20 Gases&InstrumentationMay/June 2014

F E A T U R E

Separation of Gases and Volatiles Using Gas Chromatography and PLOT Columns

By Jaap de Zeeuw

Adsorbents are capable of doing separations that cannot be done with liquid phases. Most adsorbents are available in capillary columns, allowing the combi-nation of high efficiency with selectivity. Limitations of adsorbents with respect to adsorption and particles in capillary columns have to be taken into account.

Introduction

I f gases or volatiles are analyzed, the first challenge is to retain such components using a suitable stationary phase. Retention can be realized by using very thick films of liquid

stationary phases. Very thick films of 100% polydimethyl siloxane, like Rtx-1, will provide some separation (see Figure 1). However, this often does not provide enough retention or selectivity to obtain the required separations. One can increase film thickness to increase retention, but efficiency will be compromised as the peak broadening will increase rapidly because of resistance to mass transfer in the liquid phase. This is the reason why films thicker then 7 microns are not often used in capillary separations.

One can also choose adsorption as a retention mecha-nism. With adsorption chromatography, the analysis can be done at higher temperatures and individual selectivity is possible depending on the type of adsorbent chosen.

Adsorption SeparationsAdsorption is a surface interaction process that is possible using solid stationary phases. One needs a high surface area to obtain sufficient retention and capacity. Instead of liquid, a solid is deposited inside the capillary column. To obtain sufficient retention, a layer of 5-50 microns is typi-cally applied. Such columns are generally known as porous layer open tubular (PLOT) columns. The layer is formed by particles the size of 1 micron or smaller, or by an integrated porous structure. This type of column was already introduced commercially 30 years ago and in today’s applications PLOT columns have been proven. The first type of PLOT column was based on alumina (aluminum oxide Al2O3) that was designed for the separation of hydrocarbons in the C1-C5 range. Later, several other types of PLOT columns were com-mercialized, each with a specific application field.

AluminaAlumina is a material used in the stationary phase because it is highly selective for hydrocarbons. With alumina all hydro-carbon isomers can be separated. To reduce and control the activity of the alumina, it has to be deactivated with an inorganic salt. The first alumina columns were deactivated using KCl and Na2SO4 salts. Later alumina PLOT columns also became available, showing optimized responses for polar hydrocarbons like propadiene and acetylenes. Such columns were commercialized with an acronym like Alumina MAPD. The fact is that alumina offers the best possible selectivity for C1-C5 hydrocarbon separations (see Figure 2).

The downside of alumina is that retention is strongly impacted by polar components that may be present in the sample. The major troublemaker is water. If water is intro-duced into an alumina column, the water will cover the active sites and retention times will reduce. Selectivity of the alumina will also change as peaks will shift relative to each

Figure 1. Separation of C1-C5 on a 100% polydimethyl siloxane film; Column: 60m x 0.32mm Rtx-1 df = 5 µm; Oven: 60 °C, 5 min => 175 °C @ 10 °C /min

21www.gasesmag.com May/June 2014

F E A T U R E

other. There are three ways to deal with this:

■ Heat the alumina column with each analysis to the maximum temperature to elute the water; the newest Alumina BOND MAPD columns can withstand 250°C, which will make water elute faster.

■ Use thick-film polyethylene glycol (PEG) as a pre-column to retain the water. In a PEG type phase, C1-C7 hydrocarbons will elute before the water peak. By using a valve, one can avoid letting the water from entering the alumina. This setup requires a valving system, which has to be tuned.

■ If water levels are low, the alumina can be reconditioned after a number of analy-sis. Typically, this is done if the retention time of target components are running beyond the integration window.

Alumina does work very well for light hydro-carbons, but has limitations. In addition to the impact of water, it will also adsorb many other components like CO2, oxygenates, sulfur gases etc. One must always be aware that the presence of such components in the sample will impact the absolute retention for hydrocarbons. Practi-cally, the biggest impact will be water as this is often present in small levels. Alumina columns are very robust and if a polar analyte is adsorbed by the alumina, one can usually regenerate the

alumina by conditioning it for several hours at its maximum temperature.

Molecular SievesMolecular sieves are adsorbents based on a zeolite structure. These materials have very high surface areas that allow separations of perma-nent gases like nitrogen and oxygen at ambient, or even higher temperatures. There are different molecular sieves available, of which the most useful is the molecular sieve 5 Å. When deposited in a capillary, this material allows a separation of

all permanent gases at a temperature of 30°C. Even argon/oxygen can be base-line separated (see Figure 3). Here is also an example of a metal capillary PLOT column in use. Such metal (or MXT®) are preferred when a high degree of robustness is required, as in process, portable and high temperature applications.

As with alumina, molecular sieve materi-als are also strongly impacted when water is present in the sample. Molecular sieves are also widely used as drying agents, often in gas purification. If water is injected onto a molsieve GC column, the retention will be reduced, just as with alumina. Due to the high surface area, the regeneration of mol-sieves will tolerate a higher temperature. One needs to heat to at least to 250°C to remove the water. Capillary columns can even be heated to 300°C, which then only requires a few minutes to remove the water. One can use the Molsieve 5A PLOT columns to separate hydrogen isotopes as well as hydrogen spin isomers, but then sub-ambient temperatures as low as -180°C are required.

SilicaBoth the adsorbents alumina and molsieves have unique separation characteristics, but they also have clear limitations. Silica is an interesting adsorbent as it shows much wider applicability.

Silica is used widely in liquid chromatogra-phy (LC) and it has shown its unique applica-tion as a reversed phase material. The chal-lenge in LC is that the silica will have a big impact on separations, since the surface will

Figure 2. Hydrocarbon impurities in propylene using alumina; Column: Rt-Alumina BOND / Na2SO4, 50m, 0.53mm ID, 10 µm; Oven: 45°C (1min) to 200°C @ 10°C/min. (hold 8.5 min.); Injection: split; Detection: FID; Peak identification: 1. methane 2. ethane 3. ethylene 4. propane 5. cyclopropane 6. propylene 7. isobutane 8. n-butane 9. propadi-ene 10. acetylene 11. trans-2-butene 12. 1-butene 13. isobutylene 14. cis-2-butene 15. isopentane 16. n-pentane 17. 1,3-butadiene 18. trans-2-pentene 19. 2-methyl-2-butene 20. 1-pentene 21. cis-2-pentene

Figure 3. Permanent gases using molecular sieve 5Å: Column: 30m x 0.53mm MXT-MSieve 5A, df = 50 μm; Oven: 30°C; Injection: split; Detection: μTCD

22 Gases&InstrumentationMay/June 2014

F E A T U R E

always show some interaction with analytes, despite the high coverage of C4-C18 bonds. The coverage, or “density” of surface non-polar groups, depends on many parameters like silica type, density, particle size, poros-ity, pore size, activation, purity etc. In GC, the silica surface activity can be used as a primary separation mechanism, and this is quite prom-ising. Because the surface is very well defined, and the goal is to use adsorption interactions, the activity can be controlled much better compared with its use in LC. The base material used for capillary column manufacture is high-purity silica. This silica is drawn into a capillary using drawing technology that was developed by the fiber optics industry. The difference is that for GC a capillary tube is required and there must be a temperature-stable outside coating. Figure 4 shows the take-up drum of a fused silica capillary drawing unit as is used by Restek corporation. The capillary is coated on the outside with a high temperature stable polymer, usually a polyimide.

Important for PLOT retention is to create a high-purity porous silica layer on the surface. Once that is realized it is possible to do some interesting GC separations. Silica has not only a unique selectivity for light hydrocarbons (see Figure 5), it also elutes halogenated com-pounds as well as trace sulfur compounds like H2S and COS at sub-ppm levels (see Figures 6 and 7). Important separation is the separation of COS from propane/propylene. This is done using selective detection like pulsed flame photometric detection (PFPD) or sulfur che-luminescence detection. Though these detec-tors are selective, if sub-ppm levels have to be measured, chromatographic separation is essential. Figures 6 and 7 show trace sulfurs in propane, and the matrix peak (propane) also

causes a signal. Silica BOND separates the COS and H2S away from the propane (and propyl-ene), making correct trace analysis possible.

Besides this, silica also shows a clear revers-ible interaction with water. Water elutes as a peak and retention times will minimally be affected by water. This is in sharp contrast to alumina or molsieve PLOT columns.

Porous PolymersPorous polymers are the most generic adsorp-tion materials as they are generally very inert.

Their retention is based on adsorption, but these materials elute both polar as well as non-polar components.

The challenge is to make the porous poly-mers stable inside a capillary. Due to their extreme electrostatic behavior, particles do not build stable layers. One needs stabiliz-ers or bonding procedures to make this type of column stable. Some manufacturers even supply the columns with particle traps, because particles can dislodge and move through the column. Such events will cause

Figure 4. Fused silica production at Restek Corporation, showing take-up drum; the yellow color of the tubing is caused by the outside polyimide protective coating

Figure 5. C1-C5 Hydrocarbon separation on a silica BOND PLOT column

Figure 6. Volatile sulfur compounds in propane; Column: 30m x 0.32mm Rt silica BOND; Oven: 35°C, 8 min → 200°C, 10°C/min; Injection: split; Detection: PFPD

23www.gasesmag.com May/June 2014

F E A T U R E

Figure 7. Trace COS on propane; for conditions See Figure 6

Figure 8. SEM picture of surface of Rt U-BOND porous polymer; note the absence of particles

Figure 9. Solvents on R Q-BOND: porous polymers elute a wide variety of solvents. The Q-type is 100% divinylbenzene and is similar to Porapak/Hayesep Q. Column: 30 m x 0.53mm Rt Q-BOND df = 20 μm; Oven: 120°C, 5°C/min → 250 °C; Injection: split; Detection: FID

plugging as well as detector or valve malfunc-tioning. A particle trap is a way to prevent this, but the risk of plugging will still be present. Porous polymer PLOT column layers can also be made of integrated/bonded adsorbents. Figure 8 shows an example of a very rugged, stable, porous polymer column. Here the adsorption layer is not made of particles, but forms an integrated layer. Such columns are very stable and offer a practical advantage as particle traps are not required.

Porous polymers are hydrophobic in nature, meaning they have very little affinity with

water. Water will not influence retention times. Water elutes as a nice peak, usually at a very short retention time.

Porous polymers can therefore be used for a wide variety of applications. Figure 9 show a series of solvents, indicating the different classes of compounds that can be analyzed.

Porous polymers also elute so-called “tough” components. Figure 10 shows the elution of water, methanol and formalde-hyde on a polar porous polymer column. Peak shapes are very good considering this is adsorption chromatography.

Challenges of PLOT ColumnsPLOT columns do provide unique separa-tions and they have found wide application. The retention still remains on adsorption interactions, and one has to be aware that some reactivity can be present for sensitive compounds. One other thing that has to be kept in mind is that PLOT columns are made of layers of adsorbents. This can be layers of particles, or integrated layers of so-called in-situ manufactured adsorbents. The reality is that it is always possible that some particles or sections of the adsorption layer are released and may cause a problem. With flame ionization detectors (FID) this may result in spikes (see Figure 11). When an eluting particle hits the electrodes, it will result in a signal. In FID this is generally not a problem, but in thermal conductivity detec-tors (TCD) it can contaminate the detector and peaks may be distorted as a result of activity in the detector.

To reduce the risk of particle release, one can couple a PLOT column with a particle trap. This is a 2-3 m length of polydimeth-ylsiloxane (PDMS) coated capillary coupled at the outlet of a PLOT column that traps dislodged particles. Newer generation PLOT columns, like Silica BOND, Rt Q and U-BOND, and Alumina MAPD do have increased stabil-ity, but still, less than gentle handling may dislodge particles.

It is recommended when using any type of PLOT column, when received, to condi-tion the column at twice its optimal flow and run a temperature program to its maximum temperature. Any particle dislodged during transport to the laboratory will elute. This way the PLOT column will perform very well in the application, even when valves are used.

24 Gases&InstrumentationMay/June 2014

F E A T U R E

MS ApplicationPLOT columns can also be used in MS. One

can use a 0.25mm inner dimension PLOT or a larger diameter. Here one also needs to be

aware of a possible release of particles. The vacuum in the MS presents another chal-lenge. The PLOT columns that are available today are very stable, but when exposed to a vacuum the adsorption layer may disinte-grate, resulting in particles that can enter the MS. For the MS detector, there will be no problem as principally only ions will be entering the system. The vacuum pump is the biggest challenge for the particles as it may become damaged.

To prevent elution of particles for the MS application, it is recommend to couple any PLOT column with a pressure reducing sec-tion. Typically, a section of 5m x 0.25mm or 3m x 0.15mm will act as a buffer to deal with the strong pressure change of column outlet and MS inlet. Such tubing is deactivated and will also act as a particle trap, though its primary function is to deal with the pressure gradient. Coupling can be done with universal linkage hardware like PressTights or diameter specific PressTights made by BGB Analytik.

Alternative DevelopmentsThere will always be users who do not like PLOT columns. A practical alternative may be the development of 0.53mm (wide bore capillary) columns, that are packed with adsorbents. It is possible to pack 0.53mm ID metal (MXT) with standard packing materi-als. Such micropacked columns do have a limited plate count, but they offer retention and selectivity and can be operated at flows of 2-4 mL/min . They can also be coiled in any diameter. Because of the 0.53mm ID, they can be used in any standard GC. Figure 12 shows a typical application using a ShinCar-bon type packing material.

SummaryAdsorbents are capable of doing separations that cannot be done with liquid phases. Most adsorbents are available in capillary columns, allowing the combination of high efficiency with selectivity. Limitations of adsorbents with respect to adsorption and particles in capillary columns have to be taken into account. There are practical ways to deal with all issues. G&I

JAAP DE ZEEUW is international specialist Gas chromatoGraphy at restek corporation, middelburG, the netherlands.. he can be reached at +31 118 623 151 or jaap.dezeeuw@restek.com

Figure 11. When PLOT columns are not stable, they will show spikes. Particles can elute, causing contamination, malfunction and column blockage

Figure 12. Micro-packed 0.53mm ; Column: 1m MXT x 0.53mm ID, Shincarbon 80/100; Oven 25 °C; Injection: Split; Detection: µ-TCD

Figure 10. Polar volatiles on Rt U-BOND; Column: 30m x 0.53mm Rt-U-BOND, df = 20μm; Oven: 100°C; Injection: split; Detection: FID; Peak identification: 1. N2 + O2 + CO 2. CO2 3. formaldehyde 4. water 5. methanol