Carbon and composites · Workhorse

Graphite, pyrolytic graphite, glassy carbon

Cvalues describe: iso-graphite

Big science (accelerators, light sources, public fusion)Vacuum electronics & X-ray sourcesSemiconductor equipment & lithographyFusion, private

Why it wins

Low-Z, refractory, machinable: beam dumps and strike surfaces, rotating anode heat sinks, tube grids (PG), crucibles, heater elements. The pre-tungsten fusion wall (CFC) that tritium retention retired.

Why not the alternative

Metals activate and sputter high-Z into beams and plasmas; but carbon dusts and holds gas, hence the fusion exit.

Watch out

Porous grades outgas for days: bake long, handle clean.

Properties

The values below are candidate: compiled from the sources named, not yet individually validated. Provisional provenance: ITER and JET ITER-like wall programme; CERN Accelerator School, vacuum and materials proceedings.

Wet cleaning

Recipedo not wet-clean porous grades: solvent enters the pores and never leaves; degrease dry-handled parts and vacuum fire 1000–1500 °C
Forbiddenaqueous cleaning of iso-graphite
Limitopen porosity: the bake, not the wash, is the cleaning step

Vacuum and outgassing

Outgassing, unbaked (10 h)1e-7 mbar·L/s/cm²
Outgassing, baked1e-10 mbar·L/s/cm²
Vapour pressurenone, but it holds gas in its porosity for a very long time

Temperature

Bake, assembled450 °C
Vacuum degas1500 °C
Braze / H2 firing2000 °C
Metallurgical limitsublimes above 3000 °C in vacuum

Thermal

CTE4.5 iso; 0.5 in-plane pyrolytic ppm/K
Thermal conductivity100 iso; 300–1900 in-plane pyrolytic W/m·K
Specific heat710 J/kg·K
Emissivity0.8–0.95
Melting / softening3650 sub. °C

Mechanical

Strength40–90 flex MPa
Elongationnil
Young's modulus10–15 iso GPa
Hardnesssoft, machinable
Density1.8 iso, 2.2 pyrolytic g/cm3

Electrical and magnetic

Relative permeabilitydiamagnetic
Resistivity1000–1500 µΩ·cm iso

Engineering

Corrosioninert except to oxygen hot
Joiningmechanically clamped; graphite-to-metal brazing is specialist
Process notesparticulate generation is chronic; glassy carbon is the pore-free option
Availability and costToyo Tanso and others, in stock

Brazing

Sourced pairings

WithFillerAtmosphereExpansion gap at set
Oxygen-free copper C10100 / C10200 / Kovar (FeNiCo) / Nickel and nickel plating / Austenitic stainless 304L / 316L / 316LN ESR / Titanium Gr 2 / Ti-6Al-4V / refractories (unspecified)Ticusilvacuum 1e-5 mm Hg or inert gas1.04 % at 780 °C
Oxygen-free copper C10100 / C10200 on the outside
over budget
Alumina 94 to 99.8 percent
one built object
multistage depressed collector of a traveling-wave tube, NASA Lewis Research Center: isotropic graphite electrodes joined to alumina insulators
Ticusilvacuum furnace0.14 % at 780 °C
Alumina 94 to 99.8 percent on the outside
undetermined
Alumina 94 to 99.8 percent
one built object
multistage depressed collector of a 500-W continuous-wave 4.8 to 9.6 GHz traveling-wave tube, NASA Lewis Research Center: graphite electrodes brazed to alumina insulators
BCu-1hydrogen atmosphere furnace0.39 % at 1084.62 °C
Alumina 94 to 99.8 percent on the outside
undetermined

Preparation

  • none
  • the Ticusil foil placed in direct contact with the graphite surface on a tapered joint interface, with a slight pressure applied at the interface during brazing to obtain a void-free joint. On this joint the alumina is UNMETALLIZED.
  • the graphite is metallized with a molybdenum, manganese and silica powder slurry brush-painted, dried and fired at 1600 C in dry hydrogen to form carbides with the graphite, the alumina ring is metallized on both its inner and outer circumferential surfaces, a thin copper coating overlaid on the graphite serves as the filler and protects the metallization from oxidation. Both brazes, Kovar to alumina and alumina to graphite, are made in a single operation.

Compatibility

What documents state about this material faced with a filler metal, without the other member of the joint. A statement that names the exact filler comes first. A statement about a whole AWS class comes after it: a class covers dozens of alloys, so it is the weaker of the two, and the corpus never turns one into the other.

  • does not wetBCu-1
    Schwartz states that it is essentially impossible to braze graphite with copper filler metal AWS BCu-1, because no wetting occurs.
    Conditions and source

    Book page 158, prose, in the passage that opens a new brazing technique for joining graphite to itself or to metals such as molybdenum, tungsten or copper. The sentence names no partner and no condition of atmosphere or temperature. The same passage goes on to say that an iron foil insert makes the joint possible, which is a different technique and is not recorded here. This entry was identified at lot COMPAT-2 and could not be written then, because the enumeration carried wets and not its opposite.

    [A]M. M. Schwartz, Brazing, 2nd edition, ASM International, 2003, chapter 4, Base Metals and Base-Metal Family Groups, DOI 10.1361/brse2003p063

    Chapter 4 of the second edition, book pages 63 to 162 in the copy read, which treats base metals family by family and states metallurgical reactions, atmosphere limits and filler selection guides. It is a handbook chapter, not a qualification record. The page of each statement is carried by the entry that uses it, never by this registry line.

  • listedTiCuNi
    Ticuni is listed for brazing carbon to stainless steel, to copper, to nickel, to titanium and zirconium alloys, and to carbon itself.
    Conditions and source

    Book page 153, table 4.28, Selection guide to filler metals for brazing metal-ceramic joints, carbon row. Source Ref 9 264 as cited by Schwartz, selection guide not a qualification. The identification rests on the commercial name alone: table 4.27 of the same chapter gives Ticuni as 70 Ti, 15 Cu, 15 Ni, while the corpus entry ti-cu-ni carries no composition to compare. The carbon row is marked with note (b), Includes alloys, while only note (c), Graphite and diamond, says what carbon means in this table. The corpus reads the carbon row under note (c) and records here that the document itself carries the other marker. The table does not list Ticuni against carbon and low-alloy steel, nor against refractory metals or tungsten carbide, where it lists Ticusil alone.

    [A]M. M. Schwartz, Brazing, 2nd edition, ASM International, 2003, chapter 4, Base Metals and Base-Metal Family Groups, DOI 10.1361/brse2003p063

    Chapter 4 of the second edition, book pages 63 to 162 in the copy read, which treats base metals family by family and states metallurgical reactions, atmosphere limits and filler selection guides. It is a handbook chapter, not a qualification record. The page of each statement is carried by the entry that uses it, never by this registry line.

  • listedTicusil
    Ticusil is listed for brazing carbon to stainless steel, to copper, to nickel, to titanium and zirconium alloys, and to carbon itself.
    Conditions and source

    Book page 153, table 4.28, Selection guide to filler metals for brazing metal-ceramic joints, carbon row. Source Ref 9 264 as cited by Schwartz, selection guide not a qualification. Table 4.27 of the same chapter gives Ticusil as 68.8 Ag, 26.7 Cu, 4.5 Ti, which matches the corpus entry. The carbon row is marked with note (b), Includes alloys, while only note (c), Graphite and diamond, says what carbon means in this table. The corpus reads the carbon row under note (c) and records here that the document itself carries the other marker. The same row also lists Ticusil against carbon and low-alloy steel, tool steel, nickel and cobalt alloys, refractory metals and tungsten carbide, five families the corpus does not carry under an identifier, and those cells are not written.

    [A]M. M. Schwartz, Brazing, 2nd edition, ASM International, 2003, chapter 4, Base Metals and Base-Metal Family Groups, DOI 10.1361/brse2003p063

    Chapter 4 of the second edition, book pages 63 to 162 in the copy read, which treats base metals family by family and states metallurgical reactions, atmosphere limits and filler selection guides. It is a handbook chapter, not a qualification record. The page of each statement is carried by the entry that uses it, never by this registry line.

Chemical affinity

What documents state about a BINARY SYSTEM, one element of a filler metal faced with one element of this material. An affinity is not a verdict on a joint: a joint brings many binaries together at once, some helpful and some harmful, and what follows is the list of those the corpus can name. The count below is not written anywhere. It falls out of the crossing of this material's elements with the composition of every filler the corpus holds, so it grows on its own the day a binary enters the corpus.

Elements of this material: C, from the chemical name.

interfacial reaction that wets9 fillersTi against C1 source+1 statement

The two elements react at the interface and the product of that reaction is what the rest of the filler wets. This is the mechanism that makes active metal brazing work, and it is the only one of the six that a joint wants.

On a carbide or a carbon, titanium in the filler forms titanium carbide at the interface, and that layer is what the filler wets.wantedTi against C

9 fillers: 35Au-62Cu-2Ti-1Ni, 48Ti-48Zr-4Be, 49Ti-49Cu-2Be, 56Zr-28V-16Ti, 71.5Ti-28.5Ni, Cusil-ABA, Cusin-1 ABA, Incusil-ABA, Ticusil

Conditions and verbatim

Printed page 862, closing paragraph on other nonoxide ceramics. The article states that titanium nitrides and carbides are known as the interfacial reaction products for titanium-containing active metal brazes. It immediately qualifies the reach of that statement, and the qualification is carried here rather than dropped: the wettability of such carbides and nitrides with metal liquids may be influenced by their stoichiometry, citing Ref 109. The article speaks of carbide ceramics. This corpus also holds two carbons, graphite and diamond, whose element is the same C, and the entry is served for them too because the reaction named is the one between titanium and carbon.

Source[A] Joining, in Engineered Materials Handbook Desk Edition, M. M. Gauthier editor, ASM International, 1995, pages 846 to 864, DOI 10.31399/asm.hb.emde.a0003056

Source conditionsHandbook article, consulted as a PDF whose printed page numbers run 846 to 864. The article is itself a digest of the Ceramics and Glasses volume 4 of the Engineered Materials Handbook, so its statements are second-hand summaries of the papers it cites by Ref number. The Ref numbers are kept in the conditions of each entry, because a reader who wants the primary measurement needs them.

Outgassing

0 sourced measurements for this card, 5 declared absences. A rate means nothing without the pumping time it was read at, so each one states it or declares it missing. The whole base, filters included, is at /tools/outgassing/.

  • Vitreous Carbon, Beckwith Carbon Corp., Van Nuys, California. Monolayers evolved on outgassing to 2000 K, H2 0.2, CO 0.08, CH4 much less than 1, C2H6 much less than 1. Adsorption probabilities at 300 K, H2 5e-7, CO 5e-4, O2 about 3e-9, N2 less than 5e-7.A quantity of gas evolved in monolayers is not a rate per unit area. It has no time in it, and it cannot be converted into the canonical unit of this corpus by any factor. The corpus carries no field for a quantity of gas evolved on heating, so the numbers are recorded verbatim in the text of this declaration and no value field is filled. Converting them would invent a pumping time that the source never published.
  • Pyrolytic Graphite, Super Temp, Santa Fe Springs, California. Monolayers evolved on outgassing to 2000 K, H2 1.7, CO 0.4, CH4 0.25, C2H6 0.12. Adsorption probabilities at 300 K, H2 5e-6, CO 1e-5, O2 not measured, N2 less than 1e-7.A quantity of gas evolved in monolayers is not a rate per unit area. It has no time in it, and it cannot be converted into the canonical unit of this corpus by any factor. The corpus carries no field for a quantity of gas evolved on heating, so the numbers are recorded verbatim in the text of this declaration and no value field is filled. Converting them would invent a pumping time that the source never published.
  • Grafoil, Union Carbide, New York, New York. Monolayers evolved on outgassing to 2000 K, H2 20.0, CO 1.2, CH4 3.0, C2H6 2.0. Adsorption probabilities at 300 K, H2 1.5e-3, CO 1e-2, O2 not measured, N2 less than 1e-5.A quantity of gas evolved in monolayers is not a rate per unit area. It has no time in it, and it cannot be converted into the canonical unit of this corpus by any factor. The corpus carries no field for a quantity of gas evolved on heating, so the numbers are recorded verbatim in the text of this declaration and no value field is filled. Converting them would invent a pumping time that the source never published.
  • All carbon triode gauge, collector, grid and supports in Grafoil and pyrolytic graphite. The source states that after the normal system bake-out of 300 C for 24 hr the system pressure fell below 1e-10 Torr, indicating that the graphite gauge at room temperature had a nonmeasurable outgassing rate, and that when operated as a gauge with 0.3 mA emission, 150 V filament to grid and 37 W filament power, the gauge had an outgassing rate of about 5e-9 torr l/cm2.The printed unit has no time in it. A rate per unit area is torr l per second per cm2 and the source prints torr l/cm2, so the reading is certain and the unit is not. The rule of this campaign is that a value whose unit is not certain is not written, and this one is not. The operating condition is also not a material condition, since the source expects the 37 W to raise the grid and collector to about 750 K, 130 K above the bake-out temperature.
  • Graphite grade R 8500 from SGL Carbon, isostatically pressed. The thesis measured one as received graphite sample and one sample given a heat treatment at 1000 C for 2 hours in a vacuum furnace, both on the throughput system, four samples of 75 mm by 50 mm and 10 mm thick cut by electroerosion with the block immersed in water and then ultrasonically cleaned in an ethanol solvent, 100 cm2 of surface area each. The thesis states that the untreated sample raised the system pressure by an average factor of 140 at 30 C, that the heat treated sample initially showed 12 to 20 times lower pressures at 30 C, and that the results at 200 C for the untreated sample are omitted because the test had to be stopped for a power cut.The outgassing rates of both graphite samples are published as curves against time in figures 32 and 36 and are tabulated nowhere. Reading a number off a plotted curve is not reading a published value, so no value field is filled.

Sources: S-NASA-CR-2101, S-CERN-THESIS-2019-061. Full citations and conditions at the base and in /data/outgassing.json.

Grades

Iso-graphite IG-110 / POCO EDM classfine grain machinable stock, nuclear heritage
Pyrolytic graphiteanisotropic: grid blades, anode heat spreaders
Glassy carbon (Sigradur)gas-tight, non-dusting crucibles and liners
CFC (NB31 class)legacy divertor composite, retired for tritium retention
Sources · 2
  1. JET ITER-like wall, carbon to metal transition
  2. Toyo Tanso iso-graphite data

Used by

Tools

Essays

← The Matter