600 The Growth of Metallic Eutectics .
By F. E. E. Lamplough , M.A. , and J. T. Scott , B.A. ( Communicated by C. T. Heycock , F.B.S. Beceived June 6 , \#151 ; Bead June 25 , 1914 .
) [ Plate 9 .
] Besearch on this subject originated in an attempt to discover in what way the structure of certain alloys is affected by undercooling at the first moment of solidification of the eutectic .
The formation around the primary crystals of an alloy of what may be called a halo or envelope of the second constituent , separating the primary crystals from the banded eutectic , has been considered by some workers to be the result of such superfusion .
It has , however , been suggested by Desch that this effect is due to segregation of the eutectic in the immediate neighbourhood of the primary during solidification .
At the same time it was suggested by him that , by preventing undercooling by means of inoculation , the influence of this phenomenon on the formation of such envelopes might be investigated .
The research at first followed two lines of investigation .
Experiments were carried out to determine whether in different alloys the formation of such envelopes is always accompanied by appreciable undercooling and conversely .
Further , examination was made of the effect of the prevention of superfusion by vibration or shaking and by inoculation with the secondary constituent .
In the investigation it was necessary to measure the temperature of the alloys by a method such that a rapid change of temperature was very quickly indicated by the thermometer , and further , that the mass of metal should be as nearly as possible of uniform temperature .
These conditions could only be realised by making the mass of the metal very small , the heat capacity of the thermometer inconsiderable and the conduction of heat from metal to thermometer as rapid as possible .
The alloys were investigated when possible in small test-tubes , 3 " x f " , which contained some 5 grm. of the metal .
For temperatures below 600 ' C. , copper-constantan thermo-couples were used , the two wires being fused together at the junction or brazed with silver solder .
In such cases the thermo-couple was lightly smeared with a thin paste of " Purimachos " cement and water , which after drying contaminated the surface so that the wires did not alloy with the molten metal .
The couple was very sensitive to rapid The Growth of Metallic Eutectics .
changes of temperature of the metal , for the thin covering of the junction offered extremely little resistance to conduction of heat .
The moving coil of the galvanometer was provided with bifilar suspension to give a very short period of vibration , and experiments with pure metals , such as tip , showed that after considerable superfusion the temperature indicated by the instrument rose almost instantaneously at the first moment of solidification to a maximum value which remained constant during freezing .
The temperature-time curves of the cooling alloys were recorded on a ground glass screen which moved vertically with uniform speed , whilst temperature was measured by the horizontal movement across the screen of a spot of light reflected from the galvanometer .
The tubes containing the metal were allowed to cool in an enclosure heated to a constant temperature , so that the rate of cooling and time of eutectic arrest could be readily varied as desired .
This enclosure consisted of a solid copper cylinder which was bored with holes of suitable size , one to hold the tube , wrapped with asbestos paper , the other to accommodate a mercury \lt ; thermometer or thermo-couple .
The cylinder was wrapped with asbestos cloth and nichrome electrical heating coil , and was embedded in magnesia , .
the temperature being controlled to a small fraction of a degree by an arrangement devised by one of us.* Investigations at temperatures above 400 ' C. were carried out in a small resistance furnace similar to that described , 4 Proc. Camb .
Phil. Soc./ 1911 , p. 175 , the temperature measurements above 700 ' C. being made with Pt , Pt-Rh couples protected by a silica tube , drawn out at the closed end so as just to accommodate the couple wires .
This alteration in the protection of the thermo-couple caused a considerable loss in the accuracy with which a rapid temperature change could be followed , but the arrangement was amply sufficient as a guide in " quenching " experiments .
The research originated in observation of the considerable under-cooling at the first moment of solidification of eutectic which occurs when lead is the primary constituent in lead-tin alloys ; attention was then turned to alloys of cadmium and tin in which primary crystals of cadmium are surrounded by very well marked envelopes .
Investigations were later extended to many other pairs of metals .
Attempts to connect the formation of these envelopes with the undercooling of the alloys met with no success , for it was shown that such undercooling was almost always produced when care was taken to prevent surface oxidation , and that whilst in some cases considerable superfusion existed without the formation of such structures , in others marked halos were * To be described shortly to the Chemical Society .
I 602 Messrs. F. E. E. Lamplough and J. T. Scott .
accompanied by slight superfusion only .
Moreover , when superfusion at the moment of eutectic formation was prevented either by inoculation or by shaking , the formation of the envelope was in no way affected .
If such halos were produced at the first stage of eutectic growth , then the eutectic must have begun to grow from the primary crystals as nuclei .
It was therefore determined to ascertain , by quenching at an early stage of eutectic formation , whether or not this was the case .
For this purpose a cooling curve of a given alloy was taken to determine on the recorder the exact temperature of the eutectic arrest , and its total duration ; on repeating the experiment after the solidification of any desired proportion of the eutectic , the tube was withdrawn and plunged in cold water or melted paraffin wax .
This method of investigation at once showed in the case of the cadmium-tin system that the eutectic growth did not originate at the primary cadmium crystals , and that the halo was only present round those primary crystals which had become enveloped by the growing eutectic .
Here the formation of the halo was certainly not connected with the sudden rise m temperature from undercooling at the first moment of formation of eutectic .
Other cases in which the halo had been observed were investigated in this way , and it was very clearly shown that also in the cases of Pb-Sn , Ag-Cu , and Zn-Al , the halo was produced when the eutectic growth approached very close to the primary crystal .
The further solidification of the eutectic at this point consisted in the deposition of a layer upon the surface of the primary crystals of that constituent present as primary , the other constituent of the eutectic having solidified as an envelope surrounding the primary .
This method of investigation was at once recognised as giving valuable information on the character of the eutectic during growth , and research was prosecuted on these lines .
The formation of the eutectic during its growth was ascertained in the case of numerous pairs of components .
Some of these results were brought before the British Association , 1913 , and a number of cases have since been investigated in which one constituent of the eutectic is an intermetallic compound .
From these investigations it has been possible to classify eutectics according to their character .
The greater number of eutectics fall into two classes : 1 .
Those of spherical radiating growth .
2 .
Those having well defined crystal contours , showing in section a herringbone structure .
The eutectics of Sn-Pb , Sn-Cd , Zn-Cd , Ag-Cu , Zn-Al , Cu-Cu3P , fall into Class I. In all these cases both the constituents appear as primary crystals of rounded form .
In the Zn-Cd system eutectic growth originates at primary The Growth of Metallic Eutectics .
crystals of Zn , in the Sn-Pb system possibly at crystals of Sn ; in other cases , neither constituent present as primary serves as nucleus for the eutectic growth , i.e. the eutectic growth may be said to be " independent " of both constituents .
In each case of spherical eutectic growth an envelope , or halo , was found to surround one of the constituents present as primary .
A microphotograph of a Ag-Cu alloy containing copper crystallites as primary is shown in fig. 1 .
The alloy was quenched during the eutectic arrest and a spherical eutectic grain ( the contour of which is necessarily slightly modified by the primary growth ) is clearly outlined .
The primary copper crystals are rounded , and , where surrounded by eutectic solid at the moment of quenching , are enveloped with a continuous layer of the silver constituent .
Among the second class of eutectics are those of Bi-Pb , Bi-Sn , Ag3Sb-Sb , Au-AuSb2 , Mg2Pb-Mg , Mg2Sn-Mg , ZnSb-Sb , NaCd6-Cd , NaSn2-Sn .
This class of eutectic was observed only in those systems which contained a constituent which solidified with definite crystal contour when present as primary .
Though in most systems the other constituent was of rounded form , in no case was there segregation with production of an envelope surrounding a primary .
In fact , such envelopes or halos have only been observed in conjunction with eutectics of Class I. In some cases the eutectic was " independent " of both primaries , in all other cases only those primaries of crystal shape were found to serve as nuclei or points of origin of the eutectic growth .
In no system investigated during the research was the eutectic found to originate at the primary on both sides of the eutectic .
Fig. 2 illustrates crystal grains of the eutectic Au-AuSb2 , quenched during growth .
In a few systems , Bi-Cd , Cu2Sb-Sb , SnAu-Sn , the eutectic is of a third type , neither spherical nor crystalline in contour , but of irregular form , somewhat resembling in shape a tongue of flame , projecting into the still molten mass .
In the cases investigated these were associated with one constituent crystallising as laminae , and the eutectic growth very definitely originated at one of the primaries .
Fig. 3 illustrates an example of this type , in which the grains of Sb-Cu2Sb eutectic are shown to originate at the plates of Cu2Sb , whilst fig. 4 shows the growth of the same eutectic to be independent of the primary crystals of Sb .
The authors have endeavoured to repeat Vogel 's experiment on the rapid cooling of a Zn-Cd alloy , but have failed to observe any such structure as is described by him.* In most of the systems investigated rapid cooling had * 'Zeits .
Anorg .
Chem. , ' vol. 76 , pp. 431-432 ( 1912 ) .
LAM PLOUGH \amp ; SCOTT .
Roy .
Soc. Proc. , AVo/ , PL p. The Growth of Metallic Eutectics .
merely the effect of diminishing the scale of the eutectic structure .
In a few cases , however , the rate of cooling exercised an influence on the nature of the eutectic .
This was particularly the case with bismuth-tin alloys .
With slow cooling\#151 ; a eutectic arrest of half an hour or more on masses of 5 grm.\#151 ; the eutectic solidified as separate particles of the two metals , apparently without any relation to each other .
More rapid cooling in air\#151 ; eutectic arrest about one minute\#151 ; gave a very well defined crystalline contour to the eutectic grains .
On quenching the partly solid eutectic in either case , throughout the very fine ground-mass were found to be distributed small crystals of bismuth and much-branched crystallites of tin , the latter resembling the " quenched primary " crystals often observed in quenching from the liquid state an alloy containing excess of one constituent .
In this case the regions adjacent to the eutectic grains were found to be devoid of these crystals .
Somewhat similar " quenched primary " crystals have been observed in the case of alloys of cadmium and sodium , zinc and antimony .
The authors are about to investigate the viscosity of alloys , in order to determine whether this property is connected with the form of primary and eutectic growth .
A more extended account of the above work , with full microphotographs , will be presented at a later date .

