LASER APPARATUS
The present invention relates to laser apparatus particularly but not exclusively including means for improving the spatial quality of a laser output beam at relatively high power levels.
Amongst laser apparatus currently available, it is possible to provide substantially single mode operation in relatively inexpensive devices by limiting the lasing cross-sectional area to an area of approximately 1 x 3 microns. Typically such a device is formed by wafer techniques to give a semi-conductor device having a thickness in the region of 1 micron and cross-sectional width in the region of 3 microns. Such a single mode laser will typically provide an output power of up to 200 to 300 milliwatts. If it is desired to provide a diode laser with a higher output, constructions known as broad area lasers, or diode laser arrays are provided, having approximately the same cross-sectional depth (i.e. 1 micron), but having cross-sectional widths of the order of 100 microns. Such devices can provide an output power in the region of 0.5 to 7 watts, but give multimode operation with a multilobe far-field intensity pattern in the plane of the long dimension of the device (referred to as the slow axis). Perpendicular thereto (the fast axis) the light is still essentially in a single spatial mode due to the small emitter dimension in this direction. The multilobe far-field along the slow axis is a consequence of the broader width of the diode laser medium along that axis (i.e. the stripe width), which supports oscillation of higher order transverse modes. When the intensity profile of such a broad area laser or diode laser array is measured in the far-field plane, the profile often appears to have two far-field lobes, each consisting of a combination of two-lobed cavity modes. Only in the case of single mode lasers, is a single beam extracted, namely the fundamental mode. The result of this is that the output beam is no longer substantially diffraction limited, but instead the spatial beam quality is in the order of 10 to 100 times the diffraction limit along the slow axis.
Thus the problem can be summarised as follows. To increase the optical power from a semiconductor device one can either increase the cross-section of the gain region (sometimes referred to as broad-area laser or laser diode array), or place a number of lasers next to each other (sometimes referred to as a laser diode bar). One can also do both. Unfortunately, both of these two approaches result in high-power devices that operate in several spatial and longitudinal modes simultaneously, or result in the radiation being partly incoherent due to absence of coupling between the individual laser elements. This leads to an output radiation that is many times the diffraction limit and has a very short coherence length of the order of a few hundred micrometers. This fact limits the usefulness of the high-power devices since it especially destroys the focusability of the output beam, which is important in many applications.
A number of techniques are known to reduce the transverse modes in the output beam of a diode laser. A commonly used technique is the provision of feedback into the laser by reflection from a feedback device forming an external cavity with the laser. Examples of such arrangements are disclosed for example in WO98/56087 (Torsana A/S).
The term feedback refers to the process where a fraction of the output energy returns to the active region of the laser structure, for example by means of reflection, diffraction, or scattering. The optical feedback then influences the thermal temperature distribution, the optical field and carrier distribution in the lasing cavity causing it to change behaviour. In the case of laser diodes there may be provided two different external cavity configurations, referred to herein as an "on-axis" external cavity, and an "off-axis" external cavity respectively. In the on-axis configuration the whole beam is collected and directed towards the external reflector that returns a fraction of the incident energy. In the off-axis configuration, however, only a part of the two-lobe far-field is directed towards the external reflector, which then returns all or a fraction of the energy contained in that lobe.
In the disclosure of WO98/56087, a laser assembly such as a diode laser produces an output light beam, which in the free running state without feedback has multiple lobes in its far-field intensity pattern. A light feedback device such as a plain dielectric mirror, grating, or phase conjugate mirror, forms an external cavity with the laser assembly for returning to the laser assembly light derived from a selected first lobe of the far-field intensity pattern. An optical arrangement is provided to produce an output beam, which is derived from a second lobe of the far-field intensity pattern. Most commonly the output of the laser assembly in the free running mode has two principal lobes positioned symmetrically on either side of a central axis of the laser assembly. In the prior art cited, a spatial filter is provided for restricting the transverse lasing modes of the feedback light to one or more selected transverse modes, preferably such that the laser is brought to lase in substantially a single transverse mode. The proportion of light returned to the laser assembly from the first lobe (i.e. the feed back lobe) is such that the dominant lobe is the second lobe (i.e. the output lobe) from which the output light beam is derived.
Further increases in power output can be obtained by devices of greater cross-sectional lasing area, commonly known as bar lasers or laser diode bars. Here a single monolithic structure is produced by conventional wafer technology, in which a series of spatially separated, and optically independent diode lasers are positioned side by side along the laser bar, with the long dimension of each laser cross-section aligned lengthways along the bar. The result is a laser assembly, which produces a much higher output power, of the order 10 to 100 watts, but the output beam has even poorer beam qualities due to the large cross-sectional area («1 micron by 1 cm) of the output facet. The reason is that the individual diode lasers forming the bar, are lasing independently of each other with no predetermined phase relationship between them. Consequently such devices have limited use because although the power is high the brightness of the output beam is low. A laser diode bar has characteristics similar to the characteristics of a laser diode, or broad area laser, since a bar consists of separated laser diode devices, but the beam from commonly used bars is in the
order 1000-2000 times the diffraction limit. The diffraction limit is related to the number of times poorer an arrangement is, in terms of focus diameter, compared with the theoretical best achievable result, i.e. where the beam focuses a gaussian beam with similar transverse dimension and wavelength. The term M2 is used for this quantity.
In another known disclosure, "Spectral beam combining of a broad-stripe diode laser array in an external cavity", V. Daneu, et al, Optics Letters, March 15, 2000, volume 25, no. 6, page 405-407, (also disclosed in US-A-6, 192,062 and US-A-6, 208, 679), an arrangement is proposed for combining primary beams from a bar laser by use of a grating positioned between the bar laser and an output coupler consisting of a partially reflecting planar mirror having a 10% reflection coefficient. Thus a resonant external cavity is formed between the output coupler and the bar laser and the reflective grating is positioned to reflect light between the two. A transform lens, situated one focal length away from the diode laser bar focuses the primary beams from the bar laser onto the surface of the grating, which reflects a commonly aligned beam to strike the output coupler orthogonally. The result is that the primary beams from the bar laser are co- aligned into a commonly aligned output beam, with a resulting increase in brightness. However this is achieved merely by superimposing the individual primary beams from the bar laser which result in high brightness at the expense of a corresponding increase in the spectral bandwidth of the combined output beam. Ideally the beam from the diode bar, which is approximately 1000 times the diffraction limit, would be transformed into a beam with focusing properties similar to a single component beam, i.e. 20-50 times the diffraction limit. In the prior art, no attempt is made to improve further on the focusing properties.
The focusing property is often the most important parameter for laser applications. For the telecom industry where beams are coupled into fibres, the output beam needs to be substantially diffraction limited (i.e. M2= 1). Otherwise great fibre coupling losses will occur. An example of telecommunications use is as a pump laser for Raman fibre amplifiers. Industrial applications, such as
material processing in general, rely heavily on the focusing ability combined with high output power. Examples are soldering, cutting, welding, and exposure of films. Here a great demand exists for producing multiwatt output power with near diffraction-limited performance. In these cases the spectral width is of minor importance.
Despite the techniques available for improving the spatial quality of the output beam from laser diodes, it has not been possible hitherto to produce similar high quality, near diffraction limited, beam qualities in light from larger area laser assemblies, such as laser bars.
It is an object of the present invention, at least in preferred embodiments, to provide laser apparatus in which the improvements derived in off-axis technology with broad area lasers and laser diode arrays, can also be achieved with a laser assembly such as a bar laser where a series of individual laser devices provide spaced apart component beams of unrelated phase.
According to the present invention there is provided laser apparatus comprising a laser assembly comprising a plurality of spaced apart laser devices for producing a plurality of respective primary light beams, each primary light beam having multiple lobes in its far field intensity distribution pattern, a light feedback device forming a resonant external cavity with the laser assembly for reflecting to the laser assembly a feedback portion of the light produced by the laser assembly, the light feedback device having a preferred general direction of reflection of light which gives maximum feedback; a dispersive device which for a given angle of incidence transmits or reflects light at different angles for different frequencies, the dispersive device being positioned to receive light from the laser assembly and to transmit or reflect light to the feedback device, and an imaging device for imaging the said primary beams at the dispersive device to form a combined intensity distribution pattern having multiple combined lobes corresponding to the multiple lobes of the individual primary beams, each primary light beam contributing to each combined far-field lobe of the combined
intensity distribution pattern by light from the corresponding far-field lobe of that primary light beam, the combined lobes being positioned off-axis relative to a principal axis of the laser assembly which is perpendicular to an output face thereof; the arrangement being such that each laser device lases at a different frequency such that its light is directed by the dispersive device along the said preferred direction of the light feedback device to obtain maximum feedback, and that the primary beams combine to form at least one combined feedback beam reflected back to the laser assembly and at least one combined output beam leaving the apparatus, each combined beam being formed of component beams derived from the respective laser devices and having a common axial direction; in which the light feedback device provides a different reflectance to light from one off-axis lobe of the said combined intensity distribution pattern than to light from another off-axis lobe of the said pattern.
By the term reflectance (also known as the reflection coefficient) is meant the ratio of the reflected power (or flux) to the incident power (or flux).
It is preferred that the light feedback device includes or consists of a non- adaptive reflecting surface, for example a plane mirror. However other forms of light feedback device may be used, including a curved mirror, a phase conjugate mirror, or a reflective grating. The preferred general direction of reflection of light of the light feedback device is preferably a substantially single direction of reflection of light for which an incident ray is reflected back along substantially the same light path as that of the incident ray, but in some embodiments the preferred general direction of reflection of light may be a small range of directions over which an incident ray may be reflected. For a plane mirror, the preferred direction of reflection for maximum feedback is a direction normal to the plane mirror. For other forms of feedback device, such as a curved mirror, or a phase conjugate mirror, there may be a small range of angles of reflection giving maximum feedback, constituting a generally preferred direction. For a phase conjugate mirror, a spatial filter may be used, for example a pair of spaced apart masking elements having opposed parallel straight edges, for restricting the
range of angles of reflection utilised with the phase conjugate mirror. However it is preferred that the feedback device should be such as to have as closely as possible a single preferred direction of reflection of light, so as to have a single path of the feedback beam which gives maximum feedback.
In one preferred arrangement, the light feedback device consists of a light feedback element positioned in the path of one, feedback, lobe of the combined intensity distribution pattern, the said output beam being derived from another, output, lobe of the combined intensity distribution pattern which is displaced on the opposite side of the axis from the feedback lobe. In some preferred arrangements the light feedback element provides a reflectance of substantially 100% to the feedback lobe of the combined intensity distribution pattern, and the dispersive element is arranged to direct substantially 100% of the output lobe of the combined intensity distribution pattern to form the said output beam.
In other arrangements, the light feedback device consists of a first light feedback element positioned in the path of one lobe of the combined intensity distribution pattern and another light feedback element positioned in the path of a second lobe of the combined intensity distribution pattern which is displaced on the opposite side of the axis from the said one lobe, the first light feedback element having a lower reflectance than the second feedback element, the said output beam being derived from the said one lobe of the combined intensity distribution pattern. By way of example it may be arranged that the reflectance of the first light feedback element is in the range 10% to 20% and the reflectance of the second light feedback element is in the range 50% to 100%.
It is preferred that the output beam and the feedback beam are derived from different lobes. Preferably this is done by arranging the feedback device off axis on one side of the principal longitudinal axis of the laser assembly, and arranging for extraction of the output beam from a symmetrically positioned beam on the other side of the principal axis.
It is preferred that there is included in the apparatus a spatial filter for restricting the spatial modes of the feedback light to one or more selected spatial modes from each laser device. Preferably the spatial filter is positioned in the path of light from the said one lobe of the combined intensity distribution pattern, from which the output beam is derived.
The invention has particular utility where the light produced by the laser has two predominant lobes in its far-field intensity profile, positioned symmetrically one on each side of a principal longitudinal axis perpendicular to the output facet of the laser assembly. This dual lobe profile is normally found in the output of a diode laser bar. Here each laser device contributes with one of the far-field lobes thereof to each of the two combined far-field lobes of the laser assembly.
Preferably the imaging device produces converging component beams imaged at the dispersive element. Preferably such a device includes a lens positioned one focal length thereof away from the input face (also referred to as the entrance surface) of the dispersive element. Moreover it is preferred also to have collimation means for collection and collimation of the light beams along the fast axis. This may be a cylindrical micro lens or lenslet array with high numerical aperture (NA.>0.5).
Preferably the imaging device is arranged to image the primary beams at the dispersive device in such a manner that, in respect of at least one lobe of the combined intensity distribution pattern, the primary beams overlap each other at least partially at the dispersive device. It is particularly preferred that in respect of the said at least one combined lobe, the total area of the overlapping primary beams at ATP the dispersive device is given by the expression
where N is the number of laser devices in the laser assembly and A
P is the largest area of an individual primary beam at the dispersive device for the said at least one combined lobe.
Preferably, in said at least one feedback beam, the component beams overlap each other at least partially. It is particularly preferred that in the said at least one feedback beam, the total area of the overlapping component beams (ATF) is given by the expression
where N is the number of laser devices in the laser assembly and A F is the largest area of an individual component beam in the said at least one feedback beam.
Also preferably in said at least one feedback beam the component beams thereof are co-aligned within an angle OCTF given by the expression
where N is the number of emitters in the light source and C LF is the largest angle of divergence of an individual component beam in the said at least one feedback beam.
Preferably in said at least one feedback beam, the component beams are aligned within substantially the same cross-sectional area as the cross-sectional area of the largest component beam alone.
It is also preferred that in said at least one output beam, the component beams overlap each other at least partially. Preferably the said at least one output beam, the total area of the overlapping component beams (ATO) is given by the expression
where N is the number of laser devices in the laser assembly and A|_o is the largest area of an individual component beam in the said at least one output beam.
It is also preferred that in said at least one output beam the component beams thereof are co-aligned within an angle O JO given by the expression
where N is the number of emitters in the light source and Ct|_o is the largest angle of divergence of an individual component beam in the said at least one output beam. Preferably, in said at least one output beam the component beams are aligned within substantially the same cross-sectional area as the cross- sectional area of the largest component beam alone.
In summary, it is preferred that each said combined beam is composed of overlapping component beams, each laser device contributing with one far-field lobe. Preferably, the combined component beams are aligned within
substantially the same cross-sectional area as the cross-sectional area of one component beam alone. Also it is preferred that the combined beam has a divergence angle of no more than the largest divergence angle of any one of the component beams.
In accordance with a particularly preferred feature, the dispersive device comprises at least two dispersive elements each having an inclined input face which is inclined relative to the axis of the laser assembly, the input faces being displaced from each other transversely relative to the said laser axis so that different off-axis lobes are incident on different dispersive elements, and the input faces being displaced longitudinally relative to the axis in such a manner as to allow each off-axis lobe to be imaged on an input face equidistant from the laser assembly.
The invention has particular application where the laser assembly is a bar laser comprising a plurality of laser devices each having an elongate cross- section and the devices being positioned side by side aligned along the direction of the elongate cross-section. Each laser device of such a bar laser may comprise a broad area laser diode, or a diode array having a plurality of current electrodes along the longitudinal cross-sectional dimension of the diode array.
Preferably the dispersive device provides at least one grating. Conveniently the dispersive element comprises a reflective grating, but the dispersive element may also comprise, by way of example, a reflective hologram, a transmitting grating, a transmitting hologram, or a dispersive prism. The dispersive element may comprise a photonic band gap device, also known as a super prism.
Although many different examples of a dispersive device may be used in embodiments of the invention, it is particularly preferred that the dispersive device has an entrance surface on which the component light beams impinge and the imaged device is arranged to image the component beams at the
dispersive device in such a manner that the component beams overlap on the entrance surface of the dispersive device. An example would be the use of a reflective grating with all the component beams imaged onto the front surface of the grating, so as to be diffracted from the same "point". In practical terms this means that the twenty or so diode laser array beams should preferably be overlapping on the surface of the grating. If this is not the case the feedback beam will not be a well co-aligned beam, but will consist of displaced component beams since the beams will not be diffracted from the same spatial area of the grating.
When considering the operation of multimode lasers such as a bar laser, it is well known in the art that a light beam generated has different intensity distribution patterns in two main regions of the light beam, referred to as the near-field and far-field. It is well known for example that the multiple lasing modes can be reduced to a single mode or near single mode by spatial filtering performed in the so called far-field plane. Maxwell's equations, which describe the behaviour of all electrical fields (including laser fields), can be cast into an equivalent integral equation form (a diffraction integral). In this formulation the field at a spatial point (x,y,z) can be determined when the initial field distribution at the surface of the light source is known. The complicated integral expression consists of three terms which decay differently as a function of the distance between the light source (e.g. the diode laser facet) and the point (x,y,z) considered (e.g. at a spatial filter). The far-field is reached when the two first terms of the integral expression vanish. This turns out to be the case after just a few aperture widths of the source, i.e. a few times (3-4 times) the transverse dimension of the light source. In this situation the far-field term dominates by an order of magnitude over the two near-field terms. For a diode laser with a 1 micron by 100 microns wide emitting aperture the far-field is reached a few 100 microns away (e.g. 400 microns) for the slow axis and a few microns away for the fast axis. The conclusion is that operation is nearly always in the far-field. The spatial region where there is a true near-field image, is small, for example (400 microns).
Although in some arrangements embodiments of the present invention may operate in the near-field, it is much preferred that operation takes place in the far-field. Thus it is particularly preferred that the imaging device is arranged to image at the dispersive device a combined far-field intensity distribution pattern of the component light beams.
It is to be appreciated that where features of the invention are set out herein with regard to apparatus according to the invention, such features may also be provided with regard to a method according to the invention, and vice versa. In particular, there may be provided in accordance with another aspect of the invention a method of generating laser light comprising producing a plurality of primary light beams from a laser assembly comprising a plurality of respective spaced apart laser devices, each primary light beam having multiple lobes in its far field intensity distribution pattern; reflecting to the laser assembly a feedback portion of the light produced by the laser assembly by means of a light feedback device forming a resonant external cavity with the laser assembly, the light feedback device having a preferred general direction of reflection of light which gives maximum feedback; transmitting or reflecting light from the laser assembly to the feedback device by means of a dispersive device which for a given angle of incidence transmits or reflects light at different angles for different frequencies; imaging at the dispersive device a combined intensity distribution pattern having multiple combined lobes corresponding to the multiple lobes of the individual primary beams, each primary light beam contributing to each combined far-field lobe of the combined intensity distribution pattern by light from the corresponding far-field lobe of that primary light beam, the combined lobes being positioned off- axis relative to a principal axis of the laser assembly which is perpendicular to an output face thereof; the arrangement being such that each laser device lases at a different frequency such that its light is directed by the dispersive device along the said preferred direction of the light feedback device to obtain maximum feedback, and that the primary beams combine to form at least one combined feedback beam reflected back to the laser assembly and at least one combined
output beam having the apparatus, each combined beam being formed of component beams derived from the respective laser devices and having a common axial direction; in which the method includes reflecting back to the laser assembly as feedback light, a different proportion of the light from one off-axis lobe of the said combined intensity distribution pattern, than from another off-axis lobe of the combined intensity distribution pattern.
It is to be appreciated that in some arrangements the proportion of light reflected as feedback light in a lobe can be as small as zero or as large as 100% of the light available.
There will now be described a number of preferred or optional features of the invention.
The light feedback device may consist of say a simple or plane mirror, or a phase conjugate medium, a reflective grating or a combination thereof. The spatial filter may comprise two opposing edges forming a slit. In one preferred form, the light feedback element comprises a planar mirror and the feedback light travelling between the dispersive element and the light feedback element passes along a path perpendicular to the incident surface of planar mirror.
In a one preferred embodiment an adaptive mirror is used. This may be a phase conjugate mirror formed of BaTiO3. Due to the ability of a phase conjugate mirror to retroreflect a multitude of light beams independently of the angle of incidence, aberrations and distortions coming from either production tolerances of the bar laser or misalignment of components, can be compensated. One example is a distortion known as "smile". When a bar is manufactured the typically 1 cm long stripe, with thickness 1 micron, will not be a mathematically straight line. Due to for example the packaging, the diodes (typically 20-40 in number) will be aligned along a curved line (like a smile). A "smile" effect of the order of 1-10 μm (maximum deviation from a straight line) will cause the different component beams to be imaged differently through the lens optics. This results
in a less than perfect, overlapping, combined beam, which will finally cause the output beam to be less than diffraction limited. In the case of a phase conjugate mirror the directional property is provided by an aperture, where the aperture width determines the directional reflection property of the adaptive mirror. The aperture could also in some cases be the same aperture which restricts the number of transverse modes of the laser devices of the assembly.
Previously it has not been possible to apply off axis technology to a bar laser because of the difficulty of processing the two lobes from each of the component beams from a bar laser. The present invention allows the use of a dispersive element such as a grating or prism to combine the primary beams from a bar laser into a commonly aligned beam, which is directed to the external cavity feedback element. The feedback loop forces each laser diode to tune itself to a different frequency individual to itself to obtain maximum feedback from the feed back element. The feedback beam is commonly aligned, which forces the output beam also to be commonly aligned. The technique is usable even without a spatial filter, i.e. by returning the whole feedback lobe. But in that case the beam quality (i.e. the M2 value) is restricted (e.g. M2 = 10-100) compared to the preferred embodiments.
A number of advantages arise from the technique of the present invention. In the known arrangement set out in the Daneu paper the beam quality can be improved from M2 = 2000 to M2 = 20, whereas in the present invention the beam quality (measured as M2) can be improved to be substantially equal to 1 , i.e. substantially diffraction limited. Note that for an ideal gaussian beam M2 equals 1. In the Danue paper the resulting beam is emitted in two lobes, each containing approximately 50% of the power, whereas in the present embodiment a single output beam can be emitted containing the major part of the output power. Thus, in preferred embodiments, it is possible to produce a near diffraction limited (e.g. M2=1 to 4) combined beam from a bar laser.
Moreover the usual two lobed beam profile, as seen in broad area lasers or
diode laser arrays, is changed into essentially a one lobe beam containing the major part of the output power (e.g. 75%).
Embodiments of the invention will now be described by way of example with reference to the accompanying drawings in which:-
Figure 1 illustrates a known arrangement for providing external feedback to a dual-lobe laser assembly by feedback from one off-axis lobe and output from the other off-axis lobe;
Figure 2 illustrates a known arrangement for co-aligning the primary beams from a bar laser to a single output beam whilst retaining incoherence between component beams of the bar laser;
Figure 3 is a diagrammatic representation of an embodiment of the present invention in which the component beams from a bar laser are combined and co-aligned in an output beam derived from one off-axis lobe of the bar and feedback is provided from another off-axis lobe by a light feedback device, the feedback light being restricted by a spatial filter;
Figures 3a, 3b and 3c show diagrammatically areas of various combinations of primary beams falling on the surface of a grating in the embodiment of Figure 3;
Figures 3d and 3e show diagrammatically modifications of the apparatus of Figure 3, to enable monitoring of the feedback beam, for adjustment of the apparatus;
Figure 4 is a diagrammatic representation of a modification of the embodiment of Figure 3, where two gratings are used, aligned so as to be both in the plane of focus of the two lobes of the laser assembly;
Figure 5 is a diagrammatic representation of a modification of the embodiment of Figure 3, in which no spatial filter is present, the whole of the light in the feedback lobe being returned to the laser assembly;
Figure 5a is a diagrammatic representation of a modification of the embodiment of Figure 5, in which the output beam is taken from a different off- axis lobe, compared with Figure 5;
Figure 6 is a diagrammatic representation of a modification of the embodiment of Figure 3 in which the feedback device is a phase conjugate mirror with a lens for mode matching of the feedback lobe into the phase conjugating medium;
Figure 7 is a diagrammatic representation of a modification of the embodiment of Figure 3, in which feedback is provided in both an output section as well as in a feedback section, and in which a spatial filter is situated in the output section of the apparatus; and
Figure 8 is a diagrammatic representation of the operation of a thick hologram as a dispersive device 36, in an embodiment of the invention such as shown in Figure 3 onwards.
Figure 1 shows a known arrangement, as described for example in WO98/15994, in which a GaAIAs laser diode array (or broad area laser) 11 generates light with a dual lobe far-field intensity distribution pattern. Light from one off-axis lobe is fed to a feedback element 13 which may for example be a plain mirror, a reflective grating, or may be a phase conjugate mirror, and the light is returned to the diode array 11 to provide feedback. The feedback beam 12 passes through a spatial filter 14 and a frequency selective element 15 to isolate a substantially single transverse mode and single longitudinal mode respectively. An output beam 16 derived from the second lobe of the diode array 11 is reflected from a mirror 17 to produce an output from the device. Beam
collimation and focussing is produced by a series of lenses 6, 7, 8, 9 and 10. A beam splitter 5 is provided between the lenses 6 and 7, for monitoring purposes.
In the free running mode, without feedback, the output of the laser diode array 11 is a multimode light beam consisting of a superimposition of transverse array modes. Each array mode has a double lobe intensity profile in the far-field. The array modes are distinguished by different frequencies and different radiation angles. The off-axis feedback illustrated selects and re-injects one of the lobes of a single array mode, forcing the laser diode array 11 to lase in this mode only. The resulting single mode radiation is strongly asymmetric with a dominant lobe, which forms the output beam 16, and a smaller lobe, which forms the feedback beam 12.
The off-axis configuration shown in Figure 1 has been applied successfully to broad area lasers as well as diode laser arrays. Broad area lasers support sinusoidal like near-field modes, which lead to the conventional far-field pattern with the two-lobed shape. In a laser diode array there are provided a number of individual laser elements placed next to each other (with a periodicity of the order of 10μm), which are coupled to the neighbouring elements by evanescent coupling. The entire cross-section (in the order of 100μm) can be considered as a single coupled unit. In both a broad area laser and a laser diode array, the internal cavity can support different spatial modes, also referred to as array modes. Thus, although the free running output of a broad area laser or laser diode array is a multimode output, particularly supporting different spatial modes, some coherency exists across the output aperture.
In contrast, in a laser bar, a number of diode lasers (typically 20 to 40) are fabricated monolithically on the same chip, placed next to each other. The individual diode lasers can be regarded as separate light sources. The radiation from a laser bar is therefore essentially incoherent. The far-field intensity pattern from a laser bar has a shape similar to the pattern for the individual laser devices, i.e. the conventional two-lobe shape. However, it is not possible to
apply normal off-axis technology to the output of a laser bar, because it is not practical to treat separately the two lobes from each of the component laser diode arrays or broad area lasers in the bar, which are lasing at different frequencies, and are not in phase with each other. Thus hitherto it has not been possible to apply the off-axis techniques illustrated in Figure 1 , to diode laser bars.
Figure 2 illustrates another known arrangement, relating to a different branch of laser technology, as disclosed in "Spectral beam combining of a broad- stripe diode laser array in an external cavity", V. Daneu, et al., Optics Letters, March 15, 2000, Vol. 25. No. 6, page 405-407. A system is disclosed for co- aligning the multiple beams from a laser bar to a single output beam, but allowing the incoherence between beams to remain. A laser diode bar 21 produces a plurality of spatially separated primary beams 23 of which three beams are shown by way of illustration. There will normally be twenty to forty diode lasers in the bar producing a corresponding number of primary beams. The primary beams are collimated along the fast axis by a cylindrical lens 22. Along the slow axis the beam is imaged by a transform lens 25 into a far-field distribution at which the surface of a grating 26 is situated. The reflected light indicated at 27 is passed to an output coupler 28, which allows an output beam 29 to pass through the coupler, but reflects 10% of the light back to the bar laser 21 via the grating 26 and transform lens 25.
The feedback from the output coupler 28 to the laser bar 21 will be provided at highest amplitude where the beam 27 strikes the coupler 28 orthogonally. Since diodes are tunable devices (^20-60 nm) each diode laser of the bar will adjust itself to a specific lasing wavelength, which will cause the individual emitting diode of the bar 21 to be reflected into a preferred direction, using the dispersive element 26, in such a way to obtain the highest gain feedback, that is to say orthogonal to the coupler 28. When the beam has a normal angle of incidence on the output coupler 28 the beam will be reflected directly back into the emitting diode, forming an external cavity. In summary, the
individual outputs of the diode lasers in the bar will adjust their frequencies until the beam components are co-aligned in the light beam 27.
Turning now to Figure 3, an embodiment of the present invention is shown. A laser assembly 31 consists of a laser diode bar formed of a series of side-by-side laser devices 30, each of which may be a laser diode array or a broad area laser. The laser diode bar 31 produces a plurality of spatially separated primary beams 33 which in the free running mode without feedback lase at frequencies and phases which are unrelated between the separate laser devices 30 of the laser bar 31. Three laser devices 30 and three beams 33 are shown for illustration but normally between ten and twenty devices will be present in the bar, with a corresponding number of beams 33. A first lens 32 is inserted for collection and preferably collimation of the beams from the laser devices, which are highly divergent along the fast axis (typically 30 degrees). Alternatively the lens 32 may be replaced by a lenslet array of microlenses for both fast and slow axis collimation. However, each laser device 30 of the bar produces a dual lobe output, and the overall output of the laser bar consists of a combined dual lobe output forming first and second lobe light beams 46A and 46B, where each element laser has contributed with one of its far-field lobes to each. The two combined lobes 46A and 46B are symmetrically displaced on either side of a principal axis 31 A of the laser bar 31 which is perpendicular to the output facet of the laser bar. The light beams 33 pass through a transform lens 35, to be imaged on a dispersive device 36, which may conveniently be a reflective grating 44. The grating 44 is placed one focal length away from the transform lens 35 so as to convert the linear spacing of the primary beams 33 of the diode laser bar 31 into different angles of incidence of light at the surface of the grating 44.
The grating 44 reflects the light incident thereon to form two parallel light beams 37 and 38. The light beam 37 constitutes a feedback beam, derived substantially wholly from the first lobe 46A of the dual lobe output of the laser bar 31. The light beam 38 constitutes an output beam which produces the final
output from the laser apparatus, and is derived substantially wholly from the second lobe 46B. The feedback beam 37 is fed to a feedback device 40, which conveniently may be a non-adaptive, planar (or curved) feedback mirror 45.
Between the mirror 45 and the dispersive device 36 is positioned a spatial filter 39 restricting the feedback beam 37. The spatial filter 39 is provided for restricting the transverse lasing modes of the feedback light to one or more selected transverse modes from each primary beam, preferably such that each of the laser diodes 30 is brought to lase in a substantially single transverse mode. It is preferred that the spatial filter 39 is situated at the far-field plane where the different spatial modes are most separated. Conveniently the spatial filter 39 can be provided by a pair of opposed parallel edges, for example a pair of razor blades. Preferably the spatial filter is adjustable as to the width of the slit, and the position of the slit.
In operation, the feedback mirror 45 feeds back into the laser bar 31 a proportion of the light generated by the laser bar, via the grating 44. The path between the grating 44 and the feedback mirror 45 is such that there will be a single preferred beam path for which maximum feedback will be obtained. For the example of a plane mirror 45, the maximum feedback will be obtained when the feedback beam strikes the mirror orthogonally. Because there is this common path 37, defined by the intersection point 36A on the grating 44 and the normal to the feedback mirror 45, each individual laser device 30 in the laser bar 31 will vary its frequency of lasing until it obtains maximum feedback from the mirror 45. Because each element laser 30 is imaged at slightly different angles, depending on the actual lateral position in the laser assembly 31 , on to the intersection point on the grating, each of the laser devices 30 will adopt a different lasing frequency individual thereto, such that the dispersive device 36 will take the light from that laser and combine it into a commonly aligned path leading to the feedback device, which gives the maximum feedback. Thus each individual laser device 30 in the array 31 is forced to self-tune itself to a particular frequency, which will give it maximum feedback.
It then follows that because the grating 44 acts on both lobe beams 46A and 46B , the output beam 38 is also forced to follow a co-aligned path. The feedback lobe forces the laser devices to lase at appropriate frequencies so that the component beams therefrom strike the mirror 45 orthogonally to form a single commonly aligned feedback beam 37. Because the output lobe beam 38 is produced by the same lasing modes, the frequencies of the component beams in the output beam 38 will also adopt the required frequencies such that when they strike the incident face of the grating 44, the component beams will be co- aligned and overlap in the output beam 38.
It is important that the primary beams 33 intersect each other as accurately as possible at a common point on the surface of the grating 44, so as to be diffracted from the same point. As shown in Figure 3, the upper lobe primary beams 33 strike the grating 44 at a common area 36A whereas the lower lobe primary beams strike at slightly different points on the grating 44, the points being spread over a larger area 36B. This is because the lower part of the grating 44 is further from the laser bar 31 , due to the inclination of the input face of the grating 44.
In Figure 3a there are shown diagrammatic representations of the areas of various primary beams 33 falling singly, and in combination, on the surface of the grating 44 in Figure 3. In Figure 3a the area 33A indicates the area A of a typical primary beam 33, falling on the grating 44. In Figure 3b there is shown non- perfect imaging of, for example, six beams 33B falling on the surface of the grating. As shown various of the beams 33B overlap at the surface of the grating. The overall area of all the beams 33B falling on the grating, is greater than the area A of a single beam in Figure 3a, but less than the area of the total of all six beams if none were overlapping. Figure 3c shows the ideal imaging of the six beams on the surface of the grating 44. Supposing that the beams were of equal area A, they would, if ideally intercepting at the surface of the grating, have a total area equal to the area of one beam, A.
The extent to which the primary beams 33 overlap on the surface of the grating 44, determines the extent to which the component beams within the feedback beam 37 overlap, which in turn determines the spatial quality of the output beam 38, represented by the value of the parameter M2.
Considering N independent, but otherwise identical, emitting light sources 30 in a laser assembly 31 , the combined beam from such a freely running device, will be essentially N times more diffraction limited, as compared to a single emitting source of the assembly. Thus, the more emitting light sources there are, the bigger is the M2 value, essentially in a linearly relationship.
In the present invention, if the cross-sectional area of a single primary beam 33 on the grating surface is designated A, and the total cross-sectional area of each combined lobe beam 46A, 46B, due to various unwanted aberrations, is Ac, (>A), then M2 will essentially increase by a factor of Ac/A (>1). Thus the quality of the imaging optics 32, 35, as well as the production tolerances of the laser assembly, is very important. Ideally the cross sectional areas of all the primary beams forming the lobes 46A, 46B should be equal in size and shape as well as intersect at the same point on the grating. In that case the combined beam would have the same M2 value as a single emitter of the assembly.
Considering the feedback device 40, it is much preferred that the feedback device 40 can only feed light back to the laser assembly 31 along one specific direction. If not, the various component beams would not be co-aligned well but spread according to the reflection characteristics of the feedback device and the gain profile of the individual emitters. This could be allowed to happen for certain curved mirror embodiments of the invention but this is not preferred. If the combined beam has a divergence angle θc after grating, due to these kinds of effects, the M2 value will increase by a factor of (θc/θ), θ being the divergence angle of a single component beam after the grating.
These considerations give rise to a number of preferred criteria for the parameters of the apparatus. Preferably the imaging optics 32, 35 are arranged to image the primary beams 33 at the dispersive device 36 in such a manner that, in respect of at least the first lobe beam 46A, and preferably in respect of both the lobes 46A and 46B, the primary beams 33 overlap each other as completely as possible at the surface of the grating 44. Practical criteria will now be set out in respect of the first lobe beam 46A in Figure 3, although it is to be appreciated that the same criteria are preferably also applied to the second lobe beam 46B. (One way of achieving the same criteria with respect to lobe 46B, is shown hereinafter in Figure 4).
One practical criterion for the imaging is that in respect of the first lobe beam 46A, the total area of the overlapping primary beams ATP at the dispersive device 36 is given by the expression
where N is the number of laser devices 30 in the laser assembly 31 and ALp is the largest area of an individual primary beam at the dispersive device 36 for the lobe 46A. Most preferably, all the primary beams within each lobe fall within the area of the largest primary beam, at the dispersive device 36. Preferably all primary beams have the same cross-sectional area.
It is also preferred that in at least the feedback beam 37, and preferably in both the feedback beam 37 and the output beam 38, the component beams in each combined beam overlap each other as completely as possible.
One practical criterion for this is that in the feedback beam 37, the total area of the overlapping component beams ATF is given by the expression
ATF < . A|_F (Equation 2)
where N is the number of laser devices 30 in the laser assembly 31 and ALF is the largest area of an individual component beam in the feedback beam 37.
Another practical criterion for the feedback beam 37 is that the component beams thereof are co-aligned within an angle αjF given by the expression
OCTF M N . CCLF (Equation 3)
where N is the number of emitters in the light source and F is the largest angle of divergence of an individual component beam in the feedback beam 37.
Considering now the output beam 38, one practical criterion is that in the output beam 38, the total area of the overlapping component beams Aτo is given by the expression.
where N is the number of laser devices 30 in the laser assembly 31 and ALo is the largest area of an individual component beam in the output beam 38.
Another practical criterion is that in the output beam 38 the component beams thereof are co-aligned within an angle ατo given by the expression
αTo N . CC[_o (Equation 5)
where N is the number of emitters in the light source and αLo is the largest angle of divergence of an individual component beam in the output beam 38.
Preferably these criteria are observed for both combined beams 37 and 38, and one convenient way of achieving this will be explained hereinafter with reference to Figure 4. Overall it is preferred that in each combined beam 37, 38 the component beams are aligned within substantially the same cross-sectional area and divergence angle as the largest component beam alone.
In practical terms, in an embodiment of the invention such as shown in
Figure 3, the two main criteria to be observed when setting up the apparatus are firstly that the imaged primary beams of the first lobe 46A should coincide as closely as possible at the surface of the grating 44, indicated as area 36A. The extent to which the ideal of all the primary beams overlapping within the area of the largest primary beam can be monitored by the criterion of Equation 1. Secondly, the feedback device 40 should be such as to have as closely as possible a single preferred direction of reflection of light, so as to have a single path of the feedback beam 37 which gives maximum feedback. For a plane mirror, the preferred direction of reflection for maximum feedback is a direction normal to the plane mirror. For other forms of feedback device, such as a curved mirror, or a phase conjugate mirror (as will be described hereinafter) there may be a small range of angles of reflection giving maximum feedback, constituting a generally preferred direction. The extent to which the ideal of a single preferred direction of reflection is achieved, can be monitored by the criteria of Equations 2 and 3.
In an ideal situation, which can be approached by appropriate design and adjustment of the lens system 32, 35, and the use of a plane mirror 45, the criteria for adjustment of the imaging of the primary beams and for the reflection by the feedback device can be expressed in terms of Equations 1 and 3 when:-
ATP = ALP,
ATF = A F, and TF = CtLF-
The final ideal criteria for the required output beam can be expressed as follows.
Aτo = ALo, and ατo = otLO-
There will now be described an example of practical steps of adjustment and design of the embodiment of Figure 3, the steps also being generally applicable to other embodiments of the invention.
Firstly, the criterion of Equation 1 is applied to adjust the imaging system 32, 35, to produce the required degree of overlap of the primary beams at the area 36A in Figure 3. This is done by removing the grating 44, and replacing it with a detector positioned at the area 36A for measuring the intensity profile. For example the detector may be a CCD camara or a beam scanner based on scanning knife edges. The CCD camera consists of a 2 dimensional array of light detecting elements (pixels). Typical CCD cameras might have 1024 x 1024 pixels, each pixel having a dimension of 25 μm by 25 μm. Thus the detecting area is in the order of 25 mm by 25 mm. Each pixel provides an electrical signal in proportion to the power of incident light. When a light beam is imaged onto the 2 dimensional array of detectors, the intensity profile of a light beam can be mapped, that is the power distribution in a transverse plane of the beam. The beam scanner can do the same, but in a slightly different manner. Both types of instruments are standard equipment in optical laboratories. The detector is used to measure the area of the first lobe 46A at the area 36A. In order for this to be done to determine the area ALp, (the largest area of an individual primary beam at the dispersive device 36), a masking element (not shown) is introduced between the laser diode bar 31 and the lens 32, so as to mask the output of all but one of the laser devices 30. The area of the respective laser device 31 is
measured at the position which will be occupied by the grating 44 in normal operation, and the masking element is then moved so as to reveal another single laser device 30. The measurements are taken throughout the series of laser devices 30 until the largest area of a primary beam 33 at the grating position is located. This is then noted as the area ALP. Next the detector measures the total area of the imaged first lobe 46A at the position intended for the grating 44, with all laser devices 30 unmasked. From these measurements the criterion set out in Equation 1 can be monitored, and appropriate adjustments of the lens system 32, 35, can be made to achieve the required overlap.
In Figures 3d and 3e there is shown a modification of the embodiment of Figure 3, to allow adjustment of the apparatus to achieve the preferred criterion set out in Equation 3 when applied to the feedback beam 37. There is shown in Figures 3d and 3e the insertion of a partially reflective mirror 40A in the feedback beam 37. The partially reflective mirror extracts a monitoring beam 37A which is directed to a detector 37B, which may be for example a CCD camera or a beam scanner. In order to determine the angle αLF in Equation 3, the laser diode bar 31 is blanked by a blanking plate (not shown) so that only a single beam 33 from one of the laser devices 30 is imaged by the lens system 32, 35, at the areas 36A and 36B, on the grating 44, as shown in Figure 3d. In such an arrangement, the two lobes 33A and 33B consist only of the two lobes of the single laser device 30 which is unmasked. The monitor beam 37A will have a divergence angle α which will be measured by the detector 37B. The divergence angle is measured for each of the laser devices 30 individually, and by way of example in Figure 3a the upper laser device 30 is shown as having the largest individual divergence angle CC F-
Next, the same half silvered mirror 40A is used to produce a monitor beam 37A, when all the laser devices 30 are unmasked, as shown in Figure 3e. The detector 37B then measures the divergence angle CCTF for the monitor beam 37A taken from the apparatus with all laser devices 30 unmasked.
When it is desired to monitor the apparatus utilising the criterion of Equation 5, with reference to the output beam 38, there is no need to use the partially reflecting mirror 40A. Instead the detector 37B can be placed directly in the path of the output beam 38, and can be used to measure the divergence angle of the output beam 38. This is done firstly with only a single laser device 30 unmasked, in an arrangement corresponding to that of Figure 3A. Then the measurement of the divergence angle is made for the output beam 38 when all of the laser devices 30 are unmasked.
The operation of the apparatus can also be monitored by reference to
Equation 2 or Equation 4. This can be carried out by placing the detector which has been described hereinbefore with reference to Equation 1 , in a position to monitor either the feedback beam 37 or the output beam 38. In the case of the feedback beam 37, the partially reflecting mirror 40A can be used as described above, but with an appropriate detector to detect the degree of overlap of the component beams. In the case of the output beam 38, the detector may be placed directly in the path of the output beam, so as to monitor the degree of overlap of the component beams in the output beam 38.
Thus embodiments of the invention allow an off-axis configuration to be applied to a bar laser, which would not be feasible by attempting to treat separately each of the primary beams from the bar laser, and to process the output lobe and the feedback lobe separately from each component device. Note that typically there are twenty diodes situated along a 1 cm long stripe, each diode having a typical dimension of 200 micron. The nonlasing area between two lasing elements is approximately 300 micron. Consequently the complicated structure applied to a single diode array or broad area laser in the disclosure of WO98/56087, would be extremely cumbersome if applied to each component of a laser bar. Even if this were succeed, a new difficult task would arise in combining the component beams into a commonly aligned beam since each component beam would have the same wavelength.
To summarise the embodiment of Figure 3, the laser assembly 31 comprises a plurality of spaced apart laser devices 30 for producing a plurality of respective primary light beams 33, each primary light beam contributing with its far-field lobes to the multiple lobes 46A, 47B, in its far-field intensity distribution pattern. The light feedback device 40 forms a resonant external cavity with the laser assembly 31 for reflecting to the laser assembly a feedback portion 37 of the light produced by the laser assembly, the light feedback device 40 having a preferred direction of reflection of light which gives maximum feedback. The dispersive device 36 is such that for a given angle of incidence it transmits or reflects light at different angles for different frequencies. The dispersive device 36 is positioned to receive light from the laser assembly 31 and to transmit or reflect light to the feedback device 40. The lenses 32 and 35 constitute an imaging device for imaging at the dispersive device 36 a combined far-field intensity distribution pattern having spatially separated multiple lobes 46A, 46B corresponding to the multiple lobes of the individual primary beams 33 the lobes being directed off-axis relative to the axis 31 A of the laser assembly. The arrangement is such that each laser device 30 lases at a different frequency such that its light is directed by the dispersive device 36 along the said preferred direction of the light feedback device 40 to obtain maximum feedback, and that the dispersive device 36 combines the primary 33 to form the combined feedback beam 37 reflected back to the laser assembly 31 , and to form the combined output beam 38 leaving the apparatus. The components in each combined beam overlap each other and have a common axial direction. The light feedback device 40 provides a reflectance of 100% to light in the off-axis lobe 46A of the said combined far-field intensity distribution pattern. No feedback is reflected from light in the other off-axis lobe 46B of the far-field pattern. The feedback beam 37 passes through a spatial filter 39 to isolate in the feedback beam 37 a substantially single transverse mode from each primary beam 33.
There will now be described in more detail examples of components which may be used in embodiments of the invention, exemplified with reference to Figure 3. The laser diode bar 31 may be a linear GalnAsSb/AIGaAsSb quantum-
well laser array operating at a nominal 2.05-μm wavelength. The laser bar is composed of 11 broad-area stripes of 100-μm width spaced with 300-μm centre- to-centre spacing. The array has a high-reflectance coating on the back facet and an antireflection coating on the front facet, with an estimated front facet reflectivity of <0.2%. The output is collimated in the plane perpendicular to the array (i.e. along the fast axis) by the lens 32 which is an antireflection-coated cylindrical lens having a numerical aperture of -0.7. The transform lens 35 is a spherical plano-convex lens, having a focal length of 12 cm, and is antireflection coated for 2 μm. The distance from the grating 44 to the mirror 45 is 45 cm. The Al-coated grating 36 is ruled with 600 lines/mm and has a blaze wavelength of ~1.6 μm. The grating period is chosen such that the grating has only zero- and first-order diffraction orders to eliminate efficiency losses to higher diffraction orders. The beam in the common arm of the resonator is propagating in a direction that forms a 26.7° angle with respect to the grating normal. The feedback element 40 may be a piano mirror coated with dielectric layers for high reflection (e.g. R> 99%) at 2 μm. The mirror is mounted on a stage for fine alignment of the mirror angle with respect to the direction of the zero or first order reflection of the grating. In the feedback beam 37, the spatial filter 39 is formed by two razor blades mounted on translation stages, and by adjusting the position of the blades different array modes can be selected. The size of the slit of the spatial filter depends on the lenses involved but is of the order of 1 mm. The slit is aligned so as to pass substantially only one array mode from each diode. The preferred modes from the diodes are those with high mode number. Moreover the spatial filter is located preferably substantially in the far field for good mode discrimination, i.e. next to the grating (e.g. within 20 mm).
There will now be described with reference to Figures 4 to 7 a series of modifications of the embodiment of the invention shown in Figure 3. Where components are common to the embodiment of Figure 3 and subsequent embodiments, like reference numerals will be used. For simplicity, in Figures 4 to 7, only two of the laser devices 30 are shown and only two of the beams 33 are shown.
Figure 4 is a diagrammatic representation of a modification of the embodiment of Figure 3 in which the dispersive device 36 (e.g. the grating 44 in Figure 3) is constituted by two similar dispersive elements 47 and 48 e.g. two similar gratings. By using two gratings both focusing points (indicated at 36A and 36C) of the transform lens can be located at the surfaces of the respective two dispersive elements. This results in a more precisely co-aligned out put beam 38. Thus in the embodiment of Figure 4, the dispersive device 36 comprises two dispersive elements 47 and 48 each having an inclined input face which is inclined relative to the axis 31 A of the laser assembly 31. The input faces are displaced from each other transversely relative to the said laser axis so that the two off-axis lobes 46A, 46B are incident on the two dispersive elements 47, 48 respectively, and the input faces of the two gratings are positioned longitudinally relative to the axis 31A in such a manner as to allow each off-axis lobe 46A, 46B to be imaged on an input face equidistant from the laser assembly 31 , thus giving the possibility to optimise the overlap of the component beams in each lobe 46A and 46B.
Figure 5 is a diagrammatic representation of another modification of the embodiment of Figure 3, in which the feedback beam 37 is not spatially filtered. The embodiment uses a non-adaptive planar mirror 45, and there is provided no spatial limiting device equivalent to the spatial limiter 39 in Figure 3. The outcome of this modification is that a multilobe output beam 38 is emitted from the apparatus. This may have advantage in some applications where a higher overall output power is required, without any particular need for single mode operation. In this case although no spatial filtering is taking place, the power in the feedback lobe is coupled into the output lobe, thus providing a single output beam containing the major part of the output power. Note that grating in this configuration can be translated along the principal axis 31A so as to co-align the output beam 38, so as to obtain the best M2 quality. In this case the feedback beam 37 is not co-aligned, but consists of separate (parallel) feedback beams. This is a preferred form of embodiment of figure 5. The same effect can be
achieved by placing the feedback device 40 of Figure 5 in the non-overlapping lower beam from the grating 44, as shown in Figure 5a. Here the positions of the beams 37 and 38 are interchanged.
Figure 6 is a diagrammatic representation of a modification of Figure 3 where the feedback device 40 is a phase conjugate mirror 60 (e.g. a BaTi03 crystal). For proper mode matching of the feed back beam 37 into the phase conjugate element 60 a lens 61 is included. The phase conjugate element 60 has the ability to build up gratings acting as reflectors for each primary beam 33 independently of the angle of incidence. Thereby the phase conjugate element 60 can offset different kinds of aberrations arising from production tolerances of the diode laser bar and non-perfect imaging through the optics system of the apparatus. Here the aperture 39 determines the directional property of the adaptive mirror 60.
Figure 7 is another modified embodiment of the invention. In this embodiment the primary beams 33 of the laser bar 31 are combined and co- aligned in an output beam 37B derived from a portion of the first lobe 46A of the bar laser output. The main feedback is provided at beam 38A by the second lobe 46B, and additional feedback is provided at beam 37A by a portion of the first lobe 46A. (This is in contradistinction with the embodiment of Figure 3, where the output beam is derived from one lobe, and the feedback is derived from another lobe). In Figure 7 the first lobe 46A is focussed on the incident surface of the grating 44 as in Figure 3, and passes through a spatial filter 39. However, in this embodiment an output coupler is provided by a partially reflective planar mirror 53 which reflects a proportion of the lobe 46A as feedback and transmits another proportion of the lobe 46A to become the final output beam 37B. Conveniently the partially reflecting mirror 53 is arranged to transmit 90% of the light striking it and to reflect back 10%. Thus partial feedback from the partial reflector 53 and spatial filtering at the filter 39, are provided in an output section of the feedback device 40. The main feedback is applied in a feedback section of the feedback device 40 by a mirror 52. Thus in this
embodiment, there are provided two feedback elements 52 and 53, which conveniently may be dielectric mirrors. The reflection coefficient of the output mirror 53 can be 10%, thus transmitting 90% as useful output power. In the feedback section, a high reflection mirror 52 can be used, e.g. with reflection coefficient more than 99%, reflecting effectively all the incident power. The spatial filter 39 may be a pair of opposing razor blades, forming a slit for spatial mode selection. (Alternatively the spatial filter may be provided by the entrance aperture of a wave guide, e.g. an optical fibre end). The spatial filter should preferably be situated in the far-field plane. The advantage of this arrangement over the embodiment shown in Figure 3 is two fold. Firstly, since the dispersive device 36 is positioned at the focus of the lobe 46A, aberrations are minimised. Secondly the aperture 39 acts as a limiting aperture of the output beam 37B, thus making the output beam more stable in terms of beam pointing stability. Also some compensating effect towards thermal changes in the diode laser bar is present.
In the previously illustrated embodiments, the dispersive device has been shown as a reflective grating. In such a case best results are obtained if the imaging device 32, 35, is arranged to image the primary beams 33 in such a manner that the primary beams 33 overlap on the front surface of the grating.
Many other forms of dispersive device may be used, for example a dispersive prism or a hologram. In Figure 8 there is shown diagrammatically the operation of a thick hologram as a dispersive device 36, in an embodiment of the invention such as shown in Figure 3 onwards.
Using a thick hologram (volume hologram) as the dispersive device 36, (volume hologram) the incident primary beams 33 need not overlap at the entrance surface of the hologram 36, but just need to interact in a region inside the thick hologram. Considering only two primary beams 33 (indicated at λi and λ2 ), the first primary beam λi is diffracted in one region 36D. The next primary beam λ2 passes unaffected through this region due to its different wavelength,
but interacts further inside the hologram at 36E and is diffracted along the same axis 36F as the first beam λi. Thus the two beams λi and λ2 have been combined to form a single feedback beam 37. The imaging device (32, 35 in previous figures), is arranged to image the primary beams onto the dispersive element 36 such that the beams 33 interfere along a common axis 36F inside the dispersive element 36, using the frequency tuning properties of the component diode lasers, so that a commonly aligned feedback beam 37 emerges. It should be noted that the thick hologram might also provide gratings superimposed onto each other in the same spatial region.