Thiols and Sulfides: Structure and Reactivity

Cysteine and glutathione thiols form the backbone of cellular redox homeostasis: the glutathione/glutathione-disulfide (GSH/GSSG) couple buffers oxidative stress in essentially every eukaryotic cell, and its chemistry is precisely the thiolate nucleophile / disulfide interconversion developed here. Protein disulfide bonds (Cys–S–S–Cys), formed by mild oxidation of two proximal thiols, are essential structural cross-links in extracellular and secreted proteins (insulin, antibodies, keratin) — and protein-folding machinery (protein disulfide isomerase) actively catalyzes disulfide shuffling using exactly the thiolate $S_N2$-like attack on a disulfide sulfur that you can predict from thiolate's nucleophilicity. In drug design, sulfoxide stereochemistry is not a curiosity: the proton-pump inhibitor omeprazole is administered as a racemate at sulfur, while its single-enantiomer successor esomeprazole exploits the configurational stability of the sulfoxide stereocenter (predicted directly by the high inversion barrier discussed above) to achieve improved, more predictable pharmacokinetics. Sulindac, an anti-inflammatory prodrug, relies on reversible sulfoxide/sulfide interconversion in vivo. Sulfone-containing drugs (e.g., the antibiotic dapsone, and many modern kinase inhibitors) exploit the sulfone's polarity, hydrogen-bond-accepting S=O groups, and metabolic stability, all consequences of the fully-oxidized, achiral sulfur center. Industrially, mercaptans are added at trace levels (parts per billion) to otherwise-odorless natural gas and propane specifically because the human nose detects volatile thiols far more sensitively than almost any other class of small molecule — a direct, practical consequence of the small, easily-vaporized, highly polarizable S–H-containing molecule interacting with olfactory receptors. Sulfonium salts (R$_3$S$^+$), formed when a sulfide's lone pair performs $S_N2$ attack on an alkylating agent, are biology's universal methyl-transfer reagent: S-adenosylmethionine (SAM) is a sulfonium salt whose methyl group is an outstanding electrophile precisely because a positively charged sulfur is an excellent leaving group, making SAM the cell's central methylating cofactor for DNA, RNA, protein, and lipid methylation. This sulfonium leaving-group logic, together with sulfide/sulfoxide oxidation-state control, sets the stage directly for the protecting-group strategies built on related soft-nucleophile and Lewis-acid chemistry in Alcohol Protecting Groups: Acetals and Ethers, and for the general nucleophilicity/polarizability reasoning first introduced alongside Acidity of Alcohols and the Alkoxide Ion.

Ask the tutor: Why does the sulfoxide inversion barrier (~35–45 kcal/mol) so vastly exceed the amine nitrogen inversion barrier (~5–7 kcal/mol), in terms of the orbitals actually being reorganized?

Start by comparing the atoms themselves. Sulfur (covalent radius ≈ 1.02 Å, van der Waals radius ≈ 1.80 Å) is substantially larger than oxygen (covalent radius ≈ 0.66 Å). Its valence orbitals are diffuse 3s/3p rather than compact 2s/2p. Two consequences cascade from this single fact. First, bond strength: overlap between sulfur's diffuse 3p orbital and a small, tight carbon or hydrogen 1s/2p orbital is intrinsically poorer than the analogous oxygen overlap. The S–H bond dissociation energy is about 87 kcal/mol, versus 105 kcal/mol for O–H; the C–S bond (65 kcal/mol) is weaker than C–O (86 kcal/mol). Weaker $\sigma$ bonds mean thiols ionize (heterolytically, as acids) and homolyze (in radical chemistry, e.g., thiol-ene reactions) more readily than alcohols. Second, polarizability: sulfur's outer electrons are held less tightly (lower effective nuclear charge felt by valence electrons, more shielding, larger radius), so its electron cloud deforms easily in response to an approaching electrophile. This is the direct cause of thiolate's outstanding nucleophilicity, and it is the same trend that makes I$^-$ a better nucleophile than F$^-$ despite being a far weaker base. The molecular orbital picture of a thiolate anion RS$^-$ mirrors that of alkoxide: a filled, essentially non-bonding orbital of mostly sulfur 3p character carries the negative charge, along with two more lone pairs. But because that orbital is so diffuse, the anion's charge density is spread over a larger volume — lower charge density is intrinsically more willing to bond to electrophilic carbon in the $S_N2$ transition state, where partial bond formation to the incoming carbon matters more than deep solvation stabilization.

Now consider oxidation state chemistry. A neutral sulfide R–S–R has two lone pairs, exactly like an ether oxygen, but sulfur can expand beyond an octet-like electron count in its oxidized forms by using low-lying 3d-character hybrid orbitals (more precisely: negative hyperconjugation / very polarized $\sigma^*$ acceptor orbitals) to stabilize extra oxygens as strong $S=O$ bonds. This lets sulfur cycle cleanly through three oxidation levels — sulfide, sulfoxide, sulfone — each isolable, each with its own geometry.

The acidity gap between a thiol and the corresponding alcohol can be captured by a thermochemical cycle identical in structure to the alkoxide case, but with different numbers plugged in. Gas-phase acidity is dominated by bond strength and electron affinity; solution-phase acidity adds solvation.

where:

• $BDE(X$–$H)$ is the homolytic bond dissociation energy of the acidic bond (O–H or S–H),
• $EA(X)$ is the electron affinity of the resulting radical (RO• or RS•),
• $IE(H)$ is the ionization energy of atomic hydrogen (a constant, same for both),
• $\Delta G_{\text{solv}}$ is the (differential) solvation free energy of the anion produced.

Reading this equation: acidity increases (more negative $\Delta G_{\text{acid}}$, more forward-favored ionization) when the bond is weak, when the resulting radical is electron-hungry (high EA), and when the anion is well solvated. For the O–H/S–H comparison, the bond-weakening term dominates and pushes thiols to be more acidic; solvation actually opposes this (alkoxide is better solvated than the bulkier, more diffuse thiolate) but is outweighed. The second core relationship governs nucleophilicity in the $S_N2$ transition state, expressed qualitatively through the Swain–Scott/Ritchie-type framework:

where:

• $k_{\text{Nu}}$ is the rate constant for a nucleophile Nu attacking a given substrate,
• $k_{\text{ref}}$ is the rate constant for a reference nucleophile (often water) on the same substrate,
• $n_{\text{Nu}}$ is an empirical nucleophilicity parameter for Nu,
• $s$ is a substrate sensitivity parameter.

The point of citing this relation is not the specific numbers but the structural fact it encodes: nucleophilicity parameters for sulfur nucleophiles (RS$^-$, $n \approx 5$–$6$ relative to water) vastly exceed those for oxygen nucleophiles of comparable basicity (RO$^-$), because $n_{\text{Nu}}$ correlates with polarizability and desolvation ease, not with p$K_{aH}$.

Example 1 (Foundational): Ranking acidity and predicting the dominant species at physiological pH. Rank cysteine's thiol side chain (p$K_a \approx 8.3$ for the free amino acid thiol; use 8.3), acetic acid (p$K_a = 4.76$), phenol (p$K_a = 10.0$), and ethanol (p$K_a = 16$) by acidity, and determine which are more than 50% deprotonated at physiological pH 7.4. Ordering from most to least acidic (lowest to highest p$K_a$): acetic acid (4.76) < cysteine thiol (8.3) < phenol (10.0) < ethanol (16). Using the Henderson–Hasselbalch relation,

a species is more than 50% deprotonated whenever $pH > pK_a$. At pH 7.4: acetic acid (7.4 > 4.76) is essentially fully deprotonated (as acetate); cysteine's thiol (7.4 < 8.3) is mostly protonated but with a non-trivial thiolate fraction —

so about 11% of cysteine thiols are ionized at pH 7.4 ($0.126/(1+0.126) \approx 0.112$). Phenol and ethanol remain almost entirely protonated. This 11%-ionized figure is exactly why cysteine's thiolate, though a minority species, is nonetheless the reactive nucleophile in enzyme active sites (e.g., cysteine proteases): a small equilibrium population of a very good nucleophile can dominate the kinetics.

Example 2 (Applied): Synthesizing a target sulfide and predicting its oxidation product. Devise a synthesis of benzyl methyl sulfide, PhCH$_2$–S–CH$_3$, from benzyl bromide and methanethiol, then predict the product of treatment with exactly 1.0 equivalent of mCPBA at −20 °C, and separately with 2.2 equivalents at reflux. Synthesis. Deprotonate methanethiol (p$K_a \approx 10.3$) with aqueous NaOH:

(favorable, since CH$_3$SH, p$K_a$ 10.3, is a stronger acid than water, p$K_a$ 15.7). The methanethiolate then performs $S_N2$ backside attack on benzyl bromide — an excellent $S_N2$ substrate (primary, benzylic, activated by adjacent aromatic ring stabilizing the transition state):

The bromide leaves from the back side as the thiolate's 3p lone pair pushes into the C–Br $\sigma^*$; inversion occurs at the benzylic carbon (though it is not configurationally distinguishable here since that carbon is not a stereocenter — CH$_2$). 1.0 equiv mCPBA, −20 °C. Low temperature and stoichiometric control exploit the kinetic preference for oxidizing the more nucleophilic sulfide lone pair over the less nucleophilic sulfoxide lone pair. The product is the sulfoxide, PhCH$_2$–S(=O)–CH$_3$. Because the two substituents on sulfur (benzyl and methyl) are different, and the lone pair is the fourth "ligand," this sulfoxide is a stereocenter and (if made by a chiral method) would be configurationally stable — isolable as a single enantiomer, since sulfur inversion here requires ~35–45 kcal/mol, unattainable at −20 °C or even at reflux without special catalysis. 2.2 equiv mCPBA, reflux. Excess oxidant and forcing conditions push both oxidations to completion, giving the sulfone PhCH$_2$–S(=O)$_2$–CH$_3$. This product is achiral at sulfur (two identical S=O oxygens create a local mirror plane), regardless of what stereochemistry the intermediate sulfoxide had — oxidizing further to the sulfone erases the sulfur stereocenter entirely.

Example 3 (Exam/competition level): Kinetics-based mechanistic discrimination. A chemist reacts 1-bromobutane separately with sodium ethanethiolate (NaSEt) and sodium ethoxide (NaOEt) in ethanol at 25 °C and measures pseudo-first-order rate constants (nucleophile in excess): $k_{\text{SEt}^-} = 4.5\times10^{-2}\ \text{s}^{-1}$ and $k_{\text{OEt}^-} = 8.1\times10^{-4}\ \text{s}^{-1}$. (a) Compute the rate ratio and explain it mechanistically. (b) The same chemist finds that ethanethiol (p$K_a$ 10.5, pre-mixed with excess base to ensure full deprotonation) is a weaker base than ethoxide (conjugate acid ethanol, p$K_a$ 16) by more than five p$K_a$ units, yet still reacts faster. Reconcile this with any simple "stronger base = better nucleophile" heuristic. (c) Predict, qualitatively, how the rate ratio would change if the solvent were switched to anhydrous DMSO (a polar aprotic solvent), and justify. (a) Rate ratio.

Ethanethiolate reacts roughly 56 times faster than ethoxide toward the same primary alkyl bromide in the same protic solvent, under an identical $S_N2$ mechanism (backside attack, inversion at carbon, concerted bond-making/bond-breaking with bromide as leaving group). The rate law for both is second-order overall:

consistent with a bimolecular, single-step mechanism in which nucleophile attack and leaving-group departure are concerted — the defining kinetic signature of $S_N2$. (b) Reconciling basicity and nucleophilicity. Basicity is an equilibrium/thermodynamic property measuring affinity for a proton (a very small, hard electrophile with no steric demands and instantaneous, complete bond formation, no partial-bond transition state). Nucleophilicity toward carbon in the $S_N2$ transition state is a kinetic property measuring the rate of partial bond formation to a much larger, softer, more polarizable electrophilic carbon center, one that is also more sensitive to solvation shielding of the attacking anion. Ethoxide is the stronger base because oxygen's small, compact 2p orbital binds a proton with a very strong, short, highly directional bond and because the resulting neutral (ethanol) is poorly stabilized relative to the strongly-solvated alkoxide precursor being removed from solvent shell. Ethanethiolate is the weaker base for the mirror-image reason (weaker S–H bond, less favorable protonation thermodynamics) but the better nucleophile, because (i) its diffuse, polarizable electron cloud can begin bond formation to carbon at a longer distance, lowering the transition-state energy, and (ii) it is much less strongly solvated in the protic solvent (ethanol), so less desolvation energy must be paid before it can approach the electrophilic carbon. Both factors independently favor thiolate in the rate-determining event while doing nothing to help it in the equilibrium proton-affinity comparison — hence basicity and $S_N2$ nucleophilicity diverge. (c) Solvent switch to DMSO. In protic ethanol, alkoxide is heavily hydrogen-bonded and "wrapped" by solvent, imposing a substantial desolvation penalty before it can react; thiolate, poorly hydrogen-bonded to begin with, pays much less of this penalty. Switching to anhydrous DMSO (polar aprotic, no hydrogen-bond donors) removes most of the differential solvation stabilization of alkoxide, since neither anion can be hydrogen-bonded by DMSO. Ethoxide's rate should increase dramatically (its "naked," poorly-solvated anion in DMSO is now a much better nucleophile, since desolvation is no longer a barrier), while thiolate's rate increases far less (it was barely solvated to begin with). The net effect is that the 56-fold rate ratio favoring thiolate in ethanol should shrink substantially in DMSO — potentially by one to two orders of magnitude relative to the protic case — because the "naked-anion" effect disproportionately rescues the more solvation-sensitive alkoxide. This is the same solvent-dependence logic used to explain why halide nucleophilicity orders invert between protic and aprotic solvents.

Modern mechanistic understanding of sulfur nucleophilicity owes an enormous debt to physical organic chemists who quantified nucleophile/base divergence — most centrally Jack Hine and later Ralph Pearson, whose Hard–Soft Acid–Base (HSAB) framework gave a compact conceptual language for exactly the polarizability argument used throughout this article: soft nucleophiles (large, polarizable, like RS$^-$) prefer soft electrophiles (large, polarizable carbon centers with good leaving groups), while hard nucleophiles (small, compact, like RO$^-$ or F$^-$) prefer hard electrophiles (the proton, or carbonyl carbon in some regimes). Pearson formalized HSAB in 1963, synthesizing decades of empirical nucleophilicity-scale work (Swain–Scott, Edwards) into a single qualitative principle that organic chemists still use as a first-pass heuristic for predicting chemoselectivity, including the O-vs-S nucleophile competitions central to this topic.