You are on page 1of 6

EEE

5354L Pre-Lab Reading: Oxidation

EEE 5354L: Semiconductor Device Fabrication Laboratory Fall 2017

Pre-Lab Reading:

Lab 1: Oxidation

1. Introduction

The development of a high-quality silicon dioxide (SiO2) has helped to establish the
dominance of Si in the production of commercial ICs. Among all the various oxidation
methods (thermal oxidation, electrochemical anodization, and plasma enhanced chemical
vaporization deposition (PECVD)), only thermal oxidation can provide the highest-quality
oxides having the lowest interface trap densities. PECVD has many advantages such as the
low temperature deposition, fast deposition rate. However, the quality of oxide deposited
from PECVD is bad and contains some impurity. As a result, in this module, we will
concentrate on the thermal oxidation process, kinetics and thickness measurement
techniques of SiO2 films.

2. Theory

2.1.Thermal oxidation process

There are two SiO2 growth methods, dry and wet oxidation, depending on whether dry
oxygen or water vapor is used. Dry oxidation is usually used to form thin oxides in a device
structure because of its good Si-SiO2 interface characteristics, whereas wet oxidation is
used for thicker layers because of its higher growth rate.

Dry oxidation:
+ ! !
Wet oxidation:
+ 2! Si! + 2! (!)

The oxidizing ambient may also contain several percent of hydrochloric acid. The chlorine
removes metal ions that may occur in the oxide.

Thermal oxide incorporates silicon consumed from the substrate and oxygen supplied from
the ambient. Thus, it grows both down into the wafer and up out of it. For every unit
thickness of silicon consumed, 2.27 unit thicknesses of oxide will appear (Deal & Grove,
1965).

In this experiment, we will use dry oxidation to form the SiO2 film. We will use different
oxidation time and measure the film thickness to verify the commonly used Deal-Grove
model.

1
EEE 5354L Pre-Lab Reading: Oxidation

2.2. Kinetics of growth Deal-Grove model

The kinetics of thermal oxidation of silicon can be studied based on a simple model
illustrated in Figure 1 (Deal & Grove, 1965). A silicon slice contacts the oxidizing species
(oxygen or water vapor), resulting in a surface concentration of C0 molecules/cm3 for these
species. The magnitude of C0 equals the equilibrium bulk concentration of the species at the
oxidation temperature. The equilibrium concentration generally is proportional to the
partial pressure of the oxidant adjacent to the oxide surface. At 1000 and at a pressure of
1 atm, the concentration C0 is 5.2 x 1016 molecules / cm3 for dry oxygen and 3 x 1019
molecules / cm3 for water vapor.

The oxidizing species diffuses through the silicon dioxide layer, resulting in a concentration
Cs at the surface of silicon. The flux Cs can be written as

(1)

Where D is the diffusion coefficient of the oxidizing species, and x is the thickness of the
oxide layer already present.

Figure 1. Basic model for the thermal oxidation of silicon (Deal & Grove, 1965).

2
EEE 5354L Pre-Lab Reading: Oxidation

At the silicon surface, the oxidizing species reacts chemically with silicon. Assuming the
rate of reaction is proportional to the concentration of the species at the silicon surface, the
flux F2 is given by
(2)

where k is the surface reaction rate constant for oxidation. At the steady state, F1=F2=F
Combining Eq. (1) and (2) gives

(3)

The reaction of the oxidizing species with silicon forms silicon dioxide. Let C1 be the
number of molecules of the oxidizing species in a unit volume of the oxide. There are 2.2 x
1022 silicon dioxide molecules / cm3 in the oxide, and we add one oxygen molecule (O2) to
each silicon dioxide molecule, whereas we add two water molecules (H2O) to each SiO2
molecule. Therefore, Cs for oxidation in dry oxygen is 2.2 x 1022 cm-3, and for oxidation in
water vapor it is twice this number (4.4 x 1022 cm-3). Thus, the growth rate of the oxide
layer thickness is given by
(4)

We can solve this differential equation subject to the initial condition, x(0)=d0 where d0 is
the initial oxide thickness; d0 can also be regarded as the thickness of oxide layer grown in
an earlier oxidation step. Solving Eq. (4) yields the general relationship for the oxidation of
silicon,
(5)

Where !! + 2! / ! / 2! which represents a time coordinate shift to account for


the initial oxide layer d0. The oxide thickness after an oxidizing time t is given by

(6)

For small value of t, Eq. (6) reduces to


(7)

During the early stages of oxide growth, when surface reaction is the rate-limiting factor,
the oxide thickness varies linearly with time. As the oxide layer becomes thicker, the
oxidant must diffuse through the oxide layer to react at the silicon-silicon dioxide interface,
and the reaction becomes diffusion limited. The oxide growth then becomes proportional to
the square root of the oxidizing time, which results in a parabolic growth rate.

Equation (5) is often written in a more compact form:


3
EEE 5354L Pre-Lab Reading: Oxidation

! + = ( + ) (8)

Where A = 2D / k, B = 2DC0 / C1, and B / A = kC0 / C1. Using this form, Eq. 6 and 7 can be
written as
(9)

for the linear region and as


(10)

for the parabolic region. For this reason, the term B / A is referred to as the linear rate
constant, and B is the parabolic rate constant. Experimentally measured results agree with
the predictions of this model over a wide range of oxidation conditions.

2.3. Oxidation instrument

The basic thermal oxidation instrument is shown in Figure 2. The reactor consists of a
resistance-heated furnace, a cylindrical fused-quartz tube containing the silicon wafers
held vertically in a slotted quartz boat, and a source of either pure dry oxygen or pure
water vapor. The loading end of the furnace tube protrudes into a vertical flow hood where
a filtered flow of air is maintained. Flow is a directed as shown by the arrow in Figure 2.
The hood reduces dust and particular matter in the air surrounding the wafers and
minimizes contamination during wafer loading. The oxidation temperature is generally in
the range of 900 to 1200, and the typical gas flow rate is about 1 L / min. The oxidation
system uses microprocessors to regulate the gas flow sequence, to control the automatic
insertion and removal of silicon wafers, to ramp the temperature up (i.e., to increase the
furnace temperature linearly) from a low temperature to the oxidation temperature so that
the wafers will not warp due to sudden temperature change, to maintain the oxidation
temperature to within 1 , and to ramp the temperature down when oxidation is
completed.
Filtered
air

Exhaust

Figure 2. Schematic cross section of a resistance-heated oxidation furnace (Nicollins & Brews,
1952).

4
EEE 5354L Pre-Lab Reading: Oxidation

2.4. Oxide quality

The simplest method for determining the thickness of an oxide is to compare the color of
the wafer with a reference color chart. When an oxide-coated wafer is illuminated with
white light perpendicular to the surface, the light penetrates the oxide and is reflected by
the underlying silicon wafer. Constructive interference leads to enhancement of a certain
wavelength of reflected light, and the color of the wafer corresponds to that wavelength.
For example, a wafer with a 500 nm silicon dioxide layer will appear blue green.
Clearly, color chart comparisons are subjective and are therefore not the most accurate
mechanism for determining oxide thickness. A more accurate measurement can be
obtained using techniques such as ellipsometry.

Ellipsometry is a very powerful and accurate optical method to measure the thickness of
silicon oxide lms (Pliskin & Conrad, 1964), (Archer, 1957), (Azzam & Bashara, 1977). The
technique uses plane polarized monochromatic light to illuminate the oxidized silicon
surface at an angle. The light is reected from both the silicon substrate and the oxide
surface. The direct outputs of an ellipsometric measurement are the angles j and f, which
are related to the refractive index of the silicon substrate, the refractive index of the silicon
oxide, and the thickness of the lm. The refractive index for thick silicon oxide is usually
assumed to be 1.46 for a wavelength of 632.8 nm (Reisinger, Oppolzer, & Honlein, 1992).
This technique requires that the thickness is known to be within approximately 250 nm
since the measured quantities are periodic functions of this thickness. Thicknesses down to
the very thin (<5 nm) regime may be measured. However, inaccuracies in this regime may
be present. The technique requires known refractive index of the oxide lm and assumes
that the refractive index does not vary with thickness. More complicated methods, such as
spectral or multiple wavelength ellipsometry (Reisinger, Oppolzer, & Honlein, 1992),
(Snyder, 1986) or multiple angle ellipsometry are now being used to improve the accuracy
of these measurements to <1 nm (Snyder, 1986), (Chao, Lee, & Lei, 1991).

3. References

3. Archer, R. (1957). J. Electrochem. Soc. (104), 619.


4. Azzam, R., & Bashara, N. (1977). Ellipsometry and Polarized Light. Amsterdam, North
Holland: Wiley Interscience.
5. Center, M. A. (n.d.). Retrieved 02 29, 2012, from Learning Activity 1: The Ion Implant
Process: http://matec.org/ps/library3/secure/modules/026/LA1/M026LA1.html
6. Chao, T., Lee, C., & Lei, T. (1991). J. Electrochem. Soc. , 138.
7. Deal, B., & Grove, A. (1965). General Relationship for the Thermal Oxidation of Silicon. J
Apply. Phys. , 36 (3770).
8. Doering, R., & Nishi, Y. (2008). Handbook of Semiconductor Manufacturing Technology
(Second ed.). Boca Raton, FL, USA: CRC Press.
9. Furukawa, S., Matsumura, H., & Ishiwara, H. (1972). Theoretical Considerations on
Lateral Spread of Implanted Ions. Jpn. J. Appl. Phys. (11), 134-142.
10. http://www.eeel.nist.gov/812/hall.html. (n.d.).
11. Lindhard, J., M. Scharff, & H. E. Schiott. (1963). Kgl. Dan. Vid. Selks. Mat. Fys. Medd (33),
165.
5
EEE 5354L Pre-Lab Reading: Oxidation

12. Mller, W. (2004). Fundamentals of Ion-Surface Interaction. Lecture Short Resume,


Technische Universitt Dresden, Forschungszentrum Rossendorf, Dresden.
13. May, G., & Sze, S. (2004). Fundamentals of semiconductor fabrication. Wiley.
14. Nicollins, E., & Brews, J. (1952). MOS Physics and Technology. New York: Wily.
15. Pauw, L. v. (1958). A method of Measureing the Resistivity and Hall Coefficient on
Lamellae of Arbitraty Shape. Philips Tech. Rev. 20 , pp. 220-224.
16. Pauw, L. v. (1958). A method of Measurig Specific Resistivity and Hall Effect of Discs of
Arbitraty Shapes. Philips Res. Repts. 13 , pp. 1-9.
17. Pliskin, W. A., & Conrad, E. E. (1964). IBM J. Res. Devices , 8, 4351.
18. Reisinger, H., Oppolzer, H., & Honlein, W. (1992). (35), 797803.
19. Schroder, D. K. (2006). Semiconductor material and device characterization (Third
Edition ed.). John Wiley & Sons, Inc.
20. Schroder, D. K. (1998). Semiconductor Material and Device Characterization, 2nd Ed.
21. Snyder, P. a. (1986). J. Appl. Phys. (60), 3292.
22. Velez, C., Hwang , Y.-H., Lee, H.-I., & Ren, F. (2012). Oxide Growth Procedure - Module 5 .
University of Florida, Chemical Engineering Department.