Computational and synthetic studies on the conversion of isothiocyanates into isocyanides: inspirations from the Corey–Winter olefination

Abstract

Inspired by DFT calculations on the mechanism of the Corey–Winter olefination (CWO), it seemed possible that the desulfuration of isothiocyanates using P(III)-reagents should proceed smoothly to furnish the corresponding isocyanides. A detailed theoretical study on the mechanism of the desulfuration reaction revealed similar intermediates and transition states as compared to the CWO. Here, the only intermediate results from attack of the P reagent on the thiocarbonyl group. This ylide-type structure then undergoes smooth cycloreversion to directly liberate the corresponding isocyanide. The isothiocyanate substrate scope, limitations of the P(III)-reagent, and solvent effects are evaluated computationally and compared to experimental synthetic studies. Experimentally, the functional group tolerance of the isothiocyanate desulfuration proves to be excellent, and a large variety of solvents is tolerated including solvent-free variants. The reaction is pleasing invariant to steric hinderance and proceeds under very mild conditions, rendering this method valuable especially to otherwise difficult to synthesize isocyanides. The kinetics of various desulfuration reactions was followed by NMR spectroscopy, and the scope of this reaction and its applicability to ensuing reactions with and without isolation of the isocyanides is explored.

---

Introduction

Isocyanides, combining electrophilic and nucleophilic character at the same carbon atom, are compounds with fascinating properties based on their rich chemistry, especially in the synthesis of numerous heterocyclic systems and most important due to their eminent role in multi-component reactions.^1–13 Moreover, recent investigations on isocyanides even identify them as promising lead structures in medicinal chemistry.¹⁴ Unfortunately, many of them (not all^15,16) have an unbearable odor that definitely has hampered their application in organic synthesis.¹⁷ Related to this is the fact that only a very small number of usually commercially available isocyanides have found their way to synthetic applications.⁷ Although some interesting newer developments relaying on clever protocols for the dehydration of formamides, which is still the most prominent reaction in isocyanide synthesis, have been published,^7,9,18 we think that the desulfuration of isothiocyanates is both a powerful alternative as well as a complementary entry to this important substance class. This is all the more true given the fact that there is little literature precedent for this chemistry.^19–24 Moreover, the reagents used are either unstable,^19,20,25 expensive,²² difficult to prepare and handle²⁴ and frequently of limited constitutional scope.

During an oral examination dealing with the Corey–Winter olefination (CWO), doubts arose about the generally accepted mechanism of this reaction. As a matter of fact and to our surprise, since its first report in 1963,²⁶ and despite its wide-spread usage^27–29 including natural product syntheses,^30–32 little is known about the underlying mechanism of this important transformation. Some early experimental evidence points towards phosphite ylides³³ and carbenoid intermediates,³⁴ yet we were surprised by the lack of both experimental and computational evidence in favour of the exact mechanism. Commonly found textbook representations on the CWO mechanism show that thiocarbonates are attacked by P(III)-reagents such as P(OMe)₃ (1a) at the sulfur atom to form an intermediate, which then undergoes fragmentation into a dioxo-heterocyclic carbene. Supposedly, the latter species subsequently fragments via various proposed pathways^27–29 into CO₂ and the corresponding olefin (Scheme 1).

Scheme 1 (a) Textbook opinion on the mechanism of Corey–Winter olefination and (b) revised mechanism based on quantum chemical calculations (own calculations and Zhang and Yu³⁷). For a detailed discussion on the various historically proposed pathways leading to the formation of the carbene see also ref. 37. The “true” Corey–Winter intermediates (CW-IN1+2) and transition states identified from density functional theory (DFT) calculations are labelled CW-TS1 (not shown explicitly) and CW-TS2.

However, in preliminary theoretical studies we have found that in contrast to this common belief^{27–29,35,36} the phosphorous reagent does not attack the sulfur of the thiocarbonyl group, but rather its carbon atom (Scheme 1b). The resulting Corey–Winter intermediate (ylide CW-IN1) then rearranges into a cyclic thiaphosphirane intermediate (CW-IN2), which undergoes a cycloreversion (transition state CW-TS2) liberating a dioxo-heterocyclic carbene and the O,O,O-trimethyl phosphorthioate (2a) (additional details on our computational results are given in the SI, Fig. SI1 and SI2a–d). Indeed, at the time of preparing this manuscript, a very detailed study on the Corey–Winter olefination was published by Zhang and Yu,³⁷ fully corroborating these mechanistic findings.

A closer look at the Corey–Winter intermediates (in particular CW-IN1) and the transition state CW-TS2 leading to the carbene intermediate led us to the idea to replace the two oxygens of the thiocarbonate by a double-bonded nitrogen atom. This replacement delivers isothiocyanates 3 which may be desulfurized to isocyanides 4 in an analogous manner as described for the carbene intermediate in the CWO (Scheme 2).

Scheme 2 Analogy driven proposal for a mechanism for the desulfurization of isothiocyanates 3 to isocyanides 4.

Evidence for the formation of the proposed iminoylides can be derived from two crystal structures of an iridium³⁸ and a vanadium complex³⁹ with phenyl isothiocyanate (Fig. 1). The latter has been exploited for reactions related to the desulfuration of thiocarbonyl groups.

Fig. 1 Solid state structures of an (a) iridium³⁸ (5) and (b) vanadium complex³⁹ (6) with ylide analogs (5: d_Ir–C = 1.968 and d_Ir–S = 2.485 Å; 6: d_V–C = 2.042 and d_V–S = 2.444 Å) formed from phenyl isothiocyanate.

Results and discussion

Theoretical investigations

Guided by the supposed analogy between the Corey–Winter olefination and the formation of isocyanides we decided to undertake computational studies along the same level of theory as Yu et al. used for the CWO.³⁷ As prototype substrates for modeling the isothiocyanate desulfuration in detail, we have chosen PhNCS (3a) and MeNCS (3b) as substrates and their reaction with P(OMe)₃ (1a) as well as the Corey–Hopkins reagent 1b,⁴⁰ which is known to reduce typical CWO reaction temperatures from 110–150 °C (for 1a) to 25 to 40 °C (for 1b). The computed energy profiles (ΔH and ΔG) for these model reactions are shown in Fig. 2. In any case, the most favorable pathway clearly turned out to be the addition of the P(III)-species to the carbonyl carbon of the isothiocyanate via a first transition state (TS1). This initial step yields the corresponding stable ylide-type intermediates (IN), which then undergo fragmentation (TS2) into the isonitriles 4 and the S=P(V) products 2. The energy profiles also indicate a higher reactivity of the alkyl isothiocyanate 3b compared to its aryl counterpart 3a, and as expected. The reaction barriers involving 1b (Fig. 2b) are significantly lower than those found for 1a (Fig. 2a) with the later TS2 clearly being the rate-determining step in all reaction profiles calculated here. In the model reaction of phenyl isothiocyanate (3a) with P(OMe)₃ (1a), a first potential energy-surface (PES) scan on the interaction energy of the reactants as a function of C–S and S–P distances (Fig. SI3) clearly shows that the phosphorous agent does not approach the sulfur but preferentially attacks the carbon atom of the isothiocyanate. This behaviour is in perfect agreement with the initial step in the CWO. Although, S_N2-type transition states of the P–S-attack could be located for both the CWO and the isothiocyanate desulfuration reactions (Fig. SI4), these pathways can be safely excluded due to their much higher activation barriers (ΔΔH^‡ = +50 and ΔΔG^‡ = +35 kJ mol⁻¹ for the desulfuration of MeNCS) compared to the mechanism shown in Fig. 2.

Fig. 2 Proposed mechanism for the desulfuration of isothiocyanates using P(III)-reagents. (a and b) The energies of intermediates (IN), transition states (TS) and products of the reaction of phenyl isothiocyanate (3a, black lines) and methyl isothiocyanate (3b, blue lines) with P(OMe)₃ (1a) (a) and 1b (b) are plotted along the reaction coordinate (dashed lines: ΔH values, solid lines: ΔG Gibbs free energies). (c and d) Computed molecular geometries at stationary points of the reaction coordinates above. For transition states, the atomic displacement vectors (scaled by a factor of 2.5) associated with the imaginary vibrational mode are indicated by orange vectors in the direction of the forward reaction. All energies are given in [kJ mol⁻¹] (M06-2X/def2-QZVP) relative to the sum of energies of the corresponding reagents, respectively. Characteristic interatomic distances for the dashed bonds of TS1, IN, and TS2, as well as enlarged molecular plots are provided in Table SI1a + b, respectively.

The structure plots given in Fig. 2c and d show similar molecular geometries for TS1 and IN of the desulfuration reaction as compared to CW-TS1 and CW-IN1 (Scheme 1 and Table SI1a + b). In particular IN can be regarded as a distorted trigonal–bipyramidal coordinated phosphorus, with sulfur and one methoxy group of P(OMe)₃ (1a), or the P-phenyl ring of 1b occupying the apical positions, respectively. On the other hand, TS2 is best described as a less distorted trigonal-bipyramid with the formally electron-withdrawing isonitrile leaving group and a methoxy substituent (1a) or a P–N-bond (1b) in apical positions. With a computed N–P–N bond angle of 91.4° for free 1b, this favors the required orientation of the P–N-bonds one in apical, the other in equatorial position much better than for P(OMe)₃ (1a) with average O–P–O bond angles of 98.1°. Most characteristic, TS2 features an almost linear C≡N⋯P arrangement which makes the isonitrile a predestined leaving group.

However, for the CWO two energy minimum intermediates have been identified (Scheme 1b), a ylide-type structure (CW-IN1) and a cyclic thiaphosphirane species (CW-IN2). Their interconversion (which in fact involves another transition state) even seems to be the rate-limiting step in the CWO using 1a as the desulfurization agent.³⁷ In contrast, for the isothiocyanate desulfuration the conformational rearrangement required along the path IN → TS2 – which actually is a pseudorotation on phosphorus rotating one methoxy group (1a) or the P-phenyl ring (1b) from the apical into an equatorial position – is not the rate-limiting factor. Indeed, our simulations indicate that the PX₃-fragment in IN can rotate with very low barriers about the C–P-bond (see Fig. SI5a + b for an energy profile for the S–C–P–X torsion). Due to the better leaving group quality of the isonitrile in TS2 compared to the heterocyclic 1,3-dioxo-carbene in CW-TS2, the isothiocyanate desulfuration occurs much earlier and thus takes place before a cyclic thiaphosphirane intermediate of the type CW-IN2 can form.

It is worth mentioning that we specifically searched for a structure analogous to this second intermediate CW-IN2 of the CWO along the path of the RNCS desulfuration but were never able to find it here. The reaction energy profile over the entire PhNCS (3a) + 1b intrinsic reaction coordinate (IRC)^41–44 shows a very extended, very flat high-energy plateau or hump before TS2, which, however, is not compatible with an energy minimum structure of a hypothetical intermediate of the CW-IN2 type with a three-membered C–P–S ring (Fig. 3, and Fig. SI6 + 7 in the SI). In these IRC plots, the decreasing C–P distance in TS1 (also implying a decrease in the P–S distance) indicates attack of 1b on the isothiocyanate carbon atom (Fig. 3a), concomitantly slightly increasing the C–S distance via electron donation. The short C–S, C–P, and S–P distances (all in the range of ≈2.0–2.2 Å, forming an almost equilateral triangle C–S–P in the transition state) are indicative of the cycloreversion occurring at TS2 (see circular insert in Fig. 3b) liberating the isonitrile. For all reaction coordinates shown in Fig. 2 (reactions of 1a and 1b with 3a and 3b, respectively), the corresponding IRC plots are given in Fig. SI6 + 7, and for some orbital animations of the “electron-flow” calculated along the IRC around TS2 (corresponding to the traditional “arrow-pushing”), see also the SI (MP4 files and Table SI2).

Fig. 3 Electronic energy (black curves, left ordinate, M06-2X/def2-TZVP; all energies are given as Δε₀ relative to the sum of the electronic energies of the reagents) and selected interatomic distances (colored curves, right ordinate) computed along the IRC of the (a) first (TS1) and (b) second (TS2) transition state for the reaction of phenyl isothiocyanate (3a) with the Corey–Hopkins reagent 1b. The dashed curves and arrows indicate switching of the IRC calculation to simple energy minimization.

In order to evaluate the scope of the desulfuration reaction, we computed the energy profiles along the reaction coordinate for a total of 9 isothiocyanates 3a–i with the Corey–Hopkins reagent 1b (Fig. 4). The overall shape of these profiles remains as discussed above, with TS2 always being the rate-limiting step. The entire process is quite insensitive with respect to steric hinderance with the only exception being O,O′-di-tert-butyl-phenyl isothiocyanate 3e. Even for the O,O′-dimethyl-phenyl isothiocyanate 3f or the adamantyl-derivative 3c, the decisive energy barrier is not raised significantly (note the axis break in Fig. 4b). Also, electronic effects, caused by substituents (p-OMe, p-NMe₂, or p-NO₂, 3g–i) on the substrates play a minor role only, with p-nitro-phenyl isothiocyanate 3i and MeNCS 3b being the most reactive substrates investigated.

Fig. 4 (a) Energy profiles (dashed lines: ΔH values, solid lines: ΔG Gibbs free energies) for the reaction of various isothiocyanates 3a–i with 1b. The inset (b) gives the ΔG^‡ Gibbs free energies for the rate limiting transition state TS2 in color-coded form for the different substrates shown on the right (c); all energies are given in [kJ mol⁻¹] (M06-2X/def2-TZVP). Structure plots for all transition states and intermediates calculated are shown in Table SI3.

In addition to our study on substrate scope, we have investigated various P(III)-reagents 1a–g for their suitability as desulfuration agents (Fig. 5). With the model substrate phenyl isothiocyanate 3a, P(NMe₂)₃ (1f) and PMe₃ (1g) clearly exhibit the highest barriers (TS2), whereas a much lower barrier is computed for P(OMe)₃ (1a). As expected, the reactivity of 1b and its N,O- or O,O-analogs 1c and 1d appear to be increased as compared to P(OMe)₃, though the ylide intermediate IN seems to be least favored with the O,O-reagent 1d. In summary, cyclic P(III)-reagents seem to perform better than acyclic PX₃ reagents, and at least one electronegative ring atom (N or O) favorably occupies the apical position in TS2, whereas in its pseudo-equatorial position a sterically non-demanding σ-donor seems preferred (methyl over phenyl).

Fig. 5 (a) Energy profiles (dashed lines: ΔH values, solid lines: ΔG Gibbs free energies) for the reaction of phenyl isothiocyanate (3a) with seven different P(III)-reagents 1a–g. The inset (b) gives the ΔG^‡ Gibbs free energies for the rate limiting transition state TS2. (c) Transition state molecular geometries (TS2); the corresponding (refined) TSs for the reaction of phenyl isothiocyanate (3a) with P(OMe)₃ (1a) and 1b are shown in Fig. 2c and d. The atomic displacement vectors (scaled by a factor of 2.5) associated with the imaginary vibrational mode of the TSs are indicated by orange vectors in the direction of the forward reaction. All energies are given in [kJ mol⁻¹] (M06-2X/def2-TZVP). Structure plots for all transition states and intermediates calculated are shown in Table SI4.

As solvation effects may play a major role in chemistry,^45–47 we have investigated theoretically as well as experimentally (see Application section below) the impact of various solvents with increasing polarity (benzene, diethylether, dichloromethane, acetone, and acetonitrile) on the desulfuration reactivity of PhNCS (3a) and MeNCS (3b) with 1b (Fig. 6). As expected, the polar structure of the ylide type intermediate IN is significantly stabilized by polar solvents. However, the decisive rate limiting transition state TS2 is predicted to be largely unaffected from solvent effects which is confirmed by the experimental results described below (Table 1). As all relative energies of TS2 fall in a very narrow range and well within the error margins of DFT calculations, the enthalpy and free energy of activation are given in Fig. 6 only as values averaged over all solvents along with the value predicted for the hypothetical gas phase. For further details see the Table SI7a + b.

Fig. 6 Comparison of gas phase (black curves) and solvent-corrected energies (colored lines; dashed lines: ΔH values, solid lines: ΔG Gibbs free energies for IEF-PCM solvation models, see Computational methods below) for the reaction of (a) phenyl isothiocyanate (3a) and (b) methyl isothiocyanate (3b) with the 1b. For the different solvents, only the average ΔH^‡ and ΔG^‡ values are given along with their root mean square deviations, the values in brackets refer to the gas phase. All energies are given in [kJ mol⁻¹] (M06-2X/def2-TZVP).

Table 1 Solvent compatibility of the desulfurization reaction

Entry	Solvent	Yield (4g) [%]	Entry	Solvent	Yield (4g) [%]
1	n-Hexane	75	7	Toluene	67
2	THF	70	8	Acetone	34
3	benzene	86	9	EtOAc	82
4	CH2Cl2	69	10	2-MeTHF	80
5	Et2O	73	11	MTBE	70
6	CH3CN	68

---

For PhNCS (3a) all values ΔH^‡ and ΔG^‡ (TS2) are almost invariant to solvent effects at all, and only for MeNCS (3b) the gas phase barrier of activation seems to be slightly underestimated by solvent-free calculations. Either way, the calculations show that – as long as no substrate-specific chemical reasons interfere – the desulfuration of isothiocyanates should allow for a very large variability of the solvent to be chosen.

In summary, all our computational evaluations suggest that the desulfuration reaction reported here should proceed very similar to the CWO, with one significant mechanistic difference. The thiaphosphirane CW-IN2, which is a true energy minimum structure (intermediate) in the CWO becomes a transition state TS2 here (Fig. 7). Both structures show a similar trigonal–bipyramidal coordination of the phosphorus atom, yet the isonitrile as a better leaving group as compared to a carbene causes the earlier transfer of the sulfur atom onto phosphorus. In this transition state TS2 the isonitrile is released seamlessly through a characteristic de-bending motion of the R–N≡C moiety.

Fig. 7 Analogous structures of an (a) intermediate of the Corey–Winter olefination and (b) the decisive transition state of the isothiocyanate desulfuration reported in this work (M06-2X/def2-QZVP structures). For the transition state on the right, the atomic displacement vectors (scaled by a factor of 2.5) associated with the imaginary vibrational mode of the TS are indicated by orange vectors in the forward direction of the reaction.

Based on these calculations, the reaction should obey second-order kinetics (first order in both isothiocyante and the Corey–Hopkins-reagent). We were able to confirm this by means of a time-resolved ¹H-NMR study of the reaction process. Details of this study can be found in the SI (Fig. SI9).

Applications

From the quantum chemical calculations described above, the three 1,3-diheterosubstituted phospholidines 1e, 1c and 1b have been identified as the most promising reagents for the desulfuration of isothiocyanates to isocyanides (Chart 1).

Chart 1 The most promising candidates for the desulfuration of isothiocyanates to isocyanides (ΔG^‡ values given for the rate limiting step in Fig. 5).

The diazaphospholidine 1e with the lowest barrier in the rate limiting step was not considered further because, different from phenyldichlorophosphane necessary for the synthesis of the Corey–Hopkins-reagent 1b,⁴⁰ its synthetic precursor, methyldichlorophosphane is not commercially available. The oxygen-containing heterocyclic systems, including the Mukaiyama-reagent 1c²⁰ and the dioxophospholidine 1d have proven to be rather unstable compounds which makes their handling and storage tedious.^21,25,48,49 Although 1c has been used to desulfurize an isothiocyanate in the context of a total synthesis of welwitindolinones,⁵⁰ we decided to prefer the readily available Corey–Hopkins reagent 1b as the best compromise between reactivity and ease of use. Given the numerous possibilities to synthesize isothiocyanates from different starting materials, it is an advantageous feature of the presented method that its applicability is related to these easily accessible compounds. In the sequel, we describe applications of the developed desulfuration reaction starting from a large variety of isothiocyanates.

Based on the assumption that the comparatively non-polar isothiocyanates, unlike formamides, should be soluble in a variety of solvents, it was interesting to see how solvent-tolerant the desulfurization reaction with 1b would be (Table 1).

To our delight and in accordance with our calculations (see above), the reaction turned out to be amazingly tolerant toward changes of the solvent. Apart from acetone almost every type of solvent (aliphatic/aromatic hydrocarbons, ether, ester, chlorinated solvents and acetonitrile) is compatible with the desulfurization. This is a major advantage as compared to the dehydratization of formamides which has proven to fail in many solvents except dichloromethane.⁷

Isocyanides via isothiocyanation of amines

Among the numerous possibilities to synthesize isothiocyanates,^51–54 their preparation from amines may be the most prominent one. In pleasant contrast to isocyanates, which are usually synthesized using phosgene or oligophosgenes, their production does not necessarily require thiophosgene but can also be prepared with carbon disulfide. To evaluate the constitutional scope of the desulfurization, we prepared a number of isothiocyanates from aromatic and aliphatic amines following known procedures and reacted them with the diazaphospholidine 1b (Table 2).

Table 2 Substrate scope of the desulfurization of isothiocyanates with 1b

Entry		3	R^1	R^2	R^3	4	Yield^a [%]
a Isolated yields.b Reaction time 48 h.c After sublimation from the neat reaction mixture.d Neat, 50 °C, 5 h sonification (35 kHz, 120 W).
1		3j	Me	H	H	4j	83
2		3g	OMe	H	H	4g	86
3		3k	SMe	H	H	4k	82
4		3h	NMe2	H	H	4h	77
5		3l	CN	H	H	4l	40
6		3m	Br	H	H	4m	80
7		3n	C≡C	H	H	4n	74
8		3o	OH	H	H	4o	0
9		3p	H	OMe	H	4p	72
10		3d	H	tBu	H	4d	85^b
11		3q	H	Br	H	4q	53
12		3f	H	Me	Me	4f	73^b/51^c
13		3r	H	iPr	iPr	4r	0/82^d

---

Entries 1–13 show the yields of isocyanides 4 starting from substituted phenyl isothiocyanates 3d, 3f–h and 3j–r. With the exception of the p-cyano substituted derivative 4l (entry 5) all isolated yields are around 80%. The ortho-tBu substituted substrate 3d demonstrates that considerable steric bulk is tolerable at least in cases when only one ortho position is occupied and the reaction time is increased to 48 h. But also the O,O′-dimethyl derivative 3f delivered the isonitrile 4f with a good yield of 73% under the same conditions as for 3d. The last case in this series is of special interest and importance (entry 13). Under standard conditions, even after prolonged reaction times no product was obtained from the sterically strongly congested O,O′-diisopropyl derivative 3r. This changed dramatically when the reaction was carried out without solvent (neat) under 5 h of sonication (35 kHz, 120 W). Now an impressive yield of 84% was achieved! This very beneficial effect was also observed with the 1-adamantyl isothiocyanate 3c. Here the yield increases from 57% under standard conditions to 84% under sonication. We recommend these reaction conditions (neat under sonication, 50 °C, especially with solid isothiocyanates). Apart from these aromatic substrates the desulfurization of benzylic substituted derivatives delivering the isocyanides 4v and 4w posed no problems as well as the synthesis of the aliphatic compounds 4ab and the new isocyanide 4x. The 3-isocyanopyridine 4t, isolated in 59% yield, merits special attention. Although its application in synthesis is mentioned several times, only very limited characterization data is available. Either it is described as a crude product without any analytical data,⁵⁵ or it is isolated in 30% yield as a substance that “polymerizes after removal of the solvent”.⁵⁶ Obviously, the avoidance of any kind of aqueous workup with the method described here, is very beneficial for the preparation of sensitive isocyanides.

In case of the chiral isocyanides 4w, 4z and 4aa no erosion of the optical purity occurred with 4w, but during the desulfurization reaction of the isothiocyanates leading to 4z and 4aa considerable racemization took place. Though we have not modeled this racemization process and the corresponding energy barriers in detail, it could possibly result from an intramolecular proton shift in the ylide type intermediate (IN) of the desulfuration (Scheme 3).

Scheme 3 Putative mechanism for the racemization during the desulfuration of α-amino acid derived isothiocyanates, involving the intramolecular formation of an achiral α-imino ester (inserted structure plot for the corresponding intermediate that may occur during the conversion 3aa → 4aa; M062X/def2-TZVP).

A limitation of the constitutional scope of the process is its incompatibility with OH-groups – presumably due to alcoholysis of the P(III) reagent – as exemplified by failures to prepare the phenol 4o and the aliphatic alcohol 4y.

Isocyanides via addition of HSCN to C–C-double bonds

This way to prepare isocyanides using olefins as substrates is equivalent to an otherwise difficult or even impossible to perform a hydroisocyanidation reaction. Cyanides and thiocyanates are ambident nucleophiles. This entails the difficulty to control their reactive site (N vs. C in case of the cyanides and N vs. S in case of the thiocyanates). Contrasting the situation with cyanides, this problem is much less pronounced with the thiocyanates due to the existence of many methods for the targeted synthesis of isothiocyanates.⁵¹ Moreover, the equilibrium between thiocyanates and isothiocyanates is in favour of the latter, which offers the opportunity to isomerize a mixture of both to the desired isothiocyanates.^23,57–60

As an illustrative example, we prepared the isothiocyanate 3ac following a literature procedure via hydroisothiocyanation of mesityloxide 7 by conjugate addition of in situ prepared isothiocyanic acid^61–63 (Scheme 4).

Scheme 4 Preparation of 4-isocyano-4-methylpentan-2-one 4ac.

To our delight this isothiocyanate reacted smoothly to the hitherto unknown isocyanide 4ac in 65% yield.

An attractive entry to exo-norbornylisocyanide 4ad was found via the desulfuration of the corresponding isothiocyanate 3ad which in turn was easily accessible by stereoselective hydroisothiocyanation of norbonene 8⁶⁴ (Scheme 5).

Scheme 5 A short entry to exo-norbornylisocyanide 4ad.

In this case the isothiocyanic acid was prepared in situ by grinding KSCN with potassium hydrogensulfate and its subsequent reaction with norbornene delivers 3ad in 81% yield as a pure diastereomer. After desulfurization with 1b, the desired isocyanide 4ad was obtained in 76% yield.

In situ transformations

Given the extremely unpleasant smell⁷ of most isocyanides (not all^15,16) an attractive way of using them as reactants would be to generate them “in situ”, thus avoiding their isolation (Scheme 6).

Scheme 6 In situ reactions. (a) 1b, THF, 40 °C, 18 h; (b) NiCl₂·6H₂O (1 mol%), THF, RT, 24 h.

As a first example we chose an Ugi-3-component reaction^3,8 with isothiocyanate 3f as the starting material. Its reaction with 1b delivered isonitrile 4f which was not isolated but reacted further with aqueous formaldehyde solution and diethyl amine to yield the local anesthetic lidocaine 9 in 59% yield. Tetrazole 10 was available in a one-pot-sequence starting with isothiocyanate 3g via reaction of the in situ generated isocyanide 4g with TMS-azide. Finally, we used the alanine derivative 3ae to generate the corresponding isocyanide 4ae which was, again without isolation, polymerized to the helically chiral polyisocyanide poly-4ae^65–67 in 45% yield.

Conclusions

Guided by the results of a computational study to elucidate the mechanism of the Corey–Winter Olefination, a reaction pathway describing the desulfurization of isothiocyanates was calculated. From this, the diazaphospholidine 1e, the Mukaiyama-reagent 1c and the Corey–Hopkins reagent 1b were identified as the most promising candidates for the intended isocyanide synthesis from isothiocyanates. Due to the expected difficulties associated with the synthesis of 1e and the known handling problems with 1c, caused by its instability, we have focused on using 1b for the desulfurization reactions. From our synthetic studies, it turned out that the new isocyanide synthesis offers a number of beneficial features compared to the classical formamide dehydration:

(1) The reaction is very unsensitive to solvent effects. This, together with the favorable solubility properties of the isothiocyanates, allows for the application of many different solvents of different polarities including solvent-free conditions.

(2) The constitutional scope of the reaction comprises aliphatic as well as aromatic isothiocyanates. It is surprisingly insensitive to steric hindrance and can be accelerated considerably by sonication (e.g. entry 13, Table 2).

(3) Hydroisothiocyanation followed by desulfurization effectively achieves hydroisocyanidation of olefins (Schemes 4 and 5).

(4) The reaction conditions allow for the in situ generation and application of isocyanides in subsequent reactions without their isolation (Scheme 6).

One limitation of the method is its incompatibility with OH-groups like in the valinol derivative 4y. Although not investigated, we expect that other acidic functional groups may pose problems too, although NH-groups are tolerated (4x, Table 2). Moreover, in case of α-amino acid derived isothiocyanates leading to the corresponding isocyanides 4z and 4aa (Table 2), considerable racemization took place.

From a sustainability perspective, the dehydration of formamides may be the environmentally friendlier reaction. Both, the reagent 1b as well as the byproduct 2b are used and produced in stoichiometric amounts, respectively. This means that 2b is stoichiometric waste. On the other hand, the dehydration of formamides needs stoichiometric amounts of either phosgene (or oligomers of it) or phosphorus oxychloride (POCl₃) which is not the most sustainable and harmless solution either.

Future work will be directed towards the synthesis of even more reactive P–N-based desulfurization agents based on 4-membered ring systems like cyclodiphosphazanes.^68,69 The idea behind is to reduce the reaction temperature without compromising both, configurational integrity and yield. The stereochemically advantageous effect of a reduced reaction temperature was already demonstrated by Danishefsky et al. in 2009.⁷⁰ Preliminary calculations on these systems appear very promising.

Author contributions

S. I. and M. R. have conceived and supervised the project. S. I. initiated the project, carried out the DFT-calculations and analyzed the kinetic data. D. B. did most of the experimental work and measured the NMR data. J. K. contributed experiments related to stereochemical issues of the desulfurization. D. B., S. I. and M. R. wrote the manuscript.

Conflicts of interest

There are no conflicts to declare.

Data availability

The data supporting this article have been included as part of the supplementary information (SI). Supplementary information (including computational details, experimental procedures, and NMR spectra), as well as a ZIP file (atomic coordinates of computed structures), an EXCEL file (computed energies), and three MP4 files (orbital animations) is available. See DOI: https://doi.org/10.1039/d6qo00120c.

Acknowledgements

The authors would like to thank the German Research Council (DFG) for funding under RE 1007/11-1 (MR) and IM 28/3-1 (SI). We are also grateful to the Center for Scientific Computing (CSC) of the Goethe University Frankfurt for granting access to the high-performance computing cluster and providing the CPU time required for the DFT calculations.

References

1. I. Ugi and C. Steinbrückner, Über ein neues Kondensations-Prinzip, Angew. Chem., 1960, 72, 267–268.
2. I. Ugi and C. Steinbrückner, Isonitrile.9. Alpha-Addition von Immonium-Ionen und Carbonsaure-Anionen an Isonitrile, Chem. Ber./Recl., 1961, 94, 2802–2814.
3. A. Dömling and I. Ugi, Multicomponent reactions with isocyanides, Angew. Chem., Int. Ed., 2000, 39, 3168–3210.
4. I. Ugi and S. Heck, The multicomponent reactions and their libraries for natural and preparative chemistry, Comb. Chem. High Throughput Screening, 2001, 4, 1–34.
5. I. Ugi, M. Almstetter, H. Bock, A. Domling, B. Ebert, B. Gruber, C. Hanusch-Kompa, S. Heck, K. Kehagia-Drikos, K. Lorenz, S. Papathoma, R. Raditschnig, T. Schmid, B. Werner and A. von Zychlinski, MCR XVII. Three types of MCRs and the libraries - Their chemistry of natural events and preparative chemistry, Croat. Chem. Acta, 1998, 71, 527–547.
6. S. Shaabani and A. Dömling, The Catalytic Enantioselective Ugi Four-Component Reactions, Angew. Chem., Int. Ed., 2018, 57, 16266–16268.
7. P. Patil, M. Ahmadian-Moghaddam and A. Dömling, Isocyanide 2.0, Green Chem., 2020, 22, 6902–6911.
8. A. Dömling, Recent developments in isocyanide based multicomponent reactions in applied chemistry, Chem. Rev., 2006, 106, 17–89.
9. C. G. Neochoritis, S. Stotani, B. Mishra and A. Dömling, Efficient Isocyanide - less Isocyanide - Based Multicomponent Reactions, Org. Lett., 2015, 17, 2002–2005.
10. C. G. Neochoritis, T. Zarganes-Tzitzikas, S. Stotani, A. Dömling, E. Herdtweck, K. Khoury and A. Dömling, Leuckart-Wallach Route Toward Isocyanides and Some Applications, ACS Comb. Sci., 2015, 17, 493–499.
11. Q. Wang, D. X. Wang, M. X. Wang and J. P. Zhu, Still Unconquered: Enantioselective Passerini and Ugi Multicomponent Reactions, Acc. Chem. Res., 2018, 51, 1290–1300.
12. L. Banfi, A. Basso, C. Lambruschini, L. Moni and R. Riva, The 100 facets of the Passerini reaction, Chem. Sci., 2021, 12, 15445–15472.
13. A. Pinna, A. Basso, C. Lambruschini, L. Moni, R. Riva, V. Rocca and L. Banfi, Stereodivergent access to all four stereoisomers of chiral tetrahydro-benzo[f][1,4]oxazepines, through highly diastereoselective multicomponent Ugi–Joullié reaction, RSC Adv., 2020, 10, 965–972.
14. F. Brunelli, C. Ceresa, S. Aprile, L. Coppo, B. Castiglioni, M. Bosetti, L. Fracchia and G. C. Tron, Isocyanides in med chem: A scaffold hopping approach for the identification of novel 4-isocyanophenylamides as potent antibacterial agents against methicillin-resistant Staphylococcus aureus, Eur. J. Med. Chem., 2023, 246, 114950.
15. M. C. Pirrung, S. Ghorai and T. R. Ibarra-Rivera, Multicomponent Reactions of Convertible Isonitriles, J. Org. Chem., 2009, 74, 4110–4117.
16. M. C. Pirrung and S. Ghorai, Versatile, Fragrant, Convertible Isonitriles, J. Am. Chem. Soc., 2006, 128, 11772–11773.
17. Zu weiteren Versuchen hatte ich kein Material mehr und es war mir nicht möglich, neue Mengen Cyanallyl darzustellen, weil von allen Seiten aus der Nachbarschaft des Laboratoriums Klagen über Gestank einliefen; (I had no more material for further experiments and it was not possible for me to prepare new quantities of cyanallyl because complaints about the smell were coming in from all sides in the neighborhood of the laboratory); W. Lieke, Ueber das Cyanallyl, Justus Liebigs Ann. Chem., 1859, 112, 316–321.
18. F. Brunelli, S. Aprile, C. Russo, M. Giustiniano and G. C. Tron, In-water synthesis of isocyanides under micellar conditions, Green Chem., 2022, 24, 7022–7028.
19. T. Mukaiyama and Y. Yokota, Deoxygenation of Isocyanates by 2-Phenyl-3-Methyl-1,3,2-Oxazaphospholidine. A Convenient Method for Preparation of Isonitriles, Bull. Chem. Soc. Jpn., 1965, 38, 858–859.
20. T. Mukaiyama, M. Okamoto and H. Nambu, Deoxygenations of Isocyanates and Diphenylketene by Tertiary Phosphites, J. Org. Chem., 1962, 27, 3651–3654.
21. S. V. Ley and S. J. Taylor, A polymer-supported 1,3,2 oxazaphospholidine for the conversion of isothiocyanates to isocyanides and their subsequent use in an Ugi reaction, Bioorg. Med. Chem. Lett., 2002, 12, 1813–1816.
22. Y.-S. Liu, M.-Z. Bei, Z.-H. Zhou, K. Takaki and Y. Fujiwara, SmI2 Promoted Coupling Reaction of Isocyanates to Oxamides, Chem. Lett., 1992, 21, 1143–1144.
23. Z. J. Witezak, Desulfurization of glycosyl isothiocyanates with tributyltin hybride, Tetrahedron Lett., 1986, 27, 155–158.
24. T. R. Stancil, V. A. Fortney, J. K. Murphy, M. Gembicky, A. L. Rheingold and C. S. Weinert, Desulfurization of isothiocyanates by a divalent germanium amide, Dalton Trans., 2025, 54, 70–73.
25. T. Mukaiyama, T. Fujisawa, Y. Tamura and Y. Yokota, 1,3-Addition-Type Ring-Opening Polymerization of Cyclic Phosphonites, J. Org. Chem., 1964, 29, 2572–2575.
26. E. J. Corey and R. A. E. Winter, A New, Stereospecific Olefin Synthesis from 1,2-Diols, J. Am. Chem. Soc., 1963, 85, 2677–2678.
27. D. K. Verma, Y. Dewangan and C. Verma, in Handbook of Organic Name Reactions, ed. D. K. Verma, Y. Dewangan and C. Verma, Elsevier, 2023, pp. 243–267.
28. Z. Wang, in Comprehensive Organic Name Reactions and Reagents, 2010, pp. 746–749.
29. J. J. Li, in Name Reactions: A Collection of Detailed Mechanisms and Synthetic Applications Fifth Edition, ed. J. J. Li, Springer International Publishing, Cham, 2014, pp. 182–184.
30. E. Block, in Organic Reactions, 2005, pp. 457–566.
31. T. K. M. Shing and E. K. W. Tam, (-)-Quinic acid in organic synthesis. 8. Enantiospecific syntheses of (+)-crotepoxide, (+)-boesenoxide, (+)-β-senepoxide, (+)-pipoxide acetate, (-)-crotepoxide, (-)-senepoxide, and (-)-tingtanoxide from (-)-quinic acid, J. Org. Chem., 1998, 63, 1547–1554.
32. J. H. Rigby, B. Bazin, J. H. Meyer and F. Mohammadi, Synthetic studies on the ingenane diterpenes. An improved entry into a intrabridgehead system, Org. Lett., 2002, 4, 799–801.
33. E. J. Corey and G. Märkl, Generation of phosphite ylides from trithiocarbonates and trimethyl phosphite and their application to the extension of carbon chains, Tetrahedron Lett., 1967, 8, 3201–3204.
34. D. Horton and C. G. Tindall, Jr., Synthesis and reactions of unsaturated sugars. XI. Evidence for a carbenoid intermediate in the Corey-Winter alkene synthesis, J. Org. Chem., 1970, 35, 3558–3559.
35. A. Ghosh and S. Berg, in Arrow Pushing in Inorganic Chemistry, John Wiley & Sons Inc., 2014, pp. 156–201.
36. L. Kurti and B. Czako, Strategic Applications of Named Reactions in Organic Synthesis, Elsevier Academic Press, Burlington, MA 01803, USA, 2005.
37. P. Zhang and Z.-X. Yu, Involving Carbene or Not? Mechanism of Corey–Winter Reaction, J. Am. Chem. Soc., 2025, 147, 13915–13927.
38. J. S. Merola and A. W. Grieb, Crystal structure of chlorido([eta]2-phenyl isothiocyanate-[kappa]2C,S)-mer-tris(trimethylphosphane-[kappa]P)iridium(I), Acta Crystallogr., 2014, 70, 352–354.
39. S. Gambarotta, M. L. Fiallo, C. Floriani, A. Chiesi-Villa and C. Guastini, Reactivity of decamethylvanadocene with phenyl isothiocyanate and carbonyl sulfide: reactions related to the desulfurization of the thiocarbonyl group, Inorg. Chem., 1984, 23, 3532–3537.
40. E. J. Corey and P. B. Hopkins, Seminal-Paper: A mild procedure for the conversion of 1,2-diols to olefins, Tetrahedron Lett., 1982, 23, 1979–1982.
41. C. Gonzalez and H. B. Schlegel, Reaction path following in mass-weighted internal coordinates, J. Phys. Chem., 1990, 94, 5523–5527.
42. L. Deng and T. Ziegler, The determination of intrinsic reaction coordinates by density functional theory, Int. J. Quantum Chem., 1994, 52, 731–765.
43. H. P. Hratchian and H. B. Schlegel, Accurate reaction paths using a Hessian based predictor–corrector integrator, J. Chem. Phys., 2004, 120, 9918–9924.
44. H. P. Hratchian and H. B. Schlegel, Using Hessian Updating To Increase the Efficiency of a Hessian Based Predictor-Corrector Reaction Path Following Method, J. Chem. Theory Comput., 2005, 1, 61–69.
45. C. Reichardt, Solvation Effects in Organic Chemistry: A Short Historical Overview, J. Org. Chem., 2022, 87, 1616–1629.
46. M. Das, A. R. Gogoi and R. B. Sunoj, Molecular Insights on Solvent Effects in Organic Reactions as Obtained through Computational Chemistry Tools, J. Org. Chem., 2022, 87, 1630–1640.
47. R. Kar and S. Pal, Effect of solvents having different dielectric constants on reactivity: A conceptual DFT approach, Int. J. Quantum Chem., 2010, 110, 1642–1647.
48. J. P. Dutasta, J. Martin and J. B. Robert, Phosphonites cycliques, oligomerisation, obtention de macrocycles phosphores. Etude R.M.N. et stereochimique, Heterocycles, 1980, 14, 1631.
49. R. Greenhal, J. E. Newbery, R. Woodcock and R. F. Hudson, Influence Of Ring Strain On Reactivity Of Tervalent Phosphorus Compounds, J. Chem. Soc. D, 1969, 22.
50. K. M. Allan, K. Kobayashi and V. H. Rawal, A Unified Route to the Welwitindolinone Alkaloids: Total Syntheses of (-)-N-Methylwelwitindolinone C Isothiocyanate, (-)-N-Methylwelwitindolinone C Isonitrile, and (-)-3-Hydroxy-N-methylwelwitindolinone C Isothiocyanate, J. Am. Chem. Soc., 2012, 134, 1392–1395.
51. B. Maeda and K. Murakami, Recent advancement in the synthesis of isothiocyanates, Chem. Commun., 2024, 60, 2839–2864.
52. Z. Fu, W. Yuan, N. Chen, Z. Yang and J. Xu, Na2S2O8-mediated efficient synthesis of isothiocyanates from primary amines in water, Green Chem., 2018, 20, 4484–4491.
53. Ľ. Drobnica, P. Kristián and J. Augustín, in Cyanates and Their Thio Derivatives (1977), 1977, pp. 1003–1221.
54. S. Ozaki, Recent Advances In Isocyanate Chemistry, Chem. Rev., 1972, 72, 457–496.
55. R. Neidlein, Heterocyclic isonitriles, Arch. Pharm. Ber. Dtsch. Pharm. Ges., 1966, 299, 603.
56. D. H. R. Barton, N. Ozbalik and B. Vacher, The invention of radical reactions part XVIII. A convenient solution to the 1-carbon problem (R-CO2H → R-13CO2H), Tetrahedron, 1988, 44, 3501–3512.
57. N. Taniguchi, Zinc-Catalyzed Markovnikov-Type Hydroisothiocyanation of Alkenes with Ammonium Thiocyanate, Synlett, 2022, 73–76.
58. J. M. G. Fernandez and C. O. Mellet, The Thiocarbonyl Group in Carbohydrate Chemistry, Sulfur Rep., 1996, 19, 61–159.
59. S. G. Shin, O. Hyun-Woo, K. B. Chang, O. Jaeseung, L. S. In, Y. S. Pyo, B. H. Hee, P. J. Yoon, H. J. Woo and C. H. Lee, Synthesis of multifunctionalized allyl isothiocyanates via isomerization of Morita-Baylis-Hillman adduct-derived allylic thiocyanates, Synth. Commun., 2025, 55, 251–258.
60. I. U. Hoque, A. Samanta, S. Pramanik, S. R. Chowdhury, R. Lo and S. Maity, Photocascade chemoselective controlling of ambident thio(seleno)cyanates with alkenes via catalyst modulation, Nat. Commun., 2024, 15, 5739.
61. R. P. Verma, Synthesis and reactions of 3-oxobutyl isothiocyanate (OB ITC), Eur. J. Org. Chem., 2003, 415–420.
62. A. A. Fesenko, P. A. Solovyev and A. D. Shutalev, Practical synthesis of β-isothiocyanato ketones from chalcones, Synth. Commun., 2016, 46, 678–684.
63. Y. Li, D. M. Castañeda-Bagatella, D. Kakkad, Y. Ai, H. Chen and P. A. Champagne, Synthetic and mechanistic study on the conjugate isothiocyanation of enones with trimethylsilyl isothiocyanate, Org. Biomol. Chem., 2023, 21, 9583–9590.
64. W. R. Diveley, G. A. Buntin and A. D. Lohr, Cis-Exo Addition of Isothiocyanic Acid to Norbornenes. Synthesis and Isomeric Configuration of Herbicide Norea, J. Org. Chem., 1969, 34, 616–624.
65. K. Okoshi, K. Nagai, T. Kajitani, S. I. Sakurai and E. Yashima, Anomalous Stiff Backbones of Helical Poly(phenyl isocyanide) Derivatives, Macromolecules, 2008, 41, 7752–7754.
66. M. Reller, S. Wesp, M. R. M. Koos, M. Reggelin and B. Luy, Biphasic Liquid Crystal and the Simultaneous Measurement of Isotropic and Anisotropic Parameters by Spatially Resolved NMR Spectroscopy, Chem. – Eur. J., 2017, 23, 13351–13359.
67. S. Wesp, K. Wolf, S. Immel and M. Reggelin, Poly(arylisocyanides) as Versatile, Enantiodiscriminating Alignment Media for Small Molecules, ChemPlusChem, 2022, 87, e202100507.
68. O. J. Scherer, R. Anselmann, R. T. Paine and S. Karthikeyan, in Inorg. Synth, 1989, pp. 7–12.
69. M. S. Balakrishna, Cyclodiphosphazanes with functionalities: Synthesis, reactivity and transition metal chemistry, J. Organomet. Chem., 2010, 695, 925–936.
70. J. Zhu, X. Wu and S. J. Danishefsky, On the preparation of enantiomerically pure isonitriles from amino acid esters and peptides, Tetrahedron Lett., 2009, 50, 577–579.

---