Gas Chromatography Blog

Hot Split Injections, Part 2 – Sources of Discrimination

Our discussion continues about the process of hot split injection; the most popular form of sample introduction into capillary columns.

One of the reasons that a split inlet is so useful is because it is so flexible. It accommodates a wide range of sample concentrations, solvent types, and injection volumes. It also works with almost any size of capillary column. However there are ways that you can get into trouble if you are not careful.

   The following are potential downsides of any hot sample introduction technique:

  • Discrimination
  • Decomposition
  • Overload

  Let’s cover one of them in more detail: discrimination. Discrimination was briefly discussed in last month’s article but warrants further discussion here. Discrimination in hot inlets can come from several root causes:

  • Needle discrimination
  • Losses out unintended paths (e.g., the septum purge line)
  • Flow path inhomogeneity

GC Solutions #6: Hot Split Injections, Part 2 – Sources of Discrimination
Figure 1: Forms of discrimination. The positive and negative deviations from actual sample concentrations are the potential consequences of using hot injection techniques.

   Low end discrimination (unrepresentative sampling of the most volatile sample components) can manifest as either an enhancement or a loss of components. Loss of volatile components can happen if they diffuse to the top of the inlet and then are swept out the septum purge line or blow out through a faulty septum that does not reseal quickly when the syringe is removed. The smaller and more volatile the component, the faster it can diffuse and the greater the potential loss. Losses can also happen as a result of the momentary change in split ratio (see below).

   As a solvent evaporates, it might expand several hundred fold. So, 1 µL injected liquid might expand to several hundred µL or even to more than 1 mL of gas. If the solvent evaporates quickly, there is a momentary increase of pressure in the inlet. The magnitude of over-pressure is a function of the type and amount of liquid injected, inlet temperature, inlet pressure, liner configuration and the solvent used. It is also a function of the response characteristics of the inlet pneumatics control.

   Even though pressure deviations in hot split injection are much less exaggerated than those seen with hot splitless injection, pressure spikes do occur. Any increase in pressure causes a corresponding increase in column flow and a decrease in split ratio (split flow/column flow). If the increased pressure is not immediately corrected by the inlet’s pneumatic control, the most volatile components can be over-sampled because they diffuse faster to the bottom of the inlet (head of the column) than heavier components. This is observed as higher peak areas and heights than expected for volatile components relative to less volatile ones.

GC Solutions #6: Hot Split Injections, Part 2 – Sources of Discrimination-2
Figure 2: Glass wool temperature during injection of hexane. Inlet setpoint of 250 oC. Glass wool positioned near the middle of the inlet liner cools to the boiling point of hexane (69 oC) until it has evaporated and then returns to setpoint temperature.   

 If the inlet pneumatics control responds very quickly to the increasing pressure, it does so by opening the split vent line to vent more flow. This higher flow increases the split ratio. As the most volatile components diffuse quickly to the bottom of the liner they would then be under-sampled the higher split ratio.

   There are two additional factors that affect the magnitude of the pressure surge from evaporating solvent: (1) liner dimensions and (2) glass wool. For split injections, the best liners are ones specifically designed for split injections. They are a little bit loose in the inlet (as opposed to splitless liners that are tight due to larger outer diameter; o.d.). This allows high split flows and expanding sample vapours to move more easily between the outside of the liner and the inlet wall as they travel to the split vent line, reducing the potential pressure surge.

GC Solutions #6: Hot Split Injections, Part 2 – Sources of Discrimination-3
Figure 3: Needle discrimination in a hot inlet as a function of injection speed, all other variables remaining constant. Some automatic liquid samplers are capable of completing the injection process in < 300 msec.

   Hot split liners typically have “large” internal volumes to accommodate the expanding volume of sample as it evaporates. Most hot split inlets have internal volumes approaching 1 mL (depending on the type of liner installed). The larger the internal volume of a liner, the less dramatic is the pressure spike caused by evaporating solvent.

   Even though a large liner volume is good for minimizing pressure transients, this comes with a potential downside; higher residence time of sample vapours in the inlet. This, fortunately, is not usually a problem in hot split injections. Let’s say the internal volume of an inlet is 1 mL with a given liner. Let’s also say that the total flow of carrier gas down the inlet is 100 mL/min. For an injected sample that generates 1 mL of expanded gas, the vapours would clear the inlet in approximately 1/100 or 0.01 min (0.6 sec). This provides little time for volatile sample components to back-diffuse to the top of the inlet where they would be lost out the septum purge.

   Glass wool has a dramatic affect on evaporation process, as illustrated in Figure 2. Glass wool has high surface area and a low mass. When sample is injected, it wicks into the glass wool, spreading out in the fiber matte. As solvent evaporates, the glass wool temperature quickly drops from the setpoint temperature to the boiling point of the solvent [1]. This not only slows the evaporation process, it reduces the magnitude of any potential pressure surge. As the solvent is exhausted, the glass wool temperature heats back up to the setpoint temperature.

GC Solutions #6: Hot Split Injections, Part 2 – Sources of Discrimination-4
Figure 4: Heavier sample components tend to linger near the liner wall while volatile components are evenly distributed in the gas. A tapered liner directs all components to the head of the column, minimizing discrimination.

   Even though all the discussed sources of low-end discrimination may be present, causing either exaggerated or diminished proportion of volatile components relative to that in the original sample; the probability and magnitude of low end discrimination in hot split injection actually is fairly low in practice.

   The more prevalent form of discrimination seen in split injection is loss of high end components. [I have never seen positive high-end discrimination; enhancement of high boilers] High end discrimination has three primary causes: (1) low inlet temperature, (2) needle discrimination, and (3) inhomogeneous flow effects.

   Needle discrimination comes from selective distillation and surface effects from within the metal syringe needle in the hot inlet. From the moment the needle starts to enter the hot inlet, solvent and volatile sample components start to evaporate from it. The high-boiling components concentrate in the remaining liquid. The act of forcing sample out of the syringe needle by depressing the plunger does not displace the more concentrated liquid nor does it stop the selective distillation process. In fact, that process accelerates as the needle continues to heat up. The more concentrated portion, rich in higher molecular weight components, tends to lag along the needle wall as less-viscous sample squirts through the centre of the needle.

   The slower the injection process (including inserting the syringe into and removing it from the inlet), the more pronounced is needle discrimination. Also, the higher the inlet temperature and lower the boiling point of the solvent, the more pronounced is needle discrimination.
Manual injection takes approximately 2 sec to complete. As illustrated in Figure 3, modern autosamplers with very fast injection (much faster than manual approaches can be) greatly diminish needle discrimination[2]. Some automatic liquid samplers can complete the injection process in < 300 msec.

   Discrimination resulting from flow inhomogeneity is not well documented in the literature. This is probably because there are so many variables affecting it, including:

  • Inlet temperature
  • Volatility range of the sample
  • Flow rate down the liner
  • Column flow rate
  • Column position
  • Liner design
  • Presence or lack of glass wool
  • Liner surface chemistry

   Inlet temperature and inhomogeneous flow effects conspire together to discriminate against low-volatility sample components. Higher molecular weight compounds tend to spend a disproportionate amount of time near the walls as they make their way to the bottom of the liner. As illustrated in Figure 4, the vapour cloud of evaporated sample moving toward the bottom is not axially homogeneous (there is a molecular weight dependent gradient across the liner). This inhomogeneity is quite temperature dependent. With a column positioned in the middle of the bottom of the liner, a smaller proportion of high-boiling components would enter the column than is present in the original sample. The higher molecular weight material staying near the wall all the way to the bottom of the liner would disproportionately be disproportionately swept out split vent. This effect is greatly reduced by using a liner with a taper at the bottom (Figure 4). The taper directs flow to the head of the column to minimize discrimination resulting from axial inhomogeneity. The closer the diameter of the taper is to the o.d. of the column, the more effective it is in correcting the problem.

   Although discrimination can be observed with hot split injection, proper choice of liner, temperature, sample solvent and use of glass wool can combine to reduce or eliminate it in practice. As will be seen in future articles, discrimination in hot splitless injection is much worse and is much better with cool injection techniques.

[1] Biedermann, M., Fiscalini, A. and Grob, K., Journal of Separation Science, 27: 1157–1165. (2004) doi: 10.1002/jssc.200401847
[2] US Patent 4,615,226, Francis M. DiNuzzo, James S. Fullemann; Hewlett-Packard Company, 1986

This blog article series is produced in collaboration with Dr Matthew S. Klee, internationally recognized for contributions to the theory and practice of gas chromatography. His experience in chemical, pharmaceutical and instrument companies spans over 30 years. During this time, Dr Klee’s work has focused on elucidation and practical demonstration of the many processes involved with GC analysis, with the ultimate goal of improving the ease of use of GC systems, ruggedness of methods and overall quality of results. If you have any questions about this article send them to

Improve your Knowledge & Skills with our latest 'live' and 'on-demand' educational webinars Free registration with open access >>

Subscribe to FREE learning from the experts!