Fast(er) GC: How to Decrease Analysis Time Using Existing Instrumentation? Part V: Using Smaller Bore Capillary Columns

If the use of hydrogen is not an option for speeding up, and we cannot afford to loose any efficiency, the only option we have is to use a shorter column with a smaller diameter.

The efficiency is proportional with decreasing column diameter, meaning that a column with 2x smaller diameter can be 50% shorter and will deliver exactly the same number of theoretical plates.

Fig. 1 Influence of column diameter on optimal gas velocity and HETP: The smaller the diameter, the higher the optimum linear velocity

Most widely used are the 0.25mm and 0.32mm ID columns. In order to speed up the analysis, already 25 years ago, the 0.1 mm ID columns were developed and commercialized. Compared with a 25 m column, they showed comparable efficiency and analysis time could be up to 3x shorter.  This was also due to the higher optimum linear gas velocity for the smaller bore columns, see fig. 1. The generations of GC’s, like 6890) were all developed to accommodate the application 0.1mm columns.  Pressures are much higher, gas and oven controls must be more accurate and also the detector sample rate had to fast enough to measure the narrow peaks produced by the 0.10mm.

Practically the use of 0.10mm columns did not meet expectations for many, as these columns have limitations.

 

1. For compositional analysis where we can use high split-ratio’s, the columns work fine. For trace analysis, where we have to use splitless injection, the story is different. In Splitless injections, the liner volume must be transferred on to the column. Column flow in a typical 0.1 is very low, Helium flow at a velocity of 30 cm/s at the outlet is 0.3 mL/min.  As the gas is under a pressure of 217 kPa at the inlet, it is compressed a factor 3. That means that the volumetric flow at the inlet is only 0.1 mL/min.  To transfer the full liner volume in a splitless injection, will take considerable time. This adds to analysis time but also impacts injection volume. Therefore a “pressure pulse” has to be considered.
2. Sample capacity is very low. The average 0.1 mm column can be coated with max 0.2-0.4 micrometer film.  Injection of 5 ng will often already show signs of peak skewing.
3. Using 0.1mm columns, we have to work with relative high inlet pressures: The risk for septum leaks/discrimination will  increase, especially with huge pressure pulses;
4. Because columns are very short, for optimal results very fast temperature programs are required. Ovens do have limitations there as max programming rate is dictated by oven size and design.
5. Use of MS is not always possible. Ion traps need a certain minimal flow. Also the eluting peaks can be <0.5 seconds in width. We need enough data collection speed. Newer MS systems will meet this.
6. Because of small ID and thin film, the 0.1mm ID columns need more frequent maintenance as the column  inlet will contaminate faster. Guard columns play a bigger role.

If all conditions are considered properly, one can do fast GC using the 0.1 mm columns. Fig. 2 shows a semi-volatiles analysis in only 5.5  minutes using a 10m x 0.1 mm Rxi-5Sil MS column.

Fig. 2 Semivolatiles in les then 5.5 minutes using 10m x 0.1mm capillary with 0.1 um Rxi-5Sil MS

Many of the issues listed above could be overcome by using columns of 0.15mm ID. This diameter capillary seems to provide a very practical balance between all common column parameters.  The reduction in run time we can achieve using 0.15mm, is a factor 2.

Instead of a 30m x 0.25mm, we use a 20m x 0.15mm. The efficiency of a 20m x 0.15mm is about 10% higher then the 30m x 0.25mm.  By length only, we will be able to run 66% faster if we would use the same gas velocity. Because we have 10% higher efficiency we will operate the 0.15mm column at a 30% higher velocity (50 cm/s instead of 36 cm/s). By doing this, we will loose some efficiency, but that’s acceptable as we end up with similar efficiency as the 30m x 0.25mm, but with 2x shorter run time.

For this conversion, we ideally must use columns with the same phase ratio (beta). A 0.25μm film  in a 0.25mm ID column must be replaced by a 0.15μm film in a 0.15mm ID column.

Fig 3. Formula for calculation temp. program and iso-times to get the SAME elution temperatures. Valid for columns having the SAME phase-ratio (Beta)

As we have seen in the previous blogs, when we change column length and linear gas velocity, we need to set a different temperature program, to get similar peak elution order. (we need the same elution temperatures).  Fig 3 shows an easy calculation to do that. This is generic calculation as  compressibility of gases is not included. Dr. Leonid Blumberg has done a great job making software for such conversions. (available as free-ware from the web).

Fig 4 shows a complex perfume analysis, where we converted the analysis from a 30m x 0.25mm to  20m x 0.15mm column. Conditions are listed in fig.5.  We get similar peaks sequence, but 2x shorter run time.

Fig. 5 Conditions for 30m x 0.25mm and 20m x 0.15mm Rxi-5Sil MS columns. The 0.15mm column is operated above its optimum linear velocity

Fig.4 Analysis of Perfume” Eternity Moment” on 2 systems with comparable efficiency. The 20m x 0.15mm column is 2x faster

Interesting detail we also should mention is, that the peaks from the 0.15 mm column will be 2 x higher. We can use that for sensitivity, but better may be, to inject only 50% of the sample..  By doing that, we contaminate our system also 2x less meaning we can do twice the number of analysis before maintenance..

If the sampler cannot inject 0.5 μl (instead of 1 μl), you may consider to dilute the sample 1:1 and still get the benefit.

 

Related  blogs on fast(er) GC :

Part I   : Impact of column dimensions 

Part II : Impact of higher column flow 

Part III: Using faster temperature programming 

Part IV: Using hydrogen as the carrier gas