The present invention relates to cooling of semiconductor devices.
Mobile semiconductors in the recent years have experienced a huge growth in both processing power and heat generation. Currently, cooling solutions for mobile semiconductors are in their infancy. While there are many robust semiconductor cooling solutions on the market, they are not designed for mobile devices and are very much outdated. The current cooling solutions can be summed up as too heavy, too large, not efficient, wasteful of material and in some cases consume too much power. Therefore, a miniature, light weight, low profile, efficient, passive and high performance modern mobile heat sink configuration is needed.
It is known in the art to mount a heat sink over a semiconductor package. See for example U.S. Pat. No. 8,564,114. After the semiconductor chip is packaged, a large metallic heat sink is attached over the semiconductor chip. The heat sink is generally constructed using copper and/or aluminum and incorporates arrays of heat exchangers. It is very typical to use some type of high K Thermal Interface Material (TIM) in-between the heat sink and the semiconductor chip. The TIM in general has a K value of around 9. However, it is also known to use solder as the TIM to increase the thermal conductivity. Solder has a K value of 50, and can be mixed with copper and/or silver to increase the K value up to around 80. This solution, however, is not conducive for mobile applications, where there is a need to minimize size, weight, inefficiency, material used, and power consumption in the heat dissipation solution.
The aforementioned problems and needs are addressed by a semiconductor device that includes a semiconductor chip having a first silicon substrate with opposing first and second surfaces, a semiconductor device formed at or in the first surface, a plurality of first contact pads formed at the first surface which are electrically coupled to the semiconductor device, a layer of thermal conductive material on the second surface, and a plurality of first vias formed partially through the layer of thermal conductive material.
Other objects and features of the present invention will become apparent by a review of the specification, claims and appended figures.
The present invention includes techniques and configurations for forming cooling features directly in the silicon semiconductor substrate on which the integrated circuits are formed, alone or in combination with a silicon heat sink.
A passivation layer 18 is deposited on the bottom surface of the semiconductor substrate 12. This passivation layer acts as a diffusion barrier for the semiconductor chip 10. The passivation layer 18 can be deposited by a sputtering process that is well known in the art. A preferred thickness of the passivation layer 18 is 0.1 μm or less. A preferred passivation material would have high thermal conductivity, such as tungsten, nickel, chromium and alloys of aforementioned materials or any other appropriate passivation materials that are well known in the art. A layer of material 20 with high thermal conductivity is deposited over the passivation layer 18. The thermal conductive layer 20 enhances thermal conductivity across the semiconductor chip, thus reducing hot spots. Material such as copper, silver, graphene, carbon related materials or any other well-known thermally conductive materials can be used. Deposition can be performed by Physical Vapor Deposition (PVD) or any other appropriate process for the material of choice. A separate bonding layer 22 can be applied if the thermal conductive layer 20 cannot double as a bonding layer. Metal to metal bonding is preferred for its thermal conductivity characteristic. If copper is used, then a traditional copper to copper bonding through a thermal compression bonding process can be used. If silver or indium is used, then a silver indium room temperature weld process can be used. The resulting structure is shown in
Wafer level dicing/singulation of the semiconductor chip can be done at scribe lines between the individual semiconductor devices 14 and their associated bond pads 16 with mechanical blade dicing equipment, laser cutting or any other appropriate processes. The final die structure is shown in
The exposed bottom surface portions of substrate 12 are etched through the openings in the photoresist 24 to form vias 26 (i.e. holes, slots, trenches, channels, or any other form of cavities or cavity pattern) into the bottom surface of the substrate 12. A dry etch is preferred for the etching process. The opposing walls of each via can be perpendicular or tapered. The depth, width, shape and location of the vias 26 can be random or pseudo-random. The resulting structure (with perpendicular via sidewalls) is shown in
After the photo resist 24 is removed, passivation layer 18 is deposited on the back surface of the semiconductor substrate 12 as described above. Then, thermal conductive layer 20 is deposited over the passivation layer 18 as described above. The vias 26 are either filled or coated with the thermal conductive layer 20 material. The substrate 12 is then singulated as described above, with the final structure shown in
In any of the embodiments herein, an optional thermal emissivity enhancement layer 28 can be selectively deposited randomly or pseudo-randomly over the thermal conductive layer except for the bonding area. One example is shown in
A layer of photoresist 24 is deposited on the thermal conductive layer 20. Photoresist 24 is exposed and selectively etched using a photolithographic process to leave a pattern in the photoresist that selectively exposes the underlying thermal conductive layer 20. The pattern in the photoresist can be random or pseudo-random. The resulting structure is shown in
The exposed portions of the thermal conductive layer 20 are etched through the openings in the photoresist 24 to form vias 30 (i.e. holes, slots, trenches, channels, or any other form of cavities or cavity pattern) into the thermal conductive layer 20. A dry etch is preferred for the etching process. The via walls are preferably tapered for better air flow. The depth, width, shape and location of the vias can be random or pseudo-random. Preferably, but not necessarily, the vias 30 do not extend all the way through the thermal conductive layer 20 to expose the passivation layer 18 or substrate 12. As stated above, an optional thermal emissivity enhancement layer 28 can be selectively deposited on the thermal conductive layer 20 except for bonding areas. The final structure (after photoresist 24 is removed) is shown in
For effective heat dissipation (e.g., thermal conduction and emission), it is preferable (but not necessarily required) to mount any of the above described semiconductor chips on a silicon-based heat sink. Various embodiments of silicon heat sinks are described below, which can be used with any of the above described chip configurations, or with any other semiconductor chip structure needing effective heat dissipation.
The exposed silicon substrate surface is etched to form a cavity 46 and a plurality of deep trenches, channels, holes or other cavity type voids 48 into the top surface of substrate 42. A dry etch is preferred for the etching process. The depth, width, shape and location of the cavity 46 and trenches 48 can be random or pseudo-random. The trenches 48 define cooling fins 50 of the substrate 42. The walls of the trenches 48 are preferably tapered for better air flow. The resulting structure is shown in
After photo resist 44 is removed, a thermally conductive layer 52 is deposited on the top surface of the substrate 42, including inside cavity 46 and trenches 48. Optionally, this layer can be deposited on the bottom surface of the substrate 42 as well. The purpose of layer 52 is to enhance thermal conductivity across the heat sink, thus increasing heat dissipation capabilities. Materials such as copper, silver, graphene, carbon related materials or any other well-known thermal conductive material can be used. Deposition can be by Physical Vapor Deposition (PVD) or any other appropriate process for the material of choice. A bonding layer can be applied separately if the thermal conductive layer 52 cannot double as bonding layer. Metal to metal bonding is preferred for its thermal conductivity characteristic. If copper is used, then a traditional copper to copper bonding through thermal compression bonding process can be used. If silver or indium is used, then a silver indium room temperature weld process can be used. The resulting structure is shown in
A thermal emissivity enhancement layer as described above with respect to the semiconductor chip 10 can be optionally and selectively deposited randomly or pseudo-randomly anywhere on the heat sink 40 except for the bonding area. The thermal emissivity enhancement material can be deposited directly on the silicon substrate 42 or on the thermal conductive layer 52. The final heat sink surface can be optionally, selectively chemically polished to enhance thermal emissivity. Preferably, a plurality of heat sinks 40 are formed simultaneously on a single substrate 42, which then requires singulation of the substrate 42 along scribe lines to result in individual and separate heat sinks 40, as shown in
Either before or after singulation, a semiconductor device is mounted inside cavity 46, where heat is conducted from the device, through substrate 42 and thermal conductive layer 52, and off substrate 42 by cooling fins 50. As an example,
Photoresist 62 is deposited on the bottom surface of substrate 42 (including in vias 60). Photoresist 62 is exposed and etched via photolithography to remove select portions of the photoresist at the apex portions of those vias 60 which are disposed underneath cavity 46, leaving selected portions of the bottom surface of substrate 42 exposed, as shown in
A silicon etch is then performed on the bottom surface of substrate 42 to form through-holes 64 that extend through the substrate 42 to the cavity 46. Preferably, the silicon etch is a dry etch. The through holes 64 can have vertical or tapered sidewalls.
Other variations of heat sink 40 can be used which mix and match different features of the above described heat sinks. For example,
The silicon based chips 10 and heat sinks 40 have many advantages. Using silicon to create the heat sink allows access to advanced etching processes, which enables the creation of miniature structures at a very high density and therefore enabling more surface area per volume of material. Also, miniaturization of heat sink features allows utilization of spaces that were not accessible or have been left untapped in prior art designs. For example, the sides of the semiconductor chip are usually empty space or covered by molding material. Using silicon as a heat sink substrate allows for the shrinking of the heat sink to chip level, and utilizes all available spaces with very high efficiency. Using the silicon substrate for heat dissipation, which allows for utilizing advanced silicon etching technologies, makes available many shapes and designs that were previously too costly or impossible to make using traditional heat sink manufacturing technologies, which unlocks advanced air flow designs for semiconductor heat sinks.
Thermal Interface Material (TIM) has an average thermal K value of around 9, and 80 at most. Grease and tape based TIM can wear out quickly and can be considered as a thermal insulator in comparison to the present invention. Solder based TIM only preforms at one quarter of the performance of present invention, and it is much thicker. The present invention provides metal to metal bonding between the semiconductor chip 10 and the silicon heat sink 40 with target thermal K values of 350 and above. By using silicon based heat sinks with silicon based semiconductor chips, varying coefficients of thermal expansion (CTE) are no longer an issue, thus allowing the semiconductor chip to be bonded directly to the heat sink without TIM and resulting in a drastic increase in thermal dissipation rates.
Semiconductor chips often overheat in particular areas or points, thereby creating a bottle neck to the computation performance of the entire chip, the mobile device and its cooling system. The present invention includes a thermal conductivity enhancement layer on the semiconductor chip to quickly carry heat away from the hot spots and thus enable higher computation performance and heat emission rates. The thermal conductivity enhancement layer is also applied to the silicon heat sink for the same purpose.
While some materials do extremely well in thermal conduction, they are not always the best for thermal emissivity (radiation heat transfer). For example, copper and silver have some of the best thermal conductivity and thermal emissivity coefficients. However, these metals will oxidize quickly which drastically lowers the emissivity performance while their thermal conductivity remains unharmed. It is preferred to protect the thermal conductivity layer from oxidization while retaining a high thermal emissivity coefficient. Therefore, any of the above described embodiments having a thermal conductive layer can include a thermal emissivity enhancement layer (e.g. nickel or gold for emissivity) formed on the thermal conductive layer (e.g. copper or silver for thermal conductivity). Below is a listing of materials that are ideal for the thermal emissivity enhancement layer:
When creating patterns on silicon wafers (or any other material), warpage of the silicon substrate can occur and have negative impacts on the structure. Therefore, when forming patterns of vias or trenches into substrates 12 and 42 as described above, there are several techniques that can be employed to minimize warpage of the wafer before singulation. The first technique for minimizing wafer warpage is to discontinue any vias or trenches between chips or heat sinks that are eventually singulated from each other. More specifically, those areas of the wafer where scribe lines are or will be located should be free of the vias and trenches, as shown in
A second technique for minimizing wafer warpage is to orient the via/trench pattern 70 orthogonally relative to that of the adjacent row, as shown in
A third technique for minimizing wafer warpage is to orient the via/trench pattern 70 for each device 74 orthogonally relative to that in all neighboring devices 74, as shown in
It is to be understood that the present invention is not limited to the embodiment(s) described above and illustrated herein, but encompasses any and all variations falling within the scope of the appended claims. For example, references to the present invention herein are not intended to limit the scope of any claim or claim term, but instead merely make reference to one or more features that may be covered by one or more of the claims. Materials, processes and numerical examples described above are exemplary only, and should not be deemed to limit the claims. Further, as is apparent from the claims and specification, not all method steps need be performed in the exact order illustrated or claimed, but rather in any order that allows the proper formation of the semiconductor chip and/or heat sink of the present invention. Single layers of material could be formed as multiple layers of such or similar materials, and vice versa. Lastly, while cooling fins 50 are shown and described as straight and parallel to each other, other fin configurations can be used to achieve the desired thermal emission. For example,
It should be noted that, as used herein, the terms “over” and “on” both inclusively include “directly on” (no intermediate materials, elements or space disposed therebetween) and “indirectly on” (intermediate materials, elements or space disposed therebetween). Likewise, the term “adjacent” includes “directly adjacent” (no intermediate materials, elements or space disposed therebetween) and “indirectly adjacent” (intermediate materials, elements or space disposed there between), “mounted to” includes “directly mounted to” (no intermediate materials, elements or space disposed there between) and “indirectly mounted to” (intermediate materials, elements or spaced disposed there between), and “electrically coupled” includes “directly electrically coupled to” (no intermediate materials or elements there between that electrically connect the elements together) and “indirectly electrically coupled to” (intermediate materials or elements there between that electrically connect the elements together). For example, forming an element “over a substrate” can include forming the element directly on the substrate with no intermediate materials/elements therebetween, as well as forming the element indirectly on the substrate with one or more intermediate materials/elements therebetween.
This application claims the benefit of U.S. Provisional Application No. 61/983,402, filed Apr. 23, 2014, and which is incorporated herein by reference.
Number | Name | Date | Kind |
---|---|---|---|
5151777 | Akin et al. | Sep 1992 | A |
5403783 | Nakanishi et al. | Apr 1995 | A |
6627864 | Glenn | Sep 2003 | B1 |
6777767 | Badehi | Aug 2004 | B2 |
6972480 | Zilber et al. | Dec 2005 | B2 |
7033664 | Zilber et al. | Apr 2006 | B2 |
7157742 | Badehi | Jan 2007 | B2 |
7192796 | Zilber et al. | Mar 2007 | B2 |
7265440 | Zilber et al. | Sep 2007 | B2 |
7325593 | Kujirai | Feb 2008 | B2 |
7456901 | Jeong et al. | Nov 2008 | B2 |
7495341 | Zilber et al. | Feb 2009 | B2 |
7569409 | Lin et al. | Aug 2009 | B2 |
7589422 | Lee et al. | Sep 2009 | B2 |
7642629 | Zilber et al. | Jan 2010 | B2 |
7664390 | Cho et al. | Feb 2010 | B2 |
7859033 | Brady | Dec 2010 | B2 |
8164177 | Matz | Apr 2012 | B2 |
8183579 | Wang | May 2012 | B2 |
8432011 | Oganesian | Apr 2013 | B1 |
8520388 | Kurosawa | Aug 2013 | B2 |
8564114 | Lanzone | Oct 2013 | B1 |
20040104021 | Kujirai et al. | Jun 2004 | A1 |
20040251525 | Zilber | Dec 2004 | A1 |
20050051859 | Hoffman | Mar 2005 | A1 |
20050104179 | Zilber | May 2005 | A1 |
20050104186 | Yang | May 2005 | A1 |
20050205977 | Zilber | Sep 2005 | A1 |
20070138498 | Zilber | Jun 2007 | A1 |
20070190691 | Humpston | Aug 2007 | A1 |
20070190747 | Humpston | Aug 2007 | A1 |
20070205792 | Mouli | Sep 2007 | A1 |
20080012115 | Zilber | Jan 2008 | A1 |
20080017879 | Zilber | Jan 2008 | A1 |
20080083976 | Haba | Apr 2008 | A1 |
20080083977 | Haba | Apr 2008 | A1 |
20080099900 | Oganesian | May 2008 | A1 |
20080099907 | Oganesian | May 2008 | A1 |
20080116544 | Grinman | May 2008 | A1 |
20080116545 | Grinman | May 2008 | A1 |
20080150121 | Oganesian | Jun 2008 | A1 |
20080224249 | Nabe | Sep 2008 | A1 |
20080246136 | Haba | Oct 2008 | A1 |
20080265350 | Wu et al. | Oct 2008 | A1 |
20090115047 | Haba | May 2009 | A1 |
20090160065 | Haba | Jun 2009 | A1 |
20090212381 | Crisp | Aug 2009 | A1 |
20100053407 | Crisp | Mar 2010 | A1 |
20100164093 | Mowry | Jul 2010 | A1 |
20100225006 | Haba | Sep 2010 | A1 |
20100230148 | Kariya | Sep 2010 | A1 |
20100230812 | Oganesian | Sep 2010 | A1 |
20100237452 | Hagiwara et al. | Sep 2010 | A1 |
20100276701 | Hebert et al. | Nov 2010 | A1 |
20110012259 | Grinman | Jan 2011 | A1 |
20110031629 | Haba | Feb 2011 | A1 |
20110033979 | Haba | Feb 2011 | A1 |
20110049696 | Haba | Mar 2011 | A1 |
20110108940 | Huang et al. | May 2011 | A1 |
20110149537 | Kurosawa | Jun 2011 | A1 |
20110187007 | Haba | Aug 2011 | A1 |
20120018863 | Oganesian | Jan 2012 | A1 |
20120018868 | Oganesian | Jan 2012 | A1 |
20120018893 | Oganesian | Jan 2012 | A1 |
20120018894 | Oganesian | Jan 2012 | A1 |
20120018895 | Oganesian | Jan 2012 | A1 |
20120020026 | Oganesian | Jan 2012 | A1 |
20120043635 | Yang | Feb 2012 | A1 |
20120068327 | Oganesian | Mar 2012 | A1 |
20120068330 | Oganesian | Mar 2012 | A1 |
20120068351 | Oganesian | Mar 2012 | A1 |
20120068352 | Oganesian | Mar 2012 | A1 |
20120097245 | Nishina | Apr 2012 | A1 |
20120182693 | Boday | Jul 2012 | A1 |
20120217628 | Chou | Aug 2012 | A1 |
20130280864 | Bachman | Oct 2013 | A1 |
20150084148 | Oganesian | Mar 2015 | A1 |
Number | Date | Country |
---|---|---|
02-283053 | Nov 1990 | JP |
06-252300 | Sep 1994 | JP |
06-268110 | Sep 1994 | JP |
2003-258165 | Sep 2003 | JP |
2004-134480 | Apr 2004 | JP |
2006-245560 | Sep 2006 | JP |
2007-242864 | Sep 2007 | JP |
2010-239018 | Oct 2010 | JP |
2011-134769 | Jul 2011 | JP |
2012-164956 | Aug 2012 | JP |
2013-098212 | May 2013 | JP |
02076163 | Sep 2002 | WO |
Entry |
---|
U.S. Appl. No. 13/157,193, filed Jun. 9, 2011, Oganesian, Vage. |
U.S. Appl. No. 13/157,202, filed Jun. 9, 2011, Oganesian, Vage. |
U.S. Appl. No. 13/157,207, filed Jun. 9, 2011, Oganesian, Vage. |
U.S. Appl. No. 13/186,357, filed Jul. 19, 2011, Oganesian, Vage. |
U.S. Appl. No. 13/225,092, filed Sep. 2, 2011, Oganesian, Vage. |
U.S. Appl. No. 13/301,683, filed Nov. 21, 2011, Oganesian, Vage. |
U.S. Appl. No. 13/343,682, filed Jan. 4, 2012, Oganesian, Vage. |
U.S. Appl. No. 13/427,604, filed Mar. 22, 2012, Oganesian, Vage. |
U.S. Appl. No. 13/356,328, filed Jan. 23, 2012, Oganesian, Vage. |
U.S. Appl. No. 13/468,632, filed May 10, 2012, Oganesian, Vage. |
U.S. Appl. No. 13/559,510, filed Jul. 26, 2012, Oganesian, Vage. |
U.S. Appl. No. 13/423,045, filed Mar. 16, 2012, Oganesian, Vage. |
U.S. Appl. No. 13/609,002, filed Sep. 10, 2012, Oganesian, Vage. |
U.S. Appl. No. 13/312,826, filed Dec. 2011, Oganesian. |
Number | Date | Country | |
---|---|---|---|
20150311137 A1 | Oct 2015 | US |
Number | Date | Country | |
---|---|---|---|
61983402 | Apr 2014 | US |