The pr^sgnt disclosure is related to Patent Cooperation Treaty publication WO 2011/08^592 A4 to M.Finarov, E.Tirosh and G.Dishon incorporated here by reference.

TECHNOLOGY FIELD

[001] The present method and apparatus relate to a field of manufacturing thin sheets of solid state material and particularly to manufacturing of thin wafers of crystalline silicon,

BACKGROUND

[002] photovoltaic (PV) technology employs solar panels or cells to convert sunlight into electricity. A photovoltaic system is made up of one or more photovoltaic (PV) panelj each comprising a number or cells. Silicon is the main component of the photqypltaic cells. The cells are typically designed to ensure the highest energy yield or better photo-electric conversion efficiency. Thin silicon wafers, for exarnpje, wafers 50 to 100 micron thick would have better photo-electric convf rsion efficiency than 200 micron thick wafers.

[003] ( urrently thin wafers of crystalline silicon or other solid state materials are produced by mechanical sawing a row material ingot with the help of abrasive disks or wires. Mechanical sawing is not cost or time effective, since most of the costly silicon material is wasted by kerfs losses. Also, because of significant breakage, sawinj of fragile very thin wafers is impossible.

[004] According to A. Skumanich, "The art of cutting in PV industry", Photovoltaics World magazine, May/June 2009 mechanical cutting of crystalline silicon ingots, either single crystal or polycrystalline ones, to thin wafers used as substrates for manufacturing PV solar cells the minimal wafer thickness is practically limited to abouf 120 um while about the same thickness of the material per each wafer is wastejj due to kerfs losses. Nevertheless such mechanical cutting is a dominant technoiogy in manufacturing silicon and other wafers primarily because of its mode? ate cost. [005] Mgcently another approach for preparing thin silicon wafers has appeared in the mark t. The approach is based on creating a sub-surface layer of weakened material structure by means of deep ion implantation, e.g. by hydrogen ion beams, followed by removal of the surface layer as a unit from the ingot [US 5,374,564]. Silicon Genegjg Corporation located in San-Jose (CA), USA has developed a kerf-free techn¾lpgy of cutting silicon wafers with help of ion implantation and further cleavage [US 6,013,563]. All these technologies apply high-energy ion implantation in vacuum and are very costly. The high cost prevents use of these technologies in low-cost production of solar cells.

[006] ||2009/0056513Al discloses another approach of cleaving thin wafers from single^erystal silicon ingot by illuminating the ingot with a high intensity and short pulse laser beam focused in a plane underneath the front surface of the ingot and moving a location of so heated region through that focus plane from a mechanically formeji notch along the whole surface. Such optically initiated and dynamically s read cleavage has not been practically proven yet.

[007] ξ ||11 another approach for making thin wafers is cleavage performed after creating a modifying material structure within a silicon ingot in a thin layer beneath the iij|ot surface by illuminating of high power ultra-short IR laser pulse causing mult¾photon absorption in silicon. EP2221138A1; and V.V. Parsi Sreenivas et al. Microjized subsurface modification of mono-crystalline silicon via non-linear absorption. J. Europ. Opt. Soc. Rap. Public. 7, 12035 (2012) disclose this method. For proper cleaving, the modified material structure may be formed either in a complete or partial layer. The optical coupling between air and silicon required to support this high power density was provided with the help of immersion liquid.

[008] It should be notes that after cleavage of the first wafer from the ingot the outer ingot surface could be not smooth enough and could cause significant light scattering and loss of power of laser illumination inside the silicon ingot and absorption of the laser radiation near the outer ingot surface. Patent Cooperation Treaty publication WO2011/089592 to the authors of the present application resolved these problems by illuminating the silicon ingot from an opposite ingot side.

[009] The following patents, patent applications and publications could be of interest: UJ 5,374,564; US 5,637,244; US 6,013,563; US 6,087,617; US 2009/0056513

SUMMARY 10] TJje current disclosure presents a method and apparatus for separation of a silicon wafer from an ingot by generating a few microns thick modified material structure layer located partially or entirely in the plane coplanar and located underneath the front surface of an ingot. The modified material structure layer suppers easy detaching of the wafer from the ingot. The front surface of the ingot becomes the front surface of the wafer and the thin modified material structure layer becomes the bottom surface of the wafer.

[0011] T ie presented wafer from ingot separation method performs the creation of a thin layer with modified silicon structure in a two stage process. At the first stage the ¾pt is exposed by a high laser radiation power using a large numerical aperture lens and short laser radiation pulses. The large numerical aperture lens limits, the depth of focus and supports formation of a high density laser radiation in a limited silicon material volume of about few microns size. The radiation density and pulse duration are selected to cause modification of silicon structure in a thin material layer that will become the wafer separation layer or plane. The system is configured to cause modification or defects of silicon structure at a plurality of spots or islands across the separation layer. Simple scanning methods such as galvQs!gontrolled (resonant) mirrors, mechanical moving of a large numerical aperture lens and others could be used to cause modification of the silicon structure in th separation layer.

[0012] I'he thin wafer separation layer including islands with modified silicon structure is further scanned by a scanning system using a low numerical aperture lens. The scanning system with a low numerical aperture lens also employs a high power laser beam and exposes the separation layer with short pulses of laser radiation. The throughput of the scanning system with low numerical aperture is highof than the throughput of the scanning system with large numerical aperture. The panning spot formed by a low numerical aperture system has a relatively large depth of focus. In order to ensure that during scanning with the low numerical aperture system the modified silicon structure will be created exactly in the separation layer, such scanning begins from the spots or islands located in the separation layer and created at the first step. The structural modifications or defects in these islands cause higher absorption of the used laser radiation than in rest of the ingot volume. The absorbed laser radiation causes expansion of the earlier generated and existing internal structure modifications or defects along the scanning direction. The modified internal structure of the separation layer supports easy separation of the wafer from the ingot by most of the known methods, such as cleavage of the wafer, with a reasonable throughput. BRIff LIST OF FIGURES

[0013] BIG. 1 is a schematic illustration of an example of laser radiation focusing within a silicon ingot using a lens with large numerical aperture (NA);

[0014] JG. 2 is a schematic illustration of an example of laser radiation focusing within a silicon ingot using a lens with a low numerical aperture;

[0015] IJG. 3 is a schematic illustration of an example of laser radiation scanning employing a large numerical aperture lens;

[0016] IJG. 4 is a schematic illustration of an example of laser radiation scanning employing a large numerical aperture lens and immersion liquid;

[0017]I|p . 5 is a schematic illustration of an example of laser radiation scanning employing a low numerical aperture lens;

[0018] FJ . 6 is a schematic illustration of an example of simultaneous laser radiation scanning of two planes inside a silicon ingot with a low numerical aperture lens; and

[0019] ig- 7 is a schematic diagram of an example of an apparatus for creating a wafer separation layer with a modified structure within a silicon ingot.

DETAILED DESCRIPTION

[0020] Currently thin wafers of crystalline silicon or of other solid state materials are produced by sawing a row material ingot with the help of abrasive disks or wires. The rjjain drawbacks of such technology are listed below:

• A significant amount of row material is wasted during sawing due to relatively thick kerfs;

• Mechanical stress generated during sawing limits the minimal thickness of the wafers and might cause damage to the wafers;

• After sawing the cleaved wafer contains a mechanically damaged surface layer. The layer has to be further removed by chemical etching; • The existing sawing technology utilizes significant amount of slurries and other chemicals that generate waste contaminating the environment;

• The current wafer separation processes use high cost chemicals and consumables.

[0021] ¾ffvertheless, such mechanical ingot sawing is a dominant technology in manufacturing silicon and other material wafers primarily because of its moderate cost.

[0022] 1¾e present disclosure suggests replacing the current wafer from the ingot separation process, by a method of wafer separation based on creation of a few micriy thin separation layer with a modified material structure. The layer of the modified material structure is located within the silicon ingot. Such complete or partiaj modification of the material structure of the separation layer located beneath and coplanar to the plane of the outer surface of the silicon ingot could be achieved for example, by scanning the separation layer with a focused laser beam. A contraction between the scanning throughput and the ability of focusing or concentrating the laser beam energy in the separation layer (or plane) impedes implementation of this technology.

[0023] The most efficient scanning methods such as rotating polygon, resonant mirror, e an thers apply an F-Theta focusing lens which can usually operate properly only at small or low numerical apertures (NA). Low numerical aperture lense upport a long depth of focus, although they are not suitable for focusing of the la^er beam energy into a few micron deep layer located at the selected separation plane beneath the front surface of the ingot and coplanar to the plane of the 0nt surface. Use of a large numerical aperture objective lens, usually mechanically displaced or scanned relative to the ingot with the help of X - Y motion system, limits the scanning throughput and respectively causes high manufacturing cost. Resolution of this problem could support development of new thin fer manufacturing technologies based on laser induced modified silicon structure.

[0024] F|G.l is a schematic illustration of an example of laser radiation focusing withirj a silicon ingot using a lens with large numerical aperture. Lens 100 accepts [0026] I| both cases, whether a lens with low 200 or large numerical aperture 100 is used jthe highest laser radiation density is located in focal volumes 116 and 204 extending along optical axis 136 and residing about the same plane 120. The laser radiafion density could be selected such, as to be sufficient to modify the silicon structure in plane 120 to such extent that it will allow easy separation of manufactured in this way wafer 132 from ingot 108. For example, the laser radiation density in focal volumes could be in the range of 0.1 to 10 J/cm ² in one pulsey The laser energy focused in plane 120 could be provided in pulse mode and the pul es could be short pulses for example, in the range of several femtosecond to severgl picosecond. Short duration pulses prevent power dissipation by heat conduction from the focal area/volume to adjacent ingot areas and facilitate multi- photon absorption.

[0027] ¾afer 132 separation from ingot 108 could be facilitated by modification of the siljcon structure in wafer 132 separation plane 120. The silicon structure in the separation plane could be modified completely in the whole separation plane 120 or at least in some segments or parts of separation plane 120 by generation areas or island with modified silicon structure surrounded by silicon with original or not modified structure. The depth of the islands or areas with modified silicon structure coulcj be shallow, for example a few microns.

[0028] partial or complete modification of the silicon structure in separation plane 120 using a large numerical aperture focusing optics and forming volumes or island of modified silicon structure would require relatively long time. Mechanical scanning of separation plane 120 by moving either laser beam focusing lens or the silicon ingot relative to each other would be limited by the speed of mechanical stages and frequency of the laser pulse flow. The incremental (move and stop) movement between two neighbor islands does not allow to reach the desired high throughput.

[0029] ie present method of of wafer from ingot separation by laser scanning includjs two stages. At the first stage, schematically illustrated in FIG. 3, the method employs a large numerical aperture lens 304 which focuses the laser radiation that modifies the silicon structure in a separation plane 308, which is locatfd inside ingot 312 and coplanar to ingot surface 332. The pulsed laser radiation focused by a large numerical aperture is scanned across separation plane 308 ajjd modifies the silicon structure in separation plane 308 only in a number of spots oj- number of islands 316. For example, in islands located in the nodes of a rectangular grid. Such process could be substantially faster and it could be implemented for example, by using a matrix 324 of small lenses 328 located at the nodes; §f a desired grid and in a close proximity and parallel to front surface 332 of ingot 112. Such matrix of lenses could be formed, for example, from a plastic materi l such as PMMA. A laser radiation beam 336 emitted by IR laser 340, for example, such IR laser as model Uranus 1500-1550-0800 manufactured by Polar nyx, Inc. San Jose, CA USA or model R-200 manufactured by RayDiance, PetaUma, CA USA could be oscillated around optical axis 348. An optical scarn gr, e.g. galvo-scanner 352, could oscillate a mirror 356 and scan laser beam 336 agross surface 332 of ingot 312. Other types of optical scanners could be also used, f r example, a moving focusing lens, or a number of lenses located on a line, moving a matrix of staggered lenses, acousto-optical deflector, and other laser beat f canning means. Timing and duration of the IR laser radiation pulses could synchronized with location and path of small lenses 328 assembled in the matrix 324.

[0030] I| should be noted that with any type of the optical scanning system illuminating the ingot, the illumination could be done from opposite to the front surfaejf side or back surface side (which is located more distantly from the illuminated plane than the front surface side, where further wafer separation could be df e) as it was described in Patent Cooperation Treaty Publication WO 2011^89592 to the same authors. Such method of illumination supports lower laser powef density on the first surface of the ingot as well as keeping this surface smoolJi while multiple wafer separation planes from the opposite or second ingot surfa| could be performed. In this case the focusing optics capabilities should be adapted and compensated for the thickness of the silicon ingot between the second ingot sjurface and the separation plane.

[0031] JG. 4 is a schematic illustration of an example of laser radiation scanning arrangement employing a large numerical aperture lens and immersion liquid. Use of imjnersion liquid that fills-in the space between the ingot surface and the small focusing lenses 304 allows increasing the numerical aperture of the laser radiation small focusing lenses 304. Immersion liquid 408 use could be facilitated, for example, by placing silicon ingot 312 within a container 404. Immersion liquid 408 could, fee contained in a container 404 in an amount sufficient to fill with the immejfiion liquid the gap or space between lens matrix 324 and the illuminated front |urface 332 of silicon ingot 312. Use of immersion liquid supports increase of numerical aperture of matrix 324 of lenses 328 and decrease the size of spots or islandjj with modified silicon structure. The power density of the laser radiation in the fqgused area could also be increased. The immersion liquid could be selected based, on known immersion liquids transparent in the used IR spectral range, e.g. wate¾ oil, glycerin or the like. Refraction index of these liquids is higher than the refra§ |on index of the air.

[0032] Recording to another example, the islands with modified structure could be generated with the help of a large numerical aperture lens laser illumination by scanrjjng continuously the selected plane within silicon ingot near and parallel to at least §ne edge of the ingot generating an original line of modified structure. Addit o al fast scanning with low numerical aperture lens focusing the scanning laser radiation beam could begin from this, originally focused by large numerical aperture lens laser illumination line and continue in perpendicular direction to the oppof i e edge of the ingot, extending the modified silicon structure created by the large numerical aperture optics along the scanning direction. Repeating such low numqjejcal aperture line scanning with a shift along the original line with modified silicqrj structure it is possible to cover the whole wafer area by lines with modified silicQjti structure. In one example, the lines with modified silicon structure cover only {partially the wafer area.

[0033] |G. 5 is a schematic illustration of an example of laser radiation scanning employing a low numerical aperture lens. The numerical aperture of the lens could be low _^ er than 0.5. Scanning system 500 includes a polygon prism 504 configured to rotate^ around its axis 508. The facets of the polygon prism 504 could have a refiecjjve coating. The facets 504 of the rotating around axis 508 polygon prism deflect a laser radiation beam 512 emitted by an IR laser 516 to different angles. Sho R is a central laser radiation beam 520 and marginally deflected laser radiation beamg 524 and 528. Scanning lens 532 could be an F-Theta type lens and configured to focus laser radiation beams 520a - 528a inside silicon ingot 312 in a desired wafer separation plane 308.

[0034] FIG- 5 has illustrated a one dimensional laser radiation scanning arrangement, with the laser radiation beams 520a - 528a scanned only along one axis, for example, denoted by X. Complete or partial coverage of the ingot surface by the laser Radiation beam could be facilitated provide additional scanning axis, for example, along axis Y perpendicular to axis X. Different methods such as an additional polygon mirror, acousto-optical deflector, mechanical movement of the lenseg or of the ingot could be employed for scanning along the other (Y) axis.

[0035] It should be noted that the laser radiation beam could be shaped into a symnjftric or asymmetric focused spot, e.g. as an ellipse. A longer ellipse axis ρεφς ^ίου^ to the scanning axis could provide in some examples a better scanned surfagf coverage. Additionally, the scanning speed could be increased when the longer axis of the scanning spot is parallel to the scanning axis.

[0036] ^hen a low numerical aperture lens is used, the focused segment of the laser radia fpn beam, so called waist, could extend for few hundred microns or even up to few millimeters. Scanning with low numerical aperture laser beam can be applied simu taneously to a number of different separation planes, provided that these separation planes include islands with modified material (silicon) structure, which were cheated in a previous stage of laser scanning with a large numerical aperture laser adiation beam focusing lens. This is illustrated in FIG. 6, where different planes 604 and 608 are planes with modified silicon material structure islands. Planejj 604 and 608 could be scanned simultaneously by a laser radiation beam 512 (524-5,28) or sequentially where each next plane is scanned after completion of scanrjhig of the previous plane. Generally, a plurality of planes distanced from each other by the thickness of wafers could be scanned simultaneously by a laser radiatipn beam 512. Shading of the islands located in the next (lower) planes by the islands located in an upper plane, could be prevented by locating the islands in different locations (for example, in a staggered pattern) in each of the planes and along he scanning path of the laser radiation beam focused by a low numerical aperture lens. It should be noted that in all cases of scanning with low numerical aperture each of the subsequent laser radiation pulses could overlap the spot or easy movement of the ingot being processed between the stations of the system accorjiiiig to the processing steps.

[0040] fhe incident laser radiation 728 could be an infrared radiation supplied by a laser f 32 emitting infrared radiation with wavelength of 1.1 micron or longer than 1.1 micron. This laser radiation could be directed simultaneously to the both laser scanning stations 708 and 712. For example, a beam splitting mirror 736, could direct _, the laser beam 740 to a large numerical aperture laser scanner 708 and laser beam 744 to a low numerical aperture laser scanner 712.

[0041] I rge numerical aperture laser radiation scanning system 708, for example, with numerical aperture greater than 0.6, is configured to direct the incident laser radiation 740 towards the ingot, schematically shown as a hatched square 710, and scans the laser beam across front surface of the ingot. Laser radiation focusing lens 328 (FIG. 3 and 4) of the large numerical aperture scanner is configured to focus the incident laser radiation scanned across front surface 332 of the ingot 312 in a plane §08 coplanar to the plane of front surface 332, and located beneath the front surfaej 332 of the ingot 312 as it is schematically shown in Figures 3 through 4. The injensity of the incident laser radiation scanned across the front surface of the ingot and focused into a spot 316 in the separation plane 308 (FIG. 3 and FIG. 4) is selec §d to be sufficient to cause the ingot material structural changes, e.g. in the range, of 0.1 to 10 J/cm2. In one example, laser radiation scanning system, could be configured to direct the incident laser radiation 740 and scan it over a matrix or a lineaf array of a large numerical aperture lenses as it is shown on Fig. 3. Large numqrfcal aperture lenses 328 could be located in a close proximity to front surface 332 of ingot 312 and configured to focus incident laser radiation underneath front surfagj 332 of ingot 312.

[0042] F§cused laser radiation spot could be of a symmetric shape, for example, round spot o| be of asymmetric shape, for example, elliptical or oval with the longer axis coinciding with the laser radiation scanning direction or any other direction that could be desired or required for proper operation of apparatus 700.

[0043] fhe distance of focused laser radiation spot from front surface 332 of ingot 312 coukj be varied either physically displacing the focusing lens or ingot holder with ingot 112 to regulate the distance under front surface of ingot 312 and plane 308 (FIG, ) where the focused spot of the laser radiation is located.

[0044] fanning plane 308 (FIG. 3 and FIG. 4) with a large numerical aperture laser scanning system and proper laser radiation intensity could result in formation in the separation plane 308 of a number of spots or islands of material with modified structure. The distance 128 (FIG. 1 and FIG. 2) of focused laser radiation spot from front surface of ingot 312 could actually define the thickness of the wafer to be separated from the ingot.

[0045] Large numerical aperture laser radiation scanning system 708 is configured to direcf Jjie incident laser radiation 722 and scan it across front surface 332 of ingot 312 t |ing optical or mechanical scanning devices. For example, a galvo-scanner coulcf be used to scan the laser radiation alternatively; a relative movement between focusjng lens or lenses and ingot 312 held by the ingot holder could be used to scan the lajer radiation. The scanning performed by each of the systems could be a two- dimei ional scanning supporting laser radiation scanning across the surface of ingot 12 or a one dimensional scanning. For example, optical laser radiation scanrjjng could be implemented in one dimension and mechanical scanning e.g., movement of the ingot and ingot holder could be implemented in another dimension.

[0046] numerical aperture scanning station 712 is intended to scan fast separation plane 308 located within ingot 312 with a low numerical aperture laser radia |pn beam 520-528 (FIG. 5) (for example, with numerical aperture lower than 0.5) nd operating the low numerical aperture laser radiation scanning system to affecf |he earlier created by a large numerical aperture laser scanning system spots or islands with modified material structure, thus increasing the area of the modified materja 1 structure within separation plane 308. In order to ensure fast scanning such opto-mechanical scanning mechanisms as rotating polygon mirror (FIG. 5) or a resonant (galvo) mirror could be used. Focused laser radiation spot could be a of a symmetric shape, for example, a round spot or be of asymmetric shape, for example, elliptical or oval with the longer axis coinciding with the laser radiation scannjng direction or any other direction that could be desired or required for propef operation of apparatus 700. [0047] ]|§th optical and mechanical scanning devices are configured to scan and locat focused laser radiation spots such that the spots would form a number of contiguous lines spanning from one edge to the other edge of the ingot surface and covering fully or partially the separation layer surface.

[0048] Apparatus 700 further includes a wafer separation station 716 configured to separal© the wafer from ingot 708. The wafer separation from the ingot takes place along he separation plane 308 (FIG. 3),where the ingot material structural changes took flace. The mechanism for separating the wafer from ingot 312 could include a vacuTj chuck configured to be attached to front surface of the wafer and a mechajdsm configured to move the vacuum chuck relative to the ingot to separate or detach the wafer from the ingot. The relative movements of the vacuum chuck could, be one of a group of movements consisting of lifting, tilting and rotating the wafe with respect to the ingot. Mechanism 716 for separating the wafer from ingot could also perform one of a group of treatments consisting of peeling, cleavage, thermj -cycling, microwave heating, treatment by liquid nitrogen, or a combination of thereof.

[0049] The disclosed method of wafer from ingot separation increases the speed or the separajion process, reduces material spoilage and could support development of new thin wafer manufacturing technologies based on laser induced modified silicon structure.