Industrially used alloys practically always consist of a base metal which has been alloyed with one or more elements. In liquid state, the alloying additives are in most cases soluble in the base metal. The solidification normally takes place within a solidification range which is typical of the alloying composition. Upon solidifica- tion, different solid phases are separated from the mol- ten metal, latent heat being released. By following the temperature and the duration in time of the solidifica- tion, it is possible to obtain indirectly a reference for the composition of the alloy and its manner of solidifi- cation.

The method has been standardised by the use of test cups or crucibles made of refractory material with an integrated expendable thermocouple. The method, which is called thermal analysis, is widely used for iron and alu- minium alloys. The cavity in the test cups industrially used is square or circular in cross-section and the test cups are provided with a centrally placed thermocouple.

Typical dimensions are 37 x 37 mm and a height of 40 mm.

The cups are made of shell-moulding sand and have a wall thickness of about 5 mm. The cavity is completely open upwards where the metal is poured when testing. From a test, a great deal of information can be obtained about the molten metal and its behaviour, for instance, when casting. The crucial point is to provide a high degree of repeatability of the testing. In prior art, the repeat- ability can vary, among other things, depending on the filling degree of the test cup and variations in heat

emission by radiation and convection from the upper sur- face.

One problem is that the centrally placed thermo- couple only registers temperature conditions in the centre of the cup where the molten metal is liquid for a fairly long time.

It is desirable to be able to simultaneously follow the temperature at the surface of the test cup and carry out a more detailed analysis of the test piece by com- paring the process in the centre and surface of the test cup.

Test cups with one thermocouple placed in the centre and another at the surface are already known. Thus Swiss patent specification 626 450 discloses a crucible receiv- ing a molten metal, a thermocouple being arranged in the molten metal and another in or at the wall of the cru- cible. In other known examples, use has been made of cylindrical or cubic test cups, the thermocouple at the surface being placed at a distance of 1-3 mm from the wall. One problem is that a small error when placing the peripheral thermocouple makes the measuring result un- certain.

The object of the present invention is to solve these problems and provide a device and a process for thermal analysis of molten metals providing high repeat- ability and high resolution. Hence the device and the process have the features stated in claims 1 and 4, respectively.

In the device according to the invention, the spherical cavity has a cylindrical duct which is con- nected at the top and a cylindrical part which is con- nected at the bottom.

Since the cavity is spherical, the solidification will take place in a concentric manner, which makes the impulses from the solidification to the thermocouple placed in the centre much clearer than in known cylindri- cal or cubic constructions. By arranging a cylindrical

filling duct, in which the molten metal has a shorter time of solidification than in the spherical cavity, the effect of fluctuations in the heat emission from the upper surface due to emission changes upon radiation will be eliminated. Furthermore, variations due to different degrees of filling will be eliminated since the duct is assumed to be constantly filled after the casting of a test piece.

By placing the lower thermocouple in the transition between the spherical cavity and the lower cylindrical part, the position can vary somewhat without disturbing the repeatability. The purpose of the lower cylindrical part is that the molten metal located in the same should solidify relatively rapidly and before the molten metal in the spherical cavity. Hence thermal conduction occurs in solid phase through the lower part and during the major part of the solidification in the spherical cavity.

Therefore, the lower thermocouple can indirectly register the thermal conductivity of the alloy in semi solid to solid phase.

This is useful in particular when testing cast-iron alloys where carbon is precipitated in the form of graph- ite with high thermal conductivity during the solidifica- tion. The graphite can be precipitated in different forms which affect the castability and physical properties of the alloy. If the graphite is precipitated in the form of spheroids, the alloy is called nodular iron. If the graphite is precipitated in the form of agglomerates with thin graphite flakes, the alloy is called grey cast iron or flake graphite cast iron. The thermal conductivity of flake graphite cast iron can be up to 25% higher than if the graphite has been precipitated in the form of sphe- roids. An intermediate form is the so-called dense graph- ite iron, which is distinguished by the graphite being precipitated in the form of rounded"plump"bar-like forms. Thus the thermal conductivity can be used to ana- lyse the graphite form.

According to the present invention, an indirect measure of the thermal conductivity can be obtained by measuring the difference in temperature between the ther- mocouple placed in the centre and the thermocouple placed peripherally in the spherical cavity in the transition between the spherical cavity and the cylindrical part.

According to a preferred embodiment of the invention, the difference in temperature is registered when the solidus temperature of the alloy has been reached at the thermo- couple placed in the centre.

The invention will now be described in more detail with reference to the accompanying drawing, in which Fig. 1 is a front view of a preferred embodiment of the device according to the invention, Fig. 2 is a section along the line II-II in Fig. 1 and Fig. 3 illustrates the temperature curves of the peripheral thermocouple and the central thermocouple.

With reference again to Figs 1 and 2, a device is shown comprising a mould 1 consisting of two parts of re- fractory material. The mould parts are suitably held to- gether during the casting of the test piece by means of a holder (not shown). Furthermore, the device includes a spherical cavity 2 in which a thermocouple 3 is centrally placed. Thus this thermocouple extends over the central portion of the cavity. A pouring cup 4 for pouring molten metal is arranged and this pouring cup passes into a cy- lindrical duct 5, which in turn communicates with the spherical cavity 2. A cylindrical part 7 is connected to the lower portion of the cavity and a second thermo- couple 6 is arranged in the transition between the ca- vity 2 and the cylindrical part 7. According to this illustrated preferred embodiment, the cold junctions 8 of the thermocouples are positioned along the longitudinal axis of the test cup, the one thermocouple, as mentioned above, extending over the central portion of the spheri- cal cavity 2 and the other thermocouple 6 intersecting the longitudinal axis of the test cup at a short distance

above the interface between the cavity 2 and the cylin- drical part 7. This distance is generally in the range of 0-2 mm.

The following dimensions can be mentioned as non- limiting examples. The outer dimensions of the mould are a height of 110 mm and a width of 60 mm, the thickness of each mould part being 40 mm. The pouring cup 4 has an upper diameter of 40 mm and a height of 20 mm. The con- necting duct 5 has a diameter of 20 mm and a height of 25 mm. The spherical cavity has a diameter of 40 mm and the lower cylindrical part has a diameter of 16 mm and a height of 15 mm. The thermocouples 3 and 6 are in prior- art manner made of"Chromel-Alumel"and enclosed in a tube of high purity quartz. The thermocouples are con- nected to an A/D converter in a known manner via a com- pensating circuit. When analysing a molten metal, the device is filled with molten metal by means of a casting ladle. The casting temperature for cast-iron alloys should be in the range of 1240-1350°. The temperature is preferably registered once per second. After about 250 s the molten metal is solid. Preferably, time/temperature data are analysed by means of a computer program.

Typical cooling curves are illustrated in Fig. 3.

Curve 9 shows the temperature changes of the peripheral thermocouple and curve 10 those of the central thermo- couple. The point of time of the solidus temperature of the thermocouple 3 which is placed in the centre is for this purpose defined as the minimum point of the first derivative of the time/temperature curve. The reason why this point of time has been selected is that it is not until after solidification that differences in thermal conductivity become clear. To be able to predict casting properties etc, it is important to obtain a measure of thermal conductivity at the highest possible temperature.

At this point of time, the difference 11 in temperature between the thermocouples is calculated. The difference in temperature for molten grey cast iron normally is

about 90°C and for nodular iron alloys which have infe- rior thermal conductivity about 120°C. The difference in temperature is sufficient not only for classifying the type of cast iron but also for providing information about, for instance, the nodularity of nodular iron alloys and the fraction of vermicular graphite in dense graphite alloys.