downloads.semi.orgdownloads.semi.org/.../$FILE/ATTVQURT.docx · Web view: Recipients of this...

Background Statement for SEMI Draft Document 5506ANew Standard: GUIDE FOR MEASURING FLATNESS AND SHAPE OF LOW STIFFNESS WAFERS Notice: This background statement is not part of the balloted item. It is provided solely to assist the recipient in reaching an informed decision based on the rationale of the activity that preceded the creation of this Document.

Notice: Recipients of this Document are invited to submit, with their comments, notification of any relevant patented technology or copyrighted items of which they are aware and to provide supporting documentation. In this context, “patented technology” is defined as technology for which a patent has issued or has been applied for. In the latter case, only publicly available information on the contents of the patent application is to be provided.

Background: The current metrology strategies for wafers have evolved from methods used to characterize small, low aspect ratio geometries (smaller than 300 mm and thicker than 700 µm). Conventionally, three point mounts have been used to measure flatness/warp of these wafers along with gravity compensation.

However, 3DS-IC applications use larger and thinner wafers than conventional applications. Large, thin wafers have inherently low stiffness, leading to large deflections and gravitational sag. Three-point mounts create large deflections; 4-point and ring mounts provide redundant, non-kinematic support and are sensitive to how parts are placed on the mount. The results become unreliable.

The industry, therefore, could benefit from an alternate procedure that reflects the application usage of these wafers. One such approach used in the industry is an approach that no longer tries to correct for gravitational influence, but incorporates the gravity state into the measurement. The recommended method for characterizing the geometry of low stiffness wafers is with a wire mount and a noncontact scanning method such as laser scanning interferometry, which allows the depiction of a complete picture of the wafer’s shape and geometrical parameters in the horizontal orientation supported uniformly from below. In this document:

A low-stiffness wafer – or other geometrical shape – of any material type that show deflections more than 200% of the allowable bow tolerance when tested for bow using the conventional three-point mounting method

A horizontal support structure is defined with flatness or coplanarity less than 5% of the allowable warp tolerance. The contact area is minimized and distributed uniformly across the backside of the wafer under test. This low contact area has a flatness/coplanarity less than 5% of the allowable warp tolerance for the substrate.

This guide describes a procedure that characterizes the wafer in a position that allows for a free-state profile measurement on a flat surface.

The ballot results will be reviewed and adjudicated at the meetings indicated in the table below. Check www.semi.org/standards under Calendar of Events for the latest update.

http://www.semi.org/standards

Review and Adjudication InformationTask Force Review Committee Adjudication

Group: Inspection & Metrology Task Force NA 3DS-IC (Three-dimensional Stacked Integrated Circuits) Technical Committee

Date: Tuesday, November 4th, 2014 Tuesday, November 4th, 2014Time & Timezone: 8:00 am to 10:00 am (U.S. Pacific Time) 3:00 pm to 5:00 pm (U.S. Pacific Time)Location: SEMI Headquarters

3081 Zanker RoadSEMI Headquarters3081 Zanker Road

City, State/Country: San Jose, California 95134 / USA San Jose, California 95134 / USALeader(s): David Read

Victor Vartanian, SEMATECHRichard Allen (NIST)Sesh Ramaswami (Applied Materials)Urmi Ray (Qualcomm)Chris Moore (BayTech-Resor)

Standards Staff: Paul Trio (SEMI NA)408.943.7041, [email protected]

Paul Trio (SEMI NA)408.943.7041, [email protected]

This meeting’s details are subject to change, and additional review sessions may be scheduled if necessary. Contact Standards staff for confirmation.

Telephone and web information will be distributed to interested parties as the meeting date approaches. If you will not be able to attend these meetings in person but would like to participate by telephone/web, please contact Standards staff.

DRAFTDocument Number:

Date: 5/6/23

SEMI Draft Document 5506ANew Standard: GUIDE FOR MEASURING FLATNESS AND SHAPE OF LOW STIFFNESS WAFERS 1 Purpose1.1 The control of parameters, such as thickness, total thickness variation (TTV), bow, warp, and flatness, is well known to be essential to many, if not all, semiconductor processes as reflected in a number of existing standards.

1.2 The general trend in wafer geometry is to a larger diameter but even so the thickness increases correspondingly, it may not be enough to sustain adequate stiffness of the wafer. With growing aspect ratio of a wafer (e.g. diameter/thickness), gravity compensation errors increase until they become larger than the actual wafer surface variations. This is already the case for many special purpose wafers with lower modulus of elasticity and is becoming more common in mainstream applications. For example many wafer substrates are very thin such as germanium wafers used in specialized solar applications, and many other substrates are being made thinner such as lithium tantalate, and lithium niobate used in communication filters.

1.3 Manufacturing processes incorporating bonded wafers typically require thinned wafers; these wafers are then extremely low stiffness. In addition, since non-planarity of the wafers can lead to problems in subsequent processing steps, including lithography and electrical contact between metal layers on the bonded wafers, these technologies demand higher precision wafers and more accurate depiction of the true topography of the wafer.

1.4 This guide is designed to provide definitions for describing a more suitable measurement strategy for low stiffness wafers and geometries.

1.5 The more suitable measurement process consists of an alternative mounting for wafers with high aspect ratios and the use of high resolution measurements.

1.6 This guide also provides a measurement procedure for local bow, which denotes small areas of imperfection of the otherwise flat wafer or substrate.

1.7 This guide’s alternative measurement process is suitable for use in materials acceptance and process control, but may also be useful in other applications, such as wafer design and production.

2 Scope2.1 This guide is intended in general for metrology on low stiffness substrates like wafers or other geometrical shapes such as panels. As low stiffness can occur with any material, this guide can be applied to all materials that show low stiffness in certain deformation modes.

2.2 This guide applies in particular to glass and silicon wafers with a diameter equal to or exceeding 300 mm and their thickness equal to or less than 775 µm ± 20 µm.

2.3 This guide also applies to wafers or other geometrical shapes of any material type that show deflections more than 200% of the allowable bow tolerance when tested for bow using the conventional three-point mounting method.

2.4 Although the accuracy of TTV measurement is not as dependent on deflection as bow and warp are, it makes sense to use the same wafer support configuration when measuring TTV.

2.5 This document is a guide for a non-destructive procedure that uses a semi-continuous flat mounting surface and high resolution measurement methods.

NOTICE: SEMI Standards and Safety Guidelines do not purport to address all safety issues associated with their use. It is the responsibility of the users of the documents to establish appropriate safety and health practices, and determine the applicability of regulatory or other limitations prior to use.

3 Limitations3.1 As this guide is designed for wafers with low stiffness, the following environmental considerations should be taken into account. Wafers with a high aspect ratio or low modulus of elasticity tend to be sensitive to airflow,

This is a Draft Document of the SEMI International Standards program. No material on this page is to be construed as an official or adopted Standard or Safety Guideline. Permission is granted to reproduce and/or distribute this document, in whole or in part, only within the scope of SEMI International Standards committee (document development) activity. All other reproduction and/or distribution without the prior written consent of SEMI is prohibited.

Page 1 Doc. jn l SEMI

Semiconductor Equipment and Materials International3081 Zanker RoadSan Jose, CA 95134-2127Phone: 408.943.6900, Fax: 408.943.7943


Date: 5/6/23

vibration, and acoustics from the environment. It is desirable to minimize the introduction of these influences in order to minimize movement of the specimen wafer, as this will impact the precision of the measurement.

3.2 For transparent materials, the optical thickness can be directly measured with this technique. Knowledge of the index of refraction and the group index of refraction of the material being measured is needed to properly relate the optical thickness to the physical thickness of the wafer.

3.3 This guide applies to the measurement of a single wafer or other geometrical shape of low stiffness and not to stacks although the principles of this guide could be applied to stacks as well.

4 Referenced Standards and Documents4.1 SEMI Standards and Safety Guidelines

SEMI M1 — Specifications for Polished Single Crystal Silicon Wafers

SEMI M59 — Terminology for Silicon Technology

SEMI MF533 — Test Method for Thickness and Thickness Variation of Silicon Wafers

SEMI MF534 — Test Method for Bow of Silicon Wafers

SEMI MF657 — Standard Test Method for Measuring Warp and Total Thickness Variation on Silicon Slices and Wafers by a Non-contact Scanning Method

SEMI MF1390 — Standard Test Method for Measuring Warp on Silicon Wafers by Automated Non-Contact Scanning

SEMI MF1451 — Test Method for Measuring Sori on Silicon Wafers by Automated Non-Contact Scanning

SEMI MF1530 — Standard test method for Measuring Flatness, Thickness, and Thickness Variation on Silicon Wafers by Automated Non-Contact Scanning

NOTICE: Unless otherwise indicated, all documents cited shall be the latest published versions.

5 Terminology5.1 Terms, acronyms, and symbols associated with silicon wafers and silicon technology are listed and defined in SEMI M59.

5.2 Other Abbreviations and Acronyms Used in this Standard

5.2.1 CWS — chromatic white light sensor.

5.2.2 IR — infrared.

5.2.3 FQA — fixed quality area.

5.2.4 GBIR — global backside indicated reading

5.2.5 SBIR — (s)ite (b)ackside (i)deal focal plane (r)ange

5.2.6 SFQR — (s)ite (f)rontside least s(q)uares focal plane (r)ange

5.3 Other Terms Defined in this Standard

5.3.1 aspect ratio — the ratio of the diameter to the thickness of the wafer, or the ratio of the longest length to the thickness of the substrate, for instance the longest length of a rectangle is the diagonal.

5.3.1.1 high aspect ratio — large diameter and small thickness of the wafer lead to high aspect ratios.

5.3.2 chromatic white light sensor (CWS) — a non-contact method based on the principal of confocal optics, which relies on chromatic scanning. A lens is used that refracts white light differentially based on its wavelength in order to carry out distance measurements. Resolution depends on the intensity of reflected light. [The term refers to a sensor (CWS). The rest of this section describes a method of use]

5.3.3 coplanarity — a measure of the magnitude of how far out of an ideal plane a series of points is. This is similar to flatness but measured on an incomplete surface.





Date: 5/6/23

5.3.4 correction — the mathematical compensation for outside forces, such as gravitation.

5.3.5 deflection — a change in shape induced by an outside force, for instance by gravitation.

5.3.6 full surface bow — a bow that occurs over the full surface of the substrate

5.3.7 group index of refraction — index of refraction of the beginning and the end of the scan.

5.3.8 interferometry — high resolution, non-contact measurement technique that relies on the principal of superposition of multiple beams of light to determine the difference in distance traveled between two separate optical paths. Interferometry typically has a resolution <1/100 of the wavelength of the light being evaluated.

5.3.9 local bow — an area of a small, localized deviation of the general plane of a substrate.

5.3.10 low stiffness wafer — wafers that show deflections more than 200% of the allowable bow tolerance when tested for bow using the conventional three-point mounting method.

5.3.11 non-contact-metrology — a measurement technique that allows a surface to be measured without the requiring physical contact with the surface, preventing contamination or damage to the wafer substrate. Non-contact metrology also has the advantage of placing zero force on the surface of interest preventing measurement errors due to deformation induced by the force of the measurement.

5.3.12 semi-continuous flat mounting surface — a support structure that is horizontal, and has very low flatness and coplanarity of the contact surface, with minimized contact area distributed uniformly across the back surface of the wafer under test.

5.3.13 wire grid — sometimes also called a Harp. A mounting surface created by a precise monofilament strung across the surface of an optical flat to create a semi-continuous flat mounting surface support.

6 Comprehensive Explanation of Terms Related to Low Stiffness Wafers6.1 Geometrical Parameters — the geometric parameters that are primarily measured, like bow, warp, TTV, need some further explanation. Figures 1 through 4 show schematic diagrams representing these terms. Explanations for flatness and sori are not included in this section.

6.1.1 Full Surface Bow — full surface bow is a measure of the magnitude of concavity or convexity of a wafer is in its free, unclamped state. This is typically controlled to prevent excessive film stresses in the processing of the wafer or in the interaction between multiple wafers in the case of bonded wafer stacks. Typically, this is evaluated on the median surface of the wafer, which is midway between the front and back surfaces. In most cases, bow has been evaluated by measuring the deviation of the center point of the median surface of a free, unclamped wafer from a median surface reference plane established by three points equally spaced on a circle, with diameter a specified amount less than the nominal diameter of the wafer; SEMI MF534 provides a standard test method for measuring the bow of a single wafer in this fashion. In this guide the full surface measurement is used to create a more robust measure of bow A least squares spherical fit is applied to the Fixed Quality Area of the sphere. The intersection of the edge of the FQA and the fitted sphere defines the reference plane for the Bow measurement. Bow is the height of the center point of the fitted sphere in the FQA with respect to the reference plane. Note that bow is a signed value to signify concave vs. convex. The reported bow, A, is given by the difference between the center point of a least squares spherical fit to the median surface of the wafer, B, and the center point of the intersecting plane, C, as shown in Figure 1.

Figure 1Bow





Date: 5/6/23

6.1.2 Warp — warp is the difference between the most positive, A, and most negative, B, distances of the median surface of a free, unclamped (but gravity-compensated) wafer from a reference plane. SEMI MF657 and SEMI MF1390 provide standard test methods for measuring the warp of a single wafer, they were written with low aspect ratio and therefore stiff silicon wafers in mind. Warp can be zero, even for a wafer with surface curvature, if the front surface curves are mirror images of the back surface curves. Some forms of warp (see SEMI MF657) are with respect to the same 3-point surface as bow; while in other cases (see SEMI MF1530), a least squares reference plane is fitted to the measured median surface. This measurement can also be applied to the front surface or the back surface of the wafer.

Figure 2Warp

6.1.3 Local Bow — a measure of the local curvature, and directly related to the induced stress on a film or stacked wafer. The local bow measurement is evaluated the same way as the global bow, but on a sub-section of the wafer. As is done in stepper simulation measurements (such as SFQR or SBIR) a local quality area is scanned across the surface of the wafer and a new plot is constructed showing the magnitude of the bow at every possible location on the wafer. The local quality area for local bow should be round, and the size of the local area should be at least ten times the thickness of the wafer, and ideally should be the size of an exposure site on the wafer (typically on the order of 25 mm). The increment of the local quality area evaluation can be, and should be, smaller than the quality area size. For example, a typical configuration would be to evaluate the local bow on a 25 mm round site scanned every 5 mm in x and 5 mm in y. This would yield a local bow map with reported values every 5 mm. The local bow should be evaluated statistically. The maximum, minimum and average values of local bow should be provided.

Figure 3Local Bow Measurement





Date: 5/6/23

6.1.4 Total Thickness Variation — (G)lobal (B)ackside (I)deal Focal Plane (R)ange, GBIR, is the difference between the highest and lowest elevation on the entire wafer surface. These elevations are measured with respect to an ideal back surface plane, which is parallel to the reference surface of the interferometer. The ideal back surface plane is constructed by constraining the back surface of the wafer such that it is ideally flat. This emulates the condition the wafer sees when clamped to a stepper chuck, or some other chuck for processing. Some measurement techniques use a chucking surface to construct the thickness variation measurement, though it can be constructed virtually by measuring both surfaces simultaneously.

Figure 4TTV

6.2 Considerations on Gravity Compensation and Support Structure

6.2.1 Deflection — a deflection is a change in shape induced by an outside force. In the context of this document, the deflection is induced by the force of gravity bending the wafer over the support structure of the measurement system. Since the shape and magnitude of this deflection is determined by the distribution of the force over the support points, most support configurations have a distribution of force that is unpredictable as it is a function of the topography of the wafer as well as the position of the support points. For low stiffness wafers, the variation of the deflection can be quite large, as the base deflection is quite large, therefore having a very negative impact on the measurement reproducibility. For the three point support case there is a deterministic solution for the magnitude and shape of the deflection; however, the result is heavily dependent on the load position. If the position shifts laterally a small amount, the change in deflection can be very large for low stiffness wafers.

6.2.2 Gravity compensation — a mathematically calculated correction factor that is used to remove the effects of gravity upon a warp and bow measurement.

6.2.2.1 Gravity compensation for horizontally supported wafers — gravity causes substantial deformation of large diameter wafers, whether they are supported at the edge or in the middle. Unless compensated, gravity will induce a large error in warp measurements. SEMI MF1390 describes two compensation approaches. One approach is to correct for the gravitational effect of warp measurements by inverting the wafer and simultaneously measuring both the top and bottom surfaces of the wafer. Any differences between two values at the same site are due to the effect of gravity and can be used to correct for single-side measurements. Measurements obtained on representative wafers can also be performed and the deformation resulting from gravity determined.

6.2.2.2 Gravity compensation for vertically supported wafers — on tools that support the wafer vertically, effects due to gravity are negligible and therefore do not require compensation. Bow is only measured on one side, so there is no need to reverse the wafer.

6.2.2.3 Wafer support — there are also various wafer supporting strategies, most of which were created with stiff wafers in mind (low aspect ratio).

6.2.2.3.1 Single or 3-point support — stiff wafers may be supported on a single point, which typically takes the form of a small-area chuck at the wafer center. The shape of the front side of the wafer is determined first. This measurement includes gravity-induced deflection. The wafer is then inverted and the measurement performed again. The effect of gravity is removed by subtraction. For various methods, many other support strategies are applied. In gravity compensating methods on stiff wafers, often a 3-point support is utilized and either a calculated gravity compensation, or the reversal technique is applied (measure, invert, measure), or a “representative wafer” is measured with the reversal process, and the error in this measurement is removed from subsequent measurements.



WAFER FRONT SURFACE(Exaggerated depiction)



Date: 5/6/23

Mathematical modeling can only accurately compensate for the 3-point support, as this is the only case where the part is only distorted by gravity and not by its own flatness characteristics.

6.2.2.3.2 4-Point or Ring-support — for larger numbers of support points, such as a 4-point mount or a ring mount, only an approximation for compensation of gravity can be applied. As wafers become higher and higher in their aspect ratio, the errors in the gravity compensation methods become more and more significant.

6.2.2.3.3 Semi-continuous flat mounting surface support — a support structure constructed to have very low surface contact area, with a high degree of coplanarity (flatness of the contact area) with the contact area distributed in such a way to emulate a perfectly flat support structure. The low contact area minimizes the probability of a particulate disrupting the coplanarity of the support.

6.2.2.4 Maximum spacing for semi-continuous flat mounting surface support — for a semi-continuous flat support for measurement of the unconstrained state of a low stiffness wafer, the maximum spacing between support points on the structure should be selected so that if modeled as a simply supported beam deflecting under its own weight, the maximum deflection should not exceed 5% of the warp tolerance.

6.2.2.4.1 In addition, the maximum flatness error of the support structure (co-planarity) should be <5% of the warp tolerance. In this way it is possible to ensure that the support structure does not have a significant bias from tool to tool.

6.2.2.4.2 There are various ways to construct such a support structure. It is desirable that some method of verification of the co-planarity be possible. For example, a pin chuck or a ring chuck can be measured directly using interferometry, which is a simple and convenient verification method. For other structures it may less trivial to directly measure the support structure and other techniques may be needed to ensure co-planarity.

6.2.2.4.3 The required spacing of support points is a function of the wafer stiffness. Less stiff wafers will deflect more significantly over a shorter distance, requiring finer support spacing. Some basic examples of semi-continuous flat mounting methods would be a flat pin or ring chuck (not under vacuum) or an optical flat with precise monofilament strung across the surface, or an optical flat with an array of precise diameter balls, which would form a very low contact pin-chuck-like support. The minimized contact reduces the probability of a particle influencing the coplanarity of the contact array.

6.3 Considerations on Scanning and Data Point Intervals —–for low aspect ratio and high stiffness wafers, the use of only a few data points and subsequent interpolation of data to determine warp, bow, and TTV of a wafer may be sufficient. However, the precision of interpolated data declines with increasing aspect ratio and decreasing stiffness. As the industry demands higher precision wafers, the test method should allow equally precise measurement of the entire wafer. Therefore, for low stiffness wafers, the lateral resolution of the scanning system should be on the sub-millimeter level. The density of required data points depends on the dimensions of the wafer and the system, but should not be below one data point per square millimeter. Equally important is the even distribution of data points across the surface to identify critical surface features. This is especially important for calculation of the local bow distribution.

7 Testing of Low Stiffness Wafers for Warp, Bow, and TTV using a Semi-continuous flat surface and High resolution scanning interferometry7.1 Wafers made of silicon or glass with a diameter larger than or equal to 300 mm and a thickness of less than or equal to 775 µm should be tested on a semi-continuous flat mounting surface with the use of high resolution scanning interferometry as described in this document.

7.2 Wafers in any dimensions that show deflections more than 200% of the allowable bow tolerance when tested for bow using the conventional three-point mounting method are placed on a semi-continuous flat mounting surface with the front surface up. The recommended mounting is a wire grid, but other types may also be used.

7.3 Calculation of maximum deflection Deflmax of a round plate, which is uniformly loaded by gravity and is supported by a ring, deflection can be calculated by using the equation 1

1 Richard G. Budynas - Advanced Strength and Applied Stress Analysis, McGraw-Hill Science/Engineering/Math; 2 edition (October 29, 1998)





Date: 5/6/23

Deflmax=5+ ¿

1+ ¿∗(t∗g∗¿a4∗12 ) (1−❑2)64∗E∗t3 ¿

¿ (1)

Where:

Deflmax= maximum Deflection

a = radius of wafer (diameter / 2)

= Poisson’s ratio

t = thickness

g = gravity

= density

7.4 Semi-continuous flat mounting surface for wafer support

7.4.1 The preferred support for measuring warp, bow, and TTV on low stiffness wafers is an optical flat with precise monofilament strung across the surface (wire grid), the spacing not to exceed the amount calculated from Equation (1), or an optical flat with an array of precise diameter balls, its maximum spacing not to exceed the same amount, which would form a very low contact pin-chuck-like support. The center of the wafer should not be on a wire. Figure 5 gives an illustration of a wafer on a wire support.

Figure 5Illustration of a wafer on a wire grid support

1: Wire grid is one example of a semi-continuous flat mounting surface support.

7.4.2 A horizontal support structure is defined with flatness or coplanarity less than 5% of the allowable warp tolerance. The contact area is minimized and distributed uniformly across the backside of the wafer under test. This low contact area has a flatness/coplanarity less than 5% of the allowable warp tolerance for the substrate.

7.4.3 The maximum spacing L for a semi-continuous flat mounting surface is calculated as the bending of a wafer that is supported periodically on evenly spaced points. The calculation assumes that the wafer sags between these points and based on the warp tolerance calculates the appropriate spacing of the support points to prevent the sag exceeding a certain amount of warp tolerance. Since the wafer is supported by a series of points evenly spaced, for





Date: 5/6/23

instance by the wire grid, at the support point we know that the slope is zero so this is like a fixed support on both ends of a simple beam2:

L=

4√ wtol∗384∗E∗t2

60∗ρ∗g5

(2)

Where:

wtol = warp tolerance

L = maximum spacing

E = modulus of elasticity

ρ = density

t = thickness of wafer

g = gravity

7.5 Testing

7.5.1 Testing should be nondestructive and may be used on either 100% of the wafers in a lot or on a sampling basis.

7.6 Measurements by high-resolution interferometry

7.6.1 Measurements on low stiffness wafers as described in this document should be taken by high-resolution interferometry with the highest resolution possible, which should not be less than one data point per square millimeter.

7.6.2 The optical thickness variation is measured, and simultaneously the distance between the back surface of the wafer and the optical flat is determined. The optical thickness variation can be converted to the physical thickness variation by scaling the measurement by the group index of refraction. Once the physical thickness map and the map of the back surface of the wafer are known, the front, back, and median surfaces are constructed.

7.6.3 The bow, warp, and local bow of the front, back, and/or median surface are calculated as appropriate.

7.6.4 If desired, make local bow measurements by scanning a series of round fields, 25 mm in diameter to obtain a map of the desired surface within each field. Make these measurements on a 5 mm by 5 mm grid so that all measured points lie totally within the fixed quality area. See also section 6.1.3 and Figure 3

7.7 Functional configurations

7.7.1 There are two configurations for transparent and two for non-transparent substrates, depending on the position of the gauge.

7.7.2 The following nomenclature is used:

A = distance front surface of the wafer to optical flat

B = wafer thickness

C = distance optical flat to wire surface if measured from above

Neff = group index of refraction

7.7.3 Configuration 1 — Transparent substrate measured from below with one gauge

2 Richard G. Budynas - Advanced Strength and Applied Stress Analysis, McGraw-Hill Science/Engineering/Math; 2 edition (October 29, 1998)





Date: 5/6/23

7.7.3.1 In this configuration the substrate is transparent to the measurement wavelength. Here the measurement is performed from below the semi-continuous flat mounting surface, and the flat below the wire grid is used as the reference surface. The flatness error of this reference surface can be compensated for by measuring a separate optical flat on the system.

7.7.3.2 In this configuration to generate the four surfaces (front, back, median and TTV), it is necessary to compensate for the optical density of the wafer substrate. Buncorrected is the direct optical measurement of the optical thickness of the material. In order to convert Buncorrected to the physical thickness B, the effective index of refraction Neff has to be known. For a fixed wavelength measurement technique, N eff is simply the index of refraction at the measurement wavelength. For measurements that sweep through multiple wavelengths, N eff is the group index of refraction, which accounts for the variation of index over the wavelength range of the measurement.

Figure 6Measurement of a transparent substrate optically from below

B = Buncorrected * Neff (3)

7.7.3.3 The values of A and B should be measured for every point on the FQA with a minimum sampling pitch of 1mm in X and 1 mm in Y. The array of all A values over the entire surface is Map A, and the corresponding array of all values of B is Map B.

Front Surface = -1*(Map A + Map B) (3a)

TTV Surface = Map B (3b)

Back Surface = -1 * (Map A) (3c)

Median Surface=Front Surface+Back Surface2

=−1∗Map B+Map A2 (3d)

7.7.4 Configuration 2 — Non-transparent substrate measured from top and from bottom with two gauges

7.7.4.1 In a case where optically measuring through the substrate is not possible or practical, the wafer can be measured from both sides simultaneously. For this configuration, a reference measurement of C (Figure 7) across the entire field should be performed with no wafer present to create the correction for flatness errors in either of the two reference surfaces. Also the variations of C measured around the wafer site for each measurement can be used to compensate for any change in distance C to further enhance the accuracy of the thickness variation measurement.





Date: 5/6/23

Figure 7Measurement of a non-transparent substrate optically from top and bottom

7.7.4.2 As in the earlier case, the values of Aabove, Abelow, and C should be evaluated over the FQA with a minimum spacing of 1 mm in X and 1 mm in Y, and the array of values for Aabove, Abelow, and C are Map Abelow, Map Aabove, and Map C respectively. To calculate the 4 measurement surfaces, perform the following calculations:

Front Surface = Map C - Map Aabove (4a)

TTV Surface = Map C – Map Abelow – Map Aabove (4b)

Back Surface = Map A (4c)


= MapC−Map B+Map A2 (4d)

7.7.5 Configuration 3 — Transparent substrate measured from top with one gauge

7.7.5.1 This configuration should be used in the case that the measurement system cannot measure through the support material for some reason, yet the wafer material is still transparent to the measurement system. As in configuration two, the measurement of the support surface should be taken as a reference for the measurement. That reference is referred to as C (in Figure 7).

7.7.5.2 As in configuration two, the values of A, B, and C should be evaluated over the FQA with a minimum spacing of 1 mm in X and 1 mm in Y, and the array of values for A, B, and C are Map A, Map B, and Map C respectively. See the description in Configuration 1 for details on converting BUncorrected to the physical thickness B.

7.7.5.3 To calculate the 4 measurement surfaces, perform the following calculations:

Figure 8Measurement of a transparent substrate optically from top

Front Surface = Map B - Map A (5a)

TTV Surface = Map B (5b)





Date: 5/6/23

Back Surface = Map A + Map B (5c)


= Map B+Map A2 (5d)

7.7.6 Configuration 4 — Non-transparent substrates measured from top with one gauge

7.7.6.1 This is an alternate method for non-transparent wafers. The advantage of this configuration is that it only requires a single measurement head; however, it requires two separate measurements of the wafer to construct the necessary measurements. For this measurement it is necessary to have a much finer spacing of the support points to prevent excessive deformation under the vacuum condition. Also, because the support points should be rigid for the clamped measurement, it is necessary for the measurement of the reference surface C to be measured to the tops of the pins or rings of the support structure. The points between the support points should be interpolated from the areas where the support pins are present.

7.7.6.2 As in configuration two, the values of A, B, and C should be evaluated over the FQA with a minimum spacing of 1 mm in X and 1 mm in Y, and the array of values for A, A Clamped, and C are Map A, Map AClamped, and Map C respectively. See the description in Configuration 1 for details on converting BUncorrected to the physical thickness B.

Figure 9aMeasurement of non-transparent substrate optically from top unclamped

Figure 9bMeasurement of non-transparent substrate optically from top clamped

7.7.6.3 To calculate the 4 measurement surfaces, perform the following calculations:

Front Surface = Map C - Map A (6a)

TTV Surface = Map C – Map AClamped (6b)

Back Surface = Front Surface – TTV Surface = Map AClamped – Map A (6c)





Date: 5/6/23


= MapC−2∗MapB+ Map A+Map A Clamped2

(6d)

7.8 Calculations of warp, bow, and TTV

7.8.1 The calculations are done in accordance to the specifications cited in section 4.

7.8.1.1 Determine the map of the median surface by measuring the map of front surface, back surface, and map clamped according to section 7.7.

7.8.2 Determine the warp as follows:

7.8.2.1 Construct a reference plane by making a least squares fit to all points of the median surface over the entire scan area. Calculate the warp as the difference between the most positive deviation from the reference plane and the most negative deviation from the reference plane:

Warp = Amax + |Bmin| (7)where

Amax = the most positive deviation from the reference plane and

|Bmin| = the magnitude of the most negative deviation from the reference plane.

7.8.2.2 For front surface warp, construct a reference plane by making a least squares fit to all points of the front surface over the entire scan area.

7.8.2.3 Calculate the front surface warp from equation 2 using this reference plane.

7.8.2.4 For back surface warp, construct a reference plane by making a least squares fit to all points of the back surface over the entire scan area.

7.8.2.5 Calculate the back surface warp from equation 2 using this reference plane.

7.8.3 Determine the bow as follows:

7.8.3.1 Construct a least squares spherical fit to the median surface of the wafer.

7.8.3.2 Calculate the bow as the difference between the positions of the center point of this spherical fit to the center point of an intersecting plane constructed from three points on the median surface equally spaced around the boundary of the fixed quality area:

Bow = A – C (8)where

A = the position of the center point of the least squares spherical fit to the median plane and

C = the position of the center point of the intersecting plane.

2: Bow is positive if the center of the fitted sphere lies above the intersecting plane and negative if the center of the fitted sphere lies below the intersecting plane.

7.8.3.3 For front surface bow, construct a least squares spherical fit to the front surface of the wafer.

7.8.3.4 Calculate the front surface bow from equation 3 using this spherical fit and an intersecting plane constructed from three points on the front surface equally spaced around the boundary of the fixed quality area.

7.8.3.5 For the back surface bow, construct a least squares spherical fit to the back surface of the wafer.

7.8.3.6 Calculate the back surface bow from equation 3 using this spherical fit and an intersecting plane constructed from three points on the back surface equally spaced around the boundary of the fixed quality area.

7.8.4 Calculate the total thickness variation, TTV, as follows:





Date: 5/6/23

TTV = Tmax – Tmin (9)where

Tmax = the highest elevation of the entire front surface within the fixed quality area and

Tmin = the lowest elevation of the entire front surface within the fixed quality area.

8 Related Documents8.1 SEMI Standards

SEMI 3D2 — Specification for Glass Carrier Wafers for 3D –IC applications

SEMI M23— Specification for Polished Monocrystalline Indium Phosphide Wafers

SEMI M65 — Specifications for Sapphire Substrates to Use for Compound Semiconductor Epitaxial Wafers





Date: 5/6/23

APPENDIX 1Decision Tree for Mounting of Low Stiffness WafersNOTICE: The material in this Appendix is an official part of SEMI [designation number] and was approved by full letter ballot procedures on [A&R approval date].

Low Stiffness?

Wafer ≥ 300 ±5 mm in diameter and ≤ 775 ±5

µm in thickness

Calculate stiffnessSection 7.3

Calculate minimum spacing for semi-continuous flat surface

See section 7.4.2

If desired, select conventional support type and use appropriate gravity

compensation

No

Yes

Figure A1-1Figure Decision Tree for Kind of Mounting





Date: 5/6/23

APPENDIX 2Deflection of Glass and Silicon Wafers with Certain Aspect RatiosNOTICE: The material in this Appendix is an official part of SEMI [designation number] and was approved by full letter ballot procedures on [A&R approval date].

Table A2-1 Deflection as measured on a 3-point mount (SEMI MF 534):

Material Diameter(mm) Thickness (µm) Sag (µm)

Si 50 400 0.35Si 300 690 212Si 300 700 206Si 300 710 200Si 450 700 1060Glass 300 700 404

Table A2-2 Maximum deflection as calculated per equation (1) in section 7.3

Material Diameter(mm) Thickness (μm) Deflmax (μm)

Glass 200 700 38.9

300 197.1450 998.1200 500 76.3300 386.4450 1966.3

Silicon 200 700 20.6300 104.6450 530.0200 500 40.5300 205.2450 1038.8

NOTICE: Semiconductor Equipment and Materials International (SEMI) makes no warranties or representations as to the suitability of the Standards and Safety Guidelines set forth herein for any particular application. The determination of the suitability of the Standard or Safety Guideline is solely the responsibility of the user. Users are cautioned to refer to manufacturer’s instructions, product labels, product data sheets, and other relevant literature, respecting any materials or equipment mentioned herein. Standards and Safety Guidelines are subject to change without notice.

By publication of this Standard or Safety Guideline, SEMI takes no position respecting the validity of any patent rights or copyrights asserted in connection with any items mentioned in this Standard or Safety Guideline. Users of





Date: 5/6/23

this Standard or Safety Guideline are expressly advised that determination of any such patent rights or copyrights, and the risk of infringement of such rights are entirely their own responsibility.




downloads.semi.orgdownloads.semi.org/.../$FILE/ATTVQURT.docx · Web view: Recipients of this...

Documents

Transcript of downloads.semi.orgdownloads.semi.org/.../$FILE/ATTVQURT.docx · Web view: Recipients of this...