process to a four-dimensional optimisation. If there are further unknowns, then more variables (dimensions) would have to be considered in the optimisation process.

(49) Adaptive optimisation could be performed on a point-by-point basis for every fabricated position, though it is likely to be more practical to perform fewer optimisation measurements across the field of fabrication and perform interpolation of parameters across this field. This field could exist in a line or curve, a lateral plane, or extend through three dimensions. With sufficient surface measurements, a suitably corrected focus (i.e. with a sufficiently large Strehl ratio) can be achieved. However, an optimisation procedure may be performed for each new sample.

(50) Description of Larger Scale Markings

(51) Larger scale structures and markings are also possible. This might include point like features or continuous lines to create alphanumeric characters, barcodes, QR codes or images. Such features could be grouped together to form diffractive elements, holograms, or diffraction gratings. Depth ranges and areas can be up to the size of the stone being used (typically in the range of 3 mm in the transverse x- and y-directions, and 1 mm in the propagation z-direction). The size of the features may be up to 5 μm across (in x- and y-directions) by 20 μm in the propagation z-direction. In practice, if larger features are to be created reliably, it may be achieved by stitching together combinations of smaller scale modifications. Care must be taken when generating large features to manage the stress load on the surrounding sample (e.g. diamond) to avoid large scale irregular cracks forming. This might be achieved by a sparse array of small (˜1 μm scale) features that are linked together to form a feature that looks large when viewed optically but only has minimal volume conversion of diamond to graphite.

(52) Schematic Diagram of Typical Implementation

(53) FIG. 4 shows an exemplary configuration for the adaptive fabrication system. Additional components might be added in order to, for example, aid with the aberration or position sensing, to perform additional aberration correction, or to parallelise the system and use multiple focal spots.

(54) The laser system 400 includes a laser 410, a polarizer 420, a spatial light modulator (SLM) 430, a high NA objective lens 440, and a three-dimensional translation stage 450. A sample 460 is positioned on the stage 450 at the focus of the system 400.

(55) The sample 460 is diamond and is positioned in the laser system 400 for modification by the laser 410. The sample 460 is then measured to inform the determination of a correction. Particularly the surface of the sample 460 on which the laser will be incident is measured and its inclination from transverse is determined. The transverse direction is the 2D plane perpendicular to the primary propagation direction of the laser. This is also the plane parallel to the major plane of the objective lens.

(56) The inclination of the surface of the sample 460 is used to determine the expected coma aberration which will be caused thereby on the laser focus. The expected aberration is then characterised in terms of a Zernike mode, and is communicated to the SLM. The SLM is modified to display the required phase correction to modify the laser to counteract the aberration.

(57) The pulse energy of the laser system is also determined based upon the correction. The laser is set to the required pulse energy, and is then used to modify the diamond sample.

(58) Following modification, the modified region of the sample is measured using transmission microscopy. The determined correction is then refined based upon the feedback obtained by this further measurement. The refined correction is applied to the SLM and the sample is laser modified.

EXAMPLE

(59) The diamond is mounted in the laser fabrication system. The objective lens is moved axially (i.e. in the z-direction) to initially find the diamond surface by maximising the reflected light from an LED. The diamond is moved in the transverse (x-y) directions to the desired location for fabrication.

(60) The fine positioning for the surface axial (z) location is achieved using the laser with low pulse energy (significantly below the bulk graphitisation threshold, e.g. less than 30 ml). The sample surface is found by translating the diamond axially in 100 nm steps. If using an oil immersion lens, the diamond is moved to the point at which the laser no longer boils the immersion oil. If using an air lens, the diamond is moved until the laser no longer causes any mark on the diamond surface. A further two such measurements are made, one by translating 0.2 mm in the x-direction, the other by translating 0.2 mm in the y-direction. These measurements cover an area of the sample and are used to determine the local surface tilt.

(61) The diamond is then translated axially to the desired depth for fabrication, noting that the actual fabrication depth is greater than the translation depth by a factor of approximately 2 for high NA oil lenses and about 2.7 for a high NA air lens. This is because the SLM is used to correct all aberrations caused by refraction at the sample interface, except defocus, which is more simply counteracted by axial translation of the diamond, as described above.

(62) The aberration correction is applied to the SLM based upon the surface measurements. The correction represents a spherical aberration correction based upon the translated axial depth, and coma aberration in the x and y directions based upon the measured x and y surface tilt. A predetermined pulse energy (e.g. 100 nJ for a 1.4 NA oil lens using 780 nm wavelength light with pulse duration 250 fs) is used and a burst of 5 pulses is fired into the diamond. A transmission microscope is used to verify that there has been successful modification of the diamond at the desired point. The preferred fabrication should have dimensions of approximately 0.5 μm (in the transverse direction) by approximately 1 μm (in the axial direction) and will appear dark when viewed in a transmission microscope.

(63) Aberrations also need to be compensated in the imaging in order to see the fabrication. Then it is verified that the diamond can still be modified with lower pulse energy and/or dose to the point that the modification becomes invisible. The desired pattern is then fabricated within the diamond as desired. When there is a tilt to the sample, the transverse movements of the diamond are accompanied with axial movements to ensure the fabricated points remain a constant depth beneath the diamond surface. If a 3D fabricated design is required the phase pattern displayed on the SLM is automatically updated during fabrication based upon feedback from the translation stages.

(64) If no modification is visible when the diamond is irradiated with the first burst of pulses, the sample is translated a small distance in the transverse direction (e.g. 5 μm) and the aberration modes applied to the SLM are adjusted in a systematic manner with a burst of the laser applied for each setting. The diamond is axially translated between each burst and checked to see whether the diamond is modified as desired. Once the correct phase is displayed upon the SLM the fabrication is carried out straightforwardly as described above without need for further adaptive correction.

(65) Other metrics might be used to optimise the SLM phase other than visible diamond modification, for example non-linear photoluminescence from the diamond caused by the laser focus might be optimised to correct the aberrations. For this measurement the laser pulse energy is dropped to ensure there is no fabrication (pulse energy below e.g. 20 nJ using conditions described above) or ideally by switching to a laser with a higher repetition rate and low pulse energy (80 MHz rep rate and pulse energy <20 nJ). The characteristics of the focus may be measured and the aberrations affecting it determined. Then a correction can be determined and applied to the active optical element to improve the Strehl ratio of the focus. The need for such a procedure of adaptive correction may be rare given accurate measurements of the surface.

(66) The above method may use a femtosecond infra-red fabrication laser to modify the fibre, but the techniques may also be applied to fabrication systems of any wavelength or any pulse width. For example ultraviolet (UV) and continuous-wave (CW) systems can be used. Typically the fabrication laser induces an increase in refractive index of the sample. However in some materials the laser can induce a decrease in refractive index. The optical devices manufactured may operate at a different wavelength to the writing laser. Devices may be manufactured for any operating wavelength of the optical device.