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

Chapter

I

12

a

Sample

2’1

,

ez

j:

Column

--<

1

d

1

Detector

I

b

Sample

!

\

‘I

Gas

Jl

v1

Carrier

._

I

1

-.-

I

Fig

1.7.

Accelerated backflush:

(a)

valve

1.

position

1,

inject and

forward

flow;

(b) valve

1.

posi-

tion

I.

sample load

and

backflush.

these light gases at the same temperature as the boiling point column uses for the

C3

to

C,

separation.

Figure

1.9

shows

the configuration which achieves this, with valve

2

serving

to

include

or

isolate column

3,

the porous polymer column. Restrictor

A

is

ad-

justed to give the same pneumatic resistance as column

3,

so

that the carrier gas

The

analysis

of

hydrocarbon

gases

I3

0

min

10

min

I

20

min

Fig.

1.8.

Boiling point separation with backflush. Column

1:

0.75 m

X

2

mm

i.d. Column

2:

5.25 m

X

2 mm i.d. Both containing 28% DC 200/500 on 45/60 mesh Chromosorb

P-AW.

Tem-

perature: 100°C. Carrier gas: helium at 28 ml/min.

flow

remains constant. With column

3

in series, the sample is injected via valve

1.

As

before,

C,+

is backflushed to the detector

by

returning valve

1

to the load

position.

As

soon as

all

the

C,

has passed into column

3

(found

by

trial and er-

ror), valve

2

is switched to isolate the light gases,

N,,

CO,,

C,

and

C,

in that col-

Sample

II

WLoop

,

Carrier

Gas

\

I

Column

2

Column

3

7-

Restrictor

A

Fig.

1.9.

Three

column

analyser. Valve

1,

position

2;

valve 2, position

1

References p.

40

14

Chapter

I

0

min

10

min

20

min

Fig.

1.10.

Boiling point and polymer bead

column.

Column

1:

0.75

m

X

2

mm

i.d.

Column

2:

5.25

m

X

2

mm i.d. Both containing

28%

DC

200/500

on

45/60

mesh Chromosorb

P-AW.

Col-

umn

3:

2.4

m

x

2

mm i.d.

15%

DC

200/500

on

SOB0

mesh Hayesep

N.

Temperature: 100°C.

Car-

rier

gas:

helium

at

28

ml/min.

umn. C,, C, and C, hydrocarbons emerge from columns

2

and

1

to the detector.

After n-C, has eluted, column

3

is

returned on-line by switching valve

2,

and the

light gases elute and are measured. Figure

1.10

shows a typical chromatogram.

The above configuration, with possible minor variations, is widely used for

on-line natural gas analysers, where the sample stream is connected

in

such

a

way that the possibility of contamination of the sample by air is minimal. How-

ever, any sample returned for analysis to a laboratory is prone to air contamina-

tion, which means that

0,

and

N,

must be separated if the presence

of

air is to be

recognized, and accurately measured if the air-free composition is to be recalcu-

lated.

0,

may also be present in a transmitted natural gas if air

or

N,

ballasting is

used as a means of controlling CV

or

WI.

Porous polymer bead columns will not separate air components at the tem-

peratures used for normal analysis; molecular sieves are the only materials able

to

do this, and they in turn retain CO, for

so

long as to make it unmeasurable.

The solution then is to cut the

N,

and

CH,,

with any

0,

which may be present,

onto

a

molecular sieve column, leaving the

CO,

and

C,

on the porous polymer.

Figure

1.1

1

shows the configuration, where the additional valve

3

includes

or

isolates the molecular sieve.

The procedure is similar to the previous one, starting with all columns in se-

ries.

After backflush

of

C,+, the light gases pass into the polymer bead column,

The

analysis

of

hydrocarbon

gases

15

Sampla Column

3

Carrlrr

Gas

Column

4

Restrlctor

A

(I

Reitrlctor

B

I

Column

2

Fig.

1.1 1.

Four

column analyser. Valve 1, position 2; valve

2,

position

1;

valve

3,

position 1.

but now the

N,

and

CH,

(and

0,

if present) are allowed to

go

forward to the

molecular sieve column. Both columns are now isolated, with

CO,

and

C,

in

column

3,

the porous polymer, and

0,,

N,

and

CH,

in column

4,

the molecular

sieve. After elution of

C,,

C,

and

C,

hydrocarbons, column

3

is reconnected for

elution and measurement

of

CO,

and

C,,

and finally column

4

is connected for

measurement of

0,,

N,

and

CH,.

Figure

1.12

is a typical chromatogram.

c3 c2 N2 CH4

I

0

min

10

min

20

min

Fig.

1.12. Boiling point, polymer bead and molecular sieve. Column 1: 0.75 m

X

2

mm

i.d. Col-

umn

2:

5.25

m

x

2

mm

i.d. Both containing 28% DC 200/500

on

45/60 mesh Chromosorb

P-AW.

Column

3:

2.4 m

x

2mm i.d.

15% DC 200/500

on

50/80

mesh Hayesep

N.

Column

4:

2.4 m

x

2

mm

i.d. 45/60 mesh Molecular Sieve

5A.

Temperature:

100°C.

Carrier

gas:

helium at

28

ml/min.

Re$rences

p.

40

16

Chapter

1

The above chromatograms were generated, as

is

evident from the conditions,

on a single chromatograph, configured as in Fig. 1.1

I.

The configuration of Fig.

1.9

(chromatogram

in

Fig.

1.10)

was achieved by isolating column

4

throughout,

and that of Fig.

1.7

(chromatogram in Fig.

1.8)

by isolating columns

3

and

4

throughout.

It

would be possible to optimize columns for these configurations

which gave somewhat better

N,/CH,

separation, but it

is

clear that the baseline

separations of all light components seen in Fig. 1.12 could not be matched.

The setting-up procedures described above sound more complicated than they

in

fact are and most instrument suppliers provide ready configured systems.

Once set up, their performance is usually extremely stable, since none

of

the

coiumns are used anywhere near their temperature limit. The molecular sieve

column will lose separation gradually due to slow adsorption of moisture, but

its performance does not influence other timings or separations within the

sys-

tem.

1.2.2.2

Two

detectors

The chromatogram in Fig. 1.12 illustrates the use

of

two

detectors, TCD and

FID in series. The presence of

N,

and

CO,

makes the use of the TCD necessary,

and

it

is sufficiently sensitive for measurement of the lighter hydrocarbons. To

include pentanes, however, means choosing

a

larger sample size than would be

desirable for linear measurement of the major components. Using an

FID

in

series with the TCD avoids this problem, as the FID has much higher sensitiv-

ity

for hydrocarbons.

A

smaller sample size can be used, satisfying linear detec-

tion requirements for major components and ample sensitivity for minor compo-

n

en

ts

.

1.2.2.3

C6~

detail

Separation and measurement of the individual higher hydrocarbons, repre-

sented by the C,+ backflushed peak,

is

needed both to define the composition

and hence the properties of this group, and also to provide detail for other calcu-

lations such as hydrocarbon dewpoint temperature. The complexity of the minor

alkane isomer components increases dramatically with carbon number, and this,

with the presence of cyclo-alkanes and aromatics, means that high resolution

chromatography

is

required.

Capillary columns are widely used for liquid hydrocarbon samples, and are

equally adaptable to gas analysis. The sample injected from a conventional gas

sampling valve can be split without fear of sample discrimination, as there is no

phase change on injection. Alternatively, the capillary column can be connected

directly into a micro-volume gas sampling valve, fitted with a sample loop of

some tens

of

microlitres. Chromatograms

of

natural gases with very different

isomer distributions are shown in Figs.

1.13

and

1.14.

The limit of detection for

individual components in this instance

is

around

5

parts per million molar.

The analysis

of

hydrocarbon gases

17

0

min i0'min

Fig.

1.13.

Capillary separation.

I

20

rnin

tomin

I

0

min

10

min 20'min 3Omin

Fig. 1.14. Capillary separation. Column:

50

m

X

0.2 mm

i.d.

capillary coated with OV-101. Tem-

perature: 35°C for

5

min, then 6"C/min to 220°C. Carrier gas: helium at

1.3

bar.

References

p.

40

18

Chapter

I

I.

2.2.4

Temperacure programming

Figure

1.6

shows a chromatogram with good resolution and peak shapes for

C,

to

C,

hydrocarbons, insufficient resolution

for

lighter components and long

analysis time and low, broad peaks for

C,

and heavier components. The solution

consisted of diverting groups of components onto different columns, chosen to

optimize their separation at the selected operating temperature. Another ap-

proach

is

to

extend the range by programming the temperature of a single col-

umn.

The boiling point column used

for

Fig.

1.6

would only give good separations

of the light gases by using a

low

sub-ambient column temperature, and

so

Stufkens and Bogaard

[5]

proposed the

use

of a porous polymer bead column,

which has inherently longer retentions. They chose Porapak

R,

a material with

intermediate polarity which optimizes the separation of

C,,

CO,

and

C,.

Using

a

-.IO"C

to 230°C programme,

N,,

CO,

and

C,

to

C,

hydrocarbons are well sepa-

rated for individual measurement, and

C,, C,

and

C,

hydrocarbons are measured

as groups.

n-C,

elutes at the upper temperature limit

for

the column, and any

heavier components cannot feasibly be measured. A

TCD

and an

FID

were used

in

series, with ethane acting as a bridge component to link the detector re-

sponses. Methane was measured by difference, although there is no reason

why

it

could not be measured directly. The method

is

intended for

CV

determi-

nation.

0

min

12

min

I

24

min

Fig.

1.15

Temperature programmed separation. Column:

3

m

X

2

mm

i.d.

50180

mesh Porapak

R.

Temperature: -50°C

for

2

min, then ISWmin

to

240°C.

Carrier

gas:

helium

at

30

ml/min.

The analysis

of

hydrocarbon

gases

19

Using this method, but with a sub-ambient (-50°C) start to the temperature

programme,

O,/N,

separation is achieved within the same analysis. Figure 1.15

is a typical chromatogram. IS0 6974 [6] is based on this separation, with a

starting temperature of 35"C, which eliminates the need for sub-ambient equip-

ment, but an additional separation on a molecular sieve column is required for

helium, oxygen and nitrogen.

1.2.2.5

Combined systems

The capillary chromatogram in Fig. 1.13 separates the majority of components

in a natural gas, but not those few lighter components which have the largest

concentrations. Figure 1.13 has been optimized for good detection limits over a

wide range, giving quantitative measurement from C, to C,, and beyond. Used in

conjunction with the column switching system illustrated in Fig. 1.1 1, either

C,

or

C,

components can be used as a quantitative bridge between the two analyses.

If the capillary conditions are optimized in order to include quantitative meas-

urement of C,, then the packed column TCD system for light gases can be sim-

plified to a porous polymer/molecular sieve combination. Both this and the capil-

lary/FID system can be fitted into the same chromatograph

[7].

The sample

is

injected first into the packed column system for isothermal separation of

0,,

N,,

C,, CO, and C,, and then into the capillary for temperature programmed separa-

tion of the higher hydrocarbons. There is no backflush provision to remove the

heavier hydrocarbons from the packed columns, as they are forward eluted (but

not measured) during the temperature programmed part of the cycle. Figure 1.16

illustrates the column arrangement and Fig. 1.17 a typical chromatogram.

The analysis is comprehensive, but there is no provision for a bridge compo-

nent.

C,

and C, can be distinguished on the capillary column, but it is doubtful

whether C, could be measured with sufficient accuracy. Although the analysis

takes 30 min, the cycle time, allowing for cool down and re-stabilization, is

longer. In a similar application, three sample aliquots are injected, one onto mo-

lecular sieve for

O,/N,

separation, and from which C, and above are backflushed

to vent, one onto porous polymer for

(0,

+

N,),

C,, CO, and C,, and the third

onto a capillary for temperature programmed separation

[

81.

1.2.2.6

Separation

in

backflush

In a novel application intended for process use, the natural gas is injected onto

a single, long packed column containing porous polymer and operating with a

large pressure drop.

N,,

C, and CO, are measured normally, then the entire col-

umn is backflushed to the detector for measurement of the other components.

To

understand how this works requires careful consideration of the mechanism of

backflushing in gas chromatography.

Backflushing is assumed to recombine separated or partially separated com-

ponents because they have to travel equally far in the reverse direction from that

References

p.

40

20

Sample

Chapter

1

Molecular sieve

Fig.

i

16.

Combined

system.

CH4 c4

c5

c6

Capillary column

(W

-

_____

I

15

min

30

min

0

min

Fig

1

17

Combined

s>

stem

chromatogram.

The analysis

of

hydrocarbon gases

21

in which they have been undergoing separation. This is an over-simplification. In

gas chromatography, the mobile phase is compressible, and is driven through the

column under the influence of a pressure gradient. Since each section of the col-

umn has the same mass flow of mobile phase passing through it at any time, the

actual volumetric flow at each point must vary according to the pressure at that

point.

As

the pressure decays from the beginning to the end of the column,

so

the

volumetric flow rate increases, and

so

does the linear velocity.

The mobile phase linear velocity controls the rate of travel of components,

and

so

if this increases along the length of the column, components, even in

iso-

thermal analysis, will accelerate between injection and detection. Looked at an-

other way, a component with a retention time of

x

will not have reached the mid

point of the column at

an

elapsed time of

x/2.

The larger the pressure gradient,

the greater the acceleration.

When the whole column is backflushed, the pressure gradient which exists in

forward flow is reversed, and

so

those parts of the column where the components

travelled most slowly become the high velocity areas, and vice versa. A compo-

nent which has nearly reached the end of the column at the time of backflushing

will have experienced the full acceleration. It now finds itself in the high pres-

sure, low velocity region, and has to do it all over again. Its time to elute in

backflush will be almost identical to its forward flow time. By contrast, a com-

ponent which had travelled only a short distance from the injection point by the

time of backflushing will have spent all the forward flow time in the low veloc-

ity region. After backfl ushing, it is in the high velocity region, and will be eluted

as a backflushed peak in a much shorter time than

it

spent travelling forward.

Hence there is a mechanism of separation, in reversed order, during back-

flushing. The fact that this is not evident when backflushing in applications such

as that illustrated in Figs.

1.10

and

1.12

is because in these, the backflushed sec-

tion is short. Hence, while it

is

subject to high pressure, low velocity on forward

flow, and vice versa on backflush, the pressure drop across the section itself

is

small, and backflush separation does not occur. Figure

1.18

is a chromatogram

showing the application

[9].

The advantages of this approach are that it requires

less hardware (a single ten-port valve would cover both injection and backflush-

ing) and the column can be long enough to give good

N,,

CH,,

CO,

separation

without a consequent time penalty for the heavier components. The main disad-

vantage is to do with the nominal

C,+

group. Since the backflush separation is

good enough to separate

C,

from

C,

and

C,

from

C,,

it is almost certainly per-

forming some separation of

C,, C,

and

C,.

This being

so,

the

C,+

group, which is

a sharp, distinct peak in Fig.

1.11,

becomes relatively “smeared out” and more

difficult to quantify.

I.

2.2.7

Rapid

analysis

Process natural gas analysers have analysis cycle times typically in the range

References p.

40

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

Подождите немного. Документ загружается.