HOW CAN PCBS BE ISOLATED AND RECOVERED FROM SERUM, PLASMA, OR ORGAN FOR TRACE ENVIRO-HEALTH QA?

Sample Preparation Techniques



249



TABLE 3.14

Percent Recoveries for Diazinon and Malathion Using

Hexane and MTBE as Elution Solvents in RP-SPE

Diazinon

Matrix: 10% NaCl (DDl)

1. Elute with hexane, mean of 3 SPEs

%RSD

2. Elute with MTBE, 1 SPE

Matrix: 10% NaCl, 0.2 M

1. Elute with hexane, mean of 3 SPEs

%RSD

2. Elute with MTBE, 1 SPE



Malathion



20

18.3



0

—



20

14

112



0

—

98



recover AR 1248 with nearly 100% recoveries.101 PCBs are easily released from a

serum, plasma, or tissue homogenate via probe sonification (PC), coagulation using

acetonitrile, or other water-miscible organic solvent or salt. The supernatant aqueous

phase is easily separated from the coagulated protein via mini-centrifugation.

RP-SPE serves in this case, similar to the situation when an environmental water

sample is passed through, as an on–off or digital extraction step. The dilute of the

aqueous supernatant serves to decrease the analyte–matrix interaction, as discussed

earlier, which in turn strengthens the analyte–sorbent interaction. In this case, the

combination of PS-RP-SPE with C-GC-ECD as the determinative technique eliminated the need for further cleanup. Such might not be the case if C-GC-MS or

another universal determinative technique is used. Cleanup techniques might then

need to be considered. Listed below are significant outcomes from the author’s

method development study:101

Sample



Recovery as Total PCB (%)



Spiked

Spiked

Spiked

Spiked

Spiked



102

102

115

88.2

77.7



acetonitrile 1

acetonitrile 2

rat liver 1

rat liver 2

rat liver 3



Janák and coworkers102 have discussed a similar approach to the isolation and

recovery of PCBs from whole blood. Details of their approach are outlined via a

flowchart in Scheme 3.8. Referring to Scheme 3.8A, formic acid dissolved in isopropyl alcohol (IPA) was used to coagulate protein from 5 g of whole blood prior

to bath sonication. After passing the supernatant through a conditioned RP-SPE

cartridge, the sorbent is washed with 5% IPA dissolved in water, followed by 10%

methanol in water. The sorbent is then washed again with a series of solutions as

indicated and dried. Referring to Scheme 3.8B, the sorbent is eluted with methylene

chloride (dichloromethane), then evaporated down to 10 µL, and finally reconstituted

with enough heptane to yield ~100 µL. This heptane eluent is transferred to a Pasteur

© 2006 by Taylor & Francis Group, LLC



250



Trace Environmental Quantitative Analysis, Second Edition



To 5 g whole blood is

added 10 mL HCOOH/

IPA (4:1 v/v)



DDI-distilled, deionized

water

MeOH-methanol

HCOOH-formic acid

IPA-2-propanol

DCM-dichloromethane

Cart-cartridge



To condition SPE

Isoelute ENV+

sorbent, pass the

following through:

a)

b)

c)

d)



1 mL MeOH

8 mL DCM

5 mL MeOH

3 mL DDI



Place mixture in

ultrasonic bath and

sonicate for 5 min; let

stand for 30 min



Comments: this approach

continues the central

theme of applying RP-SPE

techniques to whole blood

with a sulfuric acid oncartridge decomposition of

any extracted lipids



Add 10 mL

15% IPA in DDI

Sonicate for 5 min more; let

stand for 30 min; transfer

supernatant to conditioned

SPE cartridge; pass sample

through cartridge at

~1 mL/min



Wash cart w 15 mL 5% IPA; then 10 mL

10% MeOH; dry for 20 sec to expel nonadsorbed water droplets



Wash cart with 4 mL conc H2SO4 followed by:

a) 10 mL DDI

b) 3 mL 1M Na2CO3

c) 3 mLDDI

d) 10 mL 10% MeOH

e) run sorbent bed dry



Add a few drops (0.2 mL) of DDI: MeOH (3:7

v/v) to hasten drying; dry under N2

atmosphere for ~15 min



SCHEME 3.8A



pipette filled with mixed adsorbents and previously rinsed with heptane. The cleaned

up PCBs are subsequently eluted with a binary solvent eluent consisting of 4:1

heptane to methylene chloride. Mean percent recoveries reported from spiked wholeblood specimens were 78 ± 8%.102

RP-SPE as a sample prep technique is low cost, effective, and continues to gain

popularity as regulatory restrictions lift over time. For enviro-chemical QA considerations, Font et al.103 has reviewed developments in water pollution analysis for

PCBs and earlier published a review applicable to multiresidue pesticide analysis of

water. For enviro-health QA considerations, the comprehensive study involving

bovine and human serum conducted by the Centers for Disease Control and Prevention

© 2006 by Taylor & Francis Group, LLC



Sample Preparation Techniques



251



Elute the SPE cart with 5 mL

DCM @ ~0.5 mL/min



Evap eluent down to

10 µL then add enough

n-heptane to reach

~100 µL



Transfer eluent to a Pasteur pipet filled

with mixed adsorbents and prev rinsed

with n-heptane; apply sample, wash

w 3 mL n-heptane



Elute PCBs w 3 mL of heptaneDCM 4:1 v/v; concentrate and

transfer to a microvial



Pasteur

pipet

Anh

Na2SO4

(5 mm

high)

Silica gel w

60% H2SO4

(10 mm

high)

Anh

Na2SO4

(5 mm

high)

alumina

(30 mm

high)

Glass wool



Inject 1 µL into a

C-GC-ECD



SCHEME 3.8B



(CDC) is noteworthy.104 A plethora of analytical papers utilizing RP-SPE for up-front

extraction have appeared in recent years. We now consider a sample prep technique

developed during the late 1980s that relates to SPE, yet is applicable to solid matrices,

in particular biological tissue specimens. It is called matrix solid-phase dispersion.



72. WHAT IS MATRIX SOLID-PHASE DISPERSION AND

IS IT APPLICABLE TO BIOLOGICAL TISSUE?

In the same manner that S-LSE complements LLE, matrix solid-phase dispersion

(MSPD) has emerged as a complement to RP-SPE. MSPD is most applicable to

© 2006 by Taylor & Francis Group, LLC



252



Trace Environmental Quantitative Analysis, Second Edition



biological tissue. Biological tissue is easily “ground” into a chemically bonded silica

sorbent. Soil may prove to be a more difficult sample matrix to achieve MSPD with.

MSPD came about when it was realized that the octadecyl silyl–silica gel sorbent

used in RP-SPE could have abrasive properties, while the octadecyl ligates could

extract analytes from the biological tissue. Barker, along with coworkers, who

published the benchmark paper105 on MSPD puts in this way:106

MSPD combines aspects of several techniques for sample disruption while also generating a material that possesses unique chromatographic character for the extraction

of compounds from a given sample.



A biological sample matrix such as fish tissue is ground into conventional

RP-SPE sorbents such as C18 silica using a mortar and pestle. The cell structure is

disrupted, and organic compounds are released and partitioned into the C18 silica

based on their respective partition coefficients along the same lines as already

discussed for RP-SPE. Barker107 has compared scanning electron micrographs for

bovine liver tissue ground with underivatized silica to those of the same tissue ground

with C18 silica. Although dispersed with both silicas, the underivatized silica shows

an in-tact cell structure evenly dispersed over the material.

MSPD is straightforward to perform in the laboratory. A schematic representation of an eight-step process to perform MSPD is shown in Figure 3.26A, which

depicts the first four steps of blending the tissue with the bonded silica, while Figure

3.26B depicts the last four steps of solvent elution. A glass or agar mortar and pestle

is preferable to the more porous porcelain type to minimize analyte loss. Only gentle

blending is recommended. A ratio of 4:1, i.e., 4 g of C18 silica sorbent material to

1 g of biological tissue, has evolved as the optimum amounts to blend.108

Chemically bonded silicas serve several functions in MSPD. Bonded silicas

serve as:108

•

•

•

•



An abrasive that promotes cell disruption

A lipophilic, bound solvent that disrupts and lyses cell membranes

A sorbent capable of being packed into a column and can be eluted with

solvents of differing polarity

A bonded-phase support that enables sample fractionation



It is all too easy to compare RP-SPE with MSPD. However, after the biological

tissue has been blended with the bonded silica sorbent, a new sorbent or stationaryphase results. In RP-SPE, after a sample has passed through a column, most components of the sample accumulate at the top of the column bed. In MSPD, the

blended sample suggests that the components are more uniformly distributed

throughout the entire sorbent. This creates what Barker108 calls a “unique chromatographic phase” whose dynamic interactions are “not completely understood.”



73. WHAT FACTORS INFLUENCE PERCENT

RECOVERIES IN MSPD?

A robust sample prep technique ought to yield a decent percent recovery. What have

researchers who have sought answers to this question found?106

© 2006 by Taylor & Francis Group, LLC



Sample Preparation Techniques



253



Glass mortar



0.5 Gram sample



Add sample



Step 1



2.0 grams precleaned

and conditioned C18



Step 2



Glass

pestle



Add standards

or modifiers and

equilibrate ; blend



Laboratory

spatula



Step 4



Transfer blend

to column

Tissue/C18 blend

Step 3

Syringe barrel (10 ml)

used as MSPD column

Paper frit

Pipette

tip



FIGURE 3.26A Schematic that shows just how to conduct the first four steps of MSPD.



•

•

•

•

•



•



•

•



General principles of SPE apply.

A unique chromatographic stationary phase is created when biological

tissue is blended with chemically bonded silica.

Underivatized silanols on the surface and in the pores of the support

apparently remove water from the blend, thus yielding a drier support.

Pore size does not seem to be significant.

Particle size is relatively important since particles of 3 to 20 µm do not

permit flow of elution solvent via gravity, whereas the conventional 40to 100-µm-diameter particles do permit flow.

A lipophilic bonded phase is believed to lead to the formation of a new

phase that resembles a cell membrane bilayer assembly, giving the MSPD

material unique chromatographic properties.

The percent carbon load does not appear to have an appreciable effect.

Conditioning of the bonded silica is as essential to MSPD as it is to SPE.



© 2006 by Taylor & Francis Group, LLC



254



Trace Environmental Quantitative Analysis, Second Edition



100

90

80

70

60

50

40

30



Step 5

Compress the sample

Solvent

Modified

syringe

plunger

Add solvent(s)

Paper frit



Receiving

tube



Step 6



Elute sample;

gravity or

vacuum



Step 8

Reduce volume or

evaporate completely.

Reconstitute residue

with solvent of choice.

Filter or centrifuge

(optional).

Use co-column or

additional SPE

cleanup

Submit to analysis



Step 7

Column

eluate



Collect

sample

fractions



FIGURE 3.26B The last four steps of NSPD.



•

•



The matrix has a profound impact on percent recoveries, unlike SPE, due

to the fact that the matrix becomes part of the chromatographic phase.

Elution yields certain coeluted matrix components with certain analytes

of interest that are not well predicted based on SPE principles.



Huang and coworkers109 found that percent recoveries of two different PCB

(presumably Aroclors)-fortified fish (grass carp) samples utilizing MSPD compared

favorably with the more conventional approach of saponification, followed by LLE.

Acidified silica gel was added to the elution syringe and provided cleanup of

coextractants that resulted in the appearance of a distinct yellow color for the hexane

eluent when silica gel was not used. Ling and Huang110 followed up their benchmark

© 2006 by Taylor & Francis Group, LLC



Sample Preparation Techniques



255



paper on isolating and recovering PCBs from fish tissue using MSPD with a second

paper related to OCs as well as PCBs. These authors found an optimum cleanup

and elution solvent combination following the blending of fish tissue with octadecyl

silyl-derivatized silica (ODS), which maximized percent recoveries while minimizing coextractive interferences. For a given ODS, the following adsorbents were tried:

Florisil, acidic silica gel (44% H2SO4), and neutral alumina. The mass of fish muscle

tissue was obtained by subtracting the mass of the adsorbent-loaded-only column

from the mass of the sample-loaded column. Percent recoveries of >90% were

realized for most priority pollutant OCs using the Florisil/hexane–acetone (9:1)

combination. The authors applied MSPD to a number of fish samples from fish

caught from a river that passes through an incineration facility in Taiwan. PCB levels

were significantly higher than OC levels. One fish, caught upstream from the facility,

showed much lower concentration levels of PCBs.

MSPD provides an alternative to conventional LLE, cleanup, and fractionation

sample prep methods as applied to biological tissue. However, laboratories seem

slow to adopt MSPD, as evidenced by the discontinuance of the MSPD kit by Varian

Sample Preparation Products (the only SPE supplier who ever offered MSPD in kit

format). One more offshoot of conventional SPE is now introduced. This sample

prep technique was developed in the early 1990s and is called either solventless SPE

or, more commonly, solid-phase microextraction.



74. WHAT IS SOLID-PHASE MICROEXTRACTION?

One way to answer this question is to state that solid-phase microextraction (SPME)

is to the capillary GC column what SPE is to the HPLC column. For SPE, consider

taking a portion of the HPLC column packing, increasing the particle size, and

allowing for a wider distribution of the particle size; then pack this material into a

cartridge and use this for sample prep. For SPME, on the other hand, coat a fusedsilica capillary column on the outside, instead of the inside, and use this for sample

prep. Janusz Pawliszyn at the University of Waterloo, who understood the limitations

of SFE and SPE, was the first to modify a 7000 Series (Hamilton) liquid-handling

microsyringe by coating polydimethyl siloxane on a fine rod such as fused-silica

fiber. (See Zhang et al.115 for two of the pertinent articles published by Pawliszyn

and coworkers.) SPME is also solventless in that a sorbed analyte is easily thermally

desorbed off of the coated fiber in the injection port of a gas chromatograph.

The modified microsyringe of Pawliszyn has given way to the commercial SPME

device that incorporates a retractable fiber and is manufactured and marketed by

Supelco. The essential design of an SPME sampling device is shown in the following

schematic.

The SPME sampling device can be immersed directly into an aqueous sample,

such as groundwater, for a finite period, then withdrawn, the fiber and rod retracted

back into the needle, brought to the hot-injection port of a gas chromatograph, and

then inserted into the septum and the fiber and rod extended into the injection port

for a finite period, in which thermal desorption is accomplished. For analysis of

aqueous samples for VOCs, the fiber is inserted into the headspace, sampled,

retracted, and then injected directly into the injection port. For solids, wastewater,

© 2006 by Taylor & Francis Group, LLC



256



Trace Environmental Quantitative Analysis, Second Edition



Septum

piercing

needle



Rod

Attachment



Fiber coated

with

polydimethyl

siloxane



sludge, etc., this headspace technique is appropriate provided that analytes can

partition into the headspace from these dirty sample matrices. The rate of SPME

depends on the mass transport from a matrix to the coating. The effectiveness of

mass transport depends on the following:

1. Convective transport in air or liquid

2. Desorption rate from particulates that might be present in the sample

3. Diffusion of analytes in the coating itself

For direct SPME sampling with agitation, convective effects can be minimized.

In the absence of particulates, the mass transport rate is determined by diffusion of

analytes into the coating. For gaseous samples, the rate of mass transport is determined by diffusion of the analyte into the coating, and equilibrium can be achieved

in less than 1 min. For aqueous samples, vigorous agitation of the sample is necessary. A common technique is to simply stir the sample with a magnetic stirrer. In

this case, it takes much longer for the analyte to diffuse through the static layer of

water that surrounds the fiber. Zhang et al.115 articulated the challenge for SPME as

follows:

For volatile compounds, the release of analytes into the headspace is relatively easy

because analytes tend to vaporize once they are dissociated from their matrix. For

semivolatile compounds, the low volatility and relatively large molecular size may slow

the mass transfer from the matrix to the headspace and, in some cases, the kinetically



© 2006 by Taylor & Francis Group, LLC



Sample Preparation Techniques



257



controlled desorption or swelling process can also limit the speed of extraction, resulting in a long extraction time. When the matrix adsorbs analytes more strongly than

the extracting medium does, the analytes partition poorly into the extraction phase.

Because of the limited amount of the extraction phase in SPME (as in SPE), the

extraction will have a thermodynamic limitation. In other words, the partition coefficient, K, is too small, resulting in poor sensitivity. If the coating has a stronger ability

to adsorb analytes than the matrix does, it is only a matter of time for a substantial

amount of analytes to be extracted by the fiber coating and only kinetics plays an

important role during extraction. One of the most efficient ways to overcome the kinetic

limitation is to heat the sample to higher temperatures, which increases the vapor

pressure of analytes, provides the energy necessary for analytes to be dissociated from

the matrix, and at the same time speeds up the mass transport of analytes.



As was developed for other distribution equlibria, such as LLE [Equation (3.4)]

and static HS [Equation (3.25)], we start to discuss the principles that underlie SPME

by considering the equilibrium for the ith analyte between a sample S and the coated

fiber, f, once dynamic equilibrium has been reached.

Let us consider this in more detail. We start by defining the partition coefficient,

i

K fs(SPME), for analyte i between the fiber, denoted by f, and a sample, denoted by s,

as follows:

K ifs (SPME ) =



Cf

n /V f

=

CS C0 − (n /Vs )



(3.44)



where

n = amount of analyte i adsorbed on the SPME polymer film (in moles)

Vf = volume of coating on SPME fiber

VS = volume of sample

C0 = initial concentration of the ith analyte in the sample

Cf , CS = concentration of the ith analyte in the fiber and sample, respectively,

once equilibrium is reached

Solving Equation (3.44) for n yields

 K ifs (SPME )V f Vs

n=

i

 Vs + K fs (SPME )V f







 C0







(3.45)



Equation (3.45) presents an opportunity to make two simplifying assumptions. If

K ifs >> VS

then

n0 ≈ VS C0.



© 2006 by Taylor & Francis Group, LLC



258



Trace Environmental Quantitative Analysis, Second Edition



The amount of analyte sorbed on the fiber is directly proportional to the original

concentration of that analyte in the sample. If, however, a large sample volume is

used,

K ifs V f << VS

then

i

n0 ≈ KfsVf C0.



The amount of analyte sorbed on the fiber is also directly proportional to the

original concentration of that analyte in the sample. Both simplifying assumptions

lead to the conclusion that the amount of analyte sorbed, after equilibrium is attained,

can be related to C0. The term used to describe this is exhaustive SPME, and the

maximum number of moles of the ith analyte that can be adsorbed or partitioned

into the fiber is denoted by n0.



75. CAN WE QUANTITATE BEFORE WE REACH

EQUILIBRIUM IN SPME?

We answer the question by considering the derivation first proposed by Ai.116 The

assumption is that analyte molecules diffuse from (1) the sample matrix to the surface

of the polymer and (2) through the polymer surface to the inner layers. For a steady

state, with diffusion as the rate-controlling factor, the mass flow rate of analyte

molecules from the sample matrix to the SPME polymer surface would equal the

flow rate from the polymer surface to its inner layers. With D1 as the diffusion coefficient

of analyte molecules in the sample matrix and D2 as the diffusion coefficient for

molecules in the polymer phase whose surface area is denoted by A, and Cs and Cf

are concentrations of analyte in the sample matrix and polymer film, respectively,

Fick’s first law can be used to give the rate-determining steps governed by

∂C f

1 ∂n

∂C s

= − D1

= − D2

A ∂t

∂x

∂x

By assuming that a steady-state mass transfer occurs when agitation is effectively

applied, that the diffusion layer is a thin film, and that steady-state diffusion in this

thin film is in effect, a simplified normal differential equation can be considered:

1 dn D1

(C s − C s′ )

=

δ1

A dt

D

= 2 (C f − C ′ )

f

δ2

The terms used in Equation (3.46) are defined as follows:



© 2006 by Taylor & Francis Group, LLC



(3.46)



Sample Preparation Techniques



259



Cs = concentration of the analyte in the bulk sample matrix

C′ = surface concentration of the analyte in bulk sample matrix

s

δ1 = diffusion layer thickness in the sample matrix

δ2 = thickness of polymer film

Cf = analyte concentration in the polymer film at the surface

C′ = analyte concentration in the polymer film in contact with the silica fiber

f

Shown below is a schematic of the interface of the polymer-coated silica fiber

in contact with an aqueous solution. The illustration shows the diffusion layer

thickness in the sample matrix and the thickness of the polymer film. The slopes of

the concentration gradients are shown at the interface between the bulk sample and

polymer film surface. A steady-state diffusion is assumed when the aqueous solution

is effectively agitated. The concentration gradient in the SPME film is assumed to

be linear:



PDMS

coated

fiber



CS



Cf

C′f

C′S



δ2



δ1

0



x



Equation (3.46) can be rearranged and integrated, and within a set of boundary

conditions this ordinary differential equation can be solved for n to yield116

 KV f Vs

n = [1 − e − A{B}t ] 



 KV f + Vs





 C0







(3.47)



Equation (3.47) suggests that the number of moles of analyte sorbed depends

only on the original concentration of analyte in the sample, C0, provided that the

sampling time, t, and the rates of diffusion remain constant between samples. This

diffusion rate is reflected in the slope of the lines between Cs and C s′ and Cf and C ′ ,

f

as shown above. As the sampling time gets very large (i.e., n approaches infinity),

Equation (3.47) suggests that n approaches n0, where n0 is the number of moles

adsorbed by the coating at equilibrium, so that Equation (3.47) reduces to Equation



© 2006 by Taylor & Francis Group, LLC