GB2115825A - An antifinning agent for foundry sands - Google Patents

Abstract

An improved resistance to 'finning' or 'veining' of castings made in silica sand moulds is achieved by incorporating in the sand mixture of the mould and/or cores up to 5% of a polyvalent metal salt of a polycarboxylic acid, especially a dicarboxylic aromatic acid such as a phthalic acid. A preferred material is calcium terephthalate.

Description

SPECIFICATION An antifinning additive for foundry sands The invention is concerned with additives for granular refractory materials based on silica sand, for use in foundries. The surface quality of metal castings produced in sand moulds can be affected substantially by the type of sand binder used, and several casting defects are directly related to the properties of the sand binder and the bonded sands.

One of the defects, known as finning or veining, occurs as a thin film of metal adhereing to the casting surface and growing perpendicularly away from it into the mould or core. Such fins are undesirable and their removal involves additional fettling operations; they can be removed relatively easily provided that they occur on an accessible part of the casting, but frequently they occur in the most inaccessible places, such as waterways in engine block and cylinder head castings, and inside pump bodies, and then they can cause considerable nuisance, often resulting in a scrap casting.

Finning has been observed in the presence of all modern corebinder systems with the exception of sodium silicate bonded sands. It is particularly troublesome when phenolic resin and urethane resin binders are used.

Finning is thought to be the result of the high expansion of the silica sands normally used in core sand mixtures. When a core is surrounded by molten metal the surface layers of the core are heated rapidly and the binder begins to decompose. The initial compressive stress at the surface of the core caused by expansion of the quartz in the sand falls rapidly and is replaced by tensile stress produced by the expansion of the strong underlying layers. When this tensile stress exceeds the surface tensile strength a crack develops, into which molten metal runs and solidifies to form a fin.

The problem can be alleviated in some instances either by using a different grade of resin or by changing the ratio of resin to hardening agent used. The most popular method of reducing finning is to make a small addition to the sand of a material which reduces the total expansion of the bonded core sand. Such materials either have very low thermal expansion or they soften at high temperature, "absorbing" some of the expansion of the quartz. Compounds which have been used in this application include wood flour, starch and natural resins. The most commonly used materials, however, are iron oxides, particularly coarser grades of Fe203, and these are added normally in the range of 1-2+ per cent by weight of sand. In some instances, higher additions up to 5 per cent have been used.It is not fully understood how iron oxides reduce the finning defect.

It is believed that the tendency to finning is related to the hot strength of the core, but merely increasing the hot strength does not necessarily reduce the finning tendency. In the present state of the art, efforts to increase the hot strength of the core usually result in an increase in hot brittleness, making the bonded sand more susceptible to thermal shock and finning. Some success in reducing finning is obtained by applying strengthening coatings to the surface of the core. The coatings usually contain a refractory powder, to give some thermal insulation, and a binder in a liquid carrier. The application of such coatings involves additional processing of the cores and additional expense.

A method for reducing finning which involves an increase in the total strength of the core will not normally be sufficient to eliminate the tendency to cracking or finning, since the stress distribution in the core will remain the same.

It is an aim of the invention to provide an antifinning additive which may be used in silica and mixtures to reduce or even eliminate the finning defect.

According to the invention we propose the use, in a silica sand foundry mould- or core-making composition, of up to 5% by weight of an additive comprising at least one diva lent or trivalent salt of a polycarboxylic acid. It has been found that the surface strength of the cores, containing the additive, as measured by scratch hardness tests, is appreciably higher without a corresponding increase in overall tensile strength of the core. Additionally the hot strength is maintained for a longer period after casting.

The quantity of binder in the composition follows the usual practice and the binder can be hardened in the usual ways. It may provide some benefit to increase the amount of catalyst or hardening agent used in the mixture by a factor of 1+.

The value of using a polyvalent salt of a polyvalent acid is believed to lie in the possibility of bridging compounds being formed, ionically linking adjacent molecules. Preferably the acid is a dicarboxylic acid.

The acid is preferably an aromatic dicarboxylic acid, and in particular one of phthalic, isophthalic or terephthalic acid. Aliphatic dicarboxylic acids, for example, maleic, malonic, adipic, succinic or sebacic acids can be used but are not as satisfactory as the aromatic acids, possibly because the metal salts of aromatic acids have a better thermal stability. The metal is preferably calcium, magnesium, barium, iron or aluminium. The preferred salt, however, is calcium terephthalate. The amount of anti-finning additive used may vary between 0.1 per cent and 5 per cent by weight of the sand and is preferably in the range 0.25-0.75 per cent.

The preferred material, calcium terephthalate, may be prepared by double decomposition of sodium terephthalate (in aqueous solution) with calcium chloride, also in aqueous solution. The calcium terephthalate is precipitated as a very fine white powder which is washed thoroughly, filtered and dried at 1 00C before use in coresands.

The partice size of the calcium terephthalate has been found to have some effect on its performance as an anti-finning additive. The temperature at which the calcium terephthalate is precipitated during its manufacture has been found to be an important factor in controlling the particle size distribution of the product. The most effective materials are those produced by precipitation of the calcium terephthalate at temperatures between 600C and 800C, although materials produced outside this specification can still produce considerable reduction in finning tendency of coresands.

The invention will now be further described with reference to a number of examples and with reference to the accompanying single-figure drawing.

The examples given below illustrate the method of preparation of the calcium terephthalate, the means of use and the types of binder system in which substantial reductions in the extent of finning have been observed. These include hot-box resin process binders, shell process binders, and phenolicurethane cold-setting and gas hardened binders.

EXAMPLE 1 Calcium terephthalate was prepared as follows.

80 g of sodium hydroxide was dissolved in 2 litres of water. The sodium hydroxide solution was heated to 850C and 166 g of terephthalic acid was added slowly. The solution was stirred vigorously until all the terephthalic acid had dissolved. The temperature of the solution was adjusted to 750C and a solution of 111 g of calcium chloride in 300 ml of water was added, resulting in an immediate milky white precipitate of calcium terephthalate. The calcium salt was filtered, washed three times in hot water and finally dried in a laboratory oven at 11 OOC. The resulting fine powder was passed through a 30 mesh sieve to remove all traces of larger agglomerated particles. Yields greater than 90 per cent were readily achieved.

A number of hot-box resin bonded coresand mixtures were prepared using Chelford 60 silica sand, a commercial phenolic hot-box resin binder, appropriate catalyst and various additions of the calcium terephthalate, water and extra catalyst according to Table 1.

TABLE 1 COMPOSITION OF SAND MIXTURES

Calcium Mix Sand Resin Catalyst Water terephthalate Number kg % % % % % 1 2 1.75 0.25 - 0 2 2 1.75 0.25 ~ 0.25 3 2 1.75 0.375 0.125 0.50 4 2 1.75 0.375 0.25 0.75 5 2 1.75 0.5 0.25 1.00 The sand mixtures were prepared by procedures well known in the art and resulted in free flowing sand mixtures, suitable for the production of cores by ramming, squeezing or blowing.

Standard AFS 2.54 cm tensile test pieces were prepared from these sand mixtures and the binder was cured in a heated corebox at 2500C for either 20 s or 30 s. The tensile strengths of the prepared test specimens were determined after 24 hours storage and the surface scratch hardness of each core was also determined at this time. Results are shown in Table 2.

TABLE 2 TENSILE STRENGTHS AND SURFACE SCRATCH HARDNESS RESULTS FOR MIXES 1-5

Calcium Curing Tensile Strength Surface Mix Terephthalate Time . (24 h) Scratch Number % s kN/m2 Hardness 1 0 20 - 1833 47 30 1681 56 2 0.25 20 827 50 30 2239 62 3 0.50 20 551 54.5 30 1378 68.5 4 0.75 20 689 61.5 30 1323 63 5 1.00 20 586 55.5 30 827 62 These results show that the presence of the calcium terephthalate in the cores has had two effects:: (i) The tensile strength of the cores was reduced substantially, more particularly after the shorter curing time.

(ii) The surface scratch hardness was increased significantly with the use of the calcium terephthalate.

EXAMPLE 2 The calcium terephthalate prepared according to Example 1 was used in a series of sand mixtures, based on Chelford 60 sand with a phenolic hot-box resin binder system to assess the effect of increasing amounts of the calcium terephthalate additive on the hot properties of the bonded sands.

Sand mixtures were prepared with compositions described in Table 3 and were used to prepare 11.43 cm x 2.54 cm x 0.64 cm rectangular shaped test pieces for use in the BCIRA Hot Distortion Test.

(Described in AFS Transactions 1975. v.83. pp 73-80). The specimens were cured in a heated corebox at 2500C for 30 s. After 24 hours storage the specimens were tested in the Hot Distortion Test equipment and the experimental curves obtained are shown in the accompanying drawing.

TABLE 3

Sand Hot-box Calcium Mix Weight resin Catalyst terephthalate Number kg % 6 2.0 2.25 0.50 - 7 2.0 2.25 0.50 0.25 8 2.0 2.25 0.50 0.50 9 2.0 2.25 0.50 0.75 10 2.0 2.25 0.50 1.00 11 2.0 2.25 0.50 . 1.50 The curves showed that increasing amounts of the calcium terephthalate up to an addition of 0.75 per cent had only a very small effect on the expansion properties of the sand, but increased its hot strength considerably (i.e. the time taken for substantial distortion of the specimen was increased).

With additions in excess of 1 per cent, the sands showed greater expansion and although the hot strength was improved there was some indication that the sand had become a little less thermoplastic and potentially more brittle.

EXAMPLE 3 The calcium terephthalate prepared according to Example 1 was used in a test to establish its effectiveness in reducing finning with hot-box resin bonded sands in a specially developed test casting.

The composition of the sand is shown in Table 4.

The test casting, poured in a low phosphorus grey iron, weighed 17 kg and consisted of a block 1 8 cm x 18 cm x 7.5 cm in which four holes were produced symmetrically using four AFS 2.54 cm tensile test cores prepared from the test sand. The extent of finning within the core cavities in the casting was assessed visually after the castings had been sectioned. Results, given in Table 4, show that the castings prepared from cores with no calcium terephthalate gave rise to considerably more finning than those with calcium terepthalate present.

TABLE 4

Calcium Mix Sand Resin Catalyst Terephthalate Degree Number kg % % % of Finning 12 2 2 0.5 - Severe finning on all edges.

13 2 2 0.5 0.25 Finning much reduced, very small fins on curved surfaces.

14 2 2 0.5 - Severe finning on all faces.

15 2 1.6 0.5 0.4 Slight finning.

16 2 2 0.5 0.5 Slight finning, good surface finish.

EXAMPLE 4 Samples of calcium terephthalate were prepared according to the method described in Example 1, with the exception that the final precipitation of the calcium terephthalate was carried out over a range of temperatures from 400Cto 970C.

The calcium terephthalate products J 970C were filtered and dried as for Example 1. Three samples, namely those produced at 400C, 670C and 970C were submitted to particle size analysis and surface area measurements, the results for which are shown in Table 5. This table shows a variation in specific surface area and particle size distribution depending on the precipitation temperature.

TABLE 5 SPECIFIC SURFACE AREA AND MEAN PARTICLE DIAMETERS FOR THREE CALCIUM TEREPHTHALATE SAMPLES PRODUCED AT DIFFERENT TEMPERATURES

Sample Specific Surface Mean Particle (precipitation Areal Diameter2 temperature) m2/g Population Volume Mean Mean m m 400C 8.96 10.9 1 23.4 670C 7.45 2.9 15.8 970C 7.37 4.5 17.0 measured by multipoint B.E.T. method.

2 measured by means of HIAC particle size analyser.

The calcium terephthalate samples produced at the various precipitation temperatures were compared using the Finning Test casting procedure as described in Example 4. Hot-box resin bonded sand cores were used for these tests. The extent of finning associated with each material is given in Table 6.

TABLE 6 RESULTS OF FINNING TESTS FOR CALCIUM TEREPHTHALATE SAMPLES PREPARED AT DIFFERENT TEMPERATURES

Precipitation Temperature Extent of Casting Surface C Finning Finish 40 Moderate Poor 46 Slight Fair 53 Trace Fair 60 None Good 67 None j Good 74 None Good 80 None Good 86 Trace 1 Good 92 Trace Fair 97 Slight Fair Normal sand mix with no calcium terephthalate addition Severe 1 Very poor The best materials were those produced at temperatures in the range 60-800C, although it can be seen that some reduction in finning defects was obtained with all of the calcium terephthalate samples.

EXAMPLE 5 A sample of calcium phthalate was prepared as follows: 41.5 g of phthalic acid (0.25 moles) was dissolved in one litre of an aqueous sodium hydroxide solution (0.5 moles/litre) at 950C. When the acid had completely reacted to form the disodium salt, a solution of 28 g of calcium chloride in 100 ml water was added. The calcium phthalate was precipitated immediately as a fine white powder and was filtered, washed with water and finally dried at 1 1 OOC. The product was passed through a 30 mesh sieve to remove agglomerated particles.

A sand mixture containing a hot-box resin binder and calcium phthalate according to the proportions given below was prepared by a procedure well known in the art and resulted in a free flowing sand mixture.

Sand Mixture Composition Chelford 60 sand 1.5 kg Commercial phenolic hot-box resin binder 30 g (2%) Commercial catalyst 7.5 g (0.5%) Calcium phthalate 4.5 g (0.3%) The sand mixture was used to produce 2.54 cm AFS tensile test pieces which were hardened in a heated corebox at 2500C for either 15 seconds or 60 seconds. Tensile strengths of the prepared specimens were measured 1. immediately upon removal from the heated corebox, and 2. after 24 hours storage at 220C and 55% relative humidity.

The 2.54 cm AFS tensile test pieces were used in a finning test casting as described in Example 3 and the results are given in Table 7.

TABLE 7

Sand Time of Tensile Strength Extent of Mixture Test kg/cm2 Finning 15s 60s curing curing Mix containing "As made" 1.48 5.20 calcium Slight phthalate 24 hours 6.75 21.37 Standard "As made" 2.67 7.45 Moderate/ hot-box mixture 24 hours 17.01 18.14 Severe These results show that the calcium phthalate retarded the core curing rate but did give a significant reduction in the amount of finning.

EXAMPLE 6 A sample of calcium isophthalate was prepared from isophthalic acid according to the method given in Example 5. The same quantities of materials were used.

The prepared calcium isophthalate was tested in the same manner as the calcium phthalate of Example 5.

Results are given in Table 8.

TABLE 8

Sand Time of Tensile Strength Extent of Mixture Test kg/cm2 Finning 15s 60s curing curing Mix containing "As made" 1.83 5.27 calcium Slight isophthalate 24 hours 9.28 18.14 Standard "As made" 2.67 7.45 Moderate/ hot-box mixture 24 hours 17.01 18.14 Severe It can be seen that calcium isophthalate additions reduce the core curing rate, but significantly reduce the degree of finning.

EXAMPLE 7 Calcium sebacate was prepared from sebacic acid according to Example 5. 50.5 g were reacted with 20 g of sodium hydroxide in a 1 litre aqueous solution. A solution of 28 g of calcium chloride in 100 ml of water was added to the disodium sebacate solution and the fine precipitate of calcium sebacate which formed was filtered, washed and dried.

The calcium sebacate was used in a hot-box resin bonded sand mixture as in Example 5. Tensile test specimens were made, tensile strengths were measured and a finning test casting was made as in Example 3.

The results are given in Table 9.

TABLE 9

Sand Time of Tensile Strength Extent of Mixture 1 Test kg/cm2 Finning 15s 60s curing curing Mix containing "as made" 0.98 4.22 Slight/ calcium sebacate 24 hours 7.66 18.56 Moderate Standard "As made" 2.67 7.45 I Moderate/ hot-box mixture 24 hours 17.01 18.14 Severe EXAMPLE 8 Aluminium terephthalate was prepared by reaction of aluminium chloride with sodium terephthalate. 1 66 g of terephthalic acid was dissolved in a solution of 80 g of sodium hydroxide in 1 litre of water.The temperature was raised to 900C until the acid had dissolved and reacted to form the disodium salt. 90 g of aluminium chloride was added at 75-8O0C and a fine white precipitate of aluminium terephthalate was formed. This was filtered, washed with water and finally dried at 11 00C.

The product was a fine white powder.

A hot-box resin bonded sand mixture was prepared with the following composition.

Chelford 60 sand 1.5 kg Commercial hot-box resin 30 g (2%) Commercial catalyst 7.5 g (0.5%) Aluminium terephthalate 7.5 g (0.5%) For comparison purposes a standard mix was prepared containing only the sand, resin and catalyst additions. 2.54 cm AFS tensile test specimens were made from each sand mixture and were cured for 30 seconds at 2300 C. The tensile strengths were measured immediately after ejection from the corebox (as made) and after 24 hours storage. Finning test castings were made according to Example 3, and the results are given in Table 10.

TABLE 10

Sand Time of Tensile Strength Extent of Mixture Test kg/cm2 Finning With "As made" 7.31 Slight/ aluminium Moderate terephthalate 24 hours 1 7.01 Moderate Standard "As made" 8.86 Severe 24 hours 13.01 The aluminium terephthalate gave a substantial reduction in the extent of finning defects in the test castings, without seriously affecting the cured strengths of the bonded sand cores.

EXAMPLE 9 Ferricterephthalatewas prepared as follows.41.5 gofterephthalic acid were dissolved in 400 ml of an aqueous solution containing 20 g of sodium hydroxide. The temperature of the mixture was raised to 900C and held until all the acid had dissolved. The solution was cooled to 800C and a solution of 27 g of ferric chloride in 50 g of water was added. Immediately a gelatinous precipitate formed which flocculated on standing. The product was filtered, washed and dried and the resultant powder was a mid-brown colour.

The ferric terephthalate was used in a hot-box resin bonded sand mixture comprising 2 kg Chelford 60 sand, 40 g commercial hot-box resin, 10 g commercial catalyst and 10 g of ferric terephthalate.

2.54 cm AFS tensile test specimens were prepared from the mixture and were cured at 2300C for 30 seconds. Tensile strengths were measured immediately on ejection from the heated corebox and after 24 hours storage. Afinning test casting was made according to the method given in Example 3.

A standard sand mixture containing only the hot-box resin and its catalyst was prepared and used for comparison, the results being shown in Table 11.

TABLE 11

Sand Time of Tensile Strength ' Extent of Mixture Test kg/cm2 Finning Hot-box resin bonded sand "As made" 7.87 with ferric Slight terephthalate 24 hours 19.83 addition Standard "As made" 8.30 Moderate/ 24 hours 11.18 Severe Sand strength properties were good and the addition of the ferric terephthalate to this hot-box resin bonded sand mixture produced a significant reduction in the finning defects in the test casting.

Use in shell process sands Shell moulds and cores prepared from phenolic resin coated sands also give rise to finning defects in castings as a result of the core or mould cracking and allowing a fine film of metal to run out of the casting into the crack and form the fin. Sometimes the cracking is very severe and can result in broken moulds and metal run-out.

The shell process requires the use of precoated sands to form the moulds and cores in heated patterns. The sands are precoated with the phenolic resin and a hexamine curing agent using either a warm coating process with a liquid resin or a hot coating process with a resin in flake form.

EXAMPLE 10 A sample of calcium terephthalate, prepared according to the method in Example 1, was incorporated into a precoated shell and sand by the warm coating method.

Two coated sands were prepared by the following procedure.The fresh silica sand was preheated to 500C and 200 kg was delivered to the batch roller mill. The liquid shell resin containing methanol as solvent was added to the sand and mixed for approximately 1 minute. The dry additions, 3 per cent Sierra Leone concentrate (a commonly used commercial form of Fe2O3), 0.36 per cent hexamine, 0.5 per cent calcium stearate, and the calcium terephthalate (for the trial sand) were added and were milled with the sand/resin mixture for 3 minutes during which time warm air was blown through the sand to remove the methanol solvent. Milling was continued for a further 3 minutes period with cold air being blown through the sand, after which the sand was discharged onto a 0.32 cm screen and the fine, coated sand particles passing through the screen were bagged.

The compositions of the two sands, a standard sand (A) and the trial sand (B) were as follows: (A) Redhill H sand 200 kg Fordath shell resin (dry weight) 6 kg (3%) Sierra Leone Concentrate 6 kg (3%) Hexamine 0.72 kg (0.36%) Calcium stearate 1.0 kg (0.5%) (B) Redhill H sand 200 kg Fordath shell resin 6 kg (3%) Sierra Leone Concentrate 6 kg (3%) Hexamine 0.72 kg (0.36%) Calcium stearate 1.0 kg (0.5%) Calcium terephthalate 1.10 kg (0.55%) The coated sands (A) and (B) were used in a number of laboratory tests to examine the effect of the calcium terephthalate addition on the properties of the coated sand. These tests and the results are described in Table 12.

TABLE 12 PROPERTIES OF PRECOATED SHELL SANDS (A) AND (B)

Test Sa d Standard A Trial B Melt point 84.40C 84.40C Transverse 65.38 kg 6628 kg Breaking Load (after 30 minutes) (after 30 minutes) (2.54 cm x 2.54cm section test bar cured for 60 s at 2500C - broken on 1 6220 kg 64.92 kg 5.08 cm centres) (after 24 hours) L (after 24 hours) Tensile Strengths 1q7/cm2 "hot"** 24 hours "hot''** 24 hours Curing time 30 s 5.98 9.63 3.52 10.76 Curing time 60 s 1036 1427 11.39 14-27 11.39 14.41 Curing time 120s 10.90 47.67 13.01 45.00 Cracking Test Cracking Time Cracking Time Specimens cured for 2 minutes No cracks in 180 Z- 250 C | 146s seconds Specimens cured for 2 minutes No cracks in 180 82500C 91 s91s seconds * The BCIRA Cracking Test measures the tendency of thin shell 'biscuits' to crack when exposed to a severe thermal stress. The biscuit, prepared by curing precoated shell sand in a suitably shaped heated mould. is a 100 mm diameter disc with a minor segment removed to leave a straight edge 50 mm long, notched at its centre point. The notch, which is 3 mm deep and has a 30 mm base serves as a crack initiator when the disc is placed with its centre on a 38 mm diameter disc heated at 7000C.

** Specimen tested immediately on ejection from corebox.

The specimen is placed on the heated block and the time taken for a crack to appear in the disc is recorded. The test is repeated 3 times and the average 'cracking time' is calculated.

The addition of calcium terephthalate to the shell sand did not affect its handling and processing properties, but the resistance to cracking of the cured sand was dramatically increased (no cracks being observed in 180 s) and the cracking tendency was consequently reduced substantially.

The precoated shell sands prepared for these tests was then used in a casting trial. Tappet cores, weighing 300 g, were produced in a foundry shell-core machine. Cores produced from sand containing the calcium terephthalate were of comparable standard to those produced from the normal sand mixture.

Cores from each sand mixture were used in diesel engine cylinder head castings and after cooling and removal of the spent cores the castings were examined. The normal sand mixtures was found to give a considerable amount of finning in the most inaccessible parts of the tappet holes. Use of the calcium terephthalate-containing sand cores resulted in much cleaner castings with a substantial reduction in finning defects.

EXAMPLE 11 A sample of calcium terephthalate, prepared according to Example 1 , was used to prepare precoated shell sands by the hot coating method.

The fresh silica sand was preheated in a small fluidised bed heater to approximately 1 300C. A preweighed amount of the hot sand was run into a blade mixer. The appropriate quantity of a hexamine solution was added, followed shortly afterwards by a resin in flake form. 10 seconds after addition of the resin, 0.6 per cent of calcium terephthalate was added to the trial mix.

After mixing for a 3 minute period, the sand was passed by a screw feed device to a vibratory classifier where the partially bonded sand was reduced to single grain size, cooled and bagged.

Two precoated sands were prepared by this procedure and had the following compositions: A. Normal sand Lynn S.S. 64 kg 25% hexamine solution 720 g (0.28% hexamine by weight) Fordath TPR 72 resin 3 kg B. With calcium terephthalate Lynn S.S. 64 kg 25% hexamine solution 720 g (0.28% hexamine by weight) Fordath TPR 72 resin 3 kg Calcium terephthalate 384 g (0.6 v6) The properties of cured specimens prepared from these coated sands were as shown in Table 13.

TABLE 13

Cure Time Normal Mix Trial Mix Test s A B Melt Point l 960C l 970C Transverse Breaking load Breaking load Strengths kg kg (on 2.54 cm x 1.27 cm cross section transverse test pieces, broken on 5.08 cm Hot Cold Hot* Cold centres) (24 h) (24 h) 60 37.68 61.74 33.60 56.75 120 46.76 54.93 40.86 55.84 0.45 kg breaking load 180 46.76 57.20 41.31 55.84 Hot: Specimens were tested immediately on ejection from the corebox.

The calcium terephthalate addition slightly reduced the strengths achieved for the bonded sands but the reduction would not be sufficient to affect the handling properties of cores or moulds.

Several 64 kg batches of each type of sand were prepared, and from this sand large cores for pipe fitting castings were produced. The cores, which weighed about 1 6 kg each were produced in standard shell comprising equipment in three minutes at a temperature of 250-2700C.

39 cores were prepared from sand containing calcium terephthalate and for comparison 54 cores were produced from the normal mix. (A in Table 13).

Coatings were made in modular iron using all these cores and the finished castings were examined for defects, particularly severe cracks or fins.

The results were as shown in Table 14.

TABLE 14

Acceptable with Number Good Scrap but considerable some Cast Castings Castings fettling needed cracks No. % No. % No.

Normal shell sand 54 38 70.4 3 5.6 13 24 29.6 Trial sand with calcium terephthalate 39 34 87.2 1 2.6 4 10.2 12.8 The addition of 0.6 per cent calcium terephthalate to the sand gave a substantial reduction in finning and cracking defects.

Claims (16)

1. The use, in a silica sand foundry mould- or core-making composition, of up to 5% by weight of an additive comprising at least one divalent or trivalent salt of a polycarboxylic acid.

2. A foundry moulding composition comprising a silica sand with a binder and with the additive of claim 1.

3. A foundry moulding composition according to claim 2 in which the salt is that of a dicarboxylic acid.

4. A foundry moulding composition according to claim 2 or claim 3 in which the salt is that of an aliphatic acid.

5. Afoundry moulding composition according to claim 4 in which the salt is that of maleic, malonic, adipic, succinic or sebacic acid.

6. A foundry moulding composition according to claim 2 or claim 3 in which the salt is that of an aromatic acid.

7. A foundry moulding composition according to claim 6 in which the acid is a phthalic acid.

8. Afoundry moulding composition according to claim 7 in which the acid is a terephthalic acid.

9. Afoundry moulding composition according to any one of claims 2 to 8 in which the salt is the calcium, magnesium, barium, iron or aluminium salt.

1 0. A foundry moulding composition according to claim 9 in which the salt is the calcium salt.

11. A foundry moulding composition according to claim 10 dependent upon claim 8 in which the salt is calcium terephthalate.

12. Afoundry moulding composition according to claim 11 in which the calcium terephthalate has been produced by double decomposition from aqueous solution and precipitated out at a temperature between 600C and 800C.

13. A foundry moulding composition according to claim 12 in which the said temperature is substantially 750C.

14. A foundry moulding composition according to any one of claims 2 to 1 3 in which the additive is present to the extent of weight of between 0.1% and 5% of the weight of refractory material.

1 5. A foundry moulding composition according to claim 14 in which the additive is present to the extent by weight of 0.25% to 0.75% of the weight of refractory material.

16. Afoundry mould made using the composition of any one of claims 2 to 15.