GB2466221A - Method and apparatus for laser machining structures of different sizes by means of two different laser processes - Google Patents

Abstract

A new method that uses a common optical system and sequentially creates structures of widely different sizes in a substrate by means of two different laser processes is described. One process uses a laser beam that is tightly focussed on the substrate surface and is used for creating fine groove structures by semi-continuous direct write type beam movement. The second process uses a second laser beam that is used to form a larger size image on the substrate surface and is used to create pads and vias in the substrate in step and drill mode.

Description

METHOD AND APPARATUS FOR LASER MACHINING STRUCTURES OF

DIFFERENT SIZES BY MEANS OF TWO DIFFERENT LASER PROCESSES

TECHNICAL FIELD

This invention relates to a new method and apparatus for creating structures of different sizes in a substrate using two sequential laser processes and a common optical system. It is particularly relevant to the sequential formation of narrow groove or trench structures and larger area pad and blind via structures in the top surface of a layer of polymer for the purpose manufacturing micro-electronic circuits.

BACKGROUND ART

Lasers are widely used in the manufacture of advanced PCBs. A particularly well known example is the drilling of blind micro-vias in multi layer PCBs. In this case UV lasers are often used to drill through a top copper layer and an underlying dielectric layer to allow contact to be made to a lower copper layer. In some cases the cost effectiveness of this process is improved by using two different laser processes to remove the two different materials. A I.JV DPSS laser is usually used to drill the holes in the top copper layer to expose the lower dielectric layer and in a separate process a CO2 laser is used to remove the dielectric material exposed below each hole. Such a two stage laser process can have many economic advantages as each process can be optimized separately. In such a case however because of the very different optical requirements two physically separated optical systems are used and the substrate has to be moved between the two systems to enable the second process to be performed.

Recently a new type of high density multi-layer circuit board manufacturing technology has been proposed. US2005/0041398A1 and publication "Unveiling the next generation in substrate technology", Huemoeller et al, 2006 Pacific Micro-electronics Symposium describe the concept of "laser-embedded circuit technology".

In this new technology lasers are used to directly ablate fine grooves, larger area pads and also blind vias in organic dielectric substrates. The grooves connect to the pads and vias so that after laser processing and subsequent metal plating a complex pattern of fine conductors and pads embedded in the top surface of the dielectric layer is formed together with deeper vias connecting to a lower metal layer. Up to now pulsed UV lasers have been used to form the grooves, pads and vias in a single process using either direct write or mask imaging methods.

The direct write approach generally uses a beam scanner to move a focussed UV laser beam over the substrate surface to scribe the grooves and also create the pad and via structures. This direct write approach uses the highly focusable beam from UV lasers with high beam quality and hence is very well suited to the fine groove scribing process. It is also able to deal well with the different depth requirements associated with pad and via structures. By this method grooves, pads and vias of different depth can be readily formed. However because of the limited laser power available from highly focusable UV lasers this direct write process is slow when it comes to removing the more substantial volume of material associated with the larger area pads and vias. This direct write method also has difficulty maintaining constant depth at the intersections between grooves and pads The mask imaging approach uses an excimer laser to illuminate a mask containing the full detail of the circuit design. An image of the mask is projected onto the substrate and the mask and substrate are moved together to allow the full area of the circuit to be reproduced on the substrate. Since the whole area of the mask is scanned during the image transfer process this approach is insensitive to the area of the structures to be created and hence is well suited to creating both the fine grooves and also the larger area pads. It is also excellent at maintaining depth constancy at the intersections between grooves and pads. It is not however well suited to creating structures of different depth as required especially for the vias. In addition except in the case where the circuitry is extremely dense this mask imaging is approach is significantly more costly than the direct write approach since the purchase and operating costs of excimer lasers are both very high.

Hence it can be seen that these existing single step process methods for making advanced circuits based on this "laser-embedded circuit technology" have serious disadvantages and to improve the process rate and reduce the cost there is a requirement to use laser processes that are separately optimized for creating the different size and depth structures required. Such a two stage process using common final beam delivery optics is disclosed.

DISCLOSURE OF [NVENTION

This invention is based on the sequential use of two different laser beams using different beam motion and process methods in order to create interconnecting structures of two different sizes in the surface of a polymer layer. Both laser beams are delivered to the substrate through a common optical path so that motion of the substrate between processes is unnecessary.

The first key aspect of this invention is that fine scale groove structures are created on the substrate surface using a direct write approach. In this case the laser beam is focussed by a lens to a small spot and which is moved in a series of continuous paths to vaporize the polymer to form the required number of discrete lengths of groove.

The groove width and beam speed requirements dictate that the laser used for this direct write grooving process must operate either continuously or if pulsed must have a repetition rate that exceeds some minimum value. In an extreme case the groove width may be as small as 10pm and as up to 10 laser pulses may be required to remove material to the depth required if beam speeds up to several metres per second are used pulsed laser repetition rates exceeding several MHz are needed. For wider grooves lower repetition rates are acceptable. It is generally required that the depth of the groove be maintained constant along its length and hence the laser beam repetition rate must be sufficiently high that the distance travelled by the beam over the substrate between pulses is substantially less than the groove width. In general it is likely that repetition rates exceeding a few 100kHz will be required.

Ideal lasers for this direct write grooving process operate either continuously (CW lasers) or operate at such a high repetition rate that they behave like CW lasers. Such lasers are called quasi-continuous (QCW) and generally operate at repetition rates in the 80 to 120MHz range.

The second key aspect is that larger scale structures such as pads and blind vias are created on the substrate surface using a "step and drill" process using a pulsed laser.

In this case, rather than focussing the laser beam, an imaging method is used to form a laser spot of the required size on the substrate surface. An aperture is placed in the beam after the laser and this aperture is imaged by a lens onto the substrate to form a laser spot of well defined shape. The beam is held stationary at the required pad or via location and a "burst" of laser pulses are fired to remove material to the required depth. After a pad or via is drilled the beam is moved to the next location and the process repeated.

Such a mode of operation requires a laser that can emit sufficient energy per pulse to create an energy density on the substrate surface that exceeds the ablation threshold by some margin. In an extreme case pad diameters up to 0.3mm may be required with energy densities of a several J/cm2. Such a requirement leads to a need for a laser beam containing up to several mJ in each pulse. For smaller pads and vias correspondingly less energy per pulse is required. In general the minimum pulse energy requirement is likely to be a few tens of iJ Lasers used for this "step and drill" imaging mode process operate in pulsed mode where the pulses are emitted in a series of "bursts" each burst containing a limited number of pulses. Suitable lasers can be of high beam quality with highly focusable beams but the use of lasers with lower beam quality, so called multi-mode lasers, is preferred since these operate at much higher power levels.

In general it is expected that two different lasers will be used sequentially for the two different processes of grooving and pad or via formation but in some cases it may be appropriate to use a single laser that can operate satisfactorily in both modes. Such a laser needs to operate with high beam quality at high repetition rate or in CW mode so that is able to form narrow grooves at high beam speed and also operate at lower repetition rate with significantly higher energy in each pulse for the formation of blind holes.

Clearly a very important characteristic of the lasers used for this dual process is that they must operate at a wavelength that is strongly absorbed in the dielectric material.

Alternatively in the case of pulsed lasers then in some situations these may operate at a wavelength outside that which is strongly absorbed by the dielectric material so long as the laser pulse duration is sufficiently short such that the intensity of radiation in the focal spot is sufficiently high to cause the beam to be absorbed by non-linear processes An important aspect of this invention is the method used to move the beams over the substrate surface. The simplest method to move the laser beams is by motion of the substrate in two axes under a stationary lens. This method is generally slow and therefore the preferred method for this invention is to use a two axis beam scanner unit to deflect the beams rapidly in two orthogonal directions. Such scanner units are very well known and are generally used with a lens situated after the scanner. In this case so called f-theta lenses are often used since this type of lens is designed to operate in this mode and create as far as is possible a focal spot of constant size and shape over a flat field. In some cases however it may be appropriate to use a lens situated before the scanner unit. Such an arrangement generally gives rise to a curved focal plane and in this case the use of a dynamic variable telescope situated before the lens to adjust the focal plane is common. The use of a dynamic variable telescope with an f-theta lens situated after the scanner is also possible. Such arrangements with a dynamic variable telescope are usually referred to as three axis scanners.

Another key aspect of this invention is that the beams for the two laser processes are delivered to the substrate through a common optical path. For the case where two separate lasers are used for the two processes it is necessary to combine the beams alternately into the common beam path consisting of the scanner and the lens. One way to do this utilizes a moving mirror that is switched into one of the beam paths in order to inject the other beam. An alternative way utilizes the fact that generally most laser beams are polarized. In this case the mirror remains in position in one of the beams at all times. The mirror has a special dielectric coating that reflects preferentially a beam of one polarization and transmits preferentially a beam of the orthogonal polarization. In this case the first and second beams are arranged to have orthogonal polarizations at the beam combining mirror. Such polarization based beam combining methods are well known.

In both beam combining methods the two separate beams are adjusted to be spatially coincident at the surface of the mirror and the angle of the mirror is adjusted to ensure both beams propagate through the scanner and lens any without angular offset. This ensures that the focal spot generated by the first beam is coincident with the imaged spot generated by the second beam at the substrate surface.

The arrangement where two separate lasers are used is convenient as each path can be optically conditioned before the beams are combined. It is desirable that the beam from the laser generating the first beam for grooving has telescopic optics in its beam path to change the beam diameter so that the focal spot generated on the substrate surface is of the correct diameter. It is also desirable that the beam from the laser generating the second beam for pad and via drilling has separate telescopic optics in its beam path to change the beam diameter to match the size of the imaged aperture.

With separate beam paths the aperture can be permanently installed in the beam path and adjusted to be coaxial with the beam. In addition further optics can be installed in the second beam path after the aperture to condition the beam to allow the image of the aperture to fall in exactly the same plane as the focus of the first beam so avoiding the need to make any corrections when switching between the two process modes For the case where a single laser is used to alternately generate the first and second beams however such independent beam conditioning is not possible and mechanical movement of any beam size changing optics, the aperture and its associated optics in and out of the common beam path are necessary.

A key aspect of this invention is that since the drilling process with the second beam uses an imaging process the shape and size of the blind pads and vias produced on the substrate are defined accurately by the shape and size of an aperture situated in the beam at the object plane corresponding to the image on the substrate. In general it is expected that in most cases the pads and vias will be circular in shape and therefore a circular aperture is used in the beam path. Apertures of any shape however can be used in order to create pads and vias of non circular shape.

Apertures can consist of a simple metal plate with a hole or may be made of a transparent substrate with a patterned opaque coating. Such a substrate based device is termed a mask. Such masks can contain complex structure if required. For example a mask with a circular aperture in the opaque coating might contain an opaque region in the centre. Such a mask would give rise to a pad with annular shape on the substrate. It is also possible to envisage masks that have variable transmission of the beam within a region. The use of such masks would give rise to images on the substrate containing area of different energy density and hence different depths after laser vaporization.

In general the imaging optical system used demagnify the aperture or mask onto the substrate so that higher energy densities can be achieved on the substrate without risk of damage to the mask or aperture. Typical pad and via sizes are from a few lOs of microns up to a few lOOs of microns in diameter. Typical aperture dimensions may be up to a few mm in diameter so demagnification ratios from a few to 10 or even more are generally used.

Since pads and vias of different size are required even within individual circuits it is desirable that the aperture size can be readily changed. This is accomplished by the use of motorized variable iris if the aperture is circular or by the use of a motorized mask or aperture changer unit. Such motorized devices are well known.

Perhaps the most important aspect of this invention is that it is well suited for controlling the depth of the structure in the regions where pads and grooves intersect.

It is important from the point of view of the electrical performance of embedded conductor based circuits that the depth of the grooves is similar to the depth of the pads and that the depth in the intersection region is not significantly different to either the pad or groove. To achieve this using a direct write approach for both grooves and vias is very difficult since with such an arrangement there are many regions where the focal spot trajectory has to stop and start and highly accurate control of beam position and power is necessary. With the dual process method described in this invention the degree of control of beam position required in the pad and groove intersection regions is significantly relaxed. In general it expected that for each circuit the grooving process will take place before the pad and via drilling process but this is not essential.

A control system that is able to co-ordinate the motion of the 2 axis scanner with the motorized aperture together with the power and triggering controls of both laser types is an essential part of this invention. Such control systems are commonly used in the laser marking and micro-machining industries.

All the discussions above have concerned the use of a single scanner and lens to deliver the beams alternately to the substrate. In practice, to speed process rate, it is likely that several optical channels will be used in parallel. With sufficient laser power available from both lasers it is possible to split the beam after the combination point such that two or more scanners and lenses can operate in parallel and process different devices on the same circuit board at the same time. After these devices have been completed the substrate is stepped to a new location so that further devices can be processed. Operation over a large panel having multiple devices is thus in "step and scan" mode. Clearly devices processed in parallel using the same lasers at the same time like this must have the same circuit features.

Most circuit boards using this embedded conductor technology are constructed on a core layer with layers of different electrical circuits built up on opposite sides. The methods disclosed in the present invention can be readily extended to this situation to allow simultaneous processing of different circuits on opposite sides of the same device at the same time. In this case separate optical assemblies consisting of first and second lasers, aperture, beam shaping optics, combining mirror, scanner and lens are required for each of the two sides of the circuit board so that different circuit designs on opposite sides can be realized. A production laser tool for the high speed manufacture of multiple, multilayer, dual sided devices based on embedded conductor technology might well consist of two or more scanner and lens systems operating on one side of the circuit board and an identical combination of lasers and optics operating on the opposite side at the same time

BRIEF DESCRIPTION OF DRAWINGS

Aspects of the method and exemplary embodiments will now be described with reference to the accompanying drawings of which: Figure 1 shows how grooves are made in the substrate surface with a focused beam from a laser Figure 2 shows how blind holes are made in the substrate with a laser beam operating in imaging mode Figure 3 shows the typical optical systems used for both processes Figure 4 shows how the beams from two different lasers are combined into a common beam path and moved over the substrate surface to sequentially create groove and blind hole structures Figure 5 shows how a single laser can be used in both focussed mode and imaged mode to sequentially create grooves and blind holes in the surface of the substrate Figure 6 shows an example of an apparatus used to perform the two processes on a double sided circuit board

DETAILED DESCRIPTION OF DRAWINGS

FIGURE 1 Figure 1 illustrates the process whereby a first laser beam is used to create fine groves in the surface of a substrate. Laser beam 11 is of high quality having diffraction limited or close to diffraction limited properties. Lens 12 focuses the laser beam onto the surface of the substrate 13 to create a small spot. The energy absorbed causes the substrate material in the region of the focal spot to be vaporized. The laser beam is moved over the substrate surface so that a groove 14 is formed. A straight groove is shown in the figure but in practice grooves can be of any shape set by the motion of the beam with respect to the substrate. Grooves can be of any length set by the limits of the mechanism that moves the beam with respect to the substrate. If the laser used is of CW type then the depth of groove formed will be of uniform depth along its length so long as the laser power, beam size and speed are held constant. If the laser used is of pulsed type then to maintain uniform groove depth along its length the repetition rate must be sufficiently high that the distance travelled by the beam over the substrate between pulses is substantially less than the groove width.

FIGURE 2 Figure 2 illustrates the process whereby a second laser beam is used to create larger scale blind holes in the surface of a substrate. Lens 21 is used to image an object plane 22 in the laser beam 23 onto the surface of a substrate 24. In the case shown a circular aperture 25 is situated at the object plane so that a circular spot 26 is formed on the substrate surface. The laser beam is held stationary with respect to the substrate and a burst of pulses is fired. The laser energy absorbed during each laser pulse causes the substrate material in the image area to be vaporized to a certain depth. The cumulative effect of the burst of pulses is that a blind hole 27 of defined depth is drilled into the substrate surface. Such a blind hole in the substrate material is termed a pad. The substrate may well be of polymer type containing a buried metal layer 28 and in this case if the energy density in the laser pulses is well controlled the drilling process generally stops when the laser beam penetrates to the metal. Such a blind hole 29 connecting to a lower metal layer is termed a blind via.

FIGURE 3 Figure 3 shows an embodiment of the invention where two separate lasers are used for the two processes. First laser beam 31 and second laser beam 32 propagate through separate paths until combined into the common optical path 33 that contains the lens 34 and substrate 35. The first beam has a beam size changing telescope 36 in its path to set the beam size to give the required focal spot size at the lens focal plane.

The second beam has a beam size changing telescope 37 in its path to adjust the beam size to match the aperture 38. Optics unit 39 situated after the aperture consists of a long focal length telescope which is set so that the reduced image of the aperture 310 is situated in the focal plane of the lens. Such an imaging arrangement is well know and is often termed "infinity" imaging since the image occurs in the lens focal plane and hence the corresponding object plane is at infinity. With this arrangement of the optical systems the distance from the substrate to the lens is maintained constant during the switch from one process to the other which is very convenient FIGURE 4 Figure 4 shows one embodiment of this invention where two different lasers are used to generate the two beams required for the two different processes. The beams are combined and thereafter propagate through a common beam path consisting of a beam deflection system and a lens. Laser 41 generates a first beam that is of high quality and operates CW or with high repetition rate such that it is suitable for the grooving process. Optics unit 42 changes the laser beam size to a value that is appropriate for propagation into the common beam path. Laser 43 generates a second beam that is pulsed such that is suitable for the pad and via drilling process.

Optics unit 44 changes the laser beam size to a value that is appropriate for illumination of an aperture 45. Both first and second beams are combined at a mirror 46 that is either switched in an out of the beam path to allow either the first or second beams into the common beam path or alternatively remains in position and reflects preferentially a beam of one polarization and transmits preferentially a beam of the orthogonal polarization. In this case the first arid second beams are arranged to have orthogonal polarizations at the beam combining mirror. In the case shown the first beam passes through the mirror and therefore it is desirable that it has a polarization direction in the plane of the paper (so called p-polarization). The second beam is reflected from the mirror and therefore it is desirable that it has a polarization direction that is perpendicular to the paper surface (so called s-polarization). Such techniques for combining beams of different polarization are well known. After beam combination the two beams pass to a 2 or 3 axis scanner unit 47 and a lens 48. In some cases it may be desirable to place the lens before the scanner unit. The scanner allows movement of the beams over the process area on the substrate 49. The lens has the function of alternately focussing the first beam and also imaging the aperture in the second beam onto the substrate surface. To avoid distance changes between the lens and the substrate as the system alternately switches between the two processes additional optics 410 can be placed in the path of the second beam after the aperture.

Such optics are generally of telescopic type with the aperture situated at the effective focus and the ratio of the effective focal length of the telescope to the focal length of the lens before the substrate chosen to create a de-magnified image of the aperture of the correct size on the substrate surface. Such methods for imaging masks or apertures in laser beams are very well known FIGURE 5 Figure 5 shows another embodiment of this invention where a single laser is used to generate the two beams required for the two different processes. To form grooves laser 51 generates a first beam that is of high quality and operates CW or with high repetition rate such that it is suitable for the grooving process. Optics unit 52 changes the laser beam size to a value that is appropriate for propagation into the common beam path in order to generate the correct size of focal spot. The common beam path consists of a scanner unit 53 and a lens 54 which focuses the first beam onto the substrate 55. To form pads or vias the operation of the laser is changed to pulsed mode such that it is suitable for the imaged spot drilling process. In this case the aperture unit 56 is moved into the beam path or alternatively the aperture is closed on the beam in order to define a beam size at the substrate. The distance between the lens and the substrate may be changed between the two processes to switch between focussed beam grooving and imaged beam drilling. Alternatively telescopic type optics 57 may be moved into the beam path to allow both processes to occur with constant distance between lens and substrate. In some cases it may be desirable to replace or supplement the beam shaping optics used for the grooving process with other optics 58 to condition the laser beam to match the aperture.

FIGURE 6 Figure 6 shows one embodiment of this invention consisting of an apparatus that is appropriate for performing the dual laser process on a double sided circuit board 61 containing multiple repeating devices 62 on each board. Lasers 63 and 63' that are of the same type generate two first beams for the top and bottom side grooving processes respectively. Other lasers 64 and 64' that are also of the same type, which is different to the type of the first lasers, generate two second beams for the top and bottom side pad and via drilling processes respectively. Lasers 63 and 64 operate alternately on the top side circuits while lasers 63' and 64' operate alternately on the lower side circuits. In this manner the circuits on the top and bottom sides can be different. First and second beams are alternately injected into common paths via mirrors 65 and 65'. The top and bottom common optical paths each contain beam splitters 66 and 66' to divide the beams into two in order to feed two parallel scanner and lens units 67 and 67' that operate simultaneously on the devices on the circuit board. The two scanners on the top side process two devices at the same time with one design of circuit while the two scanners on the lower side process the opposite sides of the two devices with the same or a different circuit design. With sufficiently high power lasers division of the beams on the top and bottom sides into more than two parallel channels is possible.