U.S. patent application number 15/953564 was filed with the patent office on 2018-08-16 for semiconductor device and authentication system.
The applicant listed for this patent is Renesas Electronics Corporation. Invention is credited to Takeshi OKAGAKI.
Application Number | 20180233461 15/953564 |
Document ID | / |
Family ID | 58409863 |
Filed Date | 2018-08-16 |
United States Patent
Application |
20180233461 |
Kind Code |
A1 |
OKAGAKI; Takeshi |
August 16, 2018 |
SEMICONDUCTOR DEVICE AND AUTHENTICATION SYSTEM
Abstract
In order to realize a silicon PUF of lower power consumption, a
semiconductor device includes first and second MIS transistors of
the same conductive type in off-state coupled in series, as a PUF
element. The PUF element outputs a signal of high level or low
level depending on the potential of a connection node of the first
and the second MIS transistors. Preferably, the MIS transistors are
fin-type FETs.
Inventors: |
OKAGAKI; Takeshi; (Tokyo,
JP) |
|
Applicant: |
Name |
City |
State |
Country |
Type |
Renesas Electronics Corporation |
Tokyo |
|
JP |
|
|
Family ID: |
58409863 |
Appl. No.: |
15/953564 |
Filed: |
April 16, 2018 |
Related U.S. Patent Documents
|
|
|
|
|
|
Application
Number |
Filing Date |
Patent Number |
|
|
15212174 |
Jul 15, 2016 |
9972586 |
|
|
15953564 |
|
|
|
|
Current U.S.
Class: |
1/1 |
Current CPC
Class: |
H01L 23/576 20130101;
H01L 27/0207 20130101 |
International
Class: |
H01L 23/00 20060101
H01L023/00; H01L 27/02 20060101 H01L027/02 |
Foreign Application Data
Date |
Code |
Application Number |
Sep 24, 2015 |
JP |
2015-186993 |
Claims
1-11. (canceled)
12. A semiconductor device comprising: a plurality of cell
circuits, each of the cell circuits including first and second MIS
transistors of same conductive type, coupled between a power source
node and a ground node in series, with each gate and source coupled
together, a logic circuit that receives a potential of a connection
node of the first and the second MIS transistors as an input
signal, and a response signal creating circuit that creates a
response signal of one or some bits, based on challenge signals of
bits and output signals of the logic circuits, wherein the
respective MIS transistors forming the first and the second MIS
transistors and the logic circuit are fin-type FETs, wherein each
of the first and the second MIS transistors is a P-type transistor,
and wherein each back gate of the first and the second MIS
transistors is coupled to the power source node.
13. The semiconductor device according to claim 12, wherein each of
the cell circuits includes first and second fin wirings each
extending in a first direction, formed on an N-type well and
arranged in a second direction crossing the first direction, a
first gate wiring extending in the second direction, coupled to the
first fin wiring through a gate insulating film, a second gate
wiring extending in the second direction, coupled to the second fin
wiring through the gate insulating film, a first local wiring
extending in the second direction, directly coupled to both the
first and the second fin wirings, a second local wiring extending
in the second direction, directly coupled to the first fin wiring
at a position opposite to the first local wiring with the first
gate wiring interposed therebetween and coupled to a power source
wiring in an upper layer, a third local wiring extending in the
second direction, directly coupled to the second fin wiring at a
position opposite to the first local wiring with the second gate
wiring interposed therebetween and coupled to a ground wiring in
the upper layer, a fourth local wiring directly coupled to the
first gate wiring and the second local wiring, and a fifth local
wiring directly coupled to the second gate wiring and the first
local wiring, wherein the first fin wiring and the first gate
wiring form the first MIS transistor, and wherein the second fin
wiring and the second gate wiring form the second MIS
transistor.
14. The semiconductor device according to claim 12, wherein the
response signal creating circuit includes a selector circuit that
outputs one or some output signals selected from the output signals
of the logic circuits as the response signal, according to the
challenge signal.
15. The semiconductor device according to claim 12, comprising N
pieces (N is even number) of the cell circuits, wherein the
response signal creating circuit includes N/2 pieces of selectors
each corresponding to the two individual cell circuits, and wherein
each of the selectors outputs an output signal from one of the
corresponding two cell circuits as a value of bit corresponding to
the response signal, according to a value of the challenge signal
of the corresponding bit of the N/2 bits of the challenge
signals.
16. The semiconductor device according to claim 12, wherein the
logic circuit includes an inverter that inverts a logic level of
the input signal.
Description
CROSS-REFERENCE TO RELATED APPLICATIONS
[0001] The disclosure of Japanese Patent Application No.
2015-186993 filed on Sep. 24, 2015 including the specification,
drawings and abstract is incorporated herein by reference in its
entirety.
BACKGROUND
[0002] This invention relates to a semiconductor device and for
example, it is preferably used for a semiconductor device with a
Physical Unclonable Function (PUF) mounted.
[0003] The Physical Unclonable Function (PUF) is an artificial
object of outputting intrinsic value (response signal) in reply to
a given input (challenge signal), by using an uncontrollable small
difference of intrinsic physical properties. By using the PUF
property, a cipher and authentication function can be realized at a
low cost.
[0004] In the PUF (hereinafter, referred to as "silicon PUF") using
a semiconductor element, by using variations of the properties in
the individual semiconductor elements, intrinsic identification
information can be output from the semiconductor circuits of the
same design.
[0005] As a concrete silicon PUF, Arbiter PUF, Ring Oscillator PUF,
and Butterfly PUF are well known. The Arbiter PUF uses variations
of signal delays in a signal propagation route using semiconductor
circuits. The Ring Oscillator PUF uses variations of oscillation
frequencies in ring oscillators. The Butterfly PUF uses variations
in the initial state of a latch circuit at the power on (for
example, refer to FIGS. 4 to 6 and the related article of Japanese
Unexamined Patent Application Publication No. 2011-198317).
SUMMARY
[0006] One of the problems in the silicon PUF of the related art is
that the power consumption is apt to comparatively get larger. For
example, the above Arbiter PUF consumes power for signal
propagation. The Ring Oscillator PUF consumes power for oscillation
of a ring oscillator. When the power consumption of PUF gets too
large, its mounting on a small circuit such as an Integrated
Circuits (IC) tag becomes difficult; therefore, a PUF technique
with smaller power consumption is desired.
[0007] Other problems and novel features will be apparent from the
description of the specification and the attached drawings.
[0008] A semiconductor device according to one embodiment includes
first and second Metal Insulator Semiconductor (MIS) transistors
MN1 and MN2 of the same conductive type in off-state coupled in
series, as a PUF element CELL. The PUF element CELL outputs a
signal of high level or low level depending on the potential of a
connection node of the first and the second MIS transistors MN1 and
MN2. Preferably, the MIS transistors MN1 and MN2 are fin-type Field
Effect Transistors (FET).
[0009] According to the above embodiment, it is possible to realize
a silicon PUF of lower power consumption.
BRIEF DESCRIPTION OF THE DRAWINGS
[0010] FIG. 1 is a block diagram showing an example of a structure
of an authentication system according to a first embodiment.
[0011] FIG. 2 is a block diagram showing an example of the detailed
structure of the authentication circuit of FIG. 1.
[0012] FIG. 3 is a circuit diagram showing one example of the
structure of each cell circuit of FIG. 2.
[0013] FIG. 4 is a top plan view showing the layout of N-type MIS
transistor formed as a Fin FET on a semiconductor substrate.
[0014] FIG. 5 is a cross-sectional view taken along a cutting line
V-V in FIG. 4.
[0015] FIG. 6 is a cross-sectional view taken along a cutting line
VI-VI in FIG. 4.
[0016] FIG. 7 is a top plan view showing the layout of the cell
circuit of FIG. 3 on a semiconductor substrate.
[0017] FIG. 8 is a cross-sectional view taken along a cutting line
VIII-VIII in FIG. 7.
[0018] FIG. 9 is a cross-sectional view taken along a cutting line
IX-IX in FIG. 7.
[0019] FIG. 10 is a top plan view showing a first modified example
of the layout of FIG. 7.
[0020] FIG. 11 is a top plan view showing a second modified example
of the layout of FIG. 7.
[0021] FIGS. 12A and 12B are views for use in describing variations
of properties in the MIS transistors caused by the manufacturing
process.
[0022] FIG. 13 is a circuit diagram of a cell circuit forming a PUF
circuit in a semiconductor device according to a second
embodiment.
[0023] FIG. 14 is a top plan view showing the layout of the cell
circuit of FIG. 13 on a semiconductor substrate.
[0024] FIG. 15 is a top plan view showing a modified example of the
layout of FIG. 14.
[0025] FIG. 16 is a block diagram showing a structure of an
authentication circuit in a semiconductor device according to a
third embodiment.
[0026] FIG. 17 is a block diagram schematically showing a structure
of an automatic driving system.
DETAILED DESCRIPTION
[0027] Hereinafter, each embodiment will be described in detail
with reference to the drawings. The top plan views and
cross-sectional views used in the following description are not in
proportion to the actual measurement of each unit, for the sake of
easy understanding. In the description of each drawing, the same
reference codes are attached to the same or corresponding portions
and their description may not be repeated.
First Embodiment
[Structure of Authentication System]
[0028] FIG. 1 is a block diagram showing an example of a structure
of an authentication system according to a first embodiment. In an
authentication system 1 of FIG. 1, a computer system 20 determines
whether or not a semiconductor device 10 is a regular product.
[0029] As illustrated in FIG. 1, the semiconductor device 10
includes an authentication circuit 11 for realizing an
authentication function with the computer system 20 and a function
circuit 12 for achieving a specified function as an integrated
circuit. The semiconductor device 10 may be considered to be formed
by adding the authentication circuit 11 to the original function
circuit 12.
[0030] The semiconductor device 10 and the computer system 20 can
communicate with each other. The computer system 20 can be
considered, for example, as a server for providing an
authentication service, provided on a network.
[0031] The authentication circuit 11 creates a response signal RES
of one or plural bits based on a challenge signal CHAL of plural
bits received from the computer system 20 and outputs the created
response signal RES to the computer system 20. More specifically,
the authentication circuit 11 includes a Physical Unclonable
Function (PUF) circuit 30 and a response signal creating circuit
31. The PUF circuit 30 creates an intrinsic signal of N bits for
every semiconductor device 10. The concrete structure example of
the PUF circuit 30 used for the embodiment will be described in
FIGS. 2 to 9.
[0032] The response signal creating circuit 31 creates the response
signal RES of one or plural bits, based on the challenge signal
CHAL of plural bits and the output signal (N bits) of the PUF
circuit 30. The response signal creating circuit 31 may be, for
example, a selector circuit of selecting a part of bit or bits from
the signal of N bits output from the PUF circuit 30 according to
the challenge signal CHAL. In the embodiment, an example of forming
the response signal creating circuit 31 by the selector circuit is
illustrated in FIG. 3.
[0033] As another example, the response signal creating circuit 31
may be formed by a combined logic circuit of receiving the
challenge signal CHAL ad the output signal of the PUF circuit 30.
The output signal of the combined logic circuit is output to the
computer system 20 as the response signal RES. Alternatively, the
response signal creating circuit 31 may be regarded as a cipher
function and the output signal of the PUF circuit 30 may be
regarded as a cipher key. In this case, the challenge signal CHAL
is input from the computer system 20 as a plaintext, while in
replay to that, the response signal RES is output from the response
signal creating circuit 31 as a ciphertext. As the response signal
creating circuit 31, various circuit structures are possible other
than the above.
[0034] The computer system 20 creates the challenge signal CHAL and
outputs it to the authentication circuit 11; in reply to the
challenge signal CHAL, the computer system 20 receives the response
signal RES output from the authentication circuit 11. The computer
system 20 determines whether or not the received response signal
RES is the proper one corresponding to the challenge signal CHAL;
in the case of the proper response signal RES, it permits the
function circuit 12 to operate for the other circuits. For example,
various types of data can be exchanged between the computer system
20 and the function circuit 12 of the semiconductor device 10.
[0035] Authentication operation in the computer system 20 sometimes
requires a large size of database and does not so in other case.
Specifically, when the circuit structure of the response signal
creating circuit 31 is unclear (or when it cannot be represented as
a logical function), the response signals RES corresponding to a
lot of challenge signals CHAL are previously measured and the
measurement result of this large amount has to be stored in a
database with correlation to the individual challenge signals CHAL.
In this case, the computer system 20 selects one of the registered
challenge signals CHAL and outputs it to the authentication circuit
11, then to perform authentication by checking whether or not the
response signal RES received from the authentication circuit 11
agrees with the corresponding response signal RES stored in the
database.
[0036] When the structure of the response signal creating circuit
31 can be represented as the logical function of using the
challenge signal CHAL and the output signal of the PUF circuit 30
as a variable, it does not require a large capacity of a database.
In this case, in the computer system 20, the data representing the
output signals of the PUF circuit 30 measured previously and the
above logical function are stored. At the authentication, the
computer system 20 creates a challenge signal CHAL and outputs it
to the authentication circuit 11, and receives a response signal
RES from the authentication circuit 11. The computer system 20
substitutes the created challenge signal CHAL and the output signal
of the regular PUF circuit 30 measured previously into the above
logical function, hence to create a regular response signal RES
internally. The computer system 20 performs authentication by
checking whether or not the received response signal RES agrees
with the internally created regular response signal RES.
[Structure Example of PUF Circuit]
[0037] FIG. 2 is a block diagram showing an example of the detailed
structure of the authentication circuit of FIG. 1. With reference
to FIG. 2, the authentication circuit 11 includes the PUF circuit
30 and a selector circuit 31A as the response signal creating
circuit 31.
[0038] The PUF circuit 30 includes N pieces of cell circuits CELL_0
to CELL_N-1 (also referred to as PUF element). Each cell circuit
CELL outputs a signal OUT of high level (H level) or low level (L
level). A logic level of the output signal OUT of each cell circuit
CELL is determined according to the quality variations in the
semiconductor elements forming the respective cell circuits CELL.
As specifically described in FIG. 3, each cell circuit CELL
includes N-type Metal Insulator Semiconductor (MIS) transistors MN1
and MN2 in off-state, coupled between a power source node and a
ground node in series and an inverter INV1 which inverts the logic
level of the connection node of the MIS transistors MN1 and MN2 and
outputs the same. The output signal OUT of the inverter INV1 is
input to the selector circuit 31A.
[0039] The selector circuit 31A outputs one or plural bits selected
according to the challenge signal CHAL from the output signals
OUT_0 to OUT_N-1 of the N pieces of the cell circuit CELL_0 to
CELL_N-1. The structure of the selector circuit 31A may be
variously considered. For example, by specifying a certain M bit
(1.ltoreq.M<N), of the N bits, the selector circuit 31A may be
formed to output the output signal OUT of the cell circuit CELL
corresponding to the specified M bit as the response signal RES,
according to the challenge signal CHAL of the N bits.
Alternatively, it may be formed to output the exclusive OR of the
output signal OUT of the cell circuit CELL corresponding to the
specified M bit as the response signal RES.
[Structure of Cell Circuit]
[0040] FIG. 3 is a circuit diagram showing one example of the
structure of each cell circuit of FIG. 2. With reference to FIG. 3,
the cell circuit includes N-type Metal Insulator Semiconductor
(MIS) transistors MN1, MN2, and MN3 and P-type MIS transistor MP1.
Generally, since an oxide film (Oxide) is used as an insulating
film of the MIS transistor, the MIS transistors are often referred
to as Metal Oxide Semiconductor (MOS) transistors.
[0041] As illustrated in FIG. 3, the N-type MIS transistors MN1 and
MN2 are coupled between the power source node VDD and the ground
node VSS in series in this order from the top of the figure. The
P-type MIS transistor MP1 and the N-type MIS transistor MN3 are
coupled between the power source node VDD and the ground node VSS
in series in this order and in parallel with the MIS transistor MN1
and MN2 on the whole. The gates of the MIS transistors MP1 and MN3
are mutually coupled together and therefore, the MIS transistors
MP1 and MN3 form the inverter INV1. The output signal OUT is output
from the connection node N2 of the MIS transistors MP1 and MN3.
[0042] Instead of the above inverter INV1, a NAND gate (one input
is fixed at H level) or NOR gate (one input is fixed at L level)
may be used. More generally, the logic circuit may be whatever
structure as far as it outputs a signal of H lever or L level
depending on whether the potential of the connection node N1 of the
MIS transistors MN1 and MN2 is higher or lower than Vd/2 (Vd is the
power potential).
[0043] The N-type MIS transistors MN1 and MN2 are in the off-state
with the gate and the source (low potential side) coupled together.
The connection node N1 of the MIS transistors MN1 and MN2 is
coupled to the gate of the P-type MIS transistor MP1 and the N-type
MIS transistor MN3. The back gates (also referred to as P well,
substrate, or body) of the N-type MIS transistors MN1, MN2, and MN3
are coupled to the ground node VSS and the back gate (referred to
as N well or body) of the P-type MIS transistor MP1 is coupled to
the power source node VDD.
[Operation of Cell Circuit]
[0044] Next, the operation of the above cell circuit CELL will be
described. Since the N-type MIS transistors MN1 and MN2 are both in
off-state, the potential of the connection node N1 is determined by
a resistance ratio of the MIS transistors MN1 and MN2. The
resistance ratio of the MIS transistors MN1 and MN2 is determined
by the ratio of off-leak current. Therefore, when the off-leak
current is equal between the MIS transistors MN1 and MN2, the
potential of the power source node VDD is defined as Vd, and the
potential of the ground node VSS is defined as 0, the potential of
the connection node N1 becomes Vd/2.
[0045] Actually, however, there exist variations caused by the
manufacturing process in every MIS transistor; hence, a possibility
that the off-leak current of the MIS transistor MN1 is completely
equal to that of the MIS transistor MN2 is really low. For example,
when the off-leak current of the MIS transistor MN1 is larger than
that of the MIS transistor MN2, the potential of the connection
node N1 is larger than Vd/2. Therefore, the potential of the signal
OUT output from the node N2 gets substantially equal to 0 (L
level). On the contrary, when the off-leak current of the MIS
transistor MN1 is smaller than that of the MIS transistor MN2, the
potential of the connection node N1 is smaller than Vd/2.
Therefore, the potential of the signal OUT output from the node N2
gets substantially equal to Vd (H level).
[0046] This output signal OUT of each cell circuit CELL is
determined to be H level or L level at random because of the
property variations in the MIS transistors caused by the
manufacturing process. Accordingly, each cell circuit CELL can be
used as a PUF element.
[0047] The above circuit operation is true when the substrate bias
effect of the N-type MIS transistor MN1 hardly occurs. In the case
of the circuit structure of FIG. 3, the source potential of the MIS
transistor MN1 (the potential of the node N1) is lower than the
potential of the back gate (the potential of the ground node VSS).
Therefore, when each transistor forming the cell circuit CELL is
manufactured in the planar structure of the related art, the back
gate potential affects spreading of a depletion layer at off-time.
Specifically, in order to spread the depletion layer of the MIS
transistor MN1 more, the threshold voltage of the MIS transistor
MN1 increases and accordingly, the off-leak current of the MIS
transistor MN1 decreases. As the result, since the potential of the
node N1 becomes smaller than Vd/2, the outputs of many cell
circuits CELL are fixed at H level. In short, the average value of
the output voltages of the cell circuits CELL_0 to CELL_N-1 gets
close to the power voltage Vd.
[0048] On the contrary, in the case of the MIS transistor called
FinFET (fin-type Field Effect Transistor (FET)), the depletion
layer at the off-time spreads to the whole fin near the gate
electrode from the viewpoint of the structure and therefore, the
thickness of the depletion layer hardly changes by the effect of
the back gate potential. As the result, it is hardly subject to the
substrate bias effect, and the average value of the output voltages
of the cell circuits CELL_0 to CELL_N-1 becomes substantially equal
to Vd/2 (the output of the H level and the output of the L level
occur at random). Therefore, as the MIS transistor used for the
above cell circuit CELL, FinFET is preferably used.
[0049] Specifically, in the case of the planar type MOS transistor
of the related art (in the case of the process with the gate length
28 nm), when the power voltage is 1 V and the substrate bias is 0.4
V, a fluctuation in the threshold voltage is about 50 mV. On the
contrary, in the case of the FinFET of the same transistor size,
when the power voltage is 1 V and the substrate bias is 0.4 V, a
fluctuation in the threshold voltage is fairly small, 10 mV and
less, which can be neglected, differently from the variations in
the threshold voltage caused by the manufacturing process.
[0050] In FIG. 3, when the back gate of the N-type MIS transistor
MN1 is coupled to the source thereof, the substrate bias effect
does not occur and therefore, the MOS transistor in the planar
structure of the related art can be used. In this case, however,
the MIS transistor MN1 has to be formed within a triple well, which
makes the element structure complicated disadvantageously.
[Structure of FinFET]
[0051] Hereinafter, a structure and a manufacturing method of the
FinFET will be described in brief.
[0052] FIG. 4 is a top plan view showing the layout of the N-type
MIS transistor formed as the FinFET on a semiconductor substrate.
FIG. 5 is a cross-sectional view taken along a cutting line V-V of
FIG. 4. FIG. 6 is a cross-sectional view taken along a cutting line
VI-VI of FIG. 4.
[0053] With reference to FIGS. 4 to 6, the N-type MIS transistor
formed as the FinFET includes a fin wiring 40 formed on a P-type
well PW. The fin wiring 40 extends along the x-direction on the
substrate and protrudes in the z-direction (substrate vertical
direction). The fin wiring 40 is formed by selectively etching the
surface of the P-type well PW of the semiconductor substrate SUB.
As illustrated in FIG. 6, a silicon oxide film formed according to
a Chemical Vapor Deposition (CVD) method is provided in a space
between the adjacent fin wirings (in a portion without the fin
wiring 40) as an isolation film 92.
[0054] A gate wiring 50 as a gate electrode G is formed to cover
the top surface and the lateral surface of the fin wiring 40 with
the gate insulating film (not illustrated) intervening
therebetween. The gate wiring 50 extends in the y-direction that is
a direction of intersecting the fin wiring 40. For example,
semiconductor such as polycrystalline silicon, conductive compound
such as titanium nitride, simple metal material such as tungsten,
or a stacked film of one of them is used for the gate wiring
50.
[0055] After forming the gate wiring 50, dopant is injected into
the fin wiring 40 with the gate wiring 50 used as a mask, hence to
form a source region S and drain region D in a portion other than
the channel region surrounded by the gate wiring 50. Here, when
manufacturing an N-type MIS transistor, the fin wiring 40 is formed
on a P-type substrate or a P-type well and at the same time, N-type
dopant is injected into the fin wiring 40. On the other hand, when
manufacturing a P-type MIS transistor, the fin wiring 40 is formed
on an N-type well and at the same time, P-type dopant is injected
into the fin wiring 40.
[0056] Of the fin wiring 40, a portion where the dopant is not
injected in the dopant injection process, or a portion near the
gate electrode G is used as a channel region in the on-state and
the above becomes a depletion layer DL in the off-state. As
illustrated in FIGS. 5 and 6, the depletion layer DL expands to the
whole fin in the vicinity of the gate electrode G. In FIGS. 5 and
6, an area 90 indicated by a dotted circle is the position for
determining a threshold voltage. It is in the boundary between the
source region and the channel region and a charge amount flowing
from the source region to the channel region decides the threshold
voltage. In the FinFET, since the state of the area 90 hardly
fluctuates by the substrate bias, the threshold voltage has smaller
dependence on the substrate bias.
[0057] A first and a second local wirings (LIC: Local
Inter-Connect) not illustrated are formed using a simple metal
material, for example, tungsten, to be in ohmic contact with the
upper surface and the lateral surface of the source region S and
drain region D. These first and second local wirings (not
illustrated) work as a source electrode and a drain electrode
respectively. The gate wiring 50 and the above local wirings are
further coupled to another local wiring (not illustrated), or the
upper metal wiring layer (not illustrated) through a via hole
formed in the interlayer insulating layer not illustrated.
[0058] As illustrated in FIG. 5, another fin wiring 41 formed on
the P-type well PW is used in order to apply a voltage to the
P-type well PW corresponding to the back gate B. A silicon oxide
film is formed between the fin wirings 40 and 41 as an isolation
film 91.
[Layout Example of Cell Circuit]
[0059] Hereinafter, the layout of the cell circuit CELL of FIG. 3
on a semiconductor substrate will be described.
[0060] FIG. 7 is a top plan view showing the layout of the cell
circuit of FIG. 3 on a semiconductor substrate. FIG. 8 is a
cross-sectional view taken along a cutting line VIII-VIII of FIG.
7. FIG. 9 is a cross-sectional view taken along a cutting line
IX-IX of FIG. 7. In FIGS. 7 to 9, a direction inside the substrate
is defined as the x direction and the y direction, and the
direction vertical to the substrate is defined as the z direction.
In FIG. 7, a first metal wiring layer formed by interposing the MIS
transistors formed on the substrate surface and the interlayer
insulating layer (not illustrated) is indicated by a dashed line,
and a second metal wiring layer formed by interposing the first
metal wiring layer and the interlayer insulating layer (not
illustrated) is indicated by a chain double-dashed line.
[0061] With reference to FIGS. 7 to 9, the cell circuit CELL
includes a plurality of fin wirings 40, 42, 43, and 44 extending in
the x direction and a plurality of gate wirings 50, 51, 52, 53, and
54 extending in the y direction crossing the x direction. As
illustrated in FIG. 8, an insulating film 93 for element isolation
is formed on the lateral surfaces of the respective gate wirings 50
to 54. As illustrated in FIG. 9, an insulating film 92 for element
isolation is formed in a space between the adjacent fin wirings (a
portion where the fin wirings 40, 42 to 44 are not formed).
[0062] The fin wirings 40, 42, 43, and 44 are formed in the y
direction in this order. The fin wirings 40, 42, and 43 are formed
on the P-type well PW and the fin wiring 44 is formed on the N-type
well NW. The gate wiring 50 crosses the fin wiring 40 in plane view
and coupled to the top surface and the lateral surface of the fin
wiring 40 through a gate insulating film (not illustrated). The
gate wiring 51 crosses the fin wiring 42 in plane view and coupled
to the top surface and the lateral surface of the fin wiring 40
through the gate insulating film (not illustrated). The gate wiring
52 crosses the both fin wirings 43 and 44 and coupled to the
respective top surfaces and lateral surfaces of the respective fin
wirings 43 and 44 through the gate insulating film (not
illustrated). The gate wirings 53 and 54 are dummy wirings formed
in order to secure the pattern precision of lithography.
[0063] The N-type MIS transistor MN1 of FIG. 2 is formed by the fin
wiring 42 and the gate wiring 51. The N-type MIS transistor MN2 is
formed by the fin wiring 40 and the gate wiring 50. The N-type MIS
transistor MN3 is formed by the fin wiring 43 and the gate wiring
52. The P-type MIS transistor MP1 is formed by the fin wiring 44
and the gate wiring 52. The gate wiring 52 is shared between the
MIS transistors MN3 and MP1.
[0064] The cell circuit CELL further includes a plurality of local
wirings 60, 62, 64, 65, 66, and 67 extending in the y direction and
a plurality of local wirings 61 and 63 for each coupling the gate
and the source of the MIS transistor.
[0065] The local wiring 60 is used as the source electrode of the
MIS transistor MN2 by being coupled directly to the source region
of the fin wiring 40. Here, directly coupling means that it is
coupled not through the contact hole formed in the interlayer
insulating layer (it is true to in the following description). The
local wiring 60 is coupled to the metal wiring 70 (corresponding to
the ground wiring and ground node VSS) extending in the x
direction, provided in the first metal wiring layer through a
contact hole 110. The local wiring 60 is further coupled directly
to the gate wiring 50 through the local wiring 61.
[0066] The local wiring 62 is used as the drain electrode of the
MIS transistor MN2 by being coupled directly to the drain region of
the fin wiring 40 and at the same time, used as the source
electrode of the MIS transistor MN1 by being coupled directly to
the source region of the fin wiring 42. The local wiring 62
corresponds to the connection node N1 of FIG. 2. The local wiring
62 is further coupled directly to the gate wiring 51 through the
local wiring 63.
[0067] The local wiring 64 is used as the drain electrode of the
MIS transistor MN1 by being coupled directly to the drain region of
the fin wiring 42. The local wiring 64 is coupled to the metal
wiring 71 (corresponding to the power wiring and power source node
VDD), extending in the x direction, provided in the first metal
wiring layer through a contact hole 112.
[0068] The local wiring 65 is used as the source electrode of the
MIS transistor MN3 by being coupled directly to the source region
of the fin wiring 43. The local wiring 65 is coupled to the metal
wiring 72 (corresponding to the ground wiring and ground node VSS)
extending in the x direction, provided in the first metal wiring
layer through a contact hole 113.
[0069] The local wiring 66 is used as the source electrode of the
P-type MIS transistor MP1 by being coupled directly to the source
region of the fin wiring 44. The local wiring 66 is coupled to the
metal wiring 73 (corresponding to the power source wiring and power
source node VDD) extending in the x direction, provided in the
first metal wiring layer through a contact hole 114.
[0070] The local wiring 67 is used as the drain electrode of the
MIS transistor MN3 by being coupled directly to the drain region of
the fin wiring 43 and at the same time, it is used as the drain
electrode of the MIS transistor MP1 by being coupled directly to
the drain region of the fin wiring 44. The local wiring 67
corresponds to the connection node N2 of FIG. 2. The local wiring
67 is further coupled to the metal wiring 76 provided on the first
metal wiring layer through a contact hole 116, hence to output the
output signal OUT to the metal wiring 76.
[0071] A metal wiring 80, extending in the y direction, provided in
the second metal wiring layer is provided in order to couple the
local wiring 62 (connection node N1) and the gate wiring 52 in a
direct current way. Specifically, the local wiring 62 is coupled to
the metal wiring 74 provided in the first metal wiring layer
through a contact hole 111. The gate wiring 52 is coupled to the
metal wiring 75 provided in the first metal wiring layer through a
contact hole 115. The metal wiring 80 is coupled to the metal
wiring 74 through a contact hole 130 and at the same time, coupled
to the metal wiring 75 through a contact hole 131.
[Modified Example 1 of Layout of Cell Circuit]
[0072] FIG. 10 is a top plan view showing a first modified example
of the layout of FIG. 7. The layout of FIG. 10 is different from
the layout of FIG. 7 in that metal wirings 77 and 78 are provided
in the first metal wiring layer, instead of the local wirings 61
and 63.
[0073] Specifically, referring to FIG. 10, the metal wiring 77 is
coupled to the local wiring 60 as the source electrode through a
contact hole 117 and at the same time, coupled to the gate wiring
50 through a contact hole 118. The metal wiring 78 is coupled to
the local wiring 62 (node N1) as the source electrode through a
contact hole 119 and at the same time, coupled to the gate wiring
51 through a contact hole 120. Further, the metal wiring 80 formed
in the second metal wiring layer is coupled to the metal wiring 78
through the contact hole 130. Therefore, the metal wiring 74 of
FIG. 7 is not provided in the layout of FIG. 10.
[0074] The other points of FIG. 10 are the same as those of FIG. 7
and therefore, the same reference codes are attached to the same or
corresponding portions without repeating the description thereof.
The layout of FIG. 10 is useful for the case of supporting only one
layer of local wiring because of the restriction of the layout.
[Modified Example 2 of Layout of Cell Circuit]
[0075] FIG. 11 is a top plan view showing a second modified example
of the layout of FIG. 7. The layout of FIG. 11 is different from
that of FIG. 7 in the position of the N-type well NW.
[0076] Specifically, referring to FIG. 11, the N-type well NW is
interposed between the first P-type well PW1 and the second P-type
well PW2. The fin wirings 40 and 42 are formed on the first P-type
well PW1, the fin wiring 43 is formed on the N-type well NW, and
the fin wiring 44 is formed on the second P-type well PW2. It is
the same as the case of FIG. 7 in that the ground potential (VSS)
is given to the metal wiring 70 and that the power potential (VDD)
is given to the metal wiring 71. On the other hand, the power
potential (VDD) is given to the metal wiring 72 and the ground
potential (VSS) is given to the metal wiring 73, contrary to the
case of FIG. 7.
[0077] According to the above change, in the case of FIG. 11, the
P-type MIS transistor MP1 of FIG. 3 is formed by the fin wiring 43
and the gate wiring 52. The N-type MIS transistor MN3 is formed by
the fin wiring 44 and the gate wiring 52. The local wiring 65 forms
the source electrode of the P-type MIS transistor MP1 and the local
wiring 66 forms the source electrode of the N-type MIS transistor
MN3.
[0078] Since the other points of FIG. 11 are the same as in the
case of FIG. 7, the same reference codes are attached to the same
or corresponding portions without repeating the description
thereof.
Effect of First Embodiment
[0079] Hereinafter, the effect of the PUF circuit constituted in
the above will be described. FIGS. 12A and 12B are views for use in
describing property variations in the MIS transistor caused by the
manufacturing process.
[0080] FIG. 12A is a graph schematically illustrating a variation
degree of a threshold voltage Vth; the horizontal axis of FIG. 12A
indicates the threshold voltage Vth and the vertical axis indicates
the frequency. As illustrated in FIG. 12A, the threshold voltage
Vth of the MIS transistor is scattered in a normal distribution
shape, caused by the manufacturing process. FIG. 12B is a graph
indicating a relation between a leak current and the threshold
voltage Vth. The horizontal axis of FIG. 12B indicates the
threshold voltage Vth and the vertical axis indicates the volume of
the leak current in a log scale. As illustrated in FIG. 12B, the
leak current increases and decreases exponentially according to the
fluctuation of the threshold voltage Vth.
[0081] As mentioned above, the potential of the connection node N1
in FIG. 2 changes exponentially according to the change of the
threshold voltage Vth. This is effective in making the output
signal from the cell circuit CELL stable. When the potential of the
node N1 is close to Vd/2, the logic level of the output signal OUT
may fluctuate depending on the use environment and the operation
may be unstable. As mentioned above, the size of the off-leak
current fluctuates exponentially according to the change of the
threshold voltage Vth, hence to reduce a possibility of the
potential of the node N1 getting close to Vd/2. As the result,
according to the embodiment, the output of the PUF circuit can be
stabilized.
[0082] Further, since the current flowing in each cell circuit (PUF
element) is only the off-leak current, power consumption can be
reduced. After the power on, the output signals of the respective
cell circuits (PUF elements) are immediately fixed at H level or L
level and become readable, which enables a high speed operation of
the PUF circuits. In order to increase the number of the output
bits of the PUF circuits, the number of the cell circuits (PUF
elements) has to be increased.
Second Embodiment
[0083] In a semiconductor device of a second embodiment, an example
of changing the conductive type of the MIS transistor forming each
cell circuit (PUF element) will be described.
[Structure of Cell Circuit]
[0084] FIG. 13 is a circuit diagram of a cell circuit forming the
PUF circuit in the semiconductor device according to the second
embodiment. With reference to FIG. 13, the cell circuit includes
P-type MIS transistors MP1, MP2, and MP3 and an N-type MIS
transistor MN1.
[0085] The P-type MIS transistors MP1 and MP2 are coupled between
the power source node VDD and the ground node VSS in series in this
order. The P-type MIS transistor MP3 and the N-type MIS transistor
MN1 are coupled between the power source node VDD and the ground
node VSS in series in this order and in parallel with the MIS
transistors MP1 and MP2 on the whole. The gates of the MIS
transistors MP3 and MN1 are mutually coupled and therefore, the MIS
transistors MP3 and MN1 form the inverter INV1. The output signal
OUT is output from the connection node N2 of the MIS transistors
MP3 and MN1.
[0086] Instead of the above inverter INV1, a NAND gate (one input
is fixed at H level) or NOR gate (one input is fixed at L level)
may be used. More generally, the logic circuit may be whatever
structure as far as it outputs a signal of H lever or L level
depending on whether the potential of the connection node N1 of the
MIS transistors MP1 and MP2 is higher or lower than Vd/2 (Vd is the
power potential).
[0087] Each of the P-type MIS transistors MP1 and MP2 is in the
off-state with the gate and the source (high potential side)
coupled together. The connection node N1 of the MIS transistors MP1
and MP2 is coupled to the gate of the P-type MIS transistor MP3 and
N-type MIS transistor MN1. The back gates (also referred to as N
well, substrate, or body) of the P-type MIS transistors MP1, MP2,
and MP3 are coupled to the power source node VDD and the back gate
(referred to as P well or body) of the N-type MIS transistor MN1 is
coupled to the ground node VSS.
[Operation of Cell Circuit]
[0088] The operation of the cell circuit of FIG. 13 is the same as
in the case of FIG. 3; therefore, it will be briefly summarized in
the below. In FIG. 13, since the P-type MIS transistors MP1 and MP2
are both in the off-state, the potential of the connection node N1
is determined by a ratio of the off-leak currents of the MIS
transistors MP1 and MP2. Since the off-leak current ratio of the
MIS transistors MP1 and MP2 fluctuates at random, caused by the
manufacturing process, the cell circuits can be used as the PUF
element.
[0089] Also in the case of the circuit of FIG. 13, since the source
potential (the potential of the connection node N1) of the P-type
MIS transistor MP2 is lower than the potential of the back gate
(the potential of the power source node VDD), the substrate bias
effect of the P-type MIS transistor MP2 can be a problem. However,
the MIS transistors MP1, MP2, MP3, and MN1 are formed by the
FinFET, hence to be free from the substrate bias effect.
[0090] On the other hand, when the respective MIS transistors are
formed by the planar type MOS transistors of the related art, the
back gate of the MIS transistor MP2 is coupled to the source (the
connection node N1), hence to be free from the substrate bias
effect. In this case, however, the MIS transistor MP2 has to be
formed within a triple well.
[Example of Layout of Cell Circuit]
[0091] Hereinafter, the layout of the cell circuit CELL of FIG. 13
on a semiconductor substrate will be described.
[0092] FIG. 14 is a top plan view showing the layout of the cell
circuit of FIG. 13 on a semiconductor substrate. Since the layout
of the cell circuit CELL of FIG. 13 illustrated in FIG. 1 is
similar to the layout of the cell circuit of FIG. 3 described in
FIG. 7, only the different point from FIG. 7 will be mainly
described in the below.
[0093] In the case of FIG. 14, the fin wirings 40, 42, and 43 are
formed on the N-type well NW and the fin wiring 44 is formed on the
P-type well PW. The P-type MIS transistor MP1 of FIG. 13 is formed
by the fin wiring 42 and the gate wiring 51. The P-type MIS
transistor MP2 is formed by the fin wiring 40 and the gate wiring
50. The P-type MIS transistor MP3 is formed by the fin wiring 43
and the gate wiring 52. The N-type MIS transistor MN1 is formed by
the fin wiring 44 and the gate wiring 52. The gate wiring 52 is
shared between the MIS transistors MP3 and MN1.
[0094] In the case of FIG. 14, the ground potential (VSS) is given
to the metal wiring 70 and the power potential (VDD) is given to
the metal wiring 71, similarly to the case of FIG. 7. On the other
hand, the power potential (VDD) is given to the metal wiring 72 and
the ground potential (VSS) is given to the metal wiring 73,
contrary to the case of FIG. 7.
[0095] Since the source is on the side of high potential and the
drain is on the side of low potential in the P-type MIS transistor,
the local wiring 60 forms the drain electrode of the MIS transistor
MP2 in FIG. 14. Similarly, the local wiring 62 forms the source
electrode of the MIS transistor MP2 and at the same time, forms the
drain electrode of the MIS transistor MP1. The local wiring 64
forms the source electrode of the MIS transistor MP2. The local
wiring 65 forms the source electrode of the P-type MIS transistor
MP3. The local wiring 66 forms the source electrode of the N-type
MIS transistor MN1. The local wiring 67 forms the drain electrode
of the P-type MIS transistor MP3 and at the same time, forms the
drain electrode of the N-type MIS transistor MN1.
[0096] Further, the arrangement of the local wirings 61 and 63 in
FIG. 14 is different from the case of FIG. 7 (opposite to FIG. 7
with the gate wiring intervening therebetween). Specifically, the
local wiring 61 couples the gate wiring 50 and the local wiring 62
so as to couple the gate and the source (high potential side) of
the P-type MIS transistor MP2. The local wiring 63 couples the gate
wiring 51 and the local wiring 64 so as to couple the gate and the
source (high potential side) of the P-type MIS transistor MP1.
[0097] The other points of FIG. 14 are the same as in the case of
FIG. 7; therefore, the same reference codes are attached to the
same or corresponding portions without repeating the description
thereof.
[Modified Example of Layout of Cell Circuit]
[0098] FIG. 15 is a top plan view showing a modified example of the
layout of FIG. 14. The layout of FIG. 15 is different from that of
FIG. 14 in the arrangement of the P-type well PW.
[0099] Specifically, referring to FIG. 15, the P-type well PW is
interposed between the first N-type well NW1 and the second N-type
well NW2. The fin wirings 40 and 42 are formed on the first N-type
well NW1, the fin wiring 43 is formed on the P-type well PW, and
the fin wiring 44 is formed on the second N-type well NW2. The
ground potential (VSS) is given to the metal wiring 70 and the
power potential (VDD) is given to the metal wiring 71, similarly to
the case of FIG. 14. On the other hand, the ground potential (VSS)
is given to the metal wiring 72 and the power potential (VDD) is
given to the metal wiring 73, contrary to the case of FIG. 14.
[0100] According to the above modification, in the case of FIG. 15,
the N-type MIS transistor MN1 of FIG. 13 is formed by the fin
wiring 43 and the gate wiring 52. The P-type MIS transistor MP3 is
formed by the fin wiring 44 and the gate wiring 52. The local
wiring 65 forms the source electrode of the N-type MIS transistor
MN1 and the local wiring 66 forms the source electrode of the
P-type MIS transistor MP3.
[0101] The other points of FIG. 15 are the same as in the case of
FIG. 14; therefore, the same reference codes are attached to the
same or corresponding portions without repeating the description
thereof.
Effect of Second Embodiment
[0102] According to the second embodiment, it is possible to
stabilize the output of the PUF circuit, reduce the power
consumption, speed up the operation, and reduce the circuit size,
similarly to the first embodiment. Further, when property
variations in the P-type MIS transistors are larger than those in
the N-type MIS transistors caused by the manufacturing process, the
PUF circuit in the semiconductor device according to the second
embodiment is more stable in the output and more effective than
that in the semiconductor device according to the first
embodiment.
Third Embodiment
[0103] A semiconductor device according to a third embodiment is to
illustrate one example of the structure of the response signal
creating circuit 31 having been described in FIG. 2 according to
the first embodiment. Hereinafter, it will be specifically
described with reference to the drawings.
[Structure of Authentication Circuit]
[0104] FIG. 16 is a block diagram showing a structure of an
authentication circuit in a semiconductor device according to a
third embodiment. With reference to FIG. 16, the authentication
circuit 11 includes the PUF circuit 30 and the response signal
creating circuit 31. The PUF circuit 30 includes N pieces of the
cell circuits (PUF elements) CELL_0 to CELL_N-1 (N is even number).
The structure of the cell circuit CELL may be identical to the
structure of FIG. 3 according to the first embodiment or the
structure of FIG. 13 according to the second embodiment.
[0105] The response signal creating circuit 31 includes N/2 pieces
of selector circuits 32_0 to 32_N/2-1. Further, the response signal
creating circuit 31 receives the challenge signals CHAL_0 to
CHAL_N/2-1 of N/2 bits respectively corresponding to N/2 pieces of
the selector circuits 32_0 to 32_N/2-1. The N/2 pieces of the
selector circuits 32_0 to 32_N/2-1 correspond to the selector
circuit 31A of FIG. 2, showing the internal structure of the
selector circuit 31A.
[0106] As illustrated in FIG. 16, each selector circuit 32_i (i=0
to N/2-1) responds to two cell circuits CELL_2i and CELL_2i+1,
receiving the output signals OUT_2i and OUT_2i+1 of the
corresponding cell circuits. The selector circuit 32_i selects one
of the output signals OUT_2i and OUT_2i+1, according to the logic
level of the corresponding challenge signal CHAL_i and outputs the
selected output signal as a response signal RES_i.
Effect of Third Embodiment
[0107] According to the structure of the above response signal
creating circuit 31, the bit number of the response signal RES is
N/2 bits on the whole, and therefore, 2.sup.N/2 types of response
signals RES can be obtained. Only the N/2 pieces of the selector
circuits are required in order to create these response
signals.
[0108] On the other hand, for example, the Ring Oscillator PUF of
the related art (refer to FIG. 5 in Japanese Unexamined Patent
Application Publication No. 2011-198317) needs N pieces of
selectors, N pieces of counter circuits, and N/2 pieces of
comparators for comparing the outputs of the two corresponding
counter circuits, in order to obtain the response signals of N/2
bits. The circuit area increases in order to mount these
circuits.
[0109] As mentioned above, according to the embodiment, it is
possible to realize a silicon PUF capable of outputting a response
signal of N/2 bits in a comparatively small circuit area.
Fourth Embodiment
[0110] In a fourth embodiment, as a concrete example of the
authentication system using the semiconductor device having been
described in the first to the third embodiments, an automatic
operation system will be described.
[0111] FIG. 17 is a block diagram schematically showing the
structure of the automatic operation system. The automatic
operation system 200 of FIG. 17 includes a controller A (201) for
controlling various types of electronic control units mounted in an
automobile, in order to do automatic driving. In FIG. 17, as an
example of the electronic control unit, a brake control unit 202
for controlling a brake system 203 is representatively shown.
[0112] The controller A (201) contains the authentication circuit
11 having been described in the first to the third embodiments.
Combinations of various types of challenge signals and the
corresponding response signals from the controller A (201),
measured before the shipment of the controller A (201) are stored
in a database system 205. The database system 205 of FIG. 17
corresponds to the computer system 20 of FIG. 1.
[0113] At an actually using time, various types of electronic
control units such as the brake control unit 202 of FIG. 17 mounted
in an automobile inquire of the database system 205 provided
outside of the automobile through a network, about whether the
controller A (201) for automatic operation is a regular product or
not. The database system 201 determines whether or not the
controller A (201) is the regular product. Only when the controller
A (201) is determined as the regular product, the electronic
control units are configured to follow the instruction from the
controller A (201); otherwise, they reject the instruction from the
controller A (201). Hereinafter, the procedure of the above
authentication at the actually using time will be described in
details with reference to FIG. 17.
[0114] At first, when the controller A (201) intends to drive the
brake system 203 of the automobile, it transmits a driving
instruction of the brake system 203 to the brake control unit 202
(Step S1A). Upon receipt of the driving instruction from the
controller A (201), the brake control unit 202 inquires of the
database system 205 provided outside of the automobile about
whether the controller A is the regular product or not (Step
S2).
[0115] Upon receipt of the above inquiry, the database system 205
transmits a challenge signal to the controller A (201) through the
network (Step S3A). The authentication circuit (refer to the
reference code 11 of FIGS. 1, 2, and 16) of the controller A (201)
outputs a response signal corresponding to the challenge signal
(Step S4A). The database system 205 determines whether or not the
response signal stored in the own database correspondingly to the
transmitted challenge signal agrees with the response signal
received from the controller A (201).
[0116] In the case of agreement of the both response signals (in
other words, when the response signal from the controller A (201)
is right), the database system 205 outputs the authentication
result that the controller A (201) is the regular product to the
brake control unit 202 (Step S5). Upon receipt of the
authentication, the brake control unit 202 executes a driving
instruction from the controller A (201) (in other words, it drives
the brake system 203) (Step S6A).
[0117] On the other hand, assume that a fake controller B (210)
provided outside of the automatic operation system 200 transmits
the driving instruction of the brake system 203 to the brake
control unit 202, pretending to be the controller A (201) (Step
S1B). In this case, upon receipt of the inquiry from the brake
control unit 202 (Step S2), the database system 205 transmits the
challenge signal to the controller B (210) (Step S3B). However, the
response signal received from the controller B (210) (Step S4B)
does not agree with the response signal corresponding to the
controller A (201) stored in the database. As the result, the
database system 205 transmits the result that the controller B
(210) is not the regular product to the brake control unit 202
(S5). As the result, the brake control unit 202 does not execute
the driving instruction of the brake system 203 from the controller
B (210) (Step S6B).
[0118] As mentioned above, the above automatic operation system 200
with the authentication circuit based on the PUF circuit assembled
can identify the driving instruction from the other controller B
(210) pretending to be the controller A (201). Therefore, the
automatic operation system 200 can avoid a fraudulent operation
instruction (for example, disturbing putting on a brake) issued by
the fake controller B (210) outside of the automobile through
wireless communication.
[0119] As set forth hereinabove, the invention made by the inventor
et al. has been specifically described according to the
embodiments; the invention, however, is not restricted to the above
forms and it is needless to say that various modifications are
possible without departing from the spirit.
* * * * *