- Email: [email protected]

THERMAL HISTORY OF AMORPHOUS SOLIDS AT LOW TEMPERATURES

R.L. FAGALY S.H.E. Corporation, 4174 Sorrento Valley Blvd., San Diego, CA 92121, USA

and J.C. LASJAUNIAS Centre de Recherches sur les Trks Basses Ternp~ratures, B.P. 166X, 38042 Grenoble Cedex, France

Received 13 November 1980

As the number and type of experiments on amorphous solids grows, it has become increasingly important to know the thermal history of the sample. Enough data have been collected on one system, vitreous silica, to draw the conclusion that thermal history is important in determining at least two thermal properties of SiO2 at low temperatures [1-3]. We will discuss the results of varying the fictive temperature on the low temperature heat capacity of SiO2 above 1 K. Between 0.1 and 4.2 K, we can fit the heat capacity (Cv) of vitreous silica as [3]: Cv = C1T + C3T 3 + C s T 5 ,

(1)

although below 0.1 K a simple linear relationship fails [2]. We also observe that C3 is greater than Co, the specific heat predicted by Debye theory. However, for one type of vitreous silica, "as-received" Suprasil W (an artifically prepared glass [4]), there is a bump in the heat capacity which is not describable in terms of eq. (1). This is in contrast to all other types of vitreous silica. However, it has been shown [5] that heat treatment at 1300°C is sufficient to remove this bump and we can then describe the heat capacity in terms ofeq. (1). We consider the results of heat capacity measurements of low OH content vitreous silicas with differing heat treatments [ 1,3,5,6]. Fig. 1 shows the cubic term, C3, versus Tf, the fictive temperature, a value that is characteristic of the (heat) treatment temperature and indicates the thermal history of the sample [7]. As can be seen, the effect of decreasing Tf is'to increase C3. A weighted least squares fit [8] to the data yields A C 3 / A T f = (-0.32 -+ 0.11) erg/g K4/100°C. We also see a similar trend with higher order terms. By restricting the upper limit of temperature to 4.2 K, we can fit the data with the presence of a T s term (in addition to the contribution due to two-level systems C1 and the Cubic term 6'3). Obviously, higher order terms must be used if a higher temperature span is considered, but the trend of Cs 0 022-3093/81/0000-0000/$02.50 © North-Holland Publishing Company

R.L. Fagaly, J.C. Las]aunias / Thermal history of amorphous solids

308

-- 0 . 9

x~ --

o~

0.8

C5~

0.7

C5 4 C~ o Y ~n v

12

I

I

I

I

1

1000

II OO

1200

1300

1400

Tf (°C) Fig. 1, t~ ref. [1], • ref. [3], A refs. [5,6], The dashed lines are the best fits to the data.

should yield qualitative information on the dispersive nature of SiQ. The results also provide a test for theoretical models of amorphous solids, and allow us to reject any model that predicts an increase in C3 (and higher order terms) as Tf is increased. From the above, we draw the following conclusions. First, Suprasil W shows a significant reduction in Cv due to an initial thermal treatment [5]. Secondly, a change in Tf causes a change in Cv. It has also been shown [ 1] that samples of vitreous silica show changes in expansivity when given different values of Tf, but show no difference between OH impurity levels when Tf is the same. The implication is that thermal history is an important factor in determining thermal properties, and cannot be ignored. It should therefore be included when reporting results on thermal properties of amorphous solids at low temperatures. References [1] G.K. White and J.A. Birch, Phys. Chem. Glasses 6 (1965) 85. [2] J.C. Lasjaunias, A. Ravex, M. Vadorpe and S. Hunklinger, Solid St. Commun. 17 (1975) 1045. [3] R.L. Fagaly and R.G. Bohn, J. Non-Crystalline Solids 28 (1978) 67. [4] R. Bmckner, J. Non-Crystalline Solids 5 (1970) 123. [5] J.C. Lasjaunias, G. Penn, A. Ravex and M. Vandorpe, J. Phys. 41 (1980) L131. [6] J.C. Lasjaunias, G. Penn and M. Vandorpe, to be published. [7] G. Hetherington, K.H. Jack and J.C. Kennedy, Phys. Chem. Glasses 5 (1964) 130. [8] A. Picot, Am. J. Phys. 48 (1980) 302.