This page provides an overview of current modulation techniques as used in multi-dimensional liquid chromatography. Generally, these can be split into two categories: passive modulation, where the first-dimension effluent is transferred unmodified in its entirety to the following separation column, and active modulation, in which parts of the first-dimension effluent are modified before injection into the subsequent separation dimension.

PASSIVE MODULATION

Passive modulation is by far the most common form of modulation in two-dimensional liquid chromatography. A 10-port or 8-port 2-position switching valve connects the separation dimensions using two or more sampling loops (see figure below). At any time the first-dimension effluent is sampled by one of the modulation loops (Figure 1: Left). Once the valve switches, the contents of that loop are injected to the second-dimension column, whereas the other loop is connected to the first-dimension separation (Figure 1: Right).

While this modulation strategy is simple and effective, its use imposes a number of constraints on the analytical method. First of all, the volume of the sampling loops cannot be too large in order to prevent overloading the second-dimension column, yet cannot be too small to allow sufficient quantities of the first-dimension effluent to be sampled.

The second constraint is related to this. As a result of the modulation interface, the first-dimension effluent is essentially the injection solvent of the second dimension. Consequently, incompatibility issues may – and often do – arise, regardless whether heart-cut or comprehensive mode is used.

Finally, for comprehensive mode (LC×LC) the static modulation loops impose time constrains to the analytical method. Comprehensive mode requires the entire first-dimension effluent to be subjected to a second-dimension separation. The modulation time, the time at which the valve switches, is therefore determined by the volume of the sampling loops and the first- and second-dimension flow rates. If the first-dimension flow rate is high, the volume of the sampling loops must be large enough to store the fraction until the second-dimension system is ready to receive a new fraction. Second-dimension separations in LC×LC are therefore generally very fast often using ultra-high pressure conditions.

Schematic of a standard passive modulation interface using a 2-position 8-port switching valve. First-dimension effluent is fractioned by a modulation loop and - upon switching of the valve - injection into a second-dimension separation. Reproduced from Pirok, B.W.J.; Stoll, D. R.; Schoenmakers, P.J. Anal. Chem. 2019, 91(1), 240-263. Copyright 2019 American Chemical Society.

STATIONARY-PHASE-ASSISTED MODULATION (SPAM)

Demonstrated first by Vonk et al. [Anal. Chem. 2015], stationary-phase-assisted modulation (SPAM) is currently the most popular active-modulation technique as alternative to passive modulation [Anal. Chem. 2019]. The SPAM interface utilizes trapping units rather than storage loops, the latter of which are used in passive modulation.

Schematic of the two positions of a stationary-phase-assisted modulation (SPAM) interface. Instead of the two sampling loops, as used in passive modulation, SPAM utilizes two trapping units which essentially filter out analytes from the first-dimension effluent. To facilitate retention of analytes on the trap, SPAM interfaces typically use a dilution flow to reduce the solvation strength of the first-dimension effluent relative to the trapping unit. Reproduced from Pirok, B.W.J.; Stoll, D. R.; Schoenmakers, P.J. Anal. Chem. 2019, 91(1), 240-263. Copyright 2019 American Chemical Society.

Upon elution from the first-dimension column, analytes are retained by the traps and essentially filtered out from the effluent. Since the trapping unit, typically a small (guard) column, is of low-volume, most of the first-dimension effluent not contained by the trap and diverted to waste. Once the valve is switched, the second-dimension eluent elutes the trapped analytes as concentrated bands and introduces them into the second-dimension column.

To facilitate sufficient retention on the trapping units, SPAM interfaces generally make use of a dilution flow which is combined with the first-dimension column effluent prior to introduction into the trapping unit. In some cases, mixers are used to homogenize the resulting flow stream. This is particularly important for separations of large molecules.

of stationary-assisted modulation

  • Reduced solvent incompatibility: With the trapping units essentially filtering out the analytes, the first-dimension mobile phase is mostly removed from the effluent. Relative to passive modulation, less incompatible first-dimension solvent is thus injected into the second dimension, leading to a major reduction in solvent-incompatibility effects.
  • Improvement of detection sensitivity: By removing the first-dimension mobile phase, the analytes are essentially concentrated on the trapping units. As a result, the dilution factor imposed by the two-dimensional LC system is significantly reduced and the detector at the end of the second-dimension separation “receives” the eluting analyte bands in higher concentrations relative to passive modulation.
  • Decrease in second-dimension injection volume: With much of the first-dimension effluent discarded, the total injection volume in the second dimension is significantly reduced. This allows shorter second-dimension columns to be used without a loss in efficient, thus significantly reducing the total analytes time.
  • Decoupling of the first and second-dimension analysis times: In comprehensive mode with passive modulation, one of the major limitations is the need to timely analyze the injected first-dimension fraction in a second-dimension separation before the other storage loop is full with first-dimension effluent. Using SPAM, storage loops can essentially be overfilled as long as the trapping unit can retain the analytes.

of stationary-phase-assisted modulation

  • Premature elution of analytes from the traps: A major disadvantage of SPAM is that all analytes must universally be retained to the trapping units. For a separation of, for example, hydrophobic and hydrophilic analytes, stationary-phase and mobile-phase conditions must be selected for the trapping units which allow retention of both the hydrophobic and hydrophilic analytes. This can be highly challenging for samples comprising a large number of sample dimensions. This disadvantage must be taken into account during method development. The effluent of the valve system can be routed to a (temporary) additional detector to detect analyte bands.
  • Robustness of the trapping units: Small (guard) columns are typically used as trapping unit. These fragile columns must be able to withstand the extreme conditions of the second-dimension separation (e.g. 1200 bar backpressure and/or pressure pulses). One particular problem is a difference in performance of two traps, potentially causing in differences in the chromatographic output between the alternating dimensions. This disadvantage in our opinion is the major bottleneck for routine implementation of SPAM in high-throughput 2D-LC methods.
  • Additional complexity to method development: Of course, the SPAM interface must be tailored to the analytical method and introduces yet even more method parameters to optimize during the 2D-LC method-development process. While less daunting for experienced users, the use of this active-modulation technique may be one step too far for newcomers to the field, especially because operation and optimization are sample dependent.

ACTIVE-SOLVENT MODULATION (ASM)

Introduced by Dwight Stoll an co-workers in 2017 [Anal. Chem. 2017], active-solvent modulation was specifically developed to resolve solvent-incompatibility issues (in particular breakthrough) for 2D-LC methods using reversed-phase LC as second-dimension separation. Like SPAM, ASM is an active-modulation technique in which the conditions of the first-dimension effluent fraction are modified prior to injection into the second-dimension column.

Reproduced from Pirok, B.W.J.; Stoll, D. R.; Schoenmakers, P.J. Anal. Chem. 2019, 91(1), 240-263. Copyright 2019 American Chemical Society.

As is shown in the schematic above, two parts are added relative to a conventional modulation valve. Positions A and C are similar to the states shown for passive modulation. Positions B and D, however, depict the a state where part of the second-dimension eluent coming from the pump is used to displace the loop volume (as is done in passive modulation), and part is used to directly admix this loop volume with second-dimension eluent. Consequently, the solvation strength of the overall flow stream is weakened, allowing the analytes to be properly be retained on the column. In their paper, Stoll et al. compare passive modulation with ASM and show how ASM allowed breakthrough effects to be mitigated.

of active-solvent modulation

  • Dilutes incompatible solvent: The ASM concept relies on diluting the first-dimension effluent with weaker second-dimension eluent. Consequently, any strong solvent from the first dimension is much less likely to induce breakthrough phenomena in the second-dimension separation.
  • On-column focusing in the second dimension: Rather than relying on a trapping unit inside the loop, ASM relies on the dilution to be that sufficient that analytes introduced into the second-dimension column will immediately be retained at the head of the column. The dilution factors of ASM and SPAM have not yet been quantitatively been compared, so it is difficult to make a comparison here.
  • Robust and relatively simple implementation: One thing that clearly separates ASM from SPAM is its simplicity. From a technological perspective, ASM does not rely selecting the appropriate trapping columns, nor is there a risk of premature elution from the trapping units – potentially risking the comprehensive mode – as is the case with SPAM. While the implementation of ASM definitely introduces complexity to the method-development process, it does much less so relative to SPAM.

of active-solvent modulation

  • Modulation volume still a limiting factor: Unlike SPAM, ASM still relies on the volume of the loops to be sufficiently large to store the first-dimension effluent fraction and, consequently, the first- and second-dimension analysis times are still coupled. However, taking into account the need to prevent undersampling the first-dimension separation [J. Sep. Sci. 2018], one can question whether this disadvantage is actually a significant disadvantage.
  • Mainly useful for reversed-phase in the second dimension: ASM was specifically developed for the use of RPLC in the second dimension. While fundamentally the concept of ASM should also work for other mechanisms in the second dimension, some of its advantages, such as on-column focusing, may be voided when moving away from RPLC. This makes sense as by far the majority of 2D-LC applications utilize RPLC as their second dimension [Anal. Chem. 2019].

VACUUM-EVAPORATION MODULATION (VEM)

A completely different approach from the modulation techniques discussed above is vacuum-evaporation modulation (VEM). Developed specifically for coupling normal-phase LC with reversed-phase LC, the concept of VEM relies on the evaporation of the typically very organic effluent from the first-dimension normal-phase separation. To facilitate timely evaporation of the solvent, the storage loops can be equipped with heating wires to heat up the solvent. In addition, a vacuum is placed at the end of the modulation loop to aid evaporation.

Unlike is the case for ASM and SPAM, fundamental studies have yet to be conducted on the principles and limitations of VEM. For example, the required volatility of the to-be-evaporated solvent with respect to operation parameters is currently unknown. Moreover, no data exists on the extent of accidental loss of volatile analytes during evaporation. Finally, the rate of dissolution for large analytes must still be investigated. Once the loop contents are coupled with the second dimension, not-evaporated analytes are supposedly precipitated on the loop wall. The physicochemical properties of the second-dimension eluent can be expected to influence the rate at which the analytes are dissolved.