Patent Application
20090047748
Monitoring of copper contamination of silicon wafers has recently become increasingly important for ultra large scale (LS) integrated circuit (IC) manufacturing due to the adoption of copper as an interconnect material.
To increase speed and reduce heat dissipation, IC manufacturers have replaced aluminum (Al) interconnects with copper (Cu) interconnects. The Cu interconnect fabrication process typically involves a large number of steps during which cross-contamination of the silicon wafer by Cu can take place. Especially vulnerable are the edges, the bevel, and the backside of Si wafers, where exposed Si surfaces could meet Cu containing solutions, vapors, slurries and electroplated film residues. Due to the high solubility and diffusivity of Cu at low temperatures (e.g., up to 400° C.) used during back-end processing, this Cu contamination can diffuse into the Si wafer and spread over a distance of, e.g., a few mm. If Cu reaches the active device area, it can degrade the gate dielectric resulting in reliability and yield issues. Some of the yield issues with circuits close to the wafer edge are attributed to Cu edge contamination.
The inventors have discovered techniques for enhancing sensitivity of non-contact electrical monitoring of copper (Cu) contamination on silicon (Si) surfaces. Accordingly, methods and systems that utilize these techniques for monitoring Cu contamination are disclosed.
In general, in one aspect, the invention features a method that includes exposing a silicon wafer to a dose of radiation having wavelength, λA, measuring an electrical parameter of the silicon wafer after exposing the silicon wafer to the radiation, and determining and outputting information about copper contamination of the silicon wafer based on the measured electrical parameter. The dose provides sufficient energy at λA to substantially activate copper contaminating the silicon wafer. λA can be about 450 nm or less (e.g., about 400 nm or less, about 350 nm or less, about, 320 nm or less, about 300 nm or less, about 250 nm or less, about 200 nm or less).
In general, in another aspect, the invention features a method that includes exposing a silicon (Si) wafer to a dose of ultra-violet (UV) radiation, wherein the conditions of illumination are sufficient to substantially enhance electrical activity of Cu contaminating the Si wafer, measuring an electrical parameter of the Si wafer after the exposure, and determining and outputting information about Cu contamination of the silicon wafer surface based on a measured variance of the electrical parameter before and after activation.
Embodiments of the methods can include one or more of the following features. For example, the information about copper contamination can be determined based on a value of the surface electrical parameters of the silicon wafer measured after activation of Cu by exposing the wafer to UV radiation. In some embodiments, the electrical parameter is a surface lifetime, τS, measured by alternating current (a.c.) surface photovoltage and surface Cu concentration NCu is determined from the difference of Δ(1/τS) before and after UV Cu activation. The electrical parameter can be a low frequency saturation surface photovoltage, VSPV0 and surface Cu concentration can be determined from the difference Δ(1/VSPV0) before and after UV Cu activation. The electrical parameter can be a surface voltage, VCPD measured by contact potential difference technique CPD and the information about the Cu contamination (e.g., concentration) can be determined from the difference of surface voltage measured before and after UV Cu activation. The electrical parameter can be a semiconductor surface barrier, VSB, measured as the difference of VCPD in the dark and under illumination and the information about the Cu contamination (e.g., concentration) can be determined from the difference ΔVSB before and after UV Cu activation. The surface barrier VSB can also be measured by high frequency ac-SPV.
In some embodiments, the electrical parameter is measured at a bevel of the silicon wafer. Alternatively, or additionally, the electrical parameter can be measured at a top surface of the silicon wafer. Alternatively, or additionally, the electrical parameter can be measured at a bottom surface of the silicon wafer.
The electrical parameter can be measured using a non-contact probe (e.g., a vibrating non-contact probe), such as a Kelvin probe.
The ultra-violet radiation dose can correspond to a dose provided by a UVO-cleaner® manufactured by Jelight Company, Inc. (e.g., including a Suprasil® UV quartz lamp), for time of about 10 seconds or more (e.g., 20 seconds or more, 30 seconds or more, 60 seconds or more, 2 minutes or more, 5 minutes or more).
The light source can be broadband or narrowband (e.g., monochromatic).
The wafer can be at room temperature during the exposure and measurement.
Among other advantages, the inventive techniques can provide copper contamination methods that are both quantitative and selective. Moreover, the techniques are extremely sensitive, providing measurements of copper surface contamination in the range of 1012 atoms/cm2 or less (e.g., 1011 atoms/cm2 or less, 1010 atoms/cm2 or less, 109 atoms/cm2 or less).
Copper (Cu) on a silicon (Si) wafer surface can increase surface recombination and can change the surface charge on the wafer. These two electrical manifestations provide a basis for measuring information about the copper concentration, NCu, on a surface of a Si wafer with non-contact methods sensitive to surface recombination, such as ac-surface photovoltage or photoconductive decay method, or a vibrating Kelvin probe that measures the surface voltage and surface barrier sensitive to the surface charge.
It is believed that these approaches can be limited by the fact that typically only a small fraction of copper impurities, e.g., of about 0.3%, are in an electrically active state on the silicon surface. In general, the vast majority of surface Cu is electrically inactive, and thus invisible in electrical measurements. In other words, a very small fraction of copper contaminants typically provide surface recombination centers or charge trapping centers. Moreover, even in the electrically active state, it can be difficult to distinguish Cu from other electrically active species. This can limit Cu detection sensitivity, especially at low concentrations, making it difficult to measure Cu in the 1010 atoms/cm2 range and below which is often desired for IC process monitoring.
Accordingly, methods of determining information about relatively low concentration copper contamination of silicon wafers are disclosed. In general, the methods involve electrically activating copper contaminants on the silicon surface and measuring an electrical parameter (P) of the silicon before (P1) and after (P2) activation. A variation in the measurement parameter (e.g., ΔP=P2-PP1 or Δ1/P=1/P2-1/P1) between measurements made before and after activation provides information about the Cu contamination and serves to determine NCu. Subsequently, information about the copper contaminants is output as desired (e.g., displayed to a process engineer or used in a feedback or feedforward manner in a manufacturing process).
In general, copper activation is performed by exposing the wafer to a suitable dose of blue and/or ultra-violet radiation of an appropriate wavelength (e.g., radiation at a wavelength of about 400 nm or less, 380 nm or less, about 350 nm or less, about 300 nm or less, about 250 nm or less, about 200 nm or less). It is believed that the short wavelength radiation interacts with surface species on the wafer without causing noticeable changes in the bulk of the silicon wafer. Appropriate wavelengths include UV wavelengths from a mercury lamp or other broadband UV emitter. A suitable dose results in substantial activation of the copper electrical surface activity, i.e., that causes a significant change in a measured electrical parameter (e.g., by a factor of 3 or more, 5 or more, 10 or more, 20 or more, 50 or more, 100 or more, depending on the electrical parameter.
Electrical parameters which can manifest enhanced readings due to copper activation include, for example, the surface lifetime, τS, the low frequency saturation ac-surface photovoltage (SPV), VSPV0, the surface voltage, VCPD, and the semiconductor surface barrier, VSB or surface recombination. Techniques for measuring τS are disclosed, for example, in M. Wilson, A. Savtchouk, J. D'Amico, I. Tarasov, L. Jastrzebski, and J. Lagowski, “MANIFESTATION OF CU IMPURITIES ON SILICON SURFACES, IMPLICATION FOR MONITORING CU CONTAMINATION,” published at ECS Transactions, 11 (3) 347-361 (2007) the entire contents of which are incorporated herein by reference. Techniques for measuring surface photovoltage are disclosed, for example, in U.S. Pat. No. 6,512,384, entitled “METHOD FOR FAST AND ACCURATE DETERMINATION OF THE MINORITY CARRIER DIFFUSION LENGTH FROM SIMULTANEOUSLY MEASURED SURFACE PHOTOVOLTAGES,” the entire contents of which are incorporated herein by reference. Techniques for measuring VCPD are disclosed, for example, by J. Lagowski and P. Edelman in “CONTACT POTENTIAL DIFFERENCE METHODS FOR FULL WAFER CHARACTERIZATION OF OXIDIZED SILICON,” Inst. Phys. Conf., Ser. No. 160, p. 133-144 (1997), and by D. K. Schroder in “CONTACTLESS SURFACE CHARGE SEMICONDUCTOR CHARACTERIZATION,” Material Science and Engineering, B91-92, p. 196-210 (2002), the entire contents both of which are incorporated herein by reference. Techniques for measuring VSB are disclosed, for example, in U.S. Pat. No. 6,037,797, entitled “MEASUREMENT OF THE INTERFACE TRAP CHARGE IN AN OXIDE SEMICONDUCTOR LAYER INTERFACE,” the entire contents of which are incorporated herein by reference.
A suitable dose of radiation can change a measured value of τS by about 10 times or more (e.g., about 20 times or more, about 50 times or more). A suitable dose of radiation can change a measured value of VSPV0 by about 10 times or more (e.g., about 20 times or more, about 50 times or more, about 80 times or more). A suitable dose of radiation can change a measured value of VCPD by about 5 times or more (e.g., about 10 times or more). A suitable dose of radiation can change a measured value of VSB by about 5 times or more (e.g., about 10 times or more).
Measurements of the electrical parameter can be performed using a non contact probe, such as a vibrating Kelvin probe for VCPD or VSB and a stationary transparent SPV probe for ac-SPV measurements that provide τS and VSPV0. In some embodiments, surface properties ac-SPV measurements are performed by using modulated light (e.g., pulsed light or sinusoidally modulated light) with short penetration depth beneath the Si surface. For example, a blue LED light with the wavelength of 470 nm and penetration depth into Si of about 0.6 μm can be used.
In general, activation and/or measurements before and after activation can be performed at a variety of temperatures. In some embodiments, activation and/or measurements are performed at the room temperature (e.g., 21° C.±0.1° C.).
Referring to
In general, the location of probe 130 with respect to wafer 101 can vary as desired. Typically, probe 130 and mount 120 are arranged to position the portion of wafer 101 to be measured in sufficiently close proximity to probe 130.
In general, a variety of mounting configurations can be used. For example, mount 120 can be a stage that supports wafer 101 within tool 100. In some embodiments, as shown in
In some embodiments, probe 130 can be scanned relative to wafer 101 and measurements can be acquired for multiple different locations of the wafer surface. The multiple measurements can be used to determine a copper contamination distribution across the wafer surface (e.g., a map, such as a contour map, illustrating areas of high or low contamination). Alternatively, or additionally, probe 130 can be stationary and wafer 101 can be scanned relative to probe 130 (e.g., via wafer handling system 155).
Without wishing to be bound by theory, it is believed that the activation transfers surface Cu from an inactive to an electrically active state at the Si surface or Si-native oxide interface. For ac-SPV measurement this active state means the electron-hole surface recombination center that reduces the surface lifetime, τS, in a way such that the change in 1/τS before and after activation, Δ1/τS, becomes approximately proportional to NCu. The proportionality constant quantitatively relating NCu to Δ1/τS can be determined, for example, using Total reflection X-ray Fluorescence (TXRF) measurements in a calibration procedure.
For surface barrier voltage measurements, the active state means negatively charged centers lower the VSB in p-type silicon and correspondingly change VCPD. As a result, this method can give significant enhancement of sensitivity to measurements of these electrical parameters, and can be capable of meeting detection sensitivities of less than about 1010 atoms/cm2, desired for advanced Si microelectronics. The selective character of activation and differential measurement serves to distinguish the Cu contribution from other electrically active species.
As discussed previously, the electrical monitoring of copper is applicable to the front (the surface where the IC is formed) and back surfaces of the silicon wafer, to the wafer surface near the edge and the wafer bevel. The latter can be achieved using appropriate configuration of small area probes facing the wafer edge, e.g., for both ac-SPV and Kelvin surface voltage measurement. The activation of surface Cu should be configured to enable the UV illumination of the wafer bevel.
In general, the disclosed techniques are sensitive to a range of concentrations of Cu contaminants from about 109 atoms/cm2 to about 1013 atoms/cm2 (e.g., about 1012 atoms/cm2 or less, about 1011 atoms/cm2 or less, about 1010 atoms/cm2 or less). The concentration range below about 1011 atoms/cm2 is considered, in some respects, the most important for silicon microelectronics.
Sensitivity of various electrical parameters to Cu contamination before and after electrical activation of copper is illustrated by the data shown in
Results from a procedure for monitoring Cu using surface lifetime are shown in
Other embodiments are in the following claims.
This application claims priority to Provisional Patent Application No. 60/942,353, filed on Jun. 6, 2007, the entire contents of which are incorporated herein by reference.
| Number | Date | Country | |
|---|---|---|---|
| 60942353 | Jun 2007 | US |
