GC Bench

Choose a sample mixture from the four cards on the left. Each mixture is a blend of three or four components that are injected into the column together at time zero.
Choose a column. The more polar the column, the longer the bar under the column composition. The four stationary phases range from nonpolar to polar. A column whose polarity matches the sample's polarity retains those compounds more strongly and usually spreads them out more — try switching columns to see how retention times change.
Adjust the carrier gas flow rate. Flow rate sets the linear velocity of the carrier gas through the column. There is a sweet spot: too slow and the peaks broaden from longitudinal diffusion; too fast and they broaden from slow mass transfer between phases. Watch the plate count N in the readout as you sweep the slider to find the optimum.
Adjust the column temperature. A hotter column releases every compound faster (shorter run time) but compresses the differences between compounds, which usually lowers resolution. A cooler column separates more cleanly but takes longer. This is the classic speed-versus-resolution trade-off in GC method development.
Press Run to watch the detector trace draw left to right, like the pen on a strip-chart recorder. The chromatogram also redraws instantly any time you move a slider or change the mixture or column.
Read the peak table below the plot for each compound's retention time, area under the peak, capacity factor $k$, peak width, and its resolution to the next-eluting peak. A resolution of 1.5 or higher is generally considered fully (baseline) separated.
Goal: for each mixture, try to find a column, flow rate, and temperature that fully resolve every peak in the shortest possible run time. Then click Details to see the equations behind what you're adjusting, or Reset to create a new mixture composition.

This simulation combines four standard relationships from gas–liquid chromatography theory. The numbers are chosen for clear, physically reasonable teaching behavior rather than to match a specific real column and compound set.

Retention time

Each compound spends the dead time $t_M$ moving through the column with the carrier gas, plus additional time held up by the stationary phase, set by its capacity (retention) factor $k$:

$$ t_R = t_M\,(1+k) \qquad\qquad k = K\,\frac{V_S}{V_M} $$

$t_R$ retention time · $t_M$ dead (hold-up) time · $k$ capacity factor · $K$ partition coefficient · $V_S, V_M$ stationary- and mobile-phase volumes.

Effect of temperature

The partition coefficient follows a van't Hoff-type relationship, so $k$ falls roughly exponentially as the column heats up. Compounds with a larger effective enthalpy of retention $\Delta H^{\circ}$ (heavier, or a better polarity match to the stationary phase) are more sensitive to temperature — which is why raising the oven temperature compresses the peaks toward each other rather than shifting them all by the same amount:

$$ \ln k(T) \;=\; \frac{\Delta H^{\circ}}{R}\left(\frac{1}{T}-\frac{1}{T_{ref}}\right) + \ln k_{ref} $$

Column efficiency — the van Deemter equation

Band broadening is described by the height equivalent to a theoretical plate, $H$. It depends on the carrier-gas linear velocity $u$ through three competing mechanisms: eddy mixing ($A$), longitudinal molecular diffusion ($B/u$), and finite mass-transfer rate between phases ($Cu$):

$$ H = A + \frac{B}{u} + C u \qquad\qquad u_{opt} = \sqrt{\frac{B}{C}} \qquad\qquad N = \frac{L}{H} $$

The flow-rate slider sets $u$ from the volumetric flow $F$ and the column's internal diameter $d_c$:

$$ u = \frac{F}{\tfrac{\pi}{4}\,d_c^{2}} $$

Each peak is modeled as a Gaussian of width set by the local plate count, giving the peak's standard deviation and baseline width:

$$ \sigma = \frac{t_R}{\sqrt{N}} \qquad\qquad w_b = 4\sigma \qquad\qquad N = 16\left(\frac{t_R}{w_b}\right)^{2} $$

Resolution between neighboring peaks

Resolution combines efficiency ($N$), selectivity between the two compounds ($\alpha$, set by how differently the column retains them), and how strongly the later peak is retained ($k_2$):

$$ R_s = \frac{2\,(t_{R,2}-t_{R,1})}{w_{b,1}+w_{b,2}} \;=\; \frac{\sqrt{N}}{4}\left(\frac{\alpha - 1}{\alpha}\right)\!\left(\frac{k_2}{1+k_2}\right), \qquad \alpha = \frac{k_2}{k_1} $$

Rule of thumb used for the table's status color: $R_s \ge 1.5$ baseline resolved, $1.0 \le R_s < 1.5$ partially resolved, $R_s < 1.0$ not resolved.

How the controls map to the model

The flow-rate slider changes $u$, which moves you along the van Deemter curve for the selected column — changing $N$ and every peak width. The temperature slider changes every $k$ through the van't Hoff relation above. The column choice sets the van Deemter coefficients $A$, $B$, $C$, the internal diameter $d_c$, and how strongly the stationary phase's polarity interacts with each compound's polarity — which sets both the elution order and the selectivity $\alpha$ between compounds. The final trace sums each compound's Gaussian peak:

$$ c(t) \;=\; \sum_{i} \frac{A_i}{\sigma_i\sqrt{2\pi}}\exp\!\left[-\frac{(t-t_{R,i})^{2}}{2\sigma_i^{2}}\right] $$

Gas Chromatography

detector trace