Prep LC: How Much Can I Load?

ExperimentalSolvents are all ACS grade, purchased from Midland Scientific. Compounds were ACS grade from Sigma Aldrich. The Prep HPLC system was a CombiFlash ACCQPrep (Teledyne ISCO, Lincoln, NE, USA) system equipped with an autosampler and run with a CombiFlash RediSep® Prep C18 column (PN 69-2203-810).

Other details are provided in each section of the results and discussion.

Jack E. Silver (Jack.Silver@Teledyne.com), Chester Bailey, Deen Johnson, Steve Paeschke, and Ronald LewisTeledyne ISCO, Lincoln, NE, USA

teledyneisco.com

Results and DiscussionDetermination of Sample Load from Initial Scouting GradientThe sample loading may be estimated via a calculation. An example spreadsheet is shown below which facilitates the calculations. “Peak 1” in the calculations below refers to the first eluting peak, while “Peak 2” refers to the second eluting peak. The numerical values are from the run in figure 5.

1. Optimize gradient prior to determining the load. The gradient affects the peak resolution which affects the loading.

2. Find the nearest impurity (eluting either prior to, or after, the desired compound) and note the starting and ending times for this peak.

3. Note the starting and ending times for the peak containing the desired compound. On CombiFlash systems, this is determined by simply touching the screen and noting the time at the touch point.

4. Subtract the start time from the end time to find the peak widths for each peak (Wp).

5. Subtract the time of the end of the first eluting peak from the end of the second eluting peak to get the maximum peak width (Wmax) which is the maximum peak width that “uses” all the available resolution. This corresponds to the maximum width for the second eluting peak of interest. As shown earlier, peaks tend to creep earlier so we define the maximum peak width from the second peak.

6. Calculate the allowable increase in the peak 2 width (Wmax–Wp2) and multiply this value by 15 (the 15% rule, labeled as “Rule” in the sample worksheet) to make a multiplier which is then multiplied by the original injection volume.

7. Finally, multiply this volume by (1–0.25) to create the maximum injection volume. The 0.25 value allows for column overloading and consequent “peak creep”. If the sample is dissolved in a strong solvent such as DMSO or DMF, the “Peak Creep Correction” will need to be increased to 0.5, or perhaps larger to accommodate the loss of loading capacity from these solvents.

Note that the loop size on the preparatory system will limit the volume that can be loaded. In general, manual injections should not exceed 1/2 the loop size to avoid sample loss.

References1 Dolan, J.W. How Much Can I Inject? Part 1: Injecting in Mobile Phase. LCGC North Am. 2014 32(10), 780–785.

ConclusionsA method of calculating the sample load based on an earlier injection is presented. It is based on the resolution between the compound of interest and the nearest impurity. The calculation is based on the increase in peak width as a function of loading, modified by empirical measurements and observations of different injections. The calculation returns a volume, assuming the same injection sample is used for the initial experiment and fully loaded column.

Injection Peak 1 Peak 1 Peak 2 Peak 2 Peak 1 width Peak 2 width Max Width Width Rule Ideal Load Ideal Injection Peak Creep Injection Volume  Increase Scale-up Volume Correction Volume

Start End Start End     Wmax            

  P1Start P1End P2Start P2End P1End–P1Start P2End–P2Start P2End–P1End WMax–WidthP2   Rule * Width Ideal Load Empirically Ideal Injection volume Increase Scale-up * determined *(1–Peak Injection Volume Creep Correction)

          (4.40–3.72) (5.81–5.10) (5.80–4.40) (1.40–0.70) 15 (0.70*15) (10.5*0.5) 0.25 (5.25*(1–0.25))

0.5 3.72 4.40 5.10 5.80 0.68 0.70 1.40 0.70   10.5 5.25   3.94

Worksheet 1

by Teledyne ISCO

Factors Affecting Actual Sample Load  In real life, one finds peak width increases due to column overloading and the effects of the solvent used to dissolve the sample. In Figure 2, the sample was dissolved in a solvent weaker than the mobile phase, allowing for concentration of the sample at the head of the column. The “shark fin” peak shape suggests column overloading. As the injection volume is increased, the front of the peaks appear earlier in the column elution, while the end of the peak is unaffected.

Figure 2–Front of peaks show reduced retention as loading increased from 2.5 to 4 mL. Note end of peaks shows the same time for both runs. The sample is catechol and resorcinol in a water/methanol gradient.

0 2 4 6 8 10 12 14 16

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

Time (min)

Abs

orba

nce

(230

nm

)

0

20

40

60

80

100

% B

(met

hano

l)

0 2 4 6 8 10 12 14 16

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

Time (min)

Abs

orba

nce

(230

nm

)

0

20

40

60

80

100

% B

(met

hano

l)

Figure 5–Methyl and ethyl paraben (20 mg/mL each) dissolved in 1:1 methanol/water. 0.5 mL injected. The times depicted are used in Worksheet 1.

0 2 4 6 8

0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

Time (Minutes)

254

nm (A

U)

0

20

40

60

80

100

% B

(Met

hano

l)

3.72 min4.40 min

5.10 min

5.80 min

AbstractMeasuring sample throughput and purity of the collected compound are two parameters considered when measuring High Performance Preparative Chromatography (HPLC) efficiency. Part of achieving high efficiency is to maximize the sample loading. This is achieved by loading the sample so that the peak containing the desired compound barely touches the nearest impurity peak. An algorithm can be used to calculate maximum peak loading on a preparative column from an HPLC test run. Using the peak widths and resolution from this injection, a maximum injection volume is calculated. This calculation saves time by reducing the number of test injections to optimize loading.

BackgroundTheoretical Estimation of Maximum Sample LoadIt is important to minimize solvent usage and purification time when running preparative chromatography. One way to minimize time is to optimize the gradient so that the desired compound elutes quickly while still eluting away from the impurities. A focused gradient allows maximum resolution while minimizing run time and solvent usage. A focused gradient also allows for some variation in solvent delivery between an analytical system and a preparative system.

Another way to reduce total purification time is to maximize the loading on a column thereby reducing the number of runs required to purify a given amount of sample. In preparative chromatography, efficiency is determined by maximizing the load with less concern for peak shape. Maximum efficiency occurs when the peak for the desired compound barely touches the peak for the nearest impurity. After the gradient is adjusted, one then maximizes the sample load on a column. The amount one can load on a column depends on the resolution (distance) between two peaks and the column efficiency¹. If the sample is injected in mobile phase under isocratic conditions , without overloading, resolution is determined by the following equation:

Rs=(t2- t1)/(0.5[ω1+ ω2])

Where t2 and t1 are the retention times of the second and first eluting peak and ω2 and ω1 are their peak widths. When injecting a larger sample, dissolved in the mobile phase, into the column at the same concentration as a smaller injection, the first part of the injection behaves the same as a smaller injection. For example, if doubling the injection volume, the first half of the injection will elute down the column the same as the original injection. The rest of the injection will increase the width of the peak(s) as shown in Figure 1.

Figure 1–Increasing injection volume under conditions where the column is not overloaded. Green and purple traces shows detector overloading.

0.0 0.5 1.0 1.5 2.00.0 0.5 1.0 1.5 2.0

0

5

10

15

20

Det

ecto

r Res

pons

e

Time (arbitrary units)

1 mL 2 mL 4 mL 6 mL 8 mL

Figure 4–Mixture of methyl and ethyl paraben (20 mg/mL each) dissolved in mobile phase (left) and dissolved in DMSO (right), 3.90 mL both injections.

How Does Injection Solvent Affect Sample Load?It is common to use solvents such as DMSO and DMF to dissolve samples. However, these solvents often reduce the sample loading because they “drag” the sample down the column, smearing peaks together (Figure 4). This reduces resolution and loading. The DMSO, DMF, or other “strong” dissolution solvent acts locally as a step gradient. Although the solvent is diluted by the mobile phase, it carries the sample down the column until diluted such that the solvent strength is weaker than the elution solvent. The use of strong injection solvents thus limits the maximum injection load.

0 2 4 6 8 10 12 14

0.0

0.5

1.0

1.5

2.0

2.5

3.0

Time (min)

Abs

orba

nce

(230

nm

)

0

10

20

30

40

50

60

70

80

90

100

% B

(met

hano

l)

0 2 4 6 8 10 12 14

0.0

0.5

1.0

1.5

2.0

2.5

3.0

Time (min)

Abs

orba

nce

(230

nm

)0

20

40

60

80

100

% B

(met

hano

l)

Figure 6–Mixture of parabens used in Figure 5. 3.90 mL injected as per the calculated injection volume from Worksheet 1.

0 2 4 6 8 10 12 14

0.0

0.5

1.0

1.5

2.0

2.5

3.0

Time (min)

Abs

orba

nce

(230

nm

)

0

20

40

60

80

100

% B

(met

hano

l)

Another deviation from the ideal is caused by gradients. Shallow gradients allow for error in determining the solvent system which elutes the compound. Gradients tend to focus the sample on the column. The tailing part of the peak is in a stronger solvent system then the leading portion of the peak. This tends to cause the compound contained in that portion of the peak to run faster through the column. The front of the peak is in a weaker solvent and runs slower compared to the tail, causing the peak to become focused. Since the molecules at the tail of the peak catch up to the front of the peak we again may reach the overloaded condition described above.