A Review on the Sintering and Microstructure Development of Transparent Spinel (MgAl 2 O 4 ) Ivar Reimanis w Metallurgical and Materials Engineering Department, Colorado School of Mines, Golden, Colorado 80401 Hans-Joachim Kleebe Institut fu¨ r Angewandte Geowissenschaften, Geomaterialwissenschaft, Technische Universita¨ t Darmstadt, 64287 Darmstadt, Germany A review of the densification mechanisms and the microstruc- tural development for transparent spinel made by free sintering and by hot pressing is given. The paper is divided into two main parts. The first part considers spinel without any sintering ad- ditives because there still is some controversy concerning the role of cation stoichiometry on sintering and grain growth. The sec- ond part discusses the role of the classic sintering aid, LiF, in processing transparent spinel. LiF is shown to have multiple be- haviors: (1) it initially wets spinel and forms a liquid phase at relatively low temperatures, which affects early-stage densifi- cation and also grain growth; (2) upon cooling from intermediate temperatures, or even from higher temperatures if microstruc- ture evolution (e.g., formation of closed porosity) prevents vol- atization, the LiF-containing liquid dewets and resides in isolated pockets; (3) LiF alters the cation stoichiometry, thereby enhancing diffusion via an increase in the concentration of ox- ygen vacancies; this affects both the densification rate and grain growth; and (4) it reacts with impurities in the system, thereby acting as a cleanser. For the production of transparent spinel, it is critical that LiF or associated reaction products not be re- tained as a secondary phase. I. Introduction N UMEROUS studies have been performed to understand the densification and microstructure evolution mechanisms of magnesium aluminate spinel, MgAl 2 O 4 (hereafter termed ‘‘spi- nel’’). 1–28 Some of these are generic and focus on fundamental mechanisms that have broad applications as refractories, fusion reactor components, and electronic ceramics. Others are driven by the desire to make transparent spinel. Applications for trans- parent spinel include infrared domes and windows for missiles, transparent armor for air and ground vehicles, optical lenses, laser host materials, windows for lasers, optical heat exchangers, windows for radio frequency powder injectors, plasma diagnos- tic devices, watch crystals, sight glass for high pressure or tem- perature vessels, high pressure arc lamps, and chomatographs. Much of the research has been motivated by the need to obtain transparency across the visible and/or infrared electromagnetic radiation spectrum. 29–31 The latter need implies that a residual porosity of less than about 0.1% and pore dimensions not greater than about a tenth of the wavelength of the light of in- terest are required; for infrared applications, this means that pore diameters should not exceed about 20 nm. 29 This stringent requirement on the allowable level of porosity has provided considerable challenges for reliable manufacturing, and has driven up the cost by mandating that parts be hot-isostatically pressed. The fabrication of large parts has been particularly difficult. Recent efforts have been focused on minimizing final grain size for mechanical strength improvement. 3 A focus on late-stage sintering has occupied significant research that the present paper attempts to summarize. In addition to the removal of porosity, microstructure evolution is also governed by pro- cesses during late-stage sintering, namely dopant-grain-bound- ary and pore-grain-boundary interactions. 19 Some consideration must also be given to the source of the spinel powder, because different methods used to synthesize spinel lead to variations in powder characteristics that may gov- ern the densification behavior, microstructure development, and final properties, as with other ceramics. In particular, the par- ticle size, size distribution, morphology, purity, and stoichiome- try are known to be important. 7,8,30,32,33 Spinel starting powder may be synthesized by several chemical routes, typically using solutions of Mg and Al nitrates or sulfates, followed by copre- cipitation and calcination. The most commonly used commer- cial powder, made by Baikowski Incorporated, is synthesized through the coprecipitation of Mg and Al salts. This is followed by desulfurization, further calcining, and, finally, jet milling to adjust the particle size and surface area. Typically, small amounts of sulfur remain in the commercially available pow- der. Other techniques used to synthesize spinel powders include modified Bayer-type processes, metal alkoxide hydrolysis, var- ious sol–gel-type procedures, and a chemical technique that was shown to have tremendous versatility in terms of controlling stoichiometry, particle size, size distribution, and shape, as well as being able to achieve high levels of purity. 34 Like in many ceramics, astoichiometry typically leads to an increase in the concentration of vacancies and, hence, and in- crease in diffusion rates, 35 as discussed in more detail below. While the presence of impurities (i.e., not including intentional additives such as sintering aids) may influence the stoichiometry, its main effect appears to be that impurities act as direct light scattering sites in the final part 7,8 ; in any case, their influence on stoichiometry has not been systematically studied. Chiang and Kingery 32 report that changes in purity in their investigation had no measurable influence on diffusion kinetics during grain- growth studies, and it would seem the same should be true for sintering. On the other hand, it is possible that impurities seg- regated at grain boundaries alter the grain-boundary structure, which in turn could significantly impact grain-boundary mobil- ity. Indeed, for alumina, it has been well documented that rel- atively small amounts of impurity (hundreds of ppm) may lead to the formation of a grain-boundary liquid phase that then dramatically increases the boundary mobility. 36–38 Clearly, care- R. Bordia—contributing editor Presented at the International Conference Sintering 2008, November 16-20, 2008, San Diego, USA. This work was financially supported from the U. S. Army Research Office under grant # W911NF-06-1-0311. w Author to whom correspondence should be addressed. e-mail: reimanis@mines.edu Manuscript No. 25570. Received November 30, 2008; approved March 21, 2009. J ournal J. Am. Ceram. Soc., 92 [7] 1472–1480 (2009) DOI: 10.1111/j.1551-2916.2009.03108.x r 2009 The American Ceramic Society 1472