METHOD AND APPARATUS FOR PROCESS CONTROL IN THE SEMICONDUCTOR MANUFACTURING
FIELD OF THE INVENTION
The present invention is generally in the field of optical monitoring techniques, and relates to the control of semiconductor processing by measuring parameters of τbin films, e.g. within processing equipment (the so-called "integrated metrology").
BACKGROUND OF THE INVENTION
Optical methods for on-line or integrated measurement of the parameters of dielectric films (e.g., film thickness) are known in the art. Most of these techniques are based on reflectometry in broaden spectral range, e.g. from DUV to NIR spectral range.
The NanoSpec 9000 spectrophotometric device commercially available from Nanometrics Inc., USA (that is installed on the CVD production cluster tool Producer commercially available from Applied Materials, USA) utilizes a configuration that allows measurements of a wide range of dielectric films just after deposition within the CVD cluster tool before a processed wafer goes to an out-put cassette. The device includes a measuring tool installed in a loaάVunload block, outside a vacuum part of the cluster tool. This configuration suffers from that it applies measurements to a wafer a certain time after the wafer moves out of a reaction deposition chamber. During this time period, which is needed for fransferring the wafer within the vacuum part of the cluster tool and out of this vacuum part towards the cassette block, other
wafers of the lot pass through the deposition chamber. This time delay impedes appropriate control of the mantrfactiu±ig process. Actually, in case of malfunction or drift effect in the deposition process, measurements by a tool located out of the vacuum environment of the CVD tools arrangement will recognize this effect with a certain delay, and wafers of the lot processed after the first measured wafer will be scrapped.
Another technique is used in the NovaScan 840 integrated measuring tool, commercially available from Nova Measuring Instruments Ltd. According to this technique, a station (integrated metrology tool) installed on the CVD cluster tool as a separate vacuum chamber, or a non-operated chamber of the cluster tool is used for measurements. This is implemented by locating a measuring optical system outside the vacuum environment and separated therefrom by an optical window made in the respective chamber. The optical system utilizes a spectrophotometric measuring unit that measures the thickness of a deposited film through the optical window without affecting the deposition process. The measurements are performed on predetermined measuring sites in the wafer with a relatively small iUummating measuring spot. The typical spot size used in the system is about 15-20μm in diameter. This configuration allows recogmzing process deviations just after the first processed wafer is transferred from the processing area or chamber to the measuring area (chamber) and is measured by the integrated metrology tool. Such a fast response allows for "online" process contolling and correcting the processing parameters for the next wafer to be processed or to stop the processing at all if needed prior to processing the next wafer.
Since the above system utilizes a relatively small measuring spot and performs measurements on the predetermined sites, it requires precise positioning of the optical system relative to the wafer under measurements, as well as pattern recognition and auto-focusing techniques. A precise positioning means is used to locate the small spot on the predetermined test site using a predefined optical model (properties of all or at least some of the underneath layers of the wafer). Knowledge
of the optical model allows accurate and unambiguous mterpretation of the measured reflectance spectrum. However, this system suffers from the need for a time consuming alignment (e.g. pattern recognition, auto-focusing, and precise positioning) and additional operations or steps within the cluster tool associated with the wafer handling by an internal cluster robot that might affect the cluster tool operation sequence and its throughput especially in case of deposition of very thin films with short deposition time and respectively with high throughput of the cluster tool.
Still another approach for integrated measurements of the films' thickness, particularly apphcable to vacuum processing tools, consists of using a relatively large measuring spot (e.g., PCT publication No. WO00/12958 in the name of TEVET, or US Patent No. 5,900,633 in the name of On-Line Technologies Inc.). Such a technique does not require any pattern recognition, auto-focusing, precise positioning of a wafer, and/or movement of the optical system. Thus, the entire measuring cycle may be sufficiently reduced in order not to affect the throughput of the processing tool. Moreover, this technique provides measurements carried out during the wafer transfer from one location to the other within the processing (e.g. cluster) tool.
Measurements with a relatively large spot are implemented by averaging reflected light from a relatively large wafer's area (e.g. of a diameter d=20-30mm), i.e. shghtiy larger than the typical diagonal size of a die. Interpretation of the measured data is significantly different from that utilized in the above-indicated small tight spot based technique (e.g. 15-20μm). Averaging of reflections from different elements of the wafer pattern within a large light spot covering different optical stacks with unknown weighting makes spectrum analysis and data mterpretation very difficult, especially in those cases where there is a number of underlying layers in the wafer. Such a technique in case of multi-stack structures suffers from low confidence and low accuracy. In some cases, the contribution of the measured top layer within the relevant stack is so small that the measured reflectance spectrum is practically insensitive to this layer and cannot be measured with desired accuracy.
SUMMARY OF THE INVENTION
There is accordingly a need in the art to facilitate optical measurements of parameters of a patterned structure, such as a semiconductor wafer, by providing a novel optical system enabling measurements with measured areas of different sizes.
The main idea of the present invention consists of combining the advantages of both "large-spof and "small-spof approaches. By integrating a measurement system of the present invention with a processing tool, the accurate thickness measurements of a wafer's layer(s) can be provided with ntinimal effect on the throughput of a processing tool.
There is thus provided according to one broad aspect of the present invention, an optical system for use in a measurement system for measuring in patterned structures, the system comprising:
(a) an inuminator unit producing an iUuminating beam of tight to be directed to the structure to produce a tight beam returned from the structure;
(b) a detector unit comprising an imaging detector and a spectrophotometer detector; and
(c) a light directing assembly for directing the iUuminating beam to the structure and directing the returned beam to the detector unit, the light directing assembly defining a first optical path for the light beams propagation, optical elements accommodated in the first optical path affecting the tight beam to provide a relatively small measured area, and a second optical path outside said first optical path, such that the light beams propagation through the second optical path provides a relatively large measured area, as compared to that of the first optical path.
The term "measured area" used herein signifies a region on the structure's plane as viewed by the detector. This measured area is defined by the properties of the light directing optics and the sensitive area of the detector. The terms "small spot
operationalAmeasurement mode" and "large spot operationalAmeasurement mode" signify system operations with, respectively, relatively small and large measuring areas.
In one embodiment of the invention, the optical elements in the first optical path include an objective lens that focuses the iUunrinating beam onto the structure, while the second optical path is defined by an optical arrangement that is accommodated upstream of the objective lens with respect to the direction of the iniun ating beam propagation towards the structure's plane, and is shiftable between its operative position being in the optical path of the light beam propagating towards the objective lens and inoperative position being outside the path of the tightg beam propagating towards the objective lens. Hence, when the optical arrangement is in its operative position, it directs the iUuminating beam (and returned beam) to propagate along the second optical path aside the objective lens, thereby providing a relatively large measured size, and when the optical arrangement is in its inoperative position, the itiuminating beam (and returned beam) propagates through the objective lens, thereby resulting in a smaller measured area.
Preferably, the optical arrangement comprises first and second spaced-apart mirrors facing each other by their reflective surfaces. The first mirror is mounted stationary aside the objective lens, and the second mirror is movable between its inoperative position being outside the optical path passing through the objective lens and its operative position being inside said optical path. The optical arrangement may additionally comprise a beam-expanding unit accommodated in the path of a light beam reflected from the first mirror. til another embodiment of the invention, the optical system comprises at least two optical sub-systems, either utilizing a common inuminator and/or detector or not, wherein one sub-system is designed to provide a smaller measured area, and at least one other sub-system is designed to provide a larger measured area.
There is thus provided according to another broad aspect of the present invention, an optical system for use in a measurement system for measuring in patterned structures, the system comprising:
(a) an itiuminator unit producing an iUuminating beam of tight to be directed to the structure to produce a tight beam returned from the structure;
(b) a detector unit comprising an imaging detector and a spectrophotometer detector; and
(c) a light directing assembly for directing the iUummating beam to the structure and dkecting the returned beam to the detector unit, the tight directing assembly defining a first optical path for the light beam propagation, optical elements accommodated in the main optical path affecting the tight beam to provide a relatively small measured area, and a second optical path outside said first optical path, such that the light beam propagation through the second optical path provides a relatively large measured area, as compared to that of the first optical path.
According to yet another broad aspect of the present invention, there is provided a processing tools arrangement comprising a processing tool defining a processing region, and an integrated measurement system having the above-described optical system associated with a region within the processing tools arrangement outside said processing region.
The present invention also provides according to its yet another aspect, a method for confrolling a process apptied to a patterned structure progressing on a production tine, the method comprising selectively applying optical measurements to at least one predetermined site on the structure with measured area of different sizes.
BRIEF DESCRIPTION OF THE DRAWINGS
In order to understand the invention and to see how it may be carried out in practice, a preferred embodiment will now be described, by way of non-limiting example only, with reference to the accompanying drawings, in which:
Fig. 1 schematically illustrates an optical system according to one embodiment of the present invention, combining "small spof and "large spof operational modes in a common tight directing assembly;
Fig. 2 more specifically illustrates a beam expanding unit suitable to be used in the system of the present invention for implementing the "large spof operational mode;
Fig. 3 schematically illustrates a processing tool utilizing an integrated measurement system using the optical system of the present invention according to another embodiment, where the "small spof and "large spof operational modes are implemented by two separate sub-systems, respectively; and
Fig. 4 schematically illustrates a cluster processing tool utilizing an integrated optical measurement system using the optical system of the present invention according to yet another embodiment of the invention, where the "small spof operational mode is implemented by one sub-system and the "large spof operational mode is implemented by several other sub-systems.
DETAILED DESCRIPTION OF THE INVENTION
Referring to Fig. 1, there is illustrated an optical system, generally designated OS, according to one embodiment of the invention, apptied to a wafer W. The system OS comprises a broad-band (white) light source 10, an imaging detector 26 (CCD camera) the provision of which is optional, a spectrophotometer 30, and a light directing assembly, generally at 31. As will be described more specifically further below, the tight directing assembly is operable so as to selectively provide "large spof or "small spof measurement modes. The light directing assembly 31 defines an itiumination channel for the propagation of light produced by the tight source to the wafer's plane, and a light detection channel for the propagation of light returned from the wafer to the detectors
26 and 30, and defines two optical paths for itiummating and returned beams propagation, a first optical path for realizing the "small spof operational mode and the second optical path for realizing the "large spof operational mode. Optical elements of the assembly 31 accommodated in the inumination channel include a condenser lens 14 optionally connected to the tight source via an optic fiber 12; a beam splitter 16; a tube lens 18; and an objective lens 20 that may and may not be translatable. A pinhole mirror 22 located at the other side of the beam splitter 16 and a relay lens 24 define the tight detection channel part associated with the CCD camera 26. Another relay lens 27 and optionally a mirror 28 define the detection channel part associated with the spectrophotometer 30. All the above elements define the optical path for spectroscopic measurements with a small measuring area ("small- spof operational mode), as used for example in the NovaScan 840 ITM, commercially available from Nova Measuring mstruments, Rehovoth, Israel. Preferably, only the objective lens 20 along with the light beam deflecting element, such as a beam-sptitter or mirror (not shown) is translated in the X-Y plane parallel to the wafer's plane in combination with mirrors deflecting the collimated tight beam along X and Y axes (see US patent No. 5,764,365 assigned to Nova Measuring Instruments. Ltd.). Also, the wafer W may be moved relative the optical system OS; the movement may be carried out by X, Y or R-Θ stage or any other two-coordinate motion system.
Further provided in the tight directing assembly 31 of the optical system OS is an optical arrangement including mirrors 32 and 34, wherein mirror 32 is stationary mounted and mirror 34 is movable between its inoperative and operative positions 34 and 34' (shown in dashed tine) to be, respectively, out of and in the optical path passing through the objective lens 20. Any suitable drive (not-shown) can be used, being operated by a control unit (not shown), for providing the movement of the mirror 34, e.g. reciprocal, rotating, tilting, etc. Generally speaking, this optical arrangement of the light directing assembly provides selective propagation of the iUuminating and reflected tight beams through the first optical path passing through
the objective lens 20, or through the second optical path that does not pass through the objective lens 20. The mirrors 32 and 34, when in the operative position of the mirror 34, direct a wide collimated tight beam to the wafer W along the second optical path without the beam passage through the objective lens 20, thus providing spectroscopic measurements with a relatively large measuring area ("large-spof operational mode of the system 10). The system 10 thus can operate with two operational modes, i.e., with relatively small and large itiuntinating spots.
In the "small spof operational mode, the mirror 34 is outside the first optical path, and the system operates in the following manner. The beam splitter 16 reflects a tight beam 36 emanating from the tight source 10 towards the wafer W via lenses 18 and 20. The objective lens 20 focuses the iUuni ating tight beam 36 onto the wafer surface W. A reflected tight beam Ri is collected by the objective lens 20 and further transmitted by the lens 18 and the beam splitter 16 to the CCD camera 26 via reflective regions of the pinhole mirror 22 (outside the pinhole opening) for image acquisition procedure. It should be understood, although not specificatiy shown, that the output of the CCD camera 26 and of the spectrophotometer 30 are connectable to a control unit having suitable data processing and analyzing utilities for determining the wafer's parameters, particularly the thickness of one or more layers in the wafer. More specifically, the output of the CCD indicative of the acquired image of the uluminated site is processed by an image processor to identify the iUuminated location on the wafer W and thereby enable measurements in predetermined sites of the wafer having known optical stack(s) (model). A portion of the returned tight beam passes through the central opening in the pinhole mirror 22 and reaches the spectrophotometer 30. This tight portion is used for spectroscopic measurements. In the "small spof operational mode of the system, the measuring area on the wafer surface is defined inter alia by the pinhole size and optical magnification produced by lenses 20 and 18 and preferably is in the range of 10-20 μm.
In the "large spof operational mode, the mirror 34 is shifted into its operational position being in the first optical path. As a result, the tight beam 36
propagating from the tight source 10 and directed toward the wafer W by the beam splitter 16 and the tube lens 18 which forms the collimated beam, is reflected by the mirror 34 to propagate along the second optical path towards the mirror 32, and thus does not pass through the focusing lens 20. The mirror 32 reflects the itiunimating beam to the wafer's plane. Consequently, the itiuminating beam provides an itiuminating spot of a relatively large size (preferably, of the typical die size in the measured wafer). A tight beam R2 returned from the larger-size iUuminated spot is sequentiaUy reflected by the mirrors 32 an 34, and is then sequentially transmitted through the lens 18, beam splitter 16 and the central opening of the pinhole mirror 22 to reach the spectrophotometer 30 for spectroscopic measurements, til order to increase an input signal of the spectrophotometer 30 in the "large spof operational mode, the pinhole mirror 22 might be re-moved out of the optical path, til that case, the pinhole mirror 22 is movable to be out of or in the main optical path. If the pinhole mirror 22 is removed from the optical path, the spot size (i.e. measuring area) on the wafer surface is defined by the optical magnification produced by the lens 27, diameter of the lens 18, and by the size of detector's active area.
The present invention can be used with the NovaScan 840 ITM model, or any other Integrated Technology Measurement (ITM) tool, which operates with an image acquisition system for applying measurements through an optical window in a separate vacuum chamber or inoperative chamber of a cluster tool, as weU as any other production tool. AdditionaUy, the optical system of the present invention may comprise an auto-focusing sub-system, preferably of a dynamic type, described for example in the U.S. Patent No. 5,604,344 assigned to the assignee of the present application. The optical system of the present invention may utilize an alignment technique based on the pattern recognition, for example described in the U.S. Patents Nos. 5,682,242 and 5,867,590 assigned to the assignee of the present application. The auto-focusing and alignment techniques do not form part of the present invention and therefore need not be specificatiy described.
- li ¬
lt should be noted that in the "large spof operational mode of the system, such procedures as alignment, pattern recognition, and precise positioning of the beam on the wafer's surface, are not needed. Accordingly, the measurement time is reduced to nrtitimal, e.g. about 0.1 - 0.5 sec per measurement point, and measurements do not affect the throughput of the processing tool provided with an integrated measuring system utilizing the optical system OS.
Reference is made to Fig. 2 illustrating a beam expanding unit BE suitable to be used as part of the tight directing assembly in the above-described system OS to provide a desirably large spot size regardless of the size of the tight beam formed by the lens 18. To facilitate understating, the same reference numbers are used for identifying the common elements in the tight directing assembly 31 in Fig. 1 and the beam-expanding unit BE. The beam expanding unit BE comprises the tube lens 18, mirrors 32 (stationary mounted) and 34 (movable), and comprises two additional lenses 40 and 42 accommodated in the path of the itiuminating beam downstream of the mirror 32. The lenses 40 and 42 are designed to expand the tight beam (having, for instance, a 10mm diameter) to provide a measuring area of about 20-25mm diameter.
Turning now to Figs. 3 and 4, there are illustrated two more examples, respectively, of the present invention using "large spof and "small spof operational modes implemented using separate optical sub-systems LS and SS. Here, the optical system of the invention forms a part of an integrated measurement system.
In the example of Fig. 3, the present invention is used for contiolting a process of Chemical Mechanical Penalization (CMP). A CMP processing tool (polisher), generally designated PT, includes a processing (potishing) area 140, and an exit station 142 having a fransferring unit (e.g. robot) 144 for fransferring a wafer from an input cassette 146 to the processing area 140 for potishing, and for fransferring the polished wafer to an output cassette 152. An optical system of the present invention consists of a "large spof optical sub-system LS and a "small spof optical sub-system SS. The same robot 144 is used for transferring the wafer located
inside the exit station 142 to measurement positions with respect to sub-systems LS and SS. The "large spof sub-system LS, or only the detecting part thereof 148, is preferably installed within the exit station 142 and provides preliminary measurements (before processing) on the wafer to be potished during the wafer transfer from the input cassette 144 to the processing area 140 by the robot 144. In the present example, the detecting part 148 of the sub-system LS is connected to an optical unit OU of an external measuring unit MU via an optical fiber 149. The information about the thickness of a top layer of a wafer to be processed can be used for fitting the working parameters of the polisher, e.g. potishing time. After being polished, the wafer is supphed by the robot 144 to the measurement position of an optical system 150 of the "small spof sub-system SS located adjacent to the exit station 142 (may be mounted inside the exit station or at the location of the output cassette or adjacent thereto). The sub-system SS carries out thickness measurements, and the measurement results are used to provide a close loop control of the potishing process within the current lot of wafers. Information on the actual parameters of the processed wafer in addition to pre-potishing thickness information provide a dedicated process control within the current lot of wafers. After being measured, the wafer is transferred from the "small spof measuring sub-system SS to the output cassette 152. As will be described further-below, the "small spof measurement sub- system SS may be used for catibration of the "large spof measurement sub-system LS by performing pretiminary "catibration" measurement or measurements on at least the first wafer of the lot.
In the example of Fig. 4, the present invention is used for controlling a process of Chemical Vapor Deposition (CVD), or Physical Vapor Deposition (PVD) or etching. The CVD tools arrangement is typically a vacuum based cluster tool CT comprising several processing chambers 160 (three such chambers in the present example), a fransferring chamber 162 with a fransferring unit (internal robot) 164 and a non-operative chamber 166. In the present example, the optical system comprises a
"small spof sub-system 168 and several "large-spof sub-systems - three such subsystems 170A, 170B and 170C in the present example.
The sub-system 168 is preferably associated with the non-operative chamber 166 (measuring chamber), in a manner allowing measurements without breaking the vacuum conditions of the entire cluster tool CT. Preferably, the optical arrangement (not shown here) of the measurement sub-system 168 is located outside the vacuum chamber 166 and measurements are carried out through a transparent optical window made in the chamber 166. During the measurements, the wafer is handled by a suitable handling unit (e.g. rotatable or static chuck). The "large-spof measurement sub-system or systems may be installed within the transfer chamber 162, adjacent to the processing chambers 160 in order to perform measurements to the just processed wafer and therefore without affecting the throughput of the entire cluster tool. As shown on Fig. 4, all the sub-systems 168 and 170A-170C may be implemented as totally separate units, with a common processing unit CU and front end FE, contrary to the above-described system having common opto-electrical components, tii the present example, the measurement system utilizes a common external iUuminating unit πj for all the optical sub-systems. A corresponding number of separate specfrophotomefric units may be used, or alternatively the sub-systems may use some common components, e.g. a spectrophotometer with appropriate separate optical systems. For example, the "large spof sub-system may use optical fibers for fransntitting itiuminated and reflected tight to and from the location within the processing tool, e.g. inside the transferring chamber, etc. The "small spof sub-system is preferably installed within an exit station (interface) of the processing (cluster) tool. Single or multiple "large spof sub-systems may be located in other parts of the processing equipment, e.g., in the vicinity of a place to which the wafers are brought from the in/out cassette.
For most of wafers in the lot (usually, 25 wafers per lot), a proper process control may be carried out by measuring in a few points or even in a single (central) point of the wafer. Thus, a total effect on the processing tool throughput will be
negtigible. The measurements may be applied to the wafer while held on a robotic arm (end-effector during its movement within the cluster, so no additional wafer's handling is needed).
The important advantage of a "combined" measurement system in accordance with one aspect of the present invention (see Figs. 1 and 2) is the possibility of carrying out measurements in both the "large spof and the "small spof operational modes on the same wafer, without additional fransferring the wafer into another measuring location. The "small spof mode provides accurate measurements of the thickness of a top layer for any application (on any multi-layer stack). A combined measurement may be performed in the foUowing manner. Having performed the "small spof measurement on a predetermined site with the known optical model and calculated the thickness of the top layer, the "large spof measurement is apptied to the same location on the wafer. Data indicative of the actual thickness of the top layer obtained from the "small spof measurement can be used for optimizing the "large spof spectrum processing, e.g., by choosing (or verifying) the appropriate spectrum interpretation algorithm. It should be noted that the order of measurements, i.e. which of the two mode is used first, is not important for measurements, because the data interpretation may be carried out after both measurements (with both operational modes) have been completed. Actually, such a technique presents verification or catibration of the chosen interpretation algorithm for the "large spof operational mode, tii the case when there is no algorithm providing acceptable results (information on the top layer is lost due to averaging the signal within the large spot), it would be still possible by using only the "small spof mode. When the "large spof mode provides sufficient results, both modes may be combined. For example, the first wafer may be measured using both modes and the rest wafers in the lot (or most of them) may be measured using the "large spof mode only.
The known frequency decomposition technique (Fourier Transform) can be apptied for the interpretation of the measured spectra. In this case, a specified frequency (or frequencies window/(s)) corresponding to the top layer thickness may
be obtained from the "small spof mode measurement. This frequency, with a certain tolerance, is fuither used as a filter for exfractmg the useful data about the top layer thickness from a number of harmonic signals received as a result of the Fourier decomposition. This method ensures confident result even when the original spectrum includes a number of harmonic signals related to the tight reflection from non-relevant layers or layer stacks.
Those skilled in the art will readily appreciate that various modifications and changes can be apptied to the embodiments of the invention as hereinbefore exemplified without departing from its scope defined in and by the appended claims.