Semiconductor device manufacturing: process – With measuring or testing
Reexamination Certificate
1998-07-16
2001-04-03
Elms, Richard (Department: 2824)
Semiconductor device manufacturing: process
With measuring or testing
C438S018000, C438S017000
Reexamination Certificate
active
06210982
ABSTRACT:
BACKGROUND
1. Field of the Invention
The present invention relates generally to a method and apparatus for determining material properties using scanning microscopy. More specifically, a new method is taught for improving spatial resolution and accuracy of dopant density profiling of materials used in semiconductors when conducting scanning probe microscopy.
2. State of the Art
State of the art integrated circuit technology demonstrates that it is possible to create active and passive electrical and electronic components on a semiconductive substrate at the sub-micron level. This ability requires accurate knowledge of the spatial extent of impurity dopants that are incorporated into the semiconductive substrate. This knowledge is necessary because of the scale at which the concentration, and thus variation or profile of the dopants is operating. Essentially, in order to achieve predictability in active and passive component behavior, it is necessary to be able to accurately measure the dopant density profiles which can then be used by design engineers in design and manufacturing processes. Lack of precision in the incorporation of dopants can result in a proliferation of undesirable defects during later steps in a manufacturing process, and possibly less than adequate performance of the finished product.
Dopant regions are formed by actively injecting or passively diffusing a desired impurity into a surface of the substrate. When dealing with silicon as the semiconducting substrate, a native oxide occurs when the substrate surface is exposed to oxygen. It is possible to implant the dopants through this thin and naturally occurring oxide layer.
After forming active and/or passive components on the semiconductive substrate surface, functionality of the integrated circuit is determined by many factors. One important factor is the concentration of dopant atoms within dopant regions. Therefore, accurate profiling of the substrate is critical for accurate estimates of operating characteristics.
The state of the art is replete with different ways to characterize and thus create a profile of dopant regions. There are many one dimensional dopant profiling techniques, such as secondary ion mass spectroscopy (SIMS), spreading resistance, junction staining and anodic sectioning. Disadvantageously, these methods all fail to provide profiling in two dimensions. However, with the advent of the scanning tunneling microscope and the scanning probe microscope, new methods for dopant profiling became possible on a nanometer scale.
For example, early measurement techniques generally measured resistance and converted each resistance reading to a concentration amount.
Another technique is to profile a cross-section of the substrate which has been severed along the dopant region. Dopant concentration is then measured in two dimensions in both vertical and lateral directions.
Another way to obtain two-dimensional dopant profile measurement and inverse modeling is by scanning capacitance microscopy as disclosed in U.S. Pat. No. 5,523,700. This patent teaches how a one dimensional model is used to extract two dimensional dopant density profiles from measurements made by a scanning capacitance microscope.
Finally, one illustrated method as taught in the prior art is shown in
FIG. 1. A
probe
10
is placed on a surface
12
of a substrate material
14
to be probed. In this example, the substrate material being probed is silicon, with a layer of silicon dioxide
16
(naturally occurring or intentionally disposed thereon) covering the surface. The probe
10
is placed on and moved over the layer of silicon dioxide
16
. In this cross-sectional view, the dopant density or carrier concentration is represented in the silicon
14
as concentration contours
18
. In this method, the probe
10
is operated in what is referred to as a constant change in capacitance mode. A harmonic AC bias voltage is applied to the probe
10
. By measuring the AC voltage necessary to maintain a constant change in depletion capacitance, it is possible to determine dopant concentration using a conversion model and algorithm. The conversion algorithm relates AC bias voltage data to dopant density concentration using a physical model.
The physical model requires particular parameters to be defined in order to accurately represent the system of the substrate
14
. These parameters include an oxide dielectric constant, oxide thickness, probe tip radius, pining dopant density, pining bias voltage, and sensor probing voltage. Of these parameters, a free parameter can be the oxide dielectric constant, the oxide thickness, or the sensor probing voltage. The free parameter controls the lowest dopant density generated in the conversion. By fixing the dielectric constant and oxide thickness, it is possible to vary the sensor probing voltage for the best fit of SIMS.
FIG. 2
shows that disadvantageously, this method and model fails to take into account any dopant gradient which may exist in the material being probed. For example, the model (referred to as a first order model) incorrectly assumes that the dopant density is constant relative to the position of the probe in the material being probed. This is demonstrated by the depletion region
22
and the annular rings
24
. This assumption leads to an AC bias voltage to dopant density physical model which suffers from this inaccuracy.
It is observed in
FIG. 2
that for each annular ring, it is necessary to find the following: 1) the probe tip to silicon capacitance, 2) the oxide capacitance, 3) the silicon capacitance, and 4) the capacitance per annular ring
24
which is a series combination of the capacitances of steps 1, 2 and 3. The total capacitance is the sum contribution of all the annular rings
24
.
It would be an advantage over the state of the art to provide a new method for determining dopant density profiling which improved spatial resolution and accuracy. It would be a further advantage to provide a method for improved dopant profiling which could account for gradients in dopant density within the substrate being scanned.
OBJECTS AND SUMMARY OF THE INVENTION
It is an object of the present invention to provide a method for improving dopant profiling by increasing spatial resolution in scanning probe microscopy.
It is another object to provide a method for improving dopant profiling by increasing the accuracy of scanning probe microscopy.
It is another object to provide a method for improving dopant profiling by utilizing a new two-dimensional gradient model of interaction between the probe and the surface of the material being probed.
It is another object to utilize an iterative algorithm which utilizes the new two-dimensional gradient model to obtain a calculated dopant density which converges toward the true dopant density.
It is another object to apply this iterative dopant density determination technique for use with other microscopes.
It is another object to determine dopant density using a constant change in capacitance mode of operation of the probe.
It is another object to determine dopant density using a constant AC bias voltage mode of operation of the probe.
The presently preferred embodiment of the present invention is a method and apparatus for generating an accurate dopant density profile of a doped material using scanning probe microscopy, wherein the new method utilizes an iterative process to approach a dopant density profile having a user definable accuracy by creating a new two-dimensional gradient model of doping density within the doped material.
In a first aspect of the invention, the preferred embodiment of the method comprises the new step of utilizing a second order model to calculate a new AC bias voltage of the first order dopant density.
In a second aspect of the invention, the annular rings of the model of the doped material are divided into finite segments, thereby using a first order approximation of the dopant distribution as an input to the second order model.
In a third aspect of the invention, a newly calculated AC bias voltage is compared t
McMurray Jeffrey S.
Williams Clayton C.
Elms Richard
Morriss Bateman O'Bryant & Compagni
Smith Bradley K.
University of Utah Research Foundation
LandOfFree
Method for improving spatial resolution and accuracy in... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Method for improving spatial resolution and accuracy in..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method for improving spatial resolution and accuracy in... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2448257