Control of leachable mercury in fluorescent lamps

Description

TECHNICAL FIELD

[0001] This invention relates to mercury vapor discharge lamps and more particularly to fluorescent lamps. Still more particularly it relates to lamps that can be landfilled without leaching potentially damaging mercury into the environment. The invention relates for example to linear fluorescent lamps, including large and small diameter fluorescent lamps corresponding for example to 30 to 40 mm diameter (large) or below 30 mm (small). Typical diameters are T12 (about 38 mm, large) and T8 (about 25 mm, small) and even smaller ones.

BACKGROUND ART

[0002] Fluorescent lamps contain elemental mercury. During lamp operation, chemical reactions take place that convert some of the elemental mercury to salts or compounds, such as mercuric oxide (HgO), that are water soluble. There is a growing concern that a waste stream resulting from the disposal of fluorescent lamps may leach excessive amounts of this soluble form of mercury (Hg) into the environment. An acceptable method of measuring the amount of soluble mercury which may leach from the waste stream resulting from the disposal of fluorescent lamps is described in the Toxicity Characteristic Leaching Procedure (TCLP) prescribed on pages 26987 - 26998 of volume 55, number 126 of the June 29, 1990 issue of the Federal Register. The lamp to be tested is pulverized into granules having a surface area per gram of materials equal to or greater than 3.1 cm² or having a particle size smaller than 1 cm in its narrowest dimension. The granules are then subject to a sodium acetate buffer solution having a pH of approximately 4.9 ands a weight twenty times that of the granules. The buffer solution is then extracted, and the concentration of mercury is measured. At the present time, the United States Environmental Protection Agency (EPA) defines a maximum concentration level for mercury to be 0.2 milligram of leachable mercury per liter of leachate fluid when the TCLP is applied. According to the present standards, a fluorescent lamp is considered nonhazardous (and thus available to be conventionally land-filled) when less than 0.2 milligram per liter of leachable mercury results using the TCLP. Lamps that do not exist within the constraints of TCLP must be especially disposed of through licensed disposal operations. Lamps that have leachable mercury concentrations above the allowable limit must be especially disposed of through licensed disposal operations. Disposal operators charge a fee for disposal of lamps that are not within the EPA's limits. Therefore, customers must pay extra costs to dispose of these lamps. Customers of fluorescent lamps generally do not desire to have to contend with the EPA and disposal concern regarding mercury levels, and therefore some customers specify only those lamps which pass the TCLP standard. Customers of fluorescent lamps generally desire not to contend with disposal issues regarding mercury levels, and therefore some customers specify only those lamps which pass the TCLP standard.

[0003] Heretofore, efforts have been made to reduce the leaching of soluble mercury from fluorescent lamps during the TCLP testing as well as in landfills. Various methods have been proposed which attempt to treat or process burned-out discharge lamps or scrap lamp exhaust tubing containing mercury in order to reclaim the mercury and thereby reduce the amount of mercury-contaminated scrap.

[0004] U.S. Patent No. 5,998,927, Foust, et al., teaches a method for inhibiting the formation of leachable mercury associated with a mercury arc vapor discharge lamp when the mercury is in elemental form. The method comprises providing high-iron content metal components in the lamps, at least one of the high-iron content metal components having an amount of oxidizable iron of at least about 1 gm per kilogram of lamp weight.

[0005] What is not specifically addressed in the patent, however, is the situation in which practically all of the mercury may already be present in the soluble ionic form at the start of the TCLP testing, as a result of naturally occurring processes that take place within the fluorescent lamp during its operation.

DISCLOSURE OF INVENTION

[0006] It is, therefore, an object of the invention to obviate the disadvantages of the prior art.

[0007] It is another object of the invention to enhance the disposal of fluorescent lamps.

[0008] It is yet another object of the invention to allow conventional landfill disposal of fluorescent lamps when the mercury contained therein is in ionic form.

[0009] These objects are achieved, in one aspect of the invention, by the provision of a method for inhibiting mercury from leaching from a mercury vapor discharge lamp wherein at least a part of said mercury is present as ionic mercury, with reduced mercury leaching occasioned by the application by using a non-conductive tin oxide layer within the lamp working in conjunction with an amount of oxidizable iron.

[0010] It would be a further advance in the art if mercury leaching from large diameter (e.g., those with diameters of typically 30 to 40 mm) as well as from small diameter fluorescent lamps (e.g., those with diameters less than 26 mm) could be achieved, too.

[0011] These objects are achieved, in a further aspect of the invention, by the provision of a method for inhibiting mercury from leaching from a mercury vapor discharge lamp wherein at least a part of said mercury is present as ionic mercury, and wherein the lamp includes an amount of oxidizable iron, the amount of iron being at least 1 gram per kilogram of lamp weight, and further including within the lamp a quantity of substantially un-doped SnO₂, having a density of about 40 micrograms/cm².

[0012] The addition of the substantially un-doped stannous oxide provides a totally unexpected, synergistic effect between the stannous oxide and the oxidizable iron to inhibit mercury leaching when the mercury is present in an ionic form.

BEST MODE FOR CARRYING OUT THE INVENTION

[0013] For a better understanding of the present invention, together with other and further objects, advantages and capabilities thereof, reference is made to the following disclosure and appended claims.

[0014] A first embodiment is dealing with large diameter T12 lamps without excluding smaller diameters.

[0015] Shown in Table I are the results of a series of TCLP tests carried out with F40T12 lamps in which all of the mercury (5.0-5.5 mg) was initially present in the soluble ionic form (added as HgO). The first test was run without the addition of any metallic iron. However, the second and third tests were run with the inclusion of 4.3 cm² of 0.15 mm thick metallic iron foil, while the forth and fifth tests were run with the inclusion of 6.4 cm² of 0.15 mm thick iron foil. These quantities of metallic iron correspond to approximately 1.8 and 2.7 grams of oxidizable iron per kilogram of lamp weight, well within the range prescribed by Foust, et al in the referenced patent. Also shown in the table are the concentrations of soluble iron (i.e., the concentrations of iron in the leaching solution) at the end of each test.

TABLE I

Fe (cm2) (Surface Area)	Initial Hg (mg, as HgO)	Final Soluble Hg (mg/l)	Final soluble Fe (mg/l)
0	5.5	0.38	0.8
4.3	5.5	0.23	21
4.3	5.5	0.26	20
6.4	5.5	0.20	28
6.4	5.5	0.21	30

[0016] As shown in the table, the greater the metallic iron surface area, the greater the soluble iron concentration and the lower the concentration of soluble mercury at the end of the TCLP. However, even with the inclusion of 6.4 cm² of oxidizable metallic iron, the resulting soluble mercury concentrations can lie above the maximum allowable concentration of leachable mercury, as determined via the TCLP. Also, it should be noted that the range of soluble iron concentrations determined in these tests is in essential agreement with that represented in the referenced Foust, et al. patent.

[0017] In order to reduce the voltage necessary for ignition of certain fluorescent lamps (in particular, certain T12 lamp types that are no more than 1,3 m in length), it is known in the art to deposit transparent and electrically conductive, doped SnO₂ upon the inside surfaces of the cylindrical glass lamp envelopes. The SnO₂ is typically doped with fluorine (F) or antimony (Sb) (most typically fluorine), which dopants have the effect of greatly increasing the electrical conductivity of the material.

[0018] Previous TCLP testing of otherwise identical mercury-containing fluorescent lamps which either do or do not contain such conductive (F-doped) SnO₂ coatings on the inside surfaces of the glass envelopes have shown that the F-doped SnO₂ coating does not affect significantly the results of TCLP testing. I.e., such coatings have little or no effect upon the amount of mercury dissolved within the extraction fluid at the end of the TCLP test. This result is demonstrated in Table II wherein are listed the results of two TCLP tests carried out with F40T12 lamps in which all of the mercury (5 mg) was initially present in the soluble ionic form (added as HgO). The first test was run with a lamp which did not contain an SnO₂ coating on the glass (similar to the first test listed in Table I above), while the second test was run with glass that had been coated on the inside surface with F-doped SnO₂. As shown, the presence of the conductive, F-doped SnO₂ coating had essentially no effect upon the result of the TCLP test.

TABLE II

F-Doped SnO2 Coating	Final Soluble Hg (mg/l)
No	0.38
Yes	0.37

[0019] In view of these results, it was surprising to discover that the combination of a F-doped SnO₂ coating with a relatively small amount of oxidizable metallic iron (as used in the TCLP testing described above, Table I) is particularly efficacious in the reduction of extractable ionic mercury during the TCLP.

[0020] The apparent synergy between F-doped SnO₂ and relatively small quantities of oxidizable metallic iron is demonstrated in Table III below. Here, standard TCLP tests were run using a T12 glass envelope with or without the standard F-doped SnO₂ coating. In each case, a test was run with and without the presence of 6.4 cm² of 0.15 mm thick metallic iron foil. 4.5 mg of ionic mercury (as HgO) was also included in each test. The quantity of TCLP fluid was typical of that used in the TCLP test of a T12 fluorescent lamp (5.6 1). (A thin layer of fumed alumina was also present on the inside surface of each F-doped SnO₂ coating, as is common in the manufacture of fluorescent lamps).

TABLE III

F-Doped SnO2	Fe Foil (6.4 cm2)	Final Soluble Hg (mg/l)	Final Soluble Fe (mg/l)
No	No	0.89	<0.03
Yes	No	0.88	<0.03
No	Yes	0.61	29
Yes	Yes	0.06	27

[0021] To determine if the other transition metals were as efficacious as iron in the removal of ionic mercury, another series of tests was run employing these metals and glass coated with F-doped SnO₂. In each test, 6.4 cm² of the substituted metal was employed. As shown by the results in Table IV, all other transition metals were relatively ineffective in inhibiting the leaching of ionic mercury.

TABLE IV

Metal	Final Soluble Hg (mg/l)
Iron	0.04
Cobalt	0.45
Copper	0.50
Nickel	0.68

[0022] Additional TCLP tests were run in the same way as were those described above, except that the quantities of metallic iron foil and of F-doped SnO₂-coated glass were varied. The results of these tests are compared in Table V. As shown, the quantity of metallic iron foil that effectively prevents mercury leaching, as measured by the TCLP, decreases with increasing surface area of F-doped SnO₂-coated glass. And, similarly, the quantity of SnO₂-coated glass required to effectively prevent mercury-leaching decreases as the amount of oxidizable metallic iron foil increases.

TABLE V

Fe Area (cm2)	F-Doped SnO2 Glass Area (ft2)
	0	1	1.5
0	0.89	---	0.88
0.2	---	---	0.75
1	---	0.75	0.05
6.4	0.61	0.06	0.07

[0023] The effectiveness of the synergy between F-doped SnO₂ ands metallic iron in effecting the TCLP compliancy of a fluorescent lamp is demonstrated in Table VI. Listed here are the results of TCLP tests carried out with F40D35 lamps both with and without the standard F-doped SnO₂ coating on the inside surface of the glass envelope. For each of these two cases, TCLP results are shown for tests run which included 4.3 cm² of 0.15 mm thick metallic iron foil along with 5.5 mg of soluble ionic mercury (as HgO). The effectiveness of the metallic iron/F-doped SnO₂ synergy in promoting the reduction of soluble ionic mercury in the TCLP is apparent.

TABLE VI

F-Doped SnO2 Coated	Final Soluble Hg (mg/l)
No	0.23
No	0.23
No	0.24
No	0.26
Yes	0.16
Yes	0.16
Yes	0.17

[0024] While the results of the above-tests are startling as regards the removal of ionic mercury, there are serious disadvantages associated with the use of F-doped coatings, both in the manufacture and operation of fluorescent lamps. Hydrofluoric acid (HF) is typically used as a relatively inexpensive source of fluoride. A concentrated aqueous solution of HF is mixed with the other coating-solution components (typically SnCl₄, water, and alcohol), and the resulting mixture is sprayed onto the inside surface of the preheated glass tube. In addition to the concentration-dependent corrosive effect that HF shares with other strong acids, there is an additional extreme toxicity effect due to the fluoride. A specific property of fluoride is that it readily penetrates human skin, allowing it to destroy soft tissue and decalcify bone. The toxic effect is due to the precipitation of the available bodily calcium by the fluoride. This typically leads to a drastic drop of the calcium level, essential for most vital functions. If not promptly treated, often-fatal complications may follow (e.g., cardiac arrest). It would obviously be advantageous if the fluoride content of the SnO₂ coating could be eliminated without significantly altering the beneficial effects of the coating.

[0025] The presence of the fluoride dopant in the SnO₂ coating also exacerbates the formation of a type of lamp defect often referred to as 'black spot patches' or 'measles' which develop during lamp operation as a result of an interaction involving the conductive layer and the mercury in the arc discharge. The mercury penetrates the phosphor layer, leading to conditions that allow buildup of charge and subsequent discharge, which result in the 'measle' defect by disrupting the phosphor layer and generally forming a small crater in the glass tube. In particular, the formation of 'measles' is believed to be caused by the presence of salts in the conductive layer, salts which are formed largely by reaction between the fluoride in the conductive tin-oxide coating solution and sodium in the glass. For this reason, also, it would be advantageous if the fluoride content of the SnO₂ coating could be eliminated without significantly altering the beneficial effects of the coating.

[0026] What is disclosed herein is a method for inhibiting the leaching of mercury from mercury-containing fluorescent lamps, as determined by the TCLP. The method comprises providing a transparent, substantially undoped (and, therefore, effectively nonconducting) tin-oxide coating on the inside surfaces of the glass envelopes of said lamps, in combination with high-iron content metal components at least one of which contains an amount of oxidizable iron of at least about 1 gram per kilogram of lamp weight.

PREPARATION

[0027] The inside surfaces of two groups of standard T12 lamp envelopes about 38 mm (or 1.5 inches) in diameter and about 1,22 m (4 feet) in length were coated with tin oxide (SnO₂) using the standard spraying method with standard SnCl₄ and solvent concentrations used in each case. In one case, the sprayed solution also contained the standard concentration of hydrofluoric acid (HF). In the other case the tubes were coated using an SnCl₄ solution which did not contain any HF. Both groups of lamp envelopes were coated with the standard thickness of SnO₂.

[0028] After being subjected to the standard baking operation (designed to fuse the SnO₂ layer to the inside surface of the glass tube), the relative resistivities of the undoped and F-doped coatings were determined using point probes positioned close to the ends of the coated surfaces of each tube. The relative end-to-end film resistance of the undoped SnO₂ coating was found to be between 3 and 4 times that of the F-doped coating.

[0029] The coated surfaces also were examined analytically by two methods: energy dispersive spectroscopy (EDS) and Rutherford backscattering spectroscopy (RBS). Taken together, the results of these measurements indicated an average film thickness of about 50 nm (corresponding to a coating density of approximately 40 micrograms/cm²).

EXAMPLE I

[0030] Two TCLP tests were run, in each case using a T12 glass lamp envelope, 6.4 cm² of 0.15 mm thick metallic iron foil, and 4.5 mg of ionic mercury (as HgO). A 5.6 liter volume of TCLP extraction fluid was used in each case (the amount of fluid used in a typical T12-lamp TCLP test). A T12 glass tube with an undoped SnO₂ coating was used in the first test, while an uncoated tube was used in the other test. The results of these TCLP tests are listed in Table VII. As shown, an extracted mercury concentration well below the critical 0.2 mg/l value was obtained in the test run with the SnO₂-coated T12 glass. In contrast, an extracted mercury concentration well above the 0.2 mg/l 'passing' level was obtained when the uncoated glass was used.

TABLE VII

Undoped SnO2 Coating	Final Soluble Hg (mg/l)
Yes	0.07
No	0.63

EXAMPLE II

[0031] Two other TCLP tests were run, in each case using a 0,1 m² (or 1 ft² ) area of plate glass, 6.4 cm² of 0.15 mm thick metallic iron foil, and 4.5 mg of ionic mercury (as HgO). The glass used in one of the two tests was coated on one side with undoped SnO₂. As in Example I, a 5.6 1 volume of TCLP extraction fluid was used in each test. The resistivity of the undoped SnO₂ coating was found to be 5-to-6 times that of a SnO₂ coating of approximately the same thickness deposited using the standard F-dopant concentration. The results of these two TCLP tests are listed in Table VIII. As shown, an extracted mercury concentration well below the critical 0.2 mg/l value was obtained in the test carried out with undoped-SnO₂-coated glass. In contrast, an extracted mercury concentration much greater than the 0.2 mg/l 'passing' level was obtained when uncoated glass was used.

TABLE VIII

Undoped SnO2 Coating	Final Soluble Hg (mg/l)
Yes	0.04
No	0.55

EXAMPLE III

[0032] A series of TCLP tests were run with standard F40D830 lamp components (phosphors deposited on the inside surfaces of the T12 glass tubes, aluminum endcaps, tungsten filaments with nickel-plated iron mounts, and copper-plated outer leads with brass connecting pins). However, while some of the glass lamp envelopes were manufactured with the standard F-doped SnO₂ coating, others were coated with undoped SnO₂, and some had no SnO₂ at all. Included in each test were 4.3 cm² of oxidizable metallic iron foil and 4.5 mg of ionic mercury (as HgO). The average final soluble mercury and iron concentrations obtained in these tests are listed in Table IX. As shown, an extracted mercury concentration well below the critical 0.2 mg/l value was obtained when glass coated either with undoped SnO₂ or with F-doped SnO₂ was used, while extracted mercury concentrations greater than the 0.2 mg/l 'passing' level were obtained with uncoated glass. However, the amount of oxidized iron dissolved in the extraction fluid was nearly the same in each case.

TABLE IX

SnO2 Coating	Final Soluble Hg (mg/l)	Final Soluble Fe (mg/l)
None	0.21	19
F-doped	0.16	21
Undoped	0.14	20

[0033] As shown by the above examples, the improved method for the control of leachable mercury in a fluorescent lamp is based upon the surprising synergy that exists between substantially undoped (and, therefore, effectively nonconducting) SnO₂ (deposited upon the inside surface of the lamp's glass envelope) and a relatively small amount of oxidizable metallic iron or other high-iron content metal, to inhibit mercury leaching. Intended to be included within the scope of this method are situations in which there may possibly be slight accidental doping of the SnO₂ by impurities which inadvertently may be present within the coating-precursor solution. The high-iron content metal could be included within the lamp in a variety of ways, as suggested by the prior art.

[0034] As disclosed in the afore-mentioned embodiment, it has been discovered that replacing the electrically conductive tin oxide coating of fluorescent lamps with a non-electrically conductive tin oxide coating in conjunction with an amount of oxidizable iron will reduce the mount of leachable mercury in said lamps when that mercury is present in its ionic form.

[0035] It has furthermore discovered that applying a coating of tin oxide to envelopes having different diameters, for example 25 to 30 mm compared to 35 to 40 mm is equally efficacious in preventing the leaching of mercury from mercury-containing fluorescent lamps, as determined by TCLP, and this condition applies whether the tin oxide is conducting or non-conducting.

EXAMPLE IV

[0036] Four TCLP tests were run, in each case using the components of standard T8 fluorescent lamps having tubular glass envelopes 1,22 m long. Two additional tests were run with similar components, except that the inside surface of the tubular glass envelopes was coated with a tin oxide having the same thickness and composition as that normally used as a starting aid with T12 lamps, i.e., a density of about 40 micrograms/cm². Also included in each test was a SAES mercury dispenser/getter strip 2.5 cm² in area, a piece of metallic iron foil 0.15 mm thick and either 1.6 cm² or 2.5 cm² in area, and 4.5 mg of ionic mercury (as HgO). The results are shown in Table XI.

TABLE XI

TCLP Results for T8 Lamps having 4.5 mg of Soluble Ionic Mercury
SnO2 Coating	Fe Foil (cm2)	Final Soluble Hg Concentration (mg/l)	Final Soluble Fe Concentration (mg/l)
Yes	1.6	0.14	11
Yes	1.6	0.15	11
No	1.6	0.19	6.7
No	1.6	0.20	7.2
No	2.5	0.20	12
No	2.5	021	14

[0037] As shown by the results in Table I, extracted mercury concentrations well below the critical 0.2 mg/l value are obtained when SnO₂-coated glass was combined with only 1.6 cm² of metallic iron foil, whereas failing or nearly failing results were obtained when the standard uncoated glass tubing was used with either 1.6 or 2.5 cm² of iron foil. (Note that 1.6 cm² of 0.15 mm thick iron foil is equivalent to approximately 1.5 grams of metallic iron per kilogram of lamp weight).

EXAMPLE V

[0038] Six additional TCLP tests were run, in each case using the components of a standard T5 fluorescent lamp with a glass envelope 1,14 m (or 45 inches) long. In three of the six tests the inside surface of the tubular glass was coated with tin oxide having the same thickness and composition as that employed in Example I. Also included in each test was a quantity of ionic mercury (2 or 4 mg, as HgO) and a quantity of metallic iron foil (0.75 or 1.5 cm² in area and 0.15 mm thick). Most tests also included a SAES mercury dispenser/getter strip 0.75 cm² in area. The conditions of each test and the corresponding results are summarized in Table XII.

TABLE XII

TCLP Results for T5 Lamps with 2 or 4 mg of Soluble Ionic Mercury
SnO2 Coating	Ionic Hg (as HgO, mg)	Fe Foil (cm2)	SAES Getter	Final Soluble Hg Concn. (mg/l)	Final Soluble Fe Concn. (mg/l)
No	2	0.75	Yes	0.20	9
Yes	2	0.75	Yes	0.14	10
No	4	0.75	Yes	0.37	5
Yes	4	0.75	Yes	0.14	8
No	4	1.5	No	0.21	11
Yes	4	1.5	No	014	13

[0039] With both quantities of ionic mercury, the extracted mercury concentrations were well below the critical 0.2 mg/l value when the SnO₂-coated glass was combined with only 0.75 cm² of metallic iron foil, whereas failing results were obtained when the standard uncoated glass was used, regardless of the amount of iron foil and regardless of the amount of ionic mercury.

[0040] This method for controlling the amount of leachable mercury in fluorescent lamps with diameters less than 40 mm (or 1.5 inches) is based upon the surprising synergy that exists between SnO₂ deposited upon the inside surface of the glass envelope and a relatively small amount of oxidizable iron or other high iron content metal contained with the lamp. The high iron content metal can be included within the lamp in a variety of ways, as is known.

[0041] While a number of attempts have been made to determine experimentally the mechanism responsible for the surprising synergy between SnO₂ and oxidizable metallic iron, no completely satisfactory explanation has emerged. Nevertheless, the following hypothetical explanation is offered which is at least consistent with all of the known facts:

Step 1) Metallic iron oxidizes and dissolves in the acidic extraction fluid as Fe²⁺.

Step 2) The dissolved ferrous iron adsorbs upon the surface of the SnO₂-coated glass.

Step 3) Dissolved ionic mercury ions also adsorb upon the SnO₂-coated glass surface.

Step 4) The adsorbed ferrous iron and mercury ions interact on the surface of the SnO₂-coated glass to effect the oxidation of the ferrous iron (to Fe³⁺) with the corresponding reduction of ionic mercury to the essentially insoluble elemental form.

[0042] While there have been shown and described what are at present considered to be the preferred embodiments of the invention, it will be apparent to those skilled in the art that various changes and modification can be made herein without departing from the scope of the invention as defined by the appended claims.

Claims

1. A method for inhibiting mercury from leaching from a mercury vapor discharge lamp with an envelope wherein at least a part of said mercury is present as ionic mercury, said method comprising:

including within said lamp a quantity of SnO₂,

including within said lamp an amount of oxidizable iron, said amount of iron being at least 1 gram per kilogram of lamp weight.

2. The method of Claim 1 wherein said ionic mercury is present as mercury oxide.

3. The method of claim 1, wherein said quantity of SnO₂ is corresponding to a density of about 40 micrograms/liter.

4. The method of claim 1, wherein said quantity is included by depositing upon an interior surface of said envelope a coating of SnO₂.

5. The method of claim 4, wherein said quantity is corresponding to a density of about 40 micrograms/cm².

6. The method of Claim 4, wherein said fluorescent lamp is linear and the diameter of said envelope is at most 40 mm.

7. An arc discharge lamp including ionic mercury and at least one component comprised of oxidizable iron in amount of at least 1 gram per kilogram of lamp weight, the improvement comprising; an amount of SnO₂ within said lamp, said amount of SnO₂ being sufficient to prevent more than 0.2 mg/l of soluble mercury from forming when said lamp is pulverized and treated with a sodium acetate solution having a weight twenty times that of the pulverized lamp components and a pH of about 4.9.

8. The lamp of Claim 7 wherein said ionic mercury is present as mercury oxide.

9. The arc discharge lamp of claim 7 having an envelope with a diameter of at most 40 mm

10. The arc discharge lamp of claim 7 wherein said amount of SnO₂ being deposited as a coating of SnO₂ on an interior surface of said envelope.

11. The lamp of Claim 9 wherein said diameter of said envelope is at most 26 mm.

12. The lamp of Claim 7 wherein said coating of SnO₂ has a density of about 40 micrograms/cm².