[Enthalpy]

Hello dear friends!

I'd like to propose you a way to couple fragments of two organic molecules using (UV) light. Maybe good and new, maybe old, maybe bad - here it is anyway.

On the appended scheme, the reactant R₁X is halogenated (where X can be OH as well) while R₂H has some hydrogen.

• (UV) Light abstracts the X from R₁ while the other species absorb little.
• The remaining R₁° doesn't react much with the other R₁X and R₂H because it has the same binding energy with X and H as them. Among hydrocarbons, differences are small.
• The X° flies away and finds some molecule, sometimes an R₁X, more probably an R₂H if they're more abundent. This step 2 is as in free-radical halogenation, and helped by heat if needed, it abstract an H to form HX and leaves R₂°.
• The R₁° and R₂° pair has few chances to meet. Many pairs mix up and recombine to R₁R₂, R₁R₁ and R₂R₂ - and even R₁R₁R₂ sometimes.
• HX won't react with the other molecules nor the radicals because the binding energy to its H exceeds the one to C.

This coupling has the same limit as Wurtz. But for the products I seek (mixed oligomers of cyclobutane, spiropentane...) it's no drawback, and it avoids reactive metal. A competitor reaction would let a catalyst create HX and R₁R₂ locally.

On the appended spectra compiled from MPI's data in Mainz:
http://satellite.mpic.de/spectral_atlas (thank you!)

• Alkanes are nearly identical. Cyclopropane rings shift the absorption edge. The good Xe₂ fit few hydrocarbons.
• Double bonds compete with halogens. Local substitutions influence much, conjugated bonds even more.

Comparing the halogen radicals and assimilated, based on free radical halogenation:

• Fluorine binds too strongly with carbon and isn't useable.
• Can hydrogen rather than a hydrocarbyl be abstracted from a hydrocarbon? I've no data about the action of Kr₂ light on hydrocarbons. Cyclohexadiene, amimes, toluene...? A hydrogen atom would efficiently abstract an other hydrogen or a fluorine at step 2.
• Chlorine is efficient at step 2 but little selective. Needs light from Xe₂ (impossible with most hydrocarbons) or the less good ArF lamps. Alkenes look difficult.
• Hydroxyl improves over chlorine. Good absorption of light from Xe₂ and ArF (exceeds bromides) enables most alkanes. It must be less selective than chlorine at the target, but faster.
• Bromine enables most alkanes with light from Xe₂, ArF and the good KrCl, maybe lp-Hg. The two latter lamps enable some alkenes. Hydrogen abstration at step 2 needs heat, time, and selects favourable targets.
• Iodine doesn't abstract hydrogen at step 2 hence doesn't suit this reaction.
• Other groups may be useful, like NH₂ or, for other purposes, NO₂.

Spectra for individual halides are to come.
Marc Schaefer, aka Enthalpy

[discodermolide]

Just to clarify for me: You are talking about C-C bond formation via radical reactions?
Is that a C-H activation followed by a C-C bond formation you are looking for?

[Enthalpy]

Just to clarify for me: You are talking about C-C bond formation via radical reactions?
Is that a C-H activation followed by a C-C bond formation you are looking for?

Yes... I want to make a C-C bond between R₁ and R₂. At one substrate, light rips a halogen, at the other, the halogen atom rips a hydrogen, so that both R₁ and R₂ are radicals until they combine.

[Enthalpy]

From the spectra of chlorides here under, as well as from
http://www.chemicalforums.com/index.php?topic=77307.msg283221#msg283221 :

Unstrained alkanes, even if bigger, absorb Xe₂ light less than cyclopropane does and a bit less than alkyl chlorides. Primary chlorides absorb little more, tertiary supposedly less and cyclobutanes too. The 172nm selectivity of alkyl chlorides versus alkanes often won't suffice, imposing ArF lamps, less efficient and powerful.

Vicinal chlorines interact little, but geminal ones increase much the absorption, more so for ArF light.

Alkyl chlorides would absorb more strongly than alkenes, but I've no data on the photon's effect.

The photochlorination of alkanes can happen in the same reactor, from ArF light too since it needs fewer photons, or a different light source like mp-Hg or GaN or AlGaN Led. These Led are already good for 365nm-405nm in 2015 but still not for shorter waves.

HCl competes with the alkyl chloride. It should be removed if willing to stick to the previously described scheme.

----------

Hydroxyl binds less strongly with C but more strongly with H than chlorine does. From the joined spectra :

Xe₂ becomes possible. Still marginal for cyclopropanes, clearer for the others.

ArF light targets alcohols selectively. Even the byproduced water is little worry.

Abstraction of H at step 2 could happen at room temperature, with some luck.

I suggested hydrogen peroxide might put an alcohol on restive substrates
http://www.chemicalforums.com/index.php?topic=80085.msg296423#msg296423
and if it works, it could happen in the same reactor as the coupling, and at the same time if introducing the peroxide slowly.

Hydroxyl looks advantageous over chloride.

----------

Bromine takes heat and time for step 2, but it brings the position selectivity known from free-radical halogenation, and it absorbs light better, as the appended spectra show.

Bromides make Xe₂ lamps easier.

As well KrCl lamps are strong and efficient, and bromides absorb their light, making ArF unattractive here. Even some alkenes don't absorb the wavelength, but how will a Br atom react at the alkene? Maybe a way to conduct R₁Br addition on R₂.

Even lp-Hg could be possible.

Primary, secondary and tertiary bromides absorb nearly as much, but vicinal bromides more than double the absoprtion, and geminal ones gain much sensitivity, even at lp-Hg wavelength.
http://www.chemicalforums.com/index.php?topic=77307.msg283221#msg283221

The byproduced HBr should be removed in this scheme.

Photobromination from Br₂ in the same reactor can use the same lamp. HOBr is known for restive substrates.

Marc Schaefer, aka Enthalpy

[Enthalpy]

In a different scheme, light breaks an auxiliary molecule H-X but not the reactants R₁-H and H-R₂ which can be alkanes.

As on the appended sketch:

• The photon creates the radicals H° and °X.
• Each radical abstracts a hydrogen atom from a reactant.
  
  • X is chosen efficient (there are few): Cl, or stronger OH.
  • H (436kJ/mol H-H) is slightly more efficient than Cl (432kJ/mol H-Cl) anyway.
• The resulting radicals R₁° and °R₂ err until they meet.
  
  • R₁-R₂ is the desired product.
  • R₁-R₁ and R₂-R₂ are discarded if unwanted.
  • HX serves again, H-H is discarded.

Differing from free-radical halogenation, the hydrocarbyls R₁° and °R₂ seldom split H-X's stronger bond:
C-H 404...423kJ/mol (except methane), H-H 436kJ/mol, H-Cl 432kJ/mol, H-OH 499kJ/mol
http://www.chemicalforums.com/index.php?topic=80085.msg296423#msg296423
and even when this happens, R₁-H is the regenerated reactant while X-R₂ will be broken by light.

Because no abundent X-X is available, maybe the reactants and their radicals have time to exchange hydrogen atoms, giving some position selectivity despite H° and °X being chosen efficient.

Hydrogen atoms would split any F, Cl, OH from the reactants - but maybe amines survive. The auxiliary molecule must be tough too: as a counter-example, H-CF3 would lose F.

The reactor needs moderate heat, as for free-radical chlorination. The R₁-R₂ and like must be removed early if bigger products aren't desired. This can need a simultaneous distillation outside the reacting region and a diffusion filter to remove hydrogen.

ArF and maybe Xe₂ lamps fit H-Cl, Xe₂ and maybe ArF fit H-OH.

----------

As an example, both feeds R₁-H and H-R₂ could be cyclopropane, spiropentane..., the auxiliary molecule HCl, the lamp an ArF exciplex, and if nothing goes wrong, the product a mixture of tri- and tetra-cyclopropyl similar to Syntin.

Or both feeds could be cyclobutane, the auxiliary molecule water, the lamp a xenon excimer, and the product a mixture of di-cyclobutyl (Boctane) and varied tri-cyclobutyl.

Water catalyst and a lamp, we're getting somewhere. Aladdin walking on the Moon.
Marc Schaefer, aka Enthalpy

[Enthalpy]

Photocoupling seems possible (but interesting?) for organometallics as well. I take silanes as an example and didn't look at aluminium nor boron compounds.

The weaker Si-H bond (335kJ/mol) lets target it without damaging organic branches already present. If mere comparison of bond strengths applies, OH or H or Cl or CH₃, Br and I would be as active and selective at silanes as F, Cl and Br are at hydrocarbons. As well, Si-Me bonds absorb waves longer than alkanes do, suggesting more paths.

----------

The previous second scheme would achieve Si backbones from SiH₄. While light absorbed by HOH and HCl is deleterious to silanes, KrCl (222nm) and lp-Hg (254nm) light can atomize HBr and HI (similar to MeBr and MeI on the appended spectra) to create very reactive H° and Br° and less reactive I°. These radicals abstract H from silane and already formed oligomers, which then form bigger oligomers among them, or form Si-H back (recycled), or form Si-X (split again by the same light). Hopefully a path to di- and trisilane, and if separating disilane after a first step, n-tetrasilane. Higher silanes would be isomer mixes: even if I° shows selectivity, H° won't.

----------

As in the previous first scheme, light can separate a halogen atom from a silane, possibly substituted. The halogen targets a hydrogen from an other possibly substituted silane; Br° and I° would abstract H from Si but generally not from C. No reactive H° is produced, as an advantage over splitting auxiliary HX by light. Then the radicals recombine to a longer backbone, but randomly, which can be a drawback if they differ.

Examples:

• From MeSi(Br)(H)Me, light abstracts Br° which abstracts H only from Si to make straight oligomers of dimethylsilane.
• From Me₃Si-Br and Me₃Si-H to hexamethyldisilane.
• From I-C(R)(R')Si(R")(R'")-H, get an SiC-backbone oligomer. Though, it won't nicely alternate Si and C.

An X split from R can also target H-Si from some possibly substituted silane, but together with R-Si, one would also get R-R and Si-Si, not always wanted.

Maybe heat and catalysts do it better than light.

---------

Substituted silanes absorb longer waves than alkanes do, which supposedly breaks the Si-C. Light from Xe₂ is efficient at di-, tri- and tetramethylsilane, suggesting to create substituted Si backbones from them. For instance, irradiation of tetramethylsilane would give permethylated polysilanes. Some substitutes (CHO?) must leave more easily for refined routes.

I know, chemistry is not paperwork. Other technologies neither, by the way.
Marc Schaefer, aka Enthalpy

[Enthalpy]

Some radicals bind too weakly with H to abstract it from a C, but in a third scheme sketched here under, light couples fragments by abstracting a radical at each fragment.

Not very subtle: if the fragments differ, all combinations are produced. It also takes much light: at least one photon at each fragment, and every second time the fragment recombines with an abstracted radical. But these mild radicals accept longer wavelengths that spare fragile compounds and accept easier lamps.

Bromine is such a fragment, with little heat. Abstracted with unity quantum efficiency by the good KrCl lamps, and even by cheap lp-Hg which spares alkenes.

Chlorine may leave C-H intact under special conditions but needs hard light anyway. Hydroxyl isn't credible in this role.

Iodine fits lp-Hg well, and AlGaN Led in the future, with 3/4 of the 254nm photons abstracting an I. That wavelength spares nitro and carbonyl groups, and most amines and alkenes; less efficient XeCl and XeBr light spares dienes, heavily substituted alkenes and tertiary amines, pairs and even triplets of Br on one C.

Carbonyl is long known for abstraction by light. From aldehydes, XeBr and also XeCl light efficiently abstracts °CHO while lp-Hg favours CO leaving C-H - after a delay? Ketones break often both bonds -CO- letting both sides recombine, but can break just one bond, depending on unclear conditions.

Nitro groups absorb much light up to KrCl's 222nm, less for lp-Hg, XeBr and XeCl. While C-NO₂ is a weak bond, I've no photolysis data.

Amines absorb longer waves than alkanes do, up to KrCl or even lp-Hg for tertiary amines - but which bond breaks if any?

Many more groups are candidates: silyls, SO₂...

Marc Schaefer, aka Enthalpy

[Enthalpy]

Some schemes I suggested for silanes in Reply #5 have weaknesses.

In the second section, I hoped silane radicals would recombine among themselves, but the encounter of two radicals is rare. A radical will more probably encounter a molecule like HBr and break it to form Si-H or Si-Br.

In the third section, I didn't consider the necessary wavelength to break Si-X because I have no data. If the photon energy varies like the bond dissociation energy (no strong reason for it), then 254nm light that is moderately efficient at breaking Br-tBut translates into 199nm to break Br-SiMe₃. Such a wavelength may be harmful to the backbone of bigger silanes.