Optimized Determination ofC1–C6 Impurities in Propyleneand Ethylene UsingHP-PLOT/Al2O3 Columns

AuthorsZhenghua Ji, Imogene Chang,

and Alan Broske

AbstractImpurities in ethylene and propy-

lene (light hydrocarbon gases C1

to C6) were analyzed using HP-

PLOT/Al2O3 columns. In compari-

son with other commercially

available alumina PLOT columns,

the HP-PLOT columns showed

better selectivity, increased

capacity, and complete baseline

separation. HP-PLOT column

analyses were done using the

5890 Series II gas chromatograph

with electronic pressure control

and a flame ionization detector.

IntroductionTrace amounts of important impuri-ties in high-purity hydrocarbon prod-ucts, such as propylene and ethylene,are frequently masked by sampleoverload. Although standard commer-cially available PLOT columns1,2 pro-vide good separations for regular gasanalysis, until now they have notbeen able to completely separate C1to C6 impurities in high-purity propy-lene or ethylene with good resolution.This has limited the commercialapplication of alumina PLOTcolumns.

In comparison to these aluminaPLOT columns, the HP-PLOT/Al2O3columns showed baseline separa-tions of all C1 to C6 hydrocarbonimpurities. Improved selectivity formajor components of C1 to C6 hydro-carbons was demonstrated includingethylene, propylene, acetylene, and1,3-butadiene. In addition, in compar-ison to the other alumina/KCL andPLOT columns studied, the HP-PLOT/Al2O3 columns provided reli-able, repeatable results from run torun and column to column.

ExperimentalFour alumina PLOT columns wereused in the analyses of hydrocarbonimpurities in ethylene and propylene:

· HP-PLOT/Al2O3, S-deactivated, 50 m x 0.53 mm x 15 µm (part no. 19095P-S25)

· HP-PLOT/Al2O3, M-deactivated, 50 m x 0.53 mm x 15 µm (part no.1905P-M25)

· Alumina/KCl, 50 m x 0.53 mm x 10 µm

· Brand X, 30 m x 0.53 mm

Gas chromatography analysis of C1to C6 light hydrocarbon gases wasdone on a 5890 Series II gas chro-matograph (GC) with electronic pressure control (EPC) and a flame

ionization detector (FID). An Agilent7673 automatic sampler was used todeliver 5 µl of gas samples under asplit mode (split ratio 25–30:1). Tosimulate overloading conditions,manual or splitless injections withgas-tight syringes were used wherev-er appropriate. GC parameters arelisted in Table 1.

Except Figure 4, where a C1 to C6gas mixture was used, samples wereprepared by spiking polymer-gradepropylene (99.6% + purity) and ethyl-ene (99.8% + purity) from ScottSpecialty Gases Company with ahydrocarbons mix (RustyAssociates). The hydrocarbons mixcontained 23 components of C1 to C6hydrocarbon gases listed in Table 2.

Results and DiscussionAlumina PLOT columns have beenfrequently used to analyze C1 to C6light hydrocarbon gases due to theunique retentive properties of alu-minum oxide. However, traceamounts of critical impurities in high-purity products, such as polymer-grade propylene and ethylene, are difficult to quantify using alumina/KCl and Brand X alumina PLOTcolumns because these impurities areoften masked by sample overload.Furthermore, the selectivity of thesePLOT columns varies due to the variation in test conditions such asmoisture and carrier gas flow.

Application Note228-263

Columns and Supplies

2

Table 1. Experimental Conditions

coeluted with isobutane (peak 7) asa result of increased carrier flow(Figure 1B).

For Brand X PLOT columns, mask-ing and coelution problems alsooccurred. As shown in Figure 2A,the large tailing peak of propylenereduced the resolution of the propy-lene/isobutane pair, and isopentane(peak 15) coeluted with cis-butene(peak 14).

1. Methane 13. Isobutylene2. Ethane 14. Cis-2-Butene3. Ethylene 15. Isopentane4. Propane 16. n-Pentane5. Cyclopropane 17. 1,3-Butadiene6. Propylene 18. Propyne7. Isobutane 19. trans-2-Pentene8. n-Butane 20. 2-Methyl-Butene9. Propadiene 21. 1-Pentene

10. Acetylene 22. cis-Pentene11. trans-2-Butene 23. Hexane12. 1-Butene

In Figure 1A, cyclopropane (peak 5)was masked by the major analyte,propylene (peak 6), on the aluminaKCl column. In this run, 1,3-butadi-ene (peak 17) coeluted with n-pen-tane (peak 16) as a result of 1,3-buta-diene overload. When propylene wasnot overloaded (Figure 1B), cyclo-propane was separated from propy-lene. However, acetylene (peak 10)

Table 2. Gas Mixture of C1 to C6

Additionally, overloaded ethylene(peak 3) distorted the propane peak(peak 4) and split the latter into twoghost peaks. However, these prob-lems were not observed when over-loading did not occur (Figure 2B).

Analysis of Impurities inPropyleneFigures 3A and 3B are chro-matograms of propylene sample ana-lyzed using the HP-PLOT/Al2O3columns, S-deactivated and M-deactivated, respectively. Traceamounts (60 ppm) of cyclopropane(peak 5) were separated from propy-lene (92% purity). Baseline separa-tion was achieved for isobutane(peak 7) and propylene. All 23 com-ponents were well resolved in sharpsymmetrical peaks within 24 min-utes. Both Agilent PLOT columnsexhibited excellent separation foracetylene (peak 10) from n-butane(peak 8), for isopentane (peak 15)from cis-2-butene (peak 14), and forn-pentane (peak 16) from 1,3-butadi-ene (peak 17).

Analysis of Impurities inEthyleneFigure 4 shows the analysis of anethylene sample using the HP-PLOT/Al2O3 S-deactivated column.The concentration of ethylene, basedon peak area percentage, is 98.5%.

Within 24 minutes, all 22 impuritiesassociated with ethylene were com-pletely separated, including methane(peak 1), ethane (peak 2), andpropane (peak 4). Distortion of thepropane peak (Figure 2A) was notobserved. In this run, 1.5 cc of sam-ple was injected, demonstrating thelarge loading capacity of the HP-PLOT Al2O3 columns. Although notshown here, similar resolutions forethylene impurities were alsoachieved using the M-deactivated HP-PLOT/Al2O3 column. Both Agilentalumina PLOT columns met all of therequirements for impurity analysis ofpolymer-grade ethylene.

Retention Index andSelectivityRetention indexes (RI) are importantindicators of column selectivity.Table 3 lists the RIs of ethylene,propylene, and acetylene for the four alumina PLOT columns underevaluation.

Table 3. Retention IndexesRI RI RI

Column (Ethylene) (Propylene) (Acetylene)

Alumina/KCI 248 349 372

HP “S” 248 361 418

HP “M” 254 369 407

Brand X 263.5 380 417

High RIs represent high affinitytoward a specific “analyte.” In thecase of ethylene, the RI should bearound 250 to minimize the impact ofethylene overload on the impuritieseluting immediately after ethylene.The RI of ethylene for Brand X lieson the high end (263 verses 248 forHP “S” and the alumina/KCl). As aresult, peak distortion of propane islikely to occur.

The RIs for propylene should begreater than 350 to achieved baselineseparations of both cyclopropane and isobutane from propylene. Forthe Al2O3/KCl column, the cyclo-propane peak was masked by propy-lene because the RI for propylene istoo close to 350. On the other hand,

Gas chromatograph: Agilent 5890 Series II with EPCCarrier: He, constant flowOven: TP1 = 35°C (2 min), 5°C/min to 100°C, 10°C/min to 180°C (5 min)Injection: Split/splitless inlet, inlet: 250°CDetector: FID (250°C)

Figure Column Carrier Oven Injection Injection ModeNo. Type Flow Temperature Volume (Split Ratio)

1A Al2O3/KCI He 35 cm/s TP1 5 µl 0

1B Al2O3/KCI He 45 cm/s TP1 5 µl 0

2A Brand X He 42 cm/s TP1 0.5 cc 30:1

2B Brand X He 42 cm/s 100°C 5 µl 25:1

3A HP “S” He 35 cm/s TP1 5 µl 0

3B HP “M” He 38 cm/s TP1 5 µl 0

4 HP “S” He 42 cm/s TP1 1.5 cc 30:1

1

1

5

64

3

211

109

8

7

141312

15

16, 17, 18

19

202122 23

5 10 15 25 min

low resolution for the propylene/isobutane pair on the Brand X col-umn can be attributed to its high RIfor propylene (RI=380). Because theRIs of the HP-Al2O3/PLOT columnslie in between 350 and 380, excellentseparations were achieved for thecritical impurities, cyclopropane, andisobutane.

For the quantitation of acetylene, theRI should be greater than 400. Asshown in Figure 1A and 1B, isobu-tane and acetylene coeluted on thealumina/KCl column because its RIfor acetylene was below 400.Coelution of isobutane and acetylenewas not observed for HP-Al2O3/PLOTand Brand X columns because theirRIs for acetylene are greater than400. Furthermore, acetylene is moresensitive to small variations in GCparameters (e.g., column tempera-ture, column flow, and moisture incarrier gas) than impurities elutingnear acetylene. As a result, coelutionor change in elution order for acety-lene and these impurities may occurif the column’s RI for acetylene is notin the right range (slightly greaterthan 400).

The differences for selectivity of ethylene, propylene, and acetylenebetween the HP-PLOT/Al2O3 M- andS-deactivated columns are clearlyrepresented by their RIs. Accordingly,the resolution of impurities in propy-lene for both these columns is slightly different.

ConclusionThe new HP-PLOT/Al2O3 columnsshowed excellent selectivity, sampleloading capacity, and efficiency forthe analysis of impurities in ethyleneand propylene. Excellent resolutionwas achieved for impurities such aspropane, cyclopropane, isobutane,and acetylene in ethylene and propylene.

5

3

2

6

4 7, 10

11

9

8

1214

13 15

19

17, 18

16

212022 23

0 20

5 10 15 25 min0 20

1A

1B

Figure 1. Alumina/KCl PLOT column. (See Table 2 for peak identification.)

Figure 2. Brand X alumina PLOT column. (See Table 2 for peak identification.)

1

4

3

62

14, 15

5

8

17

10 min

18

916

2 4

12

1110

2 3

109

5

48

76

1611 12

1323

22

21

20

0

0 5 20 min

1314

15

1718

16 7

8

19

2A

2B

3

10 15

4

0

1 2 43 6

5

7 8

9 10

1612

13

14

11

15

17

18

15

19

20

21

22

23

5 10 20 25 min

Figure 4. HP-PLOT/Al2O3 columns, S-deactivated. (See Table 2 for peak identification.)

18

19

1

5

3

2

64

109

8

12

13

14

11

15

18

20

22

23

7

16

19

17

21

3A. S-Deactivated

3B. M-Deactivated

Figure 3. HP-PLOT/Al2O3 columns. (See Table 2 for peak identification.)

0 5 10 15 20 25 min

0 5 10 15 20 25 min

References1. R. C. M. De Nijs and J. De Zeeuw,“Aluminum Oxide-Coated Fused SilicaPorous-Layer Open Tubular Column forGas-Solid Chromatography of C1 to C10Hydrocarbons,” J Chromatog 279:41–48,1983.

2. J&W Catalog, 1993.

For more information, call your

local Agilent Technologies office,

and ask for an analytical

representative or go to our web

site at www.agilent.com

17

23

2221

2016

14

13

12 15

11

109

876

5

4

3

2

1

19

Copyright © 1994, 2000 Agilent TechnologiesPrinted in USA 04/00

Printed on recycled paperPublication Number 5962-8417E