Extending Reflectometry Range: A Zero-Crossing Algorithm for Thick Film Metrology

Out-of-Plane CTE Measurement Method for Freestanding Thin Films

Introduction ATHEER's gas processing plant at Habshan, Abu Dhabi is one of the world'slargest gas processing facilities. It consists of two identical trains each of270 million standard cubic feet per day (mmscfd) for processing Thammama - "C"gas 30/40 and 31/41 respectively), two trains of 350 mmscfd each for mixture ofThammama - "B" and "C" gas (Unit 32/42 and 33/43 respectively) and one train of610 mmscfd for Thammama - "F" gas (Unit - 34/44). Thammama - "B" gas isassociated gas, whereas Thammama - "C" and "F" are nonassociated gas. Eachprocessing train consists of gas sweetening unit using MDEA solvent, Dehydration unit using molecular sieve and NGL Recovery unit havingturbo-expander- recompressor, demethanizer and debutanizer. Besides, each trainof Thammama - "C" and "F" has standalone condensate stabilizer unit (Unit - 20,21 and 24), whereas two trains of Thammama ’ "B/C" have common stabilizer unit(Unit - 22). The feed gas to the plant comes from three different onshore gas sourceswith different gas compositions. In let gas conditions are widely varying andstrong interactions between different units make control of the plant operatingconditions difficult. ATHEER decided that Advanced Process Control (APC) solutions at its Habshangas processing facilities would benefit its operation and allow greater naturalgas liquids (NGL) recoveries and reduced utility consumption. The objectives ofthe APC project were:Increased NGL and condensate recovery while meeting as demand for consumersand reinjection requirementsImproved quality control of NGL and condensate productsSmoother plant operation After considering various options, GE Continental Controls (GE CCI)'s MVC®was selected to provide the APC solutions at GP&PD. MVC® is GE CII'spatented model predictive multivariable controller/optimizer with nonlinearcapability. It uses predictive control models derived from plant dynamics andrigorous simulation results. Nonlinear mathematical solutions are used tocalculate the optimal controlled variables and issue the correspondingmanipulated variable set points on each execution cycle. Automatic bias servesto maintain the predictive integrity of control models in the face varyingplant conditions. In this project, eleven (11) multivariable controller modules, one (1)predictive nonlinear level control module and fifteen (15) inferential modelsto estimate condensate Reid Vapor Pressure (RVP), NGL C2 andC5+ compositions were implemented. Original Plant Conditions The feed gas to the plant is supplied from multiple sources and is subjectto widely varying feed conditions. Strong interactions exist between differentprocess units, making control of plant conditions difficult. Lack if analyzersmakes the closed-loop control of product quality impossible with traditionalcontrol strategies. Feed Complexity and Plant Operation Variation Each feed source is supplied from multiple gas wells and is subject todisturbances in pressure, flow and composition causing widely varying feedconditions. Thammama-"B" associated gas is compressed in an external facilityand mixed with Thammama-"C" non-associated as coming from the separators ofUnit 20/21/22 before feeding into Thammama- "B/C" processing trains. Thammama-"B" and "F" gases are typically heavy and contain higher percentage ofC3+ components as compared to Thammama-"C" gas. As such, anyinstability in the external compression facility causes severe disturbances inthe operation of the downstream processing units. Thus, both feed gascomposition and flow rate to all the units were varying separators, etc. werecontrolled tightly while their outlet flows varied. This resulted in continuousvariation in stabilizer operating conditions. Lack of Analyzers There are no on-line analyzers of RVP analysis of the condensate productfrom the condensate stabilizer and debutanizer bottoms. The NGL product fromunit 45 is not analyzed for C5+. The lack of analyzers makes itdifficult for closed loop control of these compositions that are important tomaintain the condensate and NGL product specifications.

This paper describes the partnership SEMY Engineering, Inc. and White Oak Semiconductor formed to successfully provide automation and advanced process control (APC) solutions from a combination of proprietary and commercially available software systems. The White Oak factory objective was to implement a computer integrated manufacturing (CIM) system providing full availability of all necessary data for yield analysis and process control. Semiconductor production throughput and reliability at White Oak has significantly increased using this implementation. A large ROI (>500%) has already been realized in the lithography cell in the first year alone.

Shrinking technology nodes and smaller process margins require improved photolithography overlay control. Generally, overlay measurement results are modeled with Cartesian polynomial functions for both intra-field and inter-field models and the model coefficients are sent to an advanced process control (APC) system operating in an XY Cartesian basis. Dampened overlay corrections, typically via exponentially or linearly weighted moving average in time, are then retrieved from the APC system to apply on the scanner in XY Cartesian form for subsequent lot exposure. The goal of the above method is to process lots with corrections that target the least possible overlay misregistration in steady state as well as in change point situations. In this study, we model overlay errors on product using Zernike polynomials with same fitting capability as the process of reference (POR) to represent the wafer-level terms, and use the standard Cartesian polynomials to represent the field-level terms. APC calculations for wafer-level correction are performed in Zernike basis while field-level calculations use standard XY Cartesian basis. Finally, weighted wafer-level correction terms are converted to XY Cartesian space in order to be applied on the scanner, along with field-level corrections, for future wafer exposures. Since Zernike polynomials have the property of being orthogonal in the unit disk we are able to reduce the amount of collinearity between terms and improve overlay stability. Our real time Zernike modeling and feedback evaluation was performed on a 20-lot dataset in a high volume manufacturing (HVM) environment. The measured on-product results were compared to POR and showed a 7% reduction in overlay variation including a 22% terms variation. This led to an on-product raw overlay Mean + 3Sigma X&Y improvement of 5% and resulted in 0.1% yield improvement.

Slightly dispersive white-light spectral interferometry to measure distances and displacements

Stray light is light that reaches the detector that is coming from sources other than the intended observed scene and is one of the main limiting factors for the performance of Earth Observation Space Optical Instruments. It often sets a working limit on the dynamic range of the system limiting the signal-to-noise ratio and minimizing the contrast between a bright and dark scene. Therefore, there is a need for detailed knowledge of the stray light behaviour of the instrument in order to correct the stray light effects in space instruments as much as possible. Characterising stray light to a high level of detail can be a lengthy task involving detailed scans requiring tuneable lasers or other sources of monochromatic light. This work presents a new method for Space Optical Instrument stray light characterisation based on White Light Interferometry. The characterisation method consists of a broadband light source that passes through a White Light Interferometer to generate spectral fringes in the exiting beam. This light is coupled to an illuminator to the instrument under test. Using the interferometer, Optical Path Difference step-scans are made for each of the spatial points in the instrument field of view. The corresponding interferograms of each illuminated detector pixel contain the full spectral information of direct- and / or stray light reaching that pixel. The dynamic range in the spectral information is increased by multiple exposure levels that are combined to result in high dynamic range spectra. Combining interferometer and spatial scans, the full spectral and spatial sensitivity of each pixel can be determined. This is demonstrated in this paper. The dynamic range in the resulting stray light kernels is shown to exceed seven orders of magnitude, limited only by the optical flux on the detector. To verify the method a dedicated high-contrast scene was used. The same scene is synthetically constructed out of the stray light characterisation data determined with the interferometric method and is compared with the measured verification scene. Differences in stray light between the measured and synthetic verification scene varied over the instrument detector on the order of 0.1% or less close to the bright part of the scene. The study shows that the white light interferometer based stray light characterisation is a promising alternative to the traditional methods involving monochromatic sources. The stability of the white light interferometer can be controlled to such level that accurate measurements are possible.

This paper proposes a novel highly accurate virtual metrology (VM) method based on adaptive online time-series learning. The method comprises a novel time-series prediction algorithm robust against data drifts and shifts observed in semi-conductor manufacturing. The prediction algorithm incorporates time-aware data normalization and adaptive online learning. The time-aware normalizer transforms the data to suppress the effect of drifts. The adaptive online learner captures the time-varying relationship between response and inputs caused by shifts. On both simulated and real process data our approach outperforms conventional VM approaches. When applied to advanced process control (APC) on high-volume production lines at a major semiconductor manufacturer, our VM method substantially reduced variability of process outcomes.

As we are moving towards the 32nm node and beyond, the tightening of CD control requirements is becoming very challenging for the semiconductor industry. Therefore, year after year the need for accuracy in CD measurement is becoming one of the major components in process control. In order to succeed reaching in a near future the 'true CD' that will guarantee high yield and high performances, the semiconductor industry has to quickly move into a new 'Industrial Reference Metrology Model'. Today the industrial configuration of CD Metrology relies on CD-SEM or Scatterometry techniques depending on the criticality of the level that is measured. Both techniques are challenged and tend to show strong fundamental limitations in term of accuracy when being used in their conventional configuration (conventional ellipsometry or reflectometry for Scatterometry and threshold algorithm for CD-SEM). Therefore a huge effort has to be made on reference metrology inside the industrial semiconductor environment. The calibration of CD-SEM through Pitch standard is definitely not enough because it does not guarantee any accuracy onto CD measurement which is the main output. In the same way, using Scatterometry in the industrial environment without several 'golden' standards that have been calibrated with a true reference technique is also definitely limited for future technological nodes in order to avoid correlation between outputs (CD, height, Sidewall Angle) that will inevitably lead to wrong process window definition and bad process control. In this paper we will present some experimental results illustrating in practical terms the needs for future CD metrology and the current limitations of industrial techniques. For example, we will talk about emerging 3D multiwires FET devices which require specific 3D metrology. Based on the conclusion which shows the increasingly need for accuracy in the industrial environment, we will discuss about a potential new Reference Metrology Landscape that take into account the limitations of standard industrial techniques (i.e CD-SEM and Scatterometry). These complementary metrology capabilities which will become mandatory in a near future will help the semiconductor industry to corner well with accurate measurement and Reference Metrology inside the fab.