• No results found

• is the process of building devices in a single piece of silicon crystal

N/A
N/A
Protected

Academic year: 2022

Share "• is the process of building devices in a single piece of silicon crystal"

Copied!
31
0
0

Loading.... (view fulltext now)

Full text

(1)

IC Processing Technologies

(2)

FUNDAMENTAL CONSIDERATIONS FOR IC PROCESSING

IC PROCESSING:

is the process of building devices in a single piece of silicon crystal

is a process of building successive layers of insulating, conducting, and semi- conducting material.

each layer is patterned to give a distinct function and relationship with surrounding areas and subsequent layers.

the layers are produced and patterned by various techniques.

Major IC technologies that are used today are

n-channel metal-oxide - semiconductor (NMOS),

complementary metal-oxide-semiconductor (CMOS),

bipolar, and

integrated Injection-logic bipolar (I

2

L).

in order of increasing complexity of the processing.

Major IC technologies

(3)

Figure 1 shows important ways to create a layer in or on the silicon crystal.

The layers are formed by - oxidation,

- implantation, - deposition, or by

- epitaxial growth of silicon.

Each one of these layers can be formed :

- uniformly, as shown in the left side of Fig. 1, or

- selectively, as shown in the right side of the, figure.

Building Individual Layers:

FIGURE 1

Schematic representations of forming layers in silicon:

(a) uniform, (b) selective.

Fabrication processes involved in Major IC technologies

(a) (b)

(4)

1) The oxidation technique chosen depends upon - thickness and

- oxide properties required.

-if thin oxides are required then it should be grown in dry oxygen in Si.

-If thick oxides(>0.5μm) are desired, then steam is used (~1 atm or an elevated pressure of upto 25 atm).

2) if an insulating layer is needed,

- oxidation of silicon forms a layer of Si0

2

, which is an excellent insulator and can be formed with a high degree of uniformity.

- the formation of the Si0

2

layer consumes silicon and may require relatively long thermal cycles for thicker layers.

3) if a thick insulating oxide is needed,

- it requires an oxide deposition of polycrystalline silicon.

Patterning of the layers for selective formation achieved by using - photolithography and

- etching

The various techniques chosen for forming individual layers is

according to their usefulness for particular applications.

(5)

The deposited oxide has

- a poorer uniformity than the thermally grown oxide - but can be deposited at much lower temperatures.

Analysis must be made at each step in the process development

to determine which type of layer will give desired results without adversely affecting previous layers.

Oxidation and deposition are the most common ways of forming a dielectric layer .

- The oxidation can be made selective by depositing and patterning Si

3

N

4

before oxidation,

- allows the oxide to form only wherever the Si

3

N

4

is removed, since the Si

3

N

4

blocks the oxidation.

- This technique is called “local oxidation.”

Polycrystalline silicon, also called polysilicon, is a material consisting of small silicon crystals. It differs from single-crystal silicon, used for electronics and solar cells, and from amorphous silicon, used for thin film devices and solar cells.

The main advantage of polysilicon over a-Si is that the mobility can be orders of magnitude larger and the material also shows greater stability under electric field and light-induced stress.

(6)

Integrating the Process Steps: The process of building a succession of layers requires careful consideration of each layer‟s relative position to the others.

Fundamental components of all transistors are:

-The junctions formed by n-type and p-type doped regions

-The doping levels and depths of these doped regions determine the characteristics of the devices.

Ion implantation is the most important doping technique in modern ICs.

- The ion- implantation techniques and the techniques used for diffusion of the dopants determine the doping profiles.

-Selective doping, as shown in the middle of Fig. 1b, can be achieved using a variety of masking layers such as photoresist, polysilicon, or Si0

2

.

Deposition of materials can be achieved through the CVD methods

- choice of material depends on the particular application and function - most common materials being :

silicon oxides, polysilicon, aluminum, and silicon nitride.

- Some deposited materials can be deposited selectively, as shown in the bottom of Fig. 1b, which may give an advantage for some applications.

Epitaxial growth of single-crystal silicon is deposited uniformly or selectively.

- Selective epitaxial growth is not used commonly in ICs but has many advantages that may be useful for future devices.

IC Fabrication Processes: Fundamental Components and Formation Procedures

(7)

The process sequence for n-channel or metal-oxide-semiconductor (NMOS) technology is the least complex because it requires the least number of lithography levels.

In addition, the fundamental aspects of the NMOS technology are also important for 1. CMOS technology

2. MOS memory technologies.

NMOS IC TECHNOLOGY

The basic n-channel circuit consists of NMOS transistors.

Each NMOS transistor

-consisting of a source, a drainand a gate region.

-is isolated from its neighbors by a thick field oxide.

The basic device structure is shown in Fig. 1.

The device is covered with an insulating layer of phosphorus-doped Si02, called P-glass.

The device has a layer of dopant under the field oxide, -called the channel-stop region,

-serves to improve the isolation between transistors.

The NMOS transistor operates by

-causing a negative charge to move

-from the source to the drainin response to a positive charge on the gate.

-If the charge on the gate is sufficient

-increases the source-to-gate voltage above a threshold voltage VT, -electrons are attracted to the region under the gate

-forms conducting path between drainand source.

Inmost applications, two kinds ofNMOS transistors areused:

-enhancement mode devices (EMD) -depletion mode devices (DMD).

-enhancement mode device has aVT > 0 -depletion mode device has aVT < 0.

Fig. 1

(8)

NMOS Transistor

Top View and Cross-Section

N-Channel Metal-Oxide Semiconductor Transistor

n- and p-type semiconductor regions

Thick and thin oxides

Etching Openings

Polysilicon gate

Metal (Al) Interconnections

Conducting Channel Region

Figure 1

(9)

Basic NMOS Process Key Steps

•Oxidation

•Photolithograph y

•Implantation

•Diffusion

•Etching

•Film Deposition Process involved

Polysilicon-gate is doped with phosphorus to minimize mobile ion contamination, resulting in faster switching speeds and better control of threshold voltage

Thin padding layer

Starting material.

Virtually all MOS technologies use silicon waferswith the crystal surfacein the <100> orientation. This <100> orientation is preferred over the <111>

orientation because it introduces about 10 times less interface trap density, which can cause charging at the Si02-silicon interface. This charging causes variations in the threshold voltage of the devices and is therefore

undesirable. For most NMOS technologies, the siliconwill be lightly doped p-type(~1015atoms/cm3).

(10)

NMOS Fabrication Process

(a) Grow “thin-pad” layer of SiO2 and deposit Si3N4layer by CVD.

(b) (1) Mask #1 -- Photolithography and etching to define the active NMOS transistor areas.

Remove SiO2and Si3N4everywhere else by etching.

(2) Then do boron ion implantation (“field implant”).

This improves isolation of neighboring transistors from one another.

(c) (1) Grow thick oxide (“field oxide”) outside the transistor areas.

(2) Remove Si3N4and thin pad SiO2by etching.

(3) Grow thin “gate oxide” (the oxide that will separate the polysilicon gate from the transistor channel.

(4) Ion implant boron to adjust the transistor threshold voltage to the desired value (gate voltage needed to turn on the transistor).

(5) Deposit polysilicon (CVD) over entire wafer.

(d) (1) Mask #2 -- Photolithography and etching to define the polysilicon gate regions and interconnections.

Remove poly everywhere else by etching.

(2) Ion implant arsenic or phosphorus to dope the source and drain regions and the polysilicon gate material (n-type).

(3) The implanted dopant is “activated” and usually driven deeper into the wafer with a high temperature diffusion step.

(e) (1) Deposit protective oxide (SiO2) by CVD.

(2) Mask #3 -- Photolithography and etching to open contact windows.

(f) (1) Deposit metal (aluminum) over entire wafer surface by sputtering or evaporation.

(2) Mask #4 -- Photolithography and etching to pattern the aluminum metal layer to produce the desired metal interconnection pattern.

Etch away the aluminum where it is not wanted.

(11)

(g) Not Shown

(1) Deposit a passivation (protective) layer of phosphosilicate glass (PSG) or borophosphosilicate glass (BPSG) over the entire wafer surface.

(2) Mask #5 -- Photolithography and etching to open windows in the PSG or BPSG passivation layer so bonding wires can be attached to the aluminum “bonding pads” on the periphery of each IC die.

(h) Not Shown

(1) Individual IC dice are separated from one another by sawing or by scribing and breaking the wafer.

(2) Each IC die is attached to a package or “header” with an epoxy adhesive or a metal alloy solder.

(3) Wire bonding is the most common method for making electrical connections between the aluminum bonding pads on the periphery of each IC die and the package.

There are several important features of this NMOS fabrication process:

There are 5 mask steps.

Each of these mask steps uses a “subtractive” (etching) process.

i.e., the entire surface of a wafer is first coated with the desired material.

then, the material is removed from areas where it is not wanted by wet chemical etching or plasma (dry) etching.

Additional insulating layers, metal layers, and mask steps can be added to enable more complex interconnection of transistors.

(12)

Bipolar Technology

(13)

Characteristics of Bipolar Technology

Higher switching speed

Higher current drive per unit area, higher gain

Generally better noise performance and better high frequency characteristics Better analogue capability

Improved I/O speed (particularly significant with the growing importance of package limitations in high speed systems).

high power dissipation

lower input impedance (high drive current) low voltage swing logic

low packing density

low delay sensitivity to load high g

m

(g

m

Vin)

high unity gain band width (f

t

) at low currents essentially unidirectional

Advantages of Bipolar over CMOS

Other Bipolar Advantages

(14)

Bipolar IC technology preferred technology for high-speed applications. Why?

- Because in this technology transit time depends on the base width of the bipolar devices.

-The base width -is not determined by lithography (as is the channel length of MOS devices) - is determined by distance/difference between two impurity diffusion profiles.

-this distance can be precisely controlled and - base-width can be made very thin,

- allows high- speed operation of the bipolar transistors.

The processing of bipolar devices is more complicated than the MOS process sequences.

The complication arises because of the need to form a buried layer of dopant by growing an epitaxial layer of silicon on a patterned and doped silicon substrate.

Figure

Three dimensional view of oxide-isolated bipolar transistor.

Modern bipolar IC technology uses a thick silicon dioxide layer to electrically isolate the individual transistors.

Fig. 1 shows :The resulting device cross

section of an oxide isolated bipolar

transistor.

(15)

Starting material :

Starting material for these devices is a lightly doped p-type substrate with either a <111> or a <100> crystal orientation.

Buried layer and Epitaxial layer : Why ?

Transistors must have high breakdown voltage of collector.

This requires high resistivity material,

but high resistivity coupled with thickness of wafer (in earlier technology) ---- -results in excessive collector resistance that

-limits high frequency response - increases power dissipation.

Solution of this problem

Epitaxial growth of a high-resistivity layer on a low-resistivity substrate

The substrate and epitaxial layer have opposite doping types to provide isolation.

In between them is a heavily doped diffusion layer called the „Buried layer‟.

Bipolar Technology Process

(16)

-Buried layer‟s sheet resistance should be as low as possible to reduce

collector resistance of transistor . Process of Buried layer Formation:

-implanting arsenic or antimony through windows etched in a thermal oxide -forming the heavily doped n + region of the collector.

-implanted layer is driven into the substrate in an oxidizing ambient -“buried layer.‟ is formed

-All oxide is then stripped and an n-epitaxial layer is grown on the substrate.

Buried layer Step Formation & its Use Because of different rates of oxidation

-between exposed buried layer and surrounding area, a step is formed at the buried layer edge.

The step in the buried layer

-propagates up through the epitaxial layer -serves as an alignment mark

for the next lithographic level (Fig. a).

Buried layer: heavily doped diffusion layer -serves as a low resistance collector contact.

step

Buried layer mask

Buried layer=Heavily doped n+ layer

(17)

Bipolar Technology

Bipolar Transistor

Top View and Cross-Section

Bipolar Junction Transistor (BJT)

Standard Buried Collector Process (SBC)

n- and p-type semiconductor regions

Thick and thin oxides

Etching Openings

Metal (Al) Interconnections

Active Transistor Region

Figure 1.

(18)

Standard Buried Collector Process (SBC Process ): Key Steps

•Oxidation

•Photolithography

•Implantation

•Diffusion

•Etching

•Film Deposition

The basic bipolar process is more complex than the NMOS process.

Seven mask levels vs. five for NMOS.

Epitaxial growth is usually used.

(19)

CMOS Technology

(20)

First proposed in the 1960s. Was not seriously considered until the severe limitations in power density and dissipation occurred in NMOS circuits

Now the dominant technology in IC manufacturing

Employs both pMOS and nMOS transistors to form logic elements

The advantage of CMOS is that its logic elements draw significant current only during the transition from one state to another and very little current between transitions - hence power is conserved.

In the case of an inverter, in either logic state one of the transistors is off. Since the transistors are in series, (~ no) current flows.

CMOS Technology

(21)

CMOS Process

• Most logic circuits are constructed from

complementary MOS (CMOS) devices.

Why?

• This requires NMOS and PMOS transistors to be fabricated on the same wafer.

• This requires more mask levels.

• Still more mask levels are used for complex

interconnections.

(22)

CMOS Technology

N-Well Technology Cross-Section

Complementary Metal-Oxide Semiconductor Technology

Dominant Technology in Integrated Circuits Today!

Requires both NMOS and PMOS Transistors

Oxidation

Photolithography Implantation

Diffusion Etching

Film Deposition

Process involved:

(23)

Features of Polysilicon-gate CMOS Process

Consists of nine masking steps

Optimized to form complementary PMOS and NMOS transistors on a common substrate

Can also fabricate some analog circuits with slight modifications

Use (100) silicon to reduce surface state density and improve threshold voltage control

Polysilicon-gate is doped with phosphorus to minimize mobile ion contamination,

resulting in faster switching speeds and better control of threshold voltage

Overview of Polysilicon-gate CMOS Process

Start with P-substrate

Grow P-type epitaxial layer on the substrate

Create N-well regions and channel stop regions

Grow gate oxide (thin oxide) and field oxide (thick oxide)

Deposit and pattern polysilicon layer

Implant source and drain regions, substrate contacts

Create contact windows, deposit and pattern metal layer

(24)

Epitaxial Growth

P-type substrate is doped with as much boron as possible to minimize substrate resistivity

Lightly doped P-type epitaxial layer is grown on substrate

NMOS transistors are formed directly on the epi layer which serves as a backgate.

P-epi P-substrate

N-well Diffusion

The wafer is then oxidized and etched to open windows through which ion implantation deposits a controlled dose of phosphorus

A prolonged drive creates a deep lightly doped N-type region called N-well

Thermal oxidation covers the exposed silicon with thin layer of pad oxide In N-well CMOS process:

NMOS transistors occupy the epi and

PMOS transistors reside in well.

P-epi P-substrate

pad oxide

N-well

(25)

P-well Diffusion

P-well CMOS process uses N+ substrate and N- epitaxial layer and a P-well.

NMOS transistors are formed in P-well and PMOS transistors in the epi.

N-epi N-substrate

P-well

pad oxide

Why N-well CMOS Process?

The N-well process offers a slightly better NMOS transistor than P-well CMOS and also allows the use of a grounded substrate favored by most circuit designers.

In a P-well process the NMOS still outperforms its counterpart because

electrons are more mobile than holes

(26)

Inverse Moat

LOCOS process is used to define field regions and moat regions

Locally oxidized regions of the die are called „field regions‟

Areas protected from oxidation are called „moat regions‟

First a patterned nitride layer is formed by depositing nitride across entire wafer

A selective etch is used to remove nitride over the field regions

The photomask used is called inverse moat mask because it codes for areas where moat is absent

P-epi P-substrate

N-well

pad oxide photoresist

nitride

(27)

Channel Stop Implants

Channel stop implants are required to ensure that thick-field threshold exceed the operating voltages

It uses a boron implant followed by a patterned phosphorus implant

Phosphorus implant counter dopes boron implant and raises NMOS thick-field threshold above maximum operating voltage

After LOCOS oxidation, a suitable etching strips away the remnants of the nitride block mask

Curved transition region at the edges of moat results from oxidants diffusing under the edges of nitride film and is called „bird‟s-beak‟

All photoresist is stripped off from the wafer in preparation for LOCOS oxidation

A thin layer of oxide called „dummy gate oxide‟ is grown in moat regions to eliminate any nitride formed underneath the pad oxide due to „kooi effect‟

P-epi P-substrate

N-well

pad oxide nitride

photoresist Boron channel stop Phosphorus channel stop

(28)

Threshold Adjust

Method 1: Two separate implants to set PMOS V

t

and NMOS V

t

. This Method allows independent optimization of both thresholds

Method 2: Single V

t

adjust implant to simultaneously reduce PMOS threshold and increase NMOS threshold voltage

The boron V

t

adjust implant penetrates the dummy gate oxide to dope underlying silicon

After the V

t

adjust implant, dummy gate oxide is stripped away to reveal bare silicon in moat regions

P-epi P-substrate

N-well

Boron channel stop Phosphorus channel stop Field oxide

Dummy gate oxide

Boron V

t

adjust implant

(29)

Polysilicon Deposition and Patterning

Polysilicon layer used to form gate electrodes is heavily doped with

phosphorus to reduce its resistivity

The deposited polysilicon layer is patterned using polymask

P-epi P-substrate

N-well

Boron channel stop Phosphorus channel stop Field oxide

Gate oxide

Boron V

t

adjust implant

poly

(30)

Source/Drain Implants

NSD implant involves application of photoresist to the wafer, followed by patterning using the NSD mask

Heavily doped N-type regions are formed by implanting arsenic through exposed gate oxide

Photoresist residue is stripped off and a new photoresist layer is patterned using PSD mask

Heavily doped P-type region is formed by implanting Boron through exposed gate oxide

Photoresist is again stripped off and a brief anneal activates the implanted dopants

P-epi P-substrate

N-well Field oxide

Gate oxide

poly PSD NSD

(31)

Contacts

MLO (Multilevel Oxide) is deposited and the wafer is again coated with photoresist

Contact windows are created and silicide is formed in the contact openings

Thin film of refractory metal precedes a much thicker layer of copper doped aluminium

Protective Overcoat

Protective overcoat is deposited over the final layer of metallization to provide mechanical protection and to prevent contamination of the die

Selected areas of metallization are etched to attach bond-wires to the integrated

circuit

References

Related documents

Although a refined source apportionment study is needed to quantify the contribution of each source to the pollution level, road transport stands out as a key source of PM 2.5

These gains in crop production are unprecedented which is why 5 million small farmers in India in 2008 elected to plant 7.6 million hectares of Bt cotton which

Planned relocation is recognized as a possible response to rising climate risks in the Cancun Adaptation Framework under the United Nations Framework Convention for Climate Change

Harmonization of requirements of national legislation on international road transport, including requirements for vehicles and road infrastructure ..... Promoting the implementation

Angola Benin Burkina Faso Burundi Central African Republic Chad Comoros Democratic Republic of the Congo Djibouti Eritrea Ethiopia Gambia Guinea Guinea-Bissau Haiti Lesotho

The scan line algorithm which is based on the platform of calculating the coordinate of the line in the image and then finding the non background pixels in those lines and

A simple extension of Beale's and Lehmann's models for the formation of porous silicon layer on p-type silicon is proposed with a view to explain the

1 For the Jurisdiction of Commissioner of Central Excise and Service Tax, Ahmedabad South.. Commissioner of Central Excise and Service Tax, Ahmedabad South Commissioner of