Using pKa Tables Quantitatively

A single number on a page — pKa 4.76 for acetic acid, pKa 25 for acetone — encodes an entire equilibrium's worth of thermodynamics. A chemist who can read that number fluently can predict outcomes without running a single reaction. They can tell whether a base will deprotonate a substrate. They can tell which of two acidic sites will ionize first. They can also predict how large an equilibrium constant to expect for a proton transfer. This is not a memorization exercise. It is a quantitative skill: converting tabulated pKa values into free energies, into equilibrium constants, into percent conversions, and back again. At the same time, one must stay alert to the traps that make naive table-reading fail. These include solvent mismatches, degenerate protons, the wrong conjugate pair, and tables built on different reference states. The pKa table is the acid–base chemist's slide rule. Reaction mechanisms are decided by electron pushing. But which mechanism is thermodynamically downhill is decided by comparing pKa values. Comparing them correctly requires understanding exactly what the number means. It is a free energy of a specific proton-transfer half-reaction. It is referenced to water (or DMSO, or the gas phase) at 25 °C. And it applies to a specifically drawn conjugate acid/conjugate base pair. Get the pairing wrong, and the whole calculation collapses. This happens, for example, when a student subtracts the pKa of ethanol from the pKa of the ethoxide's conjugate acid, rather than ethanol itself. The error occurs even though every arithmetic step looks correct. This article is about the arithmetic and the bookkeeping. Together, they turn a printed pKa table into a working predictive tool. We build on the equilibrium-position logic of Predicting Equilibrium Position from pKa. We also build on the substituent-ranking logic of Combining Effects to Rank Acidity. Now we treat the numbers themselves as data to be manipulated. We show how to compute $K_{eq}$ for any acid–base reaction from a $\Delta$ pKa. We show how to handle polyprotic and multi-site molecules. We show how to convert between aqueous and DMSO scales. And we show how to avoid the several specific errors that pKa tables are engineered to catch you in. By the end, you should be able to look at a two-column pKa table and a proposed reaction. Within thirty seconds, you should be able to produce a number. That might be $K_{eq} = 10^{4.3}$ , or "83% deprotonated," or "this base is 400 times too weak." You should have full confidence in every digit. You should also have full awareness of when the number cannot be trusted at all.

Key takeaways

pKa is a free energy in disguise: $\Delta G^\circ = 2.303\,RT \times \text{pKa}$ , so table values convert directly to thermodynamics.
For any proton transfer, $\log K_{eq} = \text{pKa(acid you start with)} - \text{pKa(acid you end with)}$ — get the pairing right or everything is wrong.
A $\Delta$ pKa of 1 unit is a factor of 10 in $K_{eq}$ ; a $\Delta$ pKa of 3 is already a 1000-fold bias, functionally "complete."
Percent conversion follows a logistic (sigmoidal) curve in $\Delta$ pKa, not a linear one — memorize the shape, not just the endpoints.
Tables built in water, DMSO, or the gas phase are not interchangeable; DMSO pKa's are often 10–20 units higher for the same compound.
Polyprotic acids need per-step pKa's (pKa₁, pKa₂, …), and statistical/symmetry corrections matter for molecules with equivalent protons.
Key idea: a pKa table is a portable free-energy ledger — every comparison you make is really a subtraction of two $\Delta G^\circ$ values, and the antilog of that subtraction is your equilibrium constant.

Where this lives today

Quantitative pKa comparison is not a classroom abstraction; it is exactly how synthetic chemists select reagents in practice. Choosing between NaH, LDA, KHMDS, n-BuLi, and DBU for a given deprotonation is a direct application of the ΔpKa machinery developed here, and getting it wrong wastes reagent, generates side products, or — with strong, non-selective bases like n-BuLi — can deprotonate or add to sensitive functional groups elsewhere in a complex molecule. Total synthesis routes routinely list "pKa considerations" explicitly when justifying a base choice for a late-stage, precious intermediate. In medicinal chemistry, the ionization state of a drug at physiological pH (7.4) — governed by exactly the Henderson–Hasselbalch logic used above — determines membrane permeability, plasma protein binding, and renal clearance; pKa prediction and measurement is a standard early-stage ADME (absorption, distribution, metabolism, excretion) assay, and quantitative pKa tables for amine and carboxylic acid pharmacophores are consulted constantly in drug design. In biochemistry, enzyme active sites tune the local pKa of catalytic residues (aspartate, histidine, cysteine, tyrosine) by many units relative to their free-amino-acid values, through electrostatic and hydrogen-bonding microenvironments — a phenomenon only interpretable quantitatively because the baseline aqueous pKa values are known precisely and the shift can be expressed as a ΔG° via the same 1.36 kcal/mol-per-unit conversion used throughout this article. In materials and polymer chemistry, the protonation state of ionizable monomers (acrylic acid, vinylpyridine) as a function of pH — again a direct pKa table application — governs pH-responsive hydrogel swelling and controlled-release behavior, engineered explicitly using tabulated and measured pKa values.

Ask the tutor: If I have three candidate bases with conjugate-acid pKa's of 11, 25, and 38, and my substrate has pKa 20, how do I decide which single base is the best practical choice rather than just the most thermodynamically favorable?

The physical picture — reading a pKa table as a free-energy ledger

Every pKa value in a table is shorthand for a specific equilibrium constant of a specific reaction: the acid HA donating a proton to the reference solvent (water, by far the most common convention).

\text{HA} + \text{H}_2\text{O} \rightleftharpoons \text{H}_3\text{O}^+ + \text{A}^-, \qquad K_a = \frac{[\text{H}_3\text{O}^+][\text{A}^-]}{[\text{HA}]}

and pKa $= -\log_{10} K_a$ . Because $\Delta G^\circ = -RT\ln K_a$ , and $\ln x = 2.303 \log_{10} x$ , the pKa converts directly to a standard free energy of deprotonation:

\Delta G^\circ_{\text{deprotonation}} = 2.303\, RT \times \text{pKa}

At 25 °C (298 K), $2.303RT = 1.36$ kcal/mol, so each pKa unit is worth 1.36 kcal/mol of free energy. This single conversion factor is the entire "physics" underlying every quantitative pKa comparison: a $\Delta$ pKa of 10 between two acids is a 13.6 kcal/mol difference in how strongly they resist giving up a proton — an enormous, essentially insurmountable gap at room temperature. The reason pKa tables are useful rather than just decorative is that free energies are additive across a thermodynamic cycle. If acid HA₁ (pKa₁) transfers its proton to the conjugate base of HA₂ (pKa₂), the reaction free energy is simply the difference of the two half-reaction free energies, because the H⁺(aq) that appears in both half-reactions cancels:

Figure

Subtracting two deprotonation half-reactions cancels the shared proton, leaving the net proton-transfer equilibrium

This cancellation is the whole trick: the solvent-referenced proton "disappears" from the net reaction, and only the difference in pKa survives. That is why pKa tables — despite each entry being measured against water — can be used to predict reactions that never touch water at all, such as an alkoxide deprotonating an alkyne in THF. The table entries are really just relative free-energy markers on a single universal ladder. A second physical picture worth holding in your head is the percent-ionized curve. For a monoprotic acid at a given pH, the fraction ionized follows the Henderson–Hasselbalch relation, and by exact analogy, for a proton transfer between two acids the fraction of product (deprotonated acid 2) at equilibrium follows a logistic curve in $\Delta$ pKa:

f_{\text{product}} = \frac{10^{\Delta\text{pKa}}}{1 + 10^{\Delta\text{pKa}}}, \qquad \Delta\text{pKa} = \text{pKa(reactant acid)} - \text{pKa(product acid)}

Figure

Percent conversion is a logistic S-curve in ΔpKa, not linear — flat near 0% and 100%, steep through the middle

Two consequences follow immediately, and both are load-bearing for problem solving. First, you do not need a huge pKa gap for a reaction to be synthetically "complete." By ΔpKa = +3, the product ratio is already 1000:1 (99.9%), and by +4 it is 10⁴:1. Second, the region near ΔpKa = 0 is the dangerous zone: a difference of even 1–2 units still leaves a substantial equilibrium population of both species, which matters enormously if a side reaction can intercept the minor component.

The core ideas — turning pKa differences into equilibrium constants

The master equation of quantitative pKa table use, derived above from the free-energy additivity of the two half-reactions, is:

\log_{10} K_{eq} = \text{pKa}_{\text{HA (starting acid)}} - \text{pKa}_{\text{HA (conjugate acid of the base you used)}}

where:

pKa(starting acid) is the pKa of the H–A bond you are breaking (the substrate),
pKa(conjugate acid of the base) is the pKa of the acid formed when the base picks up that proton (not the pKa of the base itself — bases don't have pKa's, their conjugate acids do),
K_eq is defined for the reaction written as substrate-acid + base → substrate-conjugate-base + base's-conjugate-acid.

A companion relationship converts any $K_{eq}$ (or any single pKa) into a free energy and, if you want it, into a rate-independent statement about spontaneity:

\Delta G^\circ = -2.303\,RT \log_{10} K_{eq} = 2.303\,RT\,(\text{pKa}_{\text{product acid}} - \text{pKa}_{\text{reactant acid}})

where:

R = 1.987 cal/(mol·K), T = 298 K, giving 2.303RT = 1.364 kcal/mol,
a negative ΔG° (favorable) corresponds to the reactant acid having the lower pKa (it is the stronger acid, and giving up its proton to form the weaker, more stable conjugate base — the product acid — is downhill).

And for polyprotic or multi-site systems, the fraction of each protonation state at a given pH follows directly from stacking Henderson–Hasselbalch equations, one per ionizable site, each referencing its own microscopic or macroscopic pKa.

Derivation AFrom ΔG° additivity to the equilibrium-constant shortcut

Goal: show rigorously why $\log K_{eq} = \Delta\text{pKa}$ for a proton transfer, and fix sign conventions so they never trip you up.

Step 1 — Write both half-reactions with water as the common reference.

\text{HA} + \text{H}_2\text{O} \rightleftharpoons \text{H}_3\text{O}^+ + \text{A}^-, \qquad K_1 = 10^{-\text{pKa}_1}

\text{HB} + \text{H}_2\text{O} \rightleftharpoons \text{H}_3\text{O}^+ + \text{B}^-, \qquad K_2 = 10^{-\text{pKa}_2}

Here HA is the substrate whose proton you want to remove, and B⁻ is the base you propose to use (HB is its conjugate acid).

Step 2 — Combine to eliminate water and hydronium. Take reaction 1 forward and reaction 2 in reverse, then add:

\text{HA} + \text{B}^- \rightleftharpoons \text{A}^- + \text{HB}

This is exactly the proton-transfer reaction you care about — base B⁻ removes the proton from HA. Because H₂O and H₃O⁺ appear once on each side across the two steps, they cancel completely; no net solvent or hydronium ever appears in the target equation.

Step 3 — Combine the equilibrium constants the same way. Adding reaction 1 to the reverse of reaction 2 means multiplying $K_1$ by $1/K_2$ :

K_{eq} = \frac{K_1}{K_2} = \frac{10^{-\text{pKa}_1}}{10^{-\text{pKa}_2}} = 10^{\text{pKa}_2 - \text{pKa}_1}

Careful with the sign here — HA is the acid being deprotonated (pKa₁), and HB is the acid formed by protonating the base (pKa₂). Taking log base 10:

\log K_{eq} = \text{pKa}_2 - \text{pKa}_1 = \text{pKa(conjugate acid of base)} - \text{pKa(substrate acid)}

Equivalently, written the way most people find easier to remember, $K_{eq}$ favors formation of the weaker acid:

\log K_{eq} = \text{pKa(product, i.e., HB)} - \text{pKa(reactant, i.e., HA)}

Step 4 — Sanity-check the sign with a trivial case. If pKa(HA) = pKa(HB) exactly (substrate and base's conjugate acid equally strong), then $\log K_{eq}=0$ , i.e., $K_{eq}=1$ — the reaction sits at 50:50, exactly as intuition demands for two acids of identical strength competing for a single base.

Step 5 — Extract the free energy. Since $\Delta G^\circ = -RT\ln K_{eq} = -2.303RT\log K_{eq}$ , substituting Step 3:

\Delta G^\circ = -2.303RT\big(\text{pKa}_2 - \text{pKa}_1\big) = 2.303RT\big(\text{pKa}_1 - \text{pKa}_2\big)

A reaction is thermodynamically favorable (ΔG° < 0) exactly when pKa₁ (substrate) is smaller than pKa₂ (conjugate acid of the base) — i.e., you are always deprotonating the stronger acid present and generating the weaker one, which is simply the statement that proton transfer always runs "downhill" from strong acid to weak acid at equilibrium.

Go deeper · Derivation B · 2 minPolyprotic acids: stacking pKa steps and the statistical correction

Derivation BPolyprotic acids: stacking pKa steps and the statistical correction

Goal: extend the single-pKa machinery to molecules with more than one ionizable proton, and show why "symmetric" molecules need a numerical correction that is easy to derive from first principles.

Step 1 — Define stepwise macroscopic pKa's. For a diprotic acid H₂A, two independent equilibria are tabulated:

\text{H}_2\text{A} \rightleftharpoons \text{H}^+ + \text{HA}^-, \qquad K_{a1}=10^{-\text{pKa}_1}

\text{HA}^- \rightleftharpoons \text{H}^+ + \text{A}^{2-}, \qquad K_{a2}=10^{-\text{pKa}_2}

These are the numbers printed in tables (e.g., oxalic acid pKa₁ = 1.25, pKa₂ = 4.14; phosphoric acid pKa₁ = 2.15, pKa₂ = 7.20, pKa₃ = 12.35). Each step is always harder than the last (pKa₂ > pKa₁) because removing a second proton from an already-anionic species costs extra electrostatic work — the departing H⁺ must leave a more negatively charged, more Coulombically stabilizing environment behind, which is destabilizing to the transition state/product relative to a neutral starting acid.

Step 2 — Handle molecules with chemically equivalent protons. Consider a symmetric diacid like malonic acid's diester enolate precursor, or more simply, consider a diamine with two equivalent NH₂ groups, or citric acid's terminal CH₂-COOH groups (there are two identical ones). If a molecule has $n$ equivalent, independent protons, the microscopic pKa of any one of them, pKa $_{micro}$ , relates to the observed macroscopic pKa $_1$ (the first ionization measured experimentally) by a statistical factor.

Step 3 — Derive the statistical factor by counting pathways. The macroscopic $K_{a1}$ measures loss of any proton from $n$ equivalent sites, so there are $n$ equally likely microscopic dissociation pathways contributing to the forward rate/equilibrium, but only 1 pathway for the reverse (reprotonation can land back on any one of $n-1$ remaining equivalent sites in the product — for the first ionization step, before any site has reacted, the correction is purely on the forward side):

K_{a1}^{macro} = n \times K_{a,micro}

Taking $-\log_{10}$ of both sides:

\text{pKa}_1^{macro} = \text{pKa}_{micro} - \log_{10} n

So for $n=2$ equivalent acidic protons, $\text{pKa}_1^{macro} = \text{pKa}_{micro} - 0.301$ — the observed first pKa is about 0.3 units lower (more acidic-looking) than any single site's intrinsic pKa, purely from the statistics of there being two ways to lose a proton.

Step 4 — Show the correction reverses for the last ionization. For the final proton leaving an $n$ -fold symmetric site (going from 1 remaining acidic H to 0), there is only 1 way to lose it forward, but $n$ ways for the fully deprotonated species to be reprotonated (landing on any of the $n$ equivalent basic sites), so:

K_{a,\text{last}}^{macro} = \frac{1}{n} K_{a,micro} \quad\Rightarrow\quad \text{pKa}_{\text{last}}^{macro} = \text{pKa}_{micro} + \log_{10} n

This is exactly why, for a symmetric diprotic acid, even if both protons were intrinsically identical (same microscopic pKa), the observed pKa₁ and pKa₂ would still differ by $2\log_{10}n = 0.6$ for $n=2$ — a purely statistical/entropic splitting with no electronic origin at all. Any splitting larger than this in real diacids (e.g., oxalic acid's huge 2.9-unit gap) is the additional, genuinely electronic/inductive contribution layered on top of the statistical 0.6.

Worked examples

Example 1 (foundational): Will hydroxide deprotonate phenol, and by how much? Phenol has pKa (in water) = 10.0. Hydroxide is the conjugate base of water, pKa(H₂O) = 15.7. We want the reaction

\text{PhOH} + \text{OH}^- \rightleftharpoons \text{PhO}^- + \text{H}_2\text{O}

Here the "substrate acid" is PhOH (pKa₁ = 10.0), and the "conjugate acid of the base" is H₂O (pKa₂ = 15.7), since OH⁻ is the base and H₂O is what it becomes once protonated. Applying the master equation:

\log K_{eq} = \text{pKa}_2 - \text{pKa}_1 = 15.7 - 10.0 = 5.7

K_{eq} = 10^{5.7} \approx 5.0\times10^{5}

This is an enormous equilibrium constant: hydroxide deprotonates phenol essentially completely (>99.999%). Converting to free energy using 1.36 kcal/mol per unit:

\Delta G^\circ = -1.364 \times 5.7 = -7.8 \text{ kcal/mol}

Limiting-case check: if the substrate's pKa exactly equaled water's pKa (a hypothetical acid of pKa 15.7), ΔpKa would be 0 and $K_{eq}=1$ — hydroxide would deprotonate it to exactly 50%, consistent with the fact that OH⁻/H₂O is by definition the pKa-15.7 reference pair.

Example 2 (applied): Choosing between LDA and NaOEt for enolizing ethyl acetoacetate at the more acidic (central) CH₂. Ethyl acetoacetate's central CH₂, flanked by two carbonyls, has pKa ≈ 11 (in water; ~14.2 in DMSO, but we'll stay in the aqueous-referenced scale used by common tables for consistency). We compare two bases: Sodium ethoxide, conjugate acid ethanol, pKa(EtOH) = 16.

\log K_{eq} = 16 - 11 = 5, \qquad K_{eq}=10^{5}

This is overwhelmingly favorable — ethoxide fully deprotonates the 1,3-dicarbonyl. This matches the textbook fact that NaOEt/EtOH is the classic base/solvent pair for malonic-ester and acetoacetic-ester syntheses. LDA, conjugate acid diisopropylamine, pKa(iPr₂NH) ≈ 36 (a strongly basic amine pKa, typically quoted for DMSO/THF-referenced "non-aqueous" tables, but used here for illustration of scale).

\log K_{eq} = 36 - 11 = 25, \qquad K_{eq}=10^{25}

LDA would deprotonate ethyl acetoacetate essentially irreversibly and instantaneously — chemically correct, but it is serious overkill: LDA's real purpose is reserved for substrates with pKa ≈ 25–30 (simple ketone enolization, ester enolates) where NaOEt (pKa 16) would give only a small equilibrium population. This example demonstrates the central practical lesson of quantitative pKa comparison: bigger ΔpKa is not always "better" — match the base's strength to the substrate so that side reactions (over-deprotonation, nucleophilic addition to the carbonyl, retro-Claisen cleavage) are avoided. This foreshadows the deeper question of how to pick the right base for a given job, which is a quantitative optimization, not just a search for the largest possible $K_{eq}$ .

Example 3 (exam/competition level): Percent enolization and a mixed-solvent pKa mismatch trap. A problem gives you: cyclohexanone, pKa (DMSO) = 26.6; triethylamine, conjugate acid Et₃NH⁺, pKa(H₂O) = 10.75. You are asked: "What fraction of cyclohexanone is enolized by triethylamine in DMSO at equilibrium?"

The trap: you cannot subtract these two pKa's directly — one is referenced to DMSO, the other to water, and DMSO pKa's run 10–20 units higher than aqueous pKa's for the same functional-group class because water strongly, differentially stabilizes small anions (especially through hydrogen bonding) relative to DMSO, which cannot hydrogen-bond to anions at all. Mixing scales gives a number with no physical meaning.

Correct approach: convert Et₃NH⁺'s pKa to the DMSO scale using the known DMSO pKa of triethylammonium, which is tabulated directly as pKa(Et₃NH⁺, DMSO) ≈ 9.0 (ammonium pKa's shift far less between water and DMSO than carbon-acid or alcohol pKa's, because the charge type — cation losing a proton to give a neutral amine — is less sensitive to anion solvation; this is itself a fact you must know is tabulated separately, not assumed).

\log K_{eq} = \text{pKa(Et}_3\text{NH}^+\text{, DMSO)} - \text{pKa(cyclohexanone, DMSO)} = 9.0 - 26.6 = -17.6

K_{eq} = 10^{-17.6}

This is a vanishingly small equilibrium constant — triethylamine is nowhere near basic enough to enolize cyclohexanone to any measurable extent; essentially 0% conversion. Using the logistic conversion formula for confirmation:

f_{\text{enolate}} = \frac{10^{\Delta\text{pKa}}}{1+10^{\Delta\text{pKa}}} = \frac{10^{-17.6}}{1+10^{-17.6}} \approx 2.5\times10^{-18}

essentially zero, exactly consistent with $K_{eq}\ll1$ . The pedagogical point is double: (1) triethylamine, despite being a perfectly good base for protonated amines or carboxylic acids, is far too weak to touch a simple ketone's α-CH, and (2) the arithmetic only works because both pKa's were first converted onto the same solvent scale — mixing an aqueous amine pKa with a DMSO carbon-acid pKa would have given a wildly wrong, meaningless $K_{eq}$ (naively, $10.75-26.6=-15.85$ , which happens to look similar here but is coincidental and not generally safe; for many functional groups the aqueous/DMSO discrepancy is 10+ units and produces answers wrong by many orders of magnitude).

Limiting cases, scope & common traps

Situation	What to do	Why
Same solvent scale for both pKa's	Direct subtraction: $\log K_{eq} = \Delta\text{pKa}$	Additivity of ΔG° derived above holds exactly
Mixed scales (aqueous acid, DMSO base, or vice versa)	Convert one pKa using known cross-scale correlations, or find both values on one table	DMSO and H₂O differentially solvate anions; naive subtraction is meaningless
Symmetric/equivalent acidic protons	Apply $\log_{10}n$ statistical correction before comparing "intrinsic" acidities	Macroscopic pKa reflects both electronics and the number of equivalent sites
Polyprotic acid, want fraction in each protonation state	Use successive Henderson–Hasselbalch steps with pKa₁, pKa₂,… separately	Each step is a distinct equilibrium; cannot average pKa's
ΔpKa > +3	Treat reaction as "complete" (>99.9%) for synthetic planning	Logistic curve saturates; further base strength gives diminishing returns
ΔpKa between −1 and +1	Expect a genuine equilibrium mixture; side reactions on the minor species can matter	Near the steep part of the logistic curve, small pKa errors shift population a lot
Gas-phase "pKa" (really $\Delta H_{acid}$ ) tables	Do not mix with solution pKa without explicit conversion	No solvent reorganization energy included; rankings can even invert (e.g., alcohol vs. water acidity order flips between gas phase and water)

The chemist behind it

The modern pKa framework rests on the acid–base theory developed independently by Johannes Brønsted and Thomas Lowry in 1923, but the systematic use of pKa as a quantitative, tabulatable free-energy ledger for organic reactivity — rather than just an aqueous analytical chemistry tool — owes an enormous debt to Frederick G. Bordwell, whose decades of careful equilibrium measurements in DMSO built the modern organic pKa scale used throughout physical organic chemistry today.

Origins · 1 minFrederick G. Bordwell and how this was discovered

Frederick G. Bordwell

1916–2003 · Northwestern University, mid-to-late 20th century physical organic chemistry

Bordwell built one of the most extensively used quantitative datasets in organic chemistry: several thousand pKa values in dimethyl sulfoxide, measured by careful spectrophotometric and potentiometric equilibrium methods for carbon acids, N–H acids, O–H acids, and radical/redical-cation-related bond dissociation quantities. Because DMSO cannot hydrogen-bond to anions the way water can, Bordwell's scale isolates the intrinsic electronic and resonance effects that stabilize conjugate bases — inductive withdrawal, resonance delocalization, aromaticity of the anion — from the confounding, often dominant effect of differential solvation that plagues aqueous pKa comparisons for weakly acidic C–H bonds. His tables remain the standard reference for ranking carbanion, radical, and related stabilities quantitatively.

History

- 1923 — Brønsted and Lowry independently formalize the acid = proton-donor / base = proton-acceptor definition, giving pKa a universal conjugate-pair meaning.

- 1909 — Sørensen introduces the pH scale, the notational ancestor of pKa (-log convention).

- 1930s–1950s — Aqueous pKa tables for common organic functional groups (carboxylic acids, phenols, amines) are systematically compiled.

- 1960s–1990s — Bordwell and coworkers construct the DMSO pKa scale, extending quantitative acidity comparison to weak carbon acids inaccessible in water.

- Present — Combined aqueous, DMSO, and computational (DFT-derived) pKa datasets are cross-correlated and used routinely in synthetic route planning and mechanism prediction software.

Why does a bigger ΔpKa correspond to a "more complete" reaction, and how big does it need to be?

Because $K_{eq}=10^{\Delta pKa}$ , and equilibrium concentration ratios scale as powers of ten, even modest pKa gaps produce lopsided equilibria. At ΔpKa = 1, the ratio is 10:1 (91% conversion); at ΔpKa = 2, it's 100:1 (99%); at ΔpKa = 3, it's 1000:1 (99.9%). Synthetic chemists generally treat ΔpKa ≥ 3–4 as "complete" for practical planning purposes, since side reactions and workup losses usually dominate the small remaining equilibrium imperfection at that point. Below ΔpKa = 2, you should explicitly consider the equilibrium composition rather than assuming full conversion.

Can I mix an aqueous pKa table entry with a DMSO pKa table entry in the same calculation?

Not directly, and doing so is one of the most common quantitative errors. Aqueous and DMSO pKa's differ because water solvates small, charge-dense anions (like carboxylates or alkoxides) through hydrogen bonding far more effectively than DMSO can, artificially lowering the aqueous pKa relative to the "intrinsic" DMSO value. The shift is functional-group-dependent: alcohols and carbon acids shift by 10–15+ units between the two scales, while ammonium/amine pKa's shift by only a few units, because the relevant charge type (cation → neutral) is far less solvation-sensitive. Always find both values on a single consistent scale, or use an explicit, tabulated cross-scale correlation, before subtracting.

How do I handle a molecule with two different, non-equivalent acidic protons — say, a molecule with both an OH and an NH?

Treat each site with its own tabulated (or estimated) pKa exactly as if it were a separate molecule, then compare both pKa's against the base you're considering. The site with the lower pKa is more acidic and will be deprotonated preferentially (kinetically and, given enough time, thermodynamically) — this is the same logic used in Combining Effects to Rank Acidity for ranking which site reacts first. If the two pKa's are within about 1–2 units of each other, expect a genuine equilibrium mixture of both deprotonation regiochemistries rather than clean selectivity.

Why do textbooks sometimes list a pKa as a single number and sometimes as pKa₁, pKa₂, pKa₃…?

A single pKa describes a monoprotic acid — one ionizable proton, one equilibrium. Subscripted values (pKa₁ < pKa₂ < pKa₃…) describe successive ionization steps of a polyprotic acid, each a physically distinct equilibrium with its own conjugate base. You cannot average these or treat the molecule as having "one effective pKa" except in narrow pH windows far from any of the individual pKa values; near a given pKaᵢ, that step dominates the protonation-state chemistry.

Does a large positive ΔpKa guarantee the reaction will actually happen quickly?

No — this is a critical and frequently tested distinction. pKa and $K_{eq}$ are purely thermodynamic; they say nothing about the rate or mechanism of proton transfer. A reaction can be enormously favorable at equilibrium (huge ΔpKa) yet proceed slowly if steric bulk blocks the base from approaching the acidic proton, or if the transition state requires unfavorable geometry (e.g., poor orbital overlap for the developing anion's stabilization). Kinetic acidity and thermodynamic acidity can even rank substrates in different orders in some systems, which is why pKa tables predict position, not rate.

How is the statistical (symmetry) correction to pKa different from a genuine electronic effect?

The statistical correction, $\pm\log_{10} n$ per equivalent-proton step, arises purely from combinatorics — how many indistinguishable ways a proton can leave or return — and has zero electronic origin; it would exist even for a hypothetical molecule with perfectly non-interacting, energetically identical acidic sites. Genuine electronic effects (inductive withdrawal from an adjacent anion, through-space electrostatic repulsion, resonance) are layered on top and are usually much larger; for example, oxalic acid's pKa₁/pKa₂ gap of 2.9 units includes both the statistical 0.3 and a much larger electronic destabilization of removing a second proton from an already-negative dicarboxylate.

Why do amide bases like LDA have pKa(conjugate acid) values quoted around 36, so much higher than hydroxide's 15.7?

Nitrogen is less electronegative than oxygen, so N–H bonds are intrinsically harder to ionize than O–H bonds (the resulting amide anion, R₂N⁻, is far less stable/more basic than the corresponding alkoxide or hydroxide, because nitrogen accommodates negative charge less comfortably than oxygen). This large pKa gap is exactly why LDA is chosen deliberately for substrates like simple ketones and esters (pKa ≈ 20–30) where hydroxide or alkoxide bases would leave a substantial equilibrium population of un-deprotonated starting material.