Sampling and Laboratory Analysis for Solvent Stabilizers - Environmental Investigation and Remediation - page 204

adding salt to extract more of a water-soluble organic compound is referred to as “salting out” and
is quantii ed by the Setschenow constant, K S ,

g

Ê ˆ

w,salt S

(4.1)

log = K [salt] ,

Á ˜

g

total

Ë ¯ w

where γ w,salt is the activity coefi cient of 1,4-

dioxane in saline aqueous solution, and [salt] total is the total molar concentration of NaCl in the solution
used to measure 1,4-dioxane activity in an NaCl solution. Setschenow constants for MTBE, tetrahy-
drofuran (THF), and 1,4-dioxane determined at 70°C were 0.17 L/mol ( R 2

γ w is the activity coefi cient of 1,4-dioxane in pure water,

= 0.995), 0.16 L/mol

( R 0.999), respectively, and R 2 is the correlation coefi cient from four

data points (Jochmann et al., 2006). 1,4-Dioxane's Setschenow constant is considered exceptionally
low, which means that the addition of NaCl to a sample for extraction by SPDE will have a large
increase in sample recovery for 1,4-dioxane.
Selection of the adsorbent coating on the i bers is an important consideration for using SPME.
A study comparing the effectiveness of different SPME media determined that carboxen-poly-
dimethylsiloxane (CAR-PDMS) coating is most effective for extraction of 1,4-dioxane (Shirey,
2000). Another study contrasted the recovery of 1,4-dioxane from six different SPE media, each
receiving a uniform quality control sample spiked at 10

= 0.988), and 0.08 L/mol ( R 2 =

g/L. The best recovery was produced by
2OH Diol sorbent. All six media performed well * (Song and Zhang, 1997).
The quantity of carbon or other adsorbents used in SPME is also important. When cutting pieces
of carbon felt for in-vial elution, at least 1.5 g are needed to produce adequate recovery. Granulated
carbon has greater adsorptive power than felt, but may be short-circuited by channeling of the sam-
ple through the cartridge (Kawata et al., 2001; Isaacson et al., 2006). Samples extracted onto l exible
disks by using SPE may be eluted “in-vial” in a 2 mL autosampler vial and directly analyzed by gas
chromatography and tandem mass spectrometry (GC-MS/MS) (Isaacson et al., 2006). Desorption
of VOCs from the solid phase may use solvents such as dichloromethane (methylene chloride) or
acetone. Desorption is also done in headspace extraction by heating the i ber to an elevated tempera-
ture. A comparison of these approaches showed that heated headspace extraction at 55°C is 31%
more efi cient than direct immersion for the extraction of 1,4-dioxane from CAR-PDMS (Shirey and
Linton, 2006). Nonpolar compounds have a higher afi nity for the SPME media and may displace
polar compounds (Black and Fine, 2001).
Experiments were performed by chemists at Supelco, a manufacturer of chromatographic
columns, to determine whether high concentrations of methyl chloroform would compete with 1,4-
dioxane for sorption onto CAR-PDMS-coated i bers. Although no displacement or competitive
sorption was observed for concentrations as great as 20 mg/L methyl chloroform with 10

μ

μ g/L

1,4-dioxane and 25 g/L 1,4-dioxane-d 8 , an impurity of methyl chloroform was found to interfere

with quantitation of 1,4-dioxane. The impurity was 1,1,2-trichloroethane (Shirey and Linton, 2006;
R.E. Shirey, personal communication, 2007). A “solvent effect” on the effectiveness of SPME for
1,4-dioxane extraction was reported in a study of SPME coupled with headspace analysis that
analyzed 1,4-dioxane in cosmetics (Fuh et al., 2005). Laboratory solvents were found to decrease
the 1,4-dioxane yield from SPME by 30-40% after adding 1%, 5%, and 10% (w/v) of methanol and
acetonitrile. The SPME media used in the study were 75 μm CAR-PDMS, 85 μm polyacrylate (PA),
and 100 μm PDMS i ber assemblies made by Supelco. To avoid the solvent effect, headspace analysis

μ

* The six SPE sorbent media, sorbent manufacturers, and their performance with a 10 μg/L standard in seven replicate
tests were C 18 octadecyl (Applied Separations): 10.96 ± 0.49 μg/L; C 8 octyl (Analytichem International): 10.9 ± 0.67 μg/L;
C 2 ethyl (J.T. Baker): 11.0 ± 0.37 μg/L; SI silyca (Analytichem International): 10.7 ± 0.58 μg/L; 2OH Diol (Varian):
9.94 ± 0.58 μg/L; NH 2 -Aminopropyl (Varian): 9.65 ± 0.17 μg/L. (Kawata et al., 2001.)