`
`In re Patent of:
`U.S. Patent No.:
`Issue Date:
`Appl. Serial No.:
`Filing Date:
`Title:
`
` John Jianhong Zhu et al.
` 9,024,418 Attorney Docket No.: 39521-0045IP2
` May 5, 2015
` 13/829,864
` March 14, 2013
` LOCAL INTERCONNECT STRUCTURES FOR
`HIGH DENSITY
`
`
`
`DECLARATION OF DAVID KUAN-YU LIU, PH.D.
`
` My name is Dr. David Kuan-Yu Liu. I understand that I am
`
`submitting a declaration in connection with Inter Partes review (“IPR”)
`
`proceedings before the United States Patent and Trademark Office for U.S. Patent
`
`Number 9,024,418 (“’418 Patent”).
`
`
`
`I have been retained by Fish & Richardson P.C. (“Counsel”), on
`
`behalf of Apple Inc. to offer technical opinions with respect to the ’418 Patent and
`
`the references cited in this IPR. My compensation is not based on the outcome of
`
`my opinions.
`
` My curriculum vitae (“CV”) is provided as Exhibit 1004.
`
`
`
`I am not a lawyer. As set forth in my CV, I hold M.S. and Ph.D.
`
`degrees in Electrical Engineering from Stanford University and have 20 years of
`
`experience as an engineer and engineering manager/director of Complementary
`
`Metal Oxide Semiconductor (CMOS) technology development. I hold over 90
`
`1
`
`APPLE 1014
`
`
`
`Declaration of Dr. David Kuan-Yu Liu
`U.S. Patent No.: 9,024,418
`Attorney Docket No.: 39521-0045IP2
`
`
`U.S. patents, a large number of which are directed to CMOS processes and
`
`semiconductor technology.
`
`
`
`The vast majority of my patents are in the area of CMOS high density
`
`circuit and memory architecture. As such, I am intimately familiar with the
`
`concept of having a compact layout of transistors to implement a library of logic
`
`gate functions. For example, US Patent Nos. 6,693,027 and 8,988,103 relate to the
`
`use of novel device physics to facilitate compact and high performance logic gate
`
`functions.
`
`
`
`I have also authored several technical papers that have been published
`
`in well-respected, peer-reviewed journals, such as the IEEE Electron Device
`
`Letters, the IEEE Journal of Solid-State Circuits, and the IEEE Transactions on
`
`Electron Devices. As an example, I worked on a new conductivity-modulated
`
`Power MOSFET that features a buried minority-carrier injector to enhance the
`
`current conduction capability of the Power MOSFET. This work was published in
`
`the IEEE Transactions on Electron Devices.
`
`
`
`During my career, I have worked at some of the leading technology
`
`companies in the world, such as Texas Instruments, Advanced Micro Devices,
`
`Altera Corporation (now a subsidiary of Intel Corp.), and Xilinx. At these
`
`2
`
`
`
`Declaration of Dr. David Kuan-Yu Liu
`U.S. Patent No.: 9,024,418
`Attorney Docket No.: 39521-0045IP2
`
`
`companies, my work focused on various aspects of CMOS and semiconductor
`
`technology.
`
`
`
`The following provides a chronological highlight of my experience as
`
`discussed above and in my CV. From 1989 to 1992, I was a member of technical
`
`staff at Texas Instruments, Inc. At Texas Instruments, my job responsibilities at
`
`Texas Instruments also included process integration, device modeling, high-
`
`voltage CMOS process integration, and investigating novel source-side injection
`
`mechanisms for Flash EPROM channel hot-electron programming.
`
`
`
`From 1992 to 1995, I was a member of the technical staff at Advanced
`
`Micro Devices, Inc. (AMD) where I was a key contributor in optimizing Flash cell
`
`and periphery devices in AMD’s CMOS-based 0.5um and 0.35um Flash EPROM
`
`technology. My job responsibilities at AMD also included process integration,
`
`device modeling, and development of triple-well process technology for
`
`accommodate x-decoder transistors and high voltage transistors for negative gate
`
`erase operation. While at AMD, I was awarded a Spotlight Award for developing
`
`a method of manufacturing a self-aligned source (SAS) etch for a NOR flash
`
`memory.
`
`
`
`I spent the next five years of my career in managerial and director
`
`roles at several California-based semiconductor companies. I was responsible for
`
`3
`
`
`
`Declaration of Dr. David Kuan-Yu Liu
`U.S. Patent No.: 9,024,418
`Attorney Docket No.: 39521-0045IP2
`
`
`increasing yield and for leading teams of engineers working to develop next-
`
`generation memory devices.
`
`
`
`In 2000, I co-founded Progressant Technologies in Fremont,
`
`California. Progressant Technologies developed IP for negative differential
`
`resistance transistor technology and was eventually acquired by Synopsys, Inc.
`
` From 2000 to 2004, I was a Senior Manager at Xilinx, Incorporated,
`
`where I was responsible for developing nonvolatile memory process technology
`
`for flash and CPLD product applications, as well as advanced CMOS process
`
`technology (specifically 75nm CMOS technology node, a half node version
`
`between 90nm and 65nm).
`
` From 2004 to 2007, I was a Senior Scientist at Maxim Integrated
`
`Products where I was responsible for developing Embedded Non-volatile Memory
`
`process technology for Power Management product applications.
`
` Since 2007, I have served as a technical consultant where I have
`
`provided expert advice regarding Flash memory technology, CMOS process
`
`technology, and semiconductor device physics.
`
`
`
`
`
`I am not now, and have never been an employee of the Petitioner.
`
`In writing this Declaration, I have considered the following: my own
`
`knowledge and experience, including my work experience in the fields of
`
`4
`
`
`
`Declaration of Dr. David Kuan-Yu Liu
`U.S. Patent No.: 9,024,418
`Attorney Docket No.: 39521-0045IP2
`
`
`semiconductor design and processes; and my experience in working with others
`
`involved in those fields. I have reviewed the ’418 Patent, including the claims of
`
`the patent in view of the specification and the file history. In addition, I have
`
`reviewed the following documents:
`
` U.S. Patent No. 9,024,418 (“’418 patent” or “Ex. 1001”), and its
`
`accompanying prosecution history (“’418 Prosecution History” or
`
`“Ex. 1002”)
`
` Behzad Razavi, Design of Analog CMOS Integrated Circuits, 2001
`
`(“Razavi” or “Ex. 1008”)
`
` Chenming Hu, Modern Semiconductor Devices for Integrated
`
`Circuits, 195-258, 2010 (“Hu” or “Ex. 1009”)
`
` U.S. Patent No. 8,026,536 ("Yoshida" or "Ex. 1015")
`
` U.S. Patent No. 9,355,910 ("Liaw" or "Ex. 1016")
`
` U.S. Patent No. 4,570,176 ("Kolwicz" or "Ex. 1017")
`
` U.S. Patent No. 5,376,585 (“Lin” or “Ex. 1018”)
`
` U.S. Patent No. 7,335,583 (“Chang” or “Ex. 1019”)
`
` U.S. Patent No. 6,818,547 (“Chen” or “Ex. 1020”)
`
` U.S. Patent No. 6,767,827 (“Okada” or “Ex. 1021”)
`
` U.S. Patent No. 6,969,952 (“Gedamu or “Ex. 1022”)
`
`5
`
`
`
`Declaration of Dr. David Kuan-Yu Liu
`U.S. Patent No.: 9,024,418
`Attorney Docket No.: 39521-0045IP2
`
`
` James B. Kuo and Ker-Wei Su, CMOS VLSI Engineering Sili
`
`con-on-Insulator (SOI), 123-131, 1998 (“Kuo” or “Ex. 1023”)
`
` Counsel has informed me that I should consider these materials
`
`through the lens of one of ordinary skill in the art related to the ’418 patent at the
`
`time of the earliest effective priority date (“Critical Date”) of the ’418 Patent, and
`
`I have done so during my review of these materials. My understanding is that the
`
`Critical Date in this case is March 14, 2013, the filing date of U.S. Application No.
`
`13/829,864, from which the ’418 patent issued.
`
`
`
`I am familiar with the knowledge and capabilities of one of ordinary
`
`skill in the areas mentioned above, notably including the areas of semiconductors
`
`and semiconductor processes. My experience working in industry has allowed me
`
`to become directly and personally familiar with the level of skill of individuals and
`
`the general state of the art in these areas. Unless otherwise stated, my testimony
`
`below refers to the knowledge of one of ordinary skill (as described in Section II
`
`below) in the fields as of the Critical Date, or before.
`
` This declaration is organized as follows:
`
`I.
`
`II.
`
`Brief Overview of the ’418 Patent
`
`Level of Ordinary Skill in the Art
`
`III. Terminology
`
`6
`
`
`
`Declaration of Dr. David Kuan-Yu Liu
`U.S. Patent No.: 9,024,418
`Attorney Docket No.: 39521-0045IP2
`
`
`IV. Discussion of References
`
`V.
`
`Legal Principles
`
`VI. Conclusion and Additional Remarks
`
`I.
`
`Brief Overview of the ’418 Patent
`
` The ’418 Patent includes 20 claims, of which claim 1, 12 and 17 are
`
`independent. Claims 2-11, 13-16, and 18-20 depend directly or indirectly from
`
`claims 1, 12, and 17, respectively. The technology in the ’418 Patent generally
`
`relates to a local interconnect structure that includes a gate-directed local
`
`interconnect coupled to an adjacent gate layer through a diffusion-directed local
`
`interconnect. ’418 Patent at Abstract. For the purposes of my analysis of the art
`
`and claimed features, I have understood the term “couple” as including direct or
`
`indirect connections, which Counsel informed me is not inconsistent with how
`
`courts have interpreted “couple.”
`
`Design Trade-offs Discussed by the ’418 Patent
`
` The Background section of the ’418 Patent (“The Background”)
`
`describes the design trade-offs when trying to implement high density circuits in
`
`sub-micron processes. ’418 Patent at 1:10-2:22. The discussion of design trade-
`
`offs begins in the ’418 Patent with the need to have high density circuits while
`
`sustaining sufficient transistor strength. ’418 Patent at 1:15-25. The Background
`
`7
`
`
`
`Declaration of Dr. David Kuan-Yu Liu
`U.S. Patent No.: 9,024,418
`Attorney Docket No.: 39521-0045IP2
`
`
`notes that “[t]o achieve high density yet have adequate transistor strength, strain
`
`engineering techniques have been developed so that the crystal lattice for the
`
`semiconductor substrate is strained in the diffusion region used to form the
`
`transistor source and drains.” ’418 Patent at 1:16-20.
`
` However, one design trade-off here is that straining techniques, such
`
`as compressive strain or tensile strain, introduce “a number of constraints into the
`
`layout process.” Id. at 1:26-58. For example, shorter diffusion regions that are
`
`provided for individual transistors may relax the strain and reduce the enhancement
`
`in the transistor strength compared to longer diffusion regions, or continuous
`
`diffusion regions, that are shared by multiple transistors. Id. On the other hand,
`
`implementing continuous diffusion regions shared by multiple transistors may
`
`cause one or more of the multiple transistors to be electrically shorted. Id.
`
` One solution noted in the ’418 Patent is to use a blocking transistor to
`
`isolate two neighboring transistors. Id. at 1:66-2:10. For example, as shown in the
`
`’418 Patent’s FIG. 2 below, a blocking transistor 201 may be used to isolate
`
`transistors 100 and 101. Id. The “diffusion region 200 is continuous for both
`
`transistors such that it can develop adequate lattice strain for satisfactory transistor
`
`strength.” Id. at 1:64-66.
`
`8
`
`
`
`Declaration of Dr. David Kuan-Yu Liu
`U.S. Patent No.: 9,024,418
`Attorney Docket No.: 39521-0045IP2
`
`
`
`
` Yet, the introduction of a blocking transistor results in another trade
`
`off with chip density. In particular, the ’418 Patent explains that when blocking
`
`transistors were used, chip density decreased and layout designs became difficult
`
`to implement at the desired density. Id. at 2:10-18.
`
`Layout for Implementing High-Density Circuit with sufficient transistor
`strength
`
` To address the limitations and trade-offs noted above, the ’418 Patent
`
`discloses a circuit architecture that purportedly “provide[d] an advantageously
`
`dense local interconnect coupling for blocking transistors”. Id. at 5:49-41. An
`
`aerial layout view of the circuit implementation proposed by the ’418 Patent is
`
`provided in FIGS. 4A and 4B (reproduced below).
`
`9
`
`
`
`Declaration of Dr. David Kuan-Yu Liu
`U.S. Patent No.: 9,024,418
`Attorney Docket No.: 39521-0045IP2
`
`
`’418 Patent at FIGS. 4A (left) and 4B (right) (with colored annotations)
`
`
`
`
`
` As shown in FIGS. 4A and 4B, the ’418 Patent proposes
`
`implementing a blocking transistor 430 between two transistors, a first transistor
`
`405 and a second transistor 420, all in a common diffusion region OD 400. Id. at
`
`6:6-29. The following sentences from the ’418 Patent describe the layouts in
`
`FIGS. 4A and 4B.
`
`
`
`“A gate layer 410, a gate layer 425, and a gate layer 415 form the
`
`gates for transistor 405, blocking transistor 430, and transistor 420, respectively.”
`
`Id. at 6:15-17. “A gate-directed local interconnect 440 couples (through a
`
`corresponding level 1 gate-directed interconnect, which is not illustrated) to a drain
`
`10
`
`
`
`Declaration of Dr. David Kuan-Yu Liu
`U.S. Patent No.: 9,024,418
`Attorney Docket No.: 39521-0045IP2
`
`
`(D) for transistor 405. Similarly, a gate-directed local interconnect 435 couples to
`
`a source for transistor 420. A via V0 coupled to gate-directed local interconnect
`
`435 provides the appropriate source voltage for transistor 420.” Id. at 6:17-23.
`
`
`
`In the ’418 Patent, "gated-directed" appears to refer to layers and
`
`elements that extend according to a "gate-layer" pitch or parallel to a gate structure,
`
`and "diffusion-directed" appears to refer to layers or elements that "extend in a
`
`diffusion-directed direction that is generally orthogonal to the gate-directed
`
`direction." Id. at 2:36-44.
`
`
`
`In FIG. 4A, a diffusion-directed local interconnect 445 “couples
`
`between gate layer 425 and gate-directed local interconnect 435 to provide the bias
`
`to gate layer 425 to turn blocking transistor 430 fully off.” Id. at 6:36-39. In FIG.
`
`4B, similarly, “a diffusion-directed local interconnect 450 couples between gate-
`
`directed local interconnect 435 and gate layer 425 analogous to the coupling
`
`provided by diffusion-directed local interconnect 445 of Figure 4A.” Id. at 7:9-12.
`
`In my opinion, the main difference between FIGS. 4A and 4B is the location of the
`
`diffusion-directed local interconnect. For example, in FIG. 4B, the diffusion-
`
`directed local interconnect 450 overlaps a portion of the diffusion region OD 400.
`
`In contrast, in FIG. 4A, the diffusion-directed local interconnect 450 does not
`
`11
`
`
`
`Declaration of Dr. David Kuan-Yu Liu
`U.S. Patent No.: 9,024,418
`Attorney Docket No.: 39521-0045IP2
`
`
`overlap any portion of the diffusion region OD 400, and is placed outside of the
`
`diffusion region OD 400.
`
`
`
`In addition to the layout of transistors in a circuit, as indicated by the
`
`Abstract and Summary of the ’418 Patent, an important aspect of the ’418 Patent is
`
`the local interconnect structure “that includes a gate-directed local interconnect
`
`coupled to an adjacent gate layer through a diffusion-directed local interconnect.”
`
`’418 Patent at Abstract, 2:44-67. FIG. 3 (shown below) depicts a lateral view or
`
`vertical profile of the interconnect structures.
`
` The interconnect structure includes three levels of interconnects
`
`formed between the diffusion region 305 and a first metal layer M1. The first level
`
`
`
`12
`
`’418 Patent at FIG. 3 (with colored annotations)
`
`
`
`
`
`Declaration of Dr. David Kuan-Yu Liu
`U.S. Patent No.: 9,024,418
`Attorney Docket No.: 39521-0045IP2
`
`
`of interconnects include interconnects Llc 310 and gate layer 300. Id. at 4:5-17,
`
`4:44-46. The first level of interconnects are formed on the continuous diffusion
`
`region 305 and provide direct electrical coupling to diffusion region 305. Id.
`
`Level 1 local interconnects must be gate-directed local interconnects. Id. at 4:47-
`
`50.
`
` The second level of interconnects include interconnects Lla 315 and
`
`Llb 320. Id. at 4:5-17. The second level of interconnects are formed between the
`
`first level and the third level of interconnects and “couple to first metal layer Ml
`
`(or higher metal layers) through vias such as a via V0.” Id. at 4:19-23. Level 2
`
`local interconnects can be either gate-directed local interconnects (e.g., 315) or
`
`diffusion-directed local interconnects (e.g., 320). Id. at 4:47-54, 2:46-49. Level 2
`
`interconnects are distinguished from level 1 interconnects in that level 1
`
`interconnects are “directly couple[d] to a continuous diffusion region” and are
`
`implemented at “the level for a gate layer.” ‘418 Patent at 4:4-23.
`
` Level 3 interconnects include vias, e.g., via V0, and are arranged
`
`between level 2 interconnects and a first metal layer Ml. Id. at 4:19-23. ‘418
`
`Patent at 4:4-23. Level 3 interconnects are distinguished from level 2
`
`interconnects in that level 2 interconnects do not include vias that connect to metal
`
`layer M1. Id.
`
`13
`
`
`
`Declaration of Dr. David Kuan-Yu Liu
`U.S. Patent No.: 9,024,418
`Attorney Docket No.: 39521-0045IP2
`
`
` The ’418 Patent contends that the combination of the interconnect
`
`structure shown in FIG. 3 and the layout structure shown in FIG. 4A or FIG. 4B
`
`would address the design trade-offs and issues discussed above in paragraphs. Id.
`
`at 2:10-44, 5:1-64. In particular, a blocking transistor can be implemented between
`
`two transistors so that satisfactory transistor strength and a continuous diffusion
`
`region may be provided for multiple transistors. Id. At the same time, the
`
`interconnect structure helps with providing the requisite chip density. Id. at 6:58-
`
`7:23.
`
` The embodiments disclosed in the ’418 Patent could be implemented
`
`in various types of circuits. For instance, FIGS. 5A-5C appear to disclose a diode-
`
`connected transistor in which a diffusion-directed local interconnect is used. ’418
`
`Patent at 3:20-33. FIGS. 6A-6B disclose an inverter-to-inverter circuit in which a
`
`diffusion-directed local interconnect is used. ’418 Patent at 3:34-40. Although
`
`just limited types of circuits are described in the ’418 Patent, it is my
`
`understanding that the embodiments disclosed in the ’418 Patent could be
`
`implemented in other types of circuits as well.
`
` Notwithstanding the above, as noted in Section IV, based on my
`
`knowledge and experience in the industry, prior to the Critical Date, there existed
`
`numerous products, publications, and patents that implemented or described the
`
`14
`
`
`
`Declaration of Dr. David Kuan-Yu Liu
`U.S. Patent No.: 9,024,418
`Attorney Docket No.: 39521-0045IP2
`
`
`embodiments claimed in the ’418 patent. As detailed in Section IV below, the
`
`circuit and circuit implementations disclosed in the ’418 patent were well-known
`
`in the prior art. To the extent there was any problem to be solved in the ’418
`
`patent, it had already been solved in the prior art systems before the Critical Date.
`
`
`
`II. Level of Ordinary Skill in the Art as of the Critical Date
`
`
`
`It is my understanding that obviousness and claim interpretation,
`
`among other things, are determined, in part, from the point of view of a person of
`
`ordinary skill in the art as of the Critical Date of the ’418 Patent (“APOSITA”).
`
`APOSITA would have had a Master’s of Science Degree (or a similar technical
`
`Master’s Degree, or higher degree) in an academic area emphasizing electrical
`
`engineering or computer engineering with a concentration in semiconductors or,
`
`alternatively, a Bachelors Degree (or higher degree) in an academic area
`
`emphasizing electrical or computer engineering and having two or more years of
`
`experience in integrated circuit design and/or semiconductor processing.
`
`Additional education in a relevant field, such as computer engineering, or electrical
`
`engineering, or industry experience may compensate for a deficit in one of the
`
`other aspects of the requirements stated above. Unless noted otherwise in this
`
`15
`
`
`
`Declaration of Dr. David Kuan-Yu Liu
`U.S. Patent No.: 9,024,418
`Attorney Docket No.: 39521-0045IP2
`
`
`Petition, references to what would have been known or understood by APOSITA
`
`refer to the knowledge of APOSITA as of the Critical Date, or before.
`
`
`
`III. Terminology
`
`
`
`I am not a lawyer. However, I have been informed that, during an IPR
`
`proceeding involving the ’418 Patent, claim terminology is given the broadest
`
`reasonable interpretation (BRI) at the time of the Critical Date. Counsel has also
`
`informed me that PTAB is considering a change of the claim interpretation
`
`standard from BRI to the Phillips standard under which the words of the claims
`
`should be given their ordinary meaning.
`
`
`
`I have been requested to provide some guidance in this proceeding
`
`with respect to the terms below. In response to the request, I considered the
`
`context of the terms within the claims, use of the terms within the specification, my
`
`understanding of how APOSITA would have understood the terms as of the
`
`Critical Date, and have provided my interpretation below. I have used the Critical
`
`Date as the point in time for claim interpretation purposes, although in many cases
`
`the same analysis would hold true even at an earlier time than the Critical Date.
`
`
`
`
`
`16
`
`
`
`Declaration of Dr. David Kuan-Yu Liu
`U.S. Patent No.: 9,024,418
`Attorney Docket No.: 39521-0045IP2
`
`
`A.
`
`“ means” for coupling
`
` Based on my review, I believe the “means” in “means for coupling,”
`
`as recited in claims 17-19 encompasses a “diffusion-directed local interconnect.”
`
`
`
`I base my opinion, in part, from the disclosure in the ’418 Patent,
`
`which states that “[t]he diffusion-directed local interconnect [ ] serves as a means
`
`for coupling one of the gate-directed local interconnects to the gate layer.” ’418
`
`Patent at 5:62-64. The “diffusion-directed local interconnect 445 [ ] couples
`
`between gate layer 425 and gate-directed local interconnect 435” and “diffusion-
`
`directed local interconnect 450 couples between gate-directed local interconnect
`
`435 and gate layer 425 analogous to the coupling provided by diffusion-directed
`
`local inter-connect 445.” ’418 Patent at 6:36-38, 7:9-12; see also FIGS. 4A, 4B,
`
`5A, 5B. Based on this explicit description in the ’418 Patent, in my opinion, the
`
`“means” in “means for coupling” encompasses a “diffusion-directed local
`
`interconnect.”
`
`B.
`
`“continuous” diffusion region
`
` Based on my review of the ’418 Patent, my understanding is that the
`
`“continuous” in “continuous diffusion region,” as recited in claims 2-4, 9, 14, 15,
`
`and 17-20, encompasses a diffusion region that does not include any isolation
`
`region.
`
`17
`
`
`
`Declaration of Dr. David Kuan-Yu Liu
`U.S. Patent No.: 9,024,418
`Attorney Docket No.: 39521-0045IP2
`
`
`
`
` Here again, I base my opinion, in part, from the disclosure in the ’418
`
`Patent. A continuous diffusion region 200 is shown in FIG. 2 of the ’418 Patent
`
`(shown below) to be located between the annotated red lines and extending from
`
`the right to the left sides of FIG. 2. ’418 Patent at 1:59-1:66. The depicted
`
`continuous diffusion region 200 includes doped regions such as p-type doped
`
`regions for a PMOS transistor or n-type doped regions for a NMOS transistor, and
`
`may include multiple source and drain regions. ’418 Patent at 2:2-2:10, FIG. 2.
`
`This is a well-known composition of a continuous diffusion region, as indicated by
`
`the label of “PRIOR ART” on FIG. 2. The multiple source and drain regions and
`
`gate layers 110, 120, 205 correspond, respectively, to multiple transistors 100, 102,
`
`201 that are implemented in the single continuous diffusion region. ’418 Patent at
`
`5:9-24. For example, a single continuous diffusion region 200 is implemented for
`
`transistors 100, 101, and blocking transistor 201. ’418 Patent at 1:66-2:10.
`
`18
`
`
`
`Declaration of Dr. David Kuan-Yu Liu
`U.S. Patent No.: 9,024,418
`Attorney Docket No.: 39521-0045IP2
`
`
`Transistor
`
`Blocking
`transistor
`
` Transistor
`
`Source
`
`Drain
`
`Continuous
`Diffusion
`Region
`
`Source
`
`Drain
`
`polysilicon
`gate layer
`
`polysilicon
`gate layer
`
`polysilicon
`gate layer
`
`
`
`’418 Patent at FIGS. 2 (with colored annotations)
`
` While the ’418 Patent illustrates a continuous diffusion region 200, I
`
`could not locate an express definition of a “continuous” diffusion region in the
`
`’418 Patent. However, based on my knowledge and experience in this area, I am
`
`aware that a diffusion region is referred to as “continuous” when it does not
`
`include any isolation regions (e.g., shallow trench isolation, local oxidation of
`
`silicon (LOCOS) regions) thereby allowing the diffusion region 200 to be
`
`uninterrupted or “continuous.” My understanding of a “continuous” diffusion
`
`region, as depicted in the’418 Patent, is consistent with the interpretation other
`
`references and persons of ordinary skill in the art would have. For example, as
`
`shown in Chang’s FIG. 3A below, regions 306, 308, 310, and 312 are each
`
`continuous diffusion regions which are separated by isolation regions 314 but do
`
`19
`
`
`
`Declaration of Dr. David Kuan-Yu Liu
`U.S. Patent No.: 9,024,418
`Attorney Docket No.: 39521-0045IP2
`
`
`not include any isolation region within the respective regions 306, 308, 310, and
`
`312. Chang at 3:11-17. Accordingly, in my opinion, the “continuous” in
`
`“continuous diffusion region” recited in the’418 Patent would be correctly
`
`understood as encompassing a diffusion region that does not include any isolation
`
`region.
`
`Continuous
`diffusion
`regions
`
`Isolation
`regions
`
`
`
`Chang at FIG. 3A (with colored annotations)
`
`
`
`
`
`
`
`20
`
`
`
`Declaration of Dr. David Kuan-Yu Liu
`U.S. Patent No.: 9,024,418
`Attorney Docket No.: 39521-0045IP2
`
`
`IV. Discussion of References
`
` The following exemplary references demonstrate the prevalence of the
`
`claimed features in the prior art.
`
`A. Yoshida
`
` Yoshida describes implementations of a ring oscillator (also referred
`
`to as semiconductor device 10 in Yoshida) that includes an odd number of
`
`inverters. Yoshida at 1:7-15, 1:53-67. Yoshida describes problems with ring
`
`oscillators that are implemented in a small area, and, in particular, that the “desired
`
`operational speed [of the ring oscillator] cannot be obtained.” Id. at 1:21-25. The
`
`reduced operations speed can be attributed to parasitic capacitance between various
`
`elements in the ring oscillator. Id. at 1:25-48.
`
` To reduce parasitic capacitance and improve ring oscillator
`
`operational speed, Yoshida discloses particular implementations of ring oscillators
`
`using MOS transistors. Id. at 4:66-5:5; Abstract. An example of an
`
`implementation is shown below in Yoshida’s FIG. 4.
`
`21
`
`
`
`Declaration of Dr. David Kuan-Yu Liu
`U.S. Patent No.: 9,024,418
`Attorney Docket No.: 39521-0045IP2
`
`
`PMOS
`section
`
`NMOS
`section
`
`
`
`Yoshida at FIG. 4 (with colored annotations)
`
`Yoshida’s FIG. 4 depicts a ring oscillator that includes a PMOS section 31
`
`and a NMOS section 32. Yoshida at 8:58-65. “The PMOS section 31 and NMOS
`
`section 32 are arranged in parallel to each other in the lengthwise direction, with
`
`the pMOS transistors 14-1 to 14-3 of PMOS section 31 and the respective nMOS
`
`transistors 14-1 to 14-3 of NMOS section 32 opposing each other.” Id. at 8:65-9:3.
`
`22
`
`
`
`Declaration of Dr. David Kuan-Yu Liu
`U.S. Patent No.: 9,024,418
`Attorney Docket No.: 39521-0045IP2
`
`
`Yoshida’s FIG. 1 (included below) provides a more detailed view of the PMOS
`
`section 31.
`
`
`
`
`Dummy gates
`
`Drain regions
`
`Transistors
`
`Source
`regions
`
`Gate
`Electrode
`12
`
`
`
`Yoshida at FIG. 1 (with colored annotations)
`
` As shown in FIG. 1, a “single pMOS transistor 14 is configured by a
`
`single drain region 13b, a pair of source regions 13a on both sides of the drain
`
`region 13b, and a pair of (normal) gate electrodes 12 formed on the semiconductor
`
`substrate 11 at the portion between the source 13a and drain region 13b. The
`
`depicted portion of the semiconductor device 10 includes therein three transistors
`
`14-1, 14-2, and 14-3. The source regions 13a formed in the vicinity of both the
`
`23
`
`
`
`Declaration of Dr. David Kuan-Yu Liu
`U.S. Patent No.: 9,024,418
`Attorney Docket No.: 39521-0045IP2
`
`
`edges of the semiconductor substrate 11 outside the corresponding dummy gate
`
`electrode 12a are dummy sources.” Id. at 7:32-41. “[G]ate electrodes 12 [are]
`
`arranged at a constant pitch or constant interval on the semiconductor substrate
`
`11.” Id. at 2:5-6. “The dummy gate electrodes 12a are maintained at the same
`
`potential as the potential of the source regions 13a [ ] by using interconnects not
`
`shown.” Id. at 7:28-31. My understanding is that the lines shown in red in FIG. 1
`
`depict the connection between the dummy gate electrodes 12a and the source
`
`regions 13a, although the interconnects used to implement such connections are
`
`not depicted in Yoshida. Id. at 7:28-31.
`
` More details of Yoshida’s pMOS section 31 are provided in Yoshida’s
`
`FIG. 2, shown below, which “is a sectional view taken along II-II line in FIG. 1.”
`
`Id. at 5:37-38.
`
`first-layer source line
`
`first-layer drain line
`
`Source
`region
`
`Dummy gate
`
`Drain region Gate Electrode
`Yoshida at FIG. 2 (with colored annotations)
`
`
`
`24
`
`
`
`Declaration of Dr. David Kuan-Yu Liu
`U.S. Patent No.: 9,024,418
`Attorney Docket No.: 39521-0045IP2
`
`
` FIG. 2 depicts the drain regions 13b, source regions 13a, gate
`
`electrodes 12, dummy gate 12a formed on the semiconductor substrate 11, as
`
`described above with reference to FIG. 1. In addition, FIG. 2 shows a first-layer
`
`source line/interconnect 17a and a first-layer drain line/interconnect 17b. The
`
`“first-layer source line 17a is connected via the first-layer source contacts 16a
`
`disposed adjacent to each other to the source regions 13a.” Yoshida at 7:41-43,
`
`2:38-40. The first-layer drain line/interconnect 17b is connected via the first-layer
`
`drain contacts 16b to the drain regions 13b. Id. at 2:31-48. Lines 17a and 17b
`
`function as local interconnects that connect to the source regions 13a and drain
`
`regions 13b, respectively. See Id. at 7:46, 2:57.
`
`
`
`In my opinion, the subject matter in Yoshida is similar to the ’418
`
`Patent at least because both relate to implementing semiconductor devices in a
`
`small area, and disclose efficient semiconductor layout techniques to optimize the
`
`usage of the chip or substrate area. For instance, in Yoshida, the semiconductor
`
`device 10 (ring oscillator) includes gate electrodes 12/12a that are implemented at
`
`a constant pitch and a single dummy gate electrode 12a is shared for two different
`
`transistors. The reduction of a dummy gate electrode 12a and use of constant gate
`
`pitches are just some examples of how Yoshida reduces the chip area required to
`
`implement its semiconductor device 10. Yoshida at 2:5-6, 7:25. The following
`
`25
`
`
`
`Declaration of Dr. David Kuan-Yu Liu
`U.S. Patent No.: 9,024,418
`Attorney Docket No.: 39521-0045IP2
`
`
`explanation provides a more detailed description of Yoshida’s efficient layout
`
`implementation.
`
` Below I have provided a simplified version of Yoshida’s FIG. 1. The
`
`depiction below, consistent with Yoshida’s FIG. 1, illustrates a series of transistors
`
`14 attached to each other by dummy gate electrodes 12a and bordered at two ends
`
`by dummy sources 13a and dummy gates 12a. Yoshida at 7:36-41.
`
`
`
`Dummy
`source 13a
`
`
`
`mos
`
`p
`
`1
`
`mos
`
`p
`
`1
`
`
`
`
`
`mos
`
`p
`
`1
`
`
`
`4
`4
`4
`Dummy gate electrodes 12a
`(
`(
`(
`
`Dummy
`source 13a
`
`14-3)
`14-1)
`14-2)
` Two dummy gate electrodes 12a are disposed at two ends of each
`
`transistor cell e.g., left and right border of the transistor cell. Yoshida at 7:23-41.
`
`
`
`
`
`
`
`
`
`
`
`
`
`p
`
`1
`
`mos
`
`4
`
`Dummy gate electrodes 12a
`
`26
`
`
`
`Declaration of Dr. David Kuan-Yu Liu
`U.S. Patent No.: 9,024,418
`Attorney Docket No.: 39521-0045IP2
`
`
` The equivalent depiction of this cell in Yoshida’s FIGS. 1 and 2
`
`would appear as follows:
`1
`
`1
`
`2a
`
`12a
`
`2a
`
`12a
`
`
`
`
`12a
` (portion of FIG. 1) (portion of FIG. 2)
`
`
`
`12a
`
` To implement high density circuits with multiple transistors, Yoshida
`
`arranges multiple transistor cells next to each other to form a gate array structure.
`
`To optimize utilization of chip real estate and because the dummy gate electrode
`
`12a of each PMOS transistor is tied to a common power source line such as Vdd,
`
`two neighboring transistors cells can be combined, t