General purpose gas loop


General purpose gas loophttps://pubs.acs.org/doi/pdfplus/10.1021/ed047p653by PA Hersch - ‎1970which should be of inter...

1 downloads 91 Views 5MB Size

Paul A. Hersch 910 Franklin Terrace

Minneapolis, Minnesota 55406

General Purpose Gas Loop

In the many laboratory operations that make use of an inert carrier gas, it is customary to let that gas run to waste. Its reclamation for continuous use on a laboratory scale is probably regarded not worth the trouble, though i t can reduce the need for the transport and storage of cylinder gases and the hazards of handling them. This paper shall describe means for returning the effluentof a gas train to immediate re-use, which should be of interest to schools, small and ambulant laboratories, and where economy of space and weight is a consideration. The system described may also be recommended for gas volumetry and kinetic studies of reactions with a gaseous partner or product. Emphasis in this paper is on moist oxygen-free nitrogen as the carrier gas or reaction atmosphere, circulating at a chosen, constant rate, and continually purified from leakage oxygen and impurities picked up on its path. Gas chromatography, metabolic studies, and polarography are typical fields of application.

the nitrogen leaving the pump bubbles through the column of oxygen-absorbing liquid and returns to the pump. The rest, Fg ml/min, traverses the needle valve, the reactor, and the flowmeter, whence it returns to the pump. The flow rate F R is determined by the hydrostatic head h, of the absorber, any hydrostatic head hR in the reactor, the flow resistance of the needle valve, r,, and any flow resistance r Rof the reactor FR ( h -~ hn)l(rv m) The rate F , can be adjusted by adjusting rv. Any fluctuation in pump output affects Fz but not F R . Thus the liquid head of the absorber has a manostatic effect, stabilizing F,. A solid bed absorber would not act as a manostat. A second wash bottle, in series with the first, as shown in Figure 2, reinforces the stabilizing action, besides assisting the purification. The liquid in the second washer need not necessarily be the same as in the first one.

-

+

Basic Design

The essential requirements of the loop system are a gas recirculatmg pump, a wash bottle with an oxygenabsorbent solution, a needle valve, and a flowmeter. The loop is connected to the system under study which shall be called "the reactor." The intake side of the pump communicates with the room air through a Ushaped tube containing liquid in its lowest, capillary part. This part shall be called "the vent" (Fig. 1). In the steady state of operation, when the loop and reactor contain essentially nitrogen as the only gas, the pump dispenses as much nitrogen as it aspirates, and the vent liquid is quiescent. A portion, Fz ml/min, of Figure 2.

Gar loop for strict anaerobicity.

The small pool or slug of liquid in the vent is a mobile seal permitting the intake side of the pump to stay under ambient atmospheric pressure. All other points of the loop and reactor are a t a higher pressure. This prevents in-flow of atmospheric oxygen through defective connections; not, however, ingress by diiusion. Ingress of oxygen is also possible if the pump itself is not tight, aspirating external air in addition to internal nitrogen. Before start up, the loop and newly attached reactor contain air, say 11. Upon starting the pump, the 0.21 1 oxygen in that air is progressively removed and replaced by 0.21 1new air. Of this, 0.21'1 of oxygen is in turn removed, and so forth. Eventually all the gas in the system is nitrogen (plus inerts and water vapor). The total volume of oxygen removed adds up to figure 1.

Basis design for gas Iwp.

Volume 47, Number 9, September 1970

/

653

This process can be followed by watching the inward bubbling motion of the vent liquid. Rapid at first, it slows down after a few minutes, and ceases eventually. A continuing inward motion indicates a leak outwards from the system at some point of the loop or reactor. Conversely, escape of gas through the mobile seal indicates a leak of air into the system, owing to defects of the pump. Naturally, temperature changes in the loop or reactor, or parts thereof, also actuate the bubbler. The moment the pump is switched off, some gas is seen to escape through the vent. This is an extra mass of gas that has been drawn into the circulating atmosphere by the pump to build up pressure above barometric. When circulation ceases, the pressure differentials disappear and the extra mass is released. Pump

also be an inert electrode delivering oxygen into the outer atmosphere. In the latter case no chemical would be consumed. However, each liter of air needs as much as 56 amp-min "worth" of hydrogen for the removal of the oxygen. Start-up from air would thus be time-consuming, and the dissipation of heat from the electrolyzer and catalyst would present problems. Connections

Short sleeves of polyvinylchloride tubing make convenient connections between the various components. They are much preferable to rigid conical ground glass joints or to spherical joints with their clamps. Oxygen is measurably soluble in PVC, as in other plastics, therefore some slight contamination of the loop gas by atmospheric oxygen takes place continually. Hourever, with an absorber in a circulating system, the effect is not cumulative.

A simple membrane blower provided with an inlet leading directly into the membrane chamber sewes the purpose. A well-engineered, inexpensive fishbowl blower, traded under the name "Star Pump," can be so modified, with no leakage either inwards or outwards. (The blower is made by Etablissements Piot & Tirouflet, 89, rue de la Croix Nivert, Paris XV".) Any other small blower may be converted to a gas-tight circulator by placing it into a desiccator or similar container, with two short metal pipes seated in a rubber stopper as inlet and outlet. The inner end of the outlet pipe is to be connected to the outlet nipple of the blower. The electrical connections may also be made via the two metal pieces. One disadvantage of encasing the pump is a slow start up, owing to the large gas space around the pump. On the other hand, the space provides buffering against fluctuations of pressure and flow and against the effect of any occasional bubble of air that may enter through the vent.

The gas loop with its mobile seal or vent is the analog of a closed electrical circuit connected to ground. When the vent is made inoperative (by converting it to a manometer, or by closing the plain stopcock in Fig. 2) the system "floats" in respect to pressure-the analog of electrical potential. Floating operation over long periods of time can result in accumulation of pressure if the pump is defective at its inlet side, drawing in air. Likewise, a gradual decline of pressure may occur, owing to defects of the.pump at the outlet side, with expulsion of nitrogen, or owing to outward leaks elsewhere in the system. In the non-floating (vented) mode of operation, the reactor is always at the same pressure, slightly above barometric, even when the system is imperfect. Thus, in general, there is no advantage in closing the vent. This must, of course, be done when reactions that evolve or absorb gas are monitored manometrically.

Absorber

Loop with Strict Anaerobicity

The classic absorbent for oxygen, an alkaline solution of pyrogallol, is adequate for the absorber in Figure 1 and for wash bottle 1 in Figure 2. Pyrogallol is not highly efficient in the sense of removing all oxygen in a single pass, but this is not all-important in a circulating system. The traces of carbon monoxide normally evolved from pyrogallol may he objectionable in some instances. For these cases, oxyhydroquinone triacetate (1,2,4triacetoxybenzene = 1,2,4-phenenylacetate) in alkaline solution has been recommended (1). The popular and highly effective Keser solution (alkaline dithionite plus anthraquinone-2-sulfonate) is not recommendable in a single absorber, or the first of a pair of absorbers, because the liquid frothes very badly when air is blown through it, owing to precipitation of the quinone. Two-phase absorbers such as amalgamated zinc with acid chromous solution, also very effective in removing traces of oxygen, deteriorate at least temporarily when encountering air and may fail to provide adequate manostatic action. It is possible to remove oxygen from a circulating gas by means of cathodic evolution of hydrogen, followed by combustion a t ambient temperature at a Pd/Al2O8or other contact. The anode should be of a non-gassing type, preferably the "negative" of a storage battery in the charged state-Pb/HpSOa or Cd/KOH. I t could 654

/

Journal of Chemicol Edumtion

Floating Operation

The system shown in Figure 1 applies to cases where .complete absence of oxygen in the reactor is not critical. Changes of temperature and in the distribution of pressure can cause some momentary intake of air through the vent, reaching, in part, the reactor. Even when the vent liquid is quiescent, or the vent is closed altogether, some ingress of oxygen through junctions, and release from plastics, is unavoidable. For strict anaerobic requirements, when parts per million matter, it is necessary to place a final absorber immediately before the reactor. This absorber need not have as much capacity for quantities of oxygen as the absorber of Figure 1,but it must be capable of removing oxygen in low concentration completely in a single pass. If the final absorber is a wet scrubber, as shown in Figure 2, it adds a parasitic hydrostatic head to h,, which should be at least balanced by a second manostatic column in the left hand portion of the loop, raising h,. The final absorber should not be exposed to air and should be valved into the gas stream F, only after the whole loop and reactor have been purged and the vent liquid has ceased to move. Four-way stopcocks are convenient for inserting the absorber and other components into the stream, but a pair of three-way s t o p cocks, or three single stopcocks, may also serve the purpose.

As a final absorbent, Fieser's solution (NaOH 13.3g, Na&Oa 16 g, anthraquinone-2-sulfonate 2 g, Hz0 100 ml) (3) is very effective. Addition of antifoam is advisable, e.g., Union Carbide SAG 5440, 1 drop/100 ml. The process of absorption is auto-catalytic. A minimum supply of oxygen is necessary, most probably to build up a sufficient steady-state level of semiquinone radical which appears to be the primary absorbent. In the complete absence of oxygen, all semiquinone is reduced to the relatively inert quiuol. Paradoxically, a small seepage of oxygen through plastic junctions upstream from the scrubber maintains the solution in its most effective state. The gas-liquid interaction is best performed in a glass helix inside the scrubber. The incoming gas lifts liquid into the helix, creating a moving chain of slugs of liquid and gas bubbles. A thin liquid film, continually renewingitself, surrounds the traveling bubbles, ensuring a fast transfer of oxygen. The helical scrubber tends to go through regular fluctuations of flow rate but these are largely damped out by the needle valve and by any flow resistances in the reactor. The gas washers at the left of Figure 2, where there is no damping, should, however, not be of the helical type. The final purification may be carried out in a column of solid absorbent, with less pressure drop than in the helical scrubber, and a perfectly smooth flow. Granular mangauous oxide removes every detectable trace of oxygen at room temperature. It is self-indicating, changing color from pale green to brownish grey when used up. Th'e column is prepared from 20-30 mesh/lin. in. manganese dioxide, reduced in flowing hydrogen a t about 400°C (3). The oxide should be diluted by an inert material, otherwise it may crack its glass container through expansion during use. The dioxide is markedly more voluminous than the monoxide. The material can be rejuvenated after exhaustion by repeating the treatment with hydrogen. Wetting the spent material with some methanol and reducing in streaming nitrogen also brings back the green oxide, but this is advisable only when traces of methanol (retained by the oxide, even after purging, and slowly released) are not objectionable in the reactor. Hydrogen is not retained. Another solid absorbent operating at room temper* ture is the "BTS catalyst," grey cylindrical pellets of magnesium silicate carrying copper, with traces of other metals ( 4 ) The material, made by BASF, Ludwigshafen, Germany, is readied for use by reduction with hydrogen a t mild temperatures (over 150°C), when it turns black. There is, however, little change of color upon re-oxidation. After .reduction, the material contains some hydrogen. Later this is gradually released into the loop atmosphere. Generation of hydrogen at a cathode that is coupled with a storage battery "negative" as the anode may be of use in the final purification step. Catalytic combustion would destroy all residual oxygen. In this step the electric current requirements would be minimal. Excess hydrogen would accumulate in the loop and very gradually supplant the nitrogen in the vented mode of operation. Figure 2 shows a hypodermic needle tube seated in a silicone rubber plug above the vent. In the steady operating state the steel capillary ensures that entry of

atmospheric oxygen by diffusion from the U-piece is minimal, without impediment to outward or inward flow when the steady state builds up or is disturbed. Monitoring Residual Oxygen

To check the effectiveness of the oxygen absorbers and the tightness of the loop and reactor, a sensor continuously measuring traces of oxygen should be inserted downstream, next to the reactor. A minute flux of oxygen can be transduced to an electric current signal in a galvanic analyzer, as described elsewhere (6). The older types require means for periodic calibration, since the cell factor is diierent for each cell and dependent on age and temperature. It is hoped that an absolute device, giving a linear galvanic signal closely conforming to Faraday's law, will be generally available shortly (6). Operation in this coulometric way means that every microampere generated in the sensor indicates a flux of 3.48 nanoliter (STP) of oxygen per minute, or 0.0745 ppm Oz by volume in a stream of 50 ml gas/min measured at 20°C and 1atm. For a given mass flux of oxygen, the galvanic-coulometric cell signal is independent of cell individuality, age, and temperature. Potentiometric trace sensors based on Nernst's logarithmic law are available. They are fast and highly sensitive, but having a solid e1ectrolyte;they operate only at elevated temperatures-850°C-and slight traces of organic vapors invalidate the determination of oxygen. The so-called "polarographic" probes for oxygen, with cathodes separated from the sample gas by a membrane barrier, give a linear, diiusion-controlled signal. However, they are subject to variability and aging, and their sensitivity is limited. Procedure

The following steps should be routine when the syatern of Figure 2 is used 1) The stopcocks above the final scrubber and oxygen sensor are in by-pass position. Open the stopcock above the vent. Connect thereactor. 2) Start the pump. Air should immediately bubble inwards through the vent, unless the absorbent in wash bottle 1 is exhausted. 3) After a few minutes-the time depends on thevolume of the reador-when the bubbling has stopped, adjust the needle valve to the desired flow rate, e.g., F R = 50 ml/min. 4) Turn the stopcock above the oxygen sensor from its bypass position to the let-through position, for about one second only, then back to by-pass. Sample gas thus admitted to the sensor produces s. transient signal. The peak iis a. rough indication of the residual oxygen level whioh at this stage is usually still considerable. .-,) After w w or two nlinutrs, sample again as in 4. I f the rtcw pmk is uppreriahly lower than the firt, the absorber in bottle I is satisfactory and there is no major leak in the system. 6 ) Successive hrief sampling should show peaks declining further. Eventually the oxygen sensor oan be left inserted permanently. The reading, now continuous, should decline steadily, approaching a. final level. This indicates oxygen penetrating or emmatine from minor recesses in the svstem. throueh , . ~lsstics. , 7 , i w r r r the'fiwtl atmrrh;. Thc w m r current 4~cnil~l drop imrneJi.ttclv I , , H very I w level, of* fen microampere, arid rernaitr at this level. ~~~~

~

Atmospheres Other than Flowing, Moist Nitrogen

To fill the system of Figures 1 or 2 with, say, hydrogen, close the stopcock above the vent. Disconnect the inlet of the pump from the loop system and instead Volume 47, Number 9, September 1970

/ 655

connect that inlet with the source of the hydrogen-a lecture bottle, Kipp apparatus, or electrolyzer. Flush the entire system with the hydrogen, adjusting the needle valve so that gas bubbles through the wash bottle(s) whiie the flowmeter also indicates flow. Eventually stop the supply of hydrogen and re-connect the inlet of the pump with the pipe of the loop system from which it had been severed. Start the pump and allow a few minutes for the removal of any remaining oxygen. Readjust the flow rate F , with the needle valve. If the system of Figure 2 is used, insert the oxygen sensor, first intermittently, later permanently. Where operation in the nou-floating mode is desirable, or where the reactor consumes hydrogen, some hydrogen may continually be bled into the loop system through a T-junction inserted a t any point, the vent being kept open. If the reactor is a gas chromatograph with a thermal conductivity detector, the preferred atmosphere is helium. A small lecture bottle with this gas for priming the loop should go a long way. Occasionally bottled argon, deuterium, nitrous oxide, Freon-12 (CC12F2),sulfur hexafluoride, butane, or propane may be desirable as reactor atmosphere. Some studies require gas mixtures, either bottled or made by ad hoc blending or doping (7). Doping devices may be applied either externally when priming the loop with the host gas, or as part of the loop system after priming. The gases need not, in general, be of high purity. Oxygen and acidic impurities are being taken care of in the loop system as described, and additional purifiers, such as a column of charcoal, may be inserted. However, while recirculation ensures complete removal of external impurities, the loop atmosphere may be contaminated with by-products from the absorption media (such as carbon monoxide from pyrogallic solution), with plasticizer and grease vapor, and with organics stored and slowly released by plastic connections and other components of the loop itself. Where a dry atmosphere is essential, a freeze trap and/or a conventional moisture absorber must, of course, precede the reactor. I n this case the galvanic oxygen sensor should not be left inserted indefinitely. Exposed to a dry gas stream for hours, it would become sluggish in response and eventually refuse to respond, until it is re-humidified. Also, the wet absorbers would lose water that must be replaced eventually. For F , = 50 ml dry gas/min, the loss is less than 0.1 g HpO/hr. I n case the reactor does not require a continuing flow of gas, the loop may serve to create a controlled anaerobic atmosphere and once this has been accomplished, the pump may be switched off.

indicated by a short slug of liquid moving in a horizontal microburet, with a four-way stopcock for reversing the movement when the slug approaches the end of the graduated part. Larger changes are followed with a dual range soap bubble flowmeter. One position-the one shown in Figure 3-of the upper four-way stopcock is for gases evolving from the reactor, the other position for absorption. At least the major gas spaces of the loop system should be thermally insulated to minimize errors from temperature fluctuations. A variety of means for determining the quantity or rate of uptake or release of gases, far more sensitive than a buret or flowmeter, may be inserted in the gas loop, e.g., a thermal conductivity sensor, or an infrared or ultraviolet photometer. The galvanic monitor mentioned can record minute rates of evolution of oxygen, e.g., in experiments on photosynthesis. Conversely, the consumption of oxygen may be followed in experiments with respiratory systems, "drying" oils, corroding metals (wet or hot), autoxidation, or food degradation. The oxygen to be consumed may be dosed in by an electrolyzer, or leaked in from air through a length of thin-walled silicone rubber tubing inserted ahead of the reactor. This method should be particularly applicable where the rate of uptake is of zero order, so that a small po, can be applied, with Apa large enough relatively to be determined with accuracy. Hydrogen and organic gases evolved in the reactor can be monitored very sensitively with the same sensor, in terms of oxygen consumed in combustion. To this end, a catalytic or high temperature combustion tube must precede the sensor, receiving oxygen from an electrolyzer or diffusion bleed a t constant rate (8). Literature Cited (1) DEYEY.J. D.. Scand. J . Clin. and Lab. Inucst.. 5, 104 (19531. Chem. Ahst?.. 47, 7364h (1953). . F.,J . Am. Cham. Soe., 46,2639 (1924). (2) F l s a ~ nL. ; A. F., r l r n Swmn. S. E., "Inert Atmospheree." Butterworth, (3) W n r v ~ P. London. 1962 p. 41-5. M..Anocw. Chcm.. 70.697 (19581. (4) SOHUTZG. (5) HEnscn, P. A,. A n d . Cham., 32,1030 (1960). (6) Hemcx. P. A,. "Galvanic Analvsls." Chanter in "Advances in Analytical Chemistry and Instrumentation," J. wile,, & Sons, Ino., New York, 1964.Yol. 3,p. 214. U.S. Pat. 3,228,597. (7) H ~ n s c x . P. A,. "Analysis Instrumentation." Instrument Society of America, 1963, p. 65-71. Chcm. A h t r . , 59, 149071: Chimie Analvtipue, 46 31 (19641. J. Aiv Pollution Conlrol Association, 19, 164 (1959). (8) H ~ n s c n ,P. A,, "Galvanic-Coulometric Detectors in Gas Chromatogra~hy:' Chapter in "Lectures in G s s Chromatography," Plenum Press, New York, 1967, p. 149-181.

Uses in Gas Volumetry

Where the reactor releases a gas into the loop, or absorbs a gas from the loop atmosphere, the effect can be measured by attaching a gas buret to the vent. The buret should be purgeable (Fig. 3). The circulating loop gas may be used to stir the reaction medium if necessary. The progress of a reaction involving a gas may be followed, operating in the vented mode, with a sensitive flowmeter attached to the vent. I n the apparatus of Figure 3, small changes of the total gas volume are 656

/

Journal of Chemical Education

Figure 3.

Purgeable buret m d triple range Rowmeter.