Determination of MIB and GEO in Catfish
J. Agric. Food Chem., Vol. 44, No. 3, 1996 831
cooling unit (Y, Model RB2055AO, FTS Systems Inc., Stone
Ridge, NY). The bottom of R is packed with 500 mg of C18
functionalized silica gel (S) held in place by screen (T) or a
glass wool plug. Component P serves two purposes: (1)
pressure relief and (2) inlet for elution solvent. For P to
function as a pressure relief, nut N must not be overtightened.
(A simple rupture disk could be used in place of P if connected
via a 1/4 in. o.d. tube for a more reliable pressure relief.) If P
is removed for addition of elution solvent, it may be replaced,
and the gas pressure of 7-10 psi and 40-60 mL/min will aid
in pushing the elution solvent through W for collection in a
test tube prior to analysis. Care must be taken to ensure all
O-rings form a seal when equipment is assembled prior to
collection of a sample.
The above-described equipment design originated from
prototype configurations of the MD-SPE which occasionally
experienced explosions. Components C, D and G, in Figure
1, were fabricated from a 1 in. glass bubbler tube (Ace No.
8762-03, Ace Glass Inc., Vineland, NJ ) which was connected
1
to B via E. M was connected to the bubbler tube with a /4 in.
Teflon Swagelok union. The head (components N, O, P, and
Q) of the condenser was constructed with 1/4 in. glass and a
{ 14/20 male joint fashioned in the form of an elbow. This
was connected to M with a 1/4 in. Teflon Swagelok union. A
laboratory condenser, 13 mm i.d., with { 14/20 joints was
connected to the { 14/20 male joint of the elbow condenser
head and sealed with Teflon tape. The absorbent, S, was held
in place with glass wool. Other than these differences, the
prototype apparatus was the same as shown in Figure 1. We
have included explosion problems when they occurred with the
prototype equipment. All of the explosions experienced with
the prototype equipment resulted from high back pressure
generated from the volatilization of water that resulted in the
abrupt separation of the head assembly from B. Note: All
equ ip m en t exp osed to m icr ow a ve en er gy in sid e th e
oven sh ou ld be Teflon , gla ss, or n on -m icr ow a ve-a bsor b-
in g m a ter ia l. The equipment described as shown in Figure
1 has never experienced an explosion.
P r oced u r e. The C18 functionalized silica solid phase
adsorbent contained in the condenser was pretreated by
rinsing with 2 mL each of ethyl acetate, methanol, and
deionized water in sequence. A 20 g subsample of homog-
enized catfish tissue, spiked prior to homogenization with
DHN and NBA at 900 ppb and 5 g of sea sand, was placed in
the hydrogenation vessel, B. The microwave was activated
at 40% power for 10 min, while the sample container was
purged with argon at 40 mL/min. Plug P was then removed,
and W was rinsed with 2 mL of deionized water and then
rinsed three times with 1 mL of ethyl acetate, under argon
pressure. This extract was dried with sodium sulfate and then
injected into the gas chromatograph ion trap mass spectrom-
eter.
F igu r e 2. Relative recovery vs percent microwave power.
reduced the pressure and allowed water rather than
steam to flow through the solid phase adsorbent. This
eliminated the explosion hazards and improved recover-
ies.
Op tim iza tion . Experimental parameters consisting
of microwave power and time, solid phase adsorbent
type and amount, argon flow, and sample pretreatment
were investigated to optimize percent recoveries of the
analytes from the fish tissue. Each optimization trial
consisted of a 20 g ground fish tissue sample spiked at
1 ppm with MIB and GEO. Each optimization param-
eter consisted of three replicate trials for a given set of
conditions.
The length of microwave time was investigated at
10% power from 0 to 50 min. Only modest increases in
relative recovery for both analytes occurred at 30 min.
To reduce the analysis time, recoveries were measured
for increased microwave power levels and only a 10 min
time period. As shown in Figure 2, from 10% to 30%
power, the relative recovery is experimentally equal,
while a significant increase occurs from 30% to 40% for
MIB and from 30% to 50% for GEO. After the 50% trial,
it was discovered that pyrolysis of the fish tissue was
extensive and that a yellow oil distilled into the con-
denser. These two conditions resulted in a highly
complex chromatogram, compared to the 40% power
trial, as depicted in Figure 3. Also, a contaminated GC
inlet sleeve resulted from the injection of the 50% power
sample. The large number of peaks observed probably
relates to distilled oils and/or pyrolysates. These added
peaks only complicated the analysis of MIB and GEO.
It was decided that 40% power and 10 min of analysis
time would be an acceptable compromise.
Ca libr a tion of th e Detector . Calibration solutions were
prepared by spiking the internal standards, DHN and NBA
at 900 ppb, and MIB and GEO in a series of samples from 5
ppb to 10 ppm into the fish tissue, which was then ground.
These samples were prepared according to the described
procedure and then were injected to calibrate detector re-
sponse. A calibration curve of the area ratio of the analytes
(GEO and MIB) to the internal standards (DHN and NBA) vs
the concentration of the analytes was produced.
Solid phase adsorbents consisting of C18 functional-
ized silica gel, Amberlite XAD-2, and charcoal were
compared at 275 m2 surface area. These adsorbents
were chosen because of their affinity for organic mol-
ecules in an aqueous medium. The affinity of each
adsorbent for MIB and GEO was examined as a function
of adsorbing surface area, a material specification
obtained from the suppliers.
As illustrated in Figure 4, the C18 functionalized silica
gel was clearly superior for the recovery of MIB and
GEO in the distillate. Amberlite XAD-2 gave a relative
response approximately half that of the C18 functional-
ized silica gel, and the relative response from charcoal
RESULTS AND DISCUSSION
Ap p a r a tu s Design . Initial designs of the apparatus
did not include a thermostated condenser, rather a tube
containing the solid phase adsorbent. A high flow rate
of steam exited the tube during the microwave-on cycles.
This resulted in very low recoveries of the analytes.
Explosions of the sample container were experienced
from the high back pressure created by the solid phase
adsorbent as steam was released from the tissue sample.
The thermostated condenser was utilized to circumvent
these difficulties. The open volume of the condenser
provided an area for condensation of the steam, which