Improved Vacuum Fusion Gas Extraction and Collection Apparatus

---

ACKNOWLEDGMENT

The authors are indebted to J. A. Becker, Surface Physics Department, for his contributions to the discussion of surface sorption and to M. E. Terry, Statistics Department, for guidance

and advice in making the computstions on an IBM 704. UTERANRE CITED

(1) Beach, A. L., Guldner, W.G., ASTM Spec. Tech. Puhl. 222 (1958).

R. S., F M , E. .I., Bslis, E. W., ANAL. CKEM. 27, 1632

(2) McDonald, (1955).

RECEIVED for rwiew A ~ r i l 3 .1959.pAccepted June 4, 1959. -Pit&hurgh Conference on Analytical Chemistry snd kp plied Spectrossopy, Gas-MetA S p p c + sium, Pittsburgh, Pa., Msrct; 3, 1959.

Improved Vacuum Fusion Gas Extraction and Collection Apparatus ARTHUR LENCH and G. S. MARTIN Ausfralian Defence Scienfific Service, Defence Standads Laboratories, Maribyrnang, Vidaria, Ausfrofia

F An improved gas extraction, collection, and measurement apparatus is described for use in the vacuum fusion method far the determination of gases in metals and alloys. It features a large crucible for high sample capacity. a high speed diffusion pump for rapid gas transport, provision for gas collection and measurement between 0.05 and 38 ml., an improved O-ring vacuum valve, solenoidoperated glass-mercury valves, and a butt-jointing method of assembly for rapid removal and .replacement of defective components

CARATUS

for the vacuum fusion method of determination of gases in metals has been described (1-.(,6,7). Csually, gas extraction, collection,meas uremrnt and analysis have been carried out in the one apparatus, but, occasionally, the gas analysis section hss bern designed separately. The different apparatus have embodied t i e same principles of design, but have dzered considerably in the application of these principles. This has been reflected in variations in the operating characteristics of the individual components. Consequently, when rapid determination of the gas contents of numbers of chromium metal and chromium alloy samples was required, the opportunity was taken to assess the importance of the various design factors in vacuum fusion apparatus. As a result of this asscssmenl, assisted by experience gained in the operation of an existing vacuum fusion apparatus, an apparatus was designed for gas extraction, collection, and measurement only, incorporating a number of impmvements. 1

DESCRIPTION OF THE APPAPATUS

The general appearance of the apparatus and the arrangement of tlle components are shown by Figwes 1 to E. 0826

AN4lITIuU CHEMISTRY

Figure 1 .

Vacuum fusion apparatus

Gar exhaction, cdldon, and rneoivrernnt mils, fmnt elevation

Furnace Assembly. The large crucible with a renewable insert bas external dimensions ot iength 2'/, inches and diameter la/, inches with a wali thickness of ' / 8 inch. TKe insert, which makes a sliding fit with the crucible, bas internal dimensions of depth li/, inches and diameter at the top l'/, inches, tapering to a/, inch a t the bottom.

A large crucible assembly has the dissdvantage of giving a somewhat larger blank than 6 small crucible. For example, the assembly shown in Fiiure 6, which includes itn outer eruciblc in addition to that described above, gives a blank of 0.15 mi. of gas per hour 6. after outgwing at 2 2 M ) O C. a t 1W" In comparison, a small crccibie i inch long and a/4 inch iz e h r n d diameter

has been stated to give a blank of 0.0% mi. per hour ( I ) . On the other hand, a number oi advantages are attached to the use of 3. large crucible, including .the ability FO treat a much greater totai weigh? of metal and/or alloy per run, suitakdit3for a wider range of sample weights, and a lower rate of increase of bath viscosity due to the reduction in the ratio of crucible-bath interface area to bath volume. The insert has a working capacity of 108 grams, with a wide margin of safety to accommodate any splashing that may occur. The outer crucible is reused a number of times. &use has the advantage oi reducing outgassing time, for the gas evoived during the Iaier stages of outgassing P-as been found to be aimost. whoily due &t

gure 2. Vacuum fusicm apparatus o", cdledim,

and meawurernent unk, rear elevation

tlic pyrolysis of decomposable materials in thc crucible assrmbly: Graphite powder made from discarded outer cnicihles is very suitable for insulation of the crucible assembly from the silica furnace nall because of the comparative absence of slowly decomposable ma*rial. Furnace Capacity. The large CNriblc roables a large metal bath to he cmployed. This greatly increases useful operating time for some analyses. Homrver, a t the higher operating temprratures required for satisfactory reaction rates of the more refractory metal oxides and nitrides, bath v i 6 cosity increases rapidly and eventually rrsults in unsatisfactory gas evolution. An efficient furnace operation schedule is therefore an important requirement, particularly for such metals as chromium, titanium, and thorium, with reaction temperatures from 1 7 0 0 O to ISM)" C. In addition, although the use of the large bath reduces the rate of volatilization of volatile metals, an rfficient furnace operation schedule will minimize gettering of the gas extracted from the last samples of the run. Analleis of the gas extracted from one metal sample before proceeding with the extraction of the next sample is, in consequence, not practicable for maximum efficiency, and some means of storage of the gas extracted or of samples of the gas, must he adopted. Of two alternatives that might he employed--storage in bulbs attached to the apparatus, or transfer t o small removable gas holders by a Toepler pump

-the latter was chosen to avoid undur complication of the apparatus. This permits replicate samples of the extracted gas to be taken easily, and the samples to be analyzed in a separate gas analysis apparatus by a second operator while the 6rst is carrying out a gas extraction of the next metal sample. With this arrangement, and using a %ram steel bath, a t least 16 samples of chromium metal, total weight not exceeding 40 grams, can be t m t e d a-ithout gettering being detected. The samples of the gas extracted are analyzed by a low pressure microanalytical method, using volumes within the range of O.M and 0.5 ml. The method adopted separates carbon dioxide by freezing with liquid nitrogen, utilizes Hopcalite to oxidize ca.rbon monoxide to carbon dioxide, removes hydrogen by diffusion through heated palladium, and leaves nitrogen as the residual gas. Gas Transfer and Collection System. A four-stage mercury diffusion pump with a capacity of 70 t o 80 micron-liters per second up to a maximum hacking pressure of 35 mm. of mercury (17, Figure 3), is used for rapid gas transfer betneeii the furnace section and the gas collection and measurement section. Rapid gas transfer is essential to prevent, or a t leist minimize, gettering by traces of volatilized nirtal, such as chromium, deposited on the cooler parts of the furnace and on the furnace head. The gas collection and measurement system consists of five calibrated volumes of 584, 1430, 3850, 4540, and 5900 ml. The volume of the McLeod

gage forms part oi thc first volunic, enabling small volumrs of gas to hr measured. Thr second volume includes the volume of t h e Torplrr pump. By locating the pump in this position, small volumes of gas can be rapidly transferred to t h r rcmovable gas holdrr for admission to the gas analysis apparatus. For example, three strokes of the pump will transfer 0.04 ml. of a total volume of 0.06 ml. of gas. The maximum pressure measurable on the McLeod gage is eqnivalent to a volume of 38 ml. of gas a t normal temperature and pressure. Heating Assembly. A moto? grnerator set with a frequency of 8300 cycles per second is used as a source of power t o the high frequency induction coil. Although efficient coupling hctween the coil and the crucible is more difficult t o obtain with a fixed frequency power supply than with a power supply covering a range of frequencies, such as a spark gap converter, this is more than offset by the advantage of much closer temperature control. More efficient coupling and, a t the same time, greater heat capacity has heen secured by placing the crucible within an outer crucible of 21/,-inch external diameter and %/*inch wall thickness (Figure F). Minus 200-mesh powdered graphite made from discarded crucibles gives satisfactory insulation of the crucible assembly from the silica furnace wall. Bare copper bus bars connecting the high frequency coil t o the capacitors need nd special cooling. The coil is rentralized by guide rods (6, Figure 3). Connection of Components. Standard borosilicate industrial glass parts, with flat-ground and buttrrssed flanged ends, manufactured by Quickfit Visible Flow, Ltd. (QVF), have been used in the apparatus as much as possible, either as such or after modification. The open end of the 3inch diameter silica furnace tube has been made to standard QVF pipe end diniensions by Thermal Syndicate, Ltd. Stainless steel parts used t o connect the diffusion pumps t o the glass piping of the apparatus have the glass attachment end made to standard QVF dimensions, with the addition of an annular recess in the flabground facr for the insertion of an O-ring. The butt.jointing method of assembly with QVF fittings, as used for QVF parts, has been employed, except that the xasket has been omitted. Instead, a thin layer of high melting' Apiczon wax has h e m applied to the outsidr of the joint to give a vacuum-tight scnl. In placrs rvhrrc a joint is not ulldrr stress, backing fliiogrs hare hrrn sinittcd. The rxrlusire usr of thr huttjoint througliout thr iippsratiis (rxcrpt for caps to tliv loading trra snd magncticaIIy rontrollrd stopprr side tubr) has rlirninatcd thr rlificultirs of rrmoral and Trp~:llr~m(:lit of ronipiinrnts assoointcvl nith thr iisc nf ronc and socket or Iia11 and cup joints. I t has enabled a simple method of assembllof componrnts in linr or a t right anglrs VOL. 31, NO. 10, OCTOBER 1959

1727

to one another to be adopted (Figure 3). The suitability of this method of assembly for high vacuum work i s shown by the leakagr rate of the gas evtraction and gas collection sections, which together equals 0.004 ml. of gas per hour. For ensier dekrtion of leaks, and removal and rcbplacement of defective components, the apparatus has been left unenclosed. The water jacket and high frequency coil are moved by a rack and pinion arrangement (Figures 2 and 4). Valves. Solenoid-operated glassmercury valves have been used t o repiace the stopcocks or U-type mercury cutoffs commonly found in vacuum fusion apparatus (22, Figures 3 and 5) eliminating the difficulties frequently associated with these two latter means of pressure control. The valves have been constructed so as to be normally held in the oper, position by the tension of a light spring. When closed, a maximum pressure differential of 10 mm. of mercury can be maintained. The rate at which the valve is closed and the depth to which the cap is drawn into the rnercury in the annulus are regulated by controlling the power applied to the solenoid by means of a variable autotransformer (33, Figurr 3). A stainless steel O-ring vacuum valve located between the furnace section and the evacuation pumps (16, Figure 3) replaces the 1-in'ch-bore glass stopcock ordinarily used. I t s use has eliminated difficulties frequently encountered in the use of large bore stopcocks, such as sticking in cold weather, gas absorption upon the large area of grease exposed, and atmospheric Irakage. Provision ex-

Figure 3, 1. 2. 3. 4.

5.

6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16.

17. 18. 19. 20. 21. 22.

Scaled diagram of front elevation

Base plate Rack support rod bearing Rack support rod Indub'on coil Induction coil support inductibn coil guide rods Woter jacket Crucible Furnace tube Water jacket support Furnace head Furnace head support Stopper operation assembly Vapor shield Glass prism O-ring vacuum valve Mercury diffusion pump Edwards Speedivac 2M4 Isolation valve, Edwards Philips gage power unit Philips gage head McLeod gage Glass-mercury valve n

23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42.

Toepler pump 2-liter gas hdder Pump to furnace connection Mercury diffusion pump, Edwards Speedivac 2M2 Water vapor trap (PtOs) Rotary backing pump, Edwards Speedivac 2550 Gas sampling assembly Suspension springs Water supply taps Woter drainage outlet tank Autotransformer, Zenith 100 R Autotransformer, Zenith 200 CMH Electrical switch and pilot light assembly Solenoid 230-volt 50-cycle ax. QVF joint 3-inch diameter QVF joint 2-inch diameter QVF joint 1-inch diameter Aluminum frame Dexian angle Sample tube Crucible insert

ists for fine adjustment of the aperturc and therefore of the degree of vacuum applied to the furnace section. Tiic possibility of ejection of the graph-b. insulation powder from around hie crucible on the sudden application of high vacuum is avoided. The valve is desrribed in detail elsewhere (5,. OPERATIONAL PROCEDURE

The furnace components are as. sembled, and the silica tube is attached to the furnacc head. Samples, material t o form the bath, and a steel pusher for nonmagnetic snrnplw are loaded into the sample holding t r w

r---~-=J 9

Figure 4. Scaled diagram of furnace and elevation 1728

ANALYTICAL CHEM!STRY

Figure 5. Scaled diagram of collection and eievatiort

( A i , Figure 4). The inlet to the tree is vlcsed. Evacuation of the furnace hy the rotary backing pump (28, Figure 3) is begun through the bypass h e . When the pressure has fallen t o about 0.05 mni. of mercury, the two-stage mercury diffusion pump .(26, Figure 3) is brought into operation, and evacuation continued until the pressure has fallen to 5 X 10-5 inm. of mercury. This pump has a (6apacity of 70 to 80 micron-liters per .econd up to a maximum backing presw r e of 0.35 mm. of mercury. The four-stage mercury diffusion pump (17: Figure 3) is then brought into op(.ration, th(J valve in the bypass line vlosed (16, Figurc: 3), and the line through the gas collection system ..pened. Meanwhile, the gas collection system has been held under vacuum during removal and replacement of the Eurnacc hoad and preliminary rvaruntion of t h r furnace unit. The motor generator is now switchd on, :mrl power supplitd to the coil. The cirucihle temlwratui,e is slowly raised to 2200" C'., so that the Dressurr in the furnace unit ~icwsnot tsxcred I micron is at. any tinit,. T!iv tc~rnporatul.c~ inaintaind at 2200' C. until thtx blank m t r has fallen to a satisfactory level;.e.. ahout 0.15 ml. of gas per hour. I t is then reduced to 1600" C., and 3J grams of steel are addrtl to thc vrucihle. Rat,h t,cnipcrature is raised Lo t h e reachon level and held thert. L:ntil the f'lank rate has :wc*omt. (:oni t a n t . A samplc of t h e blank gas c~vnlved is no\\-takc,ri ior anaiysis.

vi nwtai in the run. Finally, a check is mad(. of the volume of the hlmk gas and a sample taken for analysis. This concludcs the run. The opparatcs has been successfull\. wiployd in the determination of thc gas rontcnts of a uidr vsriety of chromium metal and chromiumalloy samples, using a steel bath as the reaction inediuni. Titanium nwtal aiic! titanium alloj-s havr also h e w siircessfully :tnal!-zcd using a stwl hath. ACKNOWLEDGMENT

Th(5 authors thank the Chief Sciciitist, Acstralian Defencr Scientific Scriice, Department of Supply, Melbourne, Australia, for pvrmission to puhlish this paper. INCHES O

U

Figure 6. Scaled diagram of furnace tube and crucible assembly

Thc systc.111 is evacuatcd and the first metal samplc addcd to the hath. T h r gas eiolvtd from the metal sample is i.ollec*tedand measured, and a suitable volume transferred by the Toepler puml) to a gas holder for analysis. Resi~lualgas is pumped away. T h r sequence of addition of a metal saniple, gas rollcction, mcasurtnirnt, sampling, and wacuation of rcsitlual gas from the system is rep(xatrd for ravh sample

LITERATURE CITED

(1) Booth, E., Bryant, F.J., Parker, A , , ' i d y s t 82, 50 (1057).

( 2 ) Horton, W. S., Brady, J., ASAI,. CHEM. 25, 1891 (1953). ( 3 ) MrI)onald, R. S.,Fagel, J. E., Jhlis, E. R.,!bid., 27, 1682 (1955). ( 4 ) hlclllet,t, 31. W., 7 ' r m . v . . l r r t . SOC. Jlelals 41, 870 (1949). (.i) llartiii, G. S., Lench, Arthur, J . Sci I n s t r . 36, 141 (1059) ( 6 ) Soloman, FI. .A,, Swcinth Itvport, Hcterogencity of Stwl Ingots Committee. Iron anti Stwl liist., p. 82, 1037. ( 7 ) \Valter, i ) . I., . A s . ~ I . . (:HEM. 22, 297 (19.50).

l t s c ~ ~ v sfor u review 1 )cwnil)rr 22, 1058. Arceptcd l l q y 4, 1959.

Chlorinated Insecticides in Surface Waters A. A. ROSEN and

F. M. MIDDLETON

Robert A. Toft Engineering Center, Department of Health, Education, and Welfare, Public Health Service, Cincinnuti 26, Ohio

*

in the ewluation of the quality of jurtace waters, a general monitoring xocedure that -an detect Q variety of mrnmon inse:+wdes In concentrations betow Those 'oxic to aquatic life is perwired. This result is accomplished 2y combinirg the processes of carbon Plter sampling, adsorption chroma:ograDhy, and infrared spectrophovnetrv. The above method has >roved sufficiently sensitive TO identify PSS than 10 p.p.b. of eight chlorinated insecticides in river water. A quantitative estimate is also obtainea. 'he method qas been tested with alarin, benzene hexachloride, chlordan, DDD, DDT, dieldrin endrin. and meth7xychlor.

'7 H L O R I N A ' ~ ~ , U ,>rganic. insecticides :'an finti their way :o surface m t t w ny the ninofi of rain water from :,re:itPc! crot,!rtnti or fnrw.: areas. h;: +he

direct application to bodies of w a t u ab in mosquito control, and hy the unintentional covering of surfacr waters during airplanr spraying of adjacrnt lar,d arcas. Thv most iwmmonly observed resuit of this pollution is tktniagc. to thc fish population--1,.g., 0:. Zainting t , h c L flesh ( I ) , by tlestroying thc food supply (S), or by direct poisoning. Fish kills have h e m causeri by insrc:tii*i(lr concentrations below 10 p.p.b. (4. 16). X disagrerable tastc i s imilartcd to drinking: n.attir by 20 p.p.b. of licnzene hexachloride (i7 ) . Thv quan+,ities of rhlorinatcd insecticidcs that havt. bwn encountt,rcd s o far in surfarr waters drawn upon ior puMic iuppiirs arc n r ~ l hcloa. l \vhat &rec.oiisidcwti tn h levels toxic to humans ( 7 ) . For ilSr in the mduation of watrr quality, appropriatc analytirai ii1pthds for chlorinated insecticic!rs havo iiistinctivc requirements. Although srnsitivities to 1 CJr 2 c.p.h. :m ttrsirahlt:. ' 4 0 1 . 31, NO. 10, OCTOBER 1959

* 1729