G. de Portu et al Acta Materialia 53(2005)1511-1520 the materials of which each layer is composed [1]. How- size, 0.3 um) were used as raw ceramic materials. Pow ever, it should be noted that the residual stress field also ders, solvent(ethanol and methyl-ethylketone), surfac greatly depends on the geometry of the layered structure, tant (triolein, to enhance the powder dispersion affected by different shrinkage during sintering and CTE the mylar substrate)and 1/3 of the total binder volum o? in particular on layer thickness[2]. The overall stress field, characteristics and to facilitate removal of the tape fror mismatch between constituent phases/layers, mismatch (polyvinyl-butyral, PVB), were first ball-milled with in elastic constants between different phases/layers, and Al_O3 or ZrO, balls for 24 h. Then, the remaining part layers' geometry, may be rather complex and thus difficult of the binder and a plasticizer(dibuthyl phthalate, DBP to predict by theoretical calculations. In order to avoid were added and the mixtures ball-milled again for 24 h. cracking and delamination, a precise control of both Table I lists the formulations of the two slurries pre magnitude and distribution of residual stresses is manda- pared for tape casting, whose contents of ceramic pow- tory. In multilayered ceramic components, the develop- ders were 100% Al2O3(A) and 50 wt% Al2O3/50 wt% ment of a reliable experimental procedure for the 3Y-TZP(AZ). Slurries were filtered and degassed under evaluation of residual stresses is highly desirable. The vacuum for 5 min and tape-casting performed by means development of such a technique may also help to sub- of a doctor blade device onto a mylar substrate. Tapes stantially reduce the computational time required for (200 mm wide) were obtained at a casting speed of complete three-dimensional finite-element calculations 200 mm/min and with a blade gap of 0.8 and 1.1 mm Few techniques are available for assessing residual for A and AZ, respectively. After drying, the tapes stresses in ceramic materials, including X-ray diffraction, had an average thickness of 240 and 350 um for A neutron diffraction and piezo-spectroscopic analyses of and AZ, respectively. Dried tapes were peeled off from photo-stimulated fluorescence or Raman bands. In this the mylar and punched into 34 x 50 mm rectangular contribution, we present measurements of the residual moulds. Tapes were then stacked to form two kinds of stress and its dependence on geometry of multilayer laminates, in which the ratio between a and aZ thick Al,O3/3Y-TZP ceramics using the technique of piezo- ness was different: in one specimen(simply denominated pectroscopy applied both to the chromophoric fluores- A/AZ, henceforth) 13 layers were stacked by alternating cence of Al2O3 and to a selected Raman band of A and aZ layers according to a sequence A/AZ/. AZA; 3Y-TZP. The piezo-spectroscopic technique was first ap- another specimen(A/2AZ)consisted of 9 layers stacked plied by Grabner to the measurement of residual stresses according to the sequence A/2AZ/. /AZ/A. To ensure in Al2O3 [3]. The technique is also valid for Raman bonding among the stacked green-sheet, warm pressing assessments and it has been applied to some selected Ra- was carried out for 30 min at 80C, under a pressure man bands of ZrO: [4]. In the following section, expres- of 30 MPa. Specimens were then placed in a low sions are given relating the spectral shift of fluorescence temperature furnace at 600C for binder burnout. In and Raman bands to both magnitude and versus of the this processing step, heating and cooling rates were set mean normal stress; in addition, general piezo-spectro- at 3C/h. Finally, the laminates were sintered at 1550 scopic relationships are established according to exper oC for I h(heating and cooling rates set at 30C/h) mental calibration procedures [5]. Then, in Section 3, an After sintering the total thickness of the multilayered experimental analysis of residual stress as a function of specimens was 2.76 and 2.9 mm for the A/AZ and A layer geometry is proposed and discussed. The present 2AZ configuration, respectively. At the end of process- analysis extends and, for certain aspects, also improves ing procedure we obtained a laminated structure consist a previously presented analysis of laminate ceramics ing of 7 A and 6 AZ layers in the A/AZ material, and [6]because: (i) the effect of the elastic mismatch, in addi- 5 A and 4 AZ layers in the A/2AZ material. The final tion to CTE, on the microscopic residual stress field be- thickness for A, AZ and 2AZ laminae after sinter tween Al2O3 and 3Y-TZP phase is taken into ing was about 180, 250 and 500 um, respectively consideration;(ii the dependence on the layer geometry is studied in some detail; and (ii) experimental high-res- olution two-dimensional stress maps on areas as ex- Table I tended as the entire multilayered specimen thicknes Slip formulations of the two slurries prepared for tape casting" are. for the first time. collected Constituent AZ or 2AZ 2. Experimental and computational procedures Al,O 3Y-TZP Binder(PvB) 2. 1. Specimen preparation Plasticizer(DBP) Surfactant Al_O3 powder(Alcoa A16: average grain size, 0.3 um), Solvent and a 3Y-TZP powder (3Y-TZP Tosoh: average grain Values in grams.the materials of which each layer is composed [1]. How￾ever, it should be noted that the residual stress field also greatly depends on the geometry of the layered structure, in particular on layer thickness[2]. The overall stress field, affected by different shrinkage during sintering and CTE mismatch between constituent phases/layers, mismatch in elastic constants between different phases/layers, and layers
geometry, may be rather complex and thus difficult to predict by theoretical calculations. In order to avoid cracking and delamination, a precise control of both magnitude and distribution of residual stresses is manda￾tory. In multilayered ceramic components, the develop￾ment of a reliable experimental procedure for the evaluation of residual stresses is highly desirable. The development of such a technique may also help to sub￾stantially reduce the computational time required for complete three-dimensional finite-element calculations. Few techniques are available for assessing residual stresses in ceramic materials, including X-ray diffraction, neutron diffraction and piezo-spectroscopic analyses of photo-stimulated fluorescence or Raman bands. In this contribution, we present measurements of the residual stress and its dependence on geometry of multilayered Al2O3/3Y-TZP ceramics using the technique of piezo￾spectroscopy applied both to the chromophoric fluores￾cence of Al2O3 and to a selected Raman band of 3Y-TZP. The piezo-spectroscopic technique was first ap￾plied by Grabner to the measurement of residual stresses in Al2O3 [3]. The technique is also valid for Raman assessments and it has been applied to some selected Ra￾man bands of ZrO2 [4]. In the following section, expres￾sions are given relating the spectral shift of fluorescence and Raman bands to both magnitude and versus of the mean normal stress; in addition, general piezo-spectro￾scopic relationships are established according to experi￾mental calibration procedures [5]. Then, in Section 3, an experimental analysis of residual stress as a function of layer geometry is proposed and discussed. The present analysis extends and, for certain aspects, also improves a previously presented analysis of laminate ceramics [6] because: (i) the effect of the elastic mismatch, in addi￾tion to CTE, on the microscopic residual stress field be￾tween Al2O3 and 3Y-TZP phase is taken into consideration; (ii) the dependence on the layer geometry is studied in some detail; and (iii) experimental high-res￾olution two-dimensional stress maps on areas as ex￾tended as the entire multilayered specimen thickness are, for the first time, collected. 2. Experimental and computational procedures 2.1. Specimen preparation Al2O3 powder (Alcoa A16: average grain size, 0.3 lm), and a 3Y-TZP powder (3Y-TZP Tosoh: average grain size, 0.3 lm) were used as raw ceramic materials. Pow￾ders, solvent (ethanol and methyl-ethylketone), surfac￾tant (triolein, to enhance the powder dispersion characteristics and to facilitate removal of the tape from the mylar substrate) and 1/3 of the total binder volume (polyvinyl-butyral, PVB), were first ball-milled with Al2O3 or ZrO2 balls for 24 h. Then, the remaining part of the binder and a plasticizer (dibuthyl phthalate, DBP) were added and the mixtures ball-milled again for 24 h. Table 1 lists the formulations of the two slurries pre￾pared for tape casting, whose contents of ceramic pow￾ders were 100% Al2O3 (A) and 50 wt% Al2O3/50 wt% 3Y-TZP (AZ). Slurries were filtered and degassed under vacuum for 5 min and tape-casting performed by means of a doctor blade device onto a mylar substrate. Tapes (200 mm wide) were obtained at a casting speed of 200 mm/min and with a blade gap of 0.8 and 1.1 mm for A and AZ, respectively. After drying, the tapes had an average thickness of 240 and 350 lm for A and AZ, respectively. Dried tapes were peeled off from the mylar and punched into 34 · 50 mm rectangular moulds. Tapes were then stacked to form two kinds of laminates, in which the ratio between A and AZ thick￾ness was different: in one specimen (simply denominated A/AZ, henceforth) 13 layers were stacked by alternating A and AZ layers according to a sequence A/AZ/.../AZ/A; another specimen (A/2AZ) consisted of 9 layers stacked according to the sequence A/2AZ/.../2AZ/A. To ensure bonding among the stacked green-sheet, warm pressing was carried out for 30 min at 80 C, under a pressure of 30 MPa. Specimens were then placed in a low￾temperature furnace at 600 C for binder burnout. In this processing step, heating and cooling rates were set at 3 C/h. Finally, the laminates were sintered at 1550 C for 1 h (heating and cooling rates set at 30 C/h). After sintering the total thickness of the multilayered specimens was 2.76 and 2.9 mm for the A/AZ and A/ 2AZ configuration, respectively. At the end of process￾ing procedure we obtained a laminated structure consist￾ing of 7 A and 6 AZ layers in the A/AZ material, and 5 A and 4 AZ layers in the A/2AZ material. The final thickness for A, AZ and 2AZ laminae after sinter￾ing was about 180, 250 and 500 lm, respectively. Table 1 Slip formulations of the two slurries prepared for tape castinga Constituent Slip A AZ or 2AZ Al2O3 400 200 3Y-TZP – 200 Binder (PVB) 36 36 Plasticizer (DBP) 36 36 Surfactant 6.4 6.4 Solvent 188 188 a Values in grams. 1512 G. de Portu et al. / Acta Materialia 53 (2005) 1511–1520