溶劑化電子的傳奇歷史

貝洛尼·杰奎琳

（巴黎薩克雷大學(xué)）

1 Alkaline metal solutions in liquid ammonia

In the mid-19th century,liquid ammonia was found to be able to dissolve without reaction,unlike water,sodium and other alkaline metals in intense blue solutions that were widely used,notably by Weyl[1],to effect previously difficult chemical reductions of organic or inorganic molecules[2].However,the handling of this solvent is very complex because it is liquid only below-33°C at normal pressure.It must therefore be kept cold,but also sheltered from the atmosphere,otherwise sodium would be oxidized to the air and cold ammonia would condense water vapours.In addition,metal solutions are metastable and,even very pure,they decompose very slowly into hydrogen and amide.Despite this,Kraus,assistant to A.A.Noyes at the Massachusetts Institute of Technology,decided in1904for his thesis subject to measure the conductivity of various ions in salt solutions in liquid NH3(Fig.1)[3].

Fig.1 C.A.Kraus vacuum equipment to measure the conductivity of alkaline metal solutions in low-temperature liquid ammonia[3].

While he had already determined the conductivity of each ion of alkaline salts,Kraus also wanted to compare these results with measurements in alkaline metal solutions and,to his surprise,their conductivity was far superior to those of salts.Stranger also,conductivity did not depend on the nature of the metal.He therefore concluded in1908[4]first that alkaline metals were dissociated in ammonia into ions,and then that the common complementary anion of these solutions,responsible for most of their conductivity and blue color,could only be a solvated electron:"the anion is an electron surrounded by solvent molecules"..."a solvated electron"...

This audacious concept of a solvent electron,similar,despite its very different nature,to a monovalent anion stabilized by the polarization of solvent molecules,immediately prompted a great deal of work[2].Gibson and Argo established optical absorption spectra for the solvated electron in diluted alkaline and alkaline-earthous metals solutions in ammonia and methylamine[5-6]. They were characterized by a wide,intense and asymmetrical band in energy,with a maximum in the near IR.Later,it was shown that at lower temperature or higher pressure these spectra were moved to the visible[2],as were the halide absorption spectra.Like them,they have been attributed to a transfer of the charge to the molecules of the solvation layer(or CTTS,charge transfer to solvent).Early on,the expansion of the volume of the solution observed when the metal dissolved in the ammonia had suggested that the electron occupied a cavity much larger than its own size.This result inspired the early theoretical models of the structure of the solvated electron described by Ogg[7]and then by Jortner[8].The spectrum was attributed to an excitation of the1s→2p transition of a particle,in a cavity created by the mutual repulsion of hydrogen atoms of the polar solvent molecules oriented towards the electron.

2 Ionizing radiation

R?ntgen's[9]discovery in December1895of an invisible and very penetrating X-rays,generated by the impact of cathode rays on the anticathode of a Crookes tube,and then by H.Becquerel of the uranic rays in1896,suddenly extended the fields of photophysics and photochemistry to much more energetic radiation.Pierre and Marie Curie discovered in the pechblende in1898two new elements much more radioactive than uranium,polonium[10]and radium[11],and emitting specific radiation(Fig.2).

Fig.2 Image of a wallet on a photographic plaque in Marie Curie's thesis[12].The radiation used was γ rays emitted by a source of radium contained in glass(the radiation β was deflected by a magnetic field).The mode of penetration is very similar to that of X-rays.

The following year,they published the physicalchemical effects of these radiations on matter[13],in particular,in addition to the ionization of gases,the production of ozone from oxygen,and the production of molecular hydrogen and hydrogen peroxide in radioactive water solutions.The initial ions of the water are assumed to be H2O+and H2O-which,after a fast reaction with H2O,give birth respectively to the radicals OH·and H·[14].On the basis of this radical model,the radiolytic yields of products formed in various solutions can be gradually explained[15-18].

3 The hydrated electron

However,in the1950s,some results were beyond this[19-20].In particular in1952,G.Stein of the Hebrew University of Jerusalem found very different results for the reduction of aqueous methylene blue solutions in acidic or neutral environments[21].He concludes that in addition to the radical H-,a second reductive species should be considered.He was the first to propose the hydrated electron eaq-as an analog of the electron ammonized eam-,known in solutions of alkaline metals in liquid ammonia.The hydrated electron would come from the electron originally derived from the ionization of the water.Immediately,the theorist R.L.Platzman[21]describes in his model how this electron can escape recombination with the parent cation and thus solve the solvent.

Despite this,the hypothesis of a hydrated electron will take a decade to impose itself.The first argument against it was the stability of eam-while eaqwould be a very short-lived intermediary.Nor could any metastable blue color of the ammoniated electron be observed after liquid ammonia irradiation by a continuous accelerator[22].Prudently,the authors concluded that his concentration may have been too low.Another difficulty was that the aqueous solutions studied in radiolysis,even deaerated,were often very acidic and in this case the hydrated electron,reacting with a proton,is actually replaced by a radical H·.In addition,even in a neutral environment,many solutes are reduced indiscriminately by H·or eaq-by giving the same products.

On the other hand,G.Stein and J.Jortner observe a fleeting blue color by putting water or alcohol in contact with sodium under argon atmosphere,suggesting the existence,at least transient,of a solvated electron also in these liquids[23].In1958,chloroacetic acid radiolysis produced hydrogen in an acidic environment but chloride ions in a neutral environment[24].In1961,studies of the ionic force on the radiolysis of cationic or anionic solutes show that the predominant radical species in neutral environment carries a negative charge[25].

However,in a1961review of water radiolysis by Hart and Platzman,the hydrated electron is not even mentioned.The resolution of this controversy is also not the objective of the early experiments to detect reactional intermediaries by the pulsed radiolysis method developed by Matheson and Dorfman who observe the free radical I2[26].

Similarly,J.W.Boag[27],in his relationship of historical observations of eaq-using the pulsed radiolysis installation he had just developed at the Gray Laboratory in London,reports that E.J.Hart had in fact come to join him in order to detect free carbonate radicals.But two distinct absorption maxima are observed,one of which is found in pure water,and is moved to the infrared when the water is added with ammonia.The spectrum is therefore assigned to eaq-.The key result of this highlight was communicated to the2nd International Congress of Radiation Research in Harrogate in1962.At the same conference,J.P.Keene also reports on his observation,independently,of the new spectrum by using the pulse radiolysis installation he built at the Patterson Laboratory in Manchester,spectrum which he also assigned to the hydrated electron.Boag proposes to publish their respective results simultaneously in the same issue of the journal Nature[28-30].In the same year,J.Jortner,Ottolenghi and G.Stein observe by flash photolysis the same spectrum of eaq-produced by photodetachment from the iodide anion[31].

Fig.3 Optical absorption spectrum of hydrated electron observedforthefirsttimebypulseradiolysisinasol utionof sodiumcarbonate(withthepeaksoftheradicalCO3-·at600 nm andeaq-at700nm)(a),andin pure water(witheaq- )(b),bothdeaerated[31].Todetectananomolarconcentration of eaq-,theoptical pathwas elongatedbyplacingthe4cm optical cellatthecenterofa multiple reflectionsystem.Atthis concentration,the decay ofeaq-byrecombinationsis veryslowandis observedformorethan50ms.

4 The solvated electrons

These results had a great impact and triggered a proliferation of research.A Weyl Symposium was created,with its first edition in1963[32],to commemorate the centenary since Weyl's discovery of metal solutions in ammonia.In addition to the recent works by specialists in the field,Jortner[32]presents a review on the evolution of theories on the structure of the solvated electron that preceded its semicontinuous model,which he applied not only to eam-but also for the first time to eaq-which had just been highlighted.This structure consisted of a cavity surrounded by a few polarized solvent molecules and then a continuum of solvent,characterized in particular by its density and dielectric constants,optical and static.The model accounts for the wavelengths of the respective maxima according to temperature and pressure.Very quickly,thanks to newly installed pulse radiolysis facilities and shorter pulses,a very wide variety of liquids,more or less polar,were ionized and solvated electrons detected.Despite their instability,the very similar properties of their transient absorption spectra were compared with fast IR detections in alcohols and polyols,amines,ethers,sulphides,carbonates,ionic liquids and even hydrocarbons[33-36](Fig.4),or their reactivity with series of molecules.

Fig.4 Optical absorption spectra of electrons solvated in different liquids,calibrated into molar absorption coefficients.GLY:glycol;PD:propane diol;EG:ethylene glycol;MeOH:methanol;EtOH:ethanol;PrOH:propanol;EDA:ethane-1.2-diamine;DEA:diehanolamine;13PDA:propane-1.3-diamine;THF:tetrahydrofurane;DME:dimethylether;DEE:diethylether;Diglyme:bis(2-methoxyethyl)ether;R4NNTf2:methyl-tributyl-ammonium bis[trifluoromethyl-sulfonyl]imide;DEC:diethylcarbonate.(Adapted from[33-36]).

Hereafter,the following editions of the Weyl Symposiums[36]have brought together experimentalists and theorists discussing solvated electrons produced either by pulsed radiolysis or by dissolution of alkaline and alkaline-earth metals in liquid ammonia,amines and certain ethers.One of the interesting aspects of the solvated electron is to appear thus as the smallest solvated chemical species which,despite its transient nature,can be formed and studied in the greatest number of environments.It is also compared with electrons trapped in irradiated ices[37].

Thanks to the developments of the simulation,numerous theoretical models have been proposed to account for the properties of the solvated electron,including the hydrated electron,and to describe its Structure(Fig.5)[38].

Fig.5 Structure of the hydrated electron simulated by molecular dynamics[38].The charge of the electron is delocalized in the2.5A radius cavity.H2O molecules are oriented towards the central charge by one of their H atoms.

The pertubations of water molecules in the solvation layer are sometimes calculated[39].A recent review analyzed these models in detail.The simulated optical absorption spectrum that is closest to the experiment is composed of transitions between the fundamental state and about fifteen excited states(Fig.6)[40].

Fig.6 Comparison between the optical absorption spectra of the hydrated electron obtained by the experiment or based on simulations based on the theory of the functional of density with3or15states[40].

5 Box1-Instability of esolv-

How to explain the large difference between the eammetastability in blue metal solutions in ammonia or amines,and the rapid decay of solvated electrons produced by pulse radiolysis,including in NH3(which explains why the1953experiments failed to observe any blue color[24])?In fact,the solvated electron is in a fundamental state,which would be stable if its existence were not limited by its very high chemical reactivity that depends on the conditions(and improperly characterized by a lifetime).In NH3and amines,it is inert against alkaline or alkaline-earth cations.It therefore decreases only very slowly with impurities or acidic ions slowly released from the walls of the container(the ionization product of NH3itself is only Kion-10-23M2).While in radiolysis,eam-reacts with the NH4+cations and other radical species formed simultaneously.In addition,it was shown by pulse radiolysis that in water the dismutation reaction between two eaq-was very fast in producing dihydrogen and OH-.It also occurs during the contact between water and sodium metal and produces hydrogen explosions with oxygen from the air,well known to apprentice chemists.

But,provided that liquid ammonia is added with hydrogen to capture oxidizing radicals,and Alkaline amidurn NaNH2to capture NH4+ions,theradiationinduced electron eambecomes as stable as in metal solutions.Thanks to this inertia,the characteristic blue color was observed in this case after a simple stationary irradiation(Fig.7)[41],and also after photolysis and photodetachment from NHWhatever the mode of formation,the identity between the solvated electrons was thus definitively established.Blue eamsolutions were also obtained in NH3by electrochemistry(with alkaline halide as the electrolyte)[2,42]or by photo-injection from a semiconductor[43].

Fig.7 Optical absorption spectrum of a hydrogen and sodium amide solution in liquid ammonia irradiated at20°C(spectrum is recorded in a sealed test tube,resistant to91.19kPa[41]).

A.J.Swallow［44］even calculated that，given the formation of eaq-by irradiation by cosmic rays and by photodetchment from chlorides by the Sun's UV rays，its near-stationary concentration in the oceans would be about10-9molar（"Many of the short-lived chemical species......are known to exist naturally......"）（Fig.8）.

Fig.8 Formation of eaq-in the oceans[44].

As early as1971,Baxendale and Wardman[45]observed for the first time after a5ns pulse the solvation dynamics of an electron in the viscousnpropanol at low temperature,which slowed down molecular movements(Fig.9).The initial spectrum is located in the infrared and the absorption at1300nm then disappears in1μs while simultaneously the spectrum of the completely solvated electron develops and stabilizes with its maximum at500nm.The first spectrum is assigned to a pre-solvated electron and the solvation time is estimated atτ1/2=50ns(Fig.9,insets).

Fig.9 Solvation of es-in the n-propanol at152K after ionization by a5ns pulse of a pulse accelerator[45].Insets:pre-solvated electron decay at1300nm and growth of es-at500nm.

It will be necessary to wait for the pulses of100 femtoseconds offered by the lasers with a biphotonic UV excitation for the electron solvation,much faster at room temperature,can be observed in water(τ1/2=200fs)[46](Fig.10)or polyols(τ1/2=10ps for glycerol)[47].In liquid ammonia at-50°C,solvated electrons in blue solutions were excited in the IR band at1280 nm to eject them from their solvent cavity and the relaxation time to the solvent state wasτ1/2=150fs[48],which,given the temperature,is even faster than in the water.On the other hand,the solvation relaxation of an electron in the ionic liquid[C4mpyr][NTf2],with larger molecules and heterogeneous density domains,is distinguished by two different velocity processes(τ1/2=70.4ps andτ1/2=574ps)[49].

Fig.10 Solvation of eaq-after ionization of water molecules at 294K by a biphotonic UV excitation by a laser pulse of100fs(τ1/2=200fs)[46].

Gradually,advances in pulse radiolysis of water solutions have helped to determine the UV optical absorption spectra of H-and OH-radicals.But after subtracting these bands from the total spectrum in pure water,it remained another component,correlated with the infrared absorption band of eaq-,and shifted like it between20and34℃to the larger wavelengths,but much less intense(Fig.11,inset)[50].With this precise deconvolution,the authors concluded that the UV component must be assigned to eaq-and that it corresponds to the edge of the absorption band,shifted to the visible,of water molecules from the solvation layer,disturbed by the transfer of the electron charge.This would therefore be the first observation of a spectrum of a solvation layer.

Given the very comparable properties of all solvated electrons,the UV absorption spectra of very pure sodium solutions in ammonia were in turn examined at-50℃and20℃[51].In this case,the solution contains only eam-and Na+.The spectra display indeed UV absorption bands proportional to the concentration of the metal and also correlated with the IR bands of eam-(Fig.11).They are shifted from the band edge of pure ammonia,as well as those of eaq-at20℃and34℃compared to the pure water band edges(Fig.11,inset).These elements therefore support the previous interpretation of a spectrum due to solvation molecules,which,by analogy,should also appear in the spectra of all other solvated electrons.More generally,the'solvation'absorption band should exist for ions whose spectrum is also due to a charge transfer to solvent,such as halides,but in this case it would be masked by the main UV spectrum.

Fig.11 UV-visible optical absorption spectra of the ammoniated electron at-50℃and20°C.Inset:comparison between the UV optical absorption bands of eam-at-50℃and 20°C and eaq-at20°C and the NH3and H2O solvents;scales multiplied by4;adapted from[50]and[51].

6 Conclusion

A consensus seems to emerge that the solvated electron occupies a cavity of size and shape varying with temperature and pressure,that it attracts towards its negative charge the H atoms of one of the polar groups of the solvation molecules whose structure is perturbed,and that its optical spectrum,very dissymmetrical in energy,results from a transition between a fundamental state and several linked excited states.Actually,the solvated electron remains a particularly interesting probe of solvation in all kinds of environments,and therefore of their structure,and this strange chemical species will probably spark much more research.

(Translated with permission from L’Act.Chim.,2021,460-461,p.17-22.)