Self-aligned fuse structure and method with heat sink

[0003] This application is related to co-pending U.S. patent application Ser. Nos. 09/118,602, 09/118,231 and 09/118,230 filed concurrently herewith, which are incorporated herein by reference for all purposes.

[0004] The present invention relates to fuse structures in semiconductor devices. More particularly, the present invention relates to self-aligned fuse structures having a heat sink, methods of making such fuses, and semiconductor devices incorporating such fuses.

[0005] As semiconductor devices, such as integrated circuit chips, continue to decrease in size and increase in complexity, the likelihood of a defective chip resulting from a failed element or a defective conductor increases. One way to reduce the number of chips which must be scrapped due to fabrication defects is to manufacture fuses into semiconductor devices. Fuses may be opened to isolate defective areas and allow the rest of the circuit on a chip to be used. Fuses may also be used to trim a circuit, enable a particular mode, or enable or disable different segments of a circuit.

[0006] One example of the use of fuses in semiconductor devices is the provision of redundant elements on integrated circuits. If a primary element is defective a redundant element can be substituted for that defective element, rather than scrapping the chip. An example of a semiconductor device which uses redundant elements is electronic memory. Typical memory chips are composed of millions of equivalent memory cells arranged in addressable rows and columns. Semiconductor memory chips are typically fabricated with built-in redundancy in case one or more bits are found to be bad after fabrication. If a bad bit is found in testing following fabrication, a fuse may be blown to “program” a redundant cell to respond to the address of the defective primary cell. The use of the redundant rows and columns in memory cells increases economic efficiency by substantially raising yields. That is, an entire chip need not be rejected in the event that only one or two bits of memory are bad since those bad bits may be replaced by the redundancy built into the chip.

[0007] Conventional DRAM and SRAM stand-alone memory cell arrays are typically composed of intersecting rows and columns of polysilicon, metal or silicides. A schematic layout of a typical 64K DRAM cell 100, illustrating primary rows 102 and columns 104 of polysilicon, is shown in FIG. 1. These primary rows and columns 102 and 104 are supplemented by redundant rows 106 and columns 108. Rows and columns are typically selected by using row and column decoders 110 and 112, respectively. In the event of the failure of a particular row(s) or column(s), a fuse may be blown to disable the bad bits and enable redundant bits, as described above. Various configurations for incorporating fuses into semiconductor devices to implement redundant elements, such as in electronic memory, and for other applications are well known in the art. Examples are described in U.S. Pat. No. 5,636,172 to Prall et al., and R. T. Smith et al., Laser Programmable Redundancy and Yield Improvement in a64K DRAM, IEEE Journal of Solid-State Circuits, vol. SC-16, No. 5 (1981), the disclosures of which are incorporated by reference herein for all purposes.

[0008] Conventional semiconductor fuses may also be composed of metals or metal alloys, for example, aluminum-copper (AlCu). A typical process for blowing conventional polysilicon fuses involves the use of lasers in conjunction with a fuse structure to program the faulty element(s) out of operation and the redundant element(s) into operation. The conventional approach in using a laser to blow such fuses is to simply irradiate the fuse with a laser beam whose energy density and pulse duration are sufficient to vaporize the fuse material, thereby severing the electrical link.

[0009] A recent area in semiconductor device development is embedded memory, which is a combination of memory and logic on the same chip formed during the same fabrication process. For example, an embedded memory cell may include a single silicon layer processed to form transistor devices for both logic and memory functions covered by a series of metal layers which provide routing and interconnects to the silicon layer. Embedded memory cell arrays typically have, for example, two to six metal layers. Since the silicon logic and memory layer is overlaid by metal layers in an embedded memory cell device, it is preferred to use fuses in an upper metal layer to disable bad bits or enable redundant bits built into an embedded memory array. Redundancy is built into embedded memory arrays for the same reason as stand-alone memory, that is, to increase economic efficiency by increasing yield. The conventional process for blowing polysilicon fuses may also be applied to metal fuses. However, since metal does not absorb laser energy as efficiently as polysilicon, the process risks increasing the thermal budget for the overall fabrication process of the device, which may impair its electrical performance.

[0010] FIG. 2 is a composite figure showing a conventional semiconductor fuse technology. A cross-sectional side view of the structure of an intact conventional fuse 202 is shown at the center of the figure. A fuse material layer 204, a conductive material, typically composed of polysilicon or a metal or metal alloy, for example aluminum (Al) or aluminum-copper (AlCu), is covered by a blanket of dielectric capping material 206, typically composed of silicon dioxide (SiO₂), of substantially uniform thickness. A typical metal fuse material is about 0.5 to 1.5 μm in width and about 3000 to 5000 Å thick. For a metal fuse width of 1 μm, a recommended pitch is about 2.25 μm with a minimum length of 3.5 μm for a laser with a 2.5 μm beam diameter. As noted above, such a fuse is blown by a laser when the laser's beam incident on the fuse 202 has sufficient energy density and pulse duration to vaporize a portion of the fuse material 204 and sever the electrical connection of which the fuse is part (“blow the fuse”).

[0011] The wave form 210 at the top of FIG. 2 illustrates the energy distribution of a laser used to blow the fuse 202. A typical laser system used for this purpose is the Model 9300 High-Performance System available from Electro Scientific Industries (ESI). This laser is a diode pumped YLF with a wavelength of 1.047 μm or 1.3 μm, a pulse duration of 4-100 ns, and a maximum energy per pulse of 6 μJ. It has an adjustable beam diameter from about 2.5 to 9 μm. Alternatively, a Q-switched Nd-doped YAG laser with about a 4 μJ/pulse energy may be used.

[0012] A top view of the structure of a blown conventional fuse 202 is shown at the bottom of the figure. The laser energy distribution is represented by the larger of the two concentric circles 212. The smaller circle 214 represents the portion of the incident laser energy having sufficient energy density and pulse duration to blow the fuse.

[0013] In designing a laser-blown fuse process, one needs to consider several important topics. First, the width of the fuse, and therefore fuse-to-fuse spacing (i.e., fuse pitch), is limited by the spot (diameter) of the laser beam. Smaller spot sizes allow increased circuit density. As noted above, conventional lasers typically output beams with a minimum diameter of about 2.5 μm. Second, the fuse-blowing process should not affect the operation of the surrounding circuitry. In this respect, fuse materials with lower melting/boiling temperatures, and/or higher absorption coefficients are preferred to minimize thermal exposure. In addition, laser alignment is critical the ensure that the laser beam irradiates the fuse so that it will be blown, and not the surrounding circuitry, to which it could do thermal damage while also possibly not blowing the fuse as desired. Finally, a high-throughput fuse-blowing process is desired. These three demands can be conflicting in designing a fuse-blowing process. For example, although smaller laser-beam sizes reduce fuse pitch, they require more critical alignment, which requires more time and therefore decreases throughput.

[0014] Accordingly, semiconductor fuse structures and processes that would allow increased fuse density while reducing thermal exposure and the criticality of laser alignment would be desirable.

[0015] To achieve the foregoing, the present invention provides a self-aligned semiconductor fuse structure and method of making such a fuse structure. The fuse break point, that point at which the electrical link of which the fuse is part is severed by a laser beam, is self-aligned by the use of photolithographically patterned dielectric and a heat sink material. The self-alignment allows the size and location of the break point to be more forgiving of the laser beam size and alignment. This has several advantages, including allowing photolithographic control and effective size reduction of the laser spot irradiating the fuse material and surrounding structure. This permits reduced fuse pitch, increasing density and the efficiency of use of chip area, and results in reduced thermal exposure, which causes less damage to chip. In addition, laser alignment is less critical and therefore less time-consuming, which increases throughput in fabrication.

[0016] The present invention exploits the characteristic of most dielectric materials that they are poor conductors of thermal energy. Thermal resistance increases with the thickness of the dielectric. Thus that heat is conducted more easily and thus quickly through a relatively thin portion of dielectric than it is through a relatively thick portion of dielectric. In alternative embodiments, the present invention also exploits the characteristic of a dielectric material that its reflectance changes with its thickness due to optical interference effects. In such embodiments, the self-alignment of the fuse break point is further facilitated by the use of photolithography and anti-reflective coatings.

[0017] In one aspect, the present invention provides a semiconductor device fuse including a heat sink layer, and a dielectric layer disposed on the heat sink layer. The dielectric layer has one region bounded by one or more second regions having a thickness substantially less than that of the first region. The fuse also includes a fuse material layer disposed on the dielectric layer.

[0018] In another embodiment of this aspect of the invention, the previously described fuse also includes a second dielectric material layer disposed on the fuse material layer. The second dielectric material layer has a first thickness over the fuse material over the first region of the first dielectric material layer, and a second thickness bounding the first thickness, so that reflectance of incident laser light from the dielectric over the first portion of the fuse material is substantially less than that from the dielectric material bounding the first thickness.

[0019] In another aspect, the present invention provides a method of making a semiconductor device fuse. The method involves forming a dielectric material layer on at least a portion of a heat sink material, and depositing and patterning a photoresist mask on at least a portion of the dielectric material, so that a first region of the dielectric material overlying a first portion of the heat sink is masked. A portion of unmasked dielectric from the dielectric material layer is then removed, following which the photoresist mask from the first region of the dielectric material overlying the first portion of the heat sink is removed. Then, a fuse material layer is formed over the patterned dielectric material layer.

[0020] In another embodiment of this aspect of the invention, the previously described process further involves forming a second dielectric material layer disposed on the fuse material layer, and planarizing the second dielectric material layer, so that reflectance of incident laser light from a first thickness of the second dielectric overlying the fuse material and the first region of the first dielectric material is substantially less than that from a second thickness of the second dielectric material bounding the first thickness.

[0021] In other aspects, the invention provides semiconductor memories which incorporate fuse structures in accordance with the present invention.

[0022] These and other features and advantages of the present invention are described below with reference to the drawings.

[0023] FIG. 1 depicts a top schematic view of a conventional DRAM array, illustrating row and column redundancy.

[0024] FIG. 2 depicts a composite figure showing a conventional semiconductor fuse technology.

[0025] FIG. 3 is a graph of reflectance versus dielectric thickness which illustrates the optical interference relationship between the thickness of a dielectric, and the amount of light (energy) reflected by a dielectric-coated material.

[0026] FIG. 4 is a composite figure showing a semiconductor fuse technology in accordance with a preferred embodiment of the present invention.

[0027] FIGS. 5A through 5G illustrate the fabrication of a fuse structure in accordance with a preferred embodiment of the present invention.

[0028] FIGS. 6A and 6B illustrate fuse pitch reduction in accordance with a preferred embodiment of the present invention.

[0029] Reference will now be made in detail to preferred embodiments of the invention. Examples of the preferred embodiments are illustrated in the accompanying drawings. While the invention will be described in conjunction with these preferred embodiments, it will be understood that it is not intended to limit the invention to such preferred embodiments. On the contrary, it is intended to cover alternatives, modifications, and equivalents as may be included within the spirit and scope of the invention as defined by the appended claims. In the following description, numerous specific details are set forth in order to provide a thorough understanding of the present invention. The present invention may be practiced without some or all of these specific details. In other instances, well known process operations have not been described in detail in order not to unnecessarily obscure the present invention.

[0030] The present invention provides a self-aligned semiconductor fuse structure and method of making such a fuse structure. The fuse break point, that point at which the electrical link of which the fuse is part is severed by a laser beam, is self-aligned by the use of photolithographically patterned dielectric and a heat sink material. The self-alignment allows the size and location of the break point to be more forgiving of the laser beam size and alignment. This has several advantages, including allowing photolithographic control and effective size reduction of the laser spot irradiating the fuse material and surrounding structure. This permits reduced fuse pitch, increasing density and the efficiency of use of chip area, and results in reduced thermal exposure, which causes less damage to chip. In addition, laser alignment is less critical and therefore less time-consuming, which increases throughput in fabrication.

[0031] The present invention exploits the characteristic of most dielectric materials that they are poor conductors of thermal energy. Thermal resistance increases with the thickness of the dielectric. Thus that heat is conducted more easily and thus quickly through a relatively thin portion of dielectric than it is through a relatively thick portion of dielectric.

[0032] In alternative embodiments, the present invention exploits the characteristic of a dielectric material that its reflectance changes with its thickness due to optical interference effects. This constructive and destructive interference varies with the thickness of a dielectric according to well understood principles of optics. FIG. 3 is a graph of reflectance versus dielectric thickness which illustrates this relationship. As the dielectric thickness increases (from left to right on the x-axis of the graph 300), the reflectance varies regularly. The specific value of reflectance corresponding to maximum and minimum reflectances are in part determined by the reflective character and thickness the dielectric as well as the underlying material. Also, the periodic reflectance variation with thickness is characteristic of dielectric films that are non-absorbing. For absorbing dielectric materials, the reflectance and transmission intensities may be non-periodic with dielectric thickness. In this example, reflectance is maximal when the underlying material is not covered by dielectric and when the dielectric thickness is about one half the wavelength of the incident laser light, and minimal where the dielectric thickness is about one fourth the wavelength of the incident laser light, and so on with regular periodicity.

[0033] According to a preferred embodiment of the present invention, instead of using a blanket dielectric capping layer over the fuse material, as in the conventional approach, the fuse material is deposited over a dielectric layer which has been patterned photolithographically so that its thickness is greater in a region underlying that portion of fuse material desired to be blown. Thus, the portion of the fuse material desired to be blown is raised relative to the remainder of the fuse material. The dielectric layer underlying the fuse material is formed on a heat sink material, such as a metal (for example, a dummy metal layer or a metal layer that plays a role in another aspect of the device) or the semiconductor substrate material for the device. Upon exposure to a laser pulse, the fuse material transfers its heat down to the heat sink. The area of the fuse material that is closer to the heat sink is able to transfer its heat more efficiently to the heat sink, because it has less thermal resistance between it and the heat sink. Thus, energy is dissipated from the fuse material overlying the thicker dielectric area more slowly. This area therefore retains more energy from the laser, and thus blows before fuse material elsewhere.

[0034] In a particularly preferred alternative embodiment of the present invention, an additional layer of dielectric may also be formed over the fuse material and then planarized. Since the portion of the fuse material desired to be blown is raised due to the underlying dual-thickness dielectric, the overlying dielectric also has a dual-thickness following planarization, thinner over the portion of the fuse material to be blown and thicker elsewhere. The dual-thicknesses of the dielectric overlying the fuse material may be set by coordination with the patterning of the dielectric underlying the fuse material and control of the planarization process so that the thickness over the portion of the fuse material desired to be blown renders that portion of the dielectric anti-reflective and the thickness elsewhere is relatively reflective. Such a structure reinforces the concentration of energy in the area of the fuse material to be blown already provided by the underlying dielectric and heat sink since it not only causes heat to build up faster in the (photolithographically determined) fuse region to be blown, but also causes more laser energy to be absorbed in this area. These two independent methods for raising the heat in the fuse area relative to outside the fuse area combine to blow the fuse at a lower laser energy than could be achieved using either method alone. This configuration creates an even larger processing window for blowing fuses.

[0035] In a preferred embodiment, the fabrication is optimized so that the dielectric's thickness overlying the portion of the fuse material to be blown minimizes reflectance (maximizes transmission through the dielectric and absorption by the underlying fuse material), and the thickness of the dielectric over other portions of the fuse material maximizes reflectance (minimizes absorption by the underlying fuse material).

[0036] There are several benefits of the present invention. First, since the raised portion of the fuse material blows before the remainder, the effective laser spot size, that is, the diameter of the beam of incident laser energy of sufficient energy density and pulse duration to blow the fuse is reduced through the use of photolithographically patterned dielectric materials underlying, and also possibly overlying, the fuse material. This reduction allows a decrease in fuse pitch, which in turn allows for increased fuse density and more efficient use of chip area.

[0037] Also, the laser alignment step is less critical. While the laser equipment must gross align to the fuse area, fine alignment of the laser beam to the fuse area is facilitated by the lithography process used to form the underlying and optionally overlying dielectric layers, which is inherently more accurate. Throughput may therefore be increased by reducing the time needed for the laser system to do fine alignment. Moreover, in an embodiment which includes an overlying dual-thickness dielectric, even if the center of the laser beam is slightly misaligned to the raised fuse material area, the laser fluence can be increased (to compensate for the decreased energy exposure due to the Gaussian intensity profile for the laser—see FIG. 4) without harming the region outside of the lithographically defined fuse region, since this “auto aligned” fuse region absorbs more energy from the laser than elsewhere.

[0038] Another advantage of embodiments including an overlying dual-thickness dielectric is that the wafer's thermal exposure is reduced because of the increased reflectance outside of the patterned area. This is an important difference from the conventional approach, as the conventional approach uses a blanket dielectric coating, and therefore a larger area than that which is required to be blown to sever the electrical link (which is substantially smaller than the minimum available laser beam spot size) receives substantially equal exposure to the laser beam. In accordance with a preferred embodiment of the present invention, only the fuse material below the patterned area receives the same fluence as would have been received by the area of the fuse and surrounding area within the diameter of the beam of incident laser energy of sufficient energy density and pulse duration to blow the fuse, using the conventional method. In this way, the photolithographically-defined area “auto-aligns” the laser beam to the fuse area desired to be blown.

[0039] FIG. 4 is a composite figure showing a semiconductor fuse technology in accordance with a preferred embodiment of the present invention. A cross-sectional side view of the structure of an intact fuse 402 in accordance with this preferred embodiment of the present invention is shown at the center of the figure. The structure includes a fuse material layer 404, a conductive material, composed, for example, of polysilicon or a metal or metal alloy, such as aluminum (Al), aluminum-copper (AlCu), tungsten (W), copper (Cu), or a silicide.

[0040] Underlying the fuse material 404 is a dual-thickness layer of a dielectric 406 on a layer of heat sink material 408. The underlying dielectric 406 may be any material possessing the general properties of low thermal and electrical conductivity as described above, for example, silicon dioxide (SiO₂), silicon nitride (Si₃N₄), calcium fluoride (CaF₂), silicon oxynitrides (SiO_xN_y(H_z)), or organic polymers such as polyimide. In accordance with a preferred embodiment of the present invention, the dual thickness of the underlying dielectric layer 406 is achieved by photolithographic patterning. The underlying dielectric 406 is thicker in the region 407 underlying that portion 405 of the fuse material desired to be severed to blow the fuse (patterned dielectric area of width w₁), and thinner in the bounding region. The heat sink material 408 may be, for example, a dummy metal layer or other metal layer that plays a role in another aspect of the device, or the semiconductor substrate material for the device. Upon exposure to a laser pulse, the fuse material 404 transfers its heat down to the heat sink 408. The area of the fuse material 404 that is closer to the heat sink 408 is able to transfer its heat more efficiently to the heat sink 408, because it has less thermal resistance between it and the heat sink. Thus, energy is dissipated from the portion 405 of the fuse material 404 overlying the thicker dielectric area 407 more slowly. This portion 405 (the raised portion) of the fuse material 404 therefore retains more energy from the laser, and thus blows before fuse material elsewhere.

[0041] As described above, in a particularly preferred alternative embodiment of the present invention, an additional layer of dielectric 409 may also be formed over the fuse material 404 and then planarized. The overlying dielectric 409 may be any material possessing the general optical properties described above, for example, silicon dioxide (SiO₂), silicon nitride (Si₃N₄), calcium fluoride (CaF₂), silicon oxynitrides (SiO_xN_y(H_z)), or organic polymers such as polyimide. In accordance with a preferred embodiment of the present invention, the dual thickness of the underlying dielectric layer 406 is achieved by photolithographic patterning. Since the portion 405 of the fuse material 404 desired to be blown is raised due to the underlying dual-thickness dielectric 406, the overlying dielectric 409 also has a dual-thickness following planarization; thinner over the portion 405 of the fuse material 404 to be blown and thicker elsewhere. The dual-thicknesses of the dielectric 409 overlying the fuse material 404 may be set by coordination with the patterning of the dielectric 406 underlying the fuse material and control of the planarization process so that the thickness over the portion 405 of the fuse material 404 desired to be blown renders that portion of the dielectric 409 anti-reflective and the thickness elsewhere is relatively reflective.

[0042] The thickness of the overlying dielectric layer inside t₁and outside t₂of the patterned area w₁is such that less light is reflected inside the patterned area relative to the outside the patterned area. In a preferred embodiment, the patterning is optimized so that the dielectric's thickness outside the patterned area t₂minimizes absorption within the underlying fuse material 404, and the dielectric's thickness inside the patterned area t₁maximizes absorption in the fuse material. Thus, the raised portion 405 of the fuse material 404 under the patterned area absorbs more energy from the laser, and therefore blows sooner, than the fuse material elsewhere. Of course, less optimized configurations of the dielectric thicknesses may also be used in accordance with the present invention.

[0043] The photolithographic technique of the present invention allows the width and pitch of semiconductor fuses to be reduced, allowing for increased fuse density and more efficient use of chip area and materials. The width w₂of a fuse material layer 404 in accordance with a preferred embodiment of the present invention may have its lower limit determined by the current carrying capacity (density) of the material as specified by the application. In one preferred embodiment, for a 0.25 μm technology, the fuse material width w₂may be between about 0.25 and 1.5 μm, more preferably between about 0.4 and 1.0 μm. The thickness, t₃, of a fuse material 404 in accordance with a preferred embodiment of the present invention may be about 2000 to 10,000 Å, similar to that of conventional metal fuse materials, and preferably between about 3000 and 5000 Å.

[0044] The wave form 410 at the top of FIG. 4 illustrates the energy distribution of a laser used to blow the fuse 402. In accordance with a preferred embodiment of the present invention, a conventional laser system, such as the ESI Model 9300 High-Performance system described above may be used. Alternatively, a Q-switched Nd-doped YAG laser with about a 4 μJ/pulse energy may be used. Of course, the present invention is not limited to a particular laser system and may be adapted to any system in accordance with the principles and details described herein by one of ordinary skill in the art.

[0045] A top view of the structure of a blown fuse 402 in accordance with a preferred embodiment of the present invention is shown at the bottom of the figure. The portion of the incident laser energy of sufficient energy density and pulse duration to blow the fuse is represented by the larger of the two concentric circles 412. The smaller circle 414 represents the portion of the fuse material which receives laser energy of sufficient energy density and pulse duration to blow the fuse 402, as defined by the photolithographic patterning of the underlying dielectric 406, and optionally additionally the overlying dielectric 409.

[0046] The fuse structure 402 may be fabricated by applying photolithograhic semiconductor processing procedures. FIGS. 5A through 5G illustrate the fabrication of a fuse structure in accordance with a preferred embodiment of the present invention. Referring to FIG. 5A, a blanket layer of dielectric 504, such as SiO₂, may be deposited on a heat sink material 502, such as a dummy metal layer or other metal layer, or the semiconductor substrate material, by methods known in the art, such as CVD or sputtering. The thickness of the deposited dielectric 504 should be selected so that it substantially retards thermal energy moving through the dielectric to the heat sink 502. For example, the dielectric 504 may be in the range of about 10 to 1000 Å thick.

[0047] Then, as shown in FIG. 5B, a photolithographic process is used to generate a patterned resist mask 506 on the surface of the dielectric 504. For example, a photoresist mask may be formed on the surface of the dielectric 504 and patterned to mask a region of the dielectric which is to underlie that portion of the subsequently formed fuse material where it is desired that the fuse be severed. The unmasked region 507 of the dielectric 504 corresponds to that portion which will have a reduced thickness in the final fuse structure. Once the mask 506 is in place, an etch is used to remove the a portion of the thickness of the dielectric 504 in the unmasked region 507, as illustrated in FIG. 5C. This removal may be accomplished by a conventional etch technique, such as reactive ion etching (RIE) using fluorocarbon chemistry, such as CF₄or CHF₃, or buffer oxide etching (BOE) using HF (100:1), for example. The amount of material removed may be controlled by etch time in accordance with known etch rates for the dielectric material for the etch technique used.

[0048] FIG. 5D illustrates the partially-formed structure following removal of the resist mask 506. The heat sink 502 is covered with a dual-thickness dielectric 504. As shown in FIG. 5E, a fuse material 508 is formed on the dual-thickness dielectric 504. The fuse material may be formed according to procedures well known in the art, such as CVD or sputtering.

[0049] The structure illustrated in FIG. 5E represents one preferred embodiment of the present invention. The greater thickness of the dielectric 504 underlying a portion 509 of the fuse material 508 results in greater retention of heat energy in that (raised) portion of the fuse material when the fuse is irradiated with a laser beam. Thus the raised portion 509 of the fuse blows before the remaining fuse material.

[0050] In an alternative preferred embodiment, the fuse material 508 is further coated with a second layer of dielectric 510, as shown in FIG. 5F. The thickness t₂of the second (overlying) dielectric 510 should preferably be selected so that it reflects a substantial portion of incident laser light in a fuse blowing process, preferably the maximum amount possible for the dielectric material. The overlying dielectric layer 510 is then planarized. The planarization may be achieved according to procedures known to those of skill in the art, such as chemical mechanical polishing (CMP), reflow, or resist etch back, using known operating parameters.

[0051] FIG. 5G illustrates the fuse structure of this alternative preferred embodiment. As described above, since the portion of the fuse material desired to be blown is raised due to the underlying dual-thickness dielectric, the overlying dielectric also has a dual-thickness following planarization, thinner over the portion of the fuse material to be blown and thicker elsewhere. The dual-thicknesses of overlying dielectric 510 may be set by coordination with the patterning of the dielectric underlying the fuse material and control of the planarization process so that the thickness t₁over the portion of the fuse material desired to be blown renders that portion of the dielectric anti-reflective and the thickness t₂elsewhere is relatively reflective. In a preferred embodiment, the fabrication is optimized so that the thickness t₁of the overlying dielectric 510 in the region over the portion of the fuse material 508 to be blown minimizes reflectance (maximizes transmission through the dielectric and absorption by the underlying fuse material), and the thickness t₂of the dielectric 510 over other portions of the fuse material 508 maximizes reflectance (minimizes absorption by the underlying fuse material).

[0052] As noted above, fuse pitch may be reduced by the reduction of the effective laser beam diameter in accordance with the present invention. FIGS. 6A and 6B illustrate fuse pitch reduction according to one preferred embodiment of the present invention. A conventional fuse configuration is illustrated in FIG. 6A. In this configuration, the fuse pitch, p₁, is determined by the diameter of the laser beam used to blow the fuses, since, as described above, the laser beam (represented by the broken-line circles) should not irradiate fuses adjacent to the fuse being blown. FIG. 6B illustrates a fuse configuration in accordance with a preferred embodiment of the present invention. Since the present invention provides a window to auto-align the laser to the fuse area to be blown, the laser beam size is effectively reduced, and irradiation of fuses adjacent to that being blown is acceptable. In the preferred embodiment illustrated in FIG. 6B, windows 610, 620 and 630 of adjacent fuses 611, 621 and 631 are staggered so that the laser beam irradiating a fuse to be blown through its window may also irradiate an adjacent fuse or fuses, but not the adjacent fuse or fuses window(s). In this way, the fuses may be positioned closer together and fuse pitch, p₂, in accordance with this preferred embodiment of the present invention is less than that of the conventional configuration (p₂<p₁).

[0053] The following examples provide details concerning preferred semiconductor fuse structures in accordance with the present invention. With an understanding of the present invention, those of skill in the art may determine appropriate dielectric thicknesses, where necessary for optical effects, in accordance with the present invention by applying the principles of electromagnetic theory, for example, such as elucidated by Airy's formula. The following examples provide preferred configurations of selected fuse and dielectric materials. Of course, other fuse and dielectric materials may be substituted and the appropriate dielectric thicknesses calculated in the same manner. It should be understood the following is representative only, and that the invention is not limited by the details set forth in these examples.

[0054] The fuse material may be composed of aluminum (Al) about 4000 Å wide (w₂) and about 5000 Å thick (w₃). The underlying dielectric may be composed of silicon dioxide (SiO₂) and have a thickness of about 1000 Å under the portion of the fuse material desired to be blown, and a thickness bounding that of about 200 Å. The heat sink may be composed of aluminum and be dummy metal. The overlying dielectric may be composed of silicon dioxide (SiO₂) and have a thickness (t₁) over the portion of the fuse material desired to be blown of about 1558 Å, a thickness (t₂) bounding that of about 3266 Å, and the laser beam used the blow the fuse may have a wavelength of about 1 μm with normal incidence. With these parameters, about 83% of the incident laser light is reflected from the dielectric of thickness t₁, and about 92% of the incident laser light is reflected from the dielectric of thickness t₂.

[0055] The fuse material may be composed of aluminum (Al) about 4000 Å wide (w₂) and about 5000 Å thick (w₃). The underlying dielectric may be composed of silicon nitride (Si₃N₄) and have a thickness of about 1000 Å under the portion of the fuse material desired to be blown, and a thickness bounding that of about 200 Å. The heat sink may be composed of aluminum and be dummy metal. The overlying dielectric may be composed of silicon nitride (Si₃N₄) and have a thickness (t₁) over the portion of the fuse material desired to be blown of about 1055 Å, a thickness (t₂) bounding that of about 2261 Å, and the laser beam used the blow the fuse may have a wavelength of about 1 μm with normal incidence. With these parameters, about 70% of the incident laser light is reflected from the dielectric of thickness t₁, and about 92% of the incident laser light is reflected from the dielectric of thickness t₂.

[0056] The present invention is particularly advantageous in its application to embedded memories and integrated circuits, such as application specific integrated circuits (ASICs) which incorporate embedded memories, but may also be applied to increase densities by allowing narrower fuse pitch in stand-alone memories (e.g., SRAM, DRAM, ROM, EPROM, EEPROM, FLASH). Fuse structures in accordance with the present invention may be implemented in such semiconductor devices by replacing conventional fuse structures using the processing procedures described above in place of conventional fuse processing.

[0057] Although the foregoing invention has been described in some detail for purposes of clarity of understanding, it will be apparent that certain changes and modifications may be practiced within the scope of the appended claims. For example, the invention is not limited to the enablement of redundant memory in stand-alone and embedded electronic memories, but may also be implemented anywhere laser blown fuses are used in semiconductor devices. Accordingly, the present embodiments are to be considered as illustrative and not restrictive, and the invention is not to be limited to the details given herein, but may be modified within the scope and equivalents of the appended claims.