29

GAS ANALYSIS WI T H CARBON MOLECULAR SIEVE

Daniel M. Ottenstei n and Walter R. Supina and Patri ck W. Byrnes
Supelco, I nc. Fi sher Sci enti f i c Co.
Supelco Park 711 Forbes Avenue
Pi ttsburgh, Pa. 15219
Bel l ef onte, Pennsylvania 16823
I NTRODUCTI ON
Carbosieve B is a unique adsorbent f or gas sol i d chromatography which can be
used f or the separati on of:
(1) Permanent gases
(2) Sul fur and ni trogen oxi des
(3) Li ght hydrocarbons C to C
(4) Certai n lower mol ecul ar wefght compounds such as formaldehyde,
methanol and water
(5) Water from most organi cs
Because of i ts i nertness i t can be used f or the determi nati on i n the parts per mi l -
lion and parts per bi l l i on concentrati on l evel . Due to its unique structure, i t w i l l
el ute compounds in an order di f f erent from that obtai ned wi th other adsorbents or
packings.
Car bosi eve e i s manufactured by SUPELCO, I NC.
Kaiser' descri bed
Zl atki s et al . have al so descri bed i ts chromatographic properti es for
i ts use f or the determi nati on of trace quanti ti es of water and as a general purpose
GSC packing'.
l i ght gas anal ysi s3.
Carbosieve B is a hi ghl y pure carbon wi th a surface area of approximately 1000m2/g
and pore radi us i n the range of 10-12 Angstroms, making i t essenti al l y a carbon
molecular sieve. Its hi gh puri ty el i mi nates the adsorpti on problems that are nor-
mally encountered wi th conventi onal carbon and other adsorbents; thi s is an impor-
tant feature.
i nstances, are consi derabl y di f f erent from that obtai ned wi th ot hd adsorbents.
Its unique surface and structure cause separati ons which, i n many
Thi s adsorbent i s hi ghl y nonpol ar; .Rohrschneider Constants show thi s very wel l
and are as fol l ows:
Benzene Ethanol MEK Nitromethane Pyri di ne
(XI (Y 1 ( 2) (U) (SI
-1.03 -0.84 -0.62 -1.93 -1.51
A l l val ues are negati ve i ndi cati ng that Carbosieve B is even more nonpol ar than
squalane.
compounds, even before methane.
Because of its nonpol ar character, water is el uted before al l organi c
APPLICATIONS
PERMANENT GAS SEPARATIONS
Permanent gases such as H2, 02, N2, CO, CH
and C02 can be separated wi th
4
To do thi s i t is necessary to temperature program the column from
are not easi l y separated so that a 9 f t. x
Carbosieve B.
room temperature to 175OC.
118" stai nl ess steel column is neede3, packed wi th 120/140 mesh Carbosieve B.
column is hel d i sothermal l y f or four minutes, then programmed at the rate of 3O0C/min.
O2 and N
The
30
to 115C.
Under these condi ti ons, oxygen and argon are not separated from each otter.
3 shows a chromatogram of CO, m4 and C02 i n ai r.
SEPARATION OF HYDROCARBONS
The components are el uted i n the order of H 2, 02, N2, CO, CH
and C02.
Figure
Carbosieve B can be used to separate hydrocarbons i n the C - C range. TO do
thi s, the column must be operated above ambient temperature. 'de C54and hi gher
molecular weight hydrocarbons are not el uted from the columns. They appear to be
permanently adsorbed by the Carbosieve B. Conceivably they could be el uted i f the
column temperature were rai sed suf f i ci entl y, but thi s would resul t i n pyrol ysi s of
the sample. range, the compounds are el uted by the degree of unsat-
urati on, wi th saturated compounds el uted l ast. This i s i n contrast to conventional
pol ar columns where the order would be (a) paraf f i ni c (b) ol ef i ni c (c) acetyl eni c.
The pol ar columns would be usef ul f or traces of paraf f i ns i n ol ef i ns or ol ef i ns i n
acetyl eni c compounds, but f or traces of acetyl ene i n ol ef i ns, or ol ef i ns i n paraf f i ns,
Carbosieve B is unique.
In the C1 - C
3
C Hydrocarbons
2
The C2' s are separated i n the order acetyl ene, ethyl ene and ethane. These com-
ponents alolig wi t h methane are shown separated i n Figure 4 at 150'C wi th a 3 f t. x
2mm gl ass column.
C3 Hydrocarbons
order of methylacetylene, allene, propylene, and propane. Cyclopropane and al l ene
are not separated from each other and are el uted together.
C3's along wi th methane and the C2' s i s shown in Fi gure 5 wi th a 3 f t. column at ZOO' C.
The C ' s are al so separated by the degree of saturati on. They are el uted i n the
3
The separati on of the
C Hydrocarbons
4
I n prel i mi nary work wi th the C ' s, i t has been possi bl e to el ute is0 and n-butane,
but not the unsaturated C ' s .
4
are not resol ved wi th a 3 f t. x 2mm column at 215'C.
i n Figure 6 at 215C.
The $so and n-butane peaks are rel ati vel y broad and
Separati on of C1 - C4 is shown
Acetylene i n Ethylene
Zl atki s3 has shown that traces of acetyl ene can be determined i n ethylene.
pri nci pal advantage of Carbosieve B f or thi s anal ysi s is that acetyl ene i s el uted
before ethylene. This allows a l arge sample to be used wi thout swamping the acety-
l ene peak, as is the case where ethyl ene is el uted f i rst.
The
The separati on shown i n Figure 2 was carri ed out wi t h a 3 f t . x 2mm column,
Carbosieve B, 45/60 mesh at 14O"C, flow 4Omllmin. N and a flame i oni zati on detector
was used here. Sample size 1.5m1, sensi ti vi ty 2 x 18-" A.F.S., lmv recorder.
SEPARATION OF PERMANENT GASES AND METHANE, ETHYLENE AND ACETYLENE
I f it i s not necessary to separate O2 and N but only ai r, CO, CH , C02, acety-
methane, ethane, ethy-
2
lene, ethyl ene and ethane, the separati on can be carri ed out wi t h a 4 'ft. x 1/8"
column wi th Carbosieve B, 60180 mesh programmed from 30 to l l OC as shown i n Figure
7. If O2 and N2 are al so to be separated along wi th
C ' s el uted i n order of acetyl ene, ethyl ene and ethane.
2
lene and acetyl ene, the 9 f t. x 1/8" column used i n can be used with the
31
SEPARATION OF WATER, FORMALDEHYDE AND METHANOL
Kaiser2 f i r st reported thi s di f f i cul t separati on; i t can be readi l y made wi th
Carbosieve B as shown i n Figure 1.
ti on contai ni ng 36-38% formaldehyde, 10% methanol and the remainder water.
condi ti ons used are as follows: Column, 4 f t. x 3mmI.D. gl ass packed wi th 60/80
mesh Carbosieve B; Column Temp., 20OoC; Flow rate, 7fml/min. Hel i um: Sample si ze,
0.5 mi crol i ters.
The separati on was made of a formaldehyde solu-
The
A thermal conducti vi ty detector i s used here because i t responds to al l three
Note that the order of el uti on f or the three components
components. A flame detector i s less sui tabl e since water i s not seen at al l and
formaldehyde only poorly.
di f f ers from that obtained wi th porous polymers.
TRACE ANALYSIS
Trace Gases
Kaiser2 has shown that trace anal ysi s i s possi bl e wi th Carbosieve B, and that
i t is possi bl e to detect parts per bi l l i on of hydrocarbons using reversi on gas chro-
matography. This technique consi sts of pul l i ng a l arge, known volume of the sample
through the column, al l owi ng the i mpuri ti es to col l ect. Then carrier gas i s turned
on and the column i s heated, by a moving oven, which causes the sample to move through
the column. This technique allows f or the sample t o be hi ghl y concentrated i n the
column.
Since Carbosieve B does not bl eed, it can be used wi th ul tra hi gh sensi ti vi ty
detectors such as helium i oni zati on.
TRACE WATER ANALYSIS
Kaiser' has shown that trace quanti ti es of water can be determined using a
Carbosieve B column. Because of the l ow af f i ni ty f or water of i ts carbon surface,
water i s quickly el uted from a Carbosieve B column wi thout tai l i ng even before
methane.
from bei ng adsorbed on the wal l of the column tube.
is less adsorpti ve than gl ass or stai nl ess steel tubi ng
To carry out trace water anal ysi s, care must be taken to prevent water
Kaiser has shown that quartz
SULFUR GASES
Sul f ur di oxi de is readi l y el uted f r m Carbosieve B columns wi thout i ndi cati on
9
of tai l i ng as shown in Fi gure 8.
to 30 ppm. Bel ow that l evel the SO2 i s l ost on the column.2 Consequently, Carbosieve
B can be recommended onl y f or hi gher concentrati ons of SO2.
badl y at rel ati vel y hi gh concentrati on s o Carbosieve B cannot be recommended f or
H S anal ysi s at this ti me.
Kaiser has found that SO can be determined dawn
Hydrogen sul f i de tails
2
OXIDES OF NITROGEN
Both ni trous oxide (N 0) and ni tri c oxide (NO) can be separated from each other
as wel l as other gases wi t 2 Carbosieve B. No i nformati on of ni trogen di oxi de (NO2)
i s avai l abl e as yet.
Ni trous oxide is readi l y separated from other compounds and is el uted af ter
CO and bef ore acetyl ene.
Thfs separati on was carri ed out at 100C wi$h a 3 f t. x 118'' stai nl ess steel column
wi th 60180 mesh Carbosieve B.
Figure 9 shows N 0 separated from ai r, methane and C02.
32
Ni tri c oxide is el uted from Carbosieve more rapi dl y than N 0. A t room tempera-
2
ture i t i s qui ckl y el uted af ter N
a 6 ft. x 1/8" stai nl ess steel co?hmn wi th 60180 mesh Carbosieve B.
and before CO. This i s shown i n Figure 10 wi th
Speci al precauti ons must be observed when NO i s to be separated. The column
must be conditioned wi th hydrogen to prevent the NO peak from tai l i ng. This can be
done wi th ei ther pure hydrogen or a mixed gas such as ni trogen or helium and hydro-
gen.
f or condi ti oni ng.
essenti al l y the same resul ts.
wi th trace quanti ti es.
We have tri ed a 94% N2-6% H2 mixture and found thi s comparable to pure hydrogen
It should be noted that none of the work with NO deal t
Conditioning f or one hour at 200, 300 and 400C seemed to gi ve
I f the column i s to be operated at room temperature f or NO and then temperature
programmed to hi gher temperatures, the carri er gas should contai n hydrogen.
column i s heated wi thout hydrogen i n the carrier gas, then the NO peak wi l l tai l .
This was found wi th hel i um as the carrier gas.
ri orated f or NO when the column was repeatedl y temperature programmed to 200C to
el ute other components.
expl oratory, and at thi s poi nt we are not certai n that i t can be reproduced.
I f the
We found the column gradually dete-
Our work wi th NO and the hydrogen deacti vati on has been
Both NO and N20 can be separated from 02, N2, CO, CO and methane with a 6 f t.
2
x 118" column. The column i s operated at room temperature ti l l CO i s el uted, then
the column is programmed qui ckl y to 150C to el ute methane, C02 and N20. When the
column i s programmed, the carri er gas must be dry.
MISCELLANEOUS RETENTI ON DATA
Data prepared by Kai ser i n Table 1 compares Carbosieve B, Porapak Q and Molec-
ul ar Sieve SA. Note that the val ues are retenti on volumes, not retenti on ti mes.
Water is shown to el ute bef ore methane and CO on the Carbosieve B but it i s evi dent
that i t takes qui te some t i me f or butane to be el uted, even at 250C.
2
The data i n Table 2 were obtai ned i n our l aboratory usi ng a 3 f t. x 118" O.D.
stai nl ess steel column packed wi th 60180 Carbosieve B and operated at 40mlfmin. and
are absol ute retenti on times.
HANDLING OF CARBOSIEVE B
Carbosieve B is suppl i ed i n f i f teen cc evacuated, seal ed vi al s. when the vi al
i s opened, unused material should be transf erred to a gl ass bottl e and ti ghtl y closed.
To avoid contamination, Carbosieve B should not be exposed to the atmosphere f or pro-
longed peri ods of ti me. Because of i ts hi gh surface area, it tends to readi l y pick
up contaminants from the air; these can be removed but i t requi res extensi ve condi-
ti oni ng of the column. To avoi d oxi dati on and damage to i ts surface, do not heat
Carbosieve to over 2OO0C i n air; do not heat columns above thi s temperature i f the
carrier gas i s not f ree of oxygen. Columns that are to be temperature programmed
shoul d be conditioned f or a number of hours at or sl i ghtl y above the hi ghest tem-
perature anti ci pated.
f
Carbosieve B shoul d be used wi th a carrier gas that is f ree of oxygen to avoid
oxi dati on of i ts surface.
Oxygen can be removed catal yti cal l y by means of speci al l y designed traps which are
i nstal l ed i n the carrier gas l i ne.
I f oxi dati on shoul d occur, the peak w i l l show tai l i ng.
I f temperature programming is used, i t i s imperative that the carri er gas al so
be dry. Use of a devi ce such as a molecular si eve trap to remove water is recom-
mended. I f the carrier gas i s not dry, an i rregul ar base l i ne wi l l be obtained when
the column i s programmed wi th a broad peak f or water seen i n the vi ci ni ty of the co2
33
peak. This i s a problem wi th a thermal conducti vi ty detector, but not a flame which
does not respond to water.
A dryi ng' cartri dge should be used t o remove water from the carrier gas. One must
assume that al l carrier gases have somemoi sture i n them.
to i nstal l a s mal l dryi ng trap af ter the flow control l er i n the i nstrument.
seen i n several i nstances where it is the flow control l er bl eedi ng moi sture that causes
the base l i ne problem even though the carrier gas was being dri ed. One can also heat
the carrier gas tubi ng from the cyl i nder to the chromatograph to remove moi sture from
the wal l s of the tubing, but thi s i s unnecessary i f the trap i s pl aced af ter the con-
trol l er.
Once thi s is done, heat the column to the maximum temperature which you expect t o reach
and hold the column at that temperature f or several hours t o condi ti on it.
We al so f i nd i t necessary
We have
A short secti on of tubi ng f i l l ed wi th Molecular Si eve 5A is adequate f or thi s.
The problem of moisture i n the carri er gas is i l l ustrated i n Fi gure 11 which
The si ze and posi ti on of the water peak w i l l shows the peak f or water af ter C02.
depend on the amount of water present.
REFERENCES
(1) Kaiser, R., Chromatographia 2, 453 (1969)
(2) Kaiser, R., Chromatographia 3, 38 (1970)
3) Zl atki s, A., Kaufman, H. R. and Durbin, D. E., J . Chromatog. Sci . 8, 416 (1970)
4) Kaiser, R. , Private Communication
\
34
TABLE 1
Molecular Si eve
Carbosieve B Porapak Q 5A
O2
N2
c02
CH4
c2H2
c0
ZH4
ZH6
C3H8
qH1O
gH12
CHOH
CH30H
C2H50H
H2
150C
1.22
1.22
1.75
9.09
4.71
21.6
40.5
64.1
-
-
-
27.1
65.7
-
4.59
250C
-
-
-
1.80
1.03
4.08
6.59
9.42
54.9
126 .O
-
3.96
8.35
38.3
0.51
l5ooc -
0.28
0.28
0.28
0.78
0.40
1.11
1.04
1.23
3.09
7.83
16.72
3.78
5.59
12.35
2.61
Speci f i c Retention Volume Vg at 150C
Methane 3 . 0 ml f p
Ethane 47.4ml/gm
Propane 506. l ml f gm
15OoC
0.32
0.56
1.34
1.01
43.0
-
63.9
7.36
-
-
-
-
-
-
-
I n the tabl e above (prepared by Kai ser) retenti on volumes f or a number of compounds
are compared .for several types of columns. I n al l cases column l ength was one meter.
35
TABLE 2
RETENTION TIMES (ABSOLUTE)
HZ
O2
N2
m4
c02
N2
H2S
s02
Argon
NO
CO
Acetylene
E thy l ene
Ethane
Methyl Acetylene
Cyclopropane
Allene
Propylene
Propane
Water
Formaldehyde
Methanol
Column Temperature
Ti me i n Minutes
l0O0C 15OoC 2000c
0.23
0.23
0.23
0.23
0.23
0.23
0.5
0.92
1.45
2.4
4.7
5.4
9.8
-
-
-
-
-
-
-
-
-
0. 2
0. 2
0. 2
0. 2
0. 2
0. 2
0.3
0.4
0.6
0.85
0.450
1.6
2.6
5.0
-
12.8
15.7
-
-
0.15
0.35
1.0
0.15
0.15
0.15
0.15
0.15
0.15
0.15
0.27
0.3
0.4
0.3
0.7
1.0
1.6
3.7
4.2
4.5
6 .O
8.6
0.15
0.2
0.45
Column: 3 f t. x 118'' S.S., 60180 mesh Carbosieve B Flow: 40mlfmin.
36
_ .
YlNUlES
FIGURE 7 - Water-FormaldehydeMethenol Separation
0 2 4 6 8 10
MINUTES
FIGURE 3 - Separation of O1 N2 CO. CH4 and C02
FIGURE 2 - Trace Acetylene in Ethylene
METHANE
;I
I I ETHYLENE
ACETYLENE I ' !
.5 1.0 1.5 ' 2.0 2.5 0
r
MINUTES
FIGURE 4 - Separation of Methane and Cz's
I
i
/
I
0 1 2 3 4 5 6
MINUTES
FIGURE 5 - Separation of C,'s, C2f and C3's
FIGURE 6 - Separation of n-Paraffins - C, - C4
c I '
0 I 2 MINUTES
FIGURE 7 - Separation of Air, CO, CH4 COz C p z Cp, and
'p6
FIGURE 8 - Separation of Ai r and soz
38
1
0 I 2 MINUTES
FIGURE 9 - Separation of Ai r. CH4 C02 and N20.
2
i
0 1 2 3 MINUTES
FIGURE 10- Separation of Nr NO and CO
. ..
I , 1 1
I 2 3 4 5 6 7 YINUTES
FIGURE 11 - Separation of Or Nz CO, CH, COz and H20.