JOURNAL OF MATERIALS SCIENCE 40 (2005)6363–6365 Effect of solidification cooling rate and phosphorus inoculation on number per unit volume of primary silicon particles in hypereutectic aluminium—silicon alloys M. FARAJI, I. TODD, H. JONES Department of Engineering Materials, University of Sheffield, Mappin Street, Sheffield, S1 3JD, UK Published online: 5 October 2005 Collected data on derived number per unit volume ¯ N V of primary silicon particles in hypereutectic Al-Si alloys show a power relationship with solidification cooling rate ˙ T of the form ¯ N V = A ˙ T n where typically n ∼ 1 and A ≃ 130 mm −3 (K/s) −1 in the absence of phosphorus and A ≃ 720 mm −3 (K/s) −1 in its prescence. Significantly lower apparent values of ¯ N V from one set of results appear to stem from measurement of a mean long dimension rather than diameter of particle sections as well as lower measured undercoolings than in Bridgman experiments at similar ˙ T . C  2005 Springer Science + Business Media, Inc. Size refinement of primary silicon in hypereutectic Al- Si base alloys is a key requirement for meeting prop- erty targets [1–4] and can be achieved by inoculation with phosphorus, as is routinely applied in conven- tional foundry practice [5], or by increasing the cool- ing rate during solidification, which has been applied, for example, in the development of improved perfor- mance in alloys processed by spray forming [6], and other rapid solidification technologies [8]. Arnold and Prestley [9] showed micrographs indicating size refine- ment of primary silicon in Al-16 wt%Si with increase of solidification cooling rate ˙ T in the range 0.6 to 1.5 K/s, but the results were not quantified. Sulzer [10] reported refinement of primary silicon size from 27 ± 4 µm at 3 K/s to 15.5 ± 1.5 µm at 70 K/s for Al-20Si- 1Cu-1Mg (wt%) base alloy with 0.16 wt% P addition. Kaneko et al. [11] determined average size and num- ber per unit area of section ¯ N A of primary silicon in Al-19 wt%Si-0.02 wt%P alloy over the range of ˙ T be- tween 0.02 and 2 K/s, and obtained a linear relationship between logarithm of derived number of particles per unit volume ¯ N V and log ˙ T , with a slope of ∼1.4. Moir and Jones [12] combined their own measurements of ¯ N A versus solidification growth velocity V for two dif- ferent temperature gradients G (Bridgman solidifica- tion and tungsten inert gas weld traversing) with mea- surements by Pierantoni et al. [13] for laser surface melt traversing to show a linear relationship between log ¯ N A and log (G √ V ) with a slope of unity. Bayraktar et al. [14] combined all these results with additional Bridg- man measurements to show λ ˙ T 1/3 =250 µm (K/s) 1/3 over the range 0.02< ˙ T <10 6 K/s where λ = ¯ N −1/2 A is a measure of the primary silicon interparticle spac- ing. Mandal et al. [4] reported mean particle size ¯ D A of primary silicon versus ˙ T in the range 15 to 31 K/s for Al-17, 22 and 27 wt%Si with 0.1 and 0.2 wt% P additions. Ohmi et al. [15, 16] reported ¯ D A (but see below) of primary silicon and associated nucleation undercooling T m for Al-22 and 32 wt%Si for so- lidification cooling rates in the range 11 to 260 K/s, and showed particle size decreasing linearly with increase in undercooling and increasing with in- creased cooling rate (results show a reasonable fit with T m = A ˙ T n K /s with A = 3.5 K (K/s) −n and n = 0.6). Liang et al. [17] measured ¯ N A and ¯ D A together with formation temperature T f of primary silicon versus V and G in Bridgman in solidification of Al-18.3 wt%Si for comparison with model predictions for steady state heterogeneous nucleation of the primary silicon from the bulk melt. The experimental results show N v ∝ ˙ T 1.2 and that nucleation undercooling increases from ∼35 to ∼52 K over the range 1 < ˙ T < 20 K/s, consistent with the model prediction for a nucleation contact angle θ increasing from 26 to 36 deg over the same range of ˙ T . Most recently Kyffin et al. [18] reported the effect of phosphorus inoculation on ¯ N A of primary silicon for the range 0.8 < ˙ T < 16.5 K/s for comparison with the results of Sulzer [10] for 3 < ˙ T < 70 K/s, Kaneko et al. [11] for 0.02 < ˙ T < 2 K/s and Mandal et al. [4] for 15 < ˙ T < 30 K/s. The present purpose is to investigate the possible generality of the relationships between ¯ N V and ˙ T obtained by Kaneko et al. [11], Ohmi et al. [15, 16] and Liang et al. [17]. The available experimental data are summarised in Table I and plotted in Fig. 1 as log ¯ N V versus log ˙ T . The results for phosphorus-free samples fall into two groups. Results from Bridgman solidi- fication, TIG weld traversing and laser surface melt 0022-2461 C  2005 Springer Science + Business Media, Inc. DOI: 10.1007/s10853-005-3103-4 6363