Usage of Process Dependent Random Inaccuracies as Tracking Marks
|
|
Matthias Steiert (matthias.steiert@imtek.uni-freiburg.de) of University of Freiburg will present on Usage of Process Dependent Random Inaccuracies as Tracking Marks at the CALCE SMTA conference at College Park, Maryland on June 25th.
Abstract:
The reliable authentication and differentiation of original and fake microsystem components is important for many electronic products. Principally there are two opportunities, using extrinsic marks or intrinsic marks. Extrinsic marks are those which are additionally tagged to the microsystem or electronic component. For this purpose marks are used which are considered unforgeable like biological tags, RFID-tags, fluorescent pigments or random serial numbers. The problem thereby is: whether such an extrinsic mark gets published oven it could be also counterfeited and it follows that such a counterfeit tag can be identified first if there are both the original and the fake tag at the same time on the same place. Much better are intrinsic marks. These marks are a part of the component and they are also considered unforgeable. Examples are physical unclonable functions or IP-watermarks. In case of electronic components often geometrical data are used like the position of a chip or the geometry of lead frames. One problem of all known trademarks is that they are not useful for all integration steps, starting with bare-chips, encapsulated chips and complete electronic applications. To find intrinsic and unforgeable marks we investigated different process steps of the clean room back-end, dicing, pick and place, molding etcetera and we searched for random inaccuracies of those processes. Taking those inaccuracies into account and let us suppose there are enough inaccuracies, it can be assumed that each component is unique and by saving those data each component can be found again. We found differed inaccuracies like bubbles in die attachment layers or in mold masses. One of the most interesting marks is caused by the dicing. In case of the blade dicing small chipping at the chips edges is generated. Those chipping defects are more or less like a fingerprint of a chip.
Bio:
Matthias Steiert studied from 2003 to 2010 micro system technology with a focus on Life Science at the Albert-Ludwigs-University of Freiburg. Since May 2011, he is a research associate at the Chair of assembly and interconnection technology. In November 2013, Mr. Steiert finished his PhD with great praise on "Influence of dicing damages on the thermo-mechanical reliability of bare-chip assemblies". His current research focus is on new measuring method for shape and deformation measurement of electronic components under thermal stress. The focus here is the digital image correlation and the multi-wavelength holography.
The reliable authentication and differentiation of original and fake microsystem components is important for many electronic products. Principally there are two opportunities, using extrinsic marks or intrinsic marks. Extrinsic marks are those which are additionally tagged to the microsystem or electronic component. For this purpose marks are used which are considered unforgeable like biological tags, RFID-tags, fluorescent pigments or random serial numbers. The problem thereby is: whether such an extrinsic mark gets published oven it could be also counterfeited and it follows that such a counterfeit tag can be identified first if there are both the original and the fake tag at the same time on the same place. Much better are intrinsic marks. These marks are a part of the component and they are also considered unforgeable. Examples are physical unclonable functions or IP-watermarks. In case of electronic components often geometrical data are used like the position of a chip or the geometry of lead frames. One problem of all known trademarks is that they are not useful for all integration steps, starting with bare-chips, encapsulated chips and complete electronic applications. To find intrinsic and unforgeable marks we investigated different process steps of the clean room back-end, dicing, pick and place, molding etcetera and we searched for random inaccuracies of those processes. Taking those inaccuracies into account and let us suppose there are enough inaccuracies, it can be assumed that each component is unique and by saving those data each component can be found again. We found differed inaccuracies like bubbles in die attachment layers or in mold masses. One of the most interesting marks is caused by the dicing. In case of the blade dicing small chipping at the chips edges is generated. Those chipping defects are more or less like a fingerprint of a chip.
Bio:
Matthias Steiert studied from 2003 to 2010 micro system technology with a focus on Life Science at the Albert-Ludwigs-University of Freiburg. Since May 2011, he is a research associate at the Chair of assembly and interconnection technology. In November 2013, Mr. Steiert finished his PhD with great praise on "Influence of dicing damages on the thermo-mechanical reliability of bare-chip assemblies". His current research focus is on new measuring method for shape and deformation measurement of electronic components under thermal stress. The focus here is the digital image correlation and the multi-wavelength holography.