Avecia, Inc. v. United States (469 F. Supp. 2d 1269, 2006)

OPINION

MUSGRAVE, Judge.

Following a bench trial on whether imported ink-jet ink concentrates consisting of chromophores in deionized water should be classified for customs duty purposes under the Harmonized Tariff Schedule of the United States (“HTSUS”) as “synthetic organic coloring material,” specifically “dyes,” of heading 3204, or “printing inks” of heading 3215, the Court concludes that the products are correctly classifiable under heading 3215.

Procedural Background

This proceeding resolves a lengthy dispute between the plaintiff Avecia, Inc. and the defendant United States involving the defendant’s agency, U.S. Customs and Border Protection or its predecessor(s) (“Customs”), over the proper tariff classification of Avecia’s ink-jet ink liquids. As with other ink-jet ink classifications that Avecia has been protesting, the matter of this dispute was imported from Avecia’s related manufacturing facility at Grange-mouth, Scotland in 2003. See Pretrial Order, Schedule C (“POSC”) at ¶ 15.

Previously, in 1998, Avecia submitted to Customs protest number 1101-98-100179 on the proper classification of “Pro-Jet Fast Yellow 2” (“FY2”). Avecia argued that FY2 should be classified as a printing ink of heading 3215, HTSUS, and not as a dye (i.e., synthetic organic coloring material) of heading 3204, HTSUS. See id. at ¶ 5 (April 28, 2006). In the spring of 2000, Customs approved the protest and memorialized the approval in [¶] 962365, reclassifying Pro-Jet Fast Yellow 2 under heading 3215. See id. at ¶ 6; Pl.’s Ex. 22 at 14. After that approval, Avecia sought to apply Customs’s analysis to other ink-jet inks by filing protests on entries that Customs had classified as dyes. Customs denied the protests, and Avecia timely contested the protest denials in two actions before this court, USCIT Nos. 03-00001 and 03-00197, filed in December 2002 and April 2003, respectively. Id. at ¶ 9; Compl. & Answer I ¶ 24; Compl. & Answer II ¶ 24.

During that period, on January 2, 2003, Customs published and invited comment on its “Proposed Revocation of Ruling Letters and Treatment Relating to Tariff Classification of an Ink[-] Jet Color Preparation.” POSC at ¶ 10. The literal terms of 19 U.S.C. § 1625(c) meant that the period for submitting comments thereon was open until February 3, 2003, after which Customs was required to “take action” on any “decision to modify or revoke” the previous ruling favoring Avecia within 30 days after the close of the comment period. 19 U.S.C. § 1625(c). Cf. Harmon Dep. Tr. 80:15-82:8. Customs did not publish its general notice of revocation and ruling number [¶] 966063 relating to the tariff classification of FY2 in the Customs Bulletin until June 25, 2003. See POSC at ¶ 11. The notice and ruling concluded that FY2 is classified under subheading 3204.14.30. Id.

In July 2003, Avecia, through counsel, sent a letter to Customs asserting that Customs’s notice of revocation, which had been sent to Avecia’s counsel on June 13, 2003, was “ineffective.” Id. at ¶ 12. Some five months later, on February 25, 2004, the parties resolved Avecia’s first judicial challenges, USCIT Nos. 03-00001 and 03-00197, by entering into two stipulated judgments on agreed statements of facts. See id. at ¶ 14 (Customs noting its position that the stipulations were solely procedural and not substantive).

Avecia then sought to contest other year 2003 entries including the ink-jet inks at issue. Customs had classified the liquids under heading 3204 “synthetic organic coloring matter ... dyes” under the HTSUS. Avecia tendered the duties claimed by Customs on the subject matter at the time of entry, and Customs liquidated the entries from November 14, 2003 through August 20, 2004, as entered. Id. at ¶ 18. Avecia timely protested the classifications, again arguing that the merchandise should be classified as “printing inks” under heading 3215. Id. at ¶ 19. During June and October 2004, Customs denied each of Avecia’s protests as to the subject products. Id. at ¶20. Plaintiff then filed a summons on September 27, 2004 and initiated case No. 04-00489, and a summons on March 1, 2005 covering the other protests involved in this action. Id. at ¶ 21.

The protests cover entries of “Pro-Jet Cyan 1 Stage RO Feed,” “Pro-Jet Cyan 1 Special Liquid Feed,” “Pro-Jet Black 287 Liquid Feed,” “Pro-Jet Yellow 1 Liquid Feed,” “Pro-Jet Cyan 854 Liquid Feed,” “Pro-Jet Black 661 Liquid Feed,” and “Pro-Jet Black HS Stage.” Id. at ¶ 15. With the exception of Pro-Jet Black HS Stage, Customs considered the liquids “direct dyes” and classified them either under heading 3204.14.30 (“other ... products described in additional U.S. note 3 to section VI”), which bore a duty rate of 7.4% ad valorem, or under heading 3204.14.50 (“other ... other”), which bore a duty rate of 7.8% ad valorem. Customs considered Pro-Jet Black HS Stage, covered by entry number 916-1076747-6, a “reactive dye” and classified the liquid under heading 3204.16.30 (“other ... products described in additional U.S. note 3 to section VI”), which bore a duty rate of 7.4% ad valorem. See id. at ¶ 16; Ans. at ¶ 32. The parties do not dispute the value of the merchandise declared on the subject entries as set forth in the relevant protests. POSC at ¶ 17.

The parties severed for stipulation certain entries having dates of entry prior to August 24, 2003, when the FY2 ruling was belatedly revoked. See Order of 4/28/06 (granting motion to sever). The parties agreed that products entered before August 24, 2003 are classifiable under heading 3215.19.0060 or under heading 3215.11.0060 at a duty rate of 1.8% ad valorem pursuant to Customs’s April 2000 ruling approving Avecia’s protest as to the classification of its FY2 product. Def.’s Supp. Resp. to Interrogs. ¶ 10. Ten 2003 entries continue to be contested. See Pl.’s Exs. 21 & 22; see also Compl. & Answer I ¶ 36; Compl. & Answer II ¶ 37.

Jurisdiction

28 U.S.C. § 1581(a) confers exclusive jurisdiction in this Court on the denial of a customs classification protest.

Findings

Introduction

Avecia traces its direct roots back to Imperial Chemical Industries PLC, a U.K. dye manufacturer formed in 1926. See R 259:8-11, 260:21-22. Its witnesses therefore testified to familiarity with both dye and ink-jet ink development and manufacture. The advent of ink-jet ink printing technology prompted Avecia to concentrate on ink-jet manufacture, and the relative newness of this technology appears to have precipitated this dispute over the classification of its imports.

To discuss its products at trial, Avecia introduced the testimony of Mr. Craig Johnston, Avecia’s process technology and toll manufacturing manager, Dr. Ilesh Bidd, Avecia’s research and technology director and corporate representative, Prof. Peter Gregory, whose working life was devoted to the research department of Avecia and is now retired, and Dr. Harold Freeman, a professor of dye chemistry at North Carolina State University and Ave-cia’s expert witness. Their collective testimony, together with Avecia’s exhibits, addressed distinctions between Avecia’s products, dyes, inks, ink-jet inks, and Ave-cia’s role in the evolution of the latter. Dr. Bidd also demonstrated that the products are capable of printing in their condition as imported, using an Epson C62 printer and standard Epson photo paper, and he opined that the print “looks great.” The Court determined the success of the demonstration R 326:15-343:9, 343:24-25. See Pl.’s Ex. 95. The government asked for and received a sample of the product used in the demonstration, and, after testing same, confirmed that it was identical to the products here in dispute and did not contain additives. R 828:20-829:10.

The government introduced the testimony of Harvey Kuperstein, who is the National Import Specialist responsible for, inter alia, the classification of imported dyes under heading 3204. R 766:6-767:24, 790:23-791:2, 794:24-795:3. Although Mr. Kuperstein is not responsible for the classification of inks under heading 3215, he stated he is familiar with heading 3215 classification and represented that he has “a layman’s ability to fathom this kind of technology.” R 799:3-7.

The parties also introduced various exhibits at times throughout the trial. In addition, the parties introduced the deposition testimony of Deborah Walsh, the National Import Specialist responsible for heading 3215 during the relevant period of importation, Myles Harmon, the director of the Commercial and Trade Facilitation Division of Customs and Customs’s corporate representative as to procedural issues related to classification, Mr. Kuperstein, and Dr. Bidd, all of which the Court admitted. R 763:17-764:8, 764:18-20.

Pigment — and Dye-Based Inks; Ink-Jet Printing; Description of the Imports

All inks contain a colorant. The two organic types of colorants are pigments and dyes. R 487:3-488:22. Pigments are insoluble; in this matter, dyes refer to the dyes that are soluble in an organic solvent such as water. R 43:5-18, 276:13-25, 452:2-24, 487:3-488:22, 695:7-17.

Traditional printing inks consist of a dispersion of a pigment, such as lampblack, in a solution containing a binder which acts to adhere the pigment component to a substrate normally comprised of a hydrophilic cellulose-based material (i.e. paper). See generally R.229:9-18, 499:3-5, 598:19-24, 277:2-23, 299:23-302-7, 479:21-480:7, 695:18-696:2. Avecia’s expert witness, Dr. Freeman, explained that pigments have no affinity for a substrate and cannot penetrate the cellulosic surface of paper; accordingly, they require a binder or glue to affix them to the substrate in order to stay in place. He further explained that dyes — particularly dyes designed with ink-jet printing in mind — have built into them all of the salient properties essential for fixation to the substrate. R 487:3^88:22.

Ink-jet printing differs from traditional printing in that the ink is squirted or dropped directly onto the substrate without the involvement of, e.g., photoconduc-tors or ribbons. R 271:9-272:5, R 496:6-21, 689:13-690:24. This involves precise and controlled ejection of a liquid colorant through an ink-jet printer’s fíne nozzles onto a substrate; consequently, “inks formulated for jet printing must be very fluid, stable, and free of any particles that could cause clogging of the jet nozzles, and be capable of depositing and adhering to a substrate with a minimum of character fogging.” KirkOthmer Concise Encyclopedia of Chemical Technology 1138 (4th ed. 1999) (“Kirk-Othmer ”).

For ink-jet printing, soluble dyes offered certain advantages over pigments. Pigments, being insolvent particulates, have a propensity towards kogation, resulting in clogged ink-jet print heads. See R 697:9-24. Water was also determined to be an ideal solvent for ink-jet ink development. See R 277:24-20, 697:25-698-13. The development of aqueous dye-based ink therefore appeared obvious to original equipment manufacturers (“OEMs”) contemplating the development of ink-jet printing technologies. To meet their needs, Avecia began to research, develop, and manufacture ink-jet inks by purifying and processing water-soluble colorant molecules, or chromophores, including colorants that were commercially available at the inception of modern ink-jet printing as well as new colorants designed by or for Avecia. The qualities of each of Avecia’s ink-jet ink generations to date is a function of both the structure of the molecule and the manufacture of the ink. After investments in the late 1980s and early 1990s on ink-jet ink development, in the late 1990s Avecia sold its textile dyes business, concentrated its remaining Grangemouth facilities for ink-jet ink manufacture, and invested a substantial amount of money, especially since 1997, to anticipate growth in the ink-jet ink market. R 47:20-48:8, 51:7-53:21, 67:6-16. See Pl.’s Ex. 92; see also Pl.’s Ex. 40 at 34.

As mentioned, each of the Avecia products at issue consists of water-soluble coloring matter in water. R 233:4-14, 318:17-320:2, 608:20-24, 719:11-721:10. More precisely, they are chromophores in a solution of approximately 90 — 95% deionized water which is maintained at a precise pH equilibrium. See R 157:2-167:14. The imported Cyan 1 also has an added biocide to prevent bacterial growth. R 127:17-129:8. The products are “novel” in the sense that they consist of new or existing molecules that have been purified to an extreme state through a series of elaborate steps at the Grangemouth facility undertaken to enhance their “printing” characteristics. Some, such as Pro-Jet Fast Black, are also constructed using novel intermediates to produce the desired end-result. E.g., R 713:4-716:20; Bidd Dep. Tr. at 13. Each of the imported products is designed and/or purified to achieve a particular target of color strength, purity, solubility, environmental safety, and substrate affinity including minimum wetfastness, light fastness, and ozone fastness characteristics. See R 483:21-494:21; see also R 302:13-306:17, 507:9-508:4, 705:16-706:12, 706:15-707:19; Pl.’s Ex. 40 at 7-14. Each of the imported products is also designed to penetrate a substrate and “bind” through the evaporation of the “ink vehicle,” which in this case is water. R 352:5-21. Consequently, in contrast to pigment-based inks, the imported products do not have an added binder.

The two most crucial considerations for developing an ink-jet ink are its color and its “fastness” including substrate affinity. R 134:11-135:10, 239:15-240:15. The “active” part of the chromophore imparts col- or, and those properties are determined by the molecule’s propensity to absorb or reflect specific light waves of the visible (human) spectrum. See R 155:4; Bidd Dep. Tr., Def.’s Ex. A, at 73-74. Color may be precisely determined using so-called “LAB” values. See PL’s Ex. 91. These describe color coordinates on the Commission Internationale de l’Eclairage (CIE) L*a*b grid, a “full” color map on horizontal and vertical axes. Luminance takes values between 0 and 100, and the other coordinates take values between — 100 and 100. The a and b coordinates measure positions on the green/red and blue/yellow axes respectively. Lab values are critical specifications and are used as standard color nomenclature in the manufacture of colorants to describe the intensity of each color and exact shade of each color.fie., how green/red or blue/yellow the color appears). R 134:12-21, 291:9-20, 288:7-17, 385:16-22. Further, it is possible to (re)create about 80 to 85% of the 16.7 million colors on the color map by mixing yellow, magenta, cyan, and black into various combinations. R 709:3-8. Ideals of these four colors are therefore the targets of ink-jet ink manufacture. R 290:3-24. Cf. PL’s Ex. 91.

In addition to color, the chromophores have undergone several generations of development seeking to enhance their fastness capabilities, ie., substrate affinity and resistance to degradation from light and ozone. See, e.g., R 302:13-306:17, 456:20-457:4, 507:9-508:4, 518:19-519:15, 520:10-521:6, 700:10-17, 703:2-705:3, 704:19-705:3, 705:16-706:12, 706:15-707:19, 709:9-710:11; PL’s Ex. 40 at 7-14. Cf. 456:20-459:24, 533:12-16, 564:2-10, 566:16-574:13, 580:11-25; PL’s Exs. 66-72. These key chemical traits, which have been built into the molecules at Grangemouth, are reflected in the colors of the colorants as specified by the LAB coordinates of the inks on paper. R 133:23-134:10, 239:15-240:15. In other words, the color and color strength that the molecules display on paper is a direct reflection of the design effort that has gone into them. R 510:7-515:21; PL’s Ex. 55.

It was undisputed that the products at issue are designed for use in ink-jet printers. Upon importation into the United States, however, the liquids are not sold or used as ink-jet printing inks in their condition as imported but must undergo a finishing process involving reverse osmosis to remove certain inorganic molecules like chlorides and calcium, sulfate ions, and cations such as- iron. See R 321:11-12. The processing requires the addition of new deionized water and incidental displacement of some of the imported water. The liquids are also further processed with additives, as required, for Avecia’s OEM customers. R 169:7-173:6, 198:22-199:10.

Differences Between the Manufacture of Ink-Jet Printing Inks and the Manufacture of Textile Dyes

The testimony of Avecia’s witnesses described various differences between the manufacture of ink-jet inks and the manufacture of traditional or textile dyes, including design and customer input, raw materials, manufacturing processes, quality control, staffing differences, and management of change in addition to other dissimilarities. See Pl.’s Ex. 40 at 37-39; Pl.’s Ex. 93.

Design and Customer Input. Ink-jet inks are designed and manufactured as ink-jet inks to be printed on a substrate. R 483:21-21; 713:19-714:5. See Kuperstein Deposition Transcript (“Dep.Tr.”) 48:1-49:16; see also R 820:10-823:10; Pl.’s Ex. 93. By contrast, commercial dyes are designed to dye a particular substrate. R 528:21-529:15. See PL’s Ex. 93. In contrast to dye manufacture, the ink-jet ink business requires close collaboration with OEMs, from both a scientific and technological perspective, in order to meet specific targets and minimum standards. R 265:25-267:10. See PL’s Ex. 93. In order to meet the specific targets of OEMs, Ave-cia must build specific properties into the colorant molecules within the ink system. R 268:6-269:25. Dyes for textiles, by contrast, do not have the same strict color requirements as ink-jet inks: the textile dyes business involves very little interaction with the customers. R 267:11-269:25, 386:15-387:24. See PL’s Ex. 93. Aveeia’s textile dye experience had been that specifications or targets were not provided by customers but were decided by the company’s internal technical marketing group. Id. In contrast to ink-jet printer OEMs, textile dye customers tend to buy their requirements off-the-shelf and subject to commercial availability. See id.

Raw Materials. Significantly higher standards are required of the raw materials used in the manufacture of ink-jet ink than of those used in dye manufacture. R 59:4-12. See PL’s Ex. 40 at 38; PL’s Ex. 93. Raw materials used for ink-jet ink manufacture must in some instances be pharmaceutical grade material, whereas textile dye raw materials tend to have been commoditized. R 83:5-23. See PL’s Ex. 40 at 38; PL’s Ex. 93. Ink-jet ink raw materials are often more expensive and require more stringent storage controls than raw materials for bulk textile-dye manufacture. R 62:8-63:12, 92:14:93:2, 521:14-522:11. See PL’s Ex. 40 at 39. The choice of raw material effects the quality of the overall ink in terms of its environmental fastness and color. Ink-jet ink manufacture therefore requires raw materials that are of a much higher quality and with lower impurities than the raw materials used in dye making. R 101:16-103:15. See PL’s Ex. 93.

Manufacturing Processes. Ink-jet ink manufacture requires more steps, temperature control, pH control, and in-process testing, than are required for textile-dye manufacture. R 55:20-59:3, 59:13-23, 71:22-72:19. See PL’s Ex. 40 at 37-38; PL’s Ex. 93. In ink-jet ink manufacture, there are specifications for the temperatures, reaction time and impurities that must be met at set stages during the processing, and there is also intensive quality control and batch review at the end stage. R 59:24-61:16. See PL’s Ex. 40 at 37. The in-process testing is designed to optimize the processes for purposes of ensuring purity and avoiding kogation. R 96:13-97:7. Avoiding kogation is key to the design and manufacture of ink-jet inks, as kogation would potentially mean the recall of millions of cartridges for the impacted OEM customer. R 69:21-71:7. Batches that do not meet specifications sometimes require discard. R 61:17-62:7. See PL’s Ex. 40 at 37. In contrast, Ave-cia’s prior experience in textile dye manufacture had been that the concept of batch review was unnecessary; rather, there would be application testing at the end, and batches would be blended and shaded in order to eliminate any perturbations. R 61:17-62:7. See Pl.’s Ex. 40 at 37; Pl.’s Ex. 93.

Quality Control. In Avecia’s ink-jet ink manufacture, each batch is the final product, so blending is not performed in the way that it had been traditionally for textile dye manufacture, where standardizing tanks had been used to blend batches to ensure that they met target specifications. R 51:25-52:19, 79:19-81:18, 107:11-108:5. See PL’s Ex. 40 at 37; Pl.’s Ex. 93. In the manufacture of an ink-jet ink batch, if there are any deviations from normality at any point, a batch review is conducted which involves the quality assurance manager and the manufacturing personnel, who review all of the data, possibly gather additional data, and determine whether the material can go into the supply chain or needs to be quarantined and subjected to additional testing. R 103:16-104:8. See Pl.’s Ex. 40 at 3 8; PL’s Ex. 93. If ultimately it is determined that the material cannot be released into the supply chain, it is discarded and written off rather than reworked. R 105:14-25. See PL’s Ex. 93. With textile-dye manufacture, unless there has been a major abnormal operation, batches are capable of being reworked through blending to achieve specifications. They are not quarantined or written off. R 106:10-107:10. See PL’s Ex. 93.

Staffing Differences. For ink-jet ink manufacture, the manufacturing technicians are multi-skilled. In addition to responsibilities for process operations, they also maintain the plant and perform their own analyses of chemical purity and the chemical species present. R 90:4-23. See PL’s Ex. 40 at 39; PL’s Ex. 93. Avecia’s process technology team for ink-jet ink manufacture employs 36 individuals permanently at Grangemouth and 9 of them are Ph.D-qualified while the rest are degree-level qualified. R 90:24-91:8. See PL’s Ex. 93. The process-technology team exercises a higher level of supervision over the entire process of ink-jet ink manufacture than had been the case for dye manufacture and provides 24-hour support to ink-jet ink manufacture. R 92:3-13. See PL’s Ex. 40 at 39. When the facility had previously been manufacturing textile dyes, there had been about 10 permanent employees who oversaw a much higher average volume of products. R 91:9-23. See PL’s Ex. 40 at 39; PL’s Ex. 93.

Other Dissimilarities. Other dissimilarities between ink-jet ink and textile dye manufacture include the following. (1) There are vast differences between the volumes of product manufactured: on an equivalent-manufacturing-unit basis, Ave-cia’s production capacity was far greater for dyes than it is for ink-jet inks. R 52:4-24, 58:16-59:3, 68:20-69:9. See PL’s Ex. 40 at 37; PL’s Ex. 93.(2) In contrast to textile dye manufacture, any change in ink-jet ink manufacture processes, such as introducing new starter materials or changing a chemical process, requires more rigorous management of the change including substantial laboratory testing and possibly customer agreement in advance. R 77:2-23. See PL’s Ex. 40 at 38; PL’s Ex. 93. For example, changing a single raw material may take a year to process. Id. (3) Reconfiguring a unit line to produce a different color is a much more laborious undertaking for ink-jet ink manufacture than for textile dye manufacture. For example, in either instance the production unit must be cleaned out, but the clean-out regime for ink-jet ink manufacture is much more difficult and time consuming. R 67:20-68:17. See PL’s Ex. 40 at 37.(4) There are three laboratories at Grange-mouth for ink-jet ink, whereas there had been only one full and one partial laboratory at Grangemouth to support dye manufacture. R 100:8-14.(5) On a per-unit basis, Avecia’s manufacture of ink-jet inks is approximately ten times more expensive than had been the case for dye manufacture. R 93:3-15. See PL’s Ex. 40 at 39; Pl.’s Ex. 93.(6) In ink-jet printing, nearly all of the product is used on the substrate and none is wasted. By contrast, most of the dye of a textile dyeing operation remains in the dye bath as waste and only a small portion of the product actually dyes the substrate. R 272:6-275:11, 498:2-499:2. See PL’s Ex. 93.

Avecia’s Expert Testimony

Avecia’s expert, Dr. Harold Freeman, concluded that the subject products are properly designated as printing inks. R 479:14-20, 483:5-20, 608:10-610:2. The experiments performed by Dr. Freeman were intended to determine whether the Avecia liquid feed products — namely, Pro-Jet Cyan 1 Special Liquid Feed, Pro-Jet Cyan 854 Liquid Feed, Pro-Jet Yellow 1 Liquid Feed, Pro-Jet Black 287 Liquid Feed, Pro-Jet Black 661 Liquid Feed, and Pro-Jet Cyan 1 RO Liquid Feed — are ink-jet printing inks or direct dyes at the time of their manufacture in Grangemouth. See, e.g., R 479:14-20, 483:5-20. Dr. Freeman’s first experiment tested the properties of the products when used as ink-jet inks; his second experiment tested the properties of the products when used as textile dyes. Dr. Freeman’s first experiment demonstrated that the Avecia-Grangemouth liquid feed products can be readily printed onto paper using water as the sole non-colorant component in the ink. Dr. Freeman determined that the Avecia-L Grangemouth liquid feed products can be applied to paper without any further additives by using a standard Epson printer and also that they meet the end-use requirements of printing inks. R 483:21-24, 484:22-485:20.

To make those determinations, Dr. Freeman performed an experiment employing two sets of samples: (1) products generated in the standard manufacturing processes at Avecia-Grangemouth prior to being sent to Avecia’s outfit at New Castle, Delaware, as feed stock (liquid feeds) for commercialization, and (2) products obtained after representative further processing operations at New Castle. R 525:3-527:4. See PL’s Ex. 65. Dr. Freeman tested both the Avecia-Grangemouth sample and the Avecia-New Castle sample for each of the following products: Pro-Jet Cyan 1 Special Liquid Feed, Pro-Jet Cyan 854 Liquid Feed, Pro-Jet Yellow 1 Liquid Feed, Pro-Jet Black 287 Liquid Feed, and Pro-Jet Black 661 Liquid Feed. Id. There was no New Castle sample available for Pro-Jet Cyan 1 RO Liquid Feed, so Dr. Freeman tested only the Grange-mouth sample. Id. Dr. Freeman further testified that this fact should have no bearing on the validity of his conclusions. Id.

Samples of liquid feeds (exGrange-mouth, Scotland) and samples of final products (ex-New Castle, Delaware) were shipped to Manchester, England for Dr. Freeman’s experiments. R 525:3-527:4. He evaluated the two sets of samples based on the following qualities: (a) print-ability, (b) water fastness (a/k/a wetfastness), (c) ozone fastness, (d) light fastness, and (e) color assessment. See R 456:20-457:4, 564:2-10. His experiment for testing the properties of the subject products when used as inks corresponded to the same tests that Avecia would use in evaluating its products for printability and fastness. R 457:5-25. Dr. Freeman looked for standard test methods published by organizations such as the American Association of Textile Chemists and Colorists and the American Society for Testing and Materials but could find no such tests. R 458:13-459:4. He testified that, to his knowledge, there are no standardized methods agreed upon by all of the OEMs for the assessment of inks and their properties. R 531:17-3. Accordingly, he relied upon his experience, the feedback from Avecia as to what kinds of tests have been designated in the past by the OEMs with which they have been affiliated, and similar tests used for the evaluation of fastness properties of dyes on textiles, in order to determine appropriate test methodology. R 457:5-459:24.

Dr. Freeman first ran the sample products through an Epson C 62 printer and printed them onto “HP” plain paper and Epson photographic paper. R 533:20-534:21. He printed multiple pages of line diagrams that demonstrated the printed ink at three depths of color: 100%, 70%, and 50%. Id. The evaluations demonstrated that the subject products could, indeed, be printed readily from a standard ink-jet printer, and that there were no significant visible differences between the image created by the Grangemouth sample and the New Castle sample for each product. R 535:10-536:3, 543:5-10, 561:21-563:17; Pl.’s Exs. 58-60.

The colors of the printed inks were measured at various points. Dr. Freeman testified that he used standard equipment to measure the LAB values, the percent reflected optical density, the chroma, and the hue angle of the printed samples. R 564:2-566:13. These measurements were made (1) after printing each substrate, (2) after light fastness exposures, and (3) after ozone exposures. R 566:16-567:24; Pl.’s Exs. 66-72. He testified that: (a) color values recorded after printing and drying showed insignificant differences between the pairs of yellow, cyan, and black inks, and (b), similarly, insignificant differences in color values were observed following light and ozone exposures of commercial ink-jet paper printed with the same pairs of inks. R 567:25-574:13. See PL’s Exs. 66-72.

Dr. Freeman tested water fastness by having one-half millimeter of water applied to the samples of Avecia’s products printed on plain HP paper with the substrate placed at an angle of 45 degrees. R 574:22-575:11. These tests demonstrated that the sample from Grangemouth and the sample from New Castle for each product run to the same degree and, therefore, have similar water fastness. R 575:16-579:24, 582:24-584:3. See PL’s Exs. 61-62, 73.

Dr. Freeman tested light fastness, using an Atlas weatherometer to expose the printed samples to the equivalent of 225 kilojoules of energy over 50 hours, which is two and a half times the exposure used for typical evaluations of dyes on textiles. R 531:17-533:11. The samples exposed had been printed on plain HP paper and on photographic paper. R 580:11-25.

Dr. Freeman tested ozone fastness using a standard machine, the Hamden 903 ozone test unit, and exposing the printed samples to 1 part per million ozone, which is a typical exposure level. R 533:12-16. According to him, these tests demonstrated that the light fastness properties and the ozone fastness properties were not significantly different between the samples from Grangemouth and the samples from New Castle, with the exception of Pro-Jet Black 287 which showed a slight difference in the fading. R 580:14-581:19, 582:24-584:3; PL’s Exs. 63, 73. Overall, he opined that the further processing in New Castle did not affect the fastness properties of the yellow, cyan, and black inks other than the Black 287, and his test results may be summarized as follows:

a. The two data sets’ results for light, ozone, and water (wet) fastness testing of printed papers reflect good consistency.

b. Results from the 2 yellow inks and 5 cyan inks indicate that comparable light, ozone and water fastness values were obtained.

c. In the lone case of the Black 287 / Black 2 pairing, better light and ozone fastness was obtained from the ink derived from the New Castle sample. However, water fastness was the same for the two products.

d. Pro-Jet Black 661 gave comparable fastness properties before and after New Castle processing.

See id.; see also Pl.’s Exs. 61-68, 73. Dr. Freeman reiterated that his experiments clearly demonstrated very good consistency of the products from Grangemouth to New Castle, and that this consistency in fastness properties results from the fact that the performance properties are built into the molecule structures. R 584:10-585:5.

Dr. Freeman then testified that in order to designate the subject products as dyes, they should be able to dye something. First, he attempted to determine the most appropriate substrate to use for the experiment. R 457:5-458:12. According to Dr. Freeman, one can determine whether a dye is suitable for wood or paper or leather or plastic or nylon or wool or acrylic, etc., based upon the chemical structure of the dye. R 528:21-529:15. For example, acid dyes are used to dye substrates such as nylon, wool and silk because they are ionic and those substrates are cationic. R 489:18-490:3. See Pl.’s Ex. 49. Dispersed dyes are hydrophobic and are used to dye hydrophobic substrates such as polyester. R 490:4-16. See Pl.’s Ex. 49. Based upon the molecular structure of the AveciaGrangemouth samples, Dr. Freeman testified that in his opinion, they would only be appropriately designated as direct dyes. R 528:21-529:15. Dr. Freeman testified that, by definition, direct dyes have an affinity for eellulosic fibers; when they are used commercially, 99% of the time they are used to dye cotton. R 486:5-8. See Pl.’s Ex. 49. He further testified that he was not aware of any direct dyes that would not be suitable for dyeing cotton. R 597:19-23. Accordingly, he constructed an experiment to determine whether the Ave-cia-Grangemouth liquid feed products possessed sufficient direct dye properties for dying cotton. R 483:21-24, 485:21-486:12.

For this experiment, Dr. Freeman compared samples of the Avecia-Grangemouth products to three commonly-available examples of direct dyes that he determined would be good models to use: Direct Black 22, Direct Yellow 44, and Direct Blue 218. R 527:20-20. He stated that he used standard methods for applying dyes recommended by dye manufacturers and used in North Carolina State University laboratories, the industry, and dye houses. R 459:25-460:13. He further testified that all of the methods he used in this experiment were based on the methods commonly used and recommended by dye manufacturers, as well as his 24 years of experience in the industry. Id.

Dr. Freeman prepared dye baths containing inks from Grangemouth and dyes that he believed to be of the same general type in terms of their structural makeup. R 585:13-24. He carried out the dye experiments under the traditional conditions associated with direct dye: he used a salt to cover up the negative charges on the textile fibers so that the dyes could enter the fiber if they are behaving as direct dyes, and he conducted the dyeing at 90 degrees to ensure the swelling of the cellulose so that the dyes would have access to the amorphous regions of the fiber. R 587:23-589:5. He also used two dyeing machines that are widely used in the industry: a Gibbs dyeing machine that allows dyeing at atmospheric pressure, and a pressurized dyeing machine to accelerate the penetration of the dye in the bath, which allowed him to shorten the dye cycle. R 589:6-17. The results of his dye studies demonstrated that the cyan and the black Avecia products have no significant affinity for cotton at all. R 589:18-592:25; Pl.’s Exs. 64, 74. With the possible exception of the Yellow 1, Liquid Feed, which is one of the inks derived from off-the shelf dyes, the shade depths typical of direct dyes were not obtained using the Avecia inks. Id. Dr. Freeman testified that the conclusion was especially evident when the black inks were employed in the dyeing process, because the pastel gray shades observed on cotton that was dyed with the Avecia products contrasted starkly with the deep black shade produced from the dye Direct Black 22. See Pl.’s Ex. 64.

The results Dr. Freeman reflected in a table summarizing the absorbance values and corresponding percent exhaustion values for the various inks and direct dyes that were used. R 606:5-608:7; Pl.’s Ex. 74. The values demonstrate Avecia’s cyan and black products had significantly lower percentage of exhaustion dye uptake than the traditional dyes, and that Avecia’s yellow product, although not as low, still had a lower percentage than the corresponding traditional yellow direct dye. Id. Dr. Freeman testified that this result was to be expected because Avecia’s products were engineered with paper and ink-jet printing in mind. R 589:18-590:2. Accordingly, Dr. Freeman concluded that Avecia’s products are not direct dyes suitable for dyeing cotton. R 485:21-486:12. Mr. Kuperstein agreed with Dr. Freeman’s conclusion that the subject products are not direct dyes. R 815:7-11.

Dr. Freeman testified that even though the chemical structure of Avecia’s products would only admit possible classification of the products as direct dyes, the products are not properly classified as direct dyes because they were not designed to be direct dyes, e.g., they have no affinity for cotton, which is a characteristic of direct dyes. R 483:21-24, 485:21-486:12, 528:3-16, 528:21-529:19. He opined that the Avecia ink-jet inks have systematically evolved away from textile dye utility, to the extent that the resultant products — the products at issue here — have poor affinity for textiles but good properties as inks. R 483:21-24, 486:13-20.

The results of Dr. Freeman’s first experiment demonstrated that the products at issue can be applied to paper using water as the sole non-colorant constituent (i.e., without any additives) by using a standard printer. The resultant printouts appear sharp and crisp to the Court’s eye, and the printouts produce both text and block type prints apparently suitable for the reproduction of photographs. It was submitted that the printouts possess the degree of water fastness, light fastness, and ozone fastness that one would expect and that Avecia’s OEM customers demand. R 484:22^85:20.

The results of Dr. Freeman’s second experiment demonstrated that the products at issue are not direct dyes suitable for dyeing cotton. R 485:21-486:12. These test results formed the basis for Dr. Freeman’s expert opinion that products at issue in this case — -Pro-Jet Cyan 1 Special Liquid Feed, Pro-Jet Cyan 854 Liquid Feed, Pro-Jet Yellow 1 Liquid Feed, Pro-Jet Black 287 Liquid Feed, Pro-Jet Black 661 Liquid Feed, and Pro-Jet Cyan 1 RO Liquid Feed — are all “inks” following their manufacture at Grangemouth and prior to further processing, and that the further processing in New Castle, Delaware is unnecessary to their being designated as printing inks. R 483:10-20.

The Information at Customs’s Disposal for Classifying The Imported Merchandise

Avecia avers that Customs had all the information it required to classify the subject products as printing inks under heading 3215. Ms. Deborah Walsh, the National Import Specialist responsible for classification of merchandise under heading 3215 during the relevant time period, asserted at her deposition that if a product performs as an ink when printed, it is a printing ink under heading 3215, and that Customs does not have a list of required components that allow a product to be classified as a printing ink under 3215. Walsh Dep. Tr. at 126:10-127:9, 128:21-129:11,129:21-130:15. Ms. Walsh also testified at her deposition that Customs does not require any particular additive in order to classify a product as a printing ink under heading 3215, nor does Customs require a product to work in more than one piece of equipment to be classified under heading 3215. Walsh Dep. Tr. 134:16-135:7, 210:3-211:15, 212:7-213:6, 214:17-216:4, 216:13-21.

According to Mr. Kuperstein, Customs had samples of some of Avecia’s products at importation but did not conduct any tests comparing the binding function of Avecia’s products to the binding qualities associated with dyestuffs, nor did Customs conduct any tests comparing the subject products as manufactured at Grangemouth with the same products after leaving Ave-cia’s facility in New Castle. Kuperstein Dep. Tr. 55:3-16, 56:18-57:9.

Discussion

I

As an initial matter, the government contests the Court’s jurisdiction over entry numbers 916-1076548-8, 916-1076920-9 and 916-1076747-6. The first two were entered at the Port of Newark, New Jersey and the third was entered at the Port of Baltimore, Maryland. The protest encompassing these and other entries, Protest No. 1101-04-100239, was filed at the Port of Philadelphia, Pennsylvania. The government argues that because 19 C.F.R. § 174.12(d) requires protests to be filed with the port director whose decision on entry is being protested, and because the three entries were protested at the wrong port, their protest was not “in accordance with regulations prescribed by the Secretary.” See 19 U.S.C. § 1514(c)(1). The government therefore argues that their liquidation became final by operation of 19 U.S.C. § 1514(a). See Def.’s Br. at 7-8 (referencing Po-Chien, Inc. v. United States, 3 CIT 17, 18 (1982) (liquidation final and conclusive against the importer since it had not timely filed a valid protest with the proper Customs district) and Grover Piston Ring Co. v. United States, 752 F.2d 626, 627 (Fed.Cir.1985) (failure to include entry numbers as part of the content of the protest caused the protest to be invalid with respect to those entry numbers)). Avecia responds that the issue is only being raised for the first time in post-trial briefing and was therefore waived, or else the protest was valid. Pl.’s Reply at 2-7.

Generally speaking, the terms of the government’s consent to be sued in a particular court define the court’s jurisdiction to entertain the suit, must be strictly observed, and are not subject to implied exceptions. NEC Corp. v. United States, 806 F.2d 247, 249 (Fed.Cir.1986) (citations omitted). “[S]ubject-matter jurisdiction, because it involves the court’s power to hear a case, can never be forfeited or waived.” United States v. Cotton, 535 U.S. 625, 630, 122 S.Ct. 1781, 1785, 152 L.Ed.2d 860 (2002). Therefore, courts have an independent obligation to “police” their own subject-matter jurisdiction, even in the absence of a challenge from a party. Ruhrgas AG v. Marathon Oil Co., 526 U.S. 574, 583-84, 119 S.Ct. 1563, 1570, 143 L.Ed.2d 760 (1999) (citations omitted). Under 28 U.S.C. § 1581(a), this court possesses exclusive jurisdiction over any civil action commenced to contest the denial of a protest under 19 U.S.C. § 1515, which provides for the review of protests filed in accordance with 19 U.S.C. § 1514 concerning decisions of the U.S. Customs and Border Protection and its predecessor organization(s). The terms of section 1581(a) “limit[ ] the jurisdiction of the Court of International Trade to appeals from denials of valid protests. Thus, the court lacks jurisdiction over protests that do not satisfy the requirements of 19 U.S.C. § 1514(c)(1) and 19 C.F.R. 174.13(a).” Koike Aronson, Inc. v. United States, 165 F.3d 906, 908 (Fed.Cir.1999). Further, the burden of establishing jurisdiction lies with the party seeking to invoke the court’s jurisdiction. E.g. Takashima U.S.A., Inc. v. United States, 19 CIT 673, 677, 886 F.Supp. 858, 861 (1995); Old Republic Insurance Co. v. United States, 14 CIT 377, 379, 741 F.Supp. 1570, 1573 (1990).

Thus, the government is correct, again generally speaking, that compliance with formality is required in order to perfect a valid protest. Cf. 19 U.S.C. § 1514(c)(1) (“in accordance with regulations prescribed by the Secretary”); Noury Chemical Corp. v. United States, 4 CIT 68 (1982) (jurisdiction lacking because protest letter not filed with the proper party, the district director at Buffalo, New York); Po-Chien, supra, 3 CIT at 18 (jurisdiction lacking because protest addressed only to “U.S. Customs Service” and not to the office of the district director whose decision was being protested). But, compliance is a question of fact, see, e.g., W.R. Grace & Co. v. United States, 15 Cust.Ct. 105, C.D. 953 (1945), and

[i]t is always within the discretion of ... an administrative agency to relax or modify its procedural rules adopted for the orderly transaction of business before it when in a given case the ends of justice require it. The action of either in such a case is not reviewable except upon a showing of substantial prejudice to the complaining party.

American Farm Lines v. Black Ball Freight Service, 397 U.S. 532, 539, 90 S.Ct. 1288, 1292, 25 L.Ed.2d 547 (1970) (citation omitted). Cf. PAM S.p.A. v. United States, 463 F.3d 1345, 1348 (Fed.Cir.2006) (inquiry into substantial prejudice). In this matter, the government’s concern that Avecia did not comply with required formality to protest the three entries at their respective ports of entry is obscured by the fact that the customs officer acting on behalf of the port director did not reject consideration of those entries on that ground but rather “denied in full” the protest at issue, number 1101-04-100239 “per [¶] 967005 [dated] 5/18/04.” Nothing in the statutes indicates that only the district directors of Newark and Baltimore had authority to consider the protest. Cf. 19 U.S.C. § 1514(a) (“decision of the Customs Service ... shall be final and conclusive ... unless a protest is filed in accordance with this section”) (italics added). Customs having asserted jurisdiction over the protest’s entirety, and the improper inclusion of the three entries not having been part of the rationale for denial at least as to those three, the Court will not substitute post-hoc rationale to deny it. Avecia’s invocation of jurisdiction here over the subject matter at issue is therefore proper.

II

On another preliminary matter, Avecia challenges Customs’s published General Notice of Revocation as well as [¶] 966063 relating to the tariff classification of FY2, which concluded that FY2 is classifiable under subheading 3204.14.30. See POSC at ¶ 11. 19 U.S.C. § 1625(c) requires the Secretary of Customs to publish in the Customs Bulletin a final ruling or decision on a proposed interpretive ruling or decision that would “have the effect of modifying the treatment previously accorded by the Customs Service to substantially identical transactions” within 30 days after the close of the comment period. Avecia argues Customs did not issue [¶] 966063 (revoking [¶] 964191, 962365 and 962918) within 30 days of the close of the comment period and therefore did not conform with the requirements of 19 U.S.C. § 1625(c) for revoking a binding ruling. Avecia thus argues that the revocation ruling was ineffective, that [¶] 964191, 962365 and 962918 remain binding on Customs, and that the subject imports are therefore properly classified as printing inks based on these earlier rulings. See PL’s Br. at 22-23.

According to the first Diamond Match case, the purpose of section 1625(c) is notice, and as such does not “restrain the doing of the act after the time limit or state any consequences if action is delayed” and is therefore directory, a point with which the appellate court agreed, “particularly where ... the party in interest has been in no way prejudiced” by the inaction or delay. Diamond Match Co. v. United States, 44 Cust.Ct. 67, 68, C.D. 2154, 181 F.Supp. 952, 959 (1960), aff'd, 49 C.C.P.A. 52, C.A.D. 796 (1962). See also Dixon Ticonderoga Co. v. United States, 468 F.3d 1353 (Fed.Cir.2006) (failure of agency to follow procedural requirement does not void subsequent agency action); Intercargo Insurance Co. v. United States, 83 F.3d 391 (Fed.Cir.1996) (failure of notice from agency was harmless error). That is the apparent circumstance here, and Avecia’s procedural challenge to the revocation ruling therefore does not invalidate it.

Ill

Although the plaintiff has the burden of establishing that the government’s classification of the product was incorrect, it does not bear the burden of establishing the correct tariff classification. Jarvis Clark Co. v. United States, 733 F.2d 873, 878 (Fed.Cir.1984). Pursuant to 28 U.S.C. § 2639(a)(1), a statutory presumption of correctness is afforded to Customs’s classification decisions concerning the facts of a classification. Such presumption does not extend to questions of law. See Universal Electronics, Inc. v. United States, 112 F.3d 488, 492 (Fed.Cir.1997).

Our appellate court deems the determination of the correct tariff classification a two-step process: properly construe the relevant classification headings, and determine which one properly applies to the merchandise. See, e.g., Bausch & Lomb, Inc. v. United States, 148 F.3d 1363, 1364-65 (Fed.Cir.1998). Regardless of the number of steps involved, several legal principles have been invoked to support end results. Interpreting the meaning of a tariff provision involves statutory construction and is therefore a question of law. See, e.g., Boen Hardwood Flooring, Inc. v. United States, 357 F.3d 1262, 1264 (Fed.Cir.2004); Carl Zeiss, Inc. v. United States, 195 F.3d 1375, 1378 (Fed.Cir.1999). Determining the “nature” of merchandise to be classified is a question of fact. See, e.g., Boen Hardwood; 357 F.3d at 1264; Bausch, 148 F.3d at 1365-66. Determining whether merchandise to be classified “comes within” a properly construed tariff provision, an apparently penultimate issue, is also said to be a question of fact. See, e.g. Rollerblade, Inc. v. United States, 112 F.3d 481, 483 (Fed.Cir.1997) (referencing Totes, Inc. v. United States, 69 F.3d 495, 498 (Fed.Cir.1995)); National Advanced Systems v. United States, 26 F.3d 1107, 1109 (Fed.Cir.1994). But, determining which tariff provision imported merchandise is properly classified under, which is the ultimate issue, is a question of law. See, e.g., Sports Graphics, Inc. v. United States, 24 F.3d 1390, 1391 (Fed.Cir.1994); National Advanced Systems, 26 F.3d at 1109.

In addition to the chapter headings and relevant notes, classification of merchandise is undertaken by applying the General Rules of Interpretation to the HTSUS (“GRIs”). See, e.g., Boen Hardwood, 357 F.3d at 1264; North American Processing Co. v. United States, 236 F.3d 695, 698 (Fed.Cir.2001). GRI 1 provides that classification is to be determined according to the terms of the headings and any relevant section or chapter notes. GRIs 2 through 4 then apply, “provided such headings or notes do not otherwise require.” The section and chapter notes “are not optional interpretive rules, but are statutory law, codified at 19 U.S.C. § 1202.” Park B. Smith, Ltd. v. United States, 347 F.3d 922, 926 (Fed.Cir.2004). Thus, the first step to analyzing a classification issue is to determine the applicable heading by looking to the terms of the headings and section or chapter notes. See GRI 1. See e.g., Orlando Food Corp. v. United States, 140 F.3d 1437, 1440 (Fed.Cir.1998). If the merchandise is classifiable under more than one heading, “[t]he heading which provides the most specific description shall be preferred to headings providing a more general description.” GRI 3(a), HTSUS. See, e.g., Orlando Food, 140 F.3d at 1440. Once an imported product is determined to be classifiable under a particular heading, a court must then look to the subheadings to find the correct classification of the merchandise in question. Id.

Further, absent contrary definitions in the HTSUS or legislative history, the terms used in the headings and subheadings are to be construed according to their “common and popular meaning,” which may be drawn from a court’s own understanding of dictionaries and other reliable sources including scientific authorities. See, e.g., Russell Stadelman & Co. v. United States, 242 F.3d 1044, 1048 (Fed.Cir.2001); Carl Zeiss, supra, 195 F.3d at 1379; Medline Industries, Inc. v. United States, 62 F.3d 1407, 1409 (Fed.Cir.1995). Additionally, the Harmonized Commodity Description and Coding System Explanatory Notes (“Explanatory Notes”) may be used “to clarify the scope of HTSUS subheadings and to offer guidance in interpreting subheadings.” Mita Copystar America v. United States, 21 F.3d 1079, 1082 (Fed.Cir.1994). They are “indicative of proper interpretation” of the HTSUS but are “not legally binding[.]” Lynteq, Inc. v. United States, 976 F.2d 693, 699 (Fed.Cir.1992) (quoting H.R. Conf. Rep. No. 100-576, 100th Cong., 2d Sess. 549 (1988), reprinted in 1988 U.S.C.C.A.N. 1547,1582) (internal quotation marks omitted).

A

In its post-trial brief, the government raised the argument that the products at issue are actually separate chemically defined compounds of Chapter 29, HTSUS, and that Avecia’s proposed classification under heading 3215 is in conflict with Ciba-Geigy Corp. v. United States, 223 F.3d 1367 (Fed.Cir.2000). Def.’s Br. at 14. In Ciba-Geigy, the appellate court held that classification of a chemical powder under heading 3215 was precluded by-operation of Chapter Note 1(a) to Chapter 32, which excludes, inter alia, “separate chemically defined elements or compounds (except those of heading ... 3204[) ].” 223 F.3d at 1372-1373. In other words, heading 3215 is not mentioned “as a specific exception to the general rule that ‘separate chemically defined compounds’ are not classifiable in Chapter 32.” Id. at 1373. The government here also points to Notes 1(d) and (f) to Chapter 29, which explain that a solution of “separate chemically defined organic compounds” in water and/or with added stabilizers is nonetheless covered by Chapter 29, to argue that “the imported colorants themselves, without taking into account the water and the bio-cide (for the Cyan 1 RO feed), are ‘separate chemically defined compounds.’ ” Def.’s Br. at 15. Further, the government argues that if the record does not contain sufficient information for the Court to make that determination, the matter should be remanded to Customs to permit Avecia and Customs to develop the facts further. Id. at 14.

Avecia takes umbrage at this eleventh-hour defense. See Pl.’s Reply at 11. On the other hand, noting that the Court must “reach a correct decision in every case” (see Jarvis Clark, supra, 733 F.2d at 878), Avecia argues that Ciba-Geigy turned on the fact that the parties thereto had agreed that its imported products were described by Chapter 29, ie., were essentially “separate chemically defined compounds.” Id. Cf. 223 F.3d at 1370 (“[b]oth parties agree that [the imported products] generally fall within the definition of ‘organic chemicals’ ”). Avecia does not agree that the imported products at issue can be characterized as such. Id. Avecia contends there is sufficient evidence from the trial record to make such a determination, but if not, then the fault is that of the government: if the issue had been timely raised, Avecia argues, it would have afforded the opportunity to introduce evidence to show that ink-jet ink chromo-phores are always a mixture of individual compounds in equilibrium rather than a single mappable compound. Id. at 12.

In support of this contention, Avecia requests consideration of the declaration of Dr. Bidd attached to its reply brief rather than a remand of the matter to Customs or a reopening of the record for what it considers would amount to a superfluous evi-dentiary hearing. Id. (referencing Attachment A to Pl.’s Reply (“Declaration of Ilesh Bidd”)). Alternatively, Avecia reiterates that each of the imports at issue is a mixture (solution) of water and soluble colorant, and it stresses that Note 1(a) to Chapter 32, HTSUS, is clear in excluding from Chapter 32 classification only “separate chemically defined elements or compounds,” not “mixtures” or “solutions” of those. Id. at 7-11. Cf., e.g., HTSUS, ch. 28, notes 1 & 1(a) (stating that the headings of Chapter 28 apply only to separate chemical elements and separate chemically defined compounds “[ejxcept where the context otherwise requires”). Avecia stresses that “water” in the context of the products at issue is a “separate chemically defined compound[,]” that water and colorant form the ink-jet ink system, and that “[w]ithout the water, the system doesn’t work.” Id. (quoting inter alia R 720:11-16).

“Separate chemically defined elements or compounds” appears in both Chapters 28 and 29, HTSUS, in addition to Chapter 32. The Explanatory Notes to Chapters 28 and 29 each include the following definition:

A separate chemically defined compound is a substance which consists of one molecular species (ie., covalent or ionic) whose composition is defined by a constant ratio of elements and can be represented by a definitive structural diagram. In a crystal lattice, the molecular species corresponds to the repeating unit cell.

See 1 Explanatory Notes 260, 371 (3d ed.2002) (italics added).

A mixture of various “separate chemically defined compounds” is not itself a “separate chemically defined compound.” Metchem, Inc. v. United States, 30 CIT -, 441 F.Supp.2d 1269 (2006), reh’g denied (USCIT Nov. 3, 2006). See USR Optonix, Inc. v. United States, 29 CIT -, 362 F.Supp.2d 1365, 1370 n. 3 (2005) (construing “chemical compound” in the HTSUS term “separate chemically defined compound” as “a substance composed chemically of two or more elements in definite proportions (as opposed to a mixture)”) (quoting 3 Oxford 629; Oxford’s emphasis). Further, the Explanatory Notes to Chapters 28 and 29 indicate that those Chapters cover “separate chemically defined elements or compounds” that have been dissolved in water or dissolved in solvents other than water provided that “the solution constitutes a normal and necessary method of putting up these products adopted solely for reasons of safety or for transport (in which case the solvent must not render the product particularly suitable for some types of use rather than for general use).” Cf. 1 Explanatory Notes 261 with id. at 371 (applying “[t]he provisions in the General Explanatory Note to [c]hapter 28 concerning the addition of stabilisers ... mutatis mutandis, to the chemical compounds of’ Chapter 28) (bolding omitted). Implicitly, then, the water of a solution of a separate chemically defined compound must function for reasons of safety or transport, and it must not render such product “particularly suitable” for some types of use rather than for the general chemical use of the separate chemically defined compound. The Court is thus persuaded that Avecia is correct that water itself may be considered a “separate chemically defined compound” when acting as the “ink vehicle” of an ink-jet ink: in such a context, water performs the integral functions of delivery of the colorant to the substrate to which it must bind if it is to print, penetration, (to some degree) of the colorant into the cellulosie fiber surface of the substrate, and the resultant binding of the colorants to the substrate upon its evaporation. In such a context, water accomplishes more than “mere” preservation or stabilization for transportation.

However, that is not so with respect to the water in which the subject ink-jet inks have been imported into the United States: the products are imported in concentrate form, and upon importation deionized water is added through reverse osmosis which also displaces some of the imported water. Thus, the water employed for transport of Avecia’s ink-jet inks is not necessarily the same water that is subsequently used for printing; rather, the primary function of the water at the time of importation is to act not as the ink vehicle but to provide stabilization or preservation and to protect the products’ integrity. See infra Pl.’s Reply, Attachment A ¶ 4. Cf. R 376:17-24 (product “has to remain totally intact from this point onwards”); Mita Copystar, supra, 21 F.3d at 1082 (“[i]t is well settled law that merchandise is classified according to its condition when imported”) (citation omitted). Thus, the primary function of the water with which the products are imported does not obviate classification of the products as “separate chemically defined organic compounds” of Chapter 29.

Nonetheless, the Court finds sufficient evidence to conclude that each of the imported ink-jet inks, in their condition as imported, does not in fact consist of a single “separate chemically defined compound” as required for classification under Chapter 29. For that purpose, the Court accepts the declaration of Dr. Bidd attached as an appendix to Avecia’s reply brief. Cf. 28 U.