W h e n a steel sample is fused in a graphite crucible under v a c u u m at temperatures of 1 6 0 0 - 1 7 0 0 ° C . all oxygen-containing compounds are reduced to oxides of carbon, nitrides a n d hydrides are decomposed, and a n y dissolved gases are liberated. If t h e gases are removed quickly by a fast mercury diffusion p u m p a n d stored in a reservoir, t h e interference due t o manganese and aluminum vapors, which condense as active metal films on t h e colder p a r t s of t h e furnace and then reabsorb t h e gases, can be reduced to a minimum. T h e gases are then analyzed b y cir
culating through absorption t u b e s or t r a p s a n d noting t h e changes in pressure or the increase in weight of t h e absorption t u b e .
T h e present m e t h o d was developed b y J o r d a n ( 3 9 , 1 2 4 , 1 2 5 ) a n d his associates at the Bureau of S t a n d a r d s from 1 9 2 5 t o 1 9 3 1 . T h e history of t h e development from t h e original work of Walker a n d Patrick in 1 9 1 2 to the time ( 1 9 3 1 ) of J o r d a n ' s last article is summarized on pages 3 7 5 - 3 7 7 of reference ( 3 9 ) and additional references will be found in t h e footnotes on these pages. Subsequent improvements were m a d e b y
* A complete leak detector employing this principle is manufactured by the Radio Corporation of America, Harrison, New Jersey.
Chipman a n d F o n t a n a (4) in 1935 a n d b y N a u g h t o n a n d Uhlig (29) in 1943. Various means of simplifying the a p p a r a t u s and reducing the time required for determination of oxygen have been developed by Derge (76) and b y Alexander et al. (42).
N o s t a n d a r d procedure has been adopted for t h e v a c u u m fusion method and the literature on various modifications of t h e technique is already quite extensive. Complete details of t h e a p p a r a t u s a n d pro-cedures used in England have been appearing regularly in a series of reports b y t h e Oxygen Sub-Committee of t h e Joint C o m m i t t e e on t h e Heterogeneity of Steel Ingots, published b y the Iron a n d Steel I n s t i t u t e (28 Victoria Street, London).* Most of t h e developments in G e r m a n y appear in t h e Archiv fur das Eisenhuttenwesen in articles b y Hessenbruch a n d Oberhoffer (111), Diergarten (80), Meyer a n d Willems (146), Willems a n d co-workers (215). I n t h e United States numerous papers will be found in Metals Technology a n d t h e Transactions of the American Institute of Mining and Metallurgical Engineers as well as t h e Bureau of S t a n d a r d s publications a n d t h e Analytical Edition of Industrial and Engineering Chemistry (now Analytical Chemistry).
Oxygen has been t h e principal gaseous element of interest in steel analyses b u t hydrogen has also been t h e object of special study. Holm and T h o m p s o n (21) in 1941 recommended a low t e m p e r a t u r e (400-800°C.) extraction m e t h o d instead of fusion for hydrogen. T h e y found t h a t t h e diffusion of oxygen a n d nitrogen through t h e solid steel is so slow t h a t all of t h e hydrogen can be extracted without m u c h contamina-tion from t h e other gases. However Moore (149) a n d his co-workers h a v e studied t h e r a t e of evolution of hydrogen as a function of tempera-t u r e a n d conclude tempera-t h a tempera-t tempera-t h e extempera-tractempera-tion is notempera-t always completempera-te, even atempera-t
1000°C. A symposium on t h e determination of hydrogen in steel is reported in t h e Trans. Am. Inst. Mining Met. Engrs. (Jan. 1945) where a simplified fusion a p p a r a t u s for hydrogen is described b y Derge a n d co-workers (77). Carney, C h i p m a n a n d G r a n t (3a) h a v e described a
"tin-fusion m e t h o d " for the determination of hydrogen in steel in which t h e sample is dropped into a " l a k e " of molten tin at 1150°C.
4.2. Apparatus
A diagram of t h e a p p a r a t u s designed b y Vacher a n d J o r d a n is repro-duced in Fig. 14. C h i p m a n a n d F o n t a n a used a similar system b u t modified t h e furnace (A) t o include graphite radiation shields and a graphite " s p l a s h " plug. T h e y also chose t o introduce t h e samples b y suspending t h e m on a fine nickel wire wound on a stainless steel windlass rather t h a n using t h e Oberhoffer sample loading device (illustrated in
* First report 1937, 75 pages; second report 1939, 15 pages; third report 1941.
the article by Vacher and J o r d a n (39)). Alexander et al. (42) found t h a t t h e radiation shields and splash plug can be eliminated if t h e crucible is imbedded in graphite powder and a preliminary " l a k e " of molten metal is formed in t h e b o t t o m of t h e crucible before running t h e samples.
T h e crucible a n d sample are usually heated b y a high-frequency induction unit b u t Meyer and Willems (146) a n d Newell (158) have described graphite spiral furnaces (21, 22, 59). T h e induction unit
FIG. 1 4 . Vacuum fusion apparatus of Vacher and Jordan.
should have a capacity of about 30 k v a for fusion of 20-g. steel samples, about 20 kw being used to outgas the furnace during the blank while about 10-15 kw is sufficient to fuse the samples (22). However, only about 3 k v a is required for bringing the samples t o red heat (800°C.) in the extraction of hydrogen by the method of Holm a n d Thompson (21).
Alexander et al. (42) used a 5-kva (output) power oscillator a t 550 kc mean frequency for 5-g. steel samples.
T h e furnace t u b e is m a d e of fused silica with a wall thickness of 2 to 5 m m . T h e crucible, shields, and " s p a t t e r p l u g " (or t h e insulating powder) are turned (or filed) from Acheson graphite. T h e head piece is m a d e of brass or stainless steel and is sealed t o t h e furnace t u b e with a
vacuum wax, such as picein, Cenco Sealstix, M y v a w a x , or Apiezon W.
T h e connection t o t h e p u m p is usually located in t h e head a n d should be 20 m m . or more in diameter, if possible, similar to t h e t u b e used on Raine's carbon-spiral furnace (22) r a t h e r t h a n t h e narrow t u b e s illus
t r a t e d in the articles b y J o r d a n a n d b y Chipman a n d F o n t a n a (4). An optical pyrometer is used t o view t h e specimen through a small Pyrex glass window sealed in t h e head. Guldner and Beach (106a) have described an all-glass furnace.
T h e rest of t h e system is constructed of Pyrex glass,* although t h e mercury vapor p u m p m a y be of metal. T h e diffusion p u m p s should be equipped with t h e N a u g h t o n a n d Uhlig condenser (see Fig. 8) for pro
viding a sharp gas-vapor b o u n d a r y in t h e fore v a c u u m (29). T h e y m u s t be capable of compressing a b o u t 10 liter-millimeters (13 cc. at atmospheric pressure) of gas into t h e main reservoir. T h e general formula for storage of gas at room t e m p e r a t u r e (300°K.) is PV = 19000 m/M, where Ρ is t h e pressure in millimeters in t h e reservoir of volume V liters when filled with m grams of a gas having molecular weight Μ. T h u s if an 8-g. only one 700-ml. reservoir when analyzing 20-g. samples, a larger volume unnecessarily increasing t h e time for clean-up of oxidized gases (39). I n cases where t h e gas evolution would m a k e t h e pressure in t h e reservoirs exceed t h e forepressure limit of t h e p u m p , t h e y recommend circulating t h e gases t h r o u g h t h e absorbents for short intervals without waiting t o complete t h e extraction from t h e furnace. This is possible only when t h e gravimetric procedure is used (39). C h i p m a n and F o n t a n a adopted t h e practice of removing p a r t of t h e stored gas, after reading t h e total storage pressure, b y passing it out through t h e mechanical fore p u m p a n d t h e n analyzing only t h e remainder in cases where considerably more gas was evolved t h a n expected (4).
T h e McLeod gauge is usually specially constructed for this work with three or more scale ranges. Vacher a n d J o r d a n used a gauge covering t h e range 20 t o 0.001 m m . I n general t h e gauge should cover t h e range from 0.0001 m m . t o slightly above t h e limiting forepressure a t which t h e diffusion p u m p breaks down. A Pirani gauge m a y be included if desired
* A complete Vacuum Fusion Apparatus is manufactured by Distillation Products, Inc., Rochester, N.Y.
to aid in leak hunting and t o give a continuous indication of t h e lower pressures, b u t the analysis should be done with the McLeod.
The analytical train consists of a reservoir with inlet t u b e reaching to the bottom, a U-tube filled with fresh copper oxide heated b y an electric furnace followed b y a tube containing anhydrous magnesium perchlorate (Anhydrone) or phosphorus pentoxide for absorbing water vapor, an absorption t u b e for carbon dioxide filled with sodium hydroxide on asbestos (Ascarite) backed by a drying agent such as Anhydrone, a mercury diffusion p u m p with N a u g h t o n and Uhlig condenser for circulat
ing the gases, and the McLeod gauge for measuring t h e pressure at t h e reservoir and other points in the train. Alexander et al. (42) prepared the cupric oxide reagent by impregnating porous beryllia fragments with saturated copper nitrate solution under vacuum and t h e n heating in a muffle furnace at 450°C. until evolution of nitrogen oxides ceased, reduc
ing t o copper by a stream of hydrogen at 350°C. and t h e n reoxidizing b y a stream of oxygen at 400°C. Stopcocks and connecting tubes are pro
vided so t h a t t h e gases can be circulated through any one of a n y com
bination of absorption tubes at will and so t h a t a n y p a r t of the train m a y be isolated, removed from the system, or evacuated b y the mechanical p u m p without disturbing the pressure in t h e rest of the system. T h e absorption tubes m a y be attached b y ground glass joints and in addition t o the large tubes used for m a n y consecutive runs b y t h e volumetric method a set of small absorption tubes with matching ground glass joints should be provided for checking t h e volumetric method b y t h e procedure of weighing the absorption t u b e on an analytical balance.
I n t h e volumetric method t h e storage p a r t of t h e system must be accurately calibrated while the diffusion p u m p is operating to determine the volume included b y the reservoir, the McLeod gauge (with mercury at t h e cut-off), and the connecting tubes t o the stopcocks and the bound
ary between vapor and gas in the forepressure end of the diffusion p u m p as determined b y t h e N a u g h t o n and Uhlig condenser. An estimate is first m a d e of this volume from the dimensions of the p a r t s involved. A known volume of purified nitrogen at atmospheric pressure is t h e n admitted from a calibrated burette. T h e pressure before and after admitting the nitrogen is read on the McLeod gauge, and t h e volume of t h e storage system is t h e n readily calculated (4). T h e total volume can also be easily determined if the volume of some p a r t of the system is accurately known, such as t h e volume of t h e reservoir t o t h e stopcocks, or the volume of t h e bulb and capillary of the McLeod. Nitrogen is trapped in the known volume after measuring the pressure and the rest of the storage system is then evacuated. T h e trapped nitrogen is then expanded into the whole storage system and the pressure measured again.
4.3. Procedure
T h e references m u s t be consulted for details of t h e procedure as well as a further description of t h e a p p a r a t u s . A general idea of t h e steps involved in t h e fusion m e t h o d can be obtained from t h e following outline.
(1) Preparation a n d loading of clean samples
(2) Evacuation of system a n d degassing of furnace (1900°C.) (3) Blank run (1600°C). Check base pressure a n d rate of rise (P0) (4) Fusion of sample (1600°C). Reservoir pressure (Pi)
(5) Absorption of carbon dioxide (may be omitted) (Pi — P2) (6) Oxidation of hydrogen and carbon monoxide b y hot copper oxide (7) Absorption of water vapor in Anhydrone ( P2 — P3)
(8) Absorption of carbon dioxide in Ascarite ( P3 — P4) (9) Residual gas computed as nitrogen (P4)
(10) E v a c u a t e , check base pressure and rate of rise
T h e warm extraction m e t h o d for hydrogen differs from t h e above procedure in omitting t h e degassing of t h e furnace at high t e m p e r a t u r e a n d in heating t h e sample t o only 800°C. while t h e absorption train con-sists only of the copper oxide t u b e a n d furnace followed b y a t u b e of anhydrous magnesium perchlorate (Anhydrone).
Derge (76, 77) prefers t o analyze t h e gases b y fractional freezing rather t h a n b y absorption a n d reads t h e pressure on a butyl p h t h a l a t e manometer.*
4.4- Limitations and Accuracy
T h e v a c u u m fusion m e t h o d yields only total oxygen and nitrogen and gives no information about t h e actual compound present in t h e steel. A fractional m e t h o d suggested b y Reeve (176) is useful for estimating the different types of oxide in weld metal. Manganese and aluminum inter-fere with t h e determination of oxygen b y distilling on to the colder p a r t s of t h e furnace, forming a film t h a t readsorbs the oxygen. Previous investigators h a v e published figures of t h e order of 0 . 5 % for t h e a m o u n t of these metals t h a t can be tolerated in the crucible and t h e samples under certain conditions, b u t C h i p m a n a n d F o n t a n a maintain t h a t the error can be avoided b y providing high speed removal of t h e evolved gases and b y cleaning t h e furnace and using a new crucible after each analysis in which considerable manganese or aluminum was present (4, 22).
T h e precision a n d accuracy of t h e vacuum-fusion m e t h o d has been well established b y Chipman and F o n t a n a a n d b y cooperative analyses
* The Derge apparatus is available from the Central Scientific Co., Chicago, 111.
of samples exchanged between t h e Bureau of S t a n d a r d s a n d laboratories in England, Germany, and t h e United States. Vacher a n d J o r d a n indicate values for oxygen and nitrogen reproducible t o within 0.002%.
Chipman and F o n t a n a claim a precision of 0 . 0 0 1 % a n d a probable accuracy of ± 0 . 0 0 2 % . Hydrogen b y t h e fusion m e t h o d is not very reliable unless the content is high, b u t t h e warm extraction m e t h o d gives excellent results as indicated in t h e paper b y Holm and Thompson (21).
T h e question of t h e complete reduction of t h e oxides and nitrides present in steels has been thoroughly investigated a n d lengthy discussions will be found in the principle references already cited.
4.5. Extension to Nonferrous Metals
T h e vacuum extraction of gases from molten metals is limited only by t h e tendency of various metals to sublime t o colder p a r t s of t h e furnace a n d t h e n readsorb p a r t of t h e gas. Winterhager (217) designed a long furnace t u b e surrounded by a movable furnace coil and with connections at either end to t h e p u m p so t h a t by means of valves t h e gases could be removed in either direction. W h e n t h e metal sublimed t o an adjacent colder portion of t h e tube, the coil was moved to this spot and t h e metal resublimed. Bobalek and Shrader (57) found t h a t with fast pumping and rapid induction heating only one resublimation is necessary for extracting hydrogen, oxides of carbon, and nitrogen from magnesium alloys.
Winterhager employed his method for extracting gases from magnesium and zinc. Walter (39a) has described a modified v a c u u m fusion a p p a r a tus for the determination of oxygen in titanium. P r e s u m a b l y almost a n y metal or alloy can be analyzed for gaseous impurities when a suffi
ciently powerful induction furnace a n d fast p u m p i n g system are used.
T h e extent of t h e reduction of oxides and nitrides present m u s t be sepa
rately investigated for each metal b u t in general these compounds are completely reduced a t very high v a c u u m s a t t e m p e r a t u r e s of less t h a n 2000°C.
5. D E T E R M I N A T I O N O F C A R B O N B Y T H E L O W - P R E S S U R E C O M B U S T I O N M E T H O D