Crystal Growth and Wafer FabCrystal Growth and wafer fabricationrication

Crystal Growth and Wafer Fabrication

- Chapter 3 of Silicon VLSI Technology

- Preparation of substrates

- Czochralski (Cz) and Float Zone (FZ) techniques

- Defects and Dopants

- Measurement Techniques

Why silicon?

We have discussed several reasons as to why silicon dominates:

band-gap ?????,

p-type and n-type dopants ???????,

SiO2 has many useful qualities ???????,

The raw material is very cheap (cost of one wafer?)

In defining substrates we need to specify:

1. Doping type/level,2. Crystal orientation,3. Impurity levels, (oxygen and carbon),4. Defect levels.

What makes a crystal?

-Crystals possess long-range order,

-We may have instead poly-crystalline or even amorphous material.

Electron Micrographs of Crystal versus Poly-crystal?

Crystal Defects

vacancy

interstitial

dislocation stacking fault

precipitate

Taken from figure 3-4

Raw Material and Purification - 1

-As we have already discussed, the raw material for silicon manufacture is sand- often obtained from Australian beaches,

The sand is heated in a furnace containing a source of carbon,

2C + SiO2 (MGS) Si + 2CO,

where MGS = metallurgical grade silicon.

-Although MGS is of relatively high purity (98%)*, it still contains a number of contaminants (such iron and aluminum).

*From Roadmap

Raw Material and Purification -2

-The MGS silicon is purified to EGS silicon (Electronic Grade Silicon) using a distillation process,

-First, the MGS is reacted with HCl to form SiHCl3, (trichlorosilane) which is in liquid form at room temperature,

-Fractional distillation results in impurity segregation, and extremely pure SiHCl3.

-To convert the SiHCl3 back into purified Si a CVD (Chemical Vapor Deposition) process is used (in a hydrogen atmosphere),

SiHCl3 (gas)+H2 (gas) 2Si (solid)+ 6HCl (gas),

-The nucleation surface is thin poly-Si rod, with a final thickness of many inches in diameter,-All that is specified is impurity level, so fast deposition is possible .

Wafer growth – Czochralski Method (Cz)

-From the high purity poly-Si, single crystal silicon is required,-The Cz process is the most common for large wafer diameter production.

-The EGS is broken into small pieces and placed in an SiO2 crucible,

-In an argon ambient, the crucible is heated to just above 1417oC,

-A single crystal seed is then lowered into the melt (crystal orientation and wafer diameter determined by seed orientation and pull rate),

-Dopant is added to the melt to intentionally dope the resulting crystal,

-The oxygen and carbon (from graphite furnace components), contribute about 1017-1018cm-3 contaminants.

Wafer growth – Float Zone (FZ)

-The Float Zone (FZ) growth method is far less common, and is reserved for situations where oxygen and carbon impurities cannot be tolerated,

-The entire poly-Si rod is extracted from the EGS process as a whole.

Poly-Si

c-Si

RF coil

-The rod is clamped at each end, with one end in contact with a single crystal seed,

-An RF heating coil induces currents in the silicon, heating it beyond its melting point.

Wafer growth-Wafer Finishing

-The roughed ingot is turned into a uniform diameter using a diamond tip,

-A flat(s) is added to the ingot to indicate crystal orientation and doping type,

-Inside diameter saw is used to slice off individual wafers,

-Finally the wafers undergo a lapping and polishing stage which removes damage caused by the saw, and creates a smooth polished surface,

-Before use in the cleanroom, the wafers are subjected to a chemical clean to remove impurities on the surface.

(111), p-type (111), n-type (100), p-type

Dopant Incorporation

-Let’s now consider Cz growth, the wafer should be doped after growth (how much?),

-Dopant is introduced into the melt but control is non-trivial due to segregation,

-Any impurity will possess a certain chemical potential in silicon,

-This will be different for solid or liquid phases.

Cs

CLCs

CL

Solid

Liquid

K0 = CS/CL

-Where K0 is the segregation coefficient, for silicon K0 <1, impurities prefer to be in liquid phase!

Impurity k0

As 0.3

O 0.5

P 0.35

Sb 0.023

B 0.8

Au 0.000025

-We can relate the concentration of dopant in the solid crystal to the proportion of liquid and solid silicon by:

Cs = C0k0 (1-f)k0-1,

-where Cs is the concentration of dopant on the solid side of the interface, C0 is initial impurity concentration in the melt, and f is the ratio of solid silicon to the initial liquid volume,

-parameter of interest now in terms of known parameters!

-The concentration of impurity in the solid will change as a function of time, in other words, the distribution of dopant along the ingot will be graded.

Electrical Measurement Techniques – Hot Probe

Vm

HotCold

-The hot probe technique is used to determine the type of dopant in a wafer,

-It relies on the generation of the Seebeck voltage,

-i.e. the diffusion of free carriers as a result of a temperature gradient

e-

n-type wafer

Electrical Measurement Techniques – Four Point Probe

-The 4pp method measures the resistance of a wafer and using values of carrier mobility calculates the carrier concentration (why 4pp?),

-From geometry considerations we can show:

V= I/2rwhere r is the distance from the probe,

-in the case of 4pp this reduces to: = 2 s V/I,

which is valid if s<< t and d,

-in the case of wafers this can be reduced to:

= t/ln(2) V/I,-more conveniently:

= 1/ [q μp n + q μp p] - eqn. 1.1

V

I

St

d

= 4.532 V/I xj

s = 4.532 V/I

Bulk resistivity, cm

Sheet resistivity, /square

Electrical Measurement Techniques – Hall Effect

I

VH

B

e-

-Silicon is placed inside a magnetic field and a potential difference is placed across the material perpendicular to an applied electric field, the resultant motion of charge will experience a force perpendicular to the electric field,

-In equilibrium this results in a transverse potential difference known as the Hall voltage,

-The magnitude and sign of the Hall voltage leads directly to the carrier concentration, type, and mobility (if the resistivity is known).

RH = VH xj /B I

RH = 1/ qn (for n-type RH = μH

Hall coefficient

Giving Hall mobility