A high-speed, high-purity extraction apparatus - Analytical

---

High-speed, High-Purity Extraction Apparatus Walter A. Aue,” Marek M. Danlewski,’ Jurgen Muller, and John P. Laba Chemistry Department, Dalhousie University, Halifax, Nova Scotia, Canada

In the production of bonded phases in general, and ultrathin polymer layers (I) in particular, the synthesized materials have to be exhaustively extracted to remove any nonbonded material. In our experience this generally takes a long time, e.g., from one to a few weeks in a regular soxhlet. To cut down on this time, we switched to soxhlets which extract at the boiling point temperature of the solvent (Kontes Model K585100). Frequent use was also made of Goldfish apparatus (Fisher Laboratory Supplies),where the extraction proceeds a t boiling point temperatures and the solvent returns to the flask continuously, rather than batchwise as in a soxhlet. The latter apparatus worked quite well for bonded layer synthesis; it was also extensively used for an entirely different purpose, namely, sample extraction for pesticide residue analysis. Extraction times could be cut to about 1/4 of those necessary with the regular soxhlet. However, some problems remained. First, a still higher flow was desirable. Second, an all-glass apparatus was considered superior for reasons of purity and inertness. And third, one would want certain extractions to be conducted under nitrogen and/or at high temperatures. To meet these requirements we designed a simple, versatile extractor. Several models with different capacities, flow-rates, etc., were made, which had the following features in common: first, use of Clearseal joints (Wheaton Scientific, Millville, N.J.) to obtain a leak-tight, high-purity system free from lubricants; second, direct flow (comparable to column chromatograph) of the condensed solvent, a t or close to boiling-point temperature, through the sample; third, a built-in thermal insulation, such that a predominant fraction of the evaporated and condensed solvent actually flowed through the material to be extracted, and high-boiling solvents could be used if desirable.

RESULTS AND DISCUSSION Figures 1 and 2 show self-explanatory blueprints of two extractor units (together with more or less conventional boiling flasks and reflux condensors); the main difference between them being that in the first the extraction thimble is an integral part of the extractor; in the second, it is built to be easily inserted and removed. The advantage of the first model is that it can be easily scaled up (our largest model held a half-liter sample) and that dimensions are not critical. The advantage of the second is that a greater number of thimbles, if desired of very small volume and of different frit porosities, can be inexpensively made and used with the same extractor body. Most of our experiences were obtained on the first extractor type, but should in general be applicable to the second as well. They were as follows. First, because of the short extraction time and the use of nitrogen (which also obviates the sometimes detrimental use of boiling stones), any decomposition of labile materials is kept to a minimum. The volume of the solvent flask is easily varied, usually depending on whether the extracted material or the extract is of interest. Second, high-boiling solvents can be used; e.g., a hexadecane (bp = 287 Co) extraction under (oxygen-free) nitrogen poses no difficulties. Third, the apparatus is fairly simple to make and fairly easy Present address, Polish Academy of Science, Warsaw, Poland.

F=HOOD

R ~ ~ I Ycondenser X

with Coldtinper

c 24140 clearseal

H O I ~for vapor

24140 Clearseal

Flgure 1. Extraction apparatus with built-in thimble

Thimble

Flgure 2. Extraction apparatus with removable thimble

to clean to high-purity standards (especially important in the case of pesticide residue analysis). Customary safety precautions should be taken though in regard to the high-vacuum jacket, e.g., by applying a polymer coat. Fourth, the apparatus (especially the one shown in Figure 2) can become difficult to run with solvent mixtures such as the familiar “hexanes” used in pesticide residue analysis, ANALYTICAL CHEMISTRY, VOL. 49, NO. 9, AUGUST 1977

1465

because lower-boilingfractions may disrupt the sample packed inside the thimble. A neat solvent is definitely preferable. Fifth, the extractor shown in Figure 1 can take all the solvent vaporized from a 1-L flask by a 500 W thermowell a t full power; provided the frit is not too fine and the material above offers no unusual flow resistance. In the case of the removable thimble apparatus (Figure 2), this proviso is even more important since the frit area is limited. Rapid flow through the thimble is aided by hydrostatic pull through the long stem. In order for this effect to work well, the stem has to be a few degrees lower than the boiling point of the solvent, and this has to be taken into account in choosing the dimensions of this part of the extractor. Rapid liquid flow implies an enormous flow of vapor, and the openings in the apparatus which allow the vapor to pass through have to be rather large or liquid can build up in the condensor, especially when the thimble fills to the brim. In Figure 1,the same opening for the vapor would also serve to accommodate liquid overflow (which,however, never occurred with the materials we were extracting). If a large overflow should be encountered due to unusual flow resistance in the thimble, a second, slightly lower-placed opening on the opposite side would be desirable.

The overflow problem is potentially more acute in the removable thimble version and has been taken care of by essentially separating vapor flow from solvent flow in the critical area. If high flows are to be achieved, the reflux condensor obviously must be capable of handling them. Shown in Figures 1 and 2 are two constructions which we have used, with equal cross-sectionsof upward and downward flow in the coldfinger, and two water flows in parallel through the (high-resistance) coldfinger and the (low-resistance) outer jacket. They worked to our satisfaction; however, some types of commercial condensors might have served equally well. The obvious application of the described extractors is not the routine experiment but the unusual type of extraction, which calls for high speed, high purity, and possibly high temperature. Under these circumstances they have performed extremely well.

LITERATURE CITED (1) W. A. Aue, C. R. Hastings, and S. Kapila, Anal. Cbem., 45, 725 (1973).

RECEIVED for review March 7,1977. Accepted April 4, 1977. This study was supported by NRC Grant A-9604, AC Grant EMR-7401 and DRB Grant 9502-04.

Laboratory Real Time Clock Featuring Computer, Experiment, and Manual Interaction Robert L. Veazey and Timothy A. Nieman" School of Chemical Sciences, University of Illinois, Urbana, Illinois 6 180 1

Laboratory instruments often require a precise, highly stable time base for operation. Many experiments, particularly those which use a computer system for control and data acquisition, require this same degree of precision and stability, plus the flexibility of a timer with a widely variable period. A crystal controlled oscillator with digital counting circuitry is the only means for obtaining this combination of stability and flexibility. With the advent of microcomputers for control of instrumentation, many labs are faced with major changes in their computer systems. It is awkward and time-consuming to redesign a clock system with every major instrumentation change. This report describes a flexible clock designed and used in our laboratory which meets the criteria of high stability and adaptability, and which has several desirable functions and a unique interaction capability. Many laboratory clocks, the so-called mainframe clocks, are completely dedicated to computer-only interaction; other peripheral type clocks are devoted to one particular instrument (I). The design of our general lab clock allows computer interaction as with a mainframe clock and experimental interaction as with the peripheral clock, as well as manual interaction for setting clock parameters and for initialization. None of the control sources are restricted: every clock function can be controlled from any of the controlling elements. Output information is provided as to the status of the clock, independent of the controlling element. The design we describe here is advantageous in that the interaction is versatile enough that the clock will function with any computer-experiment setup.

GENERAL CIRCUIT DESCRIPTION A block diagram of the clock showing the important interactions is shown in Figure 1. The clock is really a dual clock designed around a pair of MOS technology counter 1466

ANALYTICAL CHEMISTRY, VOL. 49, NO. 9, AUGUST 1977

time-base chips (Mostek 5009P). A 106-Hzcrystal controlled oscillator which is accurate to 1ppm controls the chip internal oscillator circuit. As the diagram of Figure 1 shows, the two independent clocks (PG Time Base and DM Time Base) use the same crystal. The oscillator output connection of one time base (DM) allows synchronous oscillation of the second time base (PG) using the external oscillator provision of the second chip. Note that the two clock circuits are synchronous in their base frequency dependence on the same crystal, however the output frequencies are chosen independently, as are all other functions. Each clock chip has internal synchronous dividers which can be externally selected to provide frequency division ranges from 1 to 36 X los: nine division ranges from 1to los are used in the clock design. The time base chips have reset provisions to reset the internal counters to their highest or lowest states: these functions are used to ensure that clock initialization commands do not produce an initial timing error. The clock is designed so that one of the two clock oscillators, designated the Decade Marker (DM), is used merely as a decade time base and as such does not have some functions available with the other clock oscillator (designated the programmable clock, PG). The PG clock has an additional four decade counter chain (Presettable Counter) external to the actual clock chip. The 74192 BCD counters used have the capabilities of bidirectional counting, clear, and preset; these capabilities are the basis for the various counting functions available with the PG clock. The use of two independent time bases in the design is an important feature of the system. For example, using the decade marker and the programmable clock together, an experiment can be designed in which data are taken in every DM period until the PG clock overflows. The PG clock thus frees the processor from the burden of counting repeated clock interrupts for timing and, since the PG clock is designed to operate independently of