This document is the Accepted Manuscript version of
a Published Work that appeared

in final form in
Chem. Commun., 2015, 51, 15596 DOI: 10.1039/c5cc06423f

	

Mechanism	of	CO2	Hydration:	Porous	Metal	Oxide	Nanocapsule	Catalyst	Can	Mimic	
the	Biological	Carbonic	Anhydrase	Role†		
a

b	

b

Nuno	A.	G.	Bandeira, 	Somenath	Garai, Achim	Müller 	and	Carles	Bo

a,c	

	
	
	
	

The	mechanism	for	the	hydration	of	CO2	within	a	Keplerate	nanocapsule	is	presented.	A	network	of	hydrogen	bonds	across	the	water	
layers	 in	 the	 first	 metal	 coordination	 sphere	 facilitates	 the	 proton	 abstraction	 and	 nucleophilic	 addition	 of	 water.	 The	 highly	 acidic	
properties	of	the	polyoxometalate	cluster	are	crucial	in	explaining	the	catalysed	hydration.		

	

		

	

		
1

Concerns	about	global	warming,	together	with	the	incoming	necessity	to	find	alternative	feedstocks	to	fossil	fuels, 	have	boosted	interest	
2-5
in	 the	 capture	 and	 use	 of	 CO2	 as	 a	 chemical	 starting	 material. 	 Living	 organisms	 having	 the	 carbonic	 anhydrase	 enzyme	 carry	 out	 the	
simplest	CO2	transformation,	i.e.	hydration	to		
carbonic	acid,	in	an	easy	manner.	The	presence	of	an	electrophilic	Zn	center	together	with	a	network	of	water	molecules	in	the	proximity	of	
the	enzyme	site	makes	the	hydration	reaction	possible,		
	
6-9
10
which	is	rather	slow	in	the	absence	of	catalyst. The	exploration	of	carbonic	anhydrase 	and	related	analogues 	has	afforded	major	bio-

Figure 1. The pictorial representation of the {Mo132} Keplerate capsule.

inspired	 catalytic	 routes	 for	 CO2	 fixation	 over	 the	 past	 decades.	 On	 the	 other	 hand,	 synthetic	 chemistry	 afforded	 new	 transition	 metal	
11
based	 catalysts	 that	 can	 convert	 CO2	 into	 other	 chemical	 entities,	 for	 instance	 CO2	 reduction	 to	 methanol, 	 coupling	 with	 oxiranes	 to	
12-14
15,	16
produce	cyclic	carbonates,
	or	other	value	added	chemicals.
			
17
Some	 of	 us	 reported	 recently 	 a	 novel	 way	 for	 sequestrating	 and	 transforming	 CO2	 into	 carbonate	 by	 encapsulation	 within	 unique	
molybdenum	oxide	nanocapsules.	These	capsules,	belonging	to	the	Keplerate	family,	are	nano-sized	molecular	metal	oxide	spheres	with	
VI
VI
V
2
n2- 18
the	general	formula	[{(M )M 5O21(H2O)6}12{M’ 2O2X2(µ -Y)}30] 	(M=Mo,	W;	M’=Mo;	X=O,	S;	Y=bridging	ligand,	e.g.	RCOO ,SO4 ). 	This	sort	
VI
V
of	 capsule	 contains	 12	 pentagonal	 {Mo 6}	 units	 placed	 at	 the	 vertices	 of	 an	 icosahedron	 and	 linked	 by	 30	 binuclear	 {Mo 2}	 units.	 This	
arrangement	leads	to	the	formation	of	capsules	(Figure	1)	with	twenty	{M3Mo6O9}-type	pores	and	a	cavity	where	a	large	quantity	of	water	
19,	 20
VI
VI
V
2
molecules,	 anions	 or	 other	 species	 can	 be	 confined.
	 By	 bubbling	 CO2	 in	 a	 solution	 of	 (NH4)42[{(Mo )Mo 5O21(H2O)6}12{Mo 2O4(µ 21
CH3COO)}30]·	ca.	10	CH3COONH4·	ca.	300	H2O	≡	(NH4)42·Anion	1a·	ca.	10	CH3COONH4·	ca.	300	H2O	≡	Compound	1 	at	pH	7	the	carbonate	
VI
VI
V
2
derivative	 (NH4)72	 [{(Mo )Mo 5O21(H2O)6}12{Mo 2O4(µ -CO3)}30]·	 ca.	 260	 H2O	 ≡	 (NH4)72·	 Anion	 2a·	 ca.	 260	 H2O	 ≡	 Compound	 2	 was	
17
obtained. 	 The	 Figure	 2.	 The	 pictorial	 representation	 of	 the	 {Mo132}	 Keplerate	 outcome	 of	 these	 results	 urged	 the	 major	 question	 of	
whether	the	carbonate	anion	formed	in	solution	(in	minute	amounts	at	pH	7)	was	captured	by	the	Keplerate	sphere,	or	more	interestingly,	
V
VI
whether	 the	 carbonate	 anion	 formation	 took	 place	 in	 situ	 inside	 the	 capsule,	 either	 at	 the	 Mo 	 or	 Mo 	 coordination	 sites,	 by	 a	 metal	
catalysed	nucleophilic	addition	of	water	to	a	solubilised	CO2	molecule,	likewise	the	accepted	mechanism	of	carbonic	anhydrase.	

	

	

Figure 3. Several mechanistic pathways for CO2 hydration. (red) Uncatalysed reaction; (green, blue, black) Catalysed reaction. Electronic energies and
–1
Gibbs free energies in parenthesis evaluated using a partial Hessian. All energies in kJ.mol .
17

The	CO2	transformation	is	also	reversible	via	acidification	of	the	aqueous	solution	of	Compound	2.	 	 The	 results	 of	 the	 theoretical	 study	
22,	 23
presented	 herein	 suggest	 that	 this	 transformation	 of	 CO2	 to	 carbonate	 is	 actually	 the	 third	 example
	 known	 to	 date	 of	 a	 catalytic	
V
VI
process	occurring	inside	the	{Mo132}	capsule,	where	the	Mo 	and	also	the	Mo 	sites	play	a	role.	
24,	25
	 he	mechanism	of	the	hydration	of	CO2	to	form	the	carbonic	acid	has	been	a	subject	of	theoretical	studies	over	the	past	decades.
T
	The	
challenge	lies	in	the	accurate	description	of	the	explicit	water	molecules	participating	in	the	reaction	as	was	shown	by	the	latest	work	of	
26
27
Yamabe	 and	 Kawagishi. 	 The	 uppermost	 energy	 barrier	 of	 carbon	 dioxide	 hydration	 is	 always	 the	 initial	 step	 of	 water	 addition. 	 The	
24-26,	28
arrangement	of	this	initial	transition	state
	is	a	cyclic	three	water	molecular	arrangement	such	as	the	one	depicted	in	Figure	2.	We	will	
adopt	this	model	as	a	benchmark	to	compare	with	our	own	calculations	on	the	catalytic	sequestration	of	CO2	and	its	conversion	into	the	
carbonate	form.	
In	 a	 recent	 study	 we	 demonstrated	 that	 by	 using	 a	 cluster	 model	 of	 the	 {Mo132}	 nanocapsule,	 the	 reaction	 pathway	 of	 the	 reversible	
22
cleavage	of	methyl-tert-butyl	ether 	was	successfully	unravelled.	The	model	assembly	was	defined	to	mimic	the	nature	of	the	active	sites	
VI
VI
V
6+
VI
VI
of	the	Keplerate	and	it	was	formulated	as	[{(Mo )Mo 5O13(OH)8}2{Mo 2O4}] 	containing	two	pentagonal	{(Mo )Mo 5}-units	and	one	linker	
V
unit	of	the	type	{Mo 2O4}.	It	fully	retained	the	essential	characteristics	of	the	{Mo132}	reactive	sites	and	therefore	we	have	selected	that	
model	for	the	present	study.	Since	the	formation	of	the	carbonate	anion	takes	place	in	aqueous	media,	the	presence	of	water	molecules	
inside	 the	 Keplerate	 sphere	 must	 play	 an	 essential	 role	 in	 the	 reaction	 and	 therefore	 it	 is	 essential	 that	 the	 cluster	 model	 should	
incorporate	a	sufficiently	large	number	of	water	molecules.	Thus	we	included	13	additional	water	molecules	explicitly	in	this	study,	so	the	
VI
VI
V
6+
model	 used	 is	 formulated	 as	 [{(Mo )Mo 5O13(H2O)6(OH)8}2{Mo 2O4(H2O)}] ,	 which	 leaves	 one	 vacant	 coordination	 site	 for	 the	
V
interaction/coordination	 of	 CO2	 to	 one	 of	 the	 two	 Mo 	 centres,	 while	 the	 second	 one	 bears	 a	 water	 molecule	 which	 is	 supposed	 to	 be	
highly	reactive.	
V
1
2
As	expected	the	CO2	molecule	being	nonpolar,	does	not	coordinate	to	Mo 	centre	either	in	η 	or	η 	fashion.	Notwithstanding,	we	could	
V
characterize	a	weakly	bound	stationary	point	structure	in	which	CO2	is	hydrogen-bonded	to	the	water	molecule	in	one	Mo 	centre	and	to	a	
VI
water	 molecule	 on	 Mo ,	 thus	 located	 in	 the	 vicinity	 of	 the	 reactive	 centre.	 This	 will	 be	 our	 starting	 point	 (named	 Reactants)	 for	 the	
reaction	path	studies	defining	the	zero	of	energies.		
	 he	 highest	 energy	 reaction	 path	 explored	 TS1	 (Figure	 2)	 is	 perhaps	 the	 most	 intuitive	 pathway	 involving	 a	 concerted	 nucleophilic	
T
+
addition	of	an	aqua	ligand	to	CO2	followed	by	the	subsequent	proton	rejection	and	formation	of	a	local	Zundel	cation	(H5O2 )	sponsored	by	
VI
the	hydrogen	bonding	of	the	neighbouring	aqua	ligands.	Note	that	the	neighbouring	water	ligands	coordinated	to	Mo 	centres	contribute	
to	 stabilizing	 the	 rejected	 proton	 and	 the	 concomitant	 formation	 of	 bicarbonate.	 Although	 we explored multiple	 conformational	
possibilities,	a	coordinated	adduct	of	the	type	{O2C-OH2}	could	not	be	obtained.	
	 wing	to	the	accumulation	of	positive	charge	closer	to	the	metal	centres,	TS1	transition	state	is	shown	to	be	too	excessively	high	in	
O
–1
energy	(+89	kJ.mol )	to	become	a	competitive	pathway	vis-à-vis	the	unassisted	TSw	transition	state	for	hydration	of	CO2.		

	

	
	n	light	of	these	results	we	explored	a	different	route	that	yielded	a	bicarbonate	coordinated	intermediate	resulting	from	a	nucleophilic	
I
VI
V
addition	of	a	hydroxo	to	CO2.	Given	that	the	Mo 	centres	are	more	Lewis	acidic	than	Mo 	the	likely	candidate	for	a	good	reactant	would	be	
V
VI
V
VI
2b	bearing	the	{Mo (OH2)-O-Mo (OH)}	unit	rather	than	2a	({Mo (OH)-O-Mo (OH2)}).	This	is	borne	out	by	the	relative	energetics	of	the	two	
–1
V
isomers,	which	favour	2b	by	some	5	kJ.mol .	The	mechanism	should	expectedly	involve	a	proton	relay	from	the	aqua-ligand	in	the	Mo 	
centre	 concerted	 with	 the	 nucleophilic	 addition	 of	 the	 hydroxo	 group	 to	 CO2.	 The	 ΔG	 estimate	 for	 the	 2a→2b	 conversion	 is	 further	
–1
widened	to	17	kJ.mol 	in	favour	of	2b.		
	 he	bicarbonate	intermediate	undergoes	further	deprotonation	resulting	in	2d.	The	release	of	a	proton	from	2d	through	to	2e	has	a	
T
-1
-1
negligible	energy	barrier	(for	TS2d,	2	kJ.	mol 	in	electronic	or	+8	kJ.mol 	in	free	energy).	The	carbonate	intermediate	2e	is	approximately	
iso-energetic	with	its	parent	bicarbonate	2d	but	can	be	easily	converted	to	2f	with	lower	free	energy.	The	intermediate	2f	has	one	noncoordinated	water	molecule	which	stabilises	the	carbonate	ligand	via	hydrogen	bonding.	The	Mo-carbonate	bond	lengths	in	2e	are	2.392	
12
and	2.329	Å,	which	are	within	the	error	limits	of	the	experimentally	determined	values. 		
VI
The	higher	acidity	of	the	Mo 	centre	prompted	us	to	explore	another	possible	mechanistic	route	in	which	the	direct	nucleophilic	addition	

TSw	

TS2b	

Figure 4. Transition state structures for the uncatalysed CO2 + 3H2 O (TSw)
and for the catalysed reaction (TS2b). Selected distances in Å and angles in
degrees.
VI

V

to	the	CO2	molecule	takes	place	directly	by	the	hydroxo	group	coordinated	to	the	Mo 	sites	while	the	vacant	coordination	site	of	Mo 	is	
V
utilized	to	stabilize	the	transition	state.	A	subsequent	backflip	of	bicarbonate	or	carbonate	to	the	{Mo 2}-linker	would	be	necessary	to	be	
consistent	with	the	final	carbonate	adduct.	The	initial	steps	of	this	pathway	are	sketched	in	blue	in	Figure	2.	The	transition	state	TS3	has	a	
similar	energy	value	to	TSw	(the	uncatalysed	transition	state)	but	the	intermediate	3a	is	not	sufficiently	stable	to	be	considered	a	viable	
route	(see	Supplementary	material	for	these	additional	structures).				
There	are	structural	differences	between	the	catalysed	and	uncatalysed	systems	namely	with	regards	to	each	transition	state	which	are	
29
summarised	Figure	3.	The	Mayer-Mulliken	bond	orders 	(MBO)	were	also	analysed	in	the	present	case	which	reflect	the	bond	strength	
between	the	different	atoms	in	any	given	system.	The	most	striking	difference	between	TSw	and	TS2b	is	that	the	latter	is	a	slightly	“lesser	
bound”	 transition	 state	 with	 a	 reaction	 coordinate	 (C-O)	 bond	 order	 0.377	 whereas	 in	 TSw	 it	 is	 0.557.	 The	 ∠(O-C-O)	 angles	 are	 also	
considerably	different	between	TSw	(139°)	and	TS2b	(152°)	consistent	with	a	larger	electron	cloud	of	the	incoming	O(-C)	and	consequently	
a	 lower	 angular	 distortion	 of	 CO2.	 The	 leaving	 proton	 is	 also	 more	 bound	 to	 the	 oxygen	 atom	 in	 TSw	 (MBO=0.430)	 than	 in	 TS2b	
(MBO=0.250).	In	the	latter	case	the	outgoing	proton	from	the	aqua	ligand	is	already	at	a	large	distance	(1.535	Å,	see	Figure	3).	
	 inally	to	predict	the	potential	reactivity	of	related	systems,	additional	calculations	were	carried	out	on	model	analogues	of	the	{W72Mo60}	
F
30
and	{W132}	nanocapsules.	The	former	nanocapsule	has	been	characterised 	experimentally	although	the	latter	is	still	unknown.	Since	the	
VI
key	point	in	the	mechanism	is	the	generation	of	the	nucleophilic	hydroxo	species	coordinated	to	the	star-shaped	M 	moieties,	the	relative	
–1
thermodynamic	stability	of	2a	and	2b	species	was	determined.	The	calculated	ΔE(2a→2b)	is	–65	kJ.mol 	for	the	mixed	W/Mo	oxo-cluster	
–1
model	and	–85	kJ.mol 	for	the	hypothetical	full	W	system.	This	points	to	a	likely	enhanced	reactivity	of	the	heavier	metal	Keplerates	in	the	
V
VI
V
order	{Mo132}<{W72Mo60}<{W132}.	These	results	also	indicate	that	W 	centres	are	less	(Lewis)	acidic	with	respect	to	W 	than	Mo 	in	relation	
VI
to	Mo .	

Conclusions	

	

	
DFT	 based	 calculations	 enabled	 unravelling	 the	 CO2	 hydration	 reaction	 pathway	 as	 evidenced	 involving	 Compound	 1	 by	
V
considering	 the	 known	 mechanism	 in	 the	 aqueous	 solution.	 The	 in	 situ	 bicarbonate	 formation,	 promoted	 by	 the	 Mo 	 centres,	
inside	the	capsule	is	excellent	and	less	energetically	demanding	than	direct	carbonate	uptake	from	aqueous	solution.	Three	trials	
were	performed	in	the	present	work,	which	can	be	summarised	as	follows:	
‡
i)	A	neutral	charge	pathway	with	an	aqua	ligand	nucleophilic	addition	to	CO2	results	in	a	high	kinetic	barrier	ΔE =	+81	kJ/mol	and	
a	product	of	exceedingly	high	energy.	
VI
ii)	A	hydroxo	ligand	pathway	in	which	the	nucleophilic	attack	takes	place	on	a	Mo 	site.	This	is	an	endergonic	process	requiring	
+46	kJ/mol	to	form	a	product.	
VI
iii)	 A	 hydroxo	 ligand	 pathway	 where	 the	 hydroxo	 group	 in	 an	 Mo 	 centre	 will	 act	 as	 a	 proton	 acceptor	 in	 tandem	 with	 the	
V
‡
nucleophilic	 addition	 of	 CO2	 to	 an	 aqua	 ligand	 at	 the	 Mo 	 sites.	 The	 activation	 energy	 ΔE =	 +36	 kJ/mol	 is	 the	 lowest	 of	 all	 the	
trials,	even	lower	than	the	uncatalysed	hydration	reaction,	and	the	ensuing	product	assembly	is	28	kJ/mol	more	stable	than	the	
reactant	assembly.		
Therefore	the	most	plausible	mechanism	for	the	formation	of	Compound	2	will	be	the	latter	based	on	comparison	of	computed	
energies	with	respect	to	a	comparable	micro-solvated	CO2	hydration.	The	resemblance	of	the	mechanism	with	that	operating	in	
the	 carbonic	 anhydrase	 enzyme	 is	 remarkable.	 The	 subtle	 differences	 lie	 in	 the	 first	 steps	 of	 the	 latter	 mechanism:	 the	 rate[3b,4]	
limiting	 step	 is	 the	 protolysis	 of	 the	 aqua	 ligand	 in	 (His)3Zn-OH2
which	 is	 then	 followed	 by	 a	 lower	 energy	 nucleophilic	
addition	to	CO2	whereas	the	Keplerate	acts	in	a	concerted	single	step	for	both.	These	results	pave	the	way	for	defining	a	new	
application	of	Keplerate	anionic	capsules	as	CO2	storage	nanodevices.	
	
Computational	Details	
	
31
The	Amsterdam	Density	Functional	(ADF)	program	package 	version	2012.01	has	been	used	throughout.	The	Perdew,	Burke	and	
32
Ernzerhof	 (PBE) 	 gradient	 corrected	 exchange	 and	 correlation	 functionals	 were	 used	 in	 the	 calculations.	 The	 choice	 of	 this	
33
functional	is	due	to	the	fairly	accurate	description	it	provides	of	hydrogen	bonds, 	an	aspect	which	is	crucial	to	this	work.	The	
34,	 35
36
ZORA
	 scalar	 relativistic	 Hamiltonian	 was	 employed	 with	 a	 triple	 zeta	 Slater	 type	 orbital 	 (STO)	 augmented	 with	 one	
polarization	function	(TZP)	for	molybdenum,	and	double	zeta	STO	type	functions	augmented	with	d	functions	on	the	remaining	
2
10
elements.	A	small	frozen	core	was	used	for	all	the	elements	(1s 	shell	for	O	and	C,	and	3d 	shell	for	Mo)	except	hydrogen.	The	
37
geometry	optimizations	were	performed	via	the	numerical	integration	scheme	of	Versluis	and	Ziegler. 	Stationary	points	were	
located	 with	 a	 5.0	 digit	 integration	 accuracy	 whereas	 partial	 and	 full	 analytic	 hessian	 calculations	 were	 done	 with	 7.0	 digit	
38
integration	 accuracy.	 The	 COSMO 	 implicit	 solvation	 scheme	 was	 used	 throughout.	 The	 imaginary	 frequencies	 related	 to	 the	
reaction	coordinate	were	followed	by	a	small	fraction	of	their	coordinate	displacement	in	either	direction	and	subsequently	reoptimized	 to	 achieve	 the	 reactant	 and	 product.	 Partial	 hessian	 calculations	 were	 performed	 on	 the	 reaction	 site,	 i.e.	 the	 eight	
atoms	present	in	the	ensemble	{CO2+OH2+OH }	for	2a,	2b,	TS2b,	and	2c.	For	the	remainder	of	the	steps	seven	atoms	{HCO3 +OH }	
were	 used	 to	 compute	 the	 partial	 hessian.	 Electronic	 and	 free	 energies	 of	 reactions	 leading	 to	 a	 formal	 loss	 of	 a	 proton	 were	
+
calculated	via	an	explicit	Brønsted	type	equilibrium	between	two	water	molecules	and	the	Zundel	cation	(H5O2 ).	

Notes	and	references	
This	work	was	funded	by	the	Spanish	Ministerio	de	Economia	y	Competitividad	(MINECO)	through	project	CTQ2014-52824-R,	by	the	
Generalitat	de	Catalunya	project	2014SGR409,	and	by	the	ICIQ	Foundation.	The	Severo	Ochoa	Excellence	Accreditation	(SEV-20130319)	 and	 the	 COST	 Action	 CM1203	 “Polyoxometalate	 Chemistry	 for	 Molecular	 Nanoscience	 (PoCheMoN)”	 are	 gratefully	
acknowledged.	 NAGB	 gratefully	 acknowledges	 COFUND/Marie	 Curie	 action	 291787-ICIQ-IPMP	 for	 funding.	 A.M.	 acknowledges	
continuous	financial	support	by	the	Deutsche	Forschungsgemeinschaft	and	the	ERC	(Brussels)	for	an	Advanced	Grant.	
	
	Keywords:	CO2	activation	•	biological	aspects	•	DFT	mechanism	•	Keplerate	•	polyoxometalates.	
	
	
1.	
2.	

M.	Aresta,	Carbon	Dioxide	as	Chemical	Feedstock,	Wiley,	2010.	
M.	E.	Boot-Handford,	J.	C.	Abanades,	E.	J.	Anthony,	M.	J.	Blunt,	S.	Brandani,	N.	Mac	Dowell,	J.	R.	Fernández,	M.-C.	Ferrari,	R.	Gross,	J.	P.	Hallett,	R.	S.	Haszeldine,	P.	
Heptonstall,	A.	Lyngfelt,	Z.	Makuch,	E.	Mangano,	R.	T.	J.	Porter,	M.	Pourkashanian,	G.	T.	Rochelle,	N.	Shah,	J.	G.	Yao	and	P.	S.	Fennell,	Energy	Environ.	Sci.,	
2014,	7,	130-189.	
3.	 J.	J.	Vericella,	S.	E.	Baker,	J.	K.	Stolaroff,	E.	B.	Duoss,	J.	O.	Hardin	IV,	J.	Lewicki,	E.	Glogowski,	W.	C.	Floyd,	C.	A.	Valdez,	W.	L.	Smith,	J.	H.	Satcher	Jr,	W.	L.	Bourcier,	C.	M.	
Spadaccini,	J.	A.	Lewis	and	R.	D.	Aines,	Nat.	Commun.,	2015,	DOI:	10.1038/ncomms7124.	
4.	 V.	Bhola,	F.	Swalaha,	R.	Ranjith	Kumar,	M.	Singh	and	F.	Bux,	Int.	J.	Environ.	Sci.	Technol.,	2014,	11,	2103-2118.	
5.	 A.	Sanna,	M.	Uibu,	G.	Caramanna,	R.	Kuusik	and	M.	M.	Maroto-Valer,	Chem.	Soc.	Rev.,	2014,	43,	8049-8080.	
6.	 K.	D’Ambrosio,	G.	De	Simone	and	C.	T.	Supuran,	in	Carbonic	Anhydrases	as	Biocatalysts,	eds.	G.	De	Simone	and	C.	T.	Supuran,	Elsevier,	Amsterdam,	2015,	ch.	2,	pp.	1730.	
7.	 D.	N.	Silverman	and	R.	McKenna,	Acc.	Chem.	Res.,	2007,	40,	669-675.	
8.	 C.	Greco,	V.	Fourmond,	C.	Baffert,	P.-h.	Wang,	S.	Dementin,	P.	Bertrand,	M.	Bruschi,	J.	Blumberger,	L.	de	Gioia	and	C.	Leger,	Energy	Environ.	Sci.,	2014,	7,	3543-3573.	
9.	 J.	K.	J.	Yong,	G.	W.	Stevens,	F.	Caruso	and	S.	E.	Kentish,	J.	Chem.Technol.	Biotechnol.,	2015,	90,	3-10.	
10.	 G.	Parkin,	Chem.	Rev.,	2004,	104,	699-768.	
11.	 A.	Goeppert,	M.	Czaun,	J.-P.	Jones,	G.	K.	Surya	Prakash	and	G.	A.	Olah,	Chem.	Soc.	Rev.,	2014,	43,	7995-8048.	
12.	 K.	R.	Roshan,	B.	M.	Kim,	A.	C.	Kathalikkattil,	J.	Tharun,	Y.	S.	Won	and	D.	W.	Park,	Chem.	Commun.,	2014,	50,	13664-13667.	

	

	
13.	
14.	
15.	
16.	
17.	
18.	
19.	
20.	
21.	
22.	
23.	
24.	
25.	
26.	
27.	
28.	
29.	
30.	
31.	

32.	
33.	
34.	
35.	
36.	
37.	
38.	
	
	

	

F.	Castro-Gómez,	G.	Salassa,	A.	W.	Kleij	and	C.	Bo,	Chem.-Eur.	J.,	2013,	19,	6289-6298.	
C.	J.	Whiteoak,	A.	H.	Henseler,	C.	Ayats,	A.	W.	Kleij	and	M.	A.	Pericàs,	Green	Chem.,	2014,	16,	1552-1559.	
Q.	Liu,	L.	Wu,	R.	Jackstell	and	M.	Beller,	Nat.	Commun.,	2015,	DOI:	10.1038/ncomms6933.	
A.	T.	Najafabadi,	Renew.	Sustain.	Energy	Rev.,	2015,	41,	1515-1545.	
S.	Garai,	E.	T.	K.	Haupt,	H.	Bögge,	A.	Merca	and	A.	Müller,	Angew.	Chem.	Int.	Ed.	Engl.,	2012,	51,	10528-10531.	
A.	Müller	and	P.	Gouzerh,	Chem.	Soc.	Rev.,	2012,	41,	7431-7463.	
T.	Mitra,	P.	Miró,	A.-R.	Tomsa,	A.	Merca,	H.	Bögge,	J.	B.	Ávalos,	J.	M.	Poblet,	C.	Bo	and	A.	Müller,	Chem.	Eur.	J.,	2009,	15,	1844-1852.	
C.	Schäffer,	A.	M.	Todea,	H.	Bögge,	O.	A.	Petina,	D.	Rehder,	E.	T.	K.	Haupt	and	A.	Müller,	Chem.	Eur.	J.,	2011,	17,	9634-9639.	
A.	Müller,	E.	Krickemeyer,	H.	Bögge,	M.	Schmidtmann	and	F.	Peters,	Angew.	Chem.	Int.	Ed.	Engl.,	1998,	37,	3360-3363.	
S.	Kopilevich,	A.	Gil,	M.	Garcia-Ratés,	J.	Bonet-Ávalos,	C.	Bo,	A.	Müller	and	I.	A.	Weinstock,	J.	Am.	Chem.	Soc.,	2012,	134,	13082-13088.	
C.	Besson,	S.	Schmitz,	K.	M.	Capella,	S.	Kopilevich,	I.	A.	Weinstock	and	P.	Kögerler,	Dalton	Trans.,	2012,	41,	9852-9854.	
C.	S.	Tautermann,	A.	F.	Voegele,	T.	Loerting,	I.	Kohl,	A.	Hallbrucker,	E.	Mayer	and	K.	R.	Liedl,	Chem.	Eur.	J.,	2002,	8,	66-73.	
M.	T.	Nguyen,	M.	H.	Matus,	V.	E.	Jackson,	V.	T.	Ngan,	J.	R.	Rustad	and	D.	A.	Dixon,	J.	Phys.	Chem.	A,	2008,	112,	10386-10398.	
S.	Yamabe	and	N.	Kawagishi,	Theor.	Chem.	Acc.,	2011,	130,	909-918.	
M.	J.	Welch,	J.	F.	Lifton	and	J.	A.	Seck,	J.	Phys.	Chem.,	1969,	73,	3351-3356.	
G.	A.	Gallet,	F.	Pietrucci	and	W.	Andreoni,	J.	Chem.	Theory	Comput.,	2012,	8,	4029-4039.	
I.	Mayer,	Int.	J.	Quantum	Chem.,	1986,	29,	73-84.	
C.	Schäffer,	A.	Merca,	H.	Bögge,	A.	M.	Todea,	M.	L.	Kistler,	T.	Liu,	R.	Thouvenot,	P.	Gouzerh	and	A.	Müller,	Angew.	Chem.	Int.	Ed.	Engl.,	2009,	48,	149-153.	
E.	J.	Baerends,	J.	Autschbach,	A.	Bérces,	C.	Bo,	P.	M.	Boerrigter,	L.	Cavallo,	D.	P.	Chong,	L.	Deng,	R.	M.	Dickson,	D.	E.	Ellis,	M.	van	Faassen,	L.	Fan,	T.	H.	Fischer,	C.	F.	
Guerra,	S.	J.	A.	van	Gisbergen,	J.	A.	Groeneveld,	O.	V.	Gritsenko,	M.	Grüning,	F.	E.	Harris,	P.	v.	d.	Hoek,	H.	Jacobsen,	L.	Jensen,	G.	van	Kessel,	F.	Kootstra,	E.	van	
Lenthe,	D.	A.	McCormack,	A.	Michalak,	V.	P.	Osinga,	S.	Patchkovskii,	P.	H.	T.	Philipsen,	D.	Post,	C.	C.	Pye,	W.	Ravenek,	P.	Ros,	P.	R.	T.	Schipper,	G.	
Schreckenbach,	J.	G.	Snijders,	M.	Sola,	M.	Swart,	D.	Swerhone,	G.	te	Velde,	P.	Vernooijs,	L.	Versluis,	O.	Visser,	F.	Wang,	E.	van	Wezenbeek,	G.	Wiesenekker,	S.	
K.	Wolff,	T.	K.	Woo,	A.	L.	Yakovlev	and	T.	Ziegler,	http://www.scm.com,	Scientific	Computing	and	Modelling	ADF-2012.01.	
J.	P.	Perdew,	K.	Burke	and	M.	Ernzerhof,	Phys.	Rev.	Lett.,	1996,	77,	3865-3868.	
I.	Hyla-Kryspin,	G.	Haufe	and	S.	Grimme,	Chem.	Eur.	J.,	2004,	10,	3411-3422.	
E.	van	Lenthe,	E.	J.	Baerends	and	J.	G.	Snijders,	J.	Chem.	Phys.,	1993,	99,	4597-4610.	
E.	van	Lenthe,	A.	E.	Ehlers	and	E.	J.	Baerends,	J.	Chem.	Phys.,	1999,	110,	8943-8953.	
E.	Van	Lenthe	and	E.	J.	Baerends,	J.	Comput.	Chem.,	2003,	24,	1142-1156.	
L.	Versluis	and	T.	Ziegler,	J.	Chem.	Phys.,	1988,	88,	322-328.	
A.	Klamt	and	G.	Schüürmann,	J.	Chem.	Soc.,	Perkin	Trans.	2,	1993,	799-805.