You are on page 1of 7

12th A-PCNDT 2006 – Asia-Pacific Conference on NDT, 5th – 10th Nov 2006, Auckland, New Zealand

QUANTITATIVE NDE OF THE YOUNG'S MODULUS OF FCC POROUS METALS
O. Yeheskel

Nuclear Research Centre Negev P.O. Box 9001, Beer Sheva, 84190, Israel Abstract
The present work manifests the results of the dynamic Young's modulus, E, after cold compaction and after sintering of partially dense copper and silver compacts. For both metals, the results show that E in the porous compacts depends not only on the attained density, but also on the samples' consolidation route. For porous metals, it is demonstrated that the normalized longitudinal sound wave velocity, VL*, is more appropriate than porosity, for quantitative nondestructive evaluation (NDE) of the normalized Young's modulus, E*. The proposed model uses, VL*, to evaluate: E* ≅ [αVL*3+(1−α)VL*2]C(ν), where α is a constant defined from the compaction behavior, and C(ν), is a function of the Poisson’s ratio, ν. This model was further verified for the normalized Young's modulus of porous nickel compacts.

Key words: Metals, Young's modulus, sound velocity, porosity, nondestructive evaluation

1. Introduction
The elastic moduli of porous materials have been studied both theoretically and experimentally during the last six decades [1-4]. Having porosity defined as the volume fraction of pores in a solid bulk, one may intuitively assume that the elastic moduli of porous materials decrease with porosity. This assumption was verified both theoretically and experimentally [1-9]. Various equations were put forward to describe the elastic moduli-porosity relations [2-9]. Such expressions devised for use as quantitative nondestructive evaluation (QNDE) of the elastic moduli of porous materials. Recent reports clearly indicate that the elastic moduli of porous materials are dependent not only on their density but also on the processing route that was followed [4,10-12]. The outcome of this finding is that for similar porosity values, there are various elastic moduli values [3,10-12]. These findings make the NDE based on porosity, inadequate. Recently a QNDE approach for the elastic moduli of porous iron was introduced [4]. In this approach, a single sound wave velocity (SWV) is used for evaluating the elastic moduli of the porous metal. The present study uses the same approach for the evaluation of the elastic moduli of few porous face centered cubic, FCC, metals (Ag, Cu, and Ni). The Young's moduli presented in the present work were calculated based on measuring the density , ρ, and time of flight or resonant frequency. The basis of dynamic methods is measuring either the resonant frequency or the time of flight of an acoustic wave traveling in the material. Using the samples dimensions, the sound velocities are derived [13]. 2.1 SWV and elastic moduli relationships For isotropic and homogeneous polycrystalline materials the relation between the shear SWV, VT, and the shear modulus, G, is [14,15]:

VT = (G ρ ) 1 / 2

(1)

For the longitudinal SWV, VL,, the relation to the elastic properties is[15]:

VL = (E ρ)1 / 2[(1−ν ) (1 +ν ) (1− 2 )]1 2 ν

(2)

2. Methodology

where E and ν are the Young’s modulus and the Poisson’s ratio, respectively. Poisson's ratio is defined as ν=E/2G-1 [14,15]. For manifestation purpose, we use the normalized properties e.g.: the Young's modulus, E*=E/EB; SWV, VL*=VL/VLB; and relative density, ρ∗=ρ/ρth. Where EB, VLB and ρth are the properties of the

VL. where Tm is the melting temperature in K [19]. 1a and Fig. 270-405°C 2 min HIP 545°C Buch et al/ 1969. underwent sintering [10]. 945°C 0. for Cu sintered for 2. E=ρVS2 [19].8 0. In the current work. ρth.4 0. Few features are apparent in Fig. 2. or sintered according to the procedure given elsewhere [11].2 a CIP 473°C 11 min.5 1. For the former two materials.5-0.24ρ* (see later in Fig. 1a. VS.6 0. [19] for long rods. 1b show the variation of the normalized Young's modulus. Table 1. Besides these samples.1. gives the room temperature.18].3 Nickel VL* 0.22].7 ρ∗ 0.0 CIP 473°C 11 min. 1: The change of E* (a) and of VL* (b) with the relative density of Cu samples.1 Copper For the elastic moduli of Ag and Cu we used the pulse-echo method described elsewhere [10. Ag. Cu and Ni. and elastic moduli for pore-free. often refer to as "bar velocity" [15]. VL* with the relative density.8 0.21]. E*. For Ni the sonic velocity was used [20.2 b 0.9 1. using the density and the sonic velocity. Results 3. and Ni [17].0 0. [16]. More results on AAC samples appear elsewhere [22]. 470-816°C 2 min.3. the Young's modulus of our samples sintered at 0.6 E* 0.3 127. Since Buch et al.345 4726 2298 Ni 8910 80 214 0. the powder was acid treated in a process designated as acid assisted consolidated (AAC) [10. at 0.9 Tm. Table 1: The theoretical density.1 Copper Copper samples were either cold isostatically pressed (CIP'ed).2 Young's moduli measurements Ni powders were CIP'ed in the pressure range of 100-1330 MPa [20]. Cu.336 6040 3000 4900 Fig.3 Samples preparations 0. 270-405°C 2 min HIP 545°C Buch et al/ 1969. Few samples were cold pressed without AAC. sintered at 0. respectively.6 0.9 Tm for few hours. ν~0.1 0.8Tm for few minutes are similar to those of Buch et al. the elastic moduli.4. and the sound velocities of pore-free polycrystalline aggregates of: Ag [16].0 0.9 1. 2. hot isostatic pressing (HIP) at 545°C and at 50 MPa was undertaken.0 2. 945°C 0. the properties were calculated from single crystal elastic moduli constants. we use the longitudinal SWV.21]. did not provide ν values.6 0.8 0. 2. Prior to powder pressing. New Zealand pore-free solid bulk.1 0. Other results for copper attained from Buch et al. ρT G E ν VL VT VS [kgm-3] [GPa] [GPa] [ms-1] [ms-1] [ms-1] Ag 10530 29. 5th – 10th Nov 2006.7 ρ∗ 0.10+0.12th A-PCNDT 2006 – Asia-Pacific Conference on NDT. and of the normalized longitudinal SWV. we used approximate ν based on our results in the relative density range 0. Some samples either cold pressed or AAC. using the Voigt-Ruess-Hill (VRH) approximation [16].6 81. VS results.8 2.0 0. The details of sintering process appear elsewhere [20.0 Fig.8 and 16h. Sintering of the samples was conducted in the temperature range of 300-900°C. This similarity probably stem from the similar nature of the interparticle boundaries in the two . For Ni the results are from [17]. 470-816°C 2 min. In order to use Buch et al.369 3686 1677 Cu 8960 47.3.75<ρ*<1.5 0.4 0. 4a). theoretical density.1 Change of E* and VL* with relative density 3. 3. 1.2 Silver Ag samples consolidated by cold pressed at various pressures. we transform them to VL using [15]: VL = VS [(1 −ν ) (1 + ν ) (1 − 2ν )]1 2 (3) In order to transform VS to VL values one should use the Poisson's ratio..3. Auckland. For comparison we present the results of Buch et al.11.

1. Details of these functions for the various elastic moduli for both VL and VT. a description of any elastic modulus utilizing.9 1. ∆VLCS * .7 0.0 0.9 1.7 ρ∗ 0. 3.6 E* 0.9 1. 1 Fig.6 E* 0.2 0. Hence.4 0.0 Fig. New Zealand cases.8 0. is similar.1. 3. The fact that the results of Dariel et al. That this is due to the physical contact at the interparticle boundaries of the CIP'ed samples[11].8 0. 3 shows the change of the normalized Young's modulus with the normalized density of Ni samples in the sintered state. to avoid redundancy.4 0.5 1.8 0.2. The sharp increase of E* after short time sintering at 0. becomes possible.8 0. lie along the same line of our results designated as powder A'. can be determined from two independent elastic moduli [14].6 A' NT NT-S1 NT-S2 HT Anderson 1965 Dariel et al.0 0. stem from the similarity in preparation procedures. Similar description holds for the elastic moduli [4. ρ*.11].12th A-PCNDT 2006 – Asia-Pacific Conference on NDT.2 a 0.8 0.0 0. 1b). is the increase of the normalized SWV due to the sintering process. 2b. If one succeeds expressing both. in Fig.5 0. (b) The change of VL* with ρ∗ for the same samples. It is evident that the change of E*. 1999 (6) * is the normalized SWV of the CIP'ed sample and the term. The differences in samples preparation that yield the variation of the results extensively described elsewhere [10]. ρ(VL) and ν(VL). 1999 A' NT NT-S1 NT-S2 HT Anderson 1965 Dariel et al. Any elastic modulus. and VL* with ρ*. the VS* vs. Auckland. ρ* plot. 1. Another finding is that the CIP'ed samples have much lower Young's modulus than the sintered samples.21]. M. 3. 5th – 10th Nov 2006. using single SWV and one independently measured elastic modulus [14]. Similar behavior occurs for the change of VL* vs. 3. One may also represent any elastic modulus.3 Nickel C L 0.2 Silver The variation of the E*.0 0. ρ*.2 Quantitative NDE VL * ρ∗ 0.4 0. and of VL* with relative density is given in Fig.5 0. resulting in improve of the elastic moduli [10.1 Basic approach Nondestructive evaluation (NDE) is an important technique for assessing materials properties and structural integrity. are given elsewhere [4]. we omit in Fig. we rely on . We start with an expression of the density in terms of SWV: where VLS * is the normalized SWV of the sintered sample. 2: (a) The change of E* with ρ∗ of porous Ag samples. [22]. 2.6 0.8 0. and VL* vs.6 0. the term V ρ = ( aVL + b) 1. M is [4]: M = ρV L2 f LM (ν ) (5) V LS * = V LC * + ∆V LCS * (4) where f LM (ν ) is a function of the Poisson's ratio that correlates M with VL.2 b 0. The linear relation depends also on the sintering state [4]: the theory of elasticity. Based on the similarity in the change E*.0 0.0 0. It seems that there is a linear correlation between VL* and. We recently claimed that the nature of the interparticle boundary change upon heating. 2a and Fig. CIP'ed state and in two intermediated states of heating [20. respectively.0 Fig. VL .11].0 CIP Sinter 900°C Sinter 600°C Heating to 300°C Heating to 500°C 3. For quantitative NDE of the elastic properties of porous substances.6 0.7 ρ∗ 0. ρ* (Fig.5 Tm is attributed to neck formation as verified by TEM work [11].

respectively. The calculated m values are about 0. 4b and Fig.9 1.25 0.9 1.10 2 R = 0.30 0.15 0.8 0.35<VL*≤1.6 0. The power relationship is [4]: Heating to 0.0 0.00 0.095 L For the case of normalized Young's modulus.2 0. By using the normalized version.05 R2 = 0. For the purpose of our QNDE approach.996 where C(ν) = (fLE/ fLEB) where fLEB is the function for the case of pore-free material [4]. 1999 0. (2) ones gets the Young's modulus: where c and d are coefficients calculated using least square analysis.65<ρ*≤1. Fig.4 0.24ρ∗ + 0. the evaluated Poisson's ratio may be expressed as: ν EV ≅ c + dV L* (12) ρ * ≅ (αVL * + β ) (7) where α = a/(a+b) and β = b/(a+b). Introducing Eq.30 0. and b is the intercept for VL →0. in the SWV range 0.00 0.0 0.66ρ∗ .90 Poisson's ratio 0.93 ν EV ≅ ν BV L*m (11) Poisson's ratio where νEV is the evaluated Poisson's ratio of the porous material.12th A-PCNDT 2006 – Asia-Pacific Conference on NDT. 4a and Fig. For both metals.0. we get E*=E/EB: ν ≅ νΒVL*0.5 Tm Heating to 0.253V * + 0. 1999 0.8 for copper and silver.21]. where a is a slope that represents the change of VL with respect to ρ.4 VL * 0. respectively. The common assumption for porous materials is that the Poisson's ratio is insensitive to density variations [5.30 0.10 0.20 0.6 Tm Anderson 1965 ν = 0.0 a Experimental error Heating to 0.05 b 0.55 Tm CIP HIP 0. 5th – 10th Nov 2006.00 VL* 0.5 Tm Heating to 0. b) with the normalized longitudinal SWV. by either power or a linear relationships of ν and VL*. into Eq.8 Tm 0.8 1.8 1. respectively.8 Tm 0.6 0.6 Experimental error ν = 0.5 0.10 Experimental error A' NT NT-S1 NT-S2 HT Anderson 1965 Dariel et al. we get: pore-free substance. The values of c and d for both Cu and Ag appear in Fig. a) with the relative density.2. Fig.20 0. 3: The change of E* with the relative density of Ni samples [20.35 0.20 0. (8) yield: E ≅ ρ th (αVL* + β )VL2 f LE (ν ) (9) Poisson's ratio 0.6 and 0.30 Poisson's ratio 0. 4: The variation of the Poisson's ratio in porous Cu compacts.0 .05 0. 4b show the change of the Poisson's ratio in copper with the change of the relative density and the normalized SWV.2 0. 0. Fig.6 A' NT NT-S1 NT-S2 HT Anderson 1965 Dariel et al.40 ν = 0. is possible.40 0.0 Fig.20 0.2 Poisson's ratio Now we will show that ν(VL).00 0.55 Tm CIP HIP 0. From Eq.10 a 0.0. For both materials the Poisson's ratio changes by about 100% in the density range 0.81 b 0. 0.28 2 R = 0. we can fairly evaluate the Poisson's ratio.0.6 Tm Anderson 1965 0.35 0. 5b show the changes in the Poisson's ratio for silver as in Fig. Auckland.0 Experimental error ν = 0.5 0.25 0.7 ρ∗ 0. New Zealand Fig. 5b show such relations. 4b and Fig. νB is the Poisson's ratio for the ν = νBVL*0. and m is a coefficient calculated using least square analysis. It is easy to show that β = (1-α). (7) and ρ=ρthρ* . 5a and Fig.19]. 5b.10 0.34VL* + 0. 3.6 E* ≅ [αV + (1 − α )V ]C (ν ) *3 L *2 L (10) R2 = 0.7 ρ∗ 0.99 E = ρVL2 f LE (ν ) (8) where fLE(ν)=(1+ν)(1-2ν)/(1+ν).15 0.8 0. 4.

8 1. (11). 3. 3. gives very good evaluations of all the E* values. the values the slope of the sintered samples above 405°C.4<VL*<0. 3. (12).0 0.2 0.3.II.0 0.I E* QNDE . to E*=0. (10) for which we drive C(ν)based on evaluation of νEV from Eq. New Zealand Fig. Therefore.2 0.8 1.0 0. 4. 6 show the Young's moduli of the Cu compacts shown in Fig. 8 show the Young's moduli of the Ni compacts shown in Fig. Fig.92 which is inadequate.0 0.III Fig. In Fig. (10) for which omit C(ν).II E* QNDE . give good evaluations of E*.07 to 0. were used yielding α=0. The line for Case-I. describes the general change of E* vs.6 L 0. Auckland. (±0. However in the case of Ni we use the sonic velocity for which fSE = 1. 6: Quantitative NDE of E* based on the VL* of porous Cu compacts.2 0. 7. Case II . 5th – 10th Nov 2006.723.2. 7 show the Young's moduli of the Ag compacts shown in Fig.4 0. 8 we used the values of the data points of CIP'ed samples. 1a. Three evaluation lines representing three cases appear in Fig. 8. Case III .12th A-PCNDT 2006 – Asia-Pacific Conference on NDT. The entire E* data points seem to increase monotonously with VL*. (12). 1. e. 6 and in Fig.0 CIP 473°C 11 min.0 0. 8: The change of the normalized Young's moduli of porous Ni compact with the normalized sonic velocity.6 E* 0. 470-816°C 2 min. 6.0 0.8 1.2 0.8 0.3 Evaluation of Young's modulus 3. when plotted against VS*. VL*. For Case-III.4 V * 0.0 A' NT NT-S1 NT-S2 HT Anderson 1965 Dariel et al. describes all the data points with uncertainty of ± 0.1 Copper Fig.4 VL* 0. 7.05. and Case. fLEB is easily calculated by introducing νB (Table 1 ) into fLE(ν)= (1+ν)(12ν)/(1+ν). by more than 0. α=0.4<VL*<0.3 Nickel Fig.15. i.g. Eq.6 E* 0. in the range 0. 3.II E* QNDE .2 0.I E* QNDE .Stand for Eq. 7: Quantitative NDE of E* based on the VL* of porous Ag compacts. 1999 E* QNDE .Stand for Eq.6 E* 0.3.0 0.8.8. when plotted against the normalized SWV. 1.6 0. 945°C E* QNDE . when plotted against VL*.8 0. 270-405°C 2 min HIP 545°C Buch et al. α=0 58. (10) for which the C(ν) values are based on evaluation of νEV from Eq. (10) becomes: E* ≅ [αVS*3 + (1 − α )VS*2 ] (13) The line representing Eq. 1. 5: The variation of the Poisson's ratio in porous Ag compacts. but deviates downward at VL*→1. a) with the relative density.2 Silver Fig. we use an evaluated value of the Poisson's ratio from either Eq.0. (11) or Eq. the slope of the data points designated by A'.3 The term C(ν) The term needed for describing E* in terms VL*. For plotting Fig. For Case II.0 CIP Sinter 900°C Sinter 600°C Heating 300°C Heating 500°C E* QNDE Fig. For the QNDE procedure in Fig.III Fig.e. 1969. b) with the normalized longitudinal SWV. The lines representing Case-I.0 0.05.05).6 0. were used yielding.4 0. but underestimates the measured E* values in the range 0.0 0.521. The line representing Case–III.Stand for Eq.4 VS* 0. 3.8 0.2 0. (13) in Fig. in order to check its significance. Case I . deviations of E* are about 0. 7 shows large deviations of the predicted from the measured values. 2a. the line predicts the data points very well.3. is C(ν)= (fLE/ fLEB). Conclusions . While for estimation of fLE .4 0.

Ratzker. 36: 1219-1225. Academic Press. O. .alline aggregates using single-crystal data”. Dec. 11 Yeheskel. or the sonic velocity. J. "Empirical dependence of elastic moduli on porosity for ceramic materials". The total deviations between the measured and predicted values are in the range of ±0. and Eichmiller. American Ceramic Society. J... 2001. 5: 111-118 1969/70. 13 Standard practice for Measuring Ultrasonic Velocity in Materials. We here by devise the equation that relates the Young's moduli to a single sound velocity.. O.. “Elastic moduli of transparent yttria“. 1972.P. W. I. D. NDT&E International. R.J. McGraw-Hill book Co. 5th – 10th Nov 2006. Shokhat. and Reed. "The quantitative microstructurefield property correlation of multiphase and porous materials".. 2002. S. American Ceramic Society. in Mason. 82(1): 136-144. 43-95. 2 McAdam. 18 Yeheskel. Soc. 66(5): 366-370.and Dariel.. we show that the Young's moduli of some FCC porous metals (Cu. Pinkas M. 5 Wang. Engineering. 6 Rice. Physical and Chemical Reference Data.K. Goodier.19 [4]. Powder Metallurgy International . Nickel and Iron-Nickel Alloys”. 58(9): 458-459. R. 224-226: 835-840. O. 2003. There two possible ways to correlates the Poisson's ratio to VL*.. J..M.. F. W. The second is power relation given in Eq. Auckland. Kernforschungzentrum Karlsruhe March 1989.and Hsu...C. Princeton New Jersey. editor. “Determination of some uses of isotropic elastic constants of polycryst. D.. R.K. Ratzker. M. 20 Cytermann. "Effects of inhomogeneous porosity on elastic properties of ceramics". and introducing the Poisson's ratio based on SWV. Materials Letters 57: 44184423. 17. “Young’s Modulus of Porous Materials. Iron.P."A new way to investigate the dependence of elastic moduli on the microstructure of porous materials". R. Thèse de docteur ingénieur “Applications de la stereologie a l’étude des propriétés physiques et mécaniques de corps poreux”.. 10 Yeheskel. 1958. 45(2): 94-95. Proc. 7 Dean E. J. American Ceramic Society. 2004. 1984. 19: 801-808. “Some relations of powder characteristics to the elastic modulus and shrinkage of sintered ferrous compacts”. 168(4): 346-358. “Elastic Properties of Metals and Alloys. 32 (1): 17-23. Van Nostrand Co. The general equation. 8 Jeong H. M. Ni). Nondestructive Evaluation of the Dynamic Elastic 4 Moduli of Porous Iron Compacts. Materials Science. Such deviations are more than fair for quantitative NDE method. 1999. J. 16 Anderson.. 1951. This becomes possible by introducing an equation that correlates the density and the sound velocity. Key Engineering Materials. G.porosity correlation”. American Ceramic Society.. “Quantitative estimation of material properties of porous ceramics by means of composite micromechanics and ultrasonic velocity”. and Dariel. 1973.05 of E*. 29 (2): 95-101. 2-11. "Quantitative NonDestructive Evaluation (QNDE) of the Elastic Moduli of Porous Al2O3". F. 1965. Yeheskel O. "Evolution of the elastic moduli at the early stage of copper sintering".P. O. L. One is a linear relation given in Eq... 21 Cytermann. G. J. Physical Acoustics and the Properties of Solids. VS..A.. Testing and Evaluation. S. Theory of Elasticity.. 1987. P.." Elastic constants of porous silver compacts after acid assisted consolidation at room temperature". This gives an equation where E* increases monotonously with VL*. 12 Tevet. O. L’Université de Paris sud. Inc.. p. and Lopez.12th A-PCNDT 2006 – Asia-Pacific Conference on NDT. 12(3): 531-616. New York. 15 Timoshenko. A and Goldschmidt. O. (London) Series B 63. Ag. Journal of the Iron and Steel Institute. either the longitudinal SWV. VL. P. 1996. Part I: Theoretical derivation of modulus. M.-III part B. and Yeheskel. Physical Acoustics. depend besides on the density. 22 Dariel. KFK4566. New Zealand In the present study.. P. 14 Ledbetter.P. O. Reviews on Powder Metallurgy and Physical ceramics. "The elastic constants of a solid containing spherical holes". References 1 Mackenzie. 3 (3&4): 205-322. 1962. M. C. Vol. Tevet. 9 Ondrecek..1978.1950.A. Phys. (11). also on their consolidation route.P. "Influence of porosity on elastic moduli of sintered materials". J. J. 3 Knudsen. ASTM designation E 49495 March 1995. Material Science.. J. Another important finding is the proposed evaluation methods of the Poisson's ratio of porous compacts based on the normalized SWV.1983. (12). J. 19 Buch. " Effect of porosity on Young's modulus of alumina”. 1975. for QNDE of the Young's modulus of porous metallic compacts is based on replacing density by a function of SWV.D. Materials Science and. M..1987. 17 Mason W. M. 27-30 .. New York. "Acid-assisted consolidation of powder compacts: . H. J.N.

.12th A-PCNDT 2006 – Asia-Pacific Conference on NDT. 2001. Material.34: 2601-2607. Auckland. 5th – 10th Nov 2006. Science. . New Zealand cold-welding or cold sintering" J.