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

Подождите немного. Документ загружается.

Chapter

greater polarity, but still good peak shapes. Figure 1.23 shows a typical chroma-

togram.

Comparison of Figs. 1.22 and 1.23 shows the effect on acetylene and propadi-

ene, which are eluted, well resolved in the latter case, between the saturated and

unsaturated

hydrocarbons. The sodium sulphate deactivation also elutes

the

unsaturates with more equal spacing than with potassium chloride treat-

ment.

Figure

1.24

shows the use of a column deactivated with KC1 for measurement

impurities in propene. The product is about 25% propane and

75%

propene,

with other hydrocarbon gases present at parts per million concentrations. The

propene peak is clearly overloaded, giving

tailing shape, but the trace compo-

nents are easily measured. Because the temperature programme for this separa-

tion

different from that in

Fig.

1.22, the order

elution of n-butane and

acetylene can be seen to have changed.

mentioned earlier, inorganic gases are separated on

combination of po-

rous polymer and molecular sieve columns. Although

PLOT

capillary columns

are also available with these stationary phases, the extra resolving power which

they offer

not necessary for routine analysis of light gases, and

they are

rarely used. The column combination used in Fig. 1.16, which produces the

separation of light components shown in Fig. 1.17, is obviously similar to what

required here. The separation is more complicated, due to the presence of CO

n-C4

i44

_.__

min

Fig.

1.22.

[Iydrocarbons on

alumina/KCI.

Column:

0.32

i.d.

PLOT

alumina deacti-

vated

with

KC1.

Temperature:

35-200°C

at 10"Cimin. Carrier

gas:

helium at

1.0

bar.

The analysis

hydrocarbon gases

CH4

C2'

"44

i44-

1,344-

rnin

10 rnin

rnin

Fig. 1.23. Hydrocarbons on aluminaMa2SO4. Column: 25 m

0.32 mm i.d.

PLOT

alumina deac-

tivated with

Na$O,.

Temperature: 35-200°C at lO"C/min. Carrier gas: helium at 1.0

bar.

and unsaturated

hydrocarbons in refinery gas, and

the direct elution

the

gases from molecular sieve before the

CO,

reaches the end

the porous poly-

mer column

less practicable.

rnin

Fig.

1.24.

Impurities in propene on alumina/KCI. Column: 25 m

0.32 mm i.d.

PLOT

alumina

deactivated with KCI. Temperature: 100-200°C at 5"C/min. Carrier gas: helium at

1.0

bar.

References

Chapter

Figure

1.25

shows a layout which

sufficiently flexible. With valve

in the

inject position, as shown, carrier gas CG1 passes via the porous polymer and

molecular sieve columns to the detector. Column lengths are chosen

that

all

the lightest components

(02,

N,,

CH,,

CO)

are passed onto the molecular sieve

column before

CO,

elutes from the porous polymer. At this time, valve

switched to isolate ;he light components in the molecular sieve column, and

CO,

and the

hydrocarbons elute from the porous polymer column via restrictor A

to the detector. After they have been detected, valve

is returned to position

to allow the elution and measurement of the light components, and valve

also switched to position

the sample load position. This allows

CG1

to back-

flush

and heavier components from the porous polymer column to vent, while

carrier gas

CG2

continues the elution of light components from the molecular

sieve. The pressure of

CG2

and the setting of restrictor

are adjusted

that the

carrier gas flow to the detector remains constant irrespective of valve positions.

Figure

1.26

a typical chromatogram. Hydrogen, which is a component

refinery gas,

not shown here. There are two reasons for this. Firstly,

hardly retained at all on these columns, and is very difficult to trap on the mo-

lecular sieve with the other light components, and secondly because of the diffi-

culty associated with measurement of

in He carrier gas due to the non-linear

relationship between thermal conductivity and H,/He composition. The solution

using a mixed carrier gas was proposed by Purcell and Ettre

[24]

and dis-

cussed by Cowper and DeRose

[25],

but there are implications for the measure-

ment

other components, and it is not recommended for this application.

Hydrogen measurement can be carried out with a separate column/detector

system. using Ar

as carrier gas, or using He carrier, provided that the

Sample

CG1

Porous

polymer

-i

column

CG1

Porous

polymer

-i

column

Carrier

Gas

CG2

t--

-A

Mol.

sieve column

Restrictor

Vent

Fig.

1.25.

Light

gases

analyser. Valve

position

valve

position

The

analysis

hydrocarbon

gases

c2-

CH4

I I

min 5 min 10 min

15 min

Fig. 1.26. Light gases

porous polymer/molecular sieve. Column

Molecular sieve

Column

Combination

HayeSep

and

HayeSep

Temperature:

50°C.

Carrier gas: helium at

ml/min.

sample size is kept sufficiently small

that non-linearity is not significant. With

a sufficiently well equipped chromatograph, all three separation systems can be

configured into one unit.

Figure

1.26

shows that this system measures

and

hydrocarbons, which

are also covered by the alumina

PLOT

column. Since the hydrocarbons represent

the bulk of most refinery gases, the approach would normally be

measure all

the hydrocarbons on the alumina, and to use the porous polymer/molecular sieve

combination just for the inorganic gases

0,,

N,,

and

CO,.

with natural

gas analysis,

could be used as a bridge component, to allow for any drift

in relative detector response between calibration and analysis.

If the inorganic gases are present at trace levels, then a hydrogenator can be

used to convert

and

CO,

CH,,

allowing sensitive detection by

FID.

With-

out invoking exotic and rarely used detectors such as helium ionization, in-

creased sample size is the only practicable solution for

and

N,.

Figure

1.25

shows

arrangement for backflushing

and heavier hydrocar-

bons while analysing the lighter components. The same need can arise for the

alumina column;

and heavier components can be present in a refinery gas

sample, the result of which will be to require a substantially longer analysis time,

or to risk interference with subsequent analyses. If these heavier hydrocarbons

do not need to be analysed, then they can be backflushed to vent using a configu-

ration as shown in Fig.

1.27.

References

Chapter

Whereas a valve was used for backflushing in Fig. 1.25, the configuration of

Fig.

1.27

uses pressure balancing

[26].

The boiling point capillary column,

coated with a non-polar phase, separates the

hydrocarbons from any

heavier components. Pressure regulator

PR2

is set to deliver

pressure which is

slightly higher than the natural pressure at the T-junction when carrier gas is

being delivered under the influence

PRl alone. The backflush switch, shown

both the forward flow (upper) and backflush (lower) modes, either allows

carrier gas from

PRl

to flow through both columns in series, or interrupts this

flow,

that

PR2

delivers continuing flow through the alumina column and re-

versed flow through the boiling point column. The T-junction between the col-

umns is made using a glass quick-seal connector into which the capillary column

is pushed

that the polyimide coating seals to the slightly tapered internal bore.

The appropriate time at which to operate the backflush switch is found by trial

and error,

or,

further T-junction

fitted, by bleeding

small amount of the

non-polar column effluent via a restrictor to a second detector.

If there is a need to measure the heavier components, rather than just get rid

them by backflushing, the configuration shown in Fig. 1.28 can be used. In

Fig.

1.27,

the dimensions and material

the backflush switch are not very im-

portant, as it

not in the flow path encountered by measured components. When

the

backflushed components are to be measured, a different configuration is re-

Sample Valve

Vent

Carrier Gas

Boiling Point

Capillary

Alumina

PLOT

Detector

Fig.

1.27.

Rackflush

heavier

components

The analysis

hydrocarbon gases

Sample Valve

Boiling Point

Capillary

Heavier

Capillary

PR2

Components

Detector

cz]

Fig.

1.28.

Measurement

heavier components.

quired. The backflush switch is replaced by a backflush valve, which must be of

the miniature-dead-volume type, compatible with capillary columns. When the

backflush valve is operated, PR2 supplies forward flow to the alumina column

and backflush flow to the boiling point column, diverting heavier components

into a third capillary column. This is chosen according to need, and can be a

further non-polar column, or a short alumina PLOT, for rapid elution of the

heavier components or groups. PR3 supplies a flow of carrier gas to the third

capillary column during forward flow to the others.

1.3.3

Sample

handling

The range of samples varies from “pure” product to cracker streams after dif-

ferent stages of reaction, and also between those normally handled as liquids and

those which are always gases. LPG samples are normally handled entirely in the

liquid phase, and can be for the purposes of analysis. Chromatographic valves

are available which will easily cope with the pressures involved, and are capable

of injecting sample sizes

microlitres or smaller.

Since many of the streams to be analysed can only be handled as gases, it

better to standardize on treating a11 samples in the gas phase, rather than to han-

dle some as gases and some as liquids. LPG product streams should therefore, be

fully vaporized before analysis. Sampling of LPG into a cylinder

straightfor-

ward, but an ullage space must be provided before the cylinder is transported or

References

Chapter

handled any further. This can be done by venting a proportion of the cylinder

contents,

that the cylinder contains some gas as well as the original liquid.

The different components in the sample will be distributed in different propor-

tions between the gas and liquid phases, but provided that the ullage space is

small, the liquid phase, containing the bulk of the sample, will still represent the

original composition. Where available, a constant pressure cylinder, which con-

tains a sliding piston seal between the

two

ends of the cylinder, should be used.

The sample side

the cylinder can then be filled with

100%

liquid phase, and

the ballast gas, on the other side of the piston, provides the safety buffer.

Most analyses require injection onto more than one column system, even if

fitted into a single chromatograph, and it is therefore crucial that the sample is

uniform between the injections. Sufficient

LPG

should be totally vaporized to

form

a gas sample large enough and with enough pressure

purge and fill all

the sample lines and injection devices. Depending upon the composition of the

sample, it may be necessary to heat the container into which the liquid is vapor-

ized, and to heat the lines through which the gas sample is conveyed to the injec-

tors.

1.4

CONCLUSIONS

From the foregoing,

can be seen that all the gas mixtures normally encoun-

tered in the petroleum industry can be separated into their individual compo-

nents. Having thus separated components, the usual requirement is to measure

them. This quantitative aspect has only been briefly mentioned, but is, of course,

of great importance.

Gas chromatography is not an absolute technique, in that the detectors de-

scribed above do not give a predictable response to any intrinsic property

the

separated components. Quantitative information arises from comparison, the in-

strument being calibrated with one

more mixtures

known composition, then

using the resulting response factors to convert the raw data (usually peak area,

occasionally peak height) to concentration. The quality of the calculated concen-

trations is fundamentally dependent upon how well the composition of the cali-

bration gas is known. Precision of analysis depends upon instrument character-

istics; accuracy depends principally on the quality of the calibration gas.

Detailed discussion

calibration gas preparation and certification

beyond

the scope of this chapter, but a few pointers are given below. Calibration mix-

tures can be prepared by gravimetric, volumetric or manometric techniques.

Most mixtures containing percentage concentrations can be prepared directly,

whereas lower concentrations may require one

more dilution stages to be

used. Some more reactive

adsorptive components, particularly at

low

concen-

trations

(H,S

being typical), require special precautions.

The

analysis

of hydrocarbon gases

Gravimetric.

Calibration gases of the highest quality can be prepared by this

technique. Pure gases are added to a cylinder, which is weighed before and after

each addition. Although the mass

the cylinder is typically orders of magnitude

greater than the masses of the added components, the accuracy achievable, with

suitable equipment and procedures, is excellent. The advantages include the fact

that masses translate unequivocally to molar amounts, and the technique pro-

duces a substantial quantity of calibration gas, under pressure in a cylinder. For

adsorptive components, cylinders which have been specially treated internally

can be used.

Volumetric.

Volumetric techniques operate at or very close to atmospheric

pressure, and can be static or dynamic. For ideal gases, volumes at a constant

pressure are equivalent to molar amounts, and the corrections for real gases at

atmospheric pressure are small. Glass double-ended bulbs and gas-tight syringes

can be calibrated for volume from the mass of their water content. Total dis-

placement of the pure gas contents of a smaller container into a larger one, and

subsequent dilution with a complement gas gives good quality mixtures. Dy-

namic techniques are often preferred for minor and/or adsorptive components.

Various options exist, but for hydrocarbon gases, continuous injection from a

motor-driven syringe into a flowing diluent stream is usual. The uncertainties on

the composition are greater than with static methods, as measurement

flow is

less well defined than measurement of volumes, but the equipment will become

conditioned to adsorptive components.

Manometric.

with gravimetric mixtures, components are added succes-

sively to a cylinder, but the composition is calculated from pressure, rather than

mass increments. This technique allows the largest quantities of calibration gases

to be prepared, as the size of cylinder is not limited by the size of the balance, as

in gravimetric use. The accuracy is, however, relatively poor, since non-ideal

behaviour is much more pronounced at high pressure. Compression

(2)

factors

can be calculated for hydrocarbon gas mixtures with good accuracy, but these

calculations assume that the mixtures are homogeneous. The conditions under

which a cylinder

filled with different components do not encourage rapid

mixing within the cylinder. The calculated factors can therefore only be ap-

proximations.

1.5

ACKNOWLEDGEMENTS

wish to thank British Gas plc, for permission to prepare and to publish this

chapter.

would also like to thank my many colleagues who assisted in its prepa-

ration, and

Mr.

Allott and Mr. R. Jackson,

Lindsay Oil Refinery Ltd., for

their valuable discussions on refinery gas analysis.

References

1.6

REFERENCES

Chapter

A. Melvin, Natural

Gas:

Basic Science and Technology, IOP PublishingiBritish Gas plc

988).

G.J. van Rossum, ed., Gas Quality

Specification and Measurement of Physical and Chemical

Properties

Natural

Gas,

Elsevier, Amsterdam (1986).

IP 337178. Analysis

Non-Associated Natural Gas by Gas Chromatography, Institute of Pe-

troleum. London.

.4STM

D1945

1981, Analysis of Natural

Gas

by Gas Chromatography, American Society

for Testing and Materials.

J.S.

Stufkens and H.J. Bogaard,

Anal.

Chern., 47 (1975) 383.

1SO

6974

1984. Natural gas

Determination of hydrogen, inert gases and hydrocarbons up

Gas

chromatographic method, International Organisation for Standardisation.

1,.

Huber

and

14.

Obbens,

Chromatogr., 279 (1983) 167.

Varian Ltd., Application note No. 31.

IJ.S.

Patent Application 061583,469,

I.B.

Angcll. J.H. Jerman, S.C Terry and

Saadat, A Prototype Gas Analysis System using a

Miniature Gas Chromatograph,

U.S.

Department of Health and Human Services (1981).

J.B.

Angell, S.C. Terry and P.

Barth. Sci. Am., April

983) 36.

van

Es:

C. Cramers and

Rijks,

High Res. Chromatogr., 12 (1989) 303.

Chrompack

I,td.,

Application brochure 501660.

Kenter, M. Struis and A.L.C. Smit, Process Control Qual.,

(1991) 127.

IS0

Ills

10723, Natural gas

Performance evaluation of analysers, International Organi-

sation for Standardisation.

E.H. Osjord and D. Malthe-Soerenssen,

Chromatogr., 279 (1983) 219.

!I.

DiCorcia and

Samperi.

Chromatogr., 107 (1975) 99.

N.C.

Saha,

S.K.

Jain and P.K. Dua,

Chromatogr. Sci., 16 (1978) 323.

D.R. Deans and

Scott,

Anal.

Chem., 45 (1973) 1137.

C.G. Scott,

Inst. Petrol., 45 (1959) 118.

lialasz

and

Heine, Nature, 194 (1962) 971.

N.G.

McTaggart, C.A. Miller and

Pearce,

Inst. Petrol., 54 (1968) 265.

Vonk,

.I.

dc Zeeuw, M. Mohnke and

Buyten, 14th Int. Symp. on Capillary Chroma-

tography, Baltimore, MD (1992).

.I.J<.

Purcell and

L.S.

Ettre, J.

Gas

Chromatogr., 3 (1965) 69.

C.J.

Cowpcr and

A.J.

DeRose, The Analysis

Gases by Chromatography, Pergamon Press,

Oxford

(1983).

D.R.

Deans,

Chromatogr.,

(1965) 477.

E.R. Adlard (Ed.),

Chromatography

the Petroleum

Industry

Journal

Chromatography

Library

Series, Vol.

1995

Elsevier

Science

B.V.

All

rights

reserved

CHAPTER

Advances in simulated distillation

D.J.

Abbott

Esso Research Centre, Analytical Group, Abingdon, Oxfordshire

OX13

6AE,

2.1

INTRODUCTION

Distillation is of major importance in the oil and petrochemical industries,

being by far the most widely used separation process. Many refinery units are

controlled by using distillation data and many products are sold to specifications

incorporating distillation data. There are a number of laboratory distillation tests

which are routinely used to determine the boiling ranges of crude oils and their

products. Crude oils are characterized by ASTM D2892

[l],

which is a “true

boiling point” method involving a lengthy separation in a high efficiency still.

The method uses vacuum distillation to recover material with high boiling

points. A simpler method is ASTM D86, a single plate distillation applicable to

gasolines, middle distillates and similar products. There is also a low efficiency

vacuum distillation method, ASTM D1160.

Simulating distillation by gas chromatography was first reported by Eggerston

al.

in 1960 [2]. It is based on the fact that hydrocarbons are eluted from a non-

polar column in boiling point order, and the column is temperature programmed

until all the sample is eluted. Integration is done in fixed time slices, rather than

on individual peaks. The time axis is converted

temperature by running a

standard under identical conditions to that of the sample. This standard usually

consists

a range of normal paraffins whose boiling points are known accu-

rately, but it can be a mixture of aromatic compounds if highly aromatic samples

such as coal liquids are being analysed. Using data from the runs on the sample

and the standard, it

possible

calculate the percentage recovered at any tem-

perature, or vice versa. The calculation is best done by a computer, which can

also be used to measure the area slices and retention times required. Indeed, the

technique is ideally suited to automatic running, using a computer to control an

References

pp.

52-53