6.3 Fundamentals of Chromatography: Theoretical Plates, Resolution, and van Deemter Equation

6.3 Fundamentals of Chromatography: Theoretical Plates, Resolution, and van Deemter Equation

Quick Answer: Chromatography separates chemical mixtures by differential partitioning between a moving mobile phase and a stationary bed. Peak migration is governed by the retention factor $k' = (t_r - t_m)/t_m$ and column selectivity $\alpha = k'_2 / k'_1$. Band dispersion is quantified by column efficiency $N = 5.545(t_r / W_{1/2})^2$ and plate height $H = L/N$. Chromatographic resolution follows the Purnell master equation $R_s = \frac{\sqrt{N}}{4} \left( \frac{\alpha - 1}{\alpha} \right) \left( \frac{k'_2}{1 + k'_2} \right)$, where baseline separation demands $R_s \ge 1.5$. The van Deemter equation $H = A + B/u + Cu$ demonstrates that minimizing particle diameter ($d_p$) and choosing mobile phases with high diffusion rates flattens the mass transfer resistance ($Cu$) curve, enabling rapid separations.

---

Thermodynamic Basis of Chromatographic Retention

All chromatographic systems operate via continuous dynamic equilibrium as solute molecules transfer between a moving mobile phase and an immobilized stationary phase. For an analyte solute $A$:

$A_{\text{mobile}} \xrightleftharpoons[k_{\text{desorption}}]{k_{\text{adsorption}}} A_{\text{stationary}}$

The thermodynamic distribution constant (or partition coefficient, $K_D$) is defined as the ratio of equilibrium analytical molar concentrations:

$K_D = \frac{c_S}{c_M}$

where $c_S$ is the equilibrium concentration of solute in (or on) the stationary phase, and $c_M$ is its concentration in the mobile phase. Solutes with large $K_D$ values favor the stationary phase, spend a larger fraction of time retarded within the column bed, and elute at later retention times.

---

Fundamental Retention Parameters

Chromatographic data output consists of an elution profile (chromatogram) recording detector response as a function of time or mobile phase volume:

Detector Signal
      ^
      |             Peak 1 (Dead Time)            Peak 2 (Analyte)
      |                  t_m                           t_r
      |                 |---|                 |-----------------|
      |                 .   .                 .                 .
      |                 |   |                 .       /\        .
      |                 |   |                 .      /  \       .
      |                 |   |                 .     / |  \      .
      |                 |   |                 .    /  |h/2\  ---.--- Half-Height (W_1/2)
      |                 |   |                 .   /   |    \    .
   0 -+---------\\------+-+-+-----------------+--/----+-----\---.-----> Time
      |                                         |<-   W   ->|
      |                                           (Baseline)

Core Retention Equations

1. Retention Time ($t_r$): The total elapsed time between sample injection and the emergence of the analyte peak maximum at the detector.
2. Dead Time / Void Time ($t_m$ or $t_0$): The time required for an unretained solute (a molecule that does not interact with the stationary phase, $K_D = 0$, such as methane in GC or uracil in reversed-phase HPLC) to travel through the column interstitial void volume ($V_m$): $t_m = \frac{L}{u}$ where $L$ is column length and $u$ is average mobile phase linear velocity.
3. Adjusted Retention Time ($t'_r$): The net residence time spent by the analyte interacting with the stationary phase: $t'_r = t_r - t_m$
4. Retention Factor / Capacity Factor ($k'$): A dimensionless parameter describing the relative migration rate of a solute, independent of column geometry and mobile phase volumetric flow rate: $k' = \frac{t_r - t_m}{t_m} = \frac{t'_r}{t_m} = \frac{K_D V_S}{V_M} = \frac{n_S}{n_M}$ where $n_S$ and $n_M$ are the total moles of solute residing in the stationary and mobile phases, and $V_S$ and $V_M$ are the respective phase volumes.
   • Optimal Analytical Window: $1 \le k' \le 10$.
     • If $k' < 1$: Elution occurs too close to the solvent void peak; separation from unretained matrix interferences is poor.
     • If $k' > 10-20$: Excessive retention causes severe longitudinal diffusion, broad low-amplitude peaks, and unnecessarily long analysis times.
5. Selectivity Factor / Relative Retention ($\alpha$): A measure of the thermodynamic discrimination of the column chemistry between two adjacent solutes ($1$ and $2$): $\alpha = \frac{k'_2}{k'_1} = \frac{t'_{r2}}{t'_{r1}} = \frac{K_{D2}}{K_{D1}} > 1.0$ By IUPAC definition, solute $2$ elutes after solute $1$; therefore, $\alpha$ is always strictly greater than or equal to $1.0$. A value of $\alpha = 1.00$ signifies co-elution.

---

Column Efficiency: Plate Theory and Mathematical Models

Column efficiency measures the kinetic resistance of a column to zone broadening as a solute band traverses the bed. In Martin and Synge's plate theory, a chromatographic column is treated by analogy to a fractional distillation column containing a sequence of discrete theoretical equilibrium stages, termed theoretical plates ($N$).

Gaussian Solute Band: W = 4σ (Baseline Width); W_1/2 = 2.355σ (Half-Height Width)
                             /\        
                            /  \       
                           / |  \      
                          /  |h  \     
                         /   |    \    
                        /    |0.5h \   <--- W_1/2 = 2.355σ
                       /     |      \  
                      /      |       \ 
          -----------/-------+--------\----------- Baseline
                     |<-     4σ     ->|

Assuming a symmetrical Gaussian elution peak, the variance of the zone is $\sigma^2$ (in time units). The baseline width ($W$, determined by drawing tangents to the inflection points down to the baseline) spans $4\sigma$, while the width at half-maximum height ($W_{1/2}$) spans $2\sqrt{2\ln 2}\,\sigma \approx 2.3548\,\sigma$:

$N = \left( \frac{t_r}{\sigma} \right)^2 = 16 \left( \frac{t_r}{W} \right)^2 = 5.545 \left( \frac{t_r}{W_{1/2}} \right)^2$

Because baseline tangents are subjective and distorted by baseline drift, modern chromatographic software standardizes on the half-height formula ($5.545(t_r / W_{1/2})^2$).

Height Equivalent to a Theoretical Plate (Plate Height, $H$)

The physical length of column required to establish one theoretical equilibrium plate is the plate height ($H$):

$H = \frac{L}{N} = \frac{\sigma_{\text{length}}^2}{L}$

• Fundamental Maxim: A smaller plate height ($H$) corresponds to a more efficient column. High-performance capillary GC columns exhibit $H \approx 0.1-0.3\text{ mm}$ ($N > 100,000$), while conventional HPLC columns ($5\,\mu\text{m}$ particles) achieve $H \approx 10-25\,\mu\text{m}$ ($N \approx 10,000-25,000$).

Worked Example: A $15.0\text{ cm}$ ($150\text{ mm}$) reversed-phase $C_{18}$ HPLC column yields a retention time of $8.40\text{ min}$ for butylparaben. The peak baseline width ($W$) is $0.280\text{ min}$, and the width at half-height ($W_{1/2}$) is $0.165\text{ min}$. Calculate (a) the number of theoretical plates using both methods, and (b) the plate height in micrometers ($\mu\text{m}$).

• Part (a): Compute $N$ $N_{\text{baseline}} = 16 \left( \frac{8.40}{0.280} \right)^2 = 16 \times (30.0)^2 = 16 \times 900 = 14400\text{ plates}$ $N_{\text{half-height}} = 5.545 \left( \frac{8.40}{0.165} \right)^2 = 5.545 \times (50.909)^2 = 5.545 \times 2591.7 = 14371\text{ plates}$ (The two methods agree within $0.2\%$, confirming excellent Gaussian peak symmetry).
• Part (b): Compute Plate Height ($H$) $H = \frac{L}{N} = \frac{150\text{ mm}}{14400} = 0.01042\text{ mm} = 10.42\,\mu\text{m}$

---

Chromatographic Resolution ($R_s$) and the Purnell Equation

Chromatographic resolution ($R_s$) quantifies the quantitative separation of two adjacent peaks, balancing the physical distance between peak centers against their average peak widths:

$R_s = \frac{2(t_{r2} - t_{r1})}{W_1 + W_2} = \frac{t_{r2} - t_{r1}}{0.5(W_1 + W_2)} = \frac{1.177(t_{r2} - t_{r1})}{W_{1/2,1} + W_{1/2,2}}$

Quantitative Criteria for Peak Resolution

• $R_s < 1.00$: Incomplete separation. Peaks overlap significantly; quantitative integration errors exceed $5-10\%$.
• $R_s = 1.00$: Mutual peak overlap is approximately $4.0\%$. Acceptable for qualitative screening, but inadequate for trace purity certification.
• $R_s = 1.50$: Baseline separation (touching at $4\sigma$). Mutual area overlap is less than $0.3\%$. It is the minimum resolution most validated quantitative methods specify for critical peak pairs.
• $R_s \ge 2.00$: Complete, robust baseline separation. Recommended for automated industrial QA/QC assays.

   R_s = 0.75                  R_s = 1.00                  R_s = 1.50
(Severe Co-elution)          (~4% Peak Overlap)       (Baseline Separation)
      /\/\                       /\  /\                      /\      /\     
     /    \                     /  \/  \                    /  \    /  \    
    /      \                   /        \                  /    \  /    \   
   /        \                 /          \                /      \/      \  

The Fundamental Purnell Resolution Equation

By substituting expressions for efficiency, selectivity, and retention into the definition of $R_s$ (assuming narrow, closely spaced peaks where $W_1 \approx W_2$), we obtain the Purnell master equation:

$R_s = \frac{\sqrt{N}}{4} \left( \frac{\alpha - 1}{\alpha} \right) \left( \frac{k'_2}{1 + k'_2} \right)$

$\text{Resolution} = \left( \text{Efficiency Term} \right) \times \left( \text{Selectivity Term} \right) \times \left( \text{Retention Term} \right)$

Dissecting the Three Optimization Levers

1. The Efficiency Term ($\frac{\sqrt{N}}{4}$):

   • Because $R_s \propto \sqrt{N} \propto \sqrt{L}$, doubling resolution ($2 \times R_s$) requires a four-fold increase in plate count ($4 \times N$).
   • If $N$ is increased by lengthening the column ($4 \times L$), the analysis time quadruples ($4 \times t_r$) and column backpressure quadruples ($\Delta P \propto L$). Therefore, altering column length is the least energetically efficient method to resolve overlapping peaks.

2. The Selectivity Term ($\frac{\alpha - 1}{\alpha}$):

   • The most powerful chemical lever in chromatography. When $\alpha$ shifts from $1.05$ to $1.15$, the term $(\alpha - 1)/\alpha$ jumps from $0.0476$ to $0.1304$—a $174\%$ increase in resolution with virtually zero increase in retention time.
   • Selectivity is altered by changing chemical equilibria:
     • In HPLC: Changing stationary phase chemistry (e.g., $C_{18} \to$ Phenyl or Cyano), altering organic modifier ($MeOH \to ACN$ or $THF$), or adjusting mobile phase $pH$.
     • In GC: Changing stationary phase polarity (e.g., nonpolar DB-1 to polar DB-WAX) or altering oven temperature.

3. The Retention Term ($\frac{k'_2}{1 + k'_2}$):

   • Evaluates how long the compound spends in the stationary phase.
   • When $k'_2$ is very small ($k' < 1$), the term is steep: at $k' = 0.5$, $\frac{0.5}{1.5} = 0.33$; at $k' = 2$, $\frac{2}{3} = 0.67$ (a $100\%$ increase in resolution).
   • However, for $k'_2 > 5$, the function flattens asymptotically toward unity (at $k' = 5$, term $= 0.83$; at $k' = 10$, term $= 0.91$). Increasing $k'$ beyond $5-10$ yields diminishing returns in resolution while broadening peaks and wasting solvent.

Worked Example: Two pharmaceuticals elute with retention times $t_{r1} = 12.0\text{ min}$ and $t_{r2} = 12.8\text{ min}$ on a $25.0\text{ cm}$ column with an unretained void time $t_m = 1.20\text{ min}$. The baseline width of each peak is $W = 0.60\text{ min}$.

1. Calculate the current resolution ($R_s$).

2. Determine the column length required to achieve baseline resolution ($R_s = 1.50$) keeping flow rate and stationary phase identical.

• Part 1: Current Resolution $R_s = \frac{2(t_{r2} - t_{r1})}{W_1 + W_2} = \frac{2(12.8 - 12.0)}{0.60 + 0.60} = \frac{1.60}{1.20} = 1.33$
• Part 2: Required Column Length Since selectivity ($\alpha$) and retention factors ($k'$) remain constant, $R_s \propto \sqrt{N} \propto \sqrt{L}$: $\frac{R_{s,2}}{R_{s,1}} = \sqrt{\frac{L_2}{L_1}} \implies \frac{1.50}{1.333} = \sqrt{\frac{L_2}{25.0\text{ cm}}}$ $1.125 = \sqrt{\frac{L_2}{25.0}} \implies (1.125)^2 = \frac{L_2}{25.0} \implies 1.2656 = \frac{L_2}{25.0}$ $L_2 = 25.0\text{ cm} \times 1.2656 = 31.64\text{ cm}$

---

Rate Theory and the van Deemter Equation

While plate theory calculates the number of plates from an existing chromatogram, it fails to explain the physical kinetics that cause peaks to broaden. In 1956, Dutch chemical engineer Jan van Deemter formulated the van Deemter equation, which models band broadening kinetics as a function of mobile phase linear velocity ($u$):

$H = A + \frac{B}{u} + C \, u$

$\text{Plate Height } (H) = \text{Eddy Diffusion } (A) + \frac{\text{Longitudinal Diffusion } (B)}{u} + \text{Mass Transfer Resistance } (C) \cdot u$

Plate Height H
      ^
      |                   Total van Deemter Curve: H = A + B/u + C*u
      |                      \                         /
      |                       \       H_min           /
      |                        \________*____________/ <--- Mass Transfer (C*u)
      |                        /|                   /
      |                       / |                  / 
      |       B/u            /  |                 /  
      |    (Diffusion)      /   |                /   
      |                    /    |               /    
      |-------------------/-----+--------------/---------- Eddy Diffusion (A)
      |                         | 
   0 -+-------------------------+-------------------------> Linear Velocity u
                               u_opt

Dissection of the Three Physical Mechanisms

1. The $A$-Term: Eddy Diffusion (Multipath Effect):

   • Solute molecules traveling through a packed bed take tortuous, unequal flow channels of variable length and velocity around irregular packing particles. $A = 2 \, \lambda \, d_p$ where $d_p$ is average particle diameter and $\lambda$ is a packing uniformity factor (typically $\sim 0.8-1.0$ for well-packed beds).
   • The $A$-term is completely independent of mobile phase linear velocity ($u$).
   • Capillary Columns: In open tubular (capillary) GC columns, there is no packing material; therefore, $A = 0$, which explains the massive plate counts ($N > 100,000$) of capillary columns.
   • To minimize $A$ in packed HPLC columns: Use spherical, monodisperse particles with small diameters ($d_p < 2-3\,\mu\text{m}$) packed under ultra-high slurry pressures.

2. The $B$-Term: Longitudinal Diffusion:

   • Solutes concentrated at the center of a band naturally diffuse down their concentration gradient along the axial dimension of flow toward both the front and rear edges. $\frac{B}{u} = \frac{2 \, \gamma \, D_M}{u}$ where $D_M$ is the diffusion coefficient of the solute in the mobile phase, and $\gamma$ is an obstruction factor ($\sim 0.6-0.8$ for packed beds, $1.0$ for open capillaries).
   • Because diffusion is time-dependent, the $B/u$ term is inversely proportional to velocity ($u$). At extremely low flow rates ($u \to 0$), solutes spend extensive time inside the column, allowing axial diffusion to dominate and explode plate height ($H \to \infty$).
   • GC vs. HPLC Comparison: Solute diffusion in gases ($D_{M, \text{gas}} \approx 10^{-1}\text{ cm}^2/\text{s}$) is $10,000$ to $100,000$ times faster than in liquids ($D_{M, \text{liquid}} \approx 10^{-5}\text{ cm}^2/\text{s}$). Consequently, the $B$-term is vastly more significant in gas chromatography, requiring operating velocities ($20-50\text{ cm/s}$) substantially higher than HPLC ($0.5-2\text{ mm/s}$).

3. The $C$-Term: Resistance to Mass Transfer:

   • Chromatographic equilibrium is not instantaneous. As the mobile phase flows past, solute molecules must diffuse into and out of the stationary phase and across stagnant mobile phase pools: $C \, u = (C_S + C_M) \, u$ $C_S \propto \frac{d_f^2}{D_S}, \qquad C_M \propto \frac{d_p^2}{D_M}$ where $d_f$ is stationary phase film thickness, $D_S$ is solute diffusion in stationary phase, and $d_p$ is particle diameter.
   • Molecules that penetrate deeply into thick stationary films or large stagnant pores are "left behind" by the fast-moving mobile phase before they can diffuse back out, causing band tailing.
   • The $C$-term is directly proportional to velocity ($u$). At elevated flow rates ($u \gg u_{\text{opt}}$), mass transfer resistance dominates and causes plate height to rise linearly.
   • To minimize $C$:
     • In GC: Use ultra-thin stationary phase films ($d_f = 0.1-0.25\,\mu\text{m}$).
     • In HPLC: Use sub-2 $\mu\text{m}$ fully porous particles or core-shell (superficially porous) particles with short internal diffusion paths.

Optimum Linear Velocity ($u_{\text{opt}}$) and Minimum Plate Height ($H_{\min}$)

Setting the first derivative of the van Deemter equation with respect to velocity to zero ($\frac{dH}{du} = -\frac{B}{u^2} + C = 0$) identifies the mathematical optimum:

$u_{\text{opt}} = \sqrt{\frac{B}{C}}$ $H_{\min} = A + 2\sqrt{B \, C}$

At $u_{\text{opt}}$, the band broadening contributions from longitudinal diffusion ($B/u$) and mass transfer resistance ($Cu$) are exactly equal.

---

Practical Implications: Carrier Gases and UHPLC Particles

Plate Height H
      ^
      |   van Deemter Curves for Carrier Gases in Capillary GC (A = 0)
      |   
      |        \     /              N2: Lowest H_min, but very narrow u_opt range
      |         \   /                   and steep C-slope at high speed
      |          \ /                He: Safe, intermediate performance
      |           *                 H2: Highest u_opt, flattest C-slope (Ultra-fast GC)
      |            \_______
      |                    \________*________
      |                             |        \________________*_______________
      |                             |                         |               
   0 -+-----------------------------+-------------------------+--------------> Velocity u
                                u_opt(He)                  u_opt(H2)

Comparison of Carrier Gases in Gas Chromatography

Sub-2 $\mu\text{m}$ Particles in Ultra-High Performance Liquid Chromatography (UHPLC)

In conventional HPLC ($d_p = 5\,\mu\text{m}$), operating above $u_{\text{opt}}$ causes a steep rise in plate height due to $C_M \propto d_p^2 / D_M$.

By transitioning to sub-2 $\mu\text{m}$ particles ($d_p = 1.7-1.8\,\mu\text{m}$) in UHPLC:

1. The $A$-term ($2\lambda d_p$) is reduced by more than $60\%$.
2. The $C_M$-term ($d_p^2$) is reduced by a factor of $(5 / 1.7)^2 \approx 8.65$ (an 88% reduction).
3. The resulting van Deemter curve becomes nearly horizontal at high velocities, allowing chromatographers to increase flow rate by $3-5\times$ for high-throughput screening without degrading resolution.
4. Trade-off: Darcy's law dictates that backpressure scales inversely with the square of particle diameter ($\Delta P \propto 1 / d_p^2$); hence, UHPLC requires specialized instrumentation capable of operating at pressures up to $1000-1500\text{ bar}$ ($15,000-22,000\text{ psi}$).