 |
Description  |
|
|
This invention relates to a method for making a semiconductor device,
particularly an insulated gate field effect transistor.
One known method of fabricating insulated gate field effect transistors is
to form spaced source and drain regions of one type conductivity in a
semiconductor substrate of the opposite type conductivity by diffusing
conductivity modifiers into the substrate from a gaseous or solid source.
The known doped oxide diffusion method, involving diffusion from doped
oxide blocks on the surface of the substrate is then used to form lightly
doped source and drain extensions, the ends of which define the ends of
the channel region of the device and connect the channel region to more
heavily doped source and drain regions, formed for example by conventional
diffusion techniques. The more lightly doped source and drain extensions
improve the output resistance and drain-source breakdown voltage of the
device. One of the benefits of this process is that the doped oxide blocks
can also be used as a mask in forming a relatively thin gate insulator. A
gate electrode may be disposed on the gate insulator in non-critical
alignment with portions thereof overlapping the relatively thicker doped
oxide blocks. This reduces the undesirable capacitances of the portions of
the gate electrode which overlap the source and drain extensions. It is
difficult, however, to control the exact amount of conductivity modifiers
which diffuse into the substrate from the doped oxide blocks to form the
lightly doped source and drain extensions. Too heavy a doping results in
low drain to source breakdown voltage, while too light a doping introduces
significant source or drain to channel resistance.
Ion implantation has been used as a method of forming source and drain
regions and extensions thereof. Bower, U.S. Pat. No 3,615,934 describes
one method by which a so-called self-aligned gate transistor may be
fabricated. An insulated gate electrode structure is used as a mask in an
ion implantation process in which ions are implanted on both sides of the
gate electrode structure to form source and drain extensions. The ends of
the extensions which are adjacent to the gate electrode structure define
the ends of the channel region of the device. By using the gate itself as
the mask, overlap of the gate and the source and drain extensions is
minimized, leading to a considerable reduction in parasitic capacitance
compared to transistors formed by the doped oxide method. Since ion
implantation is a highly controllable process, improved control over the
amount of dopant which is introduced into the source and drain extensions
is achieved.
Ion implantation has also been used to control the doping in the channel
region of an insulated gate field effect transistor and hence the
threshold voltage of the device. Introduction of conductivity modifiers of
the same type as the substrate into the channel increases the threshold
voltage of the device, while the introduction of opposite type modifiers
decreases the threshold voltage. Processes are also known in which both
the same and opposite type modifiers are implanted into a channel, usually
for the combined effect of each. See for example Masuhara et al, IEEE
Electron Devices Meeting, 1974, pages 397-399.
Lacking in the art however, is any teaching of how ion implantation
processes may be used to provide control over all of (a) drain to source
breakdown voltage, (b) channel length, and (c) threshold voltage, in a
single device.
In the Drawings:
FIGS. 1-5 are a series of cross-sectional views illustrating a succession
of steps in the present novel process.
FIG. 5 represents, diagrammatically, an insulated gate field effect
transistor 10 formed by the present method. The transistor 10 includes a
body 12 of semiconductive material, preferably silicon, having an upper
surface 14 adjacent to which the various processing steps to be described
below are carried out. The body 12 is of one type conductivity, P type in
this example, having a resistivity between about 20 and about 50 ohm-cm.
Within the body 12 and adjacent the surface 14 thereof are spaced source
and drain regions 16 and 18. In the example illustrated, the structure is
symmetrical so that in use these regions 16 and 18 are interchangeable. As
suggested by the designation N+ within these regions 16 and 18, they are
relatively highly doped to the conductivity type opposite to that of the
body 12. Adjacent to those portions of the surface 14 which are not
occupied by the regions 16 and 18 is a zone 20 which is also of the
opposite type conductivity but contains substantially fewer modifiers than
the regions 16 and 18.
A channel region 22 is adjacent to a portion of the surface 14 which is
located between the regions 16 and 18 and has predetermined dimensions.
The width of the channel region 22 is less than the spacing between the
regions 16 and 18, i.e. the zone 20 has dimensions larger than those of
the channel region 22, so that portions 24 and 26 of the zone 20 remain as
source and drain extensions. The conductivity type of the channel region
22 may be either N type or P type, or the channel region 22 may be exactly
compensated. In the example shown this region is labelled N- or P. With N-
type conductivity the channel is operable in the depletion mode and with P
type conductivity it is operable in the enhancement mode.
On the surface 14 are layers of insulating material. On the portion of the
surface 14 which overlies the channel region 22 is an insulator 28 which
has the characteristics of a gate insulator. At other locations, there is
a relatively thick insulator 30 which, in this example, is made up of two
layers, that is, a first layer 32 adjacent to the surface 14 and a layer
34 on the layer 32. This insulator structure results from one example of
the present process as will appear below. Openings 36 are present in the
insulator 30 over the regions 16 and 18. Metal contacts 38 and 40 extend
through the openings 36 to contact the regions 16 and 18, respectively. A
gate electrode 42 is on the gate insulator 28 and will usually have
portions extending up onto the insulator 30 in the same manner as in prior
doped oxide type structures.
The transistor 10 is fabricated as follows. The specific steps
characteristic of the present process are described. Conventional steps
such as etching, washing, and photolithographic steps are not illustrated,
for convenience.
With reference to FIG. 1, the surface 14 of the body 12 is first provided
with a diffusion masking coating 44 which is provided in conventional
fashion, with openings 46 to define the locations and areas of the regions
16 and 18. These regions are then introduced into the body 12 by
conventional diffusion processes. They are spaced from each other by a
predetermined distance labeled in FIG. 1 as D.sub.1.
The next steps are illustrated in FIG. 2. The masking coating 44 is removed
from the surface 14 and is replaced by a relatively thin layer 32 of a
material which will protect the surface 14 in the ion implantation step
which is to follow. The body 12 is then placed in conventional ion
implantation apparatus and a nonselective implantation of opposite type
modifiers is done to establish the zone 20 across all of the surface 14.
If desired, this implantation may be done selectively, but the zone 20
should be formed to at least span the space between the regions 16 and 18.
The conditions of this implantation are not critical to the present
invention and may be varied as desired to control or adjust the amount of
modifiers in the zone 20, and thus its conductivity. For example, however,
where the body 12 is P type, doped with boron to a resistivity of 20
ohm-cm and where the protective coating 32 is thermally grown silicon
dioxide having a thickness of 1000A, the zone 20 may be made by ion
implantation of phosphorus at an energy of 80 KeV and a dosage of
6.8.times.10.sup. 13 modifiers/cm.sup.2.
Next, if desired, the body 12 can be heated to redistribute the modifiers
in the zone 20 and to drive the zone 20 to a desired depth within the body
12. Alternatively, a final redistribution of impurities can be done after
all implantation steps described herein.
The next steps are illustrated in FIG. 3. An insulating layer 34 is next
deposited on the oxide layer 32. Preferably, the insulating layer 34 is
silicon dioxide formed by the pyrolytic decomposition of silane
(SiH.sub.4) in oxygen, in known manner. By means of a standard
photolithographic process, an opening 50 having a width of D.sub.2 which
is less then the spacing D.sub.1 between the regions 16 and 18 is
provided. The body 12 is next subjected to a heat treatment in an
oxidizing ambient to grow the channel oxide 28. The body 12 is then again
subjected to ion implantation to introduce into the body 12, through a
portion of the surface 14 defined by the opening 50, conductivity
modifiers of the one type, that is, the same type as the body 12. The
result of this step is the establishment of a surface-adjacent region 22S,
i.e. a diffusion source region for the channel region 22. Assuming that
the above described processing conditions have taken place, the
implantation in this step may be of boron at an energy of 30 KeV and a
dosage of 1.times.10.sup.14 modifiers/cm.sup.2.
Upon completion of this implantation step, the body 12 is next subjected to
a heat treatment in an inert atmosphere to redistribute boron and
phosphorus in the region 22. The result of this step is shown in FIG. 4.
The conditions of this step are chosen such that a portion of the zone 20
which is adjacent to the portion of the surface 14 which has been defined
by the opening 50 is converted from the original N type conductivity
toward P type conductivity. As stated above, the conductivity of the
channel 22 is controllable in the present process, and if conductivity
modifiers of the one type are introduced in a relatively small amount,
this amount will be inadequate to convert the portion 22 back to P type
conductivity so that the resulting device is a depletion mode device. If
more boron is introduced, the amount will be adequate to convert the
region 22 back to P type conductivity, with the result that the device is
an enhancement device.
The process is completed in conventional fashion by providing the openings
36, that is, contact openings for the regions 16 and 18, and depositing
and defining the conductors 38, 40, and 42 respectively.
While it has been shown and is preferred that the channel region 22 is
symmetrically located between the regions 16 and 18, it is possible to
offset this region toward one or the other of the regions 16 and 18. Any
such offset will increase the resistance between the channel region 22 and
the more remote one of the regions 16 and 18 and this increased resistance
may be useful in some circuit applications.
As should be apparent, the two ion implantation steps used in this method
provide for a wide degree of control over the doping in the zone 20 and
the channel region 22 leading to wide variation in the available
drain-to-source breakdown voltage and threshold voltage of the device. A
designer is thus able to provide a device which will meet a given set of
specifications by the relatively simple expedient of changing the
implantation conditions. Moreover, the channel length of the device is
highly controllable.
* * * * *
|
|
 |
|
 |
Description |
|
