Predicting Equilibrium Position from pKa

Every retrosynthetic decision about "which base do I use to deprotonate this substrate" is a live application of $\Delta\text{p}K_a$ reasoning: choosing LDA versus NaH versus $n$-BuLi versus a milder amine base for a given enolization, as touched on in Example 3, is exactly the calculation medicinal and process chemists run mentally (or on paper) before every alkylation, aldol, or Claisen condensation is planned. The same logic underlies buffer design in biochemistry and pharmaceutical formulation, where the Henderson–Hasselbalch equation — itself a direct rearrangement of the $K_a$ expression — sets the ratio of protonated to deprotonated drug at physiological pH, controlling solubility, membrane permeability, and bioavailability. In peptide and nucleic acid chemistry, the protonation states of histidine (imidazole side chain, p$K_a \approx 6$), lysine (p$K_a \approx 10.5$), and phosphate backbones are all predicted from exactly this framework, which is why enzyme active-site mechanisms are routinely drawn with explicit judgments about which residue is protonated at physiological pH 7.4 — a direct $\Delta\text{p}K_a$ comparison against water's effective p$K_a$ of 15.7 (or, more properly, against the local $K_w$-referenced pH). In materials and battery chemistry, proton-coupled electron transfer steps in fuel cells and metal-air batteries are analyzed with combined p$K_a$/redox-potential ("square scheme") diagrams that are a direct two-dimensional generalization of the single p$K_a$-subtraction rule developed here. And in everyday process chemistry, the decision to use aqueous sodium bicarbonate (carbonic acid p$K_a$ 6.35) versus sodium hydroxide (water p$K_a$ 15.7) to quench or wash a reaction mixture containing a carboxylic acid byproduct is, again, nothing more than picking the base whose conjugate acid p$K_a$ guarantees the $\Delta\text{p}K_a$ you need for clean, complete deprotonation without also stripping protons off more sensitive functional groups nearby.

Ask the tutor: If I only have aqueous p$K_a$ values but my reaction is run in DMSO, how do I adapt this comparison safely?

Every acid-base reaction you will ever draw in organic chemistry has the same skeleton: an acid HA hands off a proton to a base B, generating a new acid HB⁺ (or HB) and a new base A⁻ (or A). Mechanistically this is a single concerted step — a lone pair on B attacks the proton on HA while the H–A bond breaks heterolytically, both electrons going to A. There is no intermediate; it is one curved arrow into the proton, one curved arrow breaking the bond behind it.

The question "which side does this sit on?" is really the question "is HB⁺ a stronger or weaker acid than HA?" If HB⁺ is the weaker acid (higher p$K_a$), it holds onto its proton more tightly than HA did, so the system prefers to sit as B–H⁺ and A⁻ — equilibrium lies to the right. If HB⁺ is the stronger acid, it would rather give the proton back to A⁻, so equilibrium sits to the left, mostly as B and HA. This is fundamentally a statement about thermodynamic stability of conjugate bases. A⁻ is more stable than B (lower in energy, better able to bear negative charge) whenever HA is the stronger acid. The reaction moves toward whichever configuration puts the negative charge (or the proton) on the more accommodating partner — exactly analogous to the way a ball rolls toward the lower of two energy wells. The size of the p$K_a$ gap is a direct read of the height of that energy difference, since p$K_a$ is a logarithmic transform of a free energy: $\Delta G^\circ = 2.303\,RT\,\text{p}K_a$.

Notice the axis: this is not literally an energy diagram with a transition state drawn (proton transfers are fast, low-barrier processes), but a schematic of relative thermodynamic depth. The product well (acetate + ammonium) is deeper because ammonium (p$K_a$ 9.25) is a much weaker acid than acetic acid (p$K_a$ 4.76) — so the equilibrium sits overwhelmingly to the right, with $K_{eq} \approx 3.1 \times 10^4$. The mechanistic bookkeeping generalizes to any pair of acids, and the key trick for organic chemists is that you rarely need a table of cross reactions — you only need each species' p$K_a$ relative to the same reference reaction (dissociation in water, or DMSO, depending on the table), because subtracting two such values cancels the common reference and gives you the reaction you actually care about.

The important qualitative habit to build: scan a p$K_a$ table, find the acid on each side of the arrow, and ask which one is lower. The lower p$K_a$ species is the stronger acid, and the reaction runs so that the stronger acid ends up as its (weaker) conjugate base while the weaker acid stays protonated. No memorized "product favored/reactant favored" rule of thumb is needed once you can do this comparison in five seconds.

The acidity constant of a generic acid HA is defined (for a dilute aqueous or common reference solvent) as

where: $[\text{H}^+]$, $[\text{A}^-]$, $[\text{HA}]$ are equilibrium concentrations (activities, strictly) of the proton, conjugate base, and acid. For a proton-transfer reaction between acid HA and base B, forming A⁻ and the conjugate acid HB⁺,

the equilibrium constant is

where: the numerator holds the product concentrations, the denominator the reactant concentrations, exactly as any $K_{eq}$ is written from the balanced equation. The central relationship of this whole topic is

where: $K_a(\text{HA})$ is the acidity constant of the acid you start with, and $K_a(\text{HB}^+)$ is the acidity constant of the acid you end up with (the conjugate acid of the base B). Taking $-\log_{10}$ of both sides converts this into the form every chemist actually uses at the bench:

where: $\text{p}K_{eq} = -\log_{10}K_{eq}$, so that

Reading this equation: if the acid being formed (HB⁺) has a higher p$K_a$ than the acid you started with (HA), the exponent is positive, $K_{eq} > 1$, and equilibrium favors products. If HB⁺ has a lower p$K_a$ than HA, the exponent is negative, $K_{eq} < 1$, and equilibrium favors reactants. The rule of thumb "equilibrium favors the weaker acid and weaker base side" is just this equation in words.

Example 1 — Foundational: acetic acid + ammonia. Predict the position of, and $K_{eq}$ for,

Identify the acid on each side: on the left it is acetic acid, p$K_a = 4.76$; on the right, the acid formed is ammonium, p$K_a(\text{NH}_4^+) = 9.25$. Apply the formula, with "acid formed" = NH₄⁺ and "acid consumed" = CH₃COOH:

Direction check: ammonium (p$K_a$ 9.25) is a much weaker acid than acetic acid (p$K_a$ 4.76), so the reaction moves toward the side with ammonium and acetate — decisively to the right, as the large positive exponent confirms. As a limiting-case sanity check, if instead you asked about acetic acid reacting with its own conjugate base acting as a base towards another acetic acid molecule ($\Delta\text{p}K_a = 0$), you would correctly recover $K_{eq}=1$, i.e. no net preference — this is the identity-reaction check built into the formula.

Example 2 — Applied: can DBU deprotonate a malonate ester enough to run a Knoevenagel condensation? DBU (1,8-diazabicyclo[5.4.0]undec-7-ene) is a common non-nucleophilic amidine base with a conjugate-acid p$K_a$ (in water, protonated-DBU-H⁺) of about 12.0. Diethyl malonate's acidic C–H (flanked by two ester carbonyls) has p$K_a \approx 13.3$ in water (or ~16.4 in DMSO — pick one solvent scale consistently; we use the aqueous scale here to match DBU's aqueous p$K_a$). We want:

The mechanism: DBU's basic imine nitrogen lone pair attacks the malonate C–H proton; the C–H σ bond breaks heterolytically, both electrons flowing onto carbon, where they are immediately delocalized by resonance into both flanking ester carbonyls (the true reason malonate C–H protons are as acidic as p$K_a$ ~13, comparable to phenols, despite being on carbon). Identify acids: acid consumed = malonate ester (p$K_a$ 13.3), acid formed = DBU-H⁺ (p$K_a$ 12.0).

The equilibrium actually lies mostly on the starting-material side — only about 5% of the malonate is deprotonated at equilibrium, since DBU-H⁺ (p$K_a$ 12.0) is a slightly stronger acid than the malonate ester (p$K_a$ 13.3), meaning DBU is a slightly weaker base than the malonate anion. This is chemically realistic and important: DBU works in Knoevenagel-type condensations not because it fully deprotonates the substrate, but because even a small equilibrium concentration of enolate is enough to feed a fast, essentially irreversible subsequent step (aldol-type addition), continually pulling the pre-equilibrium forward by Le Chatelier's principle — a subtlety that a purely qualitative "DBU deprotonates malonates" statement would miss, but the quantitative $\Delta\text{p}K_a$ calculation reveals immediately.

Example 3 — Exam/competition level: choosing between LDA and NaOEt for a mono-deprotonation, with a numeric selectivity target. A chemist needs to generate the enolate of cyclohexanone (p$K_a \approx 26$ in DMSO, using a self-consistent DMSO p$K_a$ scale since we also need an alkoxide value on the same scale) essentially quantitatively (define "essentially quantitatively" as $K_{eq}\ge 10^{6}$, i.e., forward-favored by at least a million-fold) before adding an electrophile. Two candidate bases are proposed: sodium ethoxide (conjugate acid ethanol, p$K_a \approx 29$ in DMSO) and LDA (lithium diisopropylamide, conjugate acid diisopropylamine, p$K_a \approx 36$ in DMSO). Test NaOEt first:

acid consumed = cyclohexanone (p$K_a$ 26); acid formed = EtOH (p$K_a$ 29).

$K_{eq}=1000$ sounds large, but check against the stated target: it is well short of $10^6$. At equilibrium with excess ethoxide the ketone is mostly, but not overwhelmingly, deprotonated — with 1000:1 favoring enolate at exactly stoichiometric base, a small but real equilibrium population of un-deprotonated ketone remains, which matters for reactions demanding kinetic control of enolate geometry or clean mono-alkylation without competing aldol side reactions. Now test LDA:

acid consumed = cyclohexanone (p$K_a$ 26); acid formed = diisopropylamine (p$K_a$ 36).

$K_{eq}=10^{10}$ comfortably clears the $10^6$ bar — deprotonation with LDA is, for practical purposes, complete and irreversible, which is exactly why LDA (not ethoxide) is the standard reagent for generating a specific, kinetically controlled enolate cleanly at low temperature, with no equilibrium ketone left to scramble the enolate geometry through reversible reprotonation. The mechanistic reason LDA is so much more basic than ethoxide, despite both bearing a full negative charge, is that nitrogen is less electronegative than oxygen and stabilizes negative charge far less well — the amide lone pair is higher in energy (a poorer HOMO match for a weak acid, but exactly the deep, low-affinity-for-H⁺-relative-partner needed here) and correspondingly a stronger base, consistent with diisopropylamine's very high conjugate-acid p$K_a$ of 36.

The modern p$K_a$ framework rests on the acid-base theory formalized independently by Johannes Nicolaus Brønsted and Thomas Martin Lowry in 1923, which defined acids and bases purely in terms of proton donation and acceptance — a huge simplification over the older, more restrictive Arrhenius picture that required acids and bases to be defined only in aqueous solution in terms of H⁺ and OH⁻ ions specifically. Brønsted, a Danish physical chemist working in Copenhagen, was deeply interested in the quantitative relationship between acid strength and catalytic ability (what is now called the Brønsted catalysis law, relating rate constants of acid/base-catalyzed reactions to the p$K_a$ of the catalyst) — precisely the kind of thermodynamics-meets-kinetics thinking that underlies the modern practice of scanning a p$K_a$ table to choose a base for a synthetic step.