Search Documents

By T. Johnson, B. Q. Li

"Electromagnetic stirring has found a wide application in metals processing for the benefit of obtaining refined solidification microstructures. This paper presents an experimental study on the melt flow and turbulent behavior in an electromagnetically stirred Woods metal pool. The experimental setup consists of metal bath unit equipped with temperature control to allow both cooling and heating, velocity probe positioning unit and data and signal acquisition and processing unit. The whole measuring system is under the control of an IBM personal computer. Velocity measurements are made using the incorporated magnet probes, which are calibrated using a turntable. The measured results showing the effect of the coil positions, coil arrangements and the applied currents on the melt flow are presented. Both mean flow and turbulence quantities are discussed.1. IntroductionElectromagnetic stirring is widely practiced in the metals industry for melt mixing and for casting microstructure refinement. Industrial examples are taken from induction furnaces where the electromagnetic forces are applied to generate desired melt mixing and also various types of electromagnetic stirrers used in various stages of steel and/or aluminum casting processes to produce fine grain solidification microstructure or to eliminate undesired defects such as TCG (twinned-columnar-grain) structure. For ultrahigh purity aluminum or high strength aluminum alloys containing zirconium or lithium, grain refinement can be difficult with the conventional technique, and electromagnetic stirring provides a particularly powerful means to obtain the refined solidification microstructure [1]. A fundamental understanding of the melt flow and the turbulence characteristics in these magnetically-driven flow systems is of significant engineering importance to both process design and product quality improvement.There exists extensive literature on the magnetically-driven melt flows in material processing systems. Nakanish. et al. [2] were perhaps among the early investigators to mathematically and physically model the magnetically-driven flow phenomena in molten metal systems. Subsequent theoretical and experimental work by Tarapore and Evans [3, 4], Szekely and Chang [5], Hunt and Moore [6] and Szekely and El-Kaddah [7] has contributed very significantly to our understanding of magnetically-driven flow in inductively stirred melts. They have developed mathematical models for the electromagnetic field and turbulent flow in induction furnace and validated the models by either laboratory experiments or direct measurements in the industrial-scale induction furnaces."

Jan 1, 1998

By Kenji Monden

"The creep properties of tin-based lead-free solders Sn3.0AgO.5Cu and SnO.7CuO.1Ni and lead solders 10Sn90Pb and 50Sn50Pb were investigated at temperatures between 298 K and 398 K. The creep-rupture time decreases with increasing initial stress and temperature. The analysis of the solder creep curves uses the Omega method. The creep rate E is expressed by following formula: 1n ? = 1n ?0 + O?, where ?0 and O are experimentally determined. The parameter Eo increases with increasing initial stress and temperature. The activation energy for ?0 is 108 kJ/mol for Sn3.0Ag0.5Cu and 85 kJ/mol for SnO.7CuO.INi. These energies suggest that the lattice diffusion of tin is dominant. The energy of 1OSn90Pb is 37 kJ/mol and that of 50Sn50Pb is 67 kJ/mol. The creep-rupture time is calculated using the parameters, ?0 and O. The calculated creep-rupture time is in good agreement with the measured creep-rupture time.IntroductionLead-free solders are replacing traditional tin-lead solders as part of a major trend toward protecting the earth's environment. However, tin-lead solders are currently needed for a high temperature range. Since creep is the main deformation mechanism in solder joints, understanding a constitutive model for creep and examining creep data for solders are important for reliability analysis of solder joints. A steady-state creep equation is generally used for solders [1,2], and it supposes that the secondary-, or minimum-, creep rate occupies the entire creep life. However, secondary creep accounts for only about 15 % to 40 % of the entire creep life for solders. Thus, the equation may not represent the entire creep life."

Jan 1, 2012

By T. V. Jayaraman

Fe-Ga alloys, with compositions around 27.5 at % Ga, can be heat treated to obtain ? (ordered bcc), DO19 (ordered hexagonal), and L12 (ordered fcc) ordered structures. Our earlier work using polycrystalline samples had shown that magnetostriction is strongly influenced by ordering in this alloy. Here more detailed study of the correlation between ordering and magnetostriction using single crystals of Fe 27.5 at. % Ga alloys is reported. Single crystals of Fe 27.5 at % Ga were grown using Bridgman single crystal growth technique. X-ray and electron diffraction were used to study the ordering in this alloy. Magnetostriction measurements were carried out at no load conditions. Elastic constants were measured using Resonant Ultrasound Spectroscopy.

Jan 1, 2006

By Leonard J. Friedman

There have been considerable changes in systems, equipment design and materials used in the modern Sulfuric Acid process. Many of these will shape the metallurgical sulfuric acid plant for the year 2000 and beyond. Advances have come from all over the world, and include new scrubbing area designs, catalyst contacting devices, silicon stainless alloys for unlined acid systems, energy recovery from acid heat, stainless converters and heat exchangers, etc. This paper describes each of the process "advances" and presents a review and analysis indicating advantages, operating experience and limitations.

Jan 1, 1988

By Ronald K. Riddle

The processing of lateritic ores in rotary drying and preheating/pre-reduction systems has historically been plagued by problems associated with significant dust generation. These problems include handling issues, kiln capacity limitations, negative impacts on furnace stability, a reduction in nickel recovery and higher operating costs. The dust is formed in the rotary devices during heating and tumbling due to the following three mechanisms: (I) mechanical dccrepitation of the ore as a result of tumbling and attrition, (2) thermal decrepitation during rapid vaporization of internal free and bound water, and the (3) loss of free moisture which functions as a binder. This paper investigates the causes of dust formation, provides methods to reduce dust carryover from the rotary devices, and finally offers methods to recover the dust in the existing pyrometallurgical process. Keywords: Laterite, Drying, Rotary, Dust

Jan 1, 2004

By Shigeo Asai

"Electromagnetic processing of materials is newly proposed to be considered as one of the engineering fields where functions of electromagnetism are utilized for processing electrically conducting materials, such as liquid metals, electrolytes and plasmas. The background requiring this processing and the engineering characteristics supporting it are described. The applicable functions of electromagnetism to material processing are revealed from the view point of magnetohydrodynamics. The existing processes and the proposed ones are classified into utilized functions and engineering applications are discussed. IntroductionElectric energy has been used in the metallurgical field for the processes of melting, refining and solidification. In these cases both the magnetic field which is induced by an imposed electric current, and the electric current are applied on electrically conducting materials such as molten metals. Thus, it turns out that the magnetohydrodynamic (MHD) phenomena have unconsciously been seen in metallurgical processes using electric current. The academic backbone studying MHD phenomena is the Magnetohydrodynamics which is being developed in the fields of plasma physics, astrophysics, fusion research and nuclear energy.The first attempt where magnetohydrodynamics was intentionally applied on metallurgy was a Symposium of IUTAM (The International Union of Theoretical and Applied Mechanics) in titled as ""Metallurgical Applications of Magnetohydrodynamics"" at Cambridge, UK, 1982 [1]. This timely event stimulated a few metallurgists to recognize the importance of ~MHD and was a good warning to the metal societies in the world. Especially in Japan, the Steelmaking Investigative Subcommittee of Research Planning Committee in the ISIJ (The Iron and Steel Institute of Japan) discussed the application of ~IHD to metallurgy and selected it as one of the four major innovative subjects for future steelmaking technology in Japan [2]. Indeed, the term of ""Electromagnetic Metallurgy"" itself was firstly coined by this subcommittee. Upon the request of a report from the subcommittee, the ISIJ has established a committee on ""Electromagnetic Metallurgy"" at 1985, under the author's chairmanship."

Jan 1, 1988

By R. S. Tahautdinov, A. N. Posazhennikov, V. V. Bastrygin

"The Bakal (South Ural, Russia) deposit of iron ore bearing iron carbonate (siderite) with the capacity of more than 1 billion tones belongs to the MMK. This ore cannot be fully processed via blast furnace technology because of high content of MgO. According to the investigations carried out in the USA and Japan in 1999-2004 the ITmk3 (Ironmaking Technology mark three) RHF technology is a breakthrough in Ironmaking. Four iron ore types (hematite, magnetite, high and low Al2O3/SiO2) were tested. Reduction, melting and slag removal can be achieved in just 10 min. The main objective of the investigation is to establish optimum operation conditions for the production of iron nuggets from iron carbonate bearing ore via the ITmk3 by means of the lab scale testing. Samples of iron ore and carbon reductant were chemically and physically analyzed. Green pellets were processed via a lab tube furnace to simulate RHF conditions. The nuggets and slag were chemically analyzed. This preliminary test work provides valuable information which may be used for large-scale testing in a commercially sized RHF.IntroductionThe ITmk3 process upon which this paper focuses was developed by Kobe Steel in its research facilities in 1996 [1]. After pilot testing at Kobe Steel’s Kakogava Works in 1999-2004 the pilot demonstration plant was built and operations were successfully carried out by Mesabi Nugget joint venture in 2002-2004. Steel Dynamics has committed to building the commercial plant at the Mesabi Nugget site in Hoyt Lakes (Minnesota). Only 4 iron ore types (hematite, magnetite, high Al2O3/SiO2 and low Al2O3/SiO2), as well as 4 coal types with range of volatility, ash and fixed carbon were tested. Results were consistent with nugget characteristics confirming ITmk3’s flexibility with respect to raw material inputs. At the same time, no research in the field of nugget from iron carbonate bearing ore with high content of MgO has been done. Such kind of iron ore is restricted for charging into the blast furnace due to contamination of MgO. The reason for restriction is the limitation of load of MgO in the blast furnace. If a lot of MgO enters the blast furnace, slag becomes viscous. The viscous slag is obstacle for stable operation of the blast furnace."

Jan 1, 2011

By Katsuyoshi Kondoh

The synthesis processing of Mg2Si or Mg2Si/MgO at lower temperature from the elemental magnesium alloy powder/chips and Si or SiO2 powder mixture was established via the repeated plastic working (RPW) based on Powder Metallurgy (P/M) process. RPW is much effective on the mechanical breakage of MgO surface films, which prevent from the solid-state reaction between Mg and Si powder. The solid-state synthesized Mg2Si particle of 200nm-2JIm is extremely fine, compared to that via the conventional casting process, causing the coarse one of 50-100um. This is because of not coarsening during the solid-state synthesis of Mg2Si. By using in-situ forming process of Mg2Si via RPW and pre-heating before hot working such as extrusion and forging, the magnesium composite material with Mg2Si or Mg2Si/MgO dispersions was developed in employing Si or SiO2 particles, respectively. For example, when employing AZ31 chips as raw materials of the matrix, it has low density of 1.85g/cm) and extremely superior mechanical properties, such as tensile strength over 350MPa with 5-10% elongation and Young's modulus of 47-50GPa, to the conventional Mg-Al-Zn alloys. The Mg composite alloy wires and pipes were produced via the hot extrusion process in the manufacturing plant, and also showed excellent mechanical properties.

Jan 1, 2003

By Milton Manrique, Edgar E. Vidal, Patrick R. Taylor

"Ultrafine zirconium carbide powder has been synthesized using zircon concentrate and methane in a thermal plasma reactor. The effect of several experimental variables, such as power used and zircon particle size, on the conversion has been studied. A thermodynamic analysis of the possible reactions and stable compounds was performed. A mathematical model that describes the in-flight decomposition of zircon particles was developed based on a description of the temperature and velocity profiles calculated in the reactor. It was shown that the formation of zirconium carbide from zircon and methane is thermodynamically possible. Complete 'vaporization of zircon in the reactor is predicted when the particle size is below 45flm and the effective power entering the reactor is 15 kW. The powders were collected and characterized by XRD, SEM, TEM and chemical analysis using ICP. Results of the product analysis indicated the presence of zirconium carbide, zirconium oxide, silicon carbide and silicon oxide.IntroductionIn recent years there has been an increased interest in the production of nanometric sized powders of ceramic materials for use in advanced materials applications. One possible method to produce these powders is with a thermal plasma reactor, which offers high temperatures and fast quenching rates. This technology has been called - Thermal Plasma Synthesis - [1].Zirconium Carbide (ZrCx) has typical properties of transition metal carbides: elevated hardness - between alumina and diamond - a high melting temperature of 3420°C and high chemical stability at room temperature [2]. ZrCx has a Bl crystal structure with deficiencies in the number of carbon atoms in the cell with x between 0.6 and 1 [3]. Different methods have been used to produce ZrC, one method is by direct reaction of zirconium hydride or halide with carbon. Other methods used are direct reduction and carburization of zirconium oxide (Zr02) and reaction of metallic zirconium with a caburizing gas such as acetylene [3]. Recently small films of ZrC have been obtained using a sputtering device [4,5]. ZrC has been used in nuclear plants as a low neutron absorbing and high mechanical resistant refractory material and also as a constituent in cathodes made of UC-ZrC [6]. It is also used in the form of tubes, boats and crucibles to handle molten metals and as part of electronic devices used for thermoionic transducers. Recently ZrC has been used along with zirconium diboride (ZrB2) to produce composite materials [7,8]."

Jan 1, 1996

By Guohua Bai, Lin Zhang, Shan He, Xuling Chen, Min Gan, Xiaohui Fan, Yanbin Tang

"For reducing the dust content in flue gas during copper smelting in the Ausmelt, technologies and mechanism of strengthening the pelletizing of copper concentrate were studied. The results show that adding bentonite, organic binder or composite binder can effectively enhance the strength of green pellet and reduce the rate of shedding powder in drying process, and composite binder works best. When adding 1 % of composite binder, the drop strength of green pellet increased from 1.9 times/P to 4.5 times/P, the rate of damage dropping from 2 m reduced from 50.12% to 13.08% and the rate of shedding powder reduced from 60.28% to 25.13% at drying temperature of 400111. In addition, the mechanism of binder absorbed in the surface of mineral was investigated by infrared spectroscopy. The results showed that composite binder absorbed in the surface of copper concentrate was in the form of chemisorptions, which made hydrophilic enhanced, while bentonite absorbed in the surface of copper concentrate was physisorption.IntroductionWith the technologies of copper smelting developing, the equipment of copper smelting steps into a large-scale and efficient level in recent years. As Ausmelt smelting is characterized by being efficient, environmentally-friendly and energy-saving, it is generally recognized by the enterprises of copper smelting [l]. The process of Ausmelt smelting starts with materials mixing, and then pelletizing. From this point onwards, pelletized balls are charged to the top of the furnace, undergoing drying and smelting reaction[2]. Since copper concentrate is fine-grained, directly adding into smelting furnace can make it blown out of the furnace together with the airflow upwards, rendering flue gas high in flue dust, which can cause the loss of concentrate greatly, and make smelting environment poor as well. The concentrate transported into the furnace and smelted after disc balling is conducive to improve the capacity of melting furnace during copper concentrate smelting, reduce the powder shedding rate and improve enterprise benefits."

Jan 1, 2013

By Yuri Bychkov

The Electron Beam Melting is the most efficient technology for metals and alloys refining. For ,secondary metals processing the Electron Beam Skull Melting (EBSM) with the electromagnetic stirnng (EMS) of melt in the crucible was shown to be the most appropriate. The copper produced by EBSM with EMS possesses higher density and electric conductivity in comparison with other refining methods. The details for high power electrical machines were cast of the copper waste refined by EBSM technology.

Jan 1, 1995

By S. G. Kumar

Experimental studies on high carbon steelmaking through LD Converter - Vacuum Arc Degassing - Bottom Pouring route to produce low oxygen steels were carried out. The results indicated that it is possible to produce steel with a total oxygen content of less than 15 ppm. Thermodynamic and mass balance calculations are made on low-carbon steels that are treated by a complex deoxidation practice based on primary deoxidation of aluminum, manganese and silicon additions; followed by secondary deoxidation of vacuum degassing with simultaneous additions of aluminum. The results showed that primary deoxidation is mainly due to aluminum and the effect of carbon under reduced pressures on deoxidation of molten steel that is already killed by aluminum is negligible. The effect of process parameters on total oxygen content of steel, like carbon content, slag composition and inert gas purging, are discussed.

Jan 1, 1994

By J. P. Gu, C. Beckermann

"Melt convection and macrosegregation in casting of a large steel ingot is numerically simulated. The simulation is based on a previously developed model for multicomponent steel solidification with melt convection, and involves the solution of fully coupled conservation equations for the transport phenomena in the liquid, mush, and solid. Heat transfer in the mold and insulation materials, as well as the formation of a shrinkage cavity at the top are taken into account. The numerical results show the evolution of the temperature, melt velocity, and species concentration fields during solidification. The predicted variation of the macrosegregation of carbon and sulfur along the vertical centerline is compared with measurements from an industrial steel ingot that was sectioned and analyzed. Although generally good agreement is obtained, the neglect of sedimentation of free equiaxed grains prevents the prediction of the zone of negative macrosegregation observed in the lower part of the ingot. It is also shown that the inclusion of the shrinkage cavity at the top is important in obtaining good agreement between the predictions and measurements.IntroductionMacrosegregation in large steel ingots is one of the most well-known and now classical problems in the field of solidification and casting [1]. Commonly found macrosegregation patterns in steel ingots are a positively segregated zone at the top, negative segregation near the bottom in the equiaxed zone, inverse segregation near the ingot surface, V-segregates along the centerline, and A-segregates in the columnar zone [2, 3]. It is well established that the macroscopic transport of species leading to macrosegregation can occur by a variety of mechanisms, including thermally and solutally driven natural convection of the melt in the mushy zone, flow due to solidification contraction, and the sedimentation of free equiaxed crystals. Fredriksson and coworkers [ 4,5] provide recent overviews of the solidification transport phenomena and macrosegregation in ingot casting."

Jan 1, 1999

By Takeshi Azakami

By combining the iron reduction process with the usual rotary kiln process, the direct production of metallic zinc from EAF dust is proposed. After thermodynamic estimation and desk-top experiments it is confirmed that the process is hopeful under suitable conditions. In order to get some practical test data the enlargement of the process scale was planned and the bench scale tests were carried on successfully. In these experiments a zinc ingot of PW grade was produced from EAF dust.

Jan 1, 2003

By R. M. Westcott

A two-electrode, direct current, slag resistance furnace (SRF) has been constructed and operated to selectively reduce the zinc and lead contents of non-ferrous slags. The experimental SRF is equipped with multiple methods of gas/solids injection for both mixing and reaction. A computer is used to position the electrodes and for data acquisition. A geometric factor correlation was previously developed at Columbia University for calculating the electrical resistance on the basis of furnace geometry and slag conductivity. The experimental SRF was used to test this correlation.

Jan 1, 1994

By Mei Yang, Wei Gao, Jian Yang, Qingcai Liu, Wenchang Xi

"TiO2 has been loaded on cordierite porous ceramic to prepare the SCR catalyst carrier by using both APCVD and sol-gel technology. In APCVD method, TiC14 and O 2 were used as reaction precursor. TiC14 was gasified by using N2 as carrier gas which was controlled to 500sccm. Its reaction temperature and reaction time were limited to 773K and 30min respectively. The physical properties, SEM and SEM-EDX of the TiO2-cordierite carrier and the properties of SCR catalysts prepared by these carriers have been tested. The tests showed that TiO2 loaded by APCVD method were of graininess shape and even accumulation. The loading rate of TiO2 and the specific area of carrier sample were increased 114% and 21 % respectively comparing with sol-gel technology. In the result of sol-gel method, the maximal difference of Ti content was 66.92%, while the result of APCVD method was 18.89%. It demonstrated that TiO2 was uniformly attached with the cordierite ceramic by using APCVD method, which would ensure more active ingredients well loaded on the Ti02-cordierite carrier. This difference made the denitrification ratio of catalyst prepared by APCVD method was 8.1 % higher than that of catalyst prepared by sol-gel method.IntroductionIn recent years, the emission of NOx, which were caused by coal-fired power plants, has gotten more and more attention. SCR (selective catalytic reduction) is the most effective technology to DeNOx [ll. The core technology of it is the catalyst. The SCR catalyst usually uses Ti02 as the carrier, and V2O 5 as the active component [2, 3] . However, the thermal stability and mechanical properties of TiO2 carrier are unstable, which would influence the performances of the catalyst. The commercial catalyst generally uses WO3 or MoO3 to promote its heat stability at present [4-7] Besides, TiO2 is expensive. Compared with TiO2, the cordierite is of good thermal stability and lower price. So loading Ti02 on cordierite porous ceramic as carrier to produce SCR catalyst is of good prospects [S, 91 • Currently, TiO2 is usually loaded on cordierite by the sol-gel method. But some research has shown that the sol-gel method easily reduced the loading performances of TiO2, such as distribution, combination, etc. APCVD (atmospheric pressure chemical vapor deposition) has many advantages to preparing TiO2 films, such as high bonding strength, fast reaction rate and so on [l0-13l_ In this paper, TiO2 has been loaded on cordierite porous ceramic to prepare the SCR catalyst carrier by using both APCVD and sol-gel methods. The properties of the carriers prepared by the two processes have also been compared."

Jan 1, 2011

By Bo Zhou

During secondary aluminum recovery, complex aluminum scraps are melted and refined often in a rotary melting furnace and the aluminum metal has to pass through a molten salt layer and melt down in a bottom aluminum bath. Previous experimental study and industrial observations have indicated that salt shell formation and re-melting on the metal solids in the early stage plays a critical role in the overall melting process of the scrap. In the present paper, mathematical models are developed to simulate the process of salt shell formation and its re-melting, as well as aluminum melting in the molten melt. Influence of the salt layer properties, the residence time of the metal particles in the salt layer, the particle size, the temperature of bulk melt were investigated. Reasonable agreement with the measurement data was reached. The developed melting model will be used to construct a population balance model of the scrap melting behavior for a rotary melting furnace, taking into account the distributed characteristics of the scrap. Finally, an overall process model based on Computational Fluid Dynamics (CFD) for the rotary furnace will be constructed, coupling with the population balance model for the scrap melting.

Jan 1, 2003

By Vaughan R. Voller

"Numerical methods for dealing with a temperature dependent conductivity are discussed. An approach based on a local application of a Kirchhoff transformation at the discrete level is presented. The application of this transformation - that does not require any inversion- is demonstrated on solving a two-dimensional test problem.IntroductionMuch is said in the recent literature on the modeling necessity of obtaining appropriate thermo-physical properties; in particular the requirement to account for properties that are ""strong"" functions of the depended variables, e.g., Temperature. Of course obtaining appropriate thermo-physical properties serves no purpose unless they can be accurately integrated into a discrete numerical calculation. The purpose of this paper is to examine how temperature dependent conductivities that exhibit discontinuities or rapid changes can be handled in a numerical code.The discussion will be made in the context of a control volume finite element method [1]. After a brief outline of control volume finite element fundamentals basic methods of dealing with discontinuous or rapidly changing thermal conductivity are discussed. This is followed by the presentation of the recently proposed local Kirchhoff method [2,3], The paper concludes with an investigation of the performance of the various conductivity schemes in solving two dimensional problems. This testing clearly shows the advantages (accuracy and implementation) in using the local Kirchhoff method."

Jan 1, 2001

By Hong-ying Xia, Bao-bao Wang, Jin-hui Peng, Bing-guo Liu, Li-bo Zhang, Zhi-qiang Li

Based on optimization using response surface methodology designed by Box- Behnken, to study the interaction of wet industrial salt with moisture drying temperature (50-90 ); duration (10-50s); and the weight of the materials (20-60g). The results show that: the experimental parameter conforms to the quadratic equation model to influence the moisture content and the dehydration speed., in analyzing the significance and the interaction of various factors, the optimum by the Design Expert software conditions for drying with microwave radiation include temperature of 90 , duration of 10s, and sample mass of 26g. Verified by experiment, the moisture content after drying is less than 0.25% and the dehydration rate reached 1.37g/min, between the predicted value and the actual value confirmed good fitting degree.

Jan 1, 2015

By Y. H. Su, C. Y. A. Tsao

"The solidification behaviors and velocity of atomized droplets during flight is formulated by a mathematical model, in which the concept of transient nucleation is applied. During the flight, the droplet velocity shows a maximum, followed by an inflection phenomenon. The velocity of large droplet is lower than that of small droplet at shorter flight distance, and higher than that of small droplet at longer flight distance. The effect of the gas flow pattern is more significant at longer flight distance than at shorter flight distance at a given flight distance, large droplet has higher temperature than small droplets prior to nucleation. Large droplet has higher nucleation temperature than small droplet. Medium size droplet flies over the shortest flight distance before the nucleation starts. At the end of recalescence, large droplet has less solid fraction than small droplet has.IntroductionTraditional concept of steady state nucleation was used in all the previous models describing the solidification of molten droplets. [1-5] However, during atomization, the solidification rate of the molten metal droplets is very high, and therefore there is very little time for the nucleation event to occur. In the present work, the concept of transient nucleation [6] is considered to model such short nucleation event during solidification of the atomized droplets. The solidification behaviors and velocity of atomized droplets are modeled and the effects of the atomization gas flow pattern are also presented."

Jan 1, 1997