2-Chloro-1,3,2-dithiaphospholane

• CAS: 4669-51-6
• Molecular Formula: C2H4ClPS2
• Molecular Weight: 158.6 g/mol
• SMILES: C1CSP(S1)Cl
• InChI: InChI=1S/C2H4ClPS2/c3-4-5-1-2-6-4/h1-2H2
• InChIKey: ADDKHKLPZQPSHI-UHFFFAOYSA-N

2-Chloro-1,3,2-dithiaphospholane (C₂H₄ClPS₂, MW 158.6) is a five-membered cyclic trivalent chlorophosphite in which both ring heteroatoms are sulfur. It serves as the foundational precursor for the dithiaphospholane method of phosphorodithioate synthesis, a route that is mechanistically distinct from both the oxathiaphospholane approach (phosphorothioates) and standard phosphoramidite chemistry.

2-Chloro-1,3,2-dithiaphospholane (CAS 4669-51-6): Procurement-Grade Cyclic Chlorophosphite for Phosphorodithioate Synthesis

2-Chloro-1,3,2-dithiaphospholane (C₂H₄ClPS₂, MW 158.6) is a five-membered cyclic trivalent chlorophosphite in which both ring heteroatoms are sulfur . It serves as the foundational precursor for the dithiaphospholane method of phosphorodithioate synthesis, a route that is mechanistically distinct from both the oxathiaphospholane approach (phosphorothioates) and standard phosphoramidite chemistry [1]. The compound is synthesized via condensation of PCl₃ with 1,2-ethanedithiol to give a pale yellow oil in 95% crude yield, characterized by a diagnostic ³¹P NMR resonance at δ 168.4 ppm (CDCl₃) [2]. Unlike its oxathiaphospholane and dioxaphospholane congeners, the all-sulfur ring environment confers a single preferred envelope conformation (a-Ec) and uniquely enables clean phosphorodithioate bond formation without phosphorothioate contamination [3][4].

S

Phosphorodithioate synthesis via dithiaphospholane route

W

Solid-phase oligonucleotide assembly workflow

C

Azido-nucleoside compatible chlorophosphite chemistry

• [1] McKenna, C. E.; Kashemirov, B. A.; Błażewska, K. M. Science of Synthesis, 2009, 42, 863–864. Method 2: Synthesis via 2-Sulfido-1,3,2-Dithiaphospholanes. View Source
• [2] Park, C.-M.; Xian, M. Methods in Enzymology, 2015, 554, 27–28. Section 2.3: Protocols and Data – 2-Chloro-1,3,2-dithiaphospholane. View Source
• [3] Plyamovatyi, A. Kh.; Vandyukova, I. I.; Shagidullin, R. R. Bulletin of the Academy of Sciences of the USSR, Division of Chemical Science, 1986, 35, 1011–1015. Molecular structure of five-membered organophosphorus compounds. View Source
• [4] Okruszek, A.; Sierzchala, A.; Sochacki, M.; Stec, W. J. Tetrahedron Letters, 1992, 33, 7585–7588. The synthesis of di- and oligo(deoxyribonucleoside phosphorodithioates) by dithiaphospholane method. View Source

Why 2-Chloro-1,3,2-dithiaphospholane Cannot Be Replaced by Generic Chlorophosphites or Phosphoramidites in Phosphorodithioate Chemistry

The selection of a phosphorus(III) precursor for thiophosphate synthesis is non-trivial because the ring heteroatom composition directly governs both the reaction chemoselectivity and the product class. 2-Chloro-1,3,2-dioxaphospholane and 2-chloro-1,3,2-oxathiaphospholane yield phosphotriesters and phosphorothioates, respectively, but cannot cleanly generate phosphorodithioates [1]. The phosphoramidite congener N,N-diisopropyl-1,3,2-dithiaphospholan-2-amine undergoes a deleterious side reaction with azido-containing nucleosides (e.g., AZT derivatives), whereas 2-chloro-1,3,2-dithiaphospholane avoids this pathway entirely [2]. Furthermore, organometallic reagents that fail with oxathiaphospholanes (e.g., organocerium reagents giving only ~30% yield) react efficiently with 2-chloro-1,3,2-dithiaphospholane via Grignard displacement to furnish 2-alkyl-2-thio-1,3,2-dithiaphospholanes in 70–82% yield [3]. These chemoselectivity gatekeeping properties mean that generic substitution of the phosphorus precursor leads either to the wrong product class, unacceptable side reactions, or severely compromised yield [4].

1

Oxathiaphospholane or dioxaphospholane analogs yield phosphorothioates or phosphotriesters, not clean phosphorodithioates.

2

Dithiaphospholan-2-amine congener fails with azido-nucleosides; the P–N side reaction may compromise product integrity.

3

Organolithium or organocerium alternatives may dramatically reduce C–P coupling yield relative to Grignard displacement.

• [1] Martin, S. F.; Wagman, A. S.; Zipp, G. G.; Gratchev, M. K. Journal of Organic Chemistry, 1994, 59, 7957–7958. General route to phosphonodithioic acid derivatives. View Source
• [2] McKenna, C. E.; Kashemirov, B. A.; Błażewska, K. M. Science of Synthesis, 2009, 42, 863–864. View Source
• [3] Martin, S. F.; Wagman, A. S.; Zipp, G. G.; Gratchev, M. K. Journal of Organic Chemistry, 1994, 59, 7957–7958. (Yield data: Grignard 70–82% vs. organocerium ~30%). View Source
• [4] Okruszek, A.; Sierzchala, A.; Fearon, K. L.; Stec, W. J. Journal of Organic Chemistry, 1995, 60, 6998–7005. Dithiaphospholane approach to phosphorodithioates. View Source

Head-to-Head Quantitative Evidence: Where 2-Chloro-1,3,2-dithiaphospholane Outperforms Close Analogs

Conformational Homogeneity: Single Envelope Conformer vs. Conformational Equilibrium in 2-Chloro-1,3,2-oxathiaphospholane

2-Chloro-1,3,2-dithiaphospholane adopts a single preferred envelope conformation (a-Ec, with the carbon atom departing from the ring plane) in both the gas phase and the solid state as confirmed by IR, Raman spectroscopy, and molecular mechanics calculations. In contrast, its closest analog 2-chloro-1,3,2-oxathiaphospholane exists as an equilibrium mixture of two conformers (ana-EC ⇄ e-Ec) [1][2]. Conformational homogeneity simplifies crystallography, spectroscopic characterization, and mechanistic interpretation of subsequent reactions.

Conformational Homogeneity

Head-to-head

1 conformer (a-Ec envelope) vs. 2 conformers for oxathiaphospholane analog.

Supports stereochemical-control context and mechanistic interpretation.

Data from IR, Raman, molecular mechanics.

Evidence Dimension	Number of populated conformers at ambient temperature
Target Compound Data	1 conformer (a-Ec envelope; C atom out of plane)
Comparator Or Baseline	2-Chloro-1,3,2-oxathiaphospholane: 2 conformers in equilibrium (ana-EC ⇄ e-Ec)
Quantified Difference	1 conformer vs. 2 conformers; enthalpy difference for oxathiaphospholane conformers determined
Conditions	IR, Raman spectroscopy; molecular mechanics (MM) calculations; Russian Chemical Bulletin, 1986, Vol. 35

Why This Matters

Single-conformer behavior eliminates stereochemical ambiguity in downstream phosphorodithioate bond formation, which is critical for reproducible synthesis of stereodefined oligonucleotide analogs.

• [1] Plyamovatyi, A. Kh.; Vandyukova, I. I.; Shagidullin, R. R. Bulletin of the Academy of Sciences of the USSR, Division of Chemical Science, 1986, 35, 1011–1015. View Source
• [2] Shagidullin, R. R.; Shakirov, I. Kh.; Sobolev, P.; Vandyukova, I. I.; Plyamovatii, A. Kh.; Nuretdinova, O. N. Bulletin of the Academy of Sciences of the USSR, Division of Chemical Science, 1986, 35, 1627–1633. Vibrational spectra and molecular conformations of 2-chloro-1,3,2-diheterophospholanes. View Source

Azide Compatibility: 2-Chloro-1,3,2-dithiaphospholane Avoids Phosphoramidite–Azide Side Reaction in Nucleoside Phosphorylation

When phosphorylating nucleosides bearing an azido group (e.g., 3′-azido-2′,3′-dideoxythymidine, AZT), the standard reagent N,N-diisopropyl-1,3,2-dithiaphospholan-2-amine undergoes a side reaction between the P–N bond and the azide functionality that compromises product integrity. 2-Chloro-1,3,2-dithiaphospholane circumvents this incompatibility because it lacks the reactive P–N bond, enabling clean formation of the 5′-O-(2-sulfido-1,3,2-dithiaphospholane) intermediate in 32–98% yield after sulfurization [1]. This chemoselectivity advantage is decisive for synthesizing phosphorodithioate analogs of azido-containing therapeutic nucleosides.

Azide Compatibility

Head-to-head

Compatible with AZT derivatives; phosphoramidite congener fails due to P–N side reaction.

Enables direct phosphorylation of azido-nucleosides.

Yields 32–98% after sulfurization.

Evidence Dimension	Compatibility with azido-containing nucleosides
Target Compound Data	2-Chloro-1,3,2-dithiaphospholane: Compatible; yields 32–98% after sulfurization
Comparator Or Baseline	N,N-Diisopropyl-1,3,2-dithiaphospholan-2-amine: Incompatible; side reaction with azido group
Quantified Difference	Qualitative pass/fail: phosphoramidite fails with azides; chlorophosphite succeeds with 32–98% yield range
Conditions	Nucleoside 5′-OH phosphorylation; DBU or tetrazole catalysis; Science of Synthesis (Thieme), 2009, Vol. 42

Why This Matters

This is the only dithiaphospholane-based method that allows direct phosphorylation of azido-nucleosides without protecting-group manipulation, saving 2–3 synthetic steps and improving overall yield.

• [1] McKenna, C. E.; Kashemirov, B. A.; Błażewska, K. M. Science of Synthesis, 2009, 42, 863–864. Synthesis via 2-Sulfido-1,3,2-Dithiaphospholanes. View Source

Synthesis Yield: 95% Crude Yield vs. Prior Multi-Step Routes Requiring Chromatographic Purification

The direct condensation of commercially available 1,2-ethanedithiol (3.56 mL, 42.5 mmol) with freshly distilled PCl₃ (1.75 g, 127 mmol) under argon at room temperature for 3 h, followed by simple distillation of excess PCl₃, delivers 2-chloro-1,3,2-dithiaphospholane as a pale yellow oil in 95% yield (6.4 g) [1]. An improved synthetic route has been reported with even higher yield than this previously established procedure [2]. This contrasts with the multi-step routes required for the corresponding 2-chloro-1,3,2-dioxaphospholane, which typically involve low-temperature (−5 °C) conditions and may require chromatographic purification, with reported yields around 90% and >99% purity by GC after distillation .

Synthesis Yield

Cross-study

95% crude yield from PCl₃ and 1,2-ethanedithiol at room temperature.

Supports procurement cost and scale-up reliability.

Compared to ~90% yield for dioxaphospholane.

Evidence Dimension	Isolated yield from PCl₃ and diol/dithiol
Target Compound Data	95% crude yield (6.4 g from 42.5 mmol ethanedithiol); used without further purification
Comparator Or Baseline	2-Chloro-1,3,2-dioxaphospholane: ~90% yield after distillation; requires low-temperature conditions
Quantified Difference	95% vs. ~90% (+5 percentage points); simpler room-temperature protocol
Conditions	Neat PCl₃, room temperature, 3 h, argon atmosphere; excess PCl₃ removed by distillation

Why This Matters

Higher yield, simpler protocol, and elimination of chromatographic purification reduce procurement cost and enable reliable scale-up for multi-gram laboratory syntheses.

• [1] Park, C.-M.; Xian, M. Methods in Enzymology, 2015, 554, 27–28. 2-Chloro-1,3,2-dithiaphospholane synthesis protocol. View Source
• [2] Souza, E. et al. Phosphorus, Sulfur, and Silicon and the Related Elements. New synthetic route for 2-chloro-1,3,2-dithiaphospholane with better yield than previous methods. View Source

Solid-Phase Oligonucleotide Synthesis: 5-Minute Coupling with >95% Step-Yield for Phosphorodithioate Linkages

The dithiaphospholane method, in which nucleoside 3′-O-(2-thiono-1,3,2-dithiaphospholane)s—prepared from 2-chloro-1,3,2-dithiaphospholane—are condensed with 5′-O-deprotected nucleosides using DBU, has been adapted to automated solid-phase synthesis. The process achieves a condensation step time of only 5 minutes with a step-yield exceeding 95% as measured by DMT cation assay [1]. In contrast, earlier H-phosphonothioate methods for phosphorodithioate synthesis required separate coupling and oxidation steps with longer cycle times and lower overall efficiency [2]. Pyrimidine pentamers and decamers produced by this method are free from phosphorothioate contamination, a critical purity advantage over routes that rely on sulfurization of phosphite triesters [3].

Solid-Phase Coupling

Cross-study

5 min condensation step; >95% step-yield in automated DNA synthesis.

Supports routine automated oligonucleotide synthesis workflow.

Based on DMT cation assay; DBU catalysis.

Evidence Dimension	Automated solid-phase coupling efficiency for phosphorodithioate oligonucleotides
Target Compound Data	5 min condensation step; >95% average step-yield (DMT assay)
Comparator Or Baseline	H-phosphonothioate method: multistep coupling/oxidation cycles; typically lower step-yields reported
Quantified Difference	5 min per cycle vs. longer multistep cycles; >95% vs. variable step-yields
Conditions	Automated DNA synthesizer, 1 μmol scale, DBU catalysis, LCA CPG solid support

Why This Matters

The short cycle time and high step-yield make the dithiaphospholane approach the only method currently suitable for routine automated synthesis of medium-length (10–30mer) phosphorodithioate oligonucleotides on a standard DNA synthesizer.

• [1] Okruszek, A.; Sierzchala, A.; Fearon, K. L.; Stec, W. J. Journal of Organic Chemistry, 1995, 60, 6998–7005. Synthesis of oligo(deoxyribonucleoside phosphorodithioate)s by the dithiaphospholane approach. View Source
• [2] Okruszek, A.; Sierzchala, A.; Sochacki, M.; Stec, W. J. Tetrahedron Letters, 1992, 33, 7585–7588. View Source
• [3] Okruszek, A.; Olesiak, M.; Balzarini, J. Journal of Organic Chemistry, 1997, 62, 2269–2272. Efficient one-pot synthesis of 2′-deoxyribonucleoside 3′-O- and 5′-O-phosphorodithioates. View Source

Nuclease Resistance of Downstream Products: Phosphorodithioate (SPS) Oligonucleotides Rated '+++' vs. Phosphorothioate (PS) Rated '+'

Oligonucleotides containing phosphorodithioate (SPS) internucleotide linkages, synthesized via the dithiaphospholane method starting from 2-chloro-1,3,2-dithiaphospholane, exhibit superior nuclease resistance (rated '+++' indicating total resistance) compared with phosphorothioate (PS) linkages (rated '+' indicating moderately increased resistance) in standardized nuclease susceptibility assays [1][2]. Specifically, oligomers containing alternating phosphorodithioate and phosphate linkages were highly resistant to degradation in HeLa cell nuclear and cytoplasmic extracts, human serum, and with purified nucleases S1 and DNase I [3]. This achiral, isosteric, and isopolar phosphate replacement retains RNase H activation capability while providing complete resistance to all known nucleases [4].

Nuclease Resistance

Class-level

Downstream SPS oligonucleotides rated '+++' vs. '+' for PS backbones.

Reported endpoint context for backbone stability screening.

In HeLa extracts, human serum, S1 nuclease.

Evidence Dimension	Nuclease resistance rating of modified oligonucleotide backbone
Target Compound Data	Phosphorodithioate (SPS): +++ (total resistance)
Comparator Or Baseline	Phosphorothioate (PS): + (moderately increased resistance); unmodified phosphodiester: − (no resistance)
Quantified Difference	+++ vs. + (two-tier improvement on a three-tier rating scale)
Conditions	HeLa nuclear/cytoplasmic extracts; human serum; S1 nuclease; DNase I; compiled in PMCID PMC12292337 Table 1 and PMCID PMC8341215 Table

Why This Matters

The +++ nuclease resistance rating of phosphorodithioate oligonucleotides directly justifies the selection of 2-chloro-1,3,2-dithiaphospholane as the enabling precursor for antisense therapeutics requiring extended in vivo half-life.

• [1] Biomolecules, 2025, 15, 965. Table 1: Nuclease resistance of modified oligonucleotides. Phosphorodithioate (SPS): +++; Phosphorothioate (PS): +. View Source
• [2] PMCID PMC8341215. Table: Reported synthetic non-natural internucleotide linkages tested for resistance to nuclease hydrolysis. Phosphorodithioate (SPS): +++. View Source
• [3] Cummins, L.; Graff, D.; Beaton, G.; Marshall, W. S.; Caruthers, M. H. Biochemistry, 1996, 35, 8734–8741. Biochemical and physicochemical properties of phosphorodithioate DNA. View Source
• [4] Martin, S. F.; Wagman, A. S. Journal of Organic Chemistry, 1993, 58, 5897–5899. General and efficient route to phosphorodithioate analogs. View Source

Grignard Reagent Compatibility: 70–82% Yield for C–P Bond Formation vs. 30% with Organocerium Reagents

2-Chloro-1,3,2-dithiaphospholane reacts efficiently with a panel of eight functionalized Grignard reagents (2a–h, 1.1 equiv) to give intermediate dithiophosphinates 3a–h, which upon sulfurization with elemental sulfur yield 2-alkyl-2-thio-1,3,2-dithiaphospholanes 4a–h in 70–82% isolated yield. The subsequent DBU-mediated ring-opening with choline tosylate delivers phosphonodithioates 5a–h in 67–82% yield [1]. In contrast, alkyllithium reagents fail to react with this compound, and the corresponding organocerium reagents (generated in situ from CeCl₃) give only ~30% yield [2]. This Grignard compatibility dramatically expands the scope of accessible phosphonodithioate derivatives compared to organolithium or organocerium alternatives.

Grignard Compatibility

Head-to-head

70–82% yield with Grignard reagents; ~30% with organocerium; 0% with alkyllithium.

Supports diverse phosphonodithioate library synthesis.

8 functionalized Grignard reagents tested.

Evidence Dimension	Yield of C–P bond formation with different organometallic reagents
Target Compound Data	Grignard reagents (RMgX): 70–82% yield for 3a–h; alkyllithium: no reaction
Comparator Or Baseline	Organocerium reagents (RCeCl₂): ~30% yield; alkyllithium: 0%
Quantified Difference	70–82% (Grignard) vs. ~30% (organocerium) vs. 0% (alkyllithium)
Conditions	Benzene or THF, room temperature, 1.1 equiv Grignard reagent; J. Org. Chem. 1994, 59, 7957–7958

Why This Matters

The broad functional group tolerance of Grignard reagents with 2-chloro-1,3,2-dithiaphospholane enables direct, high-yielding synthesis of structurally diverse phosphonodithioate libraries for drug discovery and chemical biology.

• [1] Martin, S. F.; Wagman, A. S.; Zipp, G. G.; Gratchev, M. K. Journal of Organic Chemistry, 1994, 59, 7957–7958. General route to phosphonodithioic acid derivatives. View Source
• [2] Martin, S. F.; Wagman, A. S. Journal of Organic Chemistry, 1993, 58, 5897–5899. Phosphorodithioate analogs of naturally occurring lipids. View Source

Optimal Procurement Scenarios: When 2-Chloro-1,3,2-dithiaphospholane Is the Scientifically Justified Choice

Synthesis of Phosphorodithioate-Modified Antisense Oligonucleotides Requiring Complete Nuclease Resistance

When designing antisense oligonucleotides for in vivo applications where nuclease degradation limits therapeutic half-life, phosphorodithioate (SPS) backbones synthesized via the dithiaphospholane method provide +++ (total) nuclease resistance — a two-tier improvement over the widely used phosphorothioate (PS) backbone rated at + [1][2]. The automated solid-phase protocol using 2-chloro-1,3,2-dithiaphospholane-derived synthons achieves 5-minute coupling cycles with >95% step-yield, making medium-length SPS-oligonucleotides accessible on standard DNA synthesizers [3]. This scenario applies to therapeutic antisense programs targeting viral mRNA, oncogenes, or other disease-associated transcripts where nuclease susceptibility is a known failure mode.

Phosphorylation of Azido-Containing Nucleosides (e.g., AZT Analogs) Without Protecting-Group Manipulation

Nucleoside analogs bearing azido groups, such as 3′-azido-2′,3′-dideoxythymidine (AZT) and its derivatives, cannot be phosphorylated using standard phosphoramidite reagents (e.g., N,N-diisopropyl-1,3,2-dithiaphospholan-2-amine) due to an inherent side reaction between the P–N bond and the azide functionality. 2-Chloro-1,3,2-dithiaphospholane is the recommended reagent for this substrate class, enabling direct 5′-O-phosphorylation followed by sulfurization to yield the 5′-O-(2-sulfido-1,3,2-dithiaphospholane) intermediate in 32–98% yield [4]. This eliminates 2–3 protecting-group manipulation steps and is particularly valuable for synthesizing phosphorodithioate prodrugs of azido-containing antiretroviral nucleosides.

Diversified Phosphonodithioate Library Synthesis via Grignard-Mediated C–P Bond Formation

For medicinal chemistry programs requiring structurally diverse phosphonodithioate esters—non-hydrolyzable phosphate mimics used as phospholipase inhibitors, catalytic antibody haptens, or phosphatase-stable probes—2-chloro-1,3,2-dithiaphospholane is the preferred electrophilic phosphorus building block. Reaction with functionalized Grignard reagents (RMgX, 1.1 equiv) followed by sulfurization provides 2-alkyl-2-thio-1,3,2-dithiaphospholanes in 70–82% yield, which are then opened with alcohols or amines under DBU catalysis to furnish the target phosphonodithioates in 67–82% yield [5][6]. Organolithium and organocerium alternatives give 0% and ~30% yield, respectively, making the Grignard-dithiaphospholane combination the only viable high-yielding route for many substrates [7].

Stereochemical Studies Requiring a Conformationally Homogeneous Phosphorus(III) Precursor

Mechanistic and stereochemical investigations of nucleophilic displacement at phosphorus benefit from the single-envelope conformer (a-Ec) of 2-chloro-1,3,2-dithiaphospholane, as established by IR, Raman, and molecular mechanics analysis [8]. In contrast, 2-chloro-1,3,2-oxathiaphospholane populates two conformers in equilibrium, and 2-chloro-1,3,2-dioxaphospholane adopts a different envelope (a-Eo), complicating interpretation of stereochemical outcomes [9]. Researchers studying P-stereogenic reaction mechanisms or developing stereocontrolled phosphorylation methods should select 2-chloro-1,3,2-dithiaphospholane when conformational homogeneity is required for kinetic or computational analysis.

Application

Selection Property

Validation Focus

Phosphorodithioate antisense oligonucleotide research

Automated solid-phase coupling efficiency

Nuclease-resistance endpoint review

Azido-nucleoside phosphorylation studies

Azide-compatible chlorophosphite reactivity

Product integrity without protecting-group steps

Diverse phosphonodithioate library synthesis

Grignard reagent compatibility for C–P coupling

Functional group tolerance and yield review

Stereochemical mechanism investigation

Single-envelope conformer homogeneity

Stereochemical outcome interpretation

• [1] Biomolecules, 2025, 15, 965. Table 1: Phosphorodithioate (SPS): +++ nuclease resistance. View Source
• [2] PMCID PMC8341215. Table: Internucleotide linkage nuclease resistance comparison. View Source
• [3] Okruszek, A.; Sierzchala, A.; Fearon, K. L.; Stec, W. J. Journal of Organic Chemistry, 1995, 60, 6998–7005. View Source
• [4] McKenna, C. E.; Kashemirov, B. A.; Błażewska, K. M. Science of Synthesis, 2009, 42, 863–864. View Source
• [5] Martin, S. F.; Wagman, A. S.; Zipp, G. G.; Gratchev, M. K. Journal of Organic Chemistry, 1994, 59, 7957–7958. View Source
• [6] Martin, S. F.; Wagman, A. S. Journal of Organic Chemistry, 1993, 58, 5897–5899. View Source
• [7] Martin, S. F.; Wagman, A. S.; Zipp, G. G.; Gratchev, M. K. Journal of Organic Chemistry, 1994, 59, 7957–7958. (Organocerium: ~30% yield). View Source
• [8] Plyamovatyi, A. Kh.; Vandyukova, I. I.; Shagidullin, R. R. Bulletin of the Academy of Sciences of the USSR, Division of Chemical Science, 1986, 35, 1011–1015. View Source
• [9] Shagidullin, R. R. et al. Bulletin of the Academy of Sciences of the USSR, Division of Chemical Science, 1986, 35, 1627–1633. View Source