Radiant energy – Photocells; circuits and apparatus – With circuit for evaluating a web – strand – strip – or sheet
Reexamination Certificate
1998-07-20
2002-12-03
Le, Que T. (Department: 2878)
Radiant energy
Photocells; circuits and apparatus
With circuit for evaluating a web, strand, strip, or sheet
C250S559190, C438S016000, C451S006000
Reexamination Certificate
active
06489624
ABSTRACT:
FIELD OF THE INVENTION
This invention pertains to apparatus and methods for polishing a workpiece such as a semiconductor wafer on which one or more thin patterned layers have been applied. More specifically, the invention pertains to such apparatus and methods that monitor, during polishing or other process step that imposes a time-varying change in thickness of such a patterned layer, the extent of thickness change imparted to the layer to determine whether a desired endpoint has been reached. The invention also pertains to apparatus and methods that measure layer and film thicknesses, especially the thickness of a layer or film on the surface of a workpiece having multiple thin-film layers, and minute surface irregularities of such films.
BACKGROUND OF THE INVENTION
Semiconductor devices continue to be developed with no apparent upper limit on the feature density of such devices. Many obstacles associated with higher feature densities have been overcome using various technologies and methods.
One significant lingering problem has been achieving, during fabrication of the semiconductor device, satisfactory planarization of wafer surfaces having comparatively large area. As the integration density of devices has increased, the wavelength used for microlithography has tended to decrease with a concomitant need to form various features in different layers in the thickness dimension of the device. Reducing the wavelength of light used for microlithography tends to lead to a reduced depth of focus of the projection-optical system used for microlithography. Hence, there is an increasing demand for more precise planarization of wafer workpieces, at least over the exposure area of the wafer, between sequential microlithography steps. Also, as greater demands are made of so-called “inlays” (i.e., implantation of metal electrode layers, plugs, or damascene in the thickness dimension), the need to remove excess metal and achieve planarization after forming each layer is increased.
With improvements in techniques used to form layers on a semiconductor wafer or the like, various methods have been proposed and implemented for achieving at least localized planarization of each layer before applying the next layer. Demand for continued improvements in planarization methods is escalating.
A commonly used planarization technique employing surficial polishing to planarize relatively large areas on a workpiece (generally termed herein a “wafer”) is termed “CMP” (Chemical Mechanical Polishing or Chemical Mechanical Planarization). CMP removes surface irregularities on the wafer by combining physical polishing with chemical action, and is effective for polishing insulating layers or conductive layers. In CMP, a polishing agent in the form of a slurry is used in which granules of an acid or alkaline abrasive (e.g., silica, alumina, and cerium oxide commonly are used) are suspended in a liquid in which the abrasive granules are at least partially soluble. Polishing proceeds by applying an amount of the slurry and an appropriate polishing surface to the wafer surface and using relative motion of the polishing surface and wafer surface. The surface of the wafer is polished uniformly by keeping the pressure and the speed of the relative motion uniform across the wafer surface.
CMP exhibits certain problems, however. One important problem is achieving accurate detection of when a polishing step should be ended (i.e., detection of the optimal polishing endpoint). There is an urgent need for accurate detection of the polishing endpoint in-situ, i.e., detecting the extent of polishing while polishing is ongoing.
One conventional method for detecting the polishing endpoint utilizes changes in the torque of a motor rotating the wafer or polishing surface during polishing; e.g., detecting a change in friction when polishing has progressed depthwise sufficiently to encounter a layer beneath the target polishing layer. Unfortunately, this method is effective only in detecting when the polishing has progressed into a different layer underlying the layer being polished. Also, this method suffers from insufficient accuracy and precision.
Another conventional method involves measurement by optical interference of the thickness of a layer during polishing of the layer. According to one approach, an optical path is provided in a polishing pad through which a light beam is irradiated on the wafer surface being polished. According to another approach, a wafer-penetrating (e.g., infrared) light beam is directed through an optical path provided in the wafer carrier that contacts the reverse surface of the wafer while the obverse surface of the wafer is being polished. Alternatively, the light beam can be directed through an optical path provided in the polishing pad.
In the interference techniques, changes in an interference pattern in reflected light are monitored over time. I.e., the interference pattern changes as the thickness of the subject layer changes. The layer thickness or amount of polishing of the subject layer that has occurred can be calculated from such data.
Employing interference to measure layer thickness (i.e., measurement of time-varying changes in an interference pattern produced using reflected laser light) suffers from unreliability and errors caused by variations in measurement position.
For example, the surface being polished normally does not always have optically uniform characteristics. For example, FIG.
1
(
a
) shows a wafer
1
(i.e., semiconductor wafer with multiple applied layers) at time of beginning polishing. The wafer
100
comprises metal conductive traces
102
or the like embedded in an insulating layer
101
. The goal of polishing is to remove the protruding portions
103
of the insulating layer
101
, thus planarizing the surface of the wafer
100
. In FIG.
1
(
b
), a portion of a metal layer
102
superposed on an insulating layer
101
is removed. With wafers such as shown in FIGS.
1
(
a
) and
1
(
b
), light passing through or reflected from the wafer surface would be influenced in many ways. For example, regions on which the metal layer
102
is present can interrupt transmitted light and produce more reflected light than regions lacking the metal layer
102
.
Further with respect to FIG.
1
(
a
), changes in layer thickness are normally limited to the protruding portions
103
. Thus, variations in the amount of transmitted or reflected light can be measured only from the protruding portions. Furthermore, when surface irregularities exist, producing a satisfactory light signal indicative of layer thickness is impossible using conventional methods.
Thus, certain conventional methods (including conventional methods that utilize interference) suffer from the problem of excess noise entering the light signal, especially when the wafer contains metal conductive traces or the like.
In situ measurements are performed while polishing is ongoing and thus require that the wafer be moving while the measurements are obtained. Such a situation generates complex signals that tend to exhibit more instability (e.g., intensity of reflected light, etc.) than when the wafer is stationary. This is due in part to the fact that the light must pass through the polishing slurry which tends to disperse the light to varying degrees when in motion.
Another problem with contemporary layer-thickness measuring methods using interference is that the surface of the wafer to be polished inevitably includes minute irregularities.
FIG. 2
shows a typical example of such irregularities in a wafer ready for polishing. The wafer comprises a Si substrate
105
and a SiO
2
layer
106
. Within the SiO
2
layer
106
are metal conductors
107
. Light reflected from the surface of such a wafer comprises the following components: interference light (designated “a”) generated by reflection from concave portions
108
and convex portions
109
, and interference light (designated “b”) generated from further interference of the “a” interference light, including a diffracted light component arising from phase change
Koyama Motoo
Matsukawa Eiji
Ueda Takehiko
Ushio Yoshijiro
Klarquist & Sparkman, LLP
Le Que T.
Luu Thanh X.
Nikon Corporation
LandOfFree
Apparatus and methods for detecting thickness of a patterned... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Apparatus and methods for detecting thickness of a patterned..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Apparatus and methods for detecting thickness of a patterned... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2995759