, 20110628, published 5 August 2013371 2013 Phil. Trans. R. Soc. A Ronald G. Lawler and Nicholas J. TurroJudy Y.-C. Chen, Yongjun Li, Michael Frunzi, Xuegong Lei, Yasujiro Murata, 

and other endofullerenes60O@C2, H60@C2Nuclear spin isomers of guest molecules in H

  

References28.full.html#ref-list-1http://rsta.royalsocietypublishing.org/content/371/1998/201106

This article cites 61 articles, 5 of which can be accessed free

Subject collections

(42 articles)spectroscopy   � (36 articles)physical chemistry   �

(93 articles)nanotechnology   � collectionsArticles on similar topics can be found in the following

Email alerting service herein the box at the top right-hand corner of the article or click Receive free email alerts when new articles cite this article - sign up

http://rsta.royalsocietypublishing.org/subscriptions go to: Phil. Trans. R. Soc. ATo subscribe to

on August 19, 2013rsta.royalsocietypublishing.orgDownloaded from

rsta.royalsocietypublishing.org

ResearchCite this article: Chen JY-C, Li Y, Frunzi M,Lei X, Murata Y, Lawler RG, Turro NJ. 2013Nuclear spin isomers of guest molecules inH2@C60, H2O@C60 and other endofullerenes.Phil Trans R Soc A 371: 20110628.http://dx.doi.org/10.1098/rsta.2011.0628

One contribution of 13 to a Theo MurphyMeeting Issue ‘Nanolaboratories: physics andchemistry of small-molecule endofullerenes’.

Subject Areas:physical chemistry, nanotechnology,spectroscopy

Keywords:spin isomers, endofullerenes, para-hydrogen,nuclear magnetic resonance, spin catalyst

Author for correspondence:Ronald G. Lawlere-mail: ronald_lawler@brown.edu

†Deceased 24 November 2012.

Nuclear spin isomers of guestmolecules in H2@C60, H2O@C60and other endofullerenesJudy Y.-C. Chen1, Yongjun Li1, Michael Frunzi1,

Xuegong Lei1, Yasujiro Murata2, Ronald G. Lawler3

and Nicholas J. Turro1,†

1Department of Chemistry, Columbia University, New York,NY 10027, USA2Institute for Chemical Research, Kyoto University,Kyoto 611-0011, Japan3Department of Chemistry, Brown University, Providence,RI 02912, USA

Spectroscopic studies of recently synthesized endo-fullerenes, in which H2, H2O and other atoms andsmall molecules are trapped in cages of carbonatoms, have shown that although the trappedmolecules interact relatively weakly with the internalenvironment they are nevertheless susceptible toappropriately applied external perturbations. Theseproperties have been exploited to isolate and studysamples of H2 in C60 and other fullerenes that arehighly enriched in the para spin isomer. Severalstrategies for spin-isomer enrichment, potentialextensions to other endofullerenes and possibleapplications of these materials are discussed.

1. Introduction

(a) Macroscopic quantum effectsQuantum effects on the macroscopic properties ofchemical substances at ordinary temperatures arerare. A well-known, easily demonstrated exampleis the influence of deuterium substitution on theelevation of the melting point and changes in otherthermodynamic properties of water [1]. Shortly after thefirst preparation of D2O, this was attributed by Bernal &Tamm [2] to mass-dependent, zero-point energy effectson intermolecular forces. Their original idea has beenextended to more generalized treatments of quantumfluctuations [3].

2013 The Author(s) Published by the Royal Society. All rights reserved.

on August 19, 2013rsta.royalsocietypublishing.orgDownloaded from

2

rsta.royalsocietypublishing.orgPhilTransRSocA371:20110628

......................................................

A

–

H H

+

yrot

(Q,F) ynuclear spin

(1,2)

180∞21

×

–J = 1

J = 0

+

+ ++ +

S

S A

H H

Figure 1. Schematic of the effect of rotation about the H–H axis on the symmetry (A or S) of (left) the rotational wave functionand (right) the spin wave function for the lowest energy para-H2 (J = 0) and ortho-H2 (J = 1) rotational states. The resultingproduct spin–rotationwave function satisfies the antisymmetry properties required of fermions such as protons. (Online versionin colour.)

120 cm–1

para-H2J = 0

ortho-H2J = 1

NMRinactive

NMRactive

Erot

ab – ba

ab + ba +

–

ynuclear spin

bb

aa

Figure 2. Energy diagram for the two lowest spin-rotational states of H2. Para-H2 (I = 0) is 120 cm−1 (0.3 kcal mol−1; 172 K)more stable than ortho-H2 (I = 1). (Online version in colour.)

A more widely studied quantum effect, of increasing practical importance, gives rise to theexistence of isolable ortho and para allotropes of molecular hydrogen that differ only in theirnuclear spin and rotational wave functions [4,5] in order to satisfy the parity requirements offermions and bosons. (It is interesting to note that symmetry and nuclear spin effects were alsomentioned, but dismissed as unimportant, in the original explanation of Bernal & Tamm [2] for theD2O/H2O melting point difference and other isotope effects in water.) Protons (IH = 1/2) obey therule for fermions requiring that the molecular wave function be antisymmetric under exchangeof equivalent particles. This leads to the result shown in figure 1 that the lowest energy rotationalstate (J = 0) of H2 is symmetric and must be combined with an antisymmetric (I = 0; para) nuclearspin state, whereas the asymmetric rotational wave function of the first excited rotational state(J = 1) corresponds to the symmetric I = 1 (ortho) spin state. A more explicit description of thenuclear spin wave functions and the corresponding energies of the two lowest energy states ofH2 is shown in figure 2.

Transitions between the ortho and para states require highly forbidden simultaneous changes inthe nuclear and rotational quantum numbers. As a consequence, a sample of molecular hydrogenoften behaves as a mixture of two separate chemical species, each with its own set of properties.

on August 19, 2013rsta.royalsocietypublishing.orgDownloaded from

3

rsta.royalsocietypublishing.orgPhilTransRSocA371:20110628

......................................................

80

equi

libri

um%

ort

ho-H

2

catalyst

no catalyst

70

E

E

E

E

60

5040

30

20

50 100 150temperature (K)

290 K(room temperature)

77 K(liquid N2)

20 K(liquid H2)

200 250 300 350 400

10

0

Figure 3. Temperature dependence of ortho-H2 composition. In the absence of a catalyst, a largemetastable excess of the orthoformmay develop. (Online version in colour.)

At room temperature, the equilibrium ortho : para ratio is 3 : 1, reflecting the statistical ratio of theI = 1 and I = 0 nuclear spin states. At easily achievable lower temperatures, however, as shown infigure 3, the equilibrium favours the para form, as implied by the energy gap of 120 cm−1 (172 K)[5] between the two forms.

(b) Para-hydrogen enrichmentExploitation of the temperature dependence of the ortho : para ratio is the basis of the mostcommonly used methods for producing samples enriched in para-H2 [5]. For example, at 20 K,the boiling point of liquid H2 at one atmosphere, the equilibrium sample is more than 99%para-H2. Under conditions where the temperature of a sample of H2 is lowered more rapidly thanthe rate of ortho–para conversion, however, a metastable state is reached in which the ortho formis many times more abundant than the equilibrium value. To avoid the build-up of potentiallydangerous levels of latent energy, large-scale users of liquid hydrogen routinely use a catalyst toensure that equilibrium is maintained as liquefaction proceeds or as hydrogen boil off occurs [6].

(c) Enrichment of less stable spin isomers(i) Chromatography and selective adsorption

The distinctly different thermodynamic properties of the two forms [5] and their mode ofinteraction with surfaces make possible separation of the hydrogen allotropes by selectiveadsorption or chromatography [7]. The separation of the allotropes of hydrogen and deuteriumillustrated in figure 4 [8] might even be described as a striking macro-scale manifestation of thePauli exclusion principle! The parity of the deuterium nucleus (ID = 1) is that of a boson andrequires that the spin–rotation wave function be symmetric to exchange of two equivalent nuclei.As with H2, this gives rise to ortho and para states for D2, except that the lowest energy, J = 0,state corresponds to symmetric nuclear spin functions I = 0 and 2 (ortho), and the J = 1 state toI = 1 (para).

By selectively trapping the eluent from the stream of separated hydrogen allotropes, it becomespossible to isolate the metastable form in useful quantities [9]. Selective adsorption of themetastable (para) form of D2 had also been accomplished by selective adsorption on γ-aluminanear the liquefaction point [10].

Selective excitation of one spin isomer with electromagnetic radiation followed by anirreversible process is a potential method of isomer separation analogous to methods appliedfor isotope enrichment. Laser-induced drift, which relies on selective excitation of a well-resolved

on August 19, 2013rsta.royalsocietypublishing.orgDownloaded from

4

rsta.royalsocietypublishing.orgPhilTransRSocA371:20110628

......................................................

He

column: 80 m × 0.25–0.3 mm, 30 µm silica coatingoven: 77.6 K (liquid nitrogen)

carrier gas: 2 ml min–1, neon

sample: 1.5 µl gas mixture

0 9 (min)

HD

o-D2

o-H2

p-D2

p-H2

Figure 4. Separation of the isotopologues and allotropes of hydrogen and deuterium by gas chromatography. The conditionsused are as shown. Adapted fromMohnke [8].

transition of one isomer, has been used, for example, for selective enrichment of spin isomersof CH3F [11]. State-selective photolysis of CH2O has been used to selectively remove one spinisomer [12], enriching the sample in the other. It was also shown that one nearly pure allotropeof H2 could be formed by this state-selective photolysis [13]. It remains to be seen whethersuch methods can be applied in condensed medium or to encapsulated molecules in fullerenes.It is encouraging, however, that the far infrared (FIR) spectrum of H2O@C60 displays well-resolved peaks of the ortho and para isomers at cryogenic temperatures [14] that would allow forselective excitation of rotational transitions. Whether such excitation would be possible at highertemperatures is not presently known.

(d) NMR applications of para-hydrogen and other long-lived nuclear spin speciesThere has been a resurgence of interest in para-H2 in recent years following the prediction [15] andaccidental discovery [16] of strongly enhanced NMR signals in the products of hydrogenationreactions carried out using H2 enriched in that allotrope. The phenomenon is commonly referredto as para-hydrogen-induced polarization (PHIP). The same methodology has been used toproduce enhanced 2H NMR signals by using enriched ortho-D2 [17]. PHIP was first observed inorganometallic compounds and continues to be used as a tool for mechanistic studies of a varietyof hydrogenation processes [18]. The use of signal enhancement via PHIP is especially promising[19] as a complement to hyperpolarization methods for improving the sensitivity of magneticresonance imaging (MRI) and analytical NMR using optical polarization of Xe [20] and dynamicnuclear polarization [21].

Interest in nuclear spin intersystem crossing processes has also been stimulated by thediscovery that surprisingly long-lived nuclear spin singlet states may be prepared in polyatomicmolecules by proper manipulation of the spins using magnetic resonance methods [22].Possible applications of selected nuclear spin states in quantum computing have not escapednotice [23–25].

2. Spin isomers of molecules encapsulated in fullerenesThe discovery that some atoms and small molecules may be captured in fullerene cages indetectable amounts [26] or placed there synthetically in high yield, including H2 [27], H2O

on August 19, 2013rsta.royalsocietypublishing.orgDownloaded from

5

rsta.royalsocietypublishing.orgPhilTransRSocA371:20110628

......................................................

[28,29], NH3 [30] and CH4 [31], all of which have spin isomers, leads one naturally to ask if theenvironment sensed by the endo molecule resembles that of the isolated molecule in the gas phase.Based on a variety of spectroscopic studies, the emerging answer is a qualified ‘yes’ [14,32–34, andreferences therein] although some coupling of rotational and internal translation modes of theendo H2 has been predicted theoretically [35] and detected spectroscopically [32]. (It has not,however, been possible to resolve by NMR the spin–rotation fine structure in H2@C60 that isresolved in gas phase molecular beam measurements [36].) Similarly, the low-temperature FIRspectrum of H2O@C60 resembles that of water in the gas phase [14]. Magic angle spinning (MAS)NMR measurements and FIR reveal splitting of the rotational energy states of the ortho form atlow temperature indicating breaking of strictly spherical symmetry in the H2O environment [14]similar to what has been observed for isolated water molecules in rare gas matrices [37].

The observations described above have prompted us to investigate the extent to which therelative isolation of H2 and other small molecules encapsulated in the fullerenes might makeit possible to prepare, store and ultimately make use of pure samples of the encapsulatedspin allotropes in condensed phases. We report and review below recent developments in themethodology for interconverting the nuclear spin isomers of small molecules in endofullerenesand suggest some promising directions for future efforts.

3. Methodologies for endofullerene spin-isomer enrichmentOur strategy for preparing and preserving endo fullerene samples enriched in the spin isomer,such as para-H2, favoured at low temperatures has required development of the followingmethodologies [34]:

(i) An accurate and reliable analytical method is developed to determine and monitor thedegree of enrichment.

(ii) The endo fullerene is subjected to intimate contact with a catalyst at a temperaturesufficiently low, and for a sufficiently long time, to achieve the desired degree ofenrichment of the equilibrium mixture in the more stable spin isomer.

(iii) The catalytic activity is removed sufficiently rapidly to preserve the enrichment, and theenriched sample is returned to the temperature where it will be studied or used, mostcommonly room temperature.

(a) AnalysisTo determine the ortho : para ratio of H2 in fullerenes, we have exploited the fact that onlythe ortho isomer gives rise to a proton NMR signal. Fortunately, the NMR peaks from endomolecules are shifted into a unique region free from interference by other signals. It then becomespossible to monitor the change in allotrope content by determining the intensity of the ortho-H2 peak. For convenience and improved accuracy, the sample also includes a portion of the samefullerene containing hydrogen deuteride, HD, incorporated at the time of endofullerene synthesis.Because endo H2 and HD have essentially the same effects on the properties of the fullerene cage(a notable exception being the differing abilities of the isotopologues of H2@C60 to quench singletoxygen [38]), the ratio of endo H2 : endo HD should be unchanged by the chemical or physicalmanipulations needed to isolate and prepare samples for analysis. Furthermore, HD has only onespin isomer and is not subject to magnetic catalysis. Changes in the ratio of the H2 : HD NMRsignals should therefore be an accurate proxy for the change in fraction of ortho-H2 in the sample.A spectrum of H2/HD@C60 before and after enrichment in para-H2 using method (i) (describedbelow) is shown in figure 5.

The high-resolution NMR strategy outlined above to monitor changes in the nuclear spinallotrope composition should be applicable to other small molecules such as D2, H2O, NH3 andCH4. In all these cases, the intrinsic NMR intensity is different for each spin isomer so that changesin composition result in changes in intensity of the mixed sample. The use of the corresponding

on August 19, 2013rsta.royalsocietypublishing.orgDownloaded from

6

rsta.royalsocietypublishing.orgPhilTransRSocA371:20110628

......................................................

(a)

(b)

1H NMR

HD@C60 H

2@C

60

–1.55–1.35 –1.40 –1.45 –1.50 ppm

Figure 5. 500 MHz 1H NMR, room temperature, of H2/HD@C60 mixture (a) before and (b) after enrichment in the para, NMRsilent, allotrope of H2 (see text). (Online version in colour.)

partly deuterated endo molecules as intensity standards should also be applicable to other smallmolecules because of the isotope shift of the 1H resonance induced by deuterium substitution.For example, it has been shown that the 1H NMR peaks from H2O@C60 and HDO@C60 are wellresolved [29].

Another magnetic resonance method that shows some promise is to monitor relative spin-isomer concentrations through changes in the intensities of peaks within multiplets arising fromcoupling of other nuclei to the spin species of interest. It may be possible, for example, to detectchanges in the intensity of the centre peak of the 17O resonance in H17

2 O@C60 which shouldpossess ortho and para spin isomers analogous to the 16O isotopologue. Figure 6 shows a 17ONMR spectrum of H2O@C60 containing 90% H17

2 O. The centre peak that is clearly resolved fromthe two outer peaks, separated from the centre by JHO ≈ ±80 Hz, arises from overlap of signalsfrom the I = 0 (para) and I = 1 (ortho) states of the two protons.

Spin isomers may, of course, be distinguished by a number of other spectroscopic methods.For example, low-temperature mid-infrared spectroscopy and inelastic neutron scattering (INS)are able to distinguish the ortho and para forms of H2@C60 [33], and the isomers of H2O@C60 havebeen distinguished by FIR and INS [14]. In principle, low-temperature MAS NMR of H2O@C60[14] and H2@C60 [39] should also detect conversion to the NMR-silent para form in solid samplesas loss of signal intensity as the fraction of the ortho form decreases. Decrease in the intensity ofthe electron–nuclear double resonance (ENDOR) signal from the radical ion of an endofullerenecontaining H2 with decreasing temperature has also been interpreted as loss of the ENDOR-activeortho isomer that is coupled to the unpaired electron via contact and dipolar interactions [40].To date, these methods have proven useful for in situ monitoring of the progress of ortho–para

on August 19, 2013rsta.royalsocietypublishing.orgDownloaded from

7

rsta.royalsocietypublishing.orgPhilTransRSocA371:20110628

......................................................

–32 –33 –34 –35 –36 –37 –38 –39

chemical shift (ppm)

H2O17@C

60

Figure 6. 17O NMR of H172 O@C60 (90%), 54.2 MHz, CDCl3/CS2 (1/1), room temperature. (Online version in colour.)

conversion. There are also reports that ortho- and para-H2O have been detected in the liquid phaseusing four-wave mixing spectroscopy [41]. It is not clear to what extent this method could beapplied to fullerene-encapsulated H2O.

(b) Paramagnetic catalysis of intersystem crossing in spin allotropesThe basic mechanism by which nuclear spin isomers may be interconverted by a magnetic catalystwas first described by Wigner [42]. This mechanism may be summarized as follows:

Transitions between nuclear spin states differing in the orientation of two (or more)geometrically equivalent spins may be achieved through magnetic symmetry breakingvia a field gradient across the spins applied by a suitably placed magnetic moment for asufficiently long time.

In Wigner’s theory, a dipolar field from a magnetic catalyst molecule colliding with H2 wasassumed. Expansions of the basic idea have included a variety of other interactions, many ofwhich pertain primarily to magnetic surfaces [43,44].

Our strategy for developing spin catalysts is based on Wigner’s mechanism. Three differentapproaches have so far been used:

(i) the endofullerene is dispersed by adsorption on a zeolite and exposed to liquid oxygenat 77 K [45,46]. This is in a sense the inverse of the commonly used method of para-H2enrichment by passing a stream of gaseous H2 over a surface-bound paramagnet such asactivated carbon [5]. In our case, the catalyst diffuses to the surface-bound H2 substrate;

(ii) a paramagnetic auxiliary group consisting of a nitroxide radical is fixed to the outsideof the fullerene cage, and a frozen solution of the modified endofullerene is cooled to afeasibly low temperature [47,48]; and

(iii) the small molecule is incorporated in a fullerene cage that forms a photo-accessible tripletstate of sufficiently long lifetime to allow interaction of the triplet magnetic moment withthe nuclei of the endo molecule [49].

A fourth proposed method which has so far not been successfully implemented for ortho–paraconversion but has been studied spectroscopically [40] is

on August 19, 2013rsta.royalsocietypublishing.orgDownloaded from

8

rsta.royalsocietypublishing.orgPhilTransRSocA371:20110628

......................................................

(iv) attachment of one or more unpaired electrons to the fullerene cage by chemical orphotochemical electron transfer to form a radical ion to provide an electron spininteraction analogous to that of a triplet state.

(c) Rapid catalyst removalEach of the catalytic methods outlined above has a corresponding means of rapid removal orinactivation of the catalyst once equilibrium has been attained:

(i) the O2 is removed by evaporation with pumping at 77 K. The sample is then warmed toroom temperature, and the adsorbed endofullerene extracted for further study [45,46];

(ii) as noted above, if necessary, the nitroxide may be rapidly converted to the correspondingdiamagnetic hydroxylamine using a reducing agent such as phenylhydrazine. Methodsof removing the auxiliary group, such as flash pyrolysis, to form the parent endofullereneare under development [47,48];

(iii) removal of the triplet state catalyst occurs spontaneously in milliseconds or less afterirradiation ceases [49]; and

(iv) the radical ion intermediate might similarly be removed almost instantaneously by backelectron transfer from a photo-generated radical ion pair, or more slowly by chemicalelectron transfer.

We have coined the term ‘magnetic switch’ to describe methods (ii) and (iii) for allotropeconversion [47].

4. Discussion of enrichment methods and results

(a) Wigner theoryIt seems likely that in all of the methods so far used with the endofullerenes the unpairedelectron(s) of the catalyst provide the field gradient necessary for intersystem crossing via themagnetic dipole–dipole interaction with the nuclei in the endo molecule. Method (ii), the use ofa paramagnetic auxiliary at a fixed distance from the centre of the fullerene cage, provides anespecially well-defined model for catalysis. Using the Wigner theory or a later formulation basedon the theory of spin relaxation [50], it can be shown [48] that the first-order rate coefficient,kpo (s−1), for a transition from the para (p) to ortho (o) states of H2 should vary as the inverse eighthpower of the distance, d, between the paramagnetic site, expressed in atomic units (0.529 Å):

kpo = Ω2d−8J(ωop, τr). (4.1a)

The prefactor Ω2 depends on the magnetic properties of the proton and paramagnet andfundamental constants (SI units):

Ω2 = 6( μ0

4π

)2[

γ 2P g2μ2

BS(S + 1)

a60

] (r0

a0

)2, (4.1b)

where r0 = 0.74 Å is the distance between the protons in the hydrogen molecule, S is the spin of theunpaired electron(s) of the paramagnet, gμB its magnetic moment and γP is the magnetogyric ratioof the proton. The spectral density, J(ωop, τr), depends on ωop = 2.3 × 1013 rad s−1, the frequencycorresponding to the 120 cm−1 energy separation of the ortho and para states, and τr, an averagelifetime during which the field gradient across the nuclei acts:

J(ωop, τr) = τr

1 + ω2opτ

2r

. (4.1c)

on August 19, 2013rsta.royalsocietypublishing.orgDownloaded from

9

rsta.royalsocietypublishing.orgPhilTransRSocA371:20110628

......................................................

perc

enta

ge o

f or

tho-

H2

80

60

40

20

0

0

O O OO

EtON

O O OOON

NO

1

2

95% para-H2

20 40 60 80 100time (days)

Figure 7. Ortho-H2 composition of a sample of two endofullerenes with paramagnetic auxiliaries cooled in liquid He forvarying times and analysed by rapid rewarming, NMRanalysis and recooling before appreciable back conversion. (Online versionin colour.)

(b) Comparison of Wigner theory with experiment: method (ii)We have measured values of kpo for several H2@C60 nitroxides which were enriched in the paraallotrope using a variation of method (i) on the diamagnetic hydroxylamine precursors of thenitroxide catalysts followed by rapid chemical oxidation to the nitroxide and measurement ofthe back conversion rate in solution at room temperature using NMR [48]. As predicted byequations (4.1), an excellent correlation is found between kpo and values of d−8 estimated frommolecular modelling. Rates for mono- and bis-nitroxides with the same value of d also confirmthe dependence on S. Comparison with calculated values based on estimates of d from molecularmodels, or more recent density functional theory calculations [51], yields estimates of τr of 0.2–0.3 ps which approximates the value of the correlation time for reorientation of the H2 moleculein C60 obtained from 1H relaxation times in H2@C60 [52,53].

(c) Enrichment at ultra low temperatureWe have recently extended method (ii) to a sample of a H2@C60 nitroxide and bis-nitroxide cooledin liquid He. Over a period of 26 days, approximately 95% enrichment in the para spin isomer isachieved as shown in figure 7.

(d) Catalysis by photoexcited tripletPhotolysis, method (iii), has been successfully used to produce enriched para-H2@C70 but forphotophysical reasons has not been successful for H2@C60 [49]. The method is quite sensitiveto the photodynamics of triplet formation and decay. The effect is enhanced in H2@C70 in frozensolution because of the long natural lifetime of the triplet and isolation of the molecules in a dilutematrix. Enrichment becomes undetectable in liquid solution, however, because of bimolecularquenching of the triplet.

on August 19, 2013rsta.royalsocietypublishing.orgDownloaded from

10

rsta.royalsocietypublishing.orgPhilTransRSocA371:20110628

......................................................

HD@C60

H2@C

60

HD@C60

-K8

H2@C

60-K8

PyPy PyO OO

O

Ph PhPhPh

PhPhN N N

S

H2@C

60-K12 H

2@C

60-K13

1H NMR

ppm–6–4 –5–2

HD@C60

-K13HD@C60

-K12

Figure 8. 500 MHz 1H NMR spectrum of isotopologue mixture H2/HD inside C60, C60-K8, C60-K12 and C60-K13 ino-dichlorobenzene-d4 at 300 K. The singlet in red corresponds to H2 and the triplet in black corresponds to HD. (Online versionin colour.)

Table 1. Lifetimes, τpo = 1/kpo, of H2, H2@C60 and H2@open-C60 as a function of concentration of the nitroxide TEMPO;o-dichlorobenzene-d4, 300 K.

5 mM 10 mM 15 mM 20 mM

lifetimes, τpo (h)

H2a 0.6 0.3 0.2 0.15. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

H2@C60 130 ± 7 140 ± 12 118 ± 7 86 ± 10. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

H2@C60-K8 82 ± 5 94 ± 10 91 ± 7 91 ± 7. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

H2@C60-K12 58 ± 3 70 ± 5 55 ± 3 67 ± 5. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

H2@C60-K13 72 ± 3 89 ± 7 70 ± 5 79 ± 5. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .aValues for H2 are estimated from kpo measured in toluene-d8 [34].

(e) Enrichment in open cage fullerenesThe remarkable sensitivity of the chemical shift of endo H2 to the cage environment has madeit possible to measure simultaneously in a single sample the rates of back conversion of thepara-enriched sample to equilibrium at room temperature for three open endofullerenes thatwere prepared as synthetic intermediates on the way to H2@C60. Their structures and the NMRspectrum of the mixture are shown in figure 8. The method used was bimolecular catalysis,induced by adding varying concentrations of the nitroxide 2,2,6,6-tetramethylpiperidine-1-oxyl(TEMPO) to a mixture in which each of the four fullerenes was enriched in the para isomer.Initial enrichment of the mixture was accomplished using adsorption on a zeolite and liquid O2as described above. Back conversion of H2@C60 and solutions of H2 induced by nitroxides hasbeen described previously [34].

The rate data, expressed as the pseudo-first-order back conversion lifetime, τpo, are givenin table 1 for different concentrations of TEMPO [54]. For comparison, the lifetime for backconversion of free H2 in solution under similar conditions is also given. As previously reported[34], catalysis of back conversion in free H2 by TEMPO is approximately two orders of magnitudemore effective than conversion in C60. This emphasizes the relative isolation of caged H2 fromexternal magnetic effects.

It can be seen from table 1 that catalysis of back conversion in H2@C60 by TEMPO is clearlyshown, as previously reported. It is apparent that the open forms back convert at a somewhatfaster rate than the parent H2@C60. This could be due to the presence of proton magnetic moments

on August 19, 2013rsta.royalsocietypublishing.orgDownloaded from

11

rsta.royalsocietypublishing.orgPhilTransRSocA371:20110628

......................................................

on the groups attached to the orifice (figure 8). Surprisingly, however, the open forms in thesame sample seem to show no consistent evidence of catalysis up to the highest concentrationsof catalyst that could be used because of solubility limits [54]. Because the back conversion ratefor all of the open forms is enhanced somewhat relative to H2@C60, in order to be detectablecatalysis would have to produce a rate comparable with that observed with H2@C60. This mightbe expected given that an opening in the cage could possibly allow more exposure to the catalystduring an encounter between the two molecules. The fact that treatment with liquid oxygen leadssimultaneously to enrichment in H2@C60 and in the open forms indicates that the opening doesnot fundamentally change the ability to catalyse conversion. The data in table 1 do, however,indicate that the dependence of bimolecular catalysis on structure is subtle.

5. Future directionsThe ability to purposefully incorporate H2, H2O, NH3 and CH4, and their various isotopologuesinto C60 or other fullerenes is now well established. Furthermore, the preparation of (H2)2@C70[55] presages the availability of whole families of endofullerenes encapsulating several smallmolecules. We list below some ways that these materials might be used to prepare and studynuclear spin isomers.

(a) H2OStudies of the translation–rotation energy levels of H2O@C60 are well underway [14]. Changesin the ortho : para ratio have been detected at cryogenic temperatures using FIR spectroscopy,and MAS NMR shows promise of being used in a similar fashion to monitor changes in theconcentration of ortho-H2O@C60.

The important question remains whether the back conversion rate in para-H2O will prove to besufficiently slow that paramagnetic catalysis can be detected and used, as in the case of H2@C60,in rationally designed schemes for producing samples enriched in the para isomer. Unlike the caseof H2, relatively rapid conversion in H2O and other polyatomic molecules may take place in theabsence of a magnetic catalyst [56]. This prediction is based, in part, on the smaller value of ωop

and the higher density of states, i.e. the number of potential pathways for conversion, availableto polyatomic molecules. The ortho–para perturbation required for transitions is provided by acombination of spin–rotation and nuclear dipole–dipole coupling [56]. Based on relaxation timemeasurements on the endo molecules in C60 and a paramagnetic derivative [57], however, theelectron–nuclear dipole interaction and reorientation correlation time for paramagnetic catalysisare similar for H2O and H2 in the same fullerene environment.

(b) D2 and D2OIt should be possible, in principle, to apply the same methods of detection and catalysis to the D2allotropes in the fullerene environment that we have used for H2. Because both the ortho and parastates of D2 are NMR-active, however, changes in intensity of the 2H NMR signal will be smallerfor the same degree of enrichment. Furthermore, the dependence of the Wigner mechanism onthe square of the nuclear moment indicates that magnetic catalysis under the same conditionsshould be at least an order of magnitude less effective at converting D2 spin isomers than hasbeen observed for H2. This may actually prove to be an asset for catalytic systems where theremoval of catalyst by chemical methods is slow compared the back conversion rate for H2.

Magnetic catalysis for D2O would be subject to the same differences relative to H2O as D2is relative to H2. Predicting the relative rates of the non-catalysed rates for D2O and H2O iscomplicated by the fact that the relative values of ωop and density of states for the two moleculesshould be opposite to the relative values of the spin–rotation and dipole–dipole interactions in

on August 19, 2013rsta.royalsocietypublishing.orgDownloaded from

12

rsta.royalsocietypublishing.orgPhilTransRSocA371:20110628

......................................................

the two molecules. That is, H2O has large ωop, smaller density of states, larger spin–rotationand dipole interaction. D2 has smaller ωop, higher density of states, smaller spin–rotation anddipole interaction.

It is also possible that the nuclear quadrupole moment of the deuteron may interact with localelectric field gradients in D2 or D2O encapsulated in the fullerenes to bring about interconversionof the spin isomers, as first proposed by Casimir [58].

(c) NH3 and CH4Almost nothing is known about the behaviour of spin isomers of these molecules in condensedphases at ordinary temperatures. However, nuclear spin conversion in CH4, both uncatalysed[59] and catalysed by O2 [60], has been studied in the solid at low temperatures.

(d) (H2)2@C70The ability to encapsulate two, or possibly more, H2 molecules in a fullerene cage, pioneeredby the preparation of (H2)2@C70 [55], suggests fascinating possibilities for studying cooperativeeffects in ortho–para conversion. Such effects have been studied, for example, as they affect theNMR spectrum of solid H2 [61]. It should be possible to enrich the para-H2 form of (H2)2@C70using the photoexcitation method which we have shown to work for H2@C70 [49]. Being able toadjust the fractions of the para : para, para : ortho and ortho : ortho forms should make it possible, forexample, to study spectroscopically at low temperatures, in condensed media, the interactionsbetween isolated hydrogen molecule pairs in carefully controlled rotational states.

6. Applications of isolated spin isomersThe ability to prepare samples of fullerenes that contain a substantial excess of one nuclearspin isomer of an encapsulated small molecule is of scientific importance simply because of theinsights it provides about subtle aspects of energy exchange in small molecules and the waythat they interact with well-defined surroundings. Nevertheless, there are also some potentiallyimportant practical aspects of research in this area. Two examples follow.

(a) NMR intensity enhancementA sample of para-H2 or para-H2O corresponds to selective population of only one of the four spinstates available to a pair of equivalent protons. On the other hand, the four levels in the samesample at equilibrium at room temperature, even in the highest available magnetic field, havepopulations that differ by only a few parts per million. The intensity of transitions among thespin levels in an NMR or MRI measurement depends on the difference in population betweenthe levels. It is clear that an enormous increase in signal intensity can be achieved if the selectivespin state population of a pure sample of the para form can be used in an NMR experiment. Asdetailed above, the basic PHIP phenomenon [15,16] accomplishes this through chemical reactionsthat break the symmetry of the spin pair without changing the relative orientations of the spins. Inother words, the NMR-silent I = 0 state of two protons is converted into the corresponding NMR-active states of two non-equivalent I = 1/2 particles with the potential for enormous increase inthe intensity of the observed transitions [20].

Similar applications of encapsulated para-H2 or H2O will require developing methods bywhich the spin states of the endo protons can communicate their selective population to theoutside world. One possibility would be to design open cages, or ways of opening closed cages,that would be able to expel enriched para molecules into the surrounding medium to be used aschemical reagents to produce enhancement via PHIP. Indeed, such chemical exchange betweenendo and exo water molecules has been measured in the NMR spectrum of an open cage fullerenecontaining water [62].

on August 19, 2013rsta.royalsocietypublishing.orgDownloaded from

13

rsta.royalsocietypublishing.orgPhilTransRSocA371:20110628

......................................................

(b) Spin-selective D2 fusionThe direct [63] and muon-catalysed [10] fusion of two deuterium nuclei is nuclear-spin-dependent. It has been proposed [64], for example, that a sample of ortho-D2, the more stableform, should produce neutrons as a by-product of nuclear fusion with a lower probability than thepara form. There has also been some interest in the possibility that confinement of D2 in clustersor other media that could be ‘crushed’ in an intense photon field would improve the efficiencyof inertial confinement methods of fusion [65]. High power laser irradiation of a beam of ortho-enriched D2@C60 molecules, which might produce a plasma of ionized carbon and deuterons,would provide an especially well-defined way of studying the feasibility of both ideas.

7. Concluding remarksBased on the above results, we are confident that developments in methodology for generatingand using samples of small molecules encapsulated in fullerenes and enriched in spin allotropesor other specific nuclear spin states should be of increasing importance in chemistry, medicineand other areas of application. The possibility of measuring the energetics and dynamics of smallmolecules in a well defined, nearly isolating, environment in condensed media also providesa set of irresistibly sweet problems that should occupy experimentalists and theorists for sometime to come.

Acknowledgements. The authors thank the National Science Foundation for its generous support through grantno. CHE-11-11392.

References1. Lewis GN, MacDonald RT. 1933 Some properties of pure (H2H2O). J. Am. Chem. Soc. 55,

3057–3059. (doi:10.1021/ja01334a515)2. Bernal JD, Tamm G. 1935 Zero point energy and physical properties of H2O and D2O. Nature

135, 229–230. (doi:10.1038/135229b0)3. Habershon S, Manolopoulos DE. 2011 Thermodynamic integration from classical to quantum

mechanics. J. Chem. Phys. 135, 224111. (doi:10.1063/1.3666011)4. Dennison DM. 1927 A note on the specific heat of the hydrogen molecule. Proc. R. Soc. Lond.

A 115, 483–486. (doi:10.1098/rspa.1927.0105)5. Farkas A. 1935 Orthohydrogen, parahydrogen and heavy hydrogen. Cambridge, UK: Cambridge

University Press.6. Schmauch GE, Singleton AH. 1964 Technical aspects of ortho–para hydrogen conversion. Ind.

Eng. Chem. 56, 20–31. (doi:10.1021/ie50653a003)7. Akhtar S, Smith HA. 1964 Separation and analysis of various forms of hydrogen by adsorption

and gas chromatography. Chem. Rev. 64, 261–276. (doi:10.1021/cr60229a003)8. Mohnke M, Saffert W. 1962 Adsorption chromatography of hydrogen isotopes with

capillary columns. In Gas chromatography (ed. M Van Swaay), pp. 216–221. Washington, DC:Butterworth.

9. Bargon J, Kandels J, Woelk K. 1990 NMR study of nuclear spin polarization during chemicalreactions with ortho hydrogen. Angew. Chem. Int. Ed. 29, 58–59. (doi:10.1002/anie.199000581)

10. Imao H, Ishida K, Kawamura N, Matsuzaki T, Matsuda Y, Toyoda A, Strasser P, IwasakiM, Nagamine K. 2008 Preparation of ortho–para ratio controlled D2 gas for muon-catalyzedfusion. Rev. Sci. Instrum. 79, 053502. (doi:10.1063/1.2918538)

11. Chapovsky PL, Krasnoperov LN, Panfilov VN, Strunin VP. 1985 Studies on separation andconversion of spin modifications of CH3F molecules. Chem. Phys. 97, 449–455. (doi:10.1016/0301-0104(85)87052-X)

12. Bechtel C, Elias E, Schramm BF. 2005 Nuclear spin symmetry state relaxation in formaldehyde.J. Mol. Struct. 741, 97–106. (doi:10.1016/j.molstruc.2005.01.056)

13. Schramm B, Bamford DJ, Moore CB. 1983 Nuclear spin state conservation in photodissociationof formaldehyde. Chem. Phys. Lett. 98, 305–309. (doi:10.1016/0009-2614(83)80212-7)

14. Beduz C et al. 2012 Quantum rotation of ortho and para-water encapsulated in a fullerene cage.Proc. Natl Acad. Sci. USA 109, 12 894–12 898. (doi:10.1073/pnas.1210790109)

on August 19, 2013rsta.royalsocietypublishing.orgDownloaded from

14

rsta.royalsocietypublishing.orgPhilTransRSocA371:20110628

......................................................

15. Bowers CR, Jones DH, Kurur ND, Labinger JA, Pravica MG, Weitekamp DP. 1990Symmetrization postulate and nuclear magnetic resonance of reacting systems. Adv. Magn.Reson. 14, 269–291. (doi:10.1016/B978-0-12-025514-6.50018-6)

16. Eisenberg R. 1991 Parahydrogen induced polarization: a new spin on reactions with H2. Acc.Chem. Res. 24, 110–116. (doi:10.1021/ar00004a004)

17. Natterer J, Greve T, Bargon J. 1998 Orthodeuterium induced polarization. Chem. Phys. Lett.293, 455–460. (doi:10.1016/S0009-2614(98)00784-2)

18. Blazina D, Duckett SB, Dunne JP, Godard C. 2004 Applications of the parahydrogenphenomenon in inorganic chemistry. Dalton Trans. 2004, 2601–2609. (doi:10.1039/B409606A)

19. Duckett SB, Mewis RE. 2012 Application of parahydrogen induced polarization techniques inNMR spectroscopy and imaging. Acc. Chem. Res. 45, 1247–1257. (doi:10.1021/ar2003094)

20. Gong Q, Gordji-Nejad A, Blümlich B, Appelt S. 2010 Trace analysis by low-field NMR:breaking the sensitivity limit. Anal. Chem. 82, 7078–7082. (doi:10.1021/ac101738f)

21. Ardenkjær-Larsen JH, Fridlund B, Gram A, Hansson G, Hansson L, Lerche MH, Servin R,Thaning M, Golman K. 2003 Increase in signal-to-noise ratio of >10,000 times in liquid stateNMR. Proc. Natl Acad. Sci. USA 100, 10 158–10 163. (doi:10.1073/pnas.1733835100)

22. Levitt MH. 2012 Singlet nuclear magnetic resonance. Annu. Rev. Phys. Chem. 63, 89–105.(doi:10.1146.annrev-physchem-032511-143724)

23. Hübler P, Bargon J, Glaser SJ. 2000 Nuclear magnetic resonance quantum computingexploiting the pure spin state of para hydrogen. J. Chem. Phys. 113, 2056–2059. (doi:10.1063/1.482015)

24. Anwar MS, Blazina D, Carteret HA, Duckett SB, Halstead TK, Jones JA, Kozak CM, TaylorRJK. 2004 Preparing high purity initial states for nuclear magnetic resonance quantumcomputing. Phys. Rev. Lett. 93, 040501. (doi:10.1103/PhysRevLett.93.040501)

25. Twamley J. 2003 Quantum cellular automata quantum computing with endohedral fullerenes.Phys. Rev. A 67, 052318. (doi:10.1103/PhysRevA.67.052318)

26. Saunders M, Cross RJ, Jimenez-Vazquez HA, Shimshi R, Khong A. 1996 Noble gas atomsinside fullerenes. Science 271, 1693–1697. (doi:10.1126/science.271.5256.1693)

27. Komatsu K, Murata M, Murata Y. 2005 Encapsulation of molecular hydrogen in fullerene C60by organic synthesis. Science 307, 238–240. (doi:10.1126/science.1106185)

28. Iwamatsu SI, Uozaki T, Kobayashi K, Re S, Nagase S, Murata S. 2004 A bowl-shaped fullereneencapsulates a water into the cage. J. Am. Chem. Soc. 126, 2668–2669. (doi:10.1021/ja038537a)

29. Kurotobi K, Murata Y. 2011 A single molecule of water encapsulated in fullerene C60. Science333, 613–616. (doi:10.1126/science.1206376)

30. Whitener Jr KE, Frunzi M, Iwamatsu S-I, Murata S, Cross RJ, Saunders M. 2008 Puttingammonia into a chemically opened fullerene. J. Am. Chem. Soc. 130, 13 996–13 999. (doi:10.1021/ja805579m)

31. Whitener Jr KE, Cross RJ, Saunders M, Iwamatsu S-I, Murata S, Mizorogi N, Nagase S.2009 Methane in an open-cage [60]fullerene. J. Am. Chem. Soc. 131, 6338–6339. (doi:10.1021/ja901383r)

32. Horsewill AJ et al. 2012 Inelastic neutron scattering investigations of the quantum moleculardynamics of a H2 molecule entrapped inside a fullerene cage. Phys. Rev. B 85, 205440.(doi:10.1103/PhysRevB.85.205440)

33. Mamone S, Chen JYC, Bhattacharyya R, Levitt MH, Lawler RG, Horsewill AJ, Rõõm T, BacicZ, Turro NJ. 2011 Theory and spectroscopy of an incarcerated quantum rotor: the infraredspectroscopy, inelastic neutron scattering and nuclear magnetic resonance of H2@C60 atcryogenic temperature. Coord. Chem. Rev. 255, 938–948. (doi:10.1016/j.ccr.2010.12.029)

34. Turro NJ et al. 2009 The spin chemistry and magnetic resonance of H2@C60. From the Pauliprinciple to trapping a long lived nuclear excited spin state inside a buckyball. Acc. Chem. Res.43, 335–345. (doi:10.1021/ar900223d)

35. Xu M, Sebastianelli F, Bacic Z, Lawler R, Turro NJ. 2008 H2, HD, and D2 inside C60: coupledtranslation–rotation eigenstates of the endohedral molecules from quantum five-dimensionalcalculations. J. Chem. Phys. 129, 064313. (doi:10.1063/1.2967858)

36. Kellogg JMB, Rabi II, Ramsey Jr NF, Zacharias JR. 1939 The magnetic moments of the protonand the deuteron: the radiofrequency spectrum of H2 in various magnetic fields. Phys. Rev. 56,728–743. (doi:10.1103/PhysRev.56.728)

37. Redington RL, Milligan DE. 1963 Molecular rotation and ortho–para nuclear spin conversionof water suspended in solid Ar, Kr and Xe. J. Chem. Phys. 39, 1276–1284. (doi:10.1063/1.1734427)

on August 19, 2013rsta.royalsocietypublishing.orgDownloaded from

15

rsta.royalsocietypublishing.orgPhilTransRSocA371:20110628

......................................................

38. Lopez-Gejo J, Marti AA, Ruzzi M, Jockusch S, Komatsu K, Tanabe F, Murata Y, Turro NJ. 2007Can H2 inside C60 communicate with the outside world? J. Am. Chem. Soc. 129, 14 554–14 555.(doi:10.1021/ja076104s)

39. Carravetta M et al. 2007 Solid-state NMR of endohedral hydrogen–fullerene complexes. Phys.Chem. Chem. Phys. 9, 4879–4894. (doi:10.1039/b707075f)

40. Zoleo A, Lawler RG, Lei X, Li Y, Murata Y, Komatsu K, Di Valentin M, Ruzzi M, Turro NJ.2012 ENDOR evidence of electron–H2 interaction in a fulleride embedding H2. J. Am. Chem.Soc. 134, 12 881–12 884. (doi:10.1021/ja305704n)

41. Bunkin AF, Pershin SM. 2008 Observation of water isotopes and spin-isomers rotationaltransitions induced by four-wave mixing in liquid. J. Raman Spectrosc. 39, 726–729. (doi:10.1002/jrs.1992)

42. Wigner E. 1933 Über die paramagnetische umwandlung von para-orthowasserstoff III. Zeit.fur Physik. Chem. B 23, 28–32.

43. Petzinger KG, Scalapino DJ. 1973 Para- to ortho-hydrogen conversion on magnetic surfaces.Phys. Rev. B 8, 266–279. (doi:10.1103/PhysRevB.8.266)

44. Ilisca E. 1992 Ortho–para conversion of hydrogen molecules physisorbed on surfaces. Prog.Surf. Sci. 41, 217–335. (doi:10.1016/0079-6816(92)90019-E)

45. Turro NJ et al. 2008 Demonstration of a chemical transformation inside a fullerene. Thereversible conversion of the allotropes of H2@C60. J. Am. Chem. Soc. 130, 10 506–10 507.(doi:10.1021/ja804311c)

46. Ge M et al. 2011 Interaction potential and infrared absorption of endohedral H2 in C60. J. Chem.Phys. 134, 054507. (doi:10.1063/1.3535598)

47. Li Y et al. 2010 A magnetic switch for spin-catalyzed interconversion of nuclear spin isomers.J. Am. Chem. Soc. 132, 4042–4043. (doi:10.1021/ja910282p)

48. Li Y, Lei X, Lawler RG, Murata Y, Komatsu K, Turro NJ. 2011 Distance-dependent para-H2-to-ortho-H2 conversion in H2@C60 derivatives covalently linked to a nitroxide radical. J. Phys.Chem. Lett. 2, 741–744. (doi:10.1021/jz200192s)

49. Frunzi M et al. 2011 A photochemical on-off switch for tuning the equilibrium mixture ofH2 nuclear spin isomers as a function of temperature. J. Am. Chem. Soc. 133, 14 232–14 235.(doi:10.1021/ja206383n)

50. Atkins PW, Clugston MJ. 1974 Ortho–para hydrogen conversion in paramagnetic solutions.Mol. Phys. 27, 1619–1631. (doi:10.1080/00268977400101351)

51. Rastrelli F, Frezzato D, Lawler RG, Li Y, Turro NJ, Bagno A. 2013 Predicting the paramagnet-enhanced NMR relaxation of H2 encapsulated in endofullerene nitroxides by density-functional theory calculations. Phil. Trans. R. Soc. A 371, 20110634. (doi:10.1098/rsta.2011.0634)

52. Sartori E, Ruzzi M, Turro NJ, Decatur JD, Doetschman DC, Lawler RG, Buchachenko AL,Murata Y, Komatsu K. 2006 Nuclear relaxation of H2 and H2@C60 in organic solvents. J. Am.Chem. Soc. 128, 14 752–14 753. (doi:10.1021/ja065172w)

53. Chen JY-C, Martí AA, Turro NJ, Komatsu K, Murata Y, Lawler RG. 2010 Comparative NMRproperties of H2 and HD in toluene-d8 and in H2/HD@C60. J. Phys. Chem. B 114, 14 689–14 695.(doi:10.1021/jp102860m)

54. Chen JY-C. 2011 Spin chemistry of guest@host systems: H2@C60 and nitroxide@octa acid. PhDdissertation, Columbia University, New York.

55. Murata M, Maeda S, Morinaka Y, Murata Y, Komatsu K. 2008 Synthesis and reaction offullerene C70 encapsulating two molecules of H2. J. Am. Chem. Soc. 130, 15 800–15 801.(doi:10.1021/ja8076846)

56. Curl Jr RF, Kasper JVV, Pitzer KS. 1967 Nuclear spin state equilibration through nonmagneticcollisions. J. Chem. Phys. 46, 3220–3228. (doi:10.1063/1.1841193)

57. Li Y, Chen JY-C, Lei X, Lawler RG, Murata Y, Komatsu K, Turro NJ. 2012 Comparison ofnuclear spin relaxation of H2O@C60 and H2@C60 and their nitroxide derivatives. J. Phys. Chem.Lett. 3, 1165–1168. (doi:10.1021/jz3002794)

58. Casimir HBG. 1940 The ortho–para conversion of deuterium and the electric quadrupolemoment of the deuteron. Physica 7, 169–176. (doi:10.1016/S0031-8914(40)90103-3)

59. Miyamoto Y, Fushitani M, Ando D, Momose T. 2008 Nuclear spin conversion of methane insolid parahydrogen. J. Chem. Phys. 128, 114502. (doi:10.1063/1.2889002)

60. Kim, JJ, Pitzer KS. 1977 Crystal field effects on oxygen in solid methane and the catalysis ofspin–species conversion of methane. J. Chem. Phys. 66, 2400–2407. (doi:10.1063/1.434277)

on August 19, 2013rsta.royalsocietypublishing.orgDownloaded from

16

rsta.royalsocietypublishing.orgPhilTransRSocA371:20110628

......................................................

61. Berlinsky AJ, Hardy WN. 1973 Theory of ortho–para conversion and its effects on theNMR spectrum of ordered solid ortho-hydrogen. Phys. Rev. B 8, 5013–5027. (doi:10.1103/PhysRevB.8.5013)

62. Frunzi M, Baldwin AM, Shibata N, Iwamatsu S-I, Lawler RG, Turro NJ. 2011 Kineticsand solvent-dependent thermodynamics of water capture by a fullerene-based hydrophobicnanocavity. J. Phys. Chem. A 115, 735–740. (doi:10.1021/jp110832m)

63. Kulsrud RM. 1987 Polarized advanced fuel reactors. Report no. PPPL-2446. Princeton, NJ: PlasmaPhysics Laboratory, Princeton University. (doi:10.2172/6169161)

64. Kulsrud RM. 1988 Application of polarized nuclei and advanced fuel fusion. Nucl. Instrum.Methods Phys. Res. A 271, 4. (doi:10.1016/0168-9002(88)91119-9)

65. Zweiback J, Smith RA, Cowan TE, Hays G, Wharton KB, Yanovsky VP, Ditmire T. 2000Nuclear fusion driven by coulomb explosions of large deuterium clusters. Phys. Rev. Lett. 84,2634–2637. (doi:10.1103/PhysRevLett.84.2634)

on August 19, 2013rsta.royalsocietypublishing.orgDownloaded from