Adlard E.R. (ed.) Chromatography in the Petroleum Industry

Подождите немного. Документ загружается.

202

Chapter

major source of

SOz

air pollution and acid rain.

For

these reasons, accurate de-

termination of the concentrations and identities of the sulfur-containing species

petroleum products

required.

Many analytical techniques have been developed for the measurement of the

sulfur content of petroleum and petroleum products. These techniques include

methods for the determination of total sulfur content including X-ray fluores-

cence [2], coulometric [3] and radiometric

[4]

techniques and chromatographic

techniques that permit the quantification of individual sulfur-containing com-

pounds. The most widely used chromatographic technique is gas chromatogra-

phy, usually coupled with sulfur-selective detectors.

8.2

SULFUR-SELECTIVE DETECTORS

FOR

GAS

CHROMATOGRAPHY

The most widely used sulfur-selective detector for gas chromatography has

been the flame photometric detector

(FPD)

IS].

Sulfur-selective detection in the

FPD

is based

combustion of sulfur-containing compounds in a hydrogen-

ricldair flame to produce diatomic sulfur in an electronically excited state

(&*).

The emission from

Sz*

is monitored using a photomultiplier tube positioned near

the flame. The problems of the

FPD

for sulfur detection are well documented

[6]

and include a non-linear response, compound-dependent response and perhaps

most important for petroleum analyses, quenching of the response due to co-

elution of hydrocarbons. Despite these limitations, the

FPD

has been success-

fully used for the determination of sulfur compounds in a wide range of petro-

leum samples for many years

[7-1

l].

The Hall electrolytic conductivity detector

(HECD)

can also be used in

sul-

fur-selective mode

[

12,133. Sulfur compounds are combusted in the presence of

oxygen in a heated nickel tube

form

SOz,

which is bubbled into a suitable

sol-

vent such as methanol

methanol/water and the electrical conductivity of the

solvent

monitored. The

HECD

not widely used for sulfur-selective detec-

tion, in part because it is viewed as being difficult to operate and maintain and

suffers from interferences and quenching of the sulfur response.

The atomic emission detector

(AED)

is a multi-element detector that can be

used for sulfur compounds

[14-16].

The

AED

has

linear response for sulfur

compounds and does not suffer from quenching, but has poor selectivity for sul-

fur

versus carbon

[

161

that limits detection of low levels of sulfur compounds in

hydrocarbon matrices. The AED is also relatively expensive and requires highly

skilled operators.

The analysis of low levels of sulfur by gas chromatography is complicated by

the reactivity of most sulfur compounds. Sulfur compounds are irreversibly ad-

The chemiluminescence detector

203

sorbed or will decompose on many surfaces, especially heated metal surfaces.

This results in the loss of sulfur compounds in sample collection equipment and

throughout the chromatographic system (syringes, injection ports, columns, and

detectors). Thus, precautions must be taken to minimize sorption and

loss

of low

level sulfur compounds in the samples by replacing metal surfaces with glass,

fused silica, Teflon, or less reactive metals such as nickel, and by conditioning

the chromatographic system by injection of high concentrations of the sulfur

compounds of interest. Sorption and

loss

are particularly noted for

H2S,

SO2

and

mercaptans, while sulfides and disulfides are less reactive.

Preparation of standards containing sulfur compounds is also a problem. For

gas standards, low level sulfur compounds can be lost on the walls of the gas

cylinders, requiring the use of specially treated cylinder and stabilization of gas

mixtures. For liquid standards, chemical reactions such as oxidation of mercap-

tans by air to form disulfides occurs even with refrigeration of the solutions.

8.3

THE SULFUR CHEMILUMINESCENCE DETECTOR

The limitations of existing sulfur-selective detectors for gas chromatography

led to the development by Benner and Stedman [17] of a new sulfur-selective

detector, the sulfur chemiluminescence detector (SCD).

Combustion of sulfur compounds

a hydrogen-rich flame produces a number

of different sulfur species including

HIS,

HS,

S2,

and

SO2,

with the ratios

of the various species determined by the fuel/air ratio and other factors [6]. Ben-

ner and Stedman [I71 noted that sulfur monoxide

(SO)

is one of the major spe-

cies formed in the combustion of sulfur compounds in reducing flames and, in

fact,

present at 10 times higher concentration than atomic sulfur. Detection of

produced in a flame should provide a more sensitive and linear detection

system for sulfur compounds.

Halstead and Thrush

[18]

reported in 1966, that sulfur monoxide undergoes a

chemiluminescent reaction with ozone. Emission in the blue and ultraviolet re-

gion of the spectrum (260480nm) was observed from the emission of elec-

tronically excited

SO2

formed from the reaction of

with

0,.

The SCD combines these reactions to provide sensitive and selective detec-

tion of sulfur compounds. Sulfur monoxide

produced by combustion of sulfur

compounds in a hydrogen-rich flame, the flame gases are collected by means of

sampling probe and transferred to a chemiluminescent reaction chamber where

they are mixed with ozone and the emission monitored using a photomultiplier

tube after passing through a wide band-pass filter.

products

lair

flame

Sulfur compounds

References

pp.

227-229

204

Chapter

-3

so2

The SCD developed by Benner and Stedman is an atmospheric monitor for

the measurement of both oxidized and reduced sulfur compounds in ambient air.

The detector has a linear response for sulfur compounds, equal molar response

and a detection limit of

0.13

ppbv

(-1

S/s).

quenching of the sulfur re-

sponse is observed due to C02 and hydrocarbons and this lack

interference

provides for improved measurement of low parts per billion level sulfur com-

pounds in ambient air compared with fluorescence and FPD sulfur monitors

At Sievers Instruments, we investigated the use of the SCD as a chroma-

tographic detector by coupling the Benner and Stedman SCD with a gas chroma-

tograph. These preliminary experiments led to the development of a commercial

detector for gas chromatography, the Model

50 Sulfur Chemiluminescence

Detector.

schematic of the SCD is shown in Fig. 8.1. It consists of a hydrogen-

rich flame, a flame sampling probe, a transfer line, an ozone generator, a chemi-

luminescent reaction cell, an optical filter, a photomultiplier tube and associated

electronics for detection of the chemiluminescent radiation and a vacuum pump.

In the GC-SCD, the flame source is a conventional flame ionization detector

operated under hydrogen rich conditions. The flow rates of hydrogen and air for

the flame in the SCD are

200

ml/min and

400

ml/min, respectively. In contrast,

normal FID flow rates are typically

mumin for hydrogen and 250-400 ml/min

for air. The more reducing flame results in the formation of sulfur monoxide

from the combustion of sulfur compounds, yet still provides a flame ionization

detector response, although the sensitivity of the FID is reduced by approxi-

mately an order of magnitude due to the higher hydrogen flow rates.

The flame sampling probe in the SCD is a high purity ceramic tube 8-1

in length

(0.5

mm i.d.) positioned

-4

mm above the FID jet.

avoid collection

of room air, only

-90%

of the flame gases are collected by the ceramic probe.

The probe is positioned in the flame using a bayonet mount to lock it into a mov-

able mounting base shown in Fig. 8.2. The position of the probe in the flame can

be optimized by either adjusting the position of the mounting base or rotating the

probe handle to one of three locking positions, thus repositioning the end of the

probe relative to the flame jet.

The vacuum pump of the SCD serves several purposes; collection of the flame

gases, transfer of the flame gases and ozone into the chemiluminescent reaction

chamber, and reduction of non-radiative collisional quenching of the emitting

species

(SO2*)

the chemiluminescent reaction cell. Under normal operating

conditions, the chemiluminescent cell pressure is

10-1

5 Torr.

chemical trap is

used to remove ozone and oxides of nitrogen from the gas stream before the vac-

uum

pump and since the flame gases contain high concentrations of water vapor,

1191.

The chemiluminescence detector

205

TRANSFER

SAMPUNQ

188EYBLY

FIO

GAS

SUPERCRITICAL

FLUID

CHROMATOGRAPH

-1

MODEL

350

SCD

FILTER

REGULATOR

RESTRICTOR

OZONE

GENERATOR

ELECTRONICS

RECORDER

MNT

VACUUM

UNE

RECORDER

Fig.

8.1.

Schematic

the

sulfur

chemiluminescence detector.

the pump is operated with the gas ballast open to vent water from the pump. To

prevent loss of oil, an oil coalescing filter is used to collect oil in the pump ex-

haust and the oil

returned to the pump.

The transfer line consists of a 1.5-m length of

4.75

mm (3/16 inch) 0.d. black

perfluoroalkoxy

(PFA)

tubing. The use of an inert material minimizes loss of

sulfur monoxide in the transfer line and the relatively large inside diameter re-

References

pp.

7-229

206

Chapter

Chemiluminescence Detector

Ceramic Probe

Fig.

8.2.

Schematic

the

fixed

position interface

for

mounting the

SCD

probe

FID.

duces condensation of water. Residence time in the transfer line and chemilumi-

nescent reaction cell is approximately

and no significant band-broadening or

peak tailing is observed.

Emission from the

SO/03

reaction occurs in the blue and ultraviolet region of

the spectrum, therefore

blue-sensitive photomultiplier tube

used

detect the

chemiluminescent reaction. The flame gases also contain relatively high concen-

trations

nitric oxide, which will also undergo a chemiluminescent reaction

with ozone. Emission from the

N0/03

reaction occurs in the red and near-

infrared region

the spectrum and therefore

optical filter that transmits ra-

diation over the wavelength region

240410

positioned in front of the

PMT

to eliminate interference from nitric oxide.

the commercial

SCDs,

two

different techniques have been used to process

the signal from the photomultiplier tube. The original detector used photon-

The chemiluminescence detector

207

counting, while newer versions employ a pico-ammeter. The principal advantage

of the analog system is a wider dynamic range and minimal distortion of the

chromatographic peak shape from the signal processing electronics.

8.4

PERFORMANCE CHARACTERISTICS

THE SCD

The SCD has a linear response for sulfur compounds over five orders of

magnitude, a specified detection limit

<5 pg

S/s

[20] with a detection limit of

<0.5

S/s

having been reported [21]. This sensitivity corresponds to -10-50 pg

of a sulfur compound or parts per billion concentrations, depending upon the

sample size, injection technique (split versus direct or splitless), and other chro-

matographic variables. The selectivity for sulfur compounds versus hydrocar-

bons of

>lo6

[20-221. The SCD response is equimolar for sulfur compounds and

the response is not quenched by co-elution of higher levels of non-sulfur-

containing compounds.

The excellent selectivity of the SCD is achieved primarily by the combustion

of the sample in the flame. Compounds that will undergo chemiluminescent re-

action with ozone, such as olefins, are converted to C02 and water, eliminating

possible interference. However, under certain circumstances, a positive hydro-

carbon response is observed. For example, when a large amount of a compound

elutes as a very sharp chromatographic peak, the compound is not completely

burned

the flame and the partial combustion products can react with ozone

and produce a response.

this situation, it is necessary to either decrease the

sample size or to change the chromatographic conditions to broaden the chroma-

tographic peak to permit complete combustion of the components.

An SCD response is also observed for arsine, phosphine and their organo-

derivatives, although the sensitivity of the SCD for these compounds has not

been determined. Nitriles also produce a small SCD response, presumably due to

CN emission however this response is much lower that the response for sulfur

compounds.

The equimolar response of the SCD for sulfur compounds has been demon-

strated for a wide range of sulfur compounds [21-241. The relative response

factors for a range of sulfur compounds (relative to phenyl sulfide) are shown in

Table

8.1.

As will be described

more detail in the applications of the SCD, the

equimolar response permits the determination of the total sulfur content (or total

volatile sulfur content) by simply summing the individual components, or when

only total sulfur content is desired, then a short chromatographic column, with

poor resolution can be employed and the overall detector response integrated for

the determination

total sulfur content.

References

pp.

227-229

208

Chapter

'TABLE

8.1

RELATIVE RESPONSE FACTORS FOR SULFUR COMPOUNDS

USING

THE SULFUR

CHEMILUMINESCENCE DETECTOR

Compound Relative response factoP

H2S

cos

so1

CS2

Methanethiol

Ethanethiol

-Propanethi01

2-Propanethiol

-Butanethiol

-Methyl-

-propanethi01

2-Methyl-

-propanethi01

2-Methyl-2-propanethiol

-Pentanethiol

-Hexanethiol

-Heptanethiol

-0ctanethiol

-Decanethiol

Dimethyl sulfide

Methyl ethyl sulfide

Diethyl sulfide

2-Methyl-2-propane sulfide

Dimethyl disulfide

Diethyl disulfide

Thiophane

Thiophene

2-Methy lthiophene

3-Methyl thiophene

2,j-Dimethyl thiophene

2-Ethyl thiophene

2-n-Propyl thiophene

2-n-Butyl-thiophene

Benzothiophene

Dibenzothiophene

0.80

0.90

.oo

.43

1.18

1.05

1.04

1.11

1.02

1.05

1.11

1.12

1.08

.oo

1.06

1.04

.oo

0.88

1.07

1.02

0.95

.oo

.04

1.08

1.10

.oo

1.05

1.03

.02

.oo

1.05

0.87

"Relative to phenyl sulfide, data

from

[21-241.

The chemiluminescence detector

209

8.5 FACTORS INFLUENCING THE SENSITIVITY

AND

SELECTIVITY

OF THE SCD

The key step in the SCD is the combustion of sulfur compounds in the hydro-

gen-rich flame to

form

sulfur monoxide and collection of

by the sampling

probe. Therefore flame conditions and the position of the probe in the flame are

the most critical factors in determining the sensitivity, selectivity, and stability.

The hydrogen/air ratio in the flame determines the products of sulfur com-

bustion. For the best SCD sensitivity and stability, the hydrogen flow rate

set

-200

ml/min and the air flow rate set to

-400

mVmin or a fuel/air ratio of

-1

:2.

Under these conditions, the tip of the sampling probe is heated to sufficient

temperatures that a dull red glow can be observed when the probe is rapidly re-

moved from the flame. These conditions also yield sulfur monoxide as the major

product of sulfur combustion.

With higher air flow rates

>400

ml/min (or lower hydrogen flow rates) sulfur

dioxide is the principal combustion product and the SCD sensitivity is decreased.

The higher air flow rates also produce a hotter flame as indicated by a white

glow at the tip of the probe, when removed from the flame.

At lower air flow rates

(<400

mumin), the yield of

in the flame is actually

increased, but several other factors make operation at lower air flow rates unde-

sirable. Whenever the sampling probe is in the flame and the ozone generator is

on, a background signal is observed for the SCD. In part, this background is due

to the so-called “ozone-wall reaction”, a background radiation observed when

ozone is passed in front of a photomultiplier tube. However, an additional back-

ground signal is observed from the reaction of the flame gases (possibly hydro-

gen atoms) with ozone. Under the normal hydrogen and air flow rates, this back-

ground signal is relatively constant and not affected by the elution of non-sulfur-

containing compounds from the GC However, when the flame for the SCD is

operated at lower air flow rates, this background signal can be “quenched” by

the elution of non-sulfur-containing compounds from the GC. This results in

negative peak for these components when the SCD is operated with low air flow

rates.

In addition to the negative SCD response for non-sulfur species, operating the

flame of the SCD at lower air flow rates produces a cooler flame (no glow ob-

served for the tip of the probe) and this lower temperature does not produce sta-

ble response over a period

several days. Thus, any increase in sensitivity

achieved at the low air flow rates is offset by the reduced stability observed.

Thus, the optimum conditions for most GCs are very close to the

200

ml/min

hydrogen flow rate and

400

ml/min air flow rate.

Examples of the SCD response obtained at different air flow rates are shown

in Fig.

8.3.

The sample is a test mixture containing five sulfur compounds (and

References

pp.

227-229

210

Chapter

%x.Q473

RESPONSE

Fig.

8.3.

Effect

hydrogen and

air

flow rates on sensitivity and selectivity

the SCD.

(A)

Chro-

matogram

sulfur

standard with low air flow

rate.

(B)

Chromatogram

sulfur

standard with

cor-

rect hydrogenhir ratio.

(C)

Chromatogram

sulfur

standard with high air flow rate.

The chemiluminescence detector

some impurities) in hexane solvent, with

10%

benzene. Figure 8.3B shows the

SCD chromatogram obtained under the proper gas flow rates (200 mumin

H2,

400

ml/min air). No response is observed for -130pg of hexane (2pl split

1O:l).

Under the chromatographic conditions employed, benzene co-elutes with thio-

phene, thus slightly distorting the thiophene peak on the SCD, however no

quenching is observed. Figure 8.3A shown the same test mixture using

200 ml/min H2 and 350 mumin of air. Under these more reducing conditions, a

negative response is observed for hexane and benzene and a greater response of

the detector for the sulfur compounds is obtained. Finally, Fig. 8.3C shows the

SCD response for the same test mixture using 200 ml/min

and

500

ml/min of

air. The response for the sulfur compounds

greatly reduced and in some cases,

a positive response for the hydrocarbons can be observed under these more oxi-

dizing conditions. This example illustrates the importance of operating the detec-

tor at the proper gas flow rates.

When the SCD is operated at the proper gas flow rates, good day-to day re-

producibility can be obtained, however, contamination of the probe from column

bleed and septa bleed can cause a loss in sensitivity. Silicone compounds from

either GC columns or septa when high injection port temperatures are used can

deposit on the tip of the probe and decrease the response of the detector for sul-

fur compounds. In most cases, the contamination can be physically removed by

simply inserting a cleaning wire into the probe. To avoid the silicone bleed con-

tamination, the columns (and septa) should be well conditioned and the column

operated at as low of temperature as possible for a given analysis. For example,

most bonded methyl silicone columns can be operated at temperatures up to

275”C, without any bleed problem, but operation at higher temperatures such as

300°C can cause contamination of the probe and reduced sensitivity. It should be

noted that the SCD is more sensitive than the FID to silicone bleed and contami-

nation can occur even though no significant baseline rise is observed on the FID.

8.6

FLAMELESS SULFUR CHXMILUMINESCENCE

dedicated ceramic burner for the formation of sulfur monoxide has been re-

cently described by Shearer [25] and a schematic of the burner is shown in Fig.

8.4.

In this “flameless” SCD, hydrogen and air are mixed with the column efflu-

ent in a heated ceramic tube and all of the combustion gases are transferred to

the chemiluminescence detector. The typical gas flow rates for the burner are

100 mumin for hydrogen and

ml/min of air, which are outside the flammabil-

ity limits for H2/air. To initiate and sustain the combustion, the burner is heated

typically to 800-900°C and direct connection of the chemiluminescence detector

(and vacuum pump) to the burner results in a reduced pressure, typically

References

pp.

227-229