©1999 Sigma-Aldrich Co.

SUPELCO

©1999 Sigma-Aldrich Co.

SUPELCO

Optimization of Extraction Conditions

and Fiber Selection for Low-Molecular

Weight Analytes and Semi-volatile

Analytes Using SPME

Robert E. Shirey and Leonard M. Sidisky

Supelco Inc., Bellefonte, PA USA

©1999 Sigma-Aldrich Co.

SUPELCO

INTRODUCTION

Developing an application using SPME can be an intimidating task
when looking at the many extraction variables and fiber coating
options. This seminar will provide a logical approach to selecting
the appropriate fiber coating and how to optimize the extraction
conditions.
To do this, two studies were investigated. One study looked at the
extraction of low-molecular weight analytes with varying
functionalities. Six different fibers were evaluated for the extraction
of these analytes. The extraction conditions such as pH of the
solution and the effects of salt were studied.
The second study looked at semi-volatile analytes using similar
parameters as in the first study. For this study nine fibers were
evaluated.

Bare fused silica

Adsorbent

Unknown

7µm Polydimethylsiloxane (PDMS)

Absorbent

Nonpolar

30µm PDMS

Absorbent

Nonpolar

100µm PDMS

Absorbent

Nonpolar

85µm Polyacrylate (PA)

Absorbent

Polar

65µm PDMS-DVB, StableFlex™

Adsorbent

Bipolar

65µm CW-DVB, StableFlex

Adsorbent

Polar

55µm/30µm DVB/Carboxen™-PDMS, StableFlex

Adsorbent

Bipolar

85µm Carboxen-PDMS, StableFlex

Adsorbent

Bipolar

Types of SPME Fibers

©1999 Sigma-Aldrich Co.

SUPELCO

The fibers can be classified by polarity or extraction type mechanism.
The polar fibers are the polyacrylate coated fibers and the Carbowax-
Divinylbenzene (CW-DVB) coated fibers. The other remaining fibers
are nonpolar or bi-polar. The nonpolar fibers have a polydimethyl-
siloxane (PDMS) coating, and the bi-polar fibers are primarily
nonpolar, but will extract some polar analytes efficiently.
The other means for classifying fibers are by extraction mechanism.
Absorbent fibers extract by partitioning into a liquid type coating. The
analytes are retained by the thickness of the coating. Adsorbent type
fibers contain porous particles suspended in a liquid phase. The
particles retain analytes in the pores or on the surface. DVB contains
primarily mesopores that extract larger analytes while Carboxen
contains more micropores which are ideal for extracting smaller
analytes. To expand the analyte range that could be extracted with one
fiber, on fiber has DVB-PDMS coated over a layer of Carboxen-
PDMS. The fibers are listed by retention strength increasing from top
to bottom. All of the adsorbent type fibers are on a StableFlex core to
reduce breakage and increase bonding strength.

Fig 1 - Analytes in Volatile Study

Isopropanol

OH

Acetone

O

Methylacetate

Propanal

O

Dichloromethane

Cl

Cl

Acetic acid

O

OH

1,4-Dioxane

O

O

Isopropylamine

Propionitrile

Nitropropane

N

O

O

58

60

74

58

NH

2

59

84

55

60

89

88

N

Pentane

72

O

O

00-0006

©1999 Sigma-Aldrich Co.

SUPELCO

Figure 1 contains the analytes in the mixture. Most of the
analytes are similar in structure but vary by functionality. All of
the analytes have a molecular weight between 58 and 89 AMU.
There are 11 organic classes of analytes represented by this
mixture of analytes. The primary purpose of this study was to
determine the relationship between analyte polarity and fiber
coating.

Analytical Conditions for Evaluation

of

Fibers with Volatile Analytes

Sample: Water containing 25% NaCl and appropriate 0.05M

phosphate buffer, spiked with analytes to a final
concentration of 2 ppm. No NaCl in DI water samples.

Extraction: 15 min with agitation, using Varian 8200autosampler,

Heated headspace done manually at 50°C

Desorption: 2 min, temperature varies, depending on fiber

Column: 30m x 0.32mm x 4.0µm SPB™-1 SULFUR

Oven: 40°C (2 min) to 140°C at 8°C/min (1 min)

Inlet: Split/splitless, closed 0.5min, 0.75mm ID liner

Detector: FID

00-0008

Fig 2 - Effects of pH and Ionic Strength

0

0

0

0

0

Acetic acid

Isopropanol

Isopropylamine

1,4 Dioxane

DI Water
pH=2
pH= 7
pH =11

Fig 2 - Effects of pH and Ionic Strength

Cont..

Fig

Acetone

Propanal

Propionitrile

DI Water
pH=2
pH= 7
pH =11

Fig 2 - Effects of pH and Ionic Strength

Cot.

Nitropropane

Methylacetate

Methylene

Chloride

Pentane

DI Water
pH=2
pH= 7
pH =11

©1999 Sigma-Aldrich Co.

SUPELCO

Figure 2 shows the comparison of the area counts from the analytes
extracted at different pH levels. To obtain the values, the area
counts from all of the fibers were averaged at each pH level.
As expected acidic analytes (acetic acid) were extracted most
efficiently from an acidic solution, pH 2, whereas bases
(isopropylamine, propionitrile) were best extracted from basic
solutions, pH 11. It was unexpected that acetone and isopropanol o
would be most efficiently extracted at pH 11. Nitropropane and
methylacetate were hydrolyzed in basic solutions which accounted
for the poor recovery at pH 11. The other analytes were not highly
affected by pH. The addition of NaCl improved the recovery of all
of the analytes especially the polar analytes.

Fig. 3 - Comparison of Area Responses by Fiber Type

316

857

20

8

311

19

2

1062

121

52

8

1306

7229

758

211

9

837

32

09

499

1671

782

9

14420

5

930

1572

0

59563

55

870

6932

9

108735

Propanal

Nitropropane

Acetone

Propionitrile

PDMS
PAcrylate
PDMS-DVB
CW-DVB
DVB-CAR
Carboxen

Fig 3 - Comparison of Area Responses by Fiber Type

Cont.

95

6

5

445

912

4

75

6

10

60

1

265

1

304

1

2

764

7

13

83

1

831

5

1

285

2

2

883

8

105

10

4

423

14

2

147

84

20

846

327

647

255

600

Methylacetate

Methylene Chloride

Pentane

PDMS

PAcrylate

PDMS-DVB

CW-DVB

DVB-CAR

Carboxen

Fig. 3 - Comparison of Area Responses by Fiber Type

Cont.

144

1445

349

155

981

101

554

1105

565

50

821

1078

1317

19548

10172

35196

26365

29276

Isopropanol

Isopropylamine

1,4 Dioxane

PDMS
PAcrylate
PDMS-DVB
CW-DVB
DVB-CAR
Carboxen

©1999 Sigma-Aldrich Co.

SUPELCO

Figure 3 shows the comparison of the area responses for the analytes
extracted by the various fibers. All of the area counts recorded were
obtained from extraction at the optimum pH level for each analyte.
The Carboxen-PDMS fiber is superior to the other fibers for
extracting these low-molecular weight analytes. This fiber extracted
more than 200 times as much of the polar analytes than the 100µm
PDMS fiber. For the nonpolar analytes the advantage was not as
great, but it was still significantly better than the other SPME fibers.
Isopropylamine was the only analyte that was not most efficiently
extracted by the Carboxen-PDMS fiber. The dual layered PDMS-
DVB over Carboxen-PDMS was best for this analyte and second best
for the other analytes. The PDMS-DVB coating has a high affinity for
amines. The combination of the high affinity of PDMS-DVB for
amines coupled with the microporosity of Carboxen, makes the DVB-
Carboxen fiber the best choice for small amines.

Fig 4A - Analyte Polarity vs. Area Response

A

ce

tic

a

ci

d

Is

o

p

ro

p

an

o

l

Is

o

p

ro

p

yl

am

in

e

A

cet

o

n

e

P

rop

an

al

1,

4 D

io

xan

e

P

ro

p

io

n

it

ri

le

N

it

ro

p

ro

p

an

e

M

et

h

yl

acet

at

e

M

eth

yl

en

e C

h

lo

ri

d

e

P

en

ta

n

e

100µm PDMS

Polyacrylate

Fig 4B - Fiber Polarity vs. Area Response

A

c

et

ic

a

ci

d

Is

o

p

ro

p

a

no

l

Is

o

p

ro

p

yl

a

m

in

e

A

ce

to

n

e

P

ro

p

a

n

a

l

1

,4

D

io

x

an

e

P

ro

p

io

n

it

ri

le

N

it

ro

p

ro

p

an

e

M

et

h

yl

ac

et

at

e

M

et

h

y

le

n

e

C

h

lo

ri

d

e

P

en

ta

n

e

PDMS-DVB

Carbowax-DVB

©1999 Sigma-Aldrich Co.

SUPELCO

Figure 4 shows analyte polarity decreasing from left to right with
respect to fiber polarity. Figure 4A contains a polar (polyacrylate)
and nonpolar (100µm PDMS) absorbent type fibers, whereas,
Figure 4B contains the adsorbent type fibers, the polar CW-DVB
and the less polar PDMS-DVB. For both types of fibers, the more
polar fibers did not extract the polar analytes better than the nonpolar
analytes. Because of the small size of these analytes, fiber polarity
had little or no influence on the extraction of the polar analytes. In
both cases, the less polar fibers were better for the extraction of the
more polar analytes. The only relationship that was seen between
fiber and analyte polarity was that the polar fibers extracted less of
the nonpolar analytes. This would provide some selectivity for the
polar fibers.

Toluene

92

o-Xylene

106

Anisole

O

108

Benzaldehyde

O

106

Phenol

OH

94

Aniline

NH

2

93

Benzoic acid

O

HO

p-Nitrophenol

OH

N

O

O

139

p-Nitroaniline

NH

2

N

O

O

1,3,5-Trinitrobenzene

N

O

O

N

O

O

N

O

O

138

122

213

Fig.5

- Semi-volatile Analytes Used in Study

G00129
8

00-0014

Fig. 5 - Semi-volatile Analytes Used in Study

Cont.

G00129
9

00-0015

Acenaphthene

N,N-Nitrosodibutylamine

N

N

O

158

194

Chrysene

Cl

Cl

Cl

Cl

Cl

Cl

Cl

Cl

Cl

Cl

154

228

Decachlorobiphenyl

499

Dimethyl phthalate

O

O

O

O

©1999 Sigma-Aldrich Co.

SUPELCO

Figure 5 shows the analytes chosen for the semi-volatile study.
Most of the analytes selected contained aromatic rings with
varying functionalities. Some had one functional group while
others had two or three functional groups to vary the polarity of
the analytes. The effects of molecular size of the analytes was a
desired input for this study. To vary the size, two PAHs and
decaclorobiphenyl were added to the mixture.

Analytical Conditions for Evaluation of

Fibers with Semi-volatile Analytes

Sample: Water containing 25% NaCl and appropriate 0.05M

phosphate buffer, spiked with analytes to a final
concentration of 75 ppb

Extraction: Directly immersed for 30 min with agitation

Heated headspace, 65°C for 30 min with agitation

Desorption: 3 min, temperature varies, depending on fiber

Column: 30m x 0.25mm x 0.25µm PTE™-5

Oven: 45°C (2 min) to 210°C at 10°C/min, then to 320°C at

20°C/min (10 min)

Inlet: Split/splitless, closed 1 min, 0.75mm ID liner

Detector: MS ion trap, m/z = 50-515 at 0.6 sec/scan

00-0017

Fig. 6 - Effects of pH

Dibutyl

nitrosoamine

Acenaphthene Decachloro

biphenyl

o-Xylene

Toluene

pH=2

pH=7

pH=11

Fig. 6 - Effects of pH (cont.)

Anisole

Benzaldehdye

Chrysene

Dimethyl

phthalate

Aniline

pH=2

pH=7

pH=11

Fig. 6 - Effects of pH (cont.)

p-Nitroaniline

1,3,5-

Trinitrobenzene

Benzoic acid

Phenol

p-Nitrophenol

pH=2

pH=7

pH=11

©1999 Sigma-Aldrich Co.

SUPELCO

Figure 6 shows that the results from extraction of the analytes at 3
pH levels were for the most part as expected. The more basic
analytes such as aniline, p-nitroaniline and nitrosodibutylamine
were best extracted at pH 11. The acidic analytes, p-nitrophenol
and benzoic acid were best extracted at pH 2. Moderately acidic
phenol, was better extracted at a neutral pH than at pH 2. This has
been shown in previous studies. Trinitrobenzene was best extracted
at pH 2. This analyte is not stable in highly basic solutions.
Dimethylphthalate extracted most efficiently from basic solutions.
This result was not expected. It appears that anisole and chrysene
are best extracted at pH 2, but these results are probably not
significant to make that conclusion. The remaining neutral analytes
were not affected by pH.

Fig 7A - Comparison of Area Responses by Fiber Type

1.

3

7

E

+

05

7

.83

E

+

04

1.

8

2

E

+

04

1.

6

6

E

+

04

5.08E+06

5.43E+06

5.54E+06

4.98E+06

o-Xylene

Toluene

Anisole

Benzaldehdye

Bare FS
7µm
30µm
100µm
Pacrylate
PDMS-DVB
CW-DVB
DVB-CAR
Carboxen

Fig 7B - Area Response vs.

Fiber Type

2.

29E

+

0

5

4.

50E

+

0

3

9.94E+06

9.14E+06

9.

04

E

+

0

5

1.02E+07

9.85E+06

4.

88E

+

0

3

Dibutylnitrosoamine

Acenaphthene

Decachlorobiphenyl

Chrysene

Bare FS

7µm

30µm

100µm

Pacrylate

PDMS-DVB

CW-DVB

DVB-CAR

Carboxen

Fig. 7C - Area Response vs. Fiber Type

5.

26

E+

03

4.

24

E+

03

5.

51

E+

04

8.

95

E

+

04

8.87E+05

1.35E+06

2.61E+06

Aniline

Dimethylphthalate

Benzoic acid

Bare FS

7µm

30µm

100µm

Pacrylate

PDMS-DVB

CW-DVB

DVB-CAR

Carboxen

Fig. 7D - Area Response vs. Fiber Type

3.

81

E

+

0

3

2.

99

E

+

0

3

0

4.

7

3

E

+

03

7

.04

E

+

03

6.27E+05

1.34E+06

6.80E+05

2.58E+05

p-Nitroaniline

p-Nitrophenol

Phenol

1,3,5-

Trinitrobenzene

Bare FS

7µm

30µm

100µm

Pacrylate

PDMS-DVB

CW-DVB

DVB-CAR

Carboxen

00-0021

©1999 Sigma-Aldrich Co.

SUPELCO

Figure 7 shows the results from the extraction of the analytes with the
nine SPME fibers.
The smaller, nonpolar analytes shown in Figure 7A are extracted best
with the Carboxen containing fibers. The porosity of Carboxen
enables it to retain these smaller analytes.
Figure 7B contains the larger nonpolar analytes and
nitrosodibutylamine. Chrysene and decachlorbiphenyl are poorly
extracted by Carboxen. These larger analytes are efficiently extracted
by PDMS fibers and polyacrylate. Bare fused silica can also extract
these analytes but not reproducibly.
Figure 7C contains more polar analytes that are best extracted with the
polar fibers CW-DVB and polyacrylate. The affect of fiber polarity is
more significant with larger analytes. Dimethylphthalate is extracted
most efficiently by the adsorbent containing fibers. It was not
extracted well by PDMS fibers.
Figure 7D contains polar analytes that are best extracted with polar
fibers. The effect of fiber polarity is significant. TNB was best
extracted by DVB-Carboxen fiber.

Fig. 8A - Analyte Polarity vs. Area Response

0.E+00

1.E+06

2.E+06

3.E+06

4.E+06

5.E+06

6.E+06

7.E+06

8.E+06

9.E+06

C

hry

sen

e

Decach

lo

ro

bi

ph

en

yl

Ac

en

ap

ht

he

ne

o-

Xy

le

ne

Tol

ue

ne

A

nis

ol

e

B

en

zal

de

hd

ye

D

im

et

hy

lp

ht

hal

at

e

D

ib

uty

ln

itr

os

oam

in

e

A

nilin

e

p-

Ni

tr

oa

nilin

e

1,

3,

5-T

ri

ni

tro

be

nz

en

e

p-

N

itro

phe

no

l

Phe

nol

B

en

zoi

c aci

d

100µm PDMS

PAcrylate

©1999 Sigma-Aldrich Co.

SUPELCO

Fig. 8B - Analyte Polarity vs. Area Response

0.E+00

1.E+06

2.E+06

3.E+06

4.E+06

5.E+06

6.E+06

7.E+06

8.E+06

9.E+06

C

hr

ysen

e

D

ecach

lo

ro

bi

phe

ny

l

A

cen

ap

ht

he

ne

o-

X

yl

ene

T

ol

ue

ne

A

ni

so

le

Be

nza

ld

eh

dy

e

D

im

et

hy

lp

ht

hal

ate

n-

Di

but

yl

ni

tr

os

oa

m

ine

A

nili

ne

p-

Ni

tr

oa

ni

line

1,

3,

5-

T

ri

ni

tr

ob

en

zen

e

p-

Ni

tr

op

he

nol

Ph

en

ol

B

en

zoic

a

ci

d

PDMS-DVB

CW-DVB

00-0023

©1999 Sigma-Aldrich Co.

SUPELCO

Figure 8 contains plots of analyte polarity, increasing from left to
right vs. fiber polarity.
Figure 8A shows the absorbent fibers,100µm PDMS and the 85µm
polyacrylate. The polar polyacrylate fiber not only extracts the polar
analytes better, in some cases 3 orders of magnitude better, but it
also extracts the nonpolar analytes better. However, the response is
only 1.5-2 times greater. The affinity that polyacrylate has for
aromatic compounds is the reason for the high extraction affinity for
these nonpolar analytes.
Figure 8B shows the adsorption type fibers. In this case the less
polar PDMS-DVB fiber extracted the nonpolar analytes better than
the more polar CW-DVB fiber. The CW-DVB fiber was more
efficient at extracting the more polar analytes than the PDMS-DVB

fiber. These results were as expected

.

©1999 Sigma-Aldrich Co.

SUPELCO

SUPELCO

Fig 9

-

Effects of Coating Thickness on Analyte

Recovery

00-0024

Tolu

ene

o-X

ylen

e

Ace

nap

hth

ene

Chr

yse

ne

Dec

ach

loro

biph

eny

l

Bare FS
7µm

30µm
100µm

PDMS Fibers

30 Min. Ext.

©1999 Sigma-Aldrich Co.

SUPELCO

Figure 9 shows the results from a 30 min. extraction of the

analytes

with the 3 PDMS and bare fused silica fibers. The 100µm PDMS
fiber extracts the lower molecular weight analytes efficiently, but the
efficiency of the larger analytes is not as good. An extraction time of
30 min.is not a sufficient time to allow the larger analytes to migrate
into the coating. The amount of analyte extracted would increase
with a longer extraction time.
The 30µm PDMS is a suitable fiber for extracting both lower and
higher molecular weight analytes within a reasonable amount of
time. This is a good fiber choice for PAHs and PCBs.
The 7µm PDMS has less capacity and poorly extracts the lower
molecular weight analytes, but it is suitable for higher molecular
weight analytes. Bare fused silica and the 7µm produced parallel
lines indicating the the extraction mechanism is similar. Most likely
the 7µm PDMS fiber extracts by both adsorption and absorption.

Fig. 10 - Analyte Size vs. Area Response

T

o

lu

en

e

o

-X

y

len

e

A

cen

ap

h

th

en

e

n

-D

ib

uty

ln

it

ro

soa

m

in

e

Di

m

et

h

yl

p

h

th

al

at

e

C

h

ryse

n

e

D

eca

ch

lo

ro

b

ip

h

e

n

yl

7µm PDMS

Carboxen-PDMS

©1999 Sigma-Aldrich Co.

SUPELCO

Figure 10 compares the extraction of analytes with the Carboxen-
PDMS fiber and the 7µm PDMS fiber. As expected, as the
molecular weight of the analyte increases, the response drops when
using the Carboxen-PDMS fiber. This is particularly true for PAHs.
Either the analytes are not being desorbed off the fiber or are too
large to be extracted. It is most likely the former, but it could be a
combination.
The responses for the analytes extracted with the 7µm PDMS fiber
increase as the molecular weight increases. The more polar analytes
are not efficiently extracted by the PDMS fiber.

©1999 Sigma-Aldrich Co.

SUPELCO

CONCLUSIONS

•

The pH of the extraction solution affects recovery of polar

analytes.

•

The addition of 25% NaCl improves analyte recovery

•

Analytes with a molecular weight of < 90 AMU are most

efficiently extracted by the Carboxen-PDMS fiber.

•

analyte polarity has little impact on fiber polarity for low

molecular weight analytes.

•

Analyte polarity and fiber polarity are directly related for analytes

with molecular weights over 90AMU.

•

Larger analytes >150 AMU and PAHs are poorly extracted by

Carboxen- PDMS fibers. Layering DVB over Carboxen expands
the range.

•

Thin absorbent type fibers are ideal for extracting nonpolar, high

molecular weight analytes