Journal of Research of the National Bureau of Standards Vol. 57, No.2, August 1956 Research Paper 2694 Thermal Properties of Aluminum Oxide From 0° to 1,200° K George T. Furukawar Thomas B. Douglasr Robert E. McCoskeYr and Defoe C . Ginnings Accurate measurements of the heat capacity of a-aluminum ox ide (corundum) from 13 0 to 1,1700 K nrc described. An adiabatic calorimeter was used from 13 0 to 3800 K and a drop method was used with a Bunsen ice calorimeter from 273 0 to 1,1700 K . The res ults a rc compa red in t he mnge 273 0 to 3800 K , where t he two methods overlap. From t he data, smoothed values of t he h eat capacity, enthalpy, entropy, and Gibbs free energy from 0 0 to 1,200 0 K are derived a nd tabul ated. 1. Introduction ideal for a heat-capacity standard over a wide tem· perature range. It is commereiall)T available in the One of the fundamental functions of the National form of syn thetic sapphire with impurities present Bureau of Standards is Lo develop new sLandards as in such small quantities that the heat capacity of the the need arises. As the science of thermodynamics sample should be the saIUe as that of a pure sample assumes new import in modern Leclluolog.\', the need wiLllin the accuracy of present calorimetric m easure­ for calorimetric standards becomes urgent. At Lhe ments. The sapphire is a cr.vstalline solid without meeting on April 2] , 1948, the Fourth Conference on knowll transiLion s or chall ges of state up to its Low Temperature Calorimetry 1 consid ered tlti s melting point (ncar 2,000° C (4)) . It is nonvolatile, problem of calorimetric standards and recommended non hygroscopic, and chemicall~ ' stable in air, and three materials to serve as heat - capacit ~T standards does not absorb carbon dioxide. Except at the over a wide Lemperature range. These maLerials lowest temperaLures, it has a high heat capacity per were benzoic acid (10° to :350° K ), 11 -JlepLane (10 ° unit volume. I t is exLremely hard and should be to 300 ° K ), and a-aluminum oxide (10° to 1,800° K ). free from mechanical effecLs such as strains clue Lo The Bureau was ash·ed to prepare very pure samples cold-working, which cause small but significant of these materi als which would be available to those changes in Lhe Lhermal properties of metals. In laboraLories illLeres ted in very precise measuremell Ls summar.v, it appears that the s.v nLhetic sapphire of heat capacity. B.\· having samples of an.\' one should be an excellent standard for heat-capacity subsLance taken from Olle so urce of vcr.\' high purity, m easureIUonts over most of Lhe temperaLure range it was hoped to Jl ave a menns of co mparing measure­ up Lo its melLing point. m ents made in different la boratori es u llcler clifIerell t The Bureau has previously made m easurements experimental co nditions. The Bureau has prepared (5) over the range 0° to 900° C on a sapphire sample samples of Lhese three materials t hat a rc not regarded (not Calorimetry Conference sample) in order to as part of the Standard Sample series of tltc Bureau, det,ermine the suitability of the material as a stand­ but will be designated here as Calorimetr.\TCo nference ard. The measurements described in t lte present samples, and has made these available without charge report are on tlte Calorimetry Conference sample and to a limited number of laboratories. .Measurements co nsist of two independent calorimetric i nvestigations have already been made at the Bureau on the using entirely differ ent methods and apparatus for the Calorimetry 'Conference sample of benzoic acid [1]/ 10w- and high-temperature ranges. In the range normal heptane [2], and aluminum oxide. A brief 13 ° to 380° K , an adiabatic calorimeter was used. summary (3) of the results of these measurements In the range 273 ° to 1,170° X, a " drop" calorimeter and details of the measurements on benzoie acid [1) was used, similar to the earlier high-temperature ex­ and normal heptane (2) have been published in other periments [5 , 6) except that an entirely new and im­ reports. It is the purpose of the present report to proved apparatus was used. give the complete results of heat capacity measure­ m ents on the Calorimetry Conference sample ot 2 . Sample aluminum oxide, which up to the present have covered the range from 13 ° to 1,173° K. Aluminum oxide ill the form of corundum The aluminum o..\ide sample investigated was (a -Alz03)3 has a number of properties that make it colorless synthetic sapphire (corundum) and was a portion of the material prepared for the Calorim­ 1 'I'he Conference on Low 'r cmpemturc CalorimetrY was renamed Ithc Ca10· 4 rimctry Conference at the meeting held on Septem ber 5, 1950, in'ord e r~LO incllld e etry. 90nference by F. W. Sch\vab of t he Chemistry other fi elds of calorimetry. DlVISlOll at the Bureau. This material, originally " Figures in hrackets indicate t ho literature re feronces at the end of this papP I.. 3 T he ti-AbOa is an impure alumina which ca n be form ed when tile mol ccn purchased from the Linde Air Products Company aluminum oxide is slowly cooled in the presence of certain impllritie<:;. 'rhe 'Y-AI2 0 a, which can be prepared by heating A l(OH h. is mct<:lstahlc, transforming III the form of splIt boules, was coated with a hard to a-Al ~ 0 3 at about 1,000° C. The a-AhOa. known as corundum, ('ontaining opaq ue form of aluminum oxide which was removed traces of chromium, is red and called ruby, while that co nta ining traces of iron and titanium is blue and c1lIed blue sapphirc. Thc synthetic corundum or by immersing in fused potassium pyrosulfate. Fol- syn Lh etic sapphire used in the preparation of the Calorimetry Conference sample was h ig hly pure and contained no coloration. , D eccascd. 38939'1-56-1 67 lowing this cleaning process, a portion (about one­ thermocouples, one of three junctions and tho other fifth of the boules was examined by C. P. Saylor of two, and tbl'ee individual heaters wore usod in of the Bureau for inclusions, and the total volume the control of the shield temperature. of the inclusions was estimated to be less than 1 The electrical power input was measured by means part per million of the volume of the aluminum of a Welmer potentiometer in conjunction with a oxide crystals. standard cell, volt box, and standard resistor. The The cleaned boules were crushed, and about 85 time interval of heating was measured by means of a percent of the material was collected in particle precision interval timer operated on a standard fre­ sizes between 0.02 and 0.08 in. The impurities quency of 60 cps furnished by the Time Section of the from the crushing and sieving processes were re­ Bureau. The timer was compared periodically with moved by digesting in hot hydrochloric acid. The standard second signals and found to vary not more material was then thoroughly washed and dried at than 0.02 sec per heating period, which was never about 300° C. This product showed no loss in less than 2 min. Temperatures were measured bv weight on subsequent drying at 110° C or heating means of a platinum-resistance thermometer and a for 2 hours at 1,200° C. To obtain the highest high-precision Mueller bridge. The platinum-re­ degree of uniformity in all samples, all the material sistance thermometer was calibrated above 90 0 K in was thoroughly mixed in a large bottle and pack­ accordance with the 1948 International T emperature aged in 70-g units of about 30 ml volume. Later Scale [9], and between 10° and 90 0 K with a provi­ some of these 70-g units were divided into smaller sional scale [10], which is maintained by a set of units. platinum-resistance thermometers which had been Spectrographic analyses made by B . F. Scribner, compared with a helium-gas thermometer. The of the Bureau, of a sample from one of the packaged provisional scale as used in the calibration of the 70-g units indicated the purity to be between 99.98 thermometer when the measurements reported in and 99.99 percent by weight. The only impurities this paper were made was based upon the value present in quantities greater than trace amounts 273 .16°K for the ice point and 90.19°K for the were sili con, 0.005 percent; iron, 0.005 percent; and temperature of the oxygen point. Above 90 0 K, the chromium, 0.002 percent. It seems likely that the temperatures in degrees Kelvin were obtained by impurities present would not afl'ect the heat capac­ adding 273.16 deg to the temperatures in degree's ity of the sample by more than 0.02 percent in the Celsius (International Temperature Scale of 1948 temperature range covered by the measurements [9]).5 All electric instruments and accessory appa­ described in this paper. ratus were calibrated at the Bureau. 3 . Low-Temperature Calorimetry 3.2. Heat-Capacity Measurements 3.1. Method and Apparatus The heat-capacity measurements on aluminum The heat-capacity measurements in the low­ oxide were made from about 13 ° to 380 0 K in sample temperature range, from about 13 ° to 380° K, were container A and calorimeter G.