Depression of melting point for protective aluminum oxide films

Depression of melting point for protective aluminum oxide films

Chemical Physics Letters 618 (2015) 63–65 Contents lists available at ScienceDirect Chemical Physics Letters journal homepage: www.elsevier.com/loca...

389KB Sizes 0 Downloads 93 Views

Chemical Physics Letters 618 (2015) 63–65

Contents lists available at ScienceDirect

Chemical Physics Letters journal homepage: www.elsevier.com/locate/cplett

Depression of melting point for protective aluminum oxide films E.L. Dreizin a,∗ , D.J. Allen b , N.G. Glumac b a b

New Jersey Institute of Technology, Newark, NJ, United States University of Illinois, Urbana-Champaign, Urbana, IL, United States

a r t i c l e

i n f o

Article history: Received 25 September 2014 In final form 28 October 2014 Available online 5 November 2014

a b s t r a c t The protective aluminum oxide film naturally formed on a surface of aluminum has a thickness in the range of 3–5 nm. Its melting causes loss of its continuity, which may significantly affect the ignition and combustion processes and their relative time scales. Melting of the alumina film also plays an important role when aluminum powders are used to prepare composites and/or being sintered. This letter quantifies depression of the melting point of an alumina film based on its nano-meter thickness. A theoretical estimate is supported by experiments relying on a detected change in the optical properties of naturally oxidized aluminum particles heated in an inert environment. © 2014 Elsevier B.V. All rights reserved.

1. Introduction Multiple models of aluminum ignition discussed in the literature rely on the melting of aluminum oxide as a demarcation event separating ignition processes from the reaction regime associated with combustion [1]. The focus has been on describing the transport rate of reagents including ions of Al and O, across the aluminum oxide layer, which is being analyzed up to its melting point, 2327 K, according to NIST Chemistry Webbook [2]. However, a small thickness of the aluminum oxide layer may be affecting its melting point. Indeed, the melting point depression has been extensively discussed in the literature for nano-materials, see for example a review in ref. [3] and, in particular, for nano-aluminum particles [4,5]. The effects become significant when the characteristic material dimension becomes close to or less than approximately 10 nm. Most nano-aluminum powders are substantially greater than this dimension, so that the effect of the particle size on the Al melting point is small. However, the thickness of the ‘natural’ aluminum oxide film is commonly in the range of 3–5 nm, e.g., [6,7]; thus, the depression of its melting point can be substantial. This letter evaluates the anticipated effect theoretically and offers an experimental validation of the depressed melting point for natural aluminum oxide layers on surface of fine aluminum powders. The effect must be accounted for in detailed models of aluminum ignition assuming a step-wise change in the protective properties of aluminum oxide upon its melting. It may also need to be considered in other

∗ Corresponding author. E-mail address: [email protected] (E.L. Dreizin). http://dx.doi.org/10.1016/j.cplett.2014.10.063 0009-2614/© 2014 Elsevier B.V. All rights reserved.

high-temperature processes, such as sintering of structural components, involving aluminum powders [8]. 2. Estimated depression of the melting point The effect can be roughly evaluated considering available theoretical models describing the depression of melting point for nano-materials. A simplified expression for the melting point of a nano-sized material is given in ref. [9]:



˛−1 D/D0 − 1

Tm = Tb exp −



(1)

where D is the film thickness, ˛ is a ratio of mean square displacements of surface atoms and interior atoms, and D0 is a critical diameter, for which all atoms are located on the surface. For lowdimensional crystals, D0 depends on the dimension d. For thin films, d = 2. The relationship for D0 and atomic diameter h is (for films) D0 = 2(3 − d)h = 2h

(2)

For surface melting, one can find ˛=

 2C

pm

3R



+1

(3)

where Cpm is heat capacity difference between liquid and solid phases at Tm . An estimate for D0 can be obtained as the average Al O bond length in such transition alumina polymorphs as  and : h = 1.9 A˚ [10]. To calculate Cpm , one can use data from the NIST Chemistry webbook [2]; bulk alumina melts at 2327 K.

64

E.L. Dreizin et al. / Chemical Physics Letters 618 (2015) 63–65

Temperature, K

2400

2000

1600

1200

800

1

2

3

4

5

6 7 8 9 10

20

Thickness, nm Figure 1. Estimated melting point as a function of thickness of aluminum oxide. Figure 2. Thermal emission spectra of 1–5 micron aluminum particles heated in argon at 1521 K and 7.5 bar. The power law fit shows a wavelength dependence of −2.94 .

For all alumina phases, the specific heat is given by Shomate equation: Cp = A + B ·

T +C · 1000

 T 2 1000

+D·

 T 3 1000

+E·

 1000 2

near 765 nm that was present in all tests. The relative emissivity is measured by comparing the experimental emission curve to the predicted blackbody emission at the temperature determined

T (4)

where Coefficient

A

B

C

D

E

Gamma Liquid

108.6830 192.464

37.22630 9.819856 × 10−8

−14.20650 −2.858928 × 10−8

−14.20650 2.929147 × 10−9

−3.209881 5.599405 × 10−8

Plugging in the numbers, gives ˛ = 4.773.

3. Experimental detection of melting point of natural aluminum oxide film The detection of the alumina melting point is based on an assumed change in the surface emissivity of aluminum particles covered with the naturally grown oxide film when the film is melting. To detect this change in emissivity, the following experiment was performed. Micron sized aluminum particles in the 3–4.5 ␮m diameter range obtained from Alfa-Aesar, A Johnson Matthey Company, were injected, aerosolized, and rapidly heated to a well characterized temperature and pressure behind the reflected shock in a heterogeneous shock tube described in full in previous publications [11,12]. The reflected shock conditions are calculated by measuring the shock velocity and the initial temperature and pressure of the argon used as the ambient test gas. In the experimentally inert environment all particles thermally equilibrate to the known gas temperature rapidly (<5 ␮s, as readily estimated based on their dimensions and thermal diffusivity) relative to the overall test time of 1.6 ms and remain at the gas temperature for the remainder of the test duration. The incandescence of the heated particles was collected using a Triax 190 spectrometer coupled to a Hamamatsu back-thinned CCD with 128 × 1044 pixels each 25 ␮m square. The resolution of the system was approximately 1.4 nm. Spectra of the thermal emission were obtained from 650–900 nm as shown in Figure 2 and were integrated over the entire duration of the test. The wavelengths from 720 to 770 nm were not used to fit an emissivity trend due to strong atomic interferences from the potassium doublet

by the shock velocity. The experimental emission curve is first calibrated to account for the instrument’s spectral response prior to comparison with the black body curve. Figure 2 shows an example power law fit (ε∼−2.94 ) for the relative wavelength dependence of the particles at 1521 K. It was previously shown that the wavelength dependence of the emissivity properties of nano-alumina particles is affected by the phase of the material [13]. A similar effect is expected for micron sized aluminum particles with nanometer thin oxide layers. In the present experiments, the shift in the emissivity wavelength dependence at different temperatures is used to infer melting of the oxide layer. Figure 3 is a plot of the power law fit at different ambient temperatures. A shift in the power law coefficient between 2000 K and 2200 K suggests that melting is occurring in this temperature range. This result is consistent with an oxide thickness in the 5–8 nm range. -2.4 -2.5

Power Law Coefficient

Finally, substituting the value of ˛ in Eq. (1), one obtains the result shown in Figure 1. For a practical range of natural alumina thickness, from ca. 3 to 5 nm, the melting point is predicted to be depressed by 400–200 K.

-2.6 -2.7 -2.8 -2.9 -3.0 -3.1 -3.2 1400

1600

1800

2000

2200

2400

Ambient Temperature (K) Figure 3. Measured power law coefficient at various ambient temperatures.

2600

E.L. Dreizin et al. / Chemical Physics Letters 618 (2015) 63–65

4. Discussion The experimental data support the initial estimate of the depressed melting point of a thin layer of aluminum oxide. As noted above, this depression is important to be accounted for when modeling ignition of aluminum particles in practical systems [1,14]. Upon melting, aluminum oxide film is expected to reduce its surface area forming a spherical inclusion and exposing aluminum surface directly to an oxidizing environment. For nano-sized aluminum particles, substantial agglomeration between particles was reported [15]. In the necking areas between agglomerated nanoparticles, the thickness of aluminum oxide may be even smaller than that on surface of the particles. It is possible that such reduced thickness, and respectively reduced melting temperature of aluminum oxide are responsible for substantial morphological changes observed for such nanoparticles heated to 1300 K [16]. From Figure 1, the above threshold temperature corresponds to melting of an oxide layer with thickness between 1 and 2 nm. For micron-sized aluminum powders used in most practical combustion system, the removal of protective oxide film leads to the onset of full-fledged vapor phase combustion. For burning aerosolized powders, this can lead to beginning of a qualitatively different flame propagation [17]. Predicting an ignition delay leading to the vapor phase combustion is important for combustion models involving aluminum as a fuel additive; thus an accurate assessment of the alumina melting point is necessary. Depression of the melting point for nano-sized aluminum oxide film should be also accounted for when composite materials involving aluminum powder are designed and manufactured, as for example is done for the aluminum-carbon nanotubes composite [18], consolidated by spark plasma sintering. Similarly, destruction of an aluminum oxide layer was of critical importance for sintering aluminum powders using combination of pulsed current and high pressure [19]. The melting of natural alumina may also be important to consider for applications involving brazing of aluminum [20].

substantial depression of its melting point is predicted. An estimate shows that the melting point can be reduced by 200–400 K for practically common natural alumina thicknesses. Experiments detected a temperature range when a shift occurred in the emissivity wavelength dependence of naturally oxidized aluminum particles heated in an inert environment. This shift was assumed to be associated with melting of natural alumina. The shift occurred at temperatures varying from 2000 to 2200 K, implying that the alumina thickness was between 5 and 8 nm. Acknowledgement This effort has been supported by Defense Threat Reduction Agency. References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18]

5. Concluding remarks Natural alumina film formed on surface of aluminum metal has thickness of about 3–5 nm. Based on its small thickness, a

65

[19] [20]

A. Ingenito, C. Bruno, J. Propuls. Power 20 (2004) 1056. M.W. Chase, J. Phys. Chem. Ref. Data, Monograph 9 (1998) 1. Q. Jiang, C.C. Yang, Curr. Nanosci. 4 (2008) 179. M.A. Trunov, S.M. Umbrajkar, M. Schoenitz, J.T. Mang, E.L. Dreizin, J. Phys. Chem. B 110 (2006) 13094. J. Sun, M.L. Pantoya, S.L. Simon, Thermochim. Acta 444 (2006) 117. V.Y. Gertsman, Q.S.M. Kwok, Microsc. Microanal. 11 (2005) 410. F. Reichel, L.P.H. Jeurgens, G. Richter, E.J. Mittemeijer, J. Appl. Phys. 103 (2008), 093515/093511-093515/093510. S.H. Huo, M. Qian, G.B. Schaffer, E. Crossin, Aluminium Powder Metallurgy, Woodhead Publishing Ltd., 2011, pp. 655. M. Zhao, Q. Jiang, Solid State Commun. 130 (2004) 37. R. Lizárraga, E. Holmström, S.C. Parker, C. Arrouvel, Phys. Rev. B – Condens. Matter Mater. Phys. 83 (2011). P. Lynch, H. Krier, N. Glumac, J. Thermophys. Heat Transf. 24 (2010) 301. N. Glumac, H. Krier, T.I.M. Bazyn, R. Eyer, Combust. Sci. Technol. 177 (2005) 485. J. Kalman, D. Allen, N. Glumac, H. Krier, J. Thermophys. Heat Transf. (2014), http://dx.doi.org/10.2514/1.T4260 (in press). M. Mittal, Chem. Eng. World 48 (2013), 38, 40, 42, 44, 46, 47. P. Chakraborty, M.R. Zachariah, Combust. Flame 161 (2014) 1408. G.C. Egan, K.T. Sullivan, T. Lagrange, B.W. Reed, M.R. Zachariah, J. Appl. Phys. 115 (2014). M. Soo, P. Julien, S. Goroshin, J.M. Bergthorson, D.L. Frost, Proc. Combust. Inst. 34 (2013) 2213. H. Kwon, M. Takamichi, A. Kawasaki, M. Leparoux, Mater. Chem. Phys. 138 (2013) 787. T. Nagae, M. Yokota, M. Nose, S. Tomida, T. Kamiya, S. Saji, Nippon Kinzoku Gakkaishi 65 (2001) 726. J. Zaehr, S. Oswald, M. Tuerpe, H.J. Ullrich, U. Fuessel, Aluminium (Hannover, Ger.) (2011) 57.