United Kingdom VAT & Duties Tribunals (Customs) Decisions

You are here: BAILII >> Databases >> United Kingdom VAT & Duties Tribunals Decisions >> United Kingdom VAT & Duties Tribunals (Customs) Decisions >> MacAnts Abrasives Ltd v Revenue & Customs [2007] UKVAT(Customs) C00243 (24 September 2007)
URL: http://www.bailii.org/uk/cases/UKVAT/Customs/2007/C00243.html
Cite as: [2007] UKVAT(Customs) C00243, [2007] UKVAT(Customs) C243

[New search] [Printable RTF version] [Help]

Mac'Ants Abrasives Ltd v Revenue & Customs [2007] UKSPC (24 September 2007)

C00243

CUSTOMS DUTY- whether product imported by the Appellant was aluminium oxide other than artificial corundum within CN code 2818.20.00 as argued by the Appellant – no - or whether it was artificial corundum falling within CN code 2818.10.90 as argued by the Respondents – yes - whether anti-dumping regulation invalid – no - appeal dismissed – Council Regulation (EEC) No. 2658/87 Annex 1 Heading 28.18; Council Regulation (EC) No. 1951/97 Art 1

LONDON TRIBUNAL CENTRE

MAC'ANTS ABRASIVES LIMITED Appellant

- and -

THE COMMISSIONERS FOR HER MAJESTY'S

REVENUE AND CUSTOMS

Respondents

Tribunal: DR A N BRICE (Chairman)

MR A MCLOUGHLIN

Sitting in London on 2 to 6 July 2007

Andrew Young of Counsel, instructed by Messrs Eversheds Solicitors, for the Appellant

Kieron Beal of Counsel, instructed by the Solicitor for HM Revenue and Customs, for the Respondents

© CROWN COPYRIGHT 2007

FINAL DECISION

The appeal

Mac'Ants Abrasives Limited (the Appellant) appeals against deemed decisions on review confirming post-clearance demands for anti-dumping duty, additional customs duty and value added tax. The first demand was dated 4 January 2001; was for £21,567.00; and was in respect of a consignment entered on 12 January 2001. The second demand was dated 28 March 2001; was for £1,342,241.71; and was in respect of consignments entered between 30 March 1998 and 19 January 2001. The third demand was dated 31 January 2002; was for £75,691.48; and was in respect of four further consignments entered between 5 June 2000 and 4 September 2000.

The three demands were issued because the Respondents were of the view that the material imported by the Appellant was brown artificial corundum; that it should be classified in the combined nomenclature under heading 2818 10 90; and accordingly that it was liable to a higher rate of customs duty, to anti-dumping duty and value added tax. The Appellant appealed because it was of the view that the imported material was aluminium oxide other than artificial corundum; that it should be classified under heading 2818 20 00; and so should not be liable to additional customs duty, anti-dumping duty and value added tax.

The progress of the appeal

One of the issues raised by the Appellant in the appeal was whether a sampling exercise conducted by the Respondents had been validly conducted under the Community Customs Code and the Implementing Code. At a hearing for directions held on 5 September 2002 it was directed that the question of the validity of the sampling exercise should be dealt with as a preliminary issue and listed for one day's hearing before a chairman alone. That hearing took place on 9 July and 29 October 2003 and the Preliminary Decision of the Tribunal was released on 12 November 2003. The Tribunal decided that the sampling exercise was validly conducted. As that decision did not determine the appeal the Tribunal directed that further dates be fixed for the hearing of the remaining issue, which was whether the material imported by the Appellant was artificial corundum or aluminium oxide other than artificial corundum. We refer to that issue as the tariff classification issue.

The Appellant appealed to the High Court from the Preliminary Decision of the Tribunal but that appeal did not proceed to a hearing. The dates fixed for the hearing of the tariff classification issue by the Tribunal were 27 and 28 June 2005 when the appeal was also listed before the chairman alone. At the beginning of the hearing on 27 June 2005 the Appellant raised a third issue, namely whether the anti-dumping regulation was invalid. We refer to that issue as the validity issue. As the parties were not then ready to argue the validity issue it was agreed that the hearing of the tariff classification issue would proceed and that, if the Appellant were not successful, a further date would be fixed for the hearing of the validity issue. The Tribunal proceeded to hear the expert evidence for the Appellant and some of the expert evidence for the Respondents. However, that evidence was not completed at the end of the hearing on 28 June 2005 and so that hearing had to be adjourned part-heard. At a directions hearing held on 19 October 2005 the Tribunal directed that both parties should, within fourteen days, provide their dates to avoid in November and December 2005 or the week beginning 9 January 2006 for the resumed hearing. However, it then became impossible to re-list the resumed hearing because of the unavailability of the parties.

By January 2007 the appeal had not been re-listed and the Tribunal formed the view that it might not be in the interests of justice to give a decision on the tariff classification issue having heard the Appellant's expert evidence in June 2005 and the rest of the expert evidence for the Respondents in 2007. Representations from the parties were invited and in January 2007 the Tribunal gave notice of a hearing for directions which was held on 10 April 2007. The Tribunal then directed that there should be one further hearing of the appeal from 2 to 6 July 2007 and that that hearing should consider the whole of the tariff classification issue, as a complete re-hearing of the argument and evidence originally heard on 27 and 28 June 2005, and should also consider the whole of the validity issue after which the final Decision of the Tribunal would be given. As the hearing on 2 to 6 July 2007 was a complete re-hearing of the tariff classification issue and the first hearing of the validity issue the Tribunal included both a chairman and member.

The legislation

The legislation relating to the tariff classification

Council Regulation (EEC) No. 2658/87 contains the provisions about the tariff classification of goods coming into the Community and the rates of duty. Article 1 establishes a combined nomenclature (CN) and Annex I sets out the combined nomenclature together with the rates of duty of the common customs tariff. Annex I is amended annually. Heading 2818 reads:

"2818

ARTIFICIAL CORUNDUM, WHETHER OR NOT CHEMICALLY DEFINED; ALUMINIUM OXIDE; ALUMINIUM HYDROXIDE

Artificial corundum, whether or not chemically defined

White, pink or ruby with an aluminium oxide content exceeding 97.5% by weight 281810 10

Other 281810 90

Aluminium oxide, other than artificial corundum 281820 00

Aluminium hydroxide 281830 00"

At the relevant time customs duty at the rate of 5.2% was charged on artificial corundum falling within codes 2818 10 10 and 2818 10 90 and at the rate of 4% on aluminium oxide falling within code 2818 20 00.

The relevant parts of the Harmonised System Explanatory Notes for heading 28.18 describe the products in the following way:

(A) ARTIFICAL CORUNDUM, WHETHER OR NOT CHEMICALLY DEFINED

Artificial corundum is formed by fusing aluminium oxide in an electric furnace. The aluminium oxide may contain small proportions of other oxides (e.g., titanium oxide, chromium oxide) either deriving from the natural starting material (bauxites) or added to improve, for example, the hardness of the fused grain or to modify the colour. However, mechanical mixtures of artificial corundum and other substances, such as zirconium dioxide, are excluded (heading 38.24).

Artificial corundum is put up in small pieces or masses, crushed or in grains; it is more resistant than ordinary aluminium oxide to the action of air and acids, and is very hard. It is used, e.g., as an abrasive, in the manufacture of refractory conglomerates (such as mullite and sillimanite, mixtures of corundum with pure refractory clay and with anhydrous aluminium silicates, respectively) or of laboratory utensils and in the electrical industry.

(B) ALUMINIUM OXIDE, OTHER THAN ARTIFICIAL CORUNDUM

Aluminium oxide (anhydrous or calcined aluminium oxide) (Al/2 O/3) is obtained by calcining the aluminium hydroxide described below, or from ammonium alum. It is a light white powder, insoluble in water, specific gravity about 3.7.

Uses include, e.g., in aluminium metallurgy, as a filler for paints, in the manufacture of abrasives and synthetic precious or semi-precious stones (rubies, sapphires, emeralds, amethysts, aquamarines, etc.), as a dehydrating agent (for drying gases), or as a catalyst (manufacture of acetone and acetic acid, cracking operations, etc.). …

This heading does not include:

(a) Natural corundum (native aluminium oxide) … "

Thus it will be seen from the HSENs that aluminium oxide other than artificial corundum is a light white powder obtained by calcining (heating). Artificial corundum is of a granular appearance and is made by fusing aluminium oxide in an electric furnace.

The legislation relating to the anti-dumping duty

The legislation relating to the anti-dumping duty is contained in Council Regulation (EC) No 1951/97 (the anti-dumping regulation) which was introduced in October 1997. Article 1 imposed an anti-dumping duty on imports of artificial corundum falling within CN codes 2818 10 10 and 2818 10 90 and originating in the People's Republic of China. Article 2 stated that the rate of anti-dumping duty was ECU 204 per tonne. Recitals (17) and (18) described the product in the following way:

"(17) The product concerned by this review is fused aluminium oxide, also known as artificial corundum falling from 1 January 1997 onwards within CN codes 2818 10 10 and 2818 10 90. It is principally produced in two basic varieties for which the chemical formula is identical (AI/2 0/3):

- brown artificial corundum consisting of 94 to 97% aluminium oxide,

- white artificial corundum consisting of 97.5% to 99.5% aluminium oxide.

Small quantities of pink artificial corundum which has a content of aluminium oxide similar to the white artificial corundum are also produced and sold in the Community. Hence the term "white artificial corundum" will be intended hereafter to include pink artificial corundum as well.

(18) The raw material for the production of brown artificial corundum is bauxite in natural or calcined form, while for the production of white artificial corundum, calcined aluminium oxide is used, which is a processed form of bauxite. Artificial corundum is manufactured by melting the raw material in electric-arc furnaces at temperatures higher than 2,000° C.

Artificial corundum, due to its specific characteristics, notably hardness, is mainly used in the production of abrasive materials such as grinding wheels, cutting wheels, sandpaper and in the production of refractory materials.

The production process generates a certain amount of lower quality artificial corundum containing less than 94% of aluminium oxide. Its use is limited to the production of resin-bonded grinding wheels and to sand-blasting purposes."

The issues

For the purposes of this Final Decision we had to determine the tariff classification issue and the validity issue. These may be stated as:

(1) whether the product imported by the Appellant was aluminium oxide falling within CN code 2818 20 00 (as argued by the Appellant) or was brown artificial corundum falling within CN code 2818 10 90 as argued by the Respondents; and

(2) whether Council Regulation (EC) No 1951/97 (the anti-dumping regulation) was invalid.

The evidence

Three bundles of documents were produced, namely a Combined Bundle (also called the Appellant's Bundle), a Respondents' Bundle and an Experts' Bundle. [In addition to the evidence, five bundles containing copies of seventy-one authorities were also produced.]

Expert evidence was given on behalf of the Appellant by Professor Brian Derby, the Professor of Materials Science at the University of Manchester Materials Science Centre. Expert evidence was given on behalf of the Respondents by Dr Richard Neil White BSc PhD, the Microstructure Laboratory Manager of Ceram Research Limited (Ceram).

The facts

The only evidence we heard was the expert evidence. The parties did not dispute the facts found in the Preliminary Decision.

For the purposes of this Decision it is relevant that the Appellant trades from premises in Dinnington, Sheffield and sells a range of abrasive cleaning materials. The Appellant also sells a complete range of fused aluminium oxide based products and has processing facilities in Europe and fusion plants in China. At the relevant time the Appellant purchased and imported the material at issue in the appeal from suppliers in The People's Republic of China. The material was in granular form. On 12 January 2000 a consignment was entered at Felixstowe as aluminium oxide under CN code 2818 20 00. On 24 January 2000 it was examined by the Respondents and four samples were taken by Mrs Josey, an Officer of the Respondents. Mrs Josey sent the samples to the Government Chemist for analysis. On 23 March 2001 Officers of the Respondents visited the premises of the Appellant and took samples from seven bags. These were also sent to the Government Chemist for analysis.

The expert evidence

Each expert witness prepared reports in 2002, 2005 and 2007. We begin our summary of the expert evidence by giving a general description of the materials mentioned in the legislation. We then refer to the results of four early analyses of the composition of the material imported by the Appellant, which analyses were referred to in the first (2002) reports of the expert witnesses. We then summarise the 2002 experts' reports. After that we refer to the results of two later analyses and then summarise the supplementary reports of the expert witnesses prepared in 2005. We then briefly summarise the second supplementary reports of the expert witnesses prepared in 2007, before summarising the oral evidence of the expert witnesses.

The materials mentioned in the legislation

Natural corundum is the description of the natural crystalline mineral form of aluminium oxide. The term artificial corundum is not in common usage in scientific circles but appears to be widely used in trading contexts. The word alumina is generally used in scientific literature to describe all materials with the chemical composition of aluminium oxide (Al/2 O/3). Corundum structure material is commonly referred to as a-alumina. However, in this Decision we use the words in the legislation, namely aluminium oxide and corundum.

Because natural corundum is found very sparingly in nature, the natural raw material of most forms of aluminium oxide is bauxite. Bauxite is found in many parts of the world and is an earthy rock consisting of hydrated aluminium hydroxide with variable proportions of iron oxides and other impurities including titanium dioxide, magnesium oxide, calcium oxide and silicon oxide.

Bauxite can be purified by dissolution in sodium hydroxide to remove the silicon, iron and titanium compounds, followed by filtration and re-precipitation. The dissolution removes most of the impurities from the bauxite and the purified product is then calcined (heated) at between 1300 and 1400 degrees centigrade. These temperatures are not high enough to melt the aluminium oxide. The calcining, or heating, produces a fine grained, low density, high surface area powder. The product produced by this process would properly be classified under heading 2818 20 00. However, this appeal is not concerned with calcined aluminium oxide but with fused aluminium oxide.

Fusion is used as an alternative method of purifying bauxite. Brown fused aluminium oxide is obtained by melting raw bauxite by the use of graphite electrodes (electric arc fusion) in a water-cooled steel shell. In order to increase the aluminium oxide content of the end product a carbon source (such as coke) and iron filings are first added to the raw bauxite to reduce some of the impurities after which a large electric current is passed through. The iron melts and the carbon (coke) reduces the impurity components to a metallic state. The level of the impurities decreases over time during the melting process and the final product is purer than the initial product. The impurities sink to the bottom of the furnace and can be tapped off. Thus most (but not all) of the impurities are removed; because titania improves the toughness of the final product the aim is to ensure that some of the titania is fused with the aluminium oxide. At the end of the process the molten aluminium oxide is cast as an ingot, crushed and graded.

White fused aluminium oxide is formed by the fusion of aluminium oxide which has been calcined without a prior chemical dissolution. An ingot is formed which is later crushed and graded to give the required product. White fused aluminium oxide is very pure but brown fused aluminium oxide contains some impurities. The material at issue in this appeal is brown fused aluminium oxide. There are other forms of aluminium oxide, including tabular aluminium oxide and sol-gel aluminium oxide neither of which are in issue in this appeal.

The first four analyses

Before 2002 reports about the composition of the material imported by the Appellant were prepared by:

(1) Sheffield Testing Laboratories Limited (Sheffield);

(2) London & Scandinavian Metallurgical Co Limited (Metallurgical);

(3) Ceram Research Limited (Ceram); and

(4) the Laboratory of the Government Chemist.

The Sheffield analyses were dated 21 October 1997, 9 March 2000 and 5 June 2000. The analysis of 21 October 1997 was a chemical analysis of a sample described as brown aluminium oxide. The conclusion was that the sample contained 97.1% by mass of aluminium oxide, the remainder being other oxides. The magnesium oxide content was 0.09% by mass. The analysis of 9 March 2000 was a chemical analysis of a sample which was also called brown aluminium oxide. The conclusion was that the sample contained 95.6% by mass of aluminium oxide, the remainder being other oxides. The magnesium oxide content was 0.40% by mass. The analysis of 5 June 2000 was a structural analysis of a sample of brown aluminium oxide (corundum). The intensity was said to be "high".

The Metallurgical analyses were dated 18 December 2000, 21 December 2000 and 15 January 2001 and related to particle size distribution. Both expert witnesses agreed that the tests about particle size had no significance and so we do not consider these any further.

The Ceram test was completed in March 2001. On 12 March 2001 a sample of granular brown material was received by Ceram from the Government Chemist. Ceram was asked to examine the material to identify it and to confirm the method and temperature of manufacture. After detailed examination a test report was prepared by Ceram on 28 March 2001. The report confirmed the identity of the material as being consistent with brown fused aluminium oxide but could not confirm either the method or temperature of production. The report added that aluminium oxide melts at some 2050 degrees centigrade and is normally fused using an electric fusion method.

On 1 November 2001 the Government Chemist produced a written report on the seven samples taken at the Appellant's premises on 23 March 2001 and also on the sample taken at Felixstowe on 24 January 2000. The relevant extracts from the report stated:

"Samples B3003857-B3003862 all confirmed to be alpha aluminium oxide, corundum, plus a small amount of silicon dioxide.

Sample B3003863 identified as iron aluminium silicate, almandine. As this sample falls outside the scope of the Regulation, the aluminium oxide content was not measured. See ** below

[percentage aluminium oxide]

B3003857 95.9

B3003858 95.5

B3003859 95.0

B3003860 93.6

B3003861 95.5

B3003862 94.3

B3003863 **

Samples B3003857-B3003862, except B3003860, lie within the range 94 to 97% aluminium oxide content as specified for brown artificial corundum in Council Regulation (EC) No. 1951/97.

The sample originally taken at Felixstowe, HMCE reference 00/0009 LGC reference A3000613 has also been confirmed to be corundum and has an aluminium oxide content of 95.4%."

The 2002 reports of the expert witnesses

Professor Derby did not himself undertake any analyses of the material imported by the Appellant; his evidence related to matters of principle and to a commentary on the analyses conducted by others. Dr White, as the Microstructure Laboratory Manager of Ceram, was connected with the analysis conducted by Ceram for the Government Chemist in March 2001. He also commented on matters of principle and on the analyses conducted by others.

Professor Derby's first report was dated 15 August 2002 and was entitled "The Identification of "Artificial Corundum"". It is relevant that, when preparing his first report, Professor Derby was shown a copy of the anti-dumping regulation but not a copy of the tariff classification or the HSENs.

The report concluded that the definition of artificial corundum in the anti-dumping regulation was ambiguous because it did not define a material but defined a method of processing the material (namely manufacturing by melting the raw material in electric furnaces at temperatures higher than 2,000 degrees centigrade). The report stated that it was difficult to devise a set of tests to determine whether a sample of a given material fitted that description. The melting point of pure aluminium oxide was difficult to measure accurately and the presence of impurities would reduce the melt temperature. It followed that it was not possible to be certain that the melt temperature was above 2000 degrees centigrade and "it was likely that the melt processing of alumina may be possible at temperatures below 2000 degrees centigrade".

Professor Derby's first report also stated that any set of tests to determine the method of manufacture (of artificial corundum) would require analytical and characterisation equipment including chemical analysis; X-ray diffraction, thermal analysis, scanning electron microscopy and transmission electron microscopy; such equipment would only be found in very well equipped research laboratories and the interpretation of the data was a highly skilled task. Even with access to all this equipment it would not be certain that the precise melt temperature of the aluminium oxide could be determined solely from tests conducted on the finished product.

On 29 August 2002 Dr White made a formal statement about the tests conducted by Ceram and about the test report dated 28 March 2001. On 25 October 2002 Dr White produced his first report which commented on Professor Derby's report of 15 August 2002 and the Sheffield and the Metallurgical analyses. Dr White concluded that both brown and white fused aluminium oxide were synonymous with artificial corundum. The analyses of the brown aluminium oxide from the Appellant showed that the material was chemically consistent with brown fused aluminium oxide. The melting point of pure corundum (a-alumina) was about 2050 degrees centigrade and calculations on the chemical analyses of the brown aluminium oxide supplied by the Appellant were consistent with a melting point in excess of 2000 degrees centigrade.

On 13 November 2002 the two experts met and on 27 November 2002 they signed an agreed report. This indicated that they agreed: that the material supplied by the Government Chemist to Ceram in March 2001 was fully consistent with brown fused aluminium oxide; that brown fused aluminium oxide was produced by a fusion process with calcined bauxite as a starting component; that the melting point could not be determined by the analyses done by Ceram; and that the chemical composition of the Appellant's material, as determined by Sheffield, fell within the range of brown fused aluminium oxide material but that there were insufficient sample data to identify the mean composition of the consignments to the Appellant. The joint report went on to state that a number of points remained in dispute. These were: the melting point of pure aluminium oxide; the calculation, from the analyses supplied by the Appellant, of the reduction in the melting point of the brown fused aluminium oxide by the presence of impurities; and the methodology, assumptions and values derived from the analyses.

The two later analyses

After 2002 further analyses were carried out on the material imported by the Appellant by Mr Robert Burton of Sheffield Hallam University and by a French firm called Setaram SA (Setaram).

Mr Burton was jointly instructed by both parties. His undated report described that four samples of fused aluminium oxide had been analysed to determine the level of impurities in each. We understand that these four samples were part of the four samples taken by Mrs Josey on 24 January 2000. The results were that the four samples contained aluminium oxide of 94.98%, 95.78%, 95.76% and 95.71% respectively with the balance being other oxides. Magnesium oxide was present as 0.32%, 0.29%, 0.29% and 0.30% respectively.

The Seteram analyses were carried out between 7 January 2005 and 18 January 2005 at the request of the Appellant alone. Setaram produced thermal analysis reports which were described as "the results of experiments carried out with a thermo-analyser to determine the melting temperatures of alumina powder". The Appellant claimed that the samples used by Seteram were sub-samples of the four samples taken by Mrs Josey on 24 January 2000 but there was no evidence to support that claim. The methodology of the tests was that the analysis was conducted in a tungsten crucible under helium with temperatures from 600 degrees centigrade to 2200 degrees centigrade. The written results in degrees centigrade were:

Sample Temperature at Temperature at

beginning of melting second step of melting

1 1980.9 2006.8

2 2023.6 (melting observed in one step)

3 1580.1 1997.7

4 1703.9 1928.9

Attached to each written report was a diagram and these gave the following temperatures in degrees centigrade:

Sample Onset point Peak Second onset point

1 1978.4 2034.7 2004.3

2 2046.4 2024.2

3 1560 2014.2 1977.6

4. 1682 2002.0 1965.8

The 2005 supplementary reports of the expert witnesses

The Setaram analyses had not been available in 2002 when the expert witnesses prepared their first reports. Those analyses were produced in January 2005. On 9 June 2005 Professor Derby signed a first supplementary report which described some of the terms used in the analysis of the thermal stability of phases as represented by the equilibrium phase diagrams used in the interpretation of the thermal analysis data supplied to the Appellant by Setaram. No summary was provided but Professor Derby concluded that, in the case of a pure material, the melting transition occurs at a unique temperature but that, where there was a mixture of components, melting occurs over a range of temperatures. Equilibrium phased diagrams were used to indicate graphically the range of composition of phases that were present in equilibrium under given conditions of temperature and pressure. Extensive regions of variable composition and temperature could have no more than two phases co-existing in equilibrium. When a two phase region consisted of a solid and a liquid phase in equilibrium the line on the diagram that separated the two phase region from the single phase liquid region was called the liquidus. Conversely, the line which separated the two phase region from regions that were completely solid was called the solidus.

Professor Derby's first supplementary report went on to state that the four samples analysed by Setaram showed a marked variation in the reported temperature of initial melting which was as low as 1580 degrees centigrade in the third sample. The Setaram data assumed the melting point of pure aluminium oxide to be in excess of 2050 degrees centigrade. The data was consistent with the chemical analysis that the samples were composed of approximately 95.5% aluminium oxide with the balance being a range of other oxides. The major impurity was titanium oxide. The presence of titanium oxide in isolation would reduce the solidus temperature to 1840 degrees centigrade and the presence of other impurities would reduce this still further. Also, the rate of heat loss in a furnace was not a simple linear function of the temperature difference because heat transfer between two bodies was proportional to the difference in temperature to the fourth power. For a small increase in temperature above 1600 degrees centigrade there was a significantly greater relative increase in the power required to heat the furnace.

On 16 June 2005 Dr White signed a first supplementary report commenting on a number of documents put to him by the Respondents, namely, the undated report by Mr Robert Burton; the Setaram thermal analysis reports; and Professor Derby's supplementary witness statement of 9 June 2005. On the subject of the Setaram results Dr White noted that the size of the samples was very small (0.04897g to 0.06964 g). No information had been given as to how such small samples had been taken from the samples taken by Mrs Josey. The second step onset points would have been the onset of the melting of the aluminium oxide. The lower first step onset point (the solidus) indicated that three samples (1, 3 and 4) contained impurities that would substantially reduce the melting point by from 25.9 to 417.6 degrees centigrade. Dr White concluded that the samples were not uniform in composition and the data was not consistent with representative sampling of the four sub-samples. The degree of variation was indicative of the sampling methodology.

The 2007 second supplementary reports of the expert witnesses

On 12 May 2007 Professor Derby signed a second supplementary report which contained some information about the melt processing of alumina from bauxite ore to produce brown or white fused alumina for industrial abrasive applications. He concluded that there was considerable variation in the scientific literature for the melting temperature of pure aluminium oxide; that there was little consensus as to the actual temperature required to produce fused alumina; that there was extreme difficulty in making measurements of temperature above 1600 degrees centigrade; that it was not necessary to obtain a temperature in excess of 2000 degrees centigrade in order to form a pourable alumina melt of sufficient quantity for fused alumina production; and that it was impossible to determine, with any certainty, the temperature at which fused alumina was produced by simple microstructural and chemical analysis. Professor Derby also expressed the opinion that, on the balance of probabilities, it would be safe to assume that a complex mixture below that of pure alumina would melt at a temperature below 2000 degrees centigrade and that, in his opinion, it was not necessary for the mixture within the melting furnace to be taken to temperatures greater than 2000 degrees centigrade in order to produce fused alumina of industrial quality because bauxite ore was likely to contain sufficient impurities to result in a melting temperature below that heat.

On 11 May 2007 Dr White produced a second supplementary report which dealt with a number of matters. These included some additional comments on the Setaram analyses and concluded that the melting point was less than that of pure alumina due to the presence of other components; the onset of the melting of the alumina within the samples was shown to be from 1928.9 to 2023.6 degrees centigrade but the two lower values were less reliable due to a higher correction factor. Also, the onset of melting temperatures represented the minimum temperature to which the samples had been heated.

The oral evidence of the expert witnesses

In oral evidence Professor Derby re-emphasised the variation in the results of, and the difficulties of measuring, the melt temperature of pure alumina and stated that this led to uncertainty. There were still many opinions but he thought that "round about 2,000 degrees centigrade" was the melting point of pure alumina. He accepted that the material imported by the Appellant was fused aluminium oxide and not anhydrous aluminium oxide or calcined aluminium oxide or aluminium hydroxide and also accepted that the material imported by the Appellant was brown fused alumina. In oral evidence Dr White expressed the opinion that the normal melt temperature of pure alumina was about 2052 degrees centigrade and that the normal melt temperature during the production process of artificial corundum was about 2020 degrees centigrade and possibly higher.

Reasons for Decision

We now consider separately each of the issues which we have to determine.

Issue (1)- Was the imported product aluminium oxide or brown artificial corundum?

The first issue is whether the product imported by the Appellant was aluminium oxide falling within CN code 2818.20.00 (as argued by the Appellant) or was brown artificial corundum falling within CN code 2818.10.90 (as argued by the Respondents).

The parties agreed that the material imported by the Appellant was aluminium oxide which had been fused in an electric furnace. It was also agreed that the material imported by the Appellant originated in the People's Republic of China. However, the Appellant argued that the reference to a temperature of "higher than 2000 degrees centigrade" in recital (18) of the anti-dumping regulation was a crucial part of the description of artificial corundum and that there was no evidence that the material imported by the Appellant had been melted at that temperature. The Respondents, on the other hand, argued that recital (18) was descriptive only and that the law was as stated in the combined nomenclature and in Articles 1 and 2 of the anti-dumping regulation which did not mention the melt temperature. The Appellant further argued that the effect of recital (17) was that fused aluminium oxide, if it consisted of less than 94% of aluminium oxide, was not artificial corundum whereas the Respondents relied on the last paragraph of recital (18) and argued that fused aluminium oxide containing less than 94% of aluminium oxide was lower quality artificial corundum but was still artificial corundum.

Thus as argued this issue raised three distinct questions, namely:

(a) whether the reference to a temperature of "higher than 2000 degrees centigrade" in recital (18) of the anti-dumping regulation was a crucial part of the description of artificial corundum;

(b) if so, whether the material imported by the Appellant had been melted at temperatures lower than 2000 degrees centigrade; and

(c) whether fused aluminium oxide, if it consisted of less than 94% of aluminium oxide, was or was not artificial corundum.

Question (a) -Is the temperature crucial?

The first question is whether the reference to a temperature of "higher than 2000 degrees centigrade" in recital (18) of the anti-dumping regulation is a crucial part of the description of artificial corundum.

In considering the arguments of the parties we recall that this appeal concerns demands for payment of both customs duty and anti-dumping duty so we start with the legislation relating to customs duty, namely, the tariff classification for heading 28.18 which is:

"28.18

ARTIFICIAL CORUNDUM, WHETHER OR NOT CHEMICALLY DEFINED; ALUMINIUM OXIDE; ALUMINIUM HYDROXIDE

Artificial corundum, whether or not chemically defined

White, pink or ruby with an aluminium oxide content exceeding 97.5% by weight 281810 10

Other 281810 90

Aluminium oxide, other than artificial corundum 281820 00

Aluminium hydroxide 281830 00"

The tariff classification thus distinguishes between artificial corundum on the one hand and aluminium oxide (other than artificial corundum) on the other. This recognises that artificial corundum is one form of aluminium oxide. It is also relevant that the heading includes the words "whether or not chemically defined". However, the tariff classification gives no further clue as to how to distinguish between aluminium oxide which is artificial corundum on the one hand and other forms of aluminium oxide on the other.

Assistance is given by the HSENs which, although they do not have binding force, are of a persuasive nature. :The HSENs make it clear that artificial corundum is formed by fusing aluminium oxide in an electric furnace. The resulting material may contain small proportions of other oxides but it still remains artificial corundum. The material occurs as small pieces or masses crushed or in grains. On the other hand aluminium oxide other than artificial corundum is obtained by calcining (heating) and is a light white powder. Thus the distinction is made between aluminium oxide which has been formed by fusion in an electric furnace (which is artificial corundum) on the one hand and other aluminium oxide which is produced as a result of calcining (heating) on the other. The product described as "aluminium oxide other than artificial corundum" in paragraph (B) of the HSEN would be classified under CN heading 2818 20.

Neither the tariff classification nor the HSENs make any reference to the temperature at which the fusion in an electric furnace has to take place. Accordingly, we conclude that, for the purposes of customs duty, the temperature is not relevant to the tariff classification. As it was not disputed that the material imported by the Appellant was aluminium oxide which had been fused in an electric furnace we conclude that, for the purposes of customs duty, it should be classified under heading 2818 10 90. It should not be classified under heading 2818 20 00 because it was not formed by calcining.

Turning to the anti-dumping legislation, it follows from Article 1 of the regulation that, because the Appellant imported artificial corundum falling within CN code 2818 10 90 which originated in the People's Republic of China, anti-dumping duty is also payable. The Article imposing the duty refers only to the CN code and does not contain any description of the method of producing the product.

The Appellant relied heavily upon recitals (17) and (18) of the anti-dumping regulation. However, these are recitals only and so are descriptive but not directive. Recital (17) repeats the provision that the product concerned is fused aluminium oxide and is "principally" produced in two basic varieties, brown and white. Recital (18) states that the raw material for brown artificial corundum is bauxite and for white artificial corundum is calcined aluminium oxide; the process for both is fusion in an electric furnace. Then appears the sentence: "Artificial corundum is manufactured by melting the raw material in electric-arc furnaces at temperatures higher than 2,000 degrees centigrade." In our view this sentence does not mean that if the temperature does not reach 2000 degrees centigrade the product cannot be artificial corundum. What it does do is to describe the normal fusion process and gives the order of magnitude of the temperature used.

We therefore conclude that the reference to a temperature of "higher than 2000 degrees centigrade" in recital (18) of the anti-dumping regulation is descriptive and not directive and so is not a crucial part of the description of artificial corundum. The crucial part of the description is that the product is formed by fusion in an electric furnace and is not formed by calcining (heating) at lower temperatures.

Having reached that conclusion we are also able to conclude our consideration of the first issue we have to determine (the tariff classification issue). We find that the product imported by the Appellant was brown artificial corundum falling within CN code 2818.10 90 and was not aluminium oxide other than artificial corundum falling within CN code 2818.20.00. That means that we do not have to answer the second questions arising from the tariff classification issue but, as lengthy expert evidence was given, we briefly express our views.

Question (b) -Was the imported material melted at temperatures below 2000 degrees?

The second question arising out of the tariff classification issue is whether the imported material was melted at temperatures below 2000 degrees centigrade. This question only arises if the reference to a temperature of "higher than 2000 degrees centigrade" in recital (18) of the anti-dumping regulation is a crucial part of the description of artificial corundum (and we do not think that it is). The burden of proof in this appeal is on the Appellant and so we ask whether the Appellant has produced evidence to satisfy the Tribunal that the material imported by it was melted at temperatures lower than 2000 degrees centigrade.

We approach this question by first considering the evidence of the expert witnesses and we will then turn to apply our conclusions to the words of recital (18).

Both expert witnesses agreed that the melting point during the fusion process of pure aluminium oxide is difficult to measure at the higher ranges and that there is a divergence of opinion. One authority gives a melt temperature of 2054 degrees centigrade, another of 2020 degrees centigrade, and others of 2015 or 2050 degrees centigrade. Professor Derby did not agree with Dr White that pure aluminium oxide melted at 2050 degrees centigrade but agreed that the melt temperature of pure alumina was about 2,000 degrees centigrade. Neither expert witness was of the view that the melting point of pure aluminium oxide was below 2000 degrees centigrade. We therefore find that in most situations the melting point of pure aluminium oxide would exceed 2000 degrees centigrade and could be in the region of 2020 to 2050 degrees centigrade.

Both expert witnesses agreed that the presence of one or more impurities in the raw material (bauxite) would reduce the melt temperature and that the presence of multiple impurities would cause a greater reduction than the presence of one. However, there was no agreement about the amount of such reduction nor about the levels of the reduced temperatures. Professor Derby expressed the opinion that "it is likely that melt-processing of aluminium oxide may be possible at temperatures below 2000 degrees centigrade". We therefore find that the presence of impurities reduces the melt temperature of pure aluminium oxide but, on the evidence before us, we are unable to state what the levels of the reduced temperature would be.

Both expert witnesses agreed that the addition of 3% of magnesium oxide to an aluminium oxide melt would reach eutectic at 1925 degrees centigrade. (Eutectic is a mixture whose constituents are in such proportion that it melts and solidifies at a single temperature that is lower than the melting point of the constituents or any other mixture of them.) However, although magnesium oxide was present in the Sheffield samples it was present in levels of 0.09% and 0.4%. In the four samples examined by Mr Burton the levels were 0.32%, 0.29%, 0.29% and 0.30%. These levels are well below 3% and so this evidence does not satisfy us that the melt temperature of the aluminium oxide in the samples was below 2000 degrees centigrade. We accept the evidence of Dr White that it is possible to form an aluminium oxide melt below 2000 degrees centigrade but that such a melt would be most unlikely to comply with the definition of brown fused artificial corundum.

Both expert witnesses agreed that during the fusion process different temperatures are present. Initially the product is in the form of a solid; at a higher temperature it turns into a liquid/solid emulsion; and at even higher temperatures it turns into a liquid when it has reached its melting point. When cooled it returns to a solid state. Thus the onset of melting starts at lower temperatures and is completed at higher temperatures and in between there is a mixture of liquids and solids. Also, the different materials melt at different temperatures. The impurities melt first and, as they melt and become separated, the aluminium oxide becomes purer thus requiring a higher melt temperature. Also, when the electric current is first passed between the electrodes at the centre of the furnace heat is produced. The product surrounding the electrodes melts first and a temperature gradient is set up within the system. The heat then extends to the product further away from the electrodes until it comes into contact with the steel shell of the furnace which is surrounded by a water cooled jacket. At this stage some of the product near the steel shell may not have melted but the core of the ingot near the electrodes could have reached temperatures in excess of 2000 degrees centigrade.

We accept the evidence of Professor Derby that it would be very difficult for an importer of artificial corundum to determine the exact temperature at which melting had taken place. Also, because the final product is purer than the initial product later tests could show a higher melt temperature than that at original fusion.

Finally we accept that, in principle, at very high temperatures heat loss in a furnace can occur and that for a small increase in temperature above 1600 degrees centigrade there is a significantly greater relative increase in the power required to heat the furnace. However there was no evidence before us about the temperatures at which the furnaces used to produce the product imported by the Appellant had been operated and there was no evidence that they had been operated at temperatures less than 2000 degrees centigrade.

In the light of our conclusions about the expert evidence we now turn to consider the words of recital (18) to the anti-dumping regulation to see whether the Appellant has produced evidence to satisfy us that the material imported by it was not "manufactured by melting the raw material in electric arc furnaces at temperatures higher than 2000 degrees centigrade".

In our view the normal meaning of those words is that it is the temperature in the furnace which has to exceed 2,000 degrees centigrade and not the temperature at which the aluminium oxide melts. We are confirmed in this view by the evidence of the expert witnesses which was that, after the event, it is very difficult to determine the precise temperature at which specific material melted. There was no evidence before us that the material imported by the Appellant was melted in furnaces which reached temperatures below 2000 degrees centigrade.

However, in case we are wrong, and in case the words in the anti-dumping regulation refer to the melting of the material, we go on to conclude that the relevant melting point would be the (higher) point at which the aluminium oxide melts and not the (lower) point of the onset of melting of the bauxite with its impurities. With those views in mind we turn to consider the expert evidence so far as it relates to the samples of the material said to be supplied by the Appellant.

Both expert witnesses agreed that the tests by Sheffield (which determined chemical compositions of 95.6% and 97.1% by mass of aluminium oxide) indicated the presence of impurities all of which would reduce the melt point. Dr White calculated that, as a result of the presence of the impurities stated in those two Sheffield reports, a reduction in the melting point of the aluminium oxide by 21 and 36 degrees centigrade respectively could be reached. We accept his opinion that, if one took the melting point of pure aluminium oxide as 2050 degrees then the estimated melting points of the two Sheffield samples which gave data relating to the amount of the impurities would be 2029 and 2040 degrees centigrade respectively. This evidence does not convince us that the material imported by the Appellant melted at temperatures below 2000 degrees centigrade.

We consider the results of the tests carried out by Mr Burton and by Seteram together. Professor Derby's views on the Seteram analyses were that each sample was probably a single coarse grain which was a very small sample. If the raw material had been pure there would have been a single temperature but, where two temperatures were stated, there would have been impurities. The first temperature would have been taken at the onset of melting when the solid was beginning to melt and the second when the material moved into a liquid state. He agreed that most people would say that the material had been processed "at about 2000 degrees centigrade" although fusion could possibly have been below 2000 degrees centigrade. It was not possible to determine the precise temperature. He also agreed that ideally there should have been a lot more samples to give confidence and the results were an incomplete data set.

On the Seteram results Dr White also had some concerns about the representative nature of the samples used. He had been told that the samples were sub-samples of the four samples taken by Mrs Josey which had been analysed by Mr Burton. Mr Burton's analysis indicated that the four samples were almost identical and so it was surprising that the Seteram results showed a wide variation in the same samples. His conclusions were: that, in respect of each of the four samples, the onset of the melting of the entire sample was in the range of 1580.1 to 2023.6 degrees centigrade; that the onset of the melting of the aluminium oxide within the samples was in the range 1928.9 to 2023.6 degrees centigrade; and that lower melting point components had been present in three of the samples. Taking the results of Mr Burton and reducing the melt temperatures by the presence of the impurities he calculated that the melting points of the aluminium oxide were 2019, 2022, 2022 and 2021 degrees centigrade respectively.

We share the concern of the expert witnesses about the representative nature of the samples and further comment that no evidence was produced to us to prove that the four Setaram samples were sub-samples of those taken by Mrs Josey. There is a further difficulty with the Seteram results which arises from the fact that they indicate that melting takes place at a number of temperatures of which they measure three, namely the onset of melting, the "second step onset point", and the peak. We accept the evidence of Dr White that the "second step onset points" were the onset of the melting of the aluminium oxide. If these were the temperatures of the onset of melting then the higher "peak" temperatures would have continued or completed the melting. All the higher "peak" temperatures exceed 2000 degrees centigrade. We recall the expert evidence of Professor Derby that it is not certain that the precise manufacturing route from raw material to finished product (including the melt temperature) can be determined from tests conducted solely on the finished product. In our view the Setaram results do not lead to the conclusion that the aluminium oxide the subject of the tests was melted at temperatures less than 2000 degrees centigrade.

We conclude that, if the words of recital (18) are crucial to the classification of the product (and we do not think that they are) then, in our view, the regulation requires the temperature of the furnace to exceed 2000 degrees centigrade and there was no evidence before us that the material imported by the Appellant was manufactured by melting the raw material in furnaces with temperatures lower than 2000 degrees centigrade. However, if the words of the regulation refer to the melt temperature of the aluminium oxide, we conclude that pure aluminium oxide melts at temperatures higher than 2000 degrees centigrade but that the presence of impurities in the raw material would reduce the melt temperature. Here we are of the view that the relevant melt temperature is that of the aluminium oxide and not that of the onset of melting of the impurities. The Appellant's best evidence is that of the Seteram tests which give temperatures for the melting of the aluminium oxide of 2006.8, 2023.6, 1997.7 and 1928.9 degrees centigrade respectively. However, these were the temperatures of the onset of the melting of the aluminium oxide, not the highest temperatures used. The peak temperatures were 2034.7, 2046.4, 2014.2 and 2002.0 respectively.

We have reservations about the validity of the Setaram sampling and also about the wide variations in results where the chemical analysis of Mr Burton in respect of the same samples indicated much less variation. We therefore find as a fact that the Setaram samples were too small, and of uncertain and unproven provenance, to be representative or acceptable as evidence. In a product such as this there will be lumps and concentrations in one place and reduced amounts of impurities elsewhere. On the evidence before us we are not therefore satisfied that, in respect of the material imported by the Appellant, the aluminium oxide was melted at temperatures below 2000 degrees centigrade.

On the second question, therefore, we conclude that the Appellant has not produced evidence to satisfy the Tribunal that the material imported by it was not "manufactured by melting the raw material in electric-arc furnaces at temperatures higher than 2,000 degrees centigrade".

Question (c)- Is a fused product which contains less than 94% aluminium oxide artificial corundum?

The third question arising out of the tariff classification issue is whether fused aluminium oxide consisting of less than 94% aluminium oxide is or is not artificial corundum. The Appellant argued that the effect of recital (17) of the anti-dumping regulation was that fused aluminium oxide, if it consisted of less than 94% of aluminium oxide, was not artificial corundum. The Respondents argued that the last paragraph of recital (18) indicated that aluminium oxide containing less that 94% of aluminium oxide was lower quality artificial corundum but was still artificial corundum.

In considering the arguments of the parties we start with the tariff classification which does not distinguish between the levels of aluminium oxide in brown artificial corundum. It only distinguishes white artificial corundum as having more than 97.5% aluminium oxide content by weight. All other artificial corundum is classified under CN code 2818.10.90. The HSENs make no distinction between the aluminium oxide content of artificial corundum. Thus for customs duty we conclude that fused aluminium oxide, even if it consists of less than 94% of aluminium oxide, is artificial corundum. .

For the purposes of anti-dumping duty the same principle applies because the charging provision in Article 1 of the anti-dumping regulation refers back to the tariff classification and recitals (17) and (18) cannot amend the tariff classification. In any event they do not support the argument of the Appellant. Recital (17) says that artificial corundum is "principally" produced in two basic varieties, brown and white, but recital (18) refers to lower quality artificial corundum as well. There is no indication in recital (17) that only brown and white artificial corundum, as there defined, constitute artificial corundum; indeed the opposite is the case because the use of the word "principally" implies that there are other varieties and in our view lower quality artificial corundum is one such other variety.

We conclude that fused aluminium oxide, if it consists of less than 94% of aluminium oxide, is artificial corundum.

Conclusion on the tariff classification issue

Our answers to the three questions raised by the tariff classification issue are:

(a) that the reference to a temperature of "higher than 2000 degrees centigrade" in recital (18) of the anti-dumping regulation is not a crucial part of the description of artificial corundum; that means that we do not have to answer the second question but, as arguments and lengthy expert evidence were put to us, we express our views which are

(b) that on the evidence before us we are not satisfied that the material imported by the Appellant was melted at temperatures lower than 2000 degrees centigrade; and