Measuring and testing – Vibration – Resonance – frequency – or amplitude study
Reexamination Certificate
2000-07-06
2002-11-12
Williams, Hezron (Department: 2856)
Measuring and testing
Vibration
Resonance, frequency, or amplitude study
C073S596000, C073S643000, C073S659000
Reexamination Certificate
active
06477898
ABSTRACT:
FIELD OF THE INVENTION
The invention relates generally to integrated circuit manufacturing and more particularly to a membrane mask stress measurement apparatus and method therefor.
RELATED ART
As semiconductor device geometries continue to decrease, more advanced lithographic techniques are required to achieve the patterning required to support such geometries. Some examples include x-ray lithography (XRL), ion beam projection lithography (IPL), extreme ultra-violet (EUV) lithography, and scattering with angular limited projection in electron-beam lithography (SCALPEL). SCALPEL, IPL and x-ray lithography techniques utilize masks that are formed of attenuating elements overlying thin membranes. The membrane thickness of a SCALPEL mask is typically in the range of 100-150 nanometers, whereas for XRL and IPL the membrane thickness is typically in the range of 2-5 microns. In order to fabricate the membrane mask, the silicon substrate is etched away to create free-standing membranes. SCALPEL masks typically include 500 to 1000 thin membrane windows approximately 1 mm×12 mm in size on a 200 mm silicon starting substrate. In contrast, XRL and IPL masks consist of one large membrane window on a silicon substrate.
SCALPEL membrane masks can suffer from distortions created through local stress on the membrane. As such, when various layers that may be included in the membrane mask are deposited, the deposition parameters may vary slightly from the ideal parameter settings such that the stress of one layer may not exactly match the expected stress of another layer. This mismatch may cause undesirable stress levels when composite layer membrane masks are formed where significant stress mismatch exists between various layers included in the composite layer structure.
In order to optimize the image placement requirements of the SCALPEL lithographic technique, the stress associated with a particular membrane mask used in the lithographic operations must be determined to lie within predetermined specification levels. By understanding what the stress level is for the membrane mask, the appropriate process alterations can be made to ensure minimal image placement deviations.
One technique used for measuring stress of thin films on substrates such as silicon wafers is a bow measurement test that measures the curvature of the substrate by reflecting light off the substrate after film deposition. The curvature of the substrate before and after deposition of the thin films is compared to determine the stress induced by the deposition of the thin films. Although such techniques work well for films having a thickness on the same order as the thickness of the underlying substrate, such techniques do not work well when the film thickness is orders of magnitude smaller than the substrate thickness. This is because the stress associated with the thick substrate renders the stress added by the thin films insignificant.
In order to avoid rendering the stress associated with the thin films insignificant, membranes made up of thin films alone may be fabricated for stress testing. One technique which utilizes such thin film membranes is bulge testing. In bulge testing, the membrane is suspended between two different pressures. The stress of the membrane can then be measured based on the “bulging” or distortion of the membrane resulting from the pressure differential.
In applications that include thin film membranes, such as SCALPEL masks, more accurate methods of stress measurement are required to ensure accurate lithographic results. One technique that has been used for measuring stress levels for SCALPEL membrane masks is based on a resonant frequency test (RFT). The RFT technique measures the stress of the free-standing membrane by inducing vibrations within the membrane under test. Such vibration inducement is accomplished in one prior art RFT technique through the use of electrostatic force. The electrostatic force is generated by applying a sinusoidal voltage to an underlying conductive chuck structure that drives the overlying membrane as well as physically supporting it. The chuck includes an array of complementary driving electrodes that correspond to the locations of membrane widows within the membrane mask. Each driving electrode stimulates a corresponding membrane widow.
A sensing electrode positioned over the membrane under test measures the voltage change generated between the electrode and the vibrating membrane. This is a capacitance-based measurement. At each membrane window test site, the voltage and frequency of the input signal applied to the drive electrodes included in the underlying chuck is manually adjusted while the output is observed on an oscilloscope or similar testing apparatus. When the amplitude of the vibrations reaches a peak level (which can be determined based on the output signal on the oscilloscope), a resonant frequency of the membrane under test is determined. The resonant frequency is then used in the following formula to determine the stress (in MPa) within the membrane.
Stress=(4f
r
2
&rgr;)/[(
m/a
)
2
+(
n/b
)
2
]
In the formula, f
r
is the resonance frequency of the membrane, &rgr; is the average film density in g/cm
3
, a and b are the rectangular edge length and width of the membrane widow in centimeters, and m and n are the number of halfwaves in the a and b directions, respectively. Membrane vibrations are detected using a capacitive measurement that senses a change in capacitance that results from the changing distance between the sensing electrode and the vibrating membrane. This particular RFT technique suffers from a number of disadvantages that may limit its effectiveness in testing membrane masks.
One fundamental problem associated with this prior art RFT technique is the low throughput in measurement. Since the capacitive sensor mounted on the micrometer is positioned inside the vacuum chamber, repositioning of the sensor to another membrane window and tuning the gap between the sensor and membrane using the micrometer requires breaking the vacuum. As a result, a considerable amount of time is required for measuring hundreds of membrane windows included on a mask. Although the time required can be reduced by installing an x-y stage mounted on a guide rail inside the vacuum chamber, this adds complexity to the measurement system as a servo motor control is required. Furthermore, the manual control in sweeping frequency around the resonance frequency and visual inspection of peak amplitude of the output signal is also time consuming.
Another drawback of prior art RFT technique is the measurement inaccuracy. Since the resonance peak is determined by visual inspection of peak amplitude of voltage, choosing the peak amplitude becomes arbitrary when the Q factor (energy storing efficiency factor) is not high, which is the case for thin membranes included in SCALPEL masks. Determination of the peak amplitude can also be problematic due to the presence of noise in the detected signal. This noise is further aggravated when the driving signal can be fed-through to the detected signal by capacitive coupling.
Furthermore, RFT measurement techniques based on the capacitive excitation cannot measure the resonance frequency of a membrane that does not have a conductive layer. In order to cause the membrane to vibrate due to electrostatic excitation, the membrane must include a conductive layer such that a ground plane is created to allow for the electrostatic excitation. This may be undesirable in some membrane structures where a conductive layer is not present or needed for lithographic purposes. For example, in one instance a single layer membrane may be measured for its stress level, such that this stress level can be compared with the stress level measured after an additional layer has been added to the initially measured layer.
Prior art RFT techniques are also limited by the need for a custom chuck that includes the drive electrodes for each substrate size and each pattern of membrane windows included on a mask. Thus, for diff
Han Sang-In
Mangat Pawitter
Clingan, Jr. James L.
Saint-Surin Jacques
Williams Hezron
LandOfFree
Membrane mask stress measurement apparatus and method therefor does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Membrane mask stress measurement apparatus and method therefor, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Membrane mask stress measurement apparatus and method therefor will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2991934