E驅≈3S Journal of the European Ceramic Society 21(2001)1027-1035 www.elsevier.com/locate/jeurc Colloidal processing of Al2O3-based composites reinforced with tiN and TiC particulates, whiskers and nanoparticles Eric laar a, Mats carlsson b,*, BenoitⅤiⅤie Mats Johnsson Mats Nygren b, Lennart Bergstrom Institute for Surface Chemistry, PO Box 5607, S-114 86 Stockholm, Swe dEpartment of Inorganic Chemistry, Stockholm University, S-1069/ Stockholm, Sweden Received 29 June 2000; received in revised form 13 October 2000: accepted 21 October 2000 o. A colloidal processing route has been developed for the preparation of dense and homogeneous Al2O3-TiN/TiC composites.The ispersion and rheological properties of mixtures of TiN or TiC particulates and Al,O3 particles were investigated using electro- kinetics and steady-shear rheology. We found that well-dispersed aqueous suspensions with low viscosity could be prepared by adding a poly(acrylic acid) dispersant and controlling ph in the alkaline range. This processing scheme was also suitable for pre paration of whisker and nanoparticle composite suspensions. The alumina-based composite suspensions with a secondary-phase concentration of 25 vol. were freeze-granulated and hot-pressed, and the resulting bodies were fully densified with well-dispersed secondary phases. Homogeneous Al2O-TiN nanoparticle composites could only be prepared with additions of up to 5 vol% nanoparticles; higher additions resulted in agglomeration and subsequent grain growth of the nanoparticles. C 2001 Elsevier Science Ltd. All rights rese Keywords: Al2Or-TiC: Al2Oy-TiN; Composites; Microstructure-final: Suspensions; TiC: TIN 1. ntroduction with homogeneous microstructure. Based on thorough powder characterization and on the results of dispersion Mechanical properties of ceramic materials are experiments, we have developed an optimized proces- strongly dependent on their microstructure. The frac- sing scheme and assessed its applicability to systems ture toughness, for instance, can be improved sub- containing TiN and TiC with a wide range of morphol- stantially by ade dding a secondary reinforcing phase to ogies. Thus, the preparation of composites containing the matrix. It has been shown that the addition of a tin micronsized TiN and TiC particles, nanosized Tin par- or TiC phase to an alumina matrix increases hardness, ticles, and TiC and Ti(C, N)whiskers has been investi fracture toughness and thermal shock resistance at gated. The results will be applied in a forthcoming temperatures up to 800oC. In order to optimize the investigation of the mechanical properties(e.g. thermal properties, the reinforcing particulate phase must be shock resistance)of Al_ O -TiN/TiC compacts. Experi- well dispersed in the matrix. Mixing, deagglomeration mentally, our study has included mainly electrokinetic and dispersion of the reinforcing components are com- and rheological characterization of aqueous composite monly performed in a liquid medium, preferably water. suspensions and microstructure evaluation of sintered Additions of dispersing agents and/ or manipulation of bodies by means of scanning electron microscopy the solution properties, e.g. the pH value, are frequently used to optimize the suspension properties In this work, we have attempted to develop an aqu- 2. Experiment eous colloidal processing route for preparating well-dis- persed powder mixtures in the AlO3-TiN/TiC systems, 2.1. Powder characterization which upon densification would yield dense compacts The manufacturer specifications of the starting pow ders are given in Table 1. For simplification we will refer to the nano-and micronsized titanium nitride powders 0955-2219/01/S. see front matter C 2001 Elsevier Science Ltd. All rights reserved. PII:S0955-2219(00)00302-2
! " # $% &' ! ( ! ) # * !" #$ % &' ( !" + ' , ( ---. ' ' / ---. / --- ' 01 2 3 ' ) )2 4 5 ' ) ) )6 ) 7 8 2 2 ) 9 '2: ) ; ) 2 8 0 ) 9 '2: . 5 2 --/ ' < 2 ' 2 )! "* 0. 0. . ! ; . < . . ! ) 2 ' ) ) ) 32 = 3 >--2/ = 32 !3 ) 5 ) 2 1 2 2 8 ' 5 ) 2 = ' ' 5 3 01 ) ; ) 2 # 3 ' ' ) ) 2 6 7 ' 2 ' 6 2 2 7 01 2 3 ) ) ) 5 ' ) )2 +, , !" - ; ' /2 ? ; -,99/,1-/1@ --/ ' < 2 ' 2 A==B <-,99/,6--7--- ( < ) / 6--/7 /-C0/-9 2 ' 2 1 1D " 2 . "" * E 2 2 6!2 72
E. Laarz et al. Journal of the European Ceramic Society 21(2001)1027-1035 TiNnano and TiN, respectively. BET-surface area and Scanning electron microscopy, SEM (JEOL 880, particle-size distribution measurements of the as- Japan), and transmission electron microscopy, TEM received TiC, TIN and TiNnano powders were obtained (JEOL 2000FX, Japan), were employed to gain infor by means of BET nitrogen adsorption analysis(Flow mation on the particle morphology(Fig. 1). X-ray Sorb lI 2300, Micromeritics, USA)and X-ray gravita- powder diffraction(XRD) patterns of the TiC and TiN tional sedimentation analysis(Sedi Graph 5100, Micro- powders were obtained using a Guinier-Hagg focusing meritics Instrument Corporation, USA), respectively. camera with subtraction geometry. Cu-Kal radiation Table I Manufacturer specifications showing particle size and chemical ana the different powders us TIN Tinman Al2O3 H C. Starck Sumitomo STD 120 Grade c AKP 30 cle size(um) (BET equivalent pherical diameter) 0.009 O(wt % Cl(wt%) <50 o'e So 9 Fig 1. Micrographs of as-received powder and whisker materials:(a)TiC; (b)TiN;(c)Ti(C N)whiskers: (d)TINnano particles
' )2 # ' ) # ) 6? < == -- ! F<7 G) ' ) 6< H 9/-- ! = F<7 ' )2 < ) < ! 6( >>- ( 7 ) ! 6( ---?G ( 7 ) ) 6? 2 /72 G) I 6G+J7 H 8* )2 )/ / ! ; ) I ! 822 < 6H )7 822 < 6H )7 822 < 6H )7 < 6( 7 A H <J /- H KA - A 67 /2L /2-9 -2- 6# 5' 7 -2 62:7 -2/9 -2> /2, 0 62:7 /,29 -2-9 -2--, 0 62:7 -2L 0 0 0 62:7 /,2/ 0 0 0 62:7 -2M /2 -29 0 62:7 0 2C 0 < 67 M- 9- N9- NL- ? 2 /2 ! ' B 67 . 67 . 6 7 6 7 . 67 2 /-> ., / - , 0 1 . $ + 2+ 3 +4 5
E. Laarz et al /Journal of the European Ceramic Society 21(2001)1027-1035 (=1.54060 A)was used, and finely powdered silicon bench for 4 h in order to equilibrate.Afterwards, sus- [a=5.430880(35)A] was added as an internal standard. pensions were sprayed through a nozzle into liquid The recorded films were evaluated in an automatic film nitrogen for granulation and were then freeze-dried canner, and the unit cell dimensions were refined with The obtained granules were sieved, and the the program PIRUM. The XRD peaks were identified 0.125<d<0.32 mm fraction was used for the sintering by matching them to ICdd data cards of TiC (o experiments. Rheological evaluation was performed on card No. 32-1383)a=4.3274 A, and TiN (ICDD card suspensions that had been equilibrated by 4 h of mag No. 38-1420)a= 4.2417(1)A. XPS analysis of the Tin netic stirring. All rheology experiments were carried out and TiC particle surface composition was performed with a controlled-stress rheometer(UDS 200, Physica using a Mg-Kg X-ray source and a magnetic collimator Messtechnik GmbH, Germany)equipped with a con- lens(AXIS-HS, Kratos Analytical, UK). The powder centric cylinder measuring geometry. solubility was assessed by determining the Ti con centration in solution after ageing in pH=0.8, 6, 10 and 23. Burnout and hot-pressing 12 solutions for 1750 h. Ti was analyzed with a DCP (direct current plasma) emission spectrometer(Spectra The burnout of paa was studied in a thermo- Span IIIB. SpectraMetrics Inc, USA). Microelec- gravimeter, TG (TAG24, Setaram, France)in Ar-6% rophoresis(Zeta Sizer 2000, Malvern Instruments, UK) H2 atmosphere. Prior to sintering, the organic content was used for electrokinetic characterization of TIN and was burned out in a graphite furnace(Thermal Tech TiC particles suspended in 0.01 M NaCI electrolyte. nology Inc, USA). Freeze-dried granules were poured Zeta-potential measurements of concentrated(5 vol % directly into the die of a hot press(Thermal Technology alumina suspensions(0.01 M NaCl background elec- Inc, USA), and the sample was subjected to 28 MPa at trolyte) were carried out with an AcoustoSizerT 1700C for 1.5 h in flowing argon atmosphere. Densities instrument (Matec Science, USA). We used of sintered bodies were measured according to archi analytical grade cher Merck AG, Germany) for medes' principle. The expected densities of the sintered adjustment of ionic Nacl) and pH (HCI and materials were calculated assuming no reactions to take NaOh) place between the components and by using the follow- TiC whiskers and Ti(C N)whiskers were synthesized ing densities: PAl203=3.965 g cm-3, PTin= 5.22 g at 1425 and 1250 C respectively via the carbothermal cm-3, Pric 4930 g cm-3. The microstructure was vapour-liquid-solid growth mechanism.4. 5 The whiskers evaluated with an SEM(880, JEOL, Japan). The com- have a length of about 10-30 um and an aspect ratio of posites containing TiNnano were thermally etched at 15-50(Fig. Ic). The TiNnano powder is pyrogenic and 1500C for 15 min in flowing Ar in order to reveal their was therefore suspended in a water-ethanol mixture microstructure (95/5 wt %)in a glove box filled with nitrogen to pro- mote a controlled surface oxidation .6 After this treat- ment, the solvent was evaporated at 50oC 3. Results and discussion 2. 2. Preparation and characterization of suspensions 3. Powder characterization Concentrated suspensions were prepa mIxIn The TiC and tin powders displayed similar particle Al2O3 powder with deionized water until the required morphology, particle size distribution, and surface area powder volume fraction was reached. The required Equiaxed grains with sharp edges were characteristic mount of poly(acrylic acid) dispersant(Dispex A40, features of both the TiN and the TiC particle morphol- llied Colloids, USA)had been dissolved in the deio- ogy(Fig. 1). The experimentally determined particle size nized water prior to mixing. The anionic polyelectrolyte distributions and specific surface areas are reported in used(denoted as PAA) is an ammonium salt of poly ( Table 2. The TiC and Tin powders were characterised acrylic acid) with a mean molecular weight of Mw= by Xrd to be monophasic with a cell axis of a= 10 000 and a polydispersity of Mw/Mn=1.56. PAa 4.32882(9)A for TiC and a=4.23910(18)A for TIN. In concentrations are given in wt. with respect to the previous XRD studies of nano-sized TiN with much alumina dry-powder weight. In a first milling step, alu- higher oxygen content (an order of magnitude higher mina and PAa were mixed for 30 min at 400 rpm in a than for the TiNnano used in our study ), traces of ana planetary ball mill(Pulverisette 6, Fritsch GmbH, Ger- tase and a substoichiometric TiO2-x phase could be many)equipped with a 250 ml SiAlON milling jar and detected. 8. 9 XPS analysis of the TiC and Tin powder SiAION milling spheres (0=10 mm). After this first surfaces indicated presence of an oxidized surface layer step, the other solid phase(TiC, TiN, TiNnano or whis- and a 20-25 at. level of surface contamination by ali- kers) was added and the sample was milled for another phatic carbon, which is substantially higher than the 3- 10 min. The obtained suspensions were rolled on a roll 10 at usually observed for high-purity ceramic powders
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
E. Laarz et al / Journal of the European Ceramic Society 21(2001)1027-1035 The solubility of TiC and Tin powders in water is neg- for TiN and TiC powders are somewhat less consistent gible (<6 ppm Ti) over practically the entire pH range: with our results: different grades of Tin powders only under very acidic conditions (pH=0. 8)significant showed pHiep in the range pH=3-5. 8,9, 13, 14and oncentrations of dissolved Ti (100 ppm) could be pHiep 2 has been found for TiC >However, since non- observed. The surface oxides passivate the TiN and Tic oxide ceramic powders are thermodynamically unstable materials, protecting them from further oxidation inin air and water, these differences are not unexpected aqueous environments. and can be attributed mainly to different degrees of surface oxidation. For example, it has been shown ear- 3.2. Dispersion in aqueous medium lier that the synthesis route and post-synthesis treatment of Si3N4 powders strongly affect the surface oxygen The oxidized TiC and tin powders exhibited iden content and the correlated area density of potential ical isoelectric points at pHicp 4.3 in NaCl electrolyte determining surface groups. 16, 17 Accordingly, increasing solutions(Fig. 2). The isoelectric point of TiC whiskers surface oxidation will shift the isoelectric point of the the isoelectric point of the a-alumina powder is much the value for the less acidic oxide TiO2(pHiep o/s ards was determined to be pHiepa43 as well. In comparison, acidic TiN and Tic powders more and more tow higher(pHiep9. 2). However, as shown in Fig. 2, upon Based on our studies of the surface chemistry of TiN addition of 0.5 wt. PAA to the alumina suspension and TiC, we developed the processing scheme depicted the pHiep shifts to acidic values(pHiep 3)due to poly- in Fig. 3, which can be employed for preparation of electrolyte adsorption 10-12 Measurements at various Al O -TiN/TiC composite suspensions and sintered ionic strengths confirmed that NaCl acts as an indiffer- bodies. As shown in detail in previous work, 8-l0 well- ent electrolyte in all cases(not shown). The zeta-poten- dispersed a-Al2O3 suspensions with an inherent suspen tial values for Al2O3 suspensions compare very well with sion pH value of A9 are obtained by simply mixing previous results. Previously reported pHicp-values deionized water, powder, and >0.2 wt. ammonium alt of PAA. according to our processing Table 2 done in the first milling step and in the next milling step Particle size distribution and specific surface area of the as-received tin or tic is added. In this manner heterocoagulation TiC and TiN powders is prevented, as can be seen from steady-shear measure TiN ments on suspensions with 20 vol. solids loading Fig. 4). Before and after adding tin or TiC, the sus- article size distribution(um) pensions show almost no shear thinning, thus indicating a well-dispersed suspension. However, if addition of PAA is omitted, a flocculated strongly shear-thinning ET surface area(m-g-) suspension is obtained(Fig 4). Clearly, the absence of BET equivalent spherical 0.45 0.026 diameter(um) flocculation is related to the electrostatic repulsion between alumina particles with adsorbed PAA layers e-TiC TiC(Whisker) Al O -o AL o,(0.5 wt% PAA) H Fig. 2. Zeta potentials of the powders used in 0.01 M NaCl electrolyte solution
) 6NM 7 ' ) 8 . ) ' ) 68R-2>7 ; ' 6/-- 7 ' 2 3 ' 3 5 ' 2 5,+, #$ 7 " 3 3 8 L2 ) 6? 2 72 8 L2 2 = 68 ,272 8 ' ? 2 -29 2: A 8 ' 68 7 ) ) 2/-/ ! ' ; I ) 6 72 ' ' ) ' 2/-/ A ' ) 8 ' . I 8 8R09 > , / /L 8 2/9 8 ' 3 ) ) I 3 ) I 3 2 ? 3 ) ) <L ) I 3) ) 2/M /C ) 3 ' 3 68 M72/> # ) ' ? 2 ) 01 2 ' >/- 8 ' , ) ) 3 W-2 2: A2 ; 3 2 = ' ) - '2: 6? 2 L72 # 2 8 ' A U ) 6? 2 L72 ) U A ) A ; ' - " 8 23 "/M /29 /2 0 "9- 2 2/ 0 ">L 92L 2L 0 # 6 / 7 2C 2/ LL2, # 5' 67 -2L9 -2C -2-M ? 2 2 S -2-/ ! ) 2 /-- ., / - , 0 1 . $ + 2+ 3 +4 5
E. Laarz et al /Journal of the European Ceramic Society 21(2001)1027-1035 and TIN or TiC particles which, according to Fig. 2, are not change the flow behavior compared to suspensions also highly negatively charged at pH=9 containing 25 vol. micron-sized TiN. a slight viscos- Viscosity measurements on 20 vol. TiC suspensions ity increase at 10 vol. TINnano concentration is obser Fig. 5)suggest that the negative surface charge is high vable, and even higher concentrations lead to nough at pH-77 to prevent homocoagulation. At substantial shear-thinning(Fig. 6). This effect might be lower pH the surface charge is apparently too low, related to a crowding effect or an induced aggregation though, and flocculation takes place. In comparison, the due to changes in suspension pH with addition of H of Tin samples must be increased to 10 to achieve a TiNnano. The pH of an alumina suspension dispersed well-dispersed suspension with a Newtonian shear with 0.5 wt. PAA decreases from 9.1 to 8.2 upon response(Fig. 5). Fig. 6 shows the flow curves of com- addition of 20 vol. TiNnano which can result in posite suspensions containing TINnano. USing nano- enhanced homocoagulation of the TiN particles(Fig. 5) sized TiN up to 5 vol. of the total solids loading does Another factor contributing to the high viscosity might be the interaction between TiNnano particles and PAA macromolecules. Previous results on the electrokinetics Mixing and Dispersing(I) Alos and rheology of TiNnano particles dispersed with poly Ho+ PAA acrylic-type dispersants 9. 4+ imply that PAA can be H 9 adsorbed on titanium nitride. These studies suggested that adsorption occurs even at a ph where both parti- Mixing and Dispersing(Il) cles and polymer carry negative charges. Hence, it is Addition of tintic possible that in our system weak adsorption ( 1.e adsorption below saturation level) of PAA on TiN causes bridging flocculation at high TiNnano particle Freeze Granulation and concentrations and, thus, a higher suspension viscosity Freeze Drying Summarizing the results of our dispersion experi ments, we can state that well-dispersed, low-viscous composite suspensions with 20-25 vol. solids loading (minimum viscosities of the composite suspensions were Sieving Granules in the range 3-4 mPa s)can easily be produced with all investigated alumina-based composite systems by using the same processing scheme. An additional advantage Hot Pressing in Ar of the chosen processing scheme is its tolerance towards at1700°C,28MPa,90 variations in degree of surface oxidation of the non- oxide components, which can be difficult to control in Fig 3. Processing scheme for composite preparation industrial practice A2O325%TN(no。PAA) 102[△42°325%TNQ5w%PA 103E°425%T(a5w%PA) a888. 0.5 wt% PAA 10 000 Rate [s" Fig 4. Steady-shear flow curves of 20 vol. suspensions at pHR9
? 2 ) ' ) 8R,2 & ) - '2: 6? 2 97 ' 8RC2C ' 2 8 ) U 2 = 8 /- ' 6? 2 972 ? 2 M U ' 2 F 9 '2: U ' 9 '2: 2 ' ) /- '2: ' ' 6? 2 M72 I I 8 2 8 -29 2: A ,2/ >2 - '2: 6? 2 972 ' ) A 2 A ' ) ) ) ) > , /L ) A 2 ' 8 ) ) ' 2 8 ) 62 2 ' 7 A U ' )2 < 3 ' -09 '2: 6 ' 0L A 7 ) ' ) ) 2 ' ' 3 3 X 2 ? 2 L2 < ) U ' - '2: 8,2 ? 2 2 A 2 ., / - , 0 1 . $ + 2+ 3 +4 5 /-/