Automated, Economical Sample Preparation: Bottle Liquid–Liquid Extraction

The extraction of organics from water is fundamental in environmental, pharmaceutical, process, food, forensic, and other laboratories. Nearly all laboratories that perform organic analyses are equipped with a battery of separatory funnel (SF) and/or continuous liquid-liquid extractors (CLLEs) to conduct liquid-liquid extraction. For environmental laboratories that conduct liquid-liquid extractions in high volume on a routine basis, this task can be a critical element, often a bottleneck of a laboratory operation. The U.S. EPA imposes a seven-day holding time on water sample preparations, imparting a significant burden on extraction technicians, especially when there are so often delays in getting the samples from the field to the laboratory. If holding times are not met, analysis results are invalid and the laboratories are commonly charged with the responsibility to resample.

Laboratories equipped to do either CLLE or SF liquid-liquid extractions (for the environmental laboratory, most commonly Method 3510C or Method 3520C1) choose one or the other by considering the advantages and disadvantages of each. An assessment is made of available labor and capacity to extract the number of samples on hand, meet holding time requirements and data quality objectives, as well as turnaround time commitments. SFs are used primarily because a few sample preparations can be completed within several hours. However, SF extraction is extremely labor intensive and often fraught with problems of precision and accuracy. Emulsions, rendering the phase separation difficult and sometimes impossible, are too frequently observed. Extraction recovery can be inherently low for some species. CLLEs are costly and slow, difficult to clean, readily broken, and subject to drip rate variation, but require much less labor. For the extraction of herbicides from water using a lighter-than-water solvent, the use of CLLE is precluded, leaving no choice. The especially cumbersome herbicide method, with its extra steps, was an additional driving force behind the development of the technique described here, so-named bottle liquid-liquid extraction (BLLE, or "Billie").

A simple concept in which water, in original sample bottles received directly from the field, is mixed with solvent and turned at a slow rate was explored and found to be markedly effective, eliminating the many problems associated with both SF and CLLE. The advantages of this new method are readily recognized by organic analysts who perform high-volume liquid-liquid extractions on a routine basis. Most notable benefits are striking simplicity and economy; radical reduction of labor, solvent, and solvent waste; ability to prepare a large number of samples at once; as well as freedom from emulsions. Extraction times can be significantly reduced while precision and overall quality are improved. There is no glassware to clean, no elaborate equipment needed, and virtually no labor required. The use of chlorinated solvents and other expensive supplies is sharply curtailed (a "green" technique).

Laboratory control sample (fortified blank) efficiencies for nearly 200 of the most important environmentally regulated semivolatile organic analytes including herbicides, pesticides, Method 6252 components, and Appendix IX3 Method 8270C target analytes were examined and compared to those obtained using conventional SF or CLLE techniques. While the study focuses on species of environmental interest, extraction efficiencies for a host of related chemical species can be readily gleaned. Also presented, an outcome of hundreds of studies conducted to validate this procedure, are little known but distinct recovery patterns and trends observed among SF, CLLE, and BLLE, which will help analysts to improve data quality. The dramatic effects of starting the semivolatile extraction at an acidic pH as opposed to base first are elucidated.

Experimental

Early experiments were conducted by rotating bottles on two ordinary barbecue rotisseries spaced a few inches apart mounted on a wooden substrate. Two-inch sections of rubber hose were cut and slid over square rotisserie rods to create a smooth, rotating surface. Clear, wide-mouth 1-L glass jars were filled approx. 90% full with tap water and fortified with several drops of a 0.1% solution of methyl yellow (p-dimethylaminoazobenzene) in methanol, followed by a few drops of 1:1 sulfuric acid. (Methyl yellow, an Appendix IX target analyte, yellow in color in basic water and bright red in acidic conditions, served as a visual indicator of extraction efficiency.) Forty milliliters of methylene chloride (DCM) was then added along with 10 g of rock salt. The jar was placed horizontally on the rotisserie rack and turned at approx. 3 rpm for 22 hr. After the 22-hr period, the water layer was completely clear and the DCM layer was bright yellow.

The experiment was repeated, altering the initial pH to basic conditions. After 22 hr, the DCM layer was observed to be bright yellow, while the water layer retained a yellow tinge.

Encouraged by this crude initial study, more elaborate experiments were designed, fortifying deionized, charcoal-filtered water in 1-L bottles with 50 µg of a 76-component mixture of semivolatile organics (Mega-Mix, catalog no. 31850, Restek Corp., Bellefonte, PA) containing acid, neutral, and basic extractables. Traditional 8270C surrogates were also added at levels customary for laboratories following environmental procedures. The pH was brought down to <2 and 40 mL DCM was added. Following a 22-hr turn at 3 rpm, the bottle contents were transferred to a 2-L separatory funnel and the DCM was collected. The water layer was returned to the bottle, the pH readjusted to >12, and a second 22-hr turn was performed. The combined acid/neutral and basic extracts were concentrated to 1 mL using a Turbovap apparatus (Zymark Corp., Hopkinton, MA) prior to analysis by the conventional environmental 8270C method. All components were recovered, most with 75% or greater recovery. Phenol, pyridine, and N-nitrosodimethylamine were low, however, recovering only at about the 10% level.

Additional studies were undertaken to improve the recovery of the phenol and amines. The ionic strength of the water was increased by adding sodium chloride, and the effect of intermixing diethyl ether and other solvents with the DCM was explored. Extracting with a 60/40-mL mix of DCM/diethyl ether in place of pure DCM was found to markedly enhance the recovery for phenol. Adding sodium chloride to the water (250 g per 1000 mL) significantly improved the recovery for the pyridine and other amines.

Figure 1 - Photograph of prototype BLLE bottle rotation device.

A heavy-duty turning apparatus was constructed using five sets of 6-ft-long galvanized rods mounted on a rack driven by an electric motor, sprockets, and chain (Figure 1). The device was engineered to turn at triple the earlier experiment rate (12 rpm) as well as to enable 30 1-L bottles to be spun at once. With this apparatus, scores of additional experiments were conducted over years of time to examine the effects of the faster spin, varied durations of spin, alternate mixtures of salt, ether, as well as reversing the extractions by initializing with base rather than acid (Method 625). A vented stopcock bottle cap adapter was designed (Figure 2) to allow the solvent to be removed directly from the bottle, thereby eliminating the need to transfer to a separatory funnel.

Figure 2 - Photograph of bottle cap attachment enabling phase separations to be conducted directly in bottles.

Using the new rotisserie, the following procedures were optimized and are offered here for use:

Procedure A: DCM/ether—Generally semivolatile acid and base-neutral extractables (Environmental Methods 8270 and 625, specifically)
Step 1: Remove the sample bottle from refrigerated storage and allow to equilibrate to room temperature.
Step 2: Shake the bottle briefly to homogenize the sample. Discard up to 250 mL and mark the water level meniscus on the bottle for subsequent accurate volume measurement.
Step 3: Fortify the sample with surrogates and quality control (QC) samples with appropriate solutions. Shake to mix.
Step 4: Adjust the pH of the sample with approx. 3 mL of 12N sulfuric acid. Check that pH is ≤2 with wide-range pH paper. (A lesser amount of a stronger acid solution may be substituted.)
Step 5: Add 100 mL of a 60/40 mix of DCM/diethyl ether (BHT-preserved ethyl ether used for this study; methanol-preserved likely as good).
Step 6: Cap bottle, place horizontally on rotation rack, and spin for 12 hr at 12 rpm. (Because the spinning is gentle, there is no pressure buildup, and venting is not necessary.)
Step 7: Attach an adapter and stopcock to the bottle or transfer the bottle contents to a conventional separatory funnel to effect phase separation. Discharge the DCM/ether layer (heavier than water) into a 200-mL bottle. Add approx. 10 g of sodium sulfate to the extract and mix.
Step 8: If using a separatory funnel, return the water to the original bottle and readjust the pH of the water to ≥12 (check with pH paper) using approx. 9 mL of 6N NaOH or a smaller quantity of 10N NaOH.
Step 9: Add 80 mL of a 50/30 DCM/ethyl ether mix and return to rotisserie for a second 12-hr period.
Step 10: Reattach an adapter and stopcock to the bottle or transfer the bottle contents to a separatory funnel to effect phase separation. Combine the DCM/ether layer with the initial DCM/ether extract.
Step 11: Allow the extract to remain in contact with sodium sulfate for at least 2 hr, mixing periodically. If the sodium sulfate forms clumps, add additional sodium sulfate until it remains free-flowing. (Adding sodium sulfate directly to the extract in place of a common practice of pouring the extract through a funnel or column of sodium sulfate serves several purposes. The sodium sulfate remains in contact with the extract for longer periods of time, the successful removal of water is visibly indicated by an absence of clumps, and much less sodium sulfate is required.)
Step 12: Drain the extract through a glass fiber filter and concentrate the extract using conventional evaporative techniques. Submit for analysis.
Step 13: Recharge the original bottle with tap water to the original mark and pour into a 1000-mL volumetric flask to record the exact volume extracted.

Procedure B: Hybrid—Generally semivolatile acid and base-neutral extractables (Environmental Methods 8270 and 625, specifically)
Steps 1-7: Identical to Procedure A 
Step 8: Returning the water to the original bottle, add 190 g of sodium chloride and shake to dissolve. 
Step 9: Readjust the pH of the water to ≥12 (check with pH paper) using approx.9 mL of 6N NaOH or a smaller quantity of 10N NaOH. 
Step 10: Add 80 mL of DCM and return to rotisserie for a second 12-hr period.
Steps 11-14: Identical to steps 10-13 of Procedure A.

Procedure C: Herbicides (Environmental Methods 8151 and 615 including hydrolysis step)
Step 1: Remove the sample bottle from refrigerated storage and allow to equilibrate to room temperature.
Step 2: Shake the bottle briefly to homogenize the sample. Discard approx. 250 mL and mark the water level for later measurement.
Step 3: Fortify the sample with surrogate and batch QC samples with appropriate solutions.
Step 4: Add 190 g of sodium chloride and shake to dissolve.
Step 5: Adjust the pH of the sample with 14 mL of 6N NaOH. (Check that pH is =12 with wide-range pH paper.)
Step 6: Cap bottle, place on rotation rack, and spin for 2 hr at 12 rpm.
Step 7: Add 100 mL DCM to the water in the bottle and turn for an additional 4 hr. Step 8: Attach an adapter and stopcock to the bottle or transfer the bottle contents to a separatory funnel to effect phase separation. Discard the DCM layer and return the water layer to the original bottle. Readjust the pH of the water to =2 (check with pH paper) using approx. 14 mL of 12N sulfuric acid.
Step 9: Add 140 mL of diethyl ether (BHT-preserved required as opposed to methanol-preserved to avoid quenching subsequent methylation) and return to the rotisserie for a 12-hr turn.
Step 10: Reattach an adapter and stopcock to the bottle or transfer the bottle contents to a separatory funnel to effect phase separation. Collect the ether layer (top layer) in an acid-washed 250-mL bottle containing approx. 10 g of acidified sodium sulfate.
Step 11: Returning the water to the original bottle, add an additional 100 mL of diethyl ether and spin for an additional 12-hr period.
Step 12: Reattach an adapter and stopcock to the bottle or transfer the bottle contents to a conventional separatory funnel to effect phase separation. Add the ether layer to the original ether extract in the 250-mL bottle containing sodium sulfate.
Step 13: Allow the extract to remain in contact with sodium sulfate for at least 2 hr, mixing periodically. If the sodium sulfate forms clumps, add additional sodium sulfate until it remains free-flowing.
Step 14: Concentrate and derivatize the extract conventionally and submit for analysis.
Step 15: Recharge the bottle with tap water to the original mark and measure the sample volume using a 1000-mL volumetric flask.

Results and discussion

Tables 1-3 summarize dozens of laboratory control sample and method blank analyses. Control matrices of purified water were fortified with a methanolic mix of compounds prior to pH change and solvent addition. Under the column header "List," analytes associated with Method 625, the Appendix IX Groundwater Monitoring List, or the Toxicity Characteristic Leaching Procedure (TCLP) are encoded with the character 6, A, or T, respectively. In each table, applicable QC limits obtained from Method 625 are provided as a frame of reference. Recovery limits for those compounds not associated with Method 625 are arbitrarily set to 40-120. Surrogate recovery limits are those established by the U.S. EPA Contract Laboratory Program (CLP) for semivolatile organics. A composite of the results for a number of associated method blanks is provided. The pound sign (#) is used to call attention to analytes of special interest.

Table 1 - Bottle extraction: Replicate results of four techniques
Table 1 - Bottle extraction: Replicate results of four techniques continued
Table 1 - Bottle extraction: Replicate results of four techniques continued