Technical Note KT030606-1: What is Loop Modulation?

Edward B. Ledford, Jr.*, Joel R. TerMaat, and Chris A. Billesbach

Zoex Corporation, 2611 West M. Court, Suite D, Lincoln, NE 68522 (402)475-7640
E-mail Us

Abstract

The heart of a GC x GC instrument is the so-called "modulator", the interface between the first and second columns of the GC x GC system.  Thermal modulators employ rapid thermal trapping and remobilization to "cut" effluent from the first column into a series of injections suitable for rapid chromatography on the second column. The so-called "Loop Modulator" is the most recent version of two-stage thermal modulator design.  It employs no moving parts near the column, employs a single low temperature valve mounted outside the GC oven, provides the narrowest chemical pulses yet observed with any modulation system, and permits in-situ measurement of carrier gas velocity in the modulator tube.  The system described in this paper employs two gas jets and a column loop to effect two-stage modulation.  Various means for cooling the cold jet are possible with this system, including LN₂ heat exchange, Joule-Thompson expansion of high pressure liquid or gaseous CO₂, wet or dry ice heat exchange, or room temperature compressed gas for modulation of low volatility compounds.

Introduction

Two-stage thermal modulation was invented by the late Prof. John B. Phillips, and his then graduate student, Zaiyou Liu, in the late 1980's. [1]  The device underwent a number of transformations [2,3,4] during a development period extending to the year 2000, when gas jet modulators were introduced [5,6] Experimentation with new thermal modulator designs continues in several laboratories, the consensus being that gas jets are the most efficacious approach to multi-stage thermal modulation.  [7,8,9]. (Figure 1). This paper describes a simple thermal modulation system, known as the "Loop Modulator," which achieves the highest performance of known modulator designs, and permits in-situ measurement of carrier gas and analyte velocities within capillary columns.

Experimental

Construction of the Loop Modulator. The loop modulator employs hot and cold jets of gas to effect two-stage thermal modulation.  The two stages are formed by looping a segment of capillary tubing through the path of a single cold jet. The tubing between the two cold spots thus formed comprises a delay loop.

The mechanical assembly of the loop modulator consists of five main components (Figure 1a).  The cold jet subassembly consists of a steel tube of approximately 3.0 mm inside diameter housed in a steel outer tube of some 19 mm outside diameter. The space between the inner and outer tubes is evacuated with a mechanical roughing pump and valved off, thereby forming a vacuum insulated housing around the cold jet.  Vacuum insulation is essential for the introduction to the GC oven of a cryogenically cooled gas stream.  A hot jet subassembly is mounted at right angles to the cold jet by means of machined brackets, which also provide means of holding the modulator tube in the paths of the gas jets.  The modulator tube is housed in a folded metal holder studded with machined "buttons" which mount in slots milled into the bracket structure.  The modulator tube is held in place within the modulator holder by a folded piece of Kapton (Dupont) film, which functions as a spring tensioner.  The modulator tube is installed and uninstalled in a matter of seconds by sliding it into or out of the slotted bracket structure.  A bottom view of the loop modulator (Figure 1b) shows the loop structure more clearly. 

Operation. The cold jet is directed vertically downward (Figure 1a) onto the modulator tube, thereby creating two cold spots each about 3.0 mm long (Figure 1b).  The cold jet runs continuously, and is diverted away from the cold spots by the hot jet stream, which is pulsed on periodically (Figure 1c).  The hot jet gases rapidly heat the cold spots, thereby remobilizing trapped analytes. At the same time, the hot jet defrosts the tip of the cold jet.  The modulation structure is fully exposed to the stirred oven air bath.

The detailed operation of the loop modulator is best described with reference to a schematic drawing (Figure 2), in which the cold jet is represented as a short nozzle pointing downward, while the hot jet is represented by an oval indicating gas flow toward the reader.  The modulator tube is shown in front view, so that the column loop is visible.  In passing twice through the cold jet path, the column loop forms the pair of cold spots necessary for two-stage thermal modulation. The hot jet bathes not only the cold spots, but upstream and downstream portions of the modulator tubes, and the hot jet tip, in heated gas whenever it fires.  These measures prevent frost build up on the cold jet and eliminate thermal edge effects in the vicinity of the cold spots, which would slow the remobilization process and result in peak tailing.

 The key to the loop modulator is the delay loop.  When material trapped in the upstream cold spot is remobilized by the firing of the hot jet, the material enters the delay loop, which is typically 0.6 m to 1.0 m long. It takes a few hundred milliseconds for material released into the delay loop to traverse around the loop.  During that delay, the hot jet is turned off, and the cold spot reforms under the action of the cold jet.  By the time the material in the delay loop reaches the second cold spot, the latter is cold enough to once again trap the material and refocus it, thereby producing the two-stage thermal modulation effect. The next firing of the hot jet releases the sharply focused chemical pulse from the downstream cold spot, and simultaneously admits another pulse of analyte material from the upstream cold spot into the delay loop. Repetitive firing of the hot jet produces a sequence of sharply focused chemical pulses, which carrier gas flow transports to downstream devices such as detectors and/or secondary columns. In effect, the loop modulator is a "continuous heart-cutter" capable of "chopping" the effluent of the first column into a series of nearly ideal injections onto a second, high speed analytical column.

Overall System Construction.  The main components of the loop modulation system are a liquid nitrogen dewar, a vacuum insulated transfer line, a modulator control module, and a jet assembly (Figure 3). There is also a valve assembly controlled by valve ports 1 through 4 of the Agilent 6890 valve control bank. One of the valves operates the hot jet, and is mounted on top of the GC oven. The other three valves are ganged into a group and bench mounted. The first of these three is a master supply valve. The second supplies cooling flow to the secondary oven (not shown in Figure 3), an accessory used for programming the temperature of the secondary column independently of the first. The third valve supplies gas flow to the cold jet, which flows from a regulated source of dry nitrogen, through a heat exchange coil mounted in the liquid nitrogen dewar, and then to the jet assembly via the vacuum insulated transfer line. The cold jet flow is adjustable by means of a manually operated needle valve incorporated into the valve bank.  The valves can be sequenced by the GC, permitting fully automatic operation of the GC x GC system, suitable for autosampler sequences.

Modulation parameters are controlled independently of the GC data system by means of a programmable logic controller, which provides a synchronous phase lock between the start of the GC detector and the start of the modulation sequence.

 Other Cooling Schemes.  It is possible to use Joule-Thompson expansion of high pressure CO₂ gas or liquid to form the cold jet rather than liquid nitrogen heat exchange. This approach simplifies apparatus and reduces the cost of cryogenic consumables, but results in a higher cold jet temperature,   -77 ^oC with CO₂, as compared to -189 ^oC with liquid nitrogen. The higher jet temperature makes it impossible to thermally modulate compounds more volatile than C₉ on uncoated capillary tubing, and C₈ on coated tubing. Nonetheless, CO₂ modulation is useful for higher boiling compounds.  It is also possible to cool the cold jet stream with dry ice or water ice heat exchange. For low volatility substances above about C₂₀, room temperature air is cool enough to effect thermal modulation.

Modulation Test Procedure.  Although the operating parameters of the thermal modulator, e.g., modulation period, hot jet block temperature, pulse duration, and gas flow rates, may be varied over wide ranges with little effect on modulator performance, it is sometimes necessary to tune or test the modulator.  For this purpose it is convenient to present a steady-state sample flux to the modulator, so that modulation peak profiles can be easily monitored as a function of experimental parameters.

A simple method for generating a continuous sample flux, which is suitable for qualitative studies of thermal modulation, is to insert a syringe loaded with a hydrocarbon such as dodecane into the GC injector operated in splitless mode, and leave the syringe in place.  After an initial surge, the sample flux into the analytical column will fall exponentially, and eventually "flatten out" into a long tail lasting several hours.  Continuous evaporation from the syringe into the injector apparently sustains this steady-state sample flux.  With the sample entering the modulator continuously, it is convenient to vary experimental parameters and measure the response of modulation pulse profiles.

A second method for generating a continuous sample flux, which is suitable for quantitative studies of thermal modulation, is to spike a carrier gas canister with a known concentration of a hydrocarbon vapor. This method provides a known sample flux to the modulator, and therefore permits quantitative studies of thermal modulation. It is convenient to swap the spiked cannister with the usual carrier gas supply by means of manually operated valves.

Velocity Measurement.  By pulsing the hot jet on for a period of time greater than the time of transit around the delay loop, analytes trapped in both cold spots may be remobilized and conveyed to the detector. This results in a so-called "double pulse" profile.  The time between pulse maxima is the transit time through the delay loop. If the length of the loop is known, then the ratio of loop length to transit time is the average velocity of the analyte through the loop.

Temperature Measurements.  To measure cold jet temperatures, a thermocouple was positioned in the positioned in the column holder in place of the modulator tube, and centered in the path of the cold jet by visual inspection.  A thermocouple reading was taken at various isothermal oven temperatures at cold jet flow rates known to produce proper thermal modulation.

Remobilization Studies.  The loop modulator was configured for single-stage thermal modulation for qualitative study of the remobilization process. Configuring the modulator for single stage modulation entailed retracting the upstream or downstream column loop into the column holder, so that only one cold spot formed on the modulator tube. The independent variables were isothermal oven temperature, accumulation time under conditions of constant sample flux, and the mode of modulator tube heating:  ambient oven or pulsed hot jet. The dependent variables were the rise-times of chemical pulses released from the cold spot, and the areas of modulation peaks.

The column set consisted of a single piece of uncoated fused silica capillary tubing.  Sample was presented continuously using the syringe method described in the tuning procedure above.  Rise times and peak areas of modulation peaks were measured using the Agilent ChemStation data analysis facilities. In each experiment,the cold jet flow was adjusted by means of a needle valve  to the minimum required to effect quantitative trapping of undecane.  Detailed experimental conditions were as follows:  GC:  Agilent 6890 with EPC. Sample:  Undecane, neat (Aldrich); m.p. -26 ^oC; b.p. 196 ^oC. Injection: Splitless, 2.5 psi head pressure, 250 ^oC. Continuous sample introduction by pulling 0.1 ྒྷl of undecane into Type B wire-in-plunger syringe, inserting needle through septum, and leaving syringe in place. Oven:  Isothermal experiments at the following temperatures: 40, 100, 160, 200 (^oC).  Modulator: A LN₂ Loop Modulator was operated with only one cold spot.   The modulator tube consisted of approximately 1.0 meter of uncoated deactivated fused silica capillary tubing, 0.1 mm i.d.  The hot jet heater was set at maximum temperature, which varied with oven temperature, because the heater block was un-insulated in this experiment.  Detector:  FID; H₂ flow 100 l/min; Air flow 800 l/min. Columns: No analytical columns were used. A bare modulator tube was connected from injector to detector via the thermal modulator.

The position of the column in the FID was optimized with respect to the width of modulation peaks. With sample entering the FID continuously, the ferrule nut on the FID was loosened enough to permit the column to be moved without disturbing the FID signal.  The modulator was run continuously at a repetition period of 3.0 to 5.0 seconds, and the column moved up and down in the FID until the sharpest peaks were observed in the Chemstation signal display window. The ferrule nut was then tightened.

The effect of ambient oven heating on modulation pulse profiles was investigated using the following procedure.  First, gas flows to both the hot and cold jets of the modulator system were turned off. Continuous presentation of the sample produced a constant offset signal. Manually turning the cold jet on for a period of time, then off again, caused sample to be accumulated during the cold jet on-time, then released by ambient oven heating.  Rise-times of chemical pulses released by ambient oven heating were measured as a function of the duration of the accumulation period, hence of the amount of material accumulated during each cycle of the cold jet, and as a function of ambient oven temperature. The accumulation periods (cold jet on-times) were 5, 10, 20, 40, and 80 seconds.  Following this series of tests, the hot jet was turned on and pulsed at regular intervals (3.0 to 5.0 seconds), thereby producing a normal modulation signal under the same conditions as the previous tests.  Chemical pulse rise times were again measured. These experiments permitted a comparison of the effects of ambient oven and pulsed hot jet heating modes upon modulation pulse rise-times.

Results & Discussion

Single Stage Thermal Modulation Experiments

Optimization.  Standard operating procedure calls for positioning the end of a capillary column about 2.0 mm from the stop position inside an FID.  In the present experiment, a sharp optimum was found in the rise-time of chemical pulses when the column end was positioned about 2.0 cm from the FID jet. The narrowest rise-times were observed when the gas flows to the FID were set at their maximum values.  Chemical pulses exhibited significant tailing in all of these experiments.  This is to be expected, because only a single cold spot, rather than two-stage thermal modulation was employed for these tests.  Tailing was not important in these experiments, because only rise-times were measured.

Temperatures.  Varying the static oven temperature from 40 ^oC to 220 ^oC caused the cold jet temperature to vary from -126 ^oC to -106 ^oC (Table 1).  The cold jet temperature is nearly independent of the oven temperature, even though it is exposed to the stirred oven bath.

• Table 1. Cold jet temperature and hot jet heating block temperature as functions of static oven temperature. All table entries are in units of ^oC.
• Oven           Hot Jet             Cold Jet 
• 40               207                  -126  
• 100              249                 -112
• 160              304                 -113
• 220              357                 -106

 Remobilization.  Single-stage thermal modulation of a continuous stream of undecane provided a number of qualitative observations consistent with the notion that the duration of modulation pulses is limited, at least in part, by the amount of time the trapped material spends in the liquid state during remobilization.  One would expect that while the material is in the liquid state, evaporation into the carrier gas will produce a vapor pulse, and that the concentration of this vapor pulse will increase as the temperature of the liquid rises in response to the hot jet. Upon complete evaporation of the liquid, one would expect a precipitous fall in the vapor concentration.  Signal profiles observed in the single-stage modulation experiment bear out these expectations.

 At ambient oven temperatures below the boiling point of undecane (195.9 ^oC), the areas of single-stage modulated peaks was directly proportional to the accumulation period under conditions of constant sample flow from the injector to the modulator (Figure 4a). The slope of this linear relationship, however, decreased monotonically with increasing temperature, indicating that the mass flow rate of sample delivered to the modulator was temperature dependent in spite of the fact that a constant mass flow rate was delivered to the head of the column by the injection technique employed in this experiment (see Experimental section).  This observation suggests that undecane vapor exiting the injector formed a condensed phase film on the interior of the uncoated tubing used in this experiment. Such film formation would be expected of course, whenever the ambient oven temperature is below the boiling point of undecane.  The thickness of this migrating film would depend upon ambient oven temperature however, due to gas-liquid partitioning.  The amount of material delivered to the single-stage modulator in a given time should be proportional to the thickness of the migrating film, which would be greater at lower temperatures than at higher temperatures. In that case, modulation peaks at lower temperatures would have greater areas than those at higher temperatures, which was the effect observed.

At isothermal oven temperatures near or above the STP boiling point of undecane, the dependence of peak area upon accumulation time exhibited curvature. The cause of this curvature was not investigated.

The areas of modulated peaks was observed to decrease monotonically with temperature when the accumulation period was fixed (Figure 4b). One might expect, under conditions of constant sample flux, that peak areas would be constant as a function of oven temperature, i.e., that the same amount of analyte should be presented to the modulator at all oven temperatures, if the sample flux is constant.  This expectation, however, assumes that the sample is fully vaporized in the capillary tubing leading from the injector to the modulator. If the analyte forms a condensed film on the interior wall of that tube, different considerations apply.  If a condensed film rather than a vapor migrates through the tube, and if the thickness of the condensed film is affected by temperature dependent gas-liquid partitioning, the modulation pulse areas would exhibit dependence upon oven temperature.  The fact that such dependence is observed, and the fact that the peak areas decrease with increasing oven temperature is again consistent with the presence of a migrating liquid film in the tubing upstream of the modulator, under the conditions of this experiment.

Two observations suggest that the amount of time the analyte spent in the liquid state during remobilization imposed a lower bound on the vapor phase pulse duration under the conditions of this experiment. (1) The rise times of modulated peaks were directly proportional to peak areas (Figure 5). Thus the time required to remobilize a peak was directly proportional to the amount of material accumulated in the modulator. This is consisent with the idea that the modulated chemical pulse is formed by evaporation of liquid material in the modulator tube during the firing of the hot jet. (2) Modulation pulse amplitudes were observed to increase monotonically with the temperature of the modulator tube, then to fall precipitously. This behavior could be explained by temperature driven vaporization of a liquid film, suddenly terminated by complete evaporation of the film.

If the amount of time the trapped analytes spend in the liquid state during remobilization limits the observed modulation pulse width, we would predict that rise times would depend significantly upon the heating rate during remobilization.  Rapid remobilization, as with a hot jet, should produce shorter rise times than the ambient oven heating.  This effect was observed (Figure 6).

Rise times obtained with hot-jet heating were significantly shorter than rise-times obtained with ambient oven heating.  Moreover, rise times obtained by ambient oven heating exhibited a strong and linear dependence upon ambient oven temperature, whereas rise times obtained with hot jet heating exhibited only a weak linear dependence. This means that the hot jet construction effectively decouples the operation of the thermal modulator from that of the ambient oven.

Although the present data do not conclusively prove that an evaporation mechanism limits the duration of modulation pulse widths in all circumstances, they suggest that such a mechanism should be considered, and that remobilization processes should be investigated more thoroughly.

Two Stage Thermal Modulation Experiments

Performance.  Loop modulation on uncoated fused silica tubing produced peaks some 36 milliseconds wide at base, the sharpest peaks so far observed using thermal modulation techniques (Figure 7). To observe such sharp peaks, it is necessary to connect the modulator directly to the detector via uncoated capillary tubing. A coated column will usually cause the peaks to broaden.  Marginally retained peaks exiting the second column of a GC x GC system, for example, range from 36 to about 60 ms in duration at baseline. Strongly retained peaks range in width from 60 ms to about 320 ms at baseline. 

With regard to tuning of the modulator, it is found that the modulation pulse shape does not depend strongly upon such parameters as flow through the column, or the duration of the hot jet pulse. The hot jet temperature is typically programmed at 100 ^oC above the oven temperature.  In general, the system is easy to operate, and tuning parameters are seldom changed once satisfactory operating conditions are established.

 The loop modulation system is found to modulate over the carbon range of C₃ to C₄₃ with no adjustment of the cold gas flow rate. However, Gaines and Frysinger have reported interesting results with a flow programmed cold jet [10].

Velocity Measurement.  Double pulse profiles obtained by turning on the hot jet for a period of time longer than the transit time through the delay loop are informative (Figure 8).  The first pulse of the double pulse profile represents the material accumulated at the second, or downstream, modulator stage. It is a sharp and generally symmetrical chemical pulse tens of milliseconds wide, which acts as a nearly ideal injection onto a downstream analytical column.  The second pulse of the double pulse profile has a more complicated structure, because it represents both the material accumulated at the head of the first, or upstream modulator stage, and unmodulated material freely transmitted through that first stage while the hot jet is pulsed on.  Under conditions in which sample is presented continuously to the modulator, this unmodulated material is apparent as a "shelf" or "tail" following the concentration pulse from first stage sample accumulation.  This "shelf" disappears rapidly once the hot jet is turned off, and the cold jet again induces accumulation in the modulator stage. The fall time of the "shelf", some 80 ms in the case of butane, illustrates rapid cooling of the capillary column by the cold jet.

The time between the concentrated pulse maxima in the double pulse profile is the transit time of analyte through the delay loop.  If the length of loop is known, then the average velocity of analyte through the delay loop is the ratio of loop length to transit time.  If the analyte is unretained, then this velocity is

that of the carrier gas. It is found that modulation performance optimizes at carrier gas velocities of about a meter per second in the delay loop. 

Single v.s.  Double Stage Modulation. Although single stage modulation is useful for characterizing the remobilization process, it offers no advantage to the GC x GC experiment, and does present disadvantages. Single stage thermal modulation does not simplify thermal modulation apparatus. Single- and two-stage configurations differ only by the position of a modulator tube segment, and an operator can switch between the two configurations in seconds, without disturbing the apparatus. However, single stage thermal modulation is in general considerably more difficult to tune than two-stage, owing to breakthrough effects, as illustrated in the double pulse profile above. If a second stage of thermal modulation is not used to focus (eliminate) the tail associated with a single stage of thermal modulation, then other techniques, such as carefully tuned modulator pulse durations, must be employed to remove it.  In general, these techniques are much more sensitive to experimental conditions than is the refocusing process employed in two-stage thermal modulation. As a result, single stage thermal modulation is difficult to tune in practice. Because single stage thermal modulation offers no simplification of apparatus, and introduces undesirable sensitivity to tuning parameters, two-stage thermal modulation is clearly a superior approach. 

Conclusion

Loop modulation is the simplest and highest performance thermal modulation yet developed for GC x GC applications. It produces the shortest modulation pulses observed with thermal modulators, is reliable and rugged, and is particularly easy to use.  Installation of the loop modulator on an Agilent 6890 GC takes approximately one hour.  The loop modulator permits direct measurement of analyte velocity within a capillary tube, a capability unique to this device, and very useful for chromatographic tuning.  The device supports detailed studies of the remobilization process, and also provides the widest range of modulation with respect to carbon number yet observed (C₃ to C₄₄). One disadvantage of the loop modulator is its reliance upon liquid nitrogen as a heat exchange medium. Alternatives to liquid nitrogen, including high pressure CO₂ gas or liquid, relieve cryogen consumption to some degree, but at the expense of modulation at lower carbon numbers. The next foreseeable advance in thermal modulation technique is the use of closed cycle cryo-refrigeration to eliminate liquid nitrogen consumption.  A second disadvantage of the thermal modulation systems so far designed is that miniaturizing such devices for use in portable analyzers would be technically challenging.

Literature References

• (a) John B. Phillips and Zaiyou Liu. "Chromatographic Technique and Apparatus", U.S. Patent No. 5,135,549; August 4, 1992. (b) John B. Phillips and Zaiyou Liu. "Apparatus and Method for Multi-dimensional Chemical Separation," U.S. Patent No. 5,196,039
• H.J. deGeuss, J. deBoer, U.A.T. Brinkman. "Development of a thermal desorption modulator for gas chromatography." J. Chromatography A. 767(1997) 137-151
• E.B. Ledford, Jr., and John B. Phillips. "Apparatus and Method for Chemical Modulation." U.S. Patent No. 6,007,602; December 28, 1999
• P.J. Marriott and R.M. Kinghorn. "Studies on cryogenic trapping of solutes during chromatographic elution in capillary gas chromatography." HRC J. High Res. Chrom.  19 (1996) 403-408
• E.B. Ledford, Jr. and C.A. Billesbach. "Jet-cooled thermal modulator for comprehensive multidimensional gas chromatography." HRC J. High Res. Chrom. 23 (2000) 205-207.
• J. Beens, M. Adachour, R.J.J.Vreuls, K. van Altena, and U.A. Brinkman. "Simple, non-moving modulation interface for comprehensive two-dimensional gas chromatography." Journal of Chromatography A. 919 (2001) 127-132.
• E.B. Ledford, Jr., Chris A. Billesbach, and Joel R. Termaat. "Transverse Thermal Modulation," U.S. Patent No. 6,547,852 B2; April 15, 2003
• E.B. Ledford, Jr.  "Method and apparatus for measuring velocity of chromatographic pulse"  PCT Patent Application No. PCT/US02/08488
• T. Hyotylainen, M. Kallio, and K. Hartonen. "Modulator design for two-dimensional comprehensive gas chromatography: Quantitative analysis of polyaromatic hydrocarbons and polychlorinated biphenyls." Anal. Chem.  74. (2002) 4441-4446.
• R.B. Gaines and G.S. Frysinger. "Temperature requirements for thermal modulation in comprehensive two-dimensional gas chromatography." Submitted to J. Sep. Sci.

©Zoex Corporation 2008