Application Note 228-387: GC/TCD Analysis of a …

1 GC/TCD Analysis of A Natural GasSample on A Single HP-PLOT Q ColumnAuthorZhenghua JiAgilent Technologies2850 Centerville RoadWilmington, DE 19808 Key WordGC/TCDN atural gasPLOT Q columnAbstractAn Agilent 6890 series gas chromato-graph (GC) equipped with a TCD(thermal conductivity detector) wasused with and HP-PLOT Q capillarycolumn for the Analysis of a naturalgas sample. Over 70 sequential runsshowed good separation for a widevariety of analytes with good gas is an important energysource and widely used as a startingmaterial for many chemical process-es. It contains mainly methane anddifferent levels of other hydrocar-bons and fixed gases such as nitro-gen, helium, and carbon heavier than C7 areusually present at ppm sulfide and other sulfurcompounds may be present, eithernaturally or as added components may includepolar compounds, such as low levelsof water, and small amounts ofmethanol and/or glycol which mayhave been added for processing pur-poses [1,2].

2 Natural gases from differ-ent sources usually have the samecomposition but different concentra-tion GC analyses, the variety of compo-nents in natural gas requires theseparation of both polar/non-polarcompounds. Multi-dimensional GC isoften required since no single columncan separate this wide variety of natural gas constituents. Nor can asingle detector detect all compoundssatisfactorily. Specifically, the separa-tion of fixed gases and water fromhydrocarbons is very difficult toobtain on most wall-coated open-tubular (WCOT) columns; and TCDhas a limited sensitivity for tracelevel compounds and GC coupled withswitching valves requires the use ofseveral different types of PLOT columns [2-4]. The HP-PLOT Al2O3column [3-4] is often used for hydrocarbon separations and thedetermination of BTUs. The HP-PLOT MoleSieve column is used forthe separation of fixed gases such asoxygen, nitrogen and helium, andeven argon, [3-4] from methane.

3 Andthe separation of polar and activecompound such as water, CO2, andodorants is obtained using a porouspolymer PLOT column, mostly Qtype [3].All three columns are connected byone or more multiple port valves andthe complete separation is obtainedby time switching eluents to eachcolumn and detector. Backflushinghydrocarbon compounds heavierthan C7is necessary in most the column interchange and connection as well as the valveand time switching make this a difficult technique to use for ideal approach would be one-dimensional GC. Natural gas analysisusing two parallel connected PLOT columns has been done with the suc-cessful separation of hydrocarbonsand oxygen and nitrogen [4].However, this method is limitedbecause the separation of polar com-pounds from hydrocarbons cannot beApplication Note228-3872prolongs the Analysis time for hydro-carbons heavier than C7and/orrequires backflushing of the hydro-carbons.

4 The limited resolution ofnitrogen/air from methane requireslow starting temperatures (@ 40 C)which increases Analysis time andaffects the accuracy of the a fraction of the nitrogen, carbon dioxide, and methane peaksoverlap, the concentration of methanewill be incorrectly HP PLOT-Q columns overcomesome of these problems making themsuitable for natural gas analyses. Thisapplication note examines a simpleGC/TCD mehtod for the Analysis ofnatural gas using a new HP porouspolymer, Q-type, PLOT column. Theresolution of nitrogen and carbondioxide from methane on differentcommercially available PLOT columns is compared and repro-ducibility and reliability are chromatography Analysis of anatural gas sample was done using anAgilent 6890 series gas chromato-graph (GC) with electronic pneumat-ics control (EPC) and a ThermalConductivity Detector (TCD).

5 Forconventional gas Analysis , a six-portvalve with an cc sampling loopwas used to introduce natural gassample onto the HP-PLOT Q columnin split mode (split ratio 18:1). TheGC parameters are listed inTable natural gas sample supplied byScott Specialty Gases, Inc,(Plumsteadville, PA) was used andthe original compounds and concen-trations are listed in Table 2. Thissample was modified by addingmethanol, water, and hydrogen sulfide. During Analysis , the possibleleaking of some air in the samplingloop may also have caused somechange in were run using an HP-PLOTQ porous polymer column (part num-achieved on these two kinds of PLOT columns. Additionally, water, CO2,and odorants deactivate Al2O3andmolesieve PLOT column , these interactions causeshifting of retention times therebyaffecting the repeatability, reliability,and accuracy of the natural gas polymer, Q-type PLOT columns combine the separation fea-tures of the Al2O3 PLOT and mole-sieve PLOT columns when separatingfeatures of the Al2O3 PLOT and mole-sieve PLOT columns when separatingalkanes and fixed gases.)

6 The PLOT-Qcoating overcomes the reproducibilityproblem caused by polar compoundsdeactivating Al2O3and molesieveabsorbent coatings in natural , PLOT-Q columns canseparate CO2, water, and odorantsfrom an alkanes matrix. Thus, theanalysis of natural gas on PLOT-Qcolumns will satisfy most the separa-tion requirements from BTUs throughhydrocarbon components and polarcompound , there are also some prob-lems associated with the use of PLOT-Q columns for natural gas gas (such as air, CO, and noblegases) cannot be separated on PLOT-Q columns at above ambient tempera-tures. The upper temperature limitsare usually low (250 C) for most com-mercial PLOT-Q some commercial PLOT-Qcolumns, loose particle binding in thecoating, high column bleed, and thelimited resolution of nitrogen/air frommethane are major problems restrict-ing their usefulness in natural gasanalyses.

7 Loose particle binding inthe coating causes baseline spikingwhen the sampling valve is operatedor fast temperature ramping is column bleed makes thesecolumns useful only at temperaturesbelow 250 C and this situation GC 6890 GC with mm x 30 m PLOT-Q columnsCarrierHelium ml/min @ 60 C, Constant flow modeOven60 C (2 min) 30 C/min to 240 C (1 min)InjectionSplit mode, 250 C, cc sampling loopeSplit flow150 ml/minValveValco 6-port valve, cc sampling loopDetectorTCDR eference flowHelium, 30 ml/minAuxilary gas flowHelium, 3 ml/minTable 1. GC Experimental ConditionsCompoundConcentration (v/v%) 2. Natural Gas Sampleber 19095P-QO4) with two otherbrands (X and Y) of Q-type PLOT columns used for resolution compar-isons. All columns were conditionedat 250 C overnight per manufacturerrecommendation to reduce and DiscussionsHP-PLOT Q type columns are coatedwith porous polymer particles madeof divinylbenzene and ethylvinylben-zene and can separate hydrocarbonsup to C14as well as some polar com-pounds.

8 Their upper isothermal andprogramming temperature limits are270 C and 290 C, separation of the constitutents inthe natural gas sample was doneusing a porous polymer HP-PLOT Qcolumn as shown in Figure 1. Theanalysis time for this run was 9 min-utes. Hydrogen sulfide, water, andmethanol were well-separated fromethane, propane and baseline spiking is com-monly associated with this analysisfor some commercially availablecolumns, no baseline spiking wasobserved with the HP-PLOT Q column, indicating that the stationaryphase of this PLOT column providesexcellent immobilization that canwithstand: fast oven temperatureramping (30 C/min), a pressurepulse generated from valve actua-tion, and carrier gas pressure ramp-ing at constant flow mode. Resultantcolumn bleed was very low. Limited resolution of nitrogen andcarbon dixoide from methane isobtained using most commercialPLOT-Q columns.

9 To evaluate the res-olution of the new HP-PLOT Q col-umn (Figure 1), an HP-PLOT Q col-umn and two other brands of PLOT-Qcolumns (brand X brand Y) werecompared. All columns were mminternal diameter. The natural gassample size was cc with a splitratio of 18:1. Peak resolutions (Rs)were calculated based on the formu-lae in (1) and the results listed inTable 3. Where taand tbare the retention times of peaks A and B Wa(1/2)and Wb(1/2)are theirpeak widths at half height, resolution for N2-air/methaneusing the HP-PLOT Q column wasgreater than which is the conven-tional requirement for base lineseparation, even at a 60 C initialoven temperature. The resolution ofcarbon dioxide from methane at60 C on the HP-PLOT Q column is40% higher than the same resolutionon the other two brands of PLOT-Qcolumns tested. This separationcapability of the HP-PLOT Q columnalso can sufficiently resolve nitrogenand carbon dioxide from methane,even if the methane peak is tailingdue to sample overload.

10 The startingtemperature of 60 C also results in a30% reduction in GC cycle of the concerns associated withusing PLOT columns for natural gasanalysis is reproducibility. It is wellknown that when using aluminaPLOT and molesieve PLOT columns the retention times for hydrocar-bons shift due to deactivation of Rs(A/B) =2* (tb- ta) * (Wa(1/2)+Wb(1/2)02468 minMethaneN2-AirCO2 Figure 1. Separation of Natural GasEthaneH2SH2OC3 Methanoli-C4n-C4neo-C5i-C5n-C5C6C7 mm x 30 m, HP-PLOT QCarrier:Helium ( ml/min @ 60 C)Oven:60 C (2 min) 30 C/min to 240 C (1 min)Detector:TCD 250 CInjection:250 C Split mode (18:1) cc natural gas sample, methane, 80% +Resolution RSHP-PLOT QBrand XBrand YRs(N2-Air/Methane, 40 C) (N2-Air/Methane, 60 C) (CO2/Methane, 40 C) (CO2/Methane, 60 C) 3. Resolution Comparisons (Sample and size, natural gas, cc)CompoundAverageRSD% 4.)