High capacity hydrogen adsorption in Cu(II) tetracarboxylate framework materials: The role of pore size, ligand functionalization, and exposed metal sites

Article abstract of DOI:10.1021/ja806624j

A series of isostructural metal-organic framework polymers of composition [Cu₂(L)(H₂O)₂](L= tetracarboxylate ligands), denoted NOTT-nnn, has been synthesized and characterized. Single crystal X-ray structures confirm the complexes to contain binuclear Cu(II) paddlewheel nodes each bridged by four carboxylate centers to give a NbO-type network of 6 ⁴? 8² topology. These complexes are activated by solvent exchange with acetone coupled to heating cycles under vacuum to afford the desolvated porous materials NOTT-100 to NOTT-109. These incorporate a vacant coordination site at each Cu(II) center and have large pore volumes that contribute to the observed high H₂ adsorption. Indeed, NOTT-103 at 77 K and 60 bar shows a very high total H₂ adsorption of 77.8 mg g ^-1 equivalent to 7.78 wt% [wt% = (weight of adsorbed H ₂)/(weight of host material)] or 7.22 wt% [wt% = 100(weight of adsorbed H₂)/(weight of host material + weight of adsorbed H ₂)]. Neutron powder diffraction studies on NOTT-101 reveal three adsorption sites for this material: at the exposed Cu(II) coordination site, at the pocket formed by three {Cu₂} paddle wheels, and at the cusp of three phenyl rings. Systematic virial analysis of the H₂ isotherms suggests that the H₂ binding energies at these sites are very similar and the differences are smaller than 1.0 kJ mol^-1, although the adsorption enthalpies for H₂ at the exposed Cu(II) site are significantly affected by pore metrics. Introducing methyl groups or using kinked ligands to create smaller pores can enhance the isosteric heat of adsorption and improve H₂ adsorption. However, although increasing the overlap of potential energy fields of pore walls increases the heat of H₂ adsorption at low pressure, it may be detrimental to the overall adsorption capacity by reducing the pore volume.

Full text of DOI:10.1021/ja806624j

Published on Web 01/21/2009
High Capacity Hydrogen Adsorption in Cu(II) Tetracarboxylate
Framework Materials: The Role of Pore Size, Ligand
Functionalization, and Exposed Metal Sites
Xiang Lin,† Irvin Telepeni,‡, Alexander J. Blake,† Anne Dailly,§ Craig M. Brown,|
Jason M. Simmons,| Marco Zoppi, Gavin S. Walker,‡ K. Mark Thomas,#
Timothy J. Mays, Peter Hubberstey,† Neil R. Champness,† and Martin Schro¨der*,†
School of Chemistry and School of Mechanical Materials & Manufacturing Engineering,
UniVersity of Nottingham, UniVersity Park, Nottingham NG7 2RD, U.K., Chemical and
EnVironmental Sciences Laboratory, General Motors Corporation, Warren, Michigan 48090,
National Institute of Standards and Technology Center for Neutron Research, 100 Bureau DriVe,
Building 235, STOP 6102, Gaithersburg, Maryland 20899, CNR-Instituto Sistemi Complessi, Via
della Madonna del Piano 10, I-50019 Sesto Fiorentino, Italy, Northern Carbon Research
Laboratories, Sir Joseph Swan Institute for Energy Research and School of Chemical Engineering
and AdVanced Materials, UniVersity of Newcastle upon Tyne, Newcastle upon Tyne NE1 7RU, U.K.,
and Department of Chemical Engineering, UniVersity of Bath, Bath BA2 7AY, U.K.
Received August 21, 2008; E-mail: M.Schroder@nottingham.ac.uk
Abstract: A series of isostructural metal-organic framework polymers of composition [Cu2(L)(H2O)2] (L)
tetracarboxylate ligands), denoted NOTT-nnn, has been synthesized and characterized. Single crystal X-ray
structures confirm the complexes to contain binuclear Cu(II) paddlewheel nodes each bridged by four
carboxylate centers to give a NbO-type network of 64 ·82 topology. These complexes are activated by solvent
exchange with acetone coupled to heating cycles under vacuum to afford the desolvated porous materials
NOTT-100 to NOTT-109. These incorporate a vacant coordination site at each Cu(II) center and have
large pore volumes that contribute to the observed high H2 adsorption. Indeed, NOTT-103 at 77 K and 60
bar shows a very high total H2 adsorption of 77.8 mg g-1 equivalent to 7.78 wt% [wt% ) (weight of adsorbed
H2)/(weight of host material)] or 7.22 wt% [wt% ) 100(weight of adsorbed H2)/(weight of host material +
weight of adsorbed H2)]. Neutron powder diffraction studies on NOTT-101 reveal three adsorption sites for
this material: at the exposed Cu(II) coordination site, at the pocket formed by three {Cu2} paddle wheels,
and at the cusp of three phenyl rings. Systematic virial analysis of the H2 isotherms suggests that the H2
binding energies at these sites are very similar and the differences are smaller than 1.0 kJ mol-1, although
the adsorption enthalpies for H2 at the exposed Cu(II) site are significantly affected by pore metrics.
Introducing methyl groups or using kinked ligands to create smaller pores can enhance the isosteric heat
of adsorption and improve H2 adsorption. However, although increasing the overlap of potential energy
fields of pore walls increases the heat of H2 adsorption at low pressure, it may be detrimental to the overall
adsorption capacity by reducing the pore volume.
desorption cycles.1,2 Adsorbents based upon carbon microporous
Introduction
8-24
materials3-7 and metal-organic frameworks
have been
Hydrogen (H2) is considered a most promising energy carrier
for vehicles and can potentially achieve zero CO2 emission at
the point of use. However a safe, economically and technically
viable solution for on-board H2 storage remains one of the major
challenges for any future “Hydrogen Economy”. Storage of H2
by physisorption on solid surfaces is an attractive option since
it provides fast kinetics and low heat effects during adsorption/
investigated widely. The latter are often crystalline and consist
(1) Hydrogen, Fuel Cells, & Infrastructure Technologies Program: Mul-
tiyear Research, Development, and Demonstration Plan (http://
www.eere.energy.gov/hydrogenandfuelcells/mypp/).
(2) U.S. Department of Energy., Energy Efficiency and Renewable Energy
(http://www.eere.energy.gov/hydrogenandfuelcells/pdfs/freedomcar-
_targets_explanations.pdf).
(3) Dillon, A. C.; Jones, K. M.; Bekkedahl, T. A.; Kiang, C. H.; Bethune,
D. S.; Heben, M. J. Nature 1997, 386, 377–379.
† School of Chemistry, University of Nottingham.
‡ School of Mechanical Materials & Manufacturing Engineering, Uni-
versity of Nottingham.
(4) Zhao, X. B.; Xiao, B.; Fletcher, A. J.; Thomas, K. M. J. Phys. Chem.
B 2005, 109, 8880–8888.
CNR-Instituto Sistemi Complessi.
(5) Schimmel, H. G.; Kearley, G. J.; Nijkamp, M. G.; Visser, C. T.; de
Jong, K. P.; Mulder, F. M. Chem.sEur. J. 2003, 9, 4764–4770.
(6) Zu¨ttel, A.; Sudan, P.; Mauron, P.; Kiyobayashi, T.; Emmenegger, C.;
Schlapbach, L. Int. J. Hydrogen Energy 2001, 27, 203–212.
(7) Yang, Z.; Xia, Y.; Mokaya, R. J. Am. Chem. Soc. 2007, 129, 1673–
1679.
§ General Motors Corporation.
| National Institute of Standards and Technology Center for Neutron
Research.
# University of Newcastle upon Tyne.
University of Bath.
9
10.1021/ja806624j CCC: $40.75
2009 American Chemical Society
J. AM. CHEM. SOC. 2009, 131, 2159–2171 2159

A R T I C L E S
Lin et al.
of extended networks assembled by the bonding of metal ions
and polyfunctional organic ligands. The development of syn-
thesis Via design over the past decade has produced examples
of metal-organic frameworks showing high porosity and
capability of adsorbing significant amounts of gases and small
molecules.24-42 These materials have been under intense
scrutiny due to their potential for high H2 adsorption capacity
and opportunities to tune and modify framework structures to
improve the properties of the resultant materials.
Physisorption of H2 in porous materials usually operates at
77 K due to the low adsorption heat for the process (typically
5-8 kJ mol-1 14,18,43-47). A higher isosteric heat of adsorption
of ∼15-20 kJ mol-1 is required for a H2 storage system
operating at room temperature and at a pressure range of 1.5-20
bar suitable for delivery directly to a fuel cell.48 Strategies to
enhance the H2 binding energy in porous materials include the
formation of narrow pores such that the overlapping potential
from two or more walls increases the binding energy between
H2 and the framework,47,49 provision of free metal coordination
sites to allow stronger binding of H2 directly to metal
nodes,14,20,46,50 and doping of frameworks with metal centers
and particles to increase H2 uptake Via a spillover mechanism.51
There is limited evidence that exposed metal sites increase the
overall H2 uptake capacities of porous metal-organic frame-
works, although binding enthalpies may well be increased by
direct interactions between H2 and metal centers.50,52-54 We
report here a comprehensive investigation of Cu(II) paddlewheel
complexes formed from a range of polyphenyl tetracarboxylate
ligands to yield materials with pore sizes ranging from 6.5 to
8.3 Å, the pore size being controlled and tuned by the
polyaromatic backbone.20 A range of different ligand spacers
(Scheme 1) has been introduced, and a series of new materials
with various pore sizes and shapes are obtained with different
functional groups projecting into the pores. Extensive charac-
terization of H2 adsorption in NOTT-101 was performed using
neutron powder diffraction, while detailed analysis of the H2
adsorption enthalpies for the series of porous complexes revealed
correlations between the exposed Cu(II) sites, organic ligands,
and the overall framework structures.
(8) Rowsell, J. L. C.; Millward, A. R.; Park, K. S.; Yaghi, O. M. J. Am.
Chem. Soc. 2004, 126, 5666–5667.
(9) Chen, B.; Ockwig, N. W.; Millward, A. R.; Contreras, D. S.; Yaghi,
O. M. Angew. Chem., Int. Ed. 2005, 44, 4745–4749.
(10) Rowsell, J. L. C.; Eckert, J.; Yaghi, O. M. J. Am. Chem. Soc. 2005,
127, 14904–14910.
(11) Furukawa, H.; Miller, M. A.; Yaghi, O. M. J. Mater. Chem. 2007,
17, 3197–3204.
(12) Wong-Foy, A. G.; Matzger, A. J.; Yaghi, O. M. J. Am. Chem. Soc.
2006, 128, 3494–3495.
(13) Rosi, N. L.; Eckert, J.; Eddaoudi, M.; Vodak, D. T.; Kim, J.; O’Keeffe,
M.; Yaghi, O. M. Science 2003, 300, 1127–1129.
(14) Dinca˘, M.; Long, J. R. J. Am. Chem. Soc. 2007, 129, 11172–11176.
(15) Kaye, S. S.; Dailly, A.; Yaghi, O. M.; Long, J. R. J. Am. Chem. Soc.
2007, 129, 14176–14177.
Experimental Section
(16) Dinca, M.; Dailly, A.; Liu, Y.; Brown, C. M.; Neumann, D. A.; Long,
J. R. J. Am. Chem. Soc. 2006, 128, 16876–16883.
All chemical reagents and gases were obtained from commercial
sources and, unless otherwise noted, were used without further
purification.
Ligand Synthesis. Preparation of Benzene-1,3-dicarboxy-
ethylester-5-boronic Acid. 1,3-Dimethyl-5-bromobenzene (100.0
g, 0.54 mol) in Et2O (800 mL) was added slowly over 2 h to Mg
(17) Kesanli, B.; Cui, Y.; Smith, M. R.; Bittner, E. W.; Bockrath, B. C.;
Lin, W. Angew. Chem., Int. Ed. 2005, 44, 72–75.
(18) Liu, Y.; Eubank, J. F.; Cairns, A. J.; Eckert, J.; Kravtsov, V. C.;
Luebke, R.; Eddaoudi, M. Angew. Chem., Int. Ed. 2007, 46, 3278–
3283.
(19) Zhao, X.; Xiao, B.; Fletcher, A. J.; Thomas, K. M.; Bradshaw, D.;
Rosseinsky, M. J. Science 2004, 306, 1012–1015.
(20) Lin, X.; Jia, J.; Zhao, X.; Thomas, K. M.; Blake, A. J.; Walker, G. S.;
Champness, N. R.; Hubberstey, P.; Schro¨der, M. Angew. Chem., Int.
Ed. 2006, 45, 7358–7364.
(38) Llewellyn, P. L.; Bourrrelly, S.; Serre, C.; Filinchuk, Y.; Fe´rey, G.
Angew. Chem., Int. Ed. 2006, 45, 7751–7754.
(39) Wang, X.-S.; Ma, S.; Forster, P. M.; Yuan, D.; Eckert, J.; Lopez, J. L.;
Murphy, B. J.; Parise, J. B.; Zhou, H.-C. Angew. Chem., Int. Ed. 2008,
47, 7263–7266.
(21) Jia, J.; Lin, X.; Wilson, C.; Blake, A. J.; Champness, N. R.; Hubberstey,
P.; Walker, G.; Cussen, E. J.; Schro¨der, M. Chem. Commun. 2007,
840–842.
(40) Lee, Y.-G.; Moon, H. R.; Cheon, Y. E.; Suh, M. P. Angew. Chem.,
Int. Ed. 2008, 47, 7741–7745.
(22) Lin, X.; Jia, J.; Hubberstey, P.; Schro¨der, M.; Champness, N. R.
CrystEngComm 2007, 9, 438–448.
(41) Wang, X.-S.; Ma, S.; Rauch, K.; Simmons, J. M.; Yuan, D.; Wang,
X.; Yildirim, T.; Cole, W. C.; Lopez, J. J.; Meijere, A.; Zhou, H.-C.
Chem. Mater. 2008, 20, 3145–3152.
(23) Fe´rey, G.; Latroche, M.; Serre, C.; Millange, F.; Loiseau, T.; Percheron-
Guegan, A. Chem. Commun. 2003, 2976–2977.
(24) Fe´rey, G. Science 2005, 310, 1119–1119.
(42) Xue, M.; Zhu, G.; Li, Y.; Zhao, X.; Jin, Z; Kang, E.; Qiu, S. Cryst.
Growth Des. 2008, 8, 2478–2483.
(25) Noro, S.-I; Kitagawa, S.; Kondo, M.; Seki, K. Angew. Chem., Int.
Ed. 2000, 39, 2082–2084.
(43) Rowsell, J. L. C.; Yaghi, O. M. J. Am. Chem. Soc. 2006, 128, 1304–
1315.
(26) Chen, B.; Eddaoudi, M.; Hyde, S. T.; O‘Keeffe, M.; Yaghi, O. M.
Science 2001, 291, 1021–1023.
(44) Kaye, S. S.; Long, J. R. J. Am. Chem. Soc. 2005, 127, 6506–6507.
(45) Dinca˘, M.; Long, J. R. J. Am. Chem. Soc. 2005, 127, 9376–9377.
(46) Dinca˘, M.; Han, W. S.; Liu, Y.; Dailly, A.; Brown, C. M.; Long, J. R.
Angew. Chem., Int. Ed. 2007, 46, 1419–1422.
(27) Eddaoudi, M.; Kim, J.; Rosi, N.; Vodak, D.; Wachter, J.; O’Keeffe,
M.; Yaghi, O. M. Science 2002, 295, 469–472.
(28) Kitaura, R.; Seki, K.; Akiyama, G.; Kitagawa, S. Angew. Chem., Int.
Ed. 2003, 42, 428–431.
(47) Yang, W.; Lin, X.; Jia, J.; Blake, A. J.; Wilson, C.; Hubberstey, P.;
Champness, N. R.; Schr¨oder, M. Chem. Commun. 2008, 359–361. Liu,
Y.; Kabbour, H.; Brown, C. M.; Neumann, D. A.; Ahn, C. C. Langmuir
2008, 24, 4772–77.
(29) Dueren, T.; Sarkisov, L.; Yaghi, O. M.; Snurr, R. Q. Langmuir 2004,
20, 2683–2689.
(30) Bourrelly, S.; Llewellyn, P. L.; Serre, C.; Millange, F.; Loiseau, T.;
Fe´rey, G. J. Am. Chem. Soc. 2005, 127, 13519–13521.
(31) Ma, S.; Sun, D.; Wang, X.-S.; Zhou, H.-C. Angew. Chem., Int. Ed.
2007, 46, 2458–2462.
(48) Bhatia, S. K.; Myers, A. L. Langmuir 2006, 22, 1688–1700.
(49) Chun, H.; Dybtsev, D. N.; Kim, H.; Kim, K. Chem.sEur. J. 2005,
11, 3521–3529.
(32) Banerjee, R.; Phan, A.; Wang, B.; Knobler, C.; Furukawa, H.;
O‘Keeffe, M.; Yaghi, O. M. Science 2008, 319, 939–943.
(33) Fletcher, A. J.; Cussen, E. J.; Bradshaw, D.; Rosseinsky, M. J.; Thomas,
K. M. J. Am. Chem. Soc. 2004, 126, 9750–9759.
(50) Dinca˘, M.; Dailly, A.; Liu, Y.; Brown, C. M.; Neumann, D. A.; Long,
J. R. J. Am. Chem. Soc. 2006, 128, 16876–83.
(51) Li, Y.; Yang, R. T. J. Am. Chem. Soc. 2006, 128, 726–727. See
also: Mulfort, K. L.; Hupp, J. T. J. Am. Chem. Soc. 2007, 129, 9604–
9605.
(34) Cussen, E. J.; Claridge, J. B.; Rosseinsky, M. J.; Kepert, C. J. J. Am.
Chem. Soc. 2002, 124, 9574–9581.
(52) Liu, Y.; Brown, C. M.; Neumann, D. A.; Peterson, V. K.; Kepert,
C. J. J. Alloys Compd. 2007, 446, 385–388.
(35) Fletcher, A. J.; Cussen, E. J.; Prior, T. J.; Rosseinsky, M. J.; Kepert,
C. J.; Thomas, K. M. J. Am. Chem. Soc. 2001, 123, 10001–10011.
(36) Lin, X.; Blake, A. J.; Wilson, C.; Sun, X. Z.; Champness, N. R.;
George, M. W.; Hubberstey, P.; Mokaya, R.; Schro¨der, M. J. Am.
Chem. Soc. 2006, 128, 10745–10753.
(53) Prestipino, C.; Regli, L.; Vitillo, J. G.; Bonino, F.; Damin, A.;
Lamberti, C.; Zecchina, A.; Solari, P. L.; Kongshaug, K. O.; Bordiga,
S. Chem. Mater. 2006, 18, 1337–1346.
(54) Forster, P. M.; Eckert, J.; Heiken, B. D.; Parise, J. B.; Yoon, J. W.;
Jhung, S. H.; Chang, J.-S.; Cheetham, A. K. J. Am. Chem. Soc. 2006,
128, 16846–16850.
(37) Ma, S.; Sun, D.; Simmons, J. M.; Collier, C. D.; Yuan, D.; Zhou,
H.-C. J. Am. Chem. Soc. 2008, 130, 1012–1016.
9
2160 J. AM. CHEM. SOC. VOL. 131, NO. 6, 2009

H2 Adsorption in Cu(II) Tetracarboxylate Materials
A R T I C L E S
Scheme 1. Synthesis of Ligands and Their Corresponding Cu Complexes [Cu(L)2]a
a
t
(i) Mg, Et2O, B(OMe)3; (ii) BuOH/H2O, KMnO4; (iii) EtOH, H2SO4; (iv) BrRBr, Pd(PPh3)4, 80 °C, 3 days; (v) 2 M NaOH, HCl.
turnings (16.0 g, 0.66 mol) with vigorous stirring under an Ar
atmosphere in the presence of 2 drops of iodomethane. The addition
of 1,3-dimethyl-5-bromobenzene was controlled so that the reaction
mixture refluxed gently without heating, and after the addition was
complete, the reaction mixture was refluxed for 6 h. Under vigorous
stirring, the resulting Grignard solution was added dropwise into a
solution of trimethyl borate (56.2 g, 0.54 mol) in Et2O (500 mL)
cooled to -60 °C using a dry ice-acetone bath. The mixture was
then allowed to warm to room temperature slowly with stirring
and carefully poured into 20% H2SO4 (500 mL) and stirred until
the two phases were clear. The Et2O layer was separated and the
water phase extracted with further aliquots of Et2O (3 portions of
100 mL), after which the ether phases were combined. The Et2O
solution was washed with water (3 portions of 100 mL) and dried
over MgSO4, and the solvent was removed by rotary evaporation
to afford a pale yellow solid (yield: 65.1 g, 82%). The crude product
of 1,3-dimethylbenzene-5-boronic acid was used in the following
reaction without further purification.
1,3-Dimethylbenzene-5-boronic acid (10.0 g, 0.068 mol) and
NaOH (5.0 g, 0.125 mol) were dissolved in tert-butanol/water (v/v
) 1:1; 500 mL). The reaction mixture was heated to 50 °C with
stirring, and small portions (ca. 1 g) of KMnO4 (totalling 65.0 g,
0.41 mol) were added to the solution once the purple color of the
previous portion had faded. After 50 g of KMnO4 had been added,
the reaction temperature was raised to 70 °C and addition of KMnO4
continued until the purple color persisted for over 3 h, indicating
full oxidation of the methyl groups. The excess KMnO4 was reduced
by addition of Na2S2O3 (1.00 g) and the solution filtered when hot.
The MnO2 cake was washed with excess boiling water (1 L) and
the aqueous filtrate concentrated to ∼150 mL by evaporation and
acidified to pH ) 1 using concentrated HCl. The white precipitate
of benzene-1,3-dicarboxylic-5-boronic triacid was collected by
filtration and washed with cold water and dried in an oven at 100
°C. The dried white solid was mixed with 98% H2SO4 (5 mL) in
anhydrous EtOH (250 mL), and the solution refluxed for 12 h. The
resulting solution was concentrated by evaporation to a volume of
∼20 mL, whereupon water (250 mL) was added to give the desired
product benzene-1,3-dicarboxyethylester-5-boronic acid as a white
precipitate which was collected by filtration and washed with water
and dried. Yield: 12.9 g, 71.3%. Anal. Calcd (Found) for
BC12O6H15: C, 54.17 (54.09); H, 5.68 (5.74) %. 1H NMR (CDCl3,
300 MHz), 8.53 (t, J ) 1.5 Hz); 8.27(d, J ) 1.5 Hz); 4.34(dd, J1
) 7.2 Hz, J2 ) 6.3 Hz); 1.35 (t, J ) 8.4 Hz). MS (ESI) m/z (M -
H+)-: 265.0871.
Preparation of Ligands. H4L1 was prepared by a modified
literature procedure55 with improved yield: 3,5,3,5-tetramethyl
biphenyl (10.0 g, 0.047 mol) was oxidized in tert-butanol/water
(v/v ) 1:1; 500 mL) containing NaOH (4.0 g, 0.1 mol) using the
above oxidation procedure. Yield: 10.9 g, 70.6%. Anal. Calcd
(Found) for C16O8H10: C, 58.19 (58.12); H, 3.05 (3.11)%. 1H NMR
(MeOD, 300 MHz): 8.64 (1H, t, J ) 1.8 Hz); 8.49 (2H, s); 4.39
(4H, q, J ) 6.9 Hz), 1.43 (6H, t, J ) 7.2 Hz).
The ligands H4L2 to H4L10 were synthesized using analogous
methodologies and experimental procedures. The synthesis of H4L4
is described in detail.
Preparation of H4L4. 2,6-Dibromonaphthalene (0.286 g, 1.0
mmol), benzene-1,3-dicarboxyethylester-5-boronic acid (0.64 g, 2.4
mmol), and K3PO4 (2.10 g, 10.0 mmol) were added to 1,4-dioxane
(30 mL), and the mixture deaerated under N2. Pd(PPh3)4 (0.05 g,
0.043 mmol) was added to the reaction mixture with stirring, and
the mixture heated to 80 °C for 3 days under N2. The resultant
mixture was evaporated to dryness and taken up in CHCl3 which
had been dried over MgSO4. The CHCl3 solution was evaporated
to dryness and the residue washed briefly with EtOH (10 mL). The
resulting crude product (mainly tetra-ethyl esters of the target ligand)
was hydrolyzed by refluxing the crude product in 2 M aqueous
NaOH followed by acidification with 37% HCl to afford H4L. The
crude product was recrystallized from DMF/H2O. Yield: 0.28 g,
1
65.2%. H NMR (DMSO-d6, 300 MHz), H4L4: 8.56 (d, 2H, J )
3.2 Hz), 8.50 (dd, 1H, J1 ) 13.92 Hz, J2 ) 3.0), 8.42 (s, 1H), 8.24
(d, 1H, J ) 6.1 Hz), 7.96 (d, 1H, J ) 12.7 Hz). Anal. Calcd (Found)
for H4L4, C26H16O8: C, 68.42(68.03); H, 3.53(3.79)%. MS (ESI)
m/z (M - H+)-: 455.0752.
H4L2. 1H NMR (DMSO-d6, 300 MHz): 8.72 (t, 1H, J ) 6.3 Hz),
8.55 (d, 2H, J ) 1.4), 7.86 (s, 2H). Anal. Calcd (Found) for H4L2
(C22H14O8): C, 65.03 (65.23); H, 3.47 (3.40). MS (ESI) m/z (M -
H+)-: 405.0594.
H4L3. 1H NMR (DMSO-d6, 300 MHz): 8.49 (t, 1H, J ) 1.8 Hz),
8.44 (d, 2H, J ) 1.5 Hz), 7.92 (d, 4H, J ) 0.9 Hz). Anal. Calcd
(Found) for H4L3 (C28H18O8): C, 69.71 (69.52); H, 3.76(3.79)%.
MS (ESI) m/z (M - H+)-: 481.0920.
(55) Coles, S. J.; Holmes, R.; Hursthouse, M. B.; Price, D. J. Acta
Crystallogr., Sect. E 2002, 58, o626–o628.
9
J. AM. CHEM. SOC. VOL. 131, NO. 6, 2009 2161

A R T I C L E S
Lin et al.
1
H4L5. H NMR (DMSO-d6 300 MHz), H4L4: 8.56 (d, 2H, J )
solution was heated without stirring in an 85 °C oil bath for 16 h.
The blue crystalline product was separated by filtration when the
solution was still warm (ca. 50 °C) and washed with warm DMF
and dried briefly in air. Yield: 0.929 g (72.6%). Anal. Calcd (Found)
for C42.5H54.5Cu2N1.5O19.5: C, 49.78 (49.60); H, 5.36 (4.56); N, 2.05
(1.58%). Selected IR(KBr): ν (cm-1) ) 1616(s), 1558(s), 1505(s),
1442(s), 1384(vs), 1261(w), 1020(w), 824(w), 767(m), 730(m),
669(m), 630(w).
Preparation of NOTT-103, [Cu2(C26O8H12)(H2O)2]·3DMF·
3H2O. Cu(NO3)2 ·2.5H2O (2.62 g, 10 mmol) was dissolved in DMF
(200 mL) containing H4L4 (1.30 g, 2.85 mmol) to afford a blue
solution. Dioxane (50 mL) and distilled water (100 mL) were added,
and a small amount of blue precipitate formed at this point. On
addition of HCl (37%, 0.6 mL) the reaction mixture turned back
to a clear blue-green solution, which was heated without stirring
in an 85 °C oil bath for 16 h. The product was separated by filtration
and washed with warm DMF to give the complex (1.74 g, 69.0%
yield based on H4L4) as a blue crystalline product. Anal. Calcd
(Found) for C35H43Cu2N3O16: C, 47.34 (48.71); H, 4.88 (5.14); N,
4.74 (4.42)%. Selected IR(KBr): ν (cm-1) ) 1618(s), 1502(vs),
1410(s), 1366(vs), 1295(m), 1152(m), 1107(m), 823(s), 746(s),
685(s).
Preparation of NOTT-104, [Cu4(C34O8H18)2(H2O)4]·DMF·
3H2O. Cu(NO3)2 ·2.5H2O (0.30 g, 1.15 mmol) was dissolved in
DMF (40 mL) containing H4L5 (0.18 g, 0.32 mol) to afford a blue
solution. To the solution were added distilled water (20 mL), an
aqueous solution of LiCl (0.1 mL, 0.1 M), and 1 drop of HCl (37%).
The resulting green solution was heated to 85 °C in a programmed
oil bath at a heating rate of 0.1 °C/min and kept at 85 °C for 24 h.
The green crystalline product was separated by filtration to give
0.15 g of product (Yield: 55.5%). Anal. Calcd (Found) for
C74H70Cu4N2O28: C, 52.66 (54.09); H, 4.18 (4.09); N, 1.66 (1.92)%.
Selected IR(KBr): ν (cm-1) ) 1645(s), 1622(s), 1561(s), 1493(m),
1438(s), 1359(vs), 1220(s), 1101(m), 1003(w), 914(w), 838(s),
770(vs), 689(s), 670(s), 637(m).
Preparation of NOTT-105, [Cu2(C22O8H8F2)(H2O)2]·2DMF·
C4O2H8 ·4H2O. Cu(NO3)2 ·2.5H2O (0.50 g, 1.92 mmol) was dis-
solved in DMF (40 mL) containing H4L6 (0.280 g, 0.50 mmol) to
afford a blue solution. Distilled water (20 mL) was added together
with HCl (37%, 0.05 mL). The resulting green solution was heated
to 85 °C in a programmed oil bath with a heating rate of 0.1 °C
min-1 and kept at 85 °C. After 24 h a green crystalline product
(Yield: 0.369 g, 82.3%) was separated by filtration. Anal. Calcd
(Found) for C32H42Cu2F2N2O18: C, 42.38 (42.79); H, 4.67 (4.74);
N, 3.09 (2.88)%. Selected IR(KBr): ν (cm-1) ) 1667(s), 1562(s),
1510(s), 1369(vs), 1221(m), 1172(m), 1111(m), 884(m), 775(s),
729(vs), 685(m), 668(m), 619(m), 601(m).
Preparation of NOTT-106, [Cu2(C24H14O8)(H2O)2]·0.25DMF ·
H2O. Cu(NO3)2 ·2.5H2O (0.50 g, 1.92 mmol) was dissolved DMF
(40 mL) containing H4L7 (0.220 g, 0.50 mmol) to afford a blue
solution. To the solution were added distilled water (20 mL) and
HCl (37%, 0.05 mL). The resultant green solution was heated to
85 °C in a programmed oil bath with a heating rate of 0.1 °C min-1
and kept at 85 °C. After 24 h, a green crystalline product (Yield:
0.203 g, 64.5%) was separated by filtration. Anal. Calcd (Found)
for C24.75H21.75Cu2N0.25O11.25: C, 47.28 (47.09); H, 3.49 (3.64); N,
0.56 (0.70)%. Selected IR(KBr): ν (cm-1) ) 1644(s), 1632(s),
1561(s), 1423(s), 1362(vs), 1221(s), 1108(m), 1090(m), 922(w),
889(w), 774(s), 729(s), 691(w), 642(m), 625(w).
Preparation of NOTT-107, [Cu2(C26H20O8)(H2O)2]·0.5DMF ·
H2O. The procedure used for the preparation of NOTT-104 was
employed, but only a microcrystalline product was obtained. The
crystals were too small for single crystal diffraction, but the powder
X-ray diffraction pattern could be indexed using the unit cell for
NOTT-106, which confirmed NOTT-107 as isostructural with
NOTT-106. Anal. Calcd (Found): C, 48.70 (49.02); H, 4.38 (4.32);
N, 1.03 (1.26)%. Selected IR(KBr): ν (cm-1) ) 1626(vs), 1578(s),
1497(w), 1432(s), 1367(vs), 1297(m), 1253(m), 1098(m), 1060(w),
920(w), 773(s), 728(s), 696(m), 662(s).
3.2 Hz), 8.50 (dd, 1H, J1 ) 13.92 Hz, J2 ) 3.0), 7.6-7.8 (m, 6H).
Anal. Calcd (Found) for H4L5 (C34H22O8): C, 73.11 (73.67); H,
3.97(4.08)%. MS (ESI) m/z (M - H+)-: 557.1235.
H4L6. 1H NMR (DMSO-d6 300 MHz): 8.67 (t, 1H, J ) 6.2 Hz),
8.38 (d, 2H, J ) 1.8 Hz), 7.29 (t, 1H, J ) 8.1 Hz). Anal. Calcd
(Found) for H4L6 (C22H12F2O8): C, 59.74 (59.26); H, 2.73(2.64)%.
MS (ESI) m/z (M - H+)-: 422.0506.
H4L7. 1H NMR (DMSO-d6 300 MHz): 8.71 (t, 1H, J ) 3.6 Hz),
8.26 (d, 2H, J ) 1.6 Hz), 7.21 (s, 1H), 2.29 (s, 3H). Anal. Calcd
(Found) for H4L7, C24H18O8: C, 66.36 (66.10); H, 4.18(4.32)%. MS
(ESI) m/z (M - H+)-: 433.0916.
1
H4L8. H NMR (DMSO-d6 300 MHz), H4L4: 8.72 (t, 1H, J )
3.6 Hz), 8.09 (d, 2H, J ) 1.7 Hz), 1.94 (s, 6H). Anal. Calcd (Found)
for H4L8, (C26H22O8): C, 67.23 (67.44); H, 5.21 (5.34)%. MS (ESI)
m/z (M - H+)-: 461.1223.
H4L9. 1H NMR (DMSO-d6 300 MHz): 8.87 (t, 1H, J ) 3.0 Hz),
8.57 (d, 2H, J ) 0.6 Hz). Anal. Calcd (Found) for H4L9
(C22H10F4O8): C, 55.24 (55.17); H, 2.11 (1.94)%. MS (ESI) m/z
(M - H+)-: 477.0230.
H4L10. H NMR (DMSO-d6 300 MHz): 8.67 (t, 1H, J ) 3.0
1
Hz), 8.31 (d, 2H, J ) 0.6 Hz), 7.91 (q, 1H, J ) 3.2 Hz), 7.48 (s,
1H), 7.45 (q, 1H, J ) 3.0 Hz). Anal. Calcd (Found) for H4L10,
C26H16O8: C, 68.42 (68.77); H, 3.53 (3.20)%. MS (ESI) m/z (M -
H+)-: 455.0758.
Preparation of Metal Complexes. Most materials were obtained
in microcrystalline form and good phase purity via the same
experimental procedure as described below. By using a programmed
heating rate (0.1 °C min-1) and a prolonged heating time (3 days)
at 85 °C, large high-quality crystals suitable for single crystal X-ray
diffraction can be produced. However, this route also produces a
solid byproduct which comprises dark-red Cu2O and other green
amorphous copper-containing materials. Samples were solvent-
exchanged with acetone, which was repeated 5 times over 2 days,
and the sample was kept under acetone. For gas adsorption and
neutron powder diffraction experiments, the acetone-exchanged
material was desolvated at 120 °C under dynamic vacuum to afford
porous materials.
Preparation of NOTT-100, [Cu2(C16H6O8)(H2O)2]·2.5DMF ·
4H2O. H4L1 (1.65 g, 5.00 mmol) and Cu(NO3)2 ·2.5H2O (5.00 g,
10.0 mmol) were mixed and dispersed in DMF/H2O (350 mL, 2:1
v/v). HCl (37%, 1.0 mL) was added to the mixture and mixed
thoroughly. The solution was heated without stirring in an 85 °C
oil bath for 24 h, and a large amount of microcrystalline product
precipitated. The blue crystalline product was separated by filtration
when the solution was still warm (ca. 50 °C), washed with warm
DMF, and dried briefly in air. Yield: 2.540 g (69.8%). Anal. Calcd
(Found) for C23.5H35.5Cu2N2.5O15.5: C, 38.81 (39.02); H, 4.92 (4.88);
N, 4.82 (5.05)%. Selected IR(KBr): ν (cm-1) ) 1634(vs), 1588(s),
1552(w), 1418(s), 1404(m), 1365(vs), 1319(m), 1237(m), 1222(s),
1082(m), 936(w), 906(m), 827(w), 771(s), 731(vs), 678(m), 664(s).
Preparation of NOTT-101, [Cu2(C22H10O8)(H2O)2]·1.5DMF ·
2.5(C4O2H8) · H2O. H4L2 (0.50 g, 1.23 mmol) and Cu(NO3)2 ·
2.5H2O (1.00 g, 4.30 mmol) were mixed and dispersed in DMF/
1,4-dioxane/H2O (200 mL, 2:1:1 v/v/v). The resulting blue-green
slurry turned clear upon addition of HCl (37%, 0.3 mL). The
solution was heated without stirring in an 85 °C oil bath for 16 h.
The blue crystalline product was separated by filtration when the
solution was still warm (ca. 50 °C) and washed with warm DMF
and dried briefly in air. Yield: 0.777 g (69.2%). Anal. Calcd (Found)
for C36.50H46.5Cu2N1.5O17.5: C, 47.99 (48.52); H 5.13 (5.03); N, 2.30
(2.07)%. Selected IR(KBr): ν (cm-1) ) 1618(vs), 1566(vs), 1451(s),
1406(vs), 1384(vs), 1303(m), 1261(m), 1111(w), 1080(w), 838(m),
772(m), 729(m).
Preparation of NOTT-102, [Cu2(C28O8H14)(H2O)2] ·
1.5(DMF)·2.5(C4O2H8)·3(H2O). H4L3 (0.60 g, 1.25 mmol) and
Cu(NO3)2 ·2.5H2O (1.00 g, 4.30 mmol) were mixed and dispersed
in DMF/1,4-dioxane/H2O (200 mL, 2:1:1 v/v/v). The resulting blue-
green slurry turned clear upon addition of HCl (37%, 0.3 mL). The
9
2162 J. AM. CHEM. SOC. VOL. 131, NO. 6, 2009

H2 Adsorption in Cu(II) Tetracarboxylate Materials
A R T I C L E S
Table 1. Summary Data for the Crystal Structures
NOTT-103
NOTT-104
NOTT-105
NOTT-106
NOTT-108
NOTT-109
formula
C35H43Cu2N3O16
888.82
C74H70Cu4N2O28
1689.53
C32H42Cu2F2N2O18
907.77
C24.75H21.75Cu2N0.25O11.25
629.77
C25H19Cu2F4NO12
728.50
C44H66Cu2N6O20
1126.11
formula mass
(g mol-1
)
j
j
j
j
j
space group
a, b (Å)
c (Å)
R3m
R3m
R3m
R3m
R3m
I4/mmm
18.852(3)
27.417(6)
9744(3)
8
1.535
0.959
4720
18.513(1)
45.354(5)
13462(4)
9
0.987
0.760
18.348(1)
65.094(2)
18978(2)
9
1.337
1.075
18.594(2)
38.671(4)
11579(3)
9
1.172
0.892
18.504(2)
39.035(5)
11575(4)
9
0.813
0.858
18.525(1)
38.830(2)
11540(2)
9
0.943
0.879
V (Å3)
Z
Dcalcd (g cm-3
)
µ (mm-1
F(000)
)
4140
7812
4211
2880
3293
total/unique
reflections
Rint
49 853/3327
47 634/2454
23 279/3251
22 575/3814
54 877/3521
17 010/1386
0.059
0.113
0.067
0.069
0.074
0.081
R1, wR2 (%)
goodness-of-fit
difference
11.67, 33.64
1.02
0.87, -0.33
13.22, 40.23
1.22
0.57, -0.69
3.68, 9.17
0.91
0.49, - 0.50
6.51, 19.57
0.98
0.63, -0.54
5.89, 17.21
1.04
0.76, -0.56
6.82, 19.49
1.09
0.62, -0.52
Fourier map,
max, min (e Å-3
)
Preparation of NOTT-108, [Cu2(C22O8H6F4)(H2O)2]·DMF·
H2O. The preparation of NOTT-108 does not require the addition
of HCl. Cu(NO3)2 ·2.5H2O (0.30 g, 1.15 mmol) was dissolved in
DMF (40 mL) containing H4L9 (0.15 g, 0.31 mmol) to afford a
blue solution. To the solution was added distilled water (20 mL),
and the resulting mixture was heated at 85 °C for 24 h. Only a few
crystals separated from the reaction mixture, and most of the ligand
remained unreacted. Single crystals suitable for X-ray structural
analysis were obtained, but it proved impossible to isolate NOTT-
108 as a pure, single-phase, bulk material for gas adsorption
experiments. Selected IR(KBr): ν (cm-1) ) 1623(vs), 1487(s),
1376(s), 1329(m), 1164(vs), 1061(m), 982(s), 920(m), 895(m),
789(s), 697(m), 669(vs), 605(m).
Preparation of NOTT-109, [Cu2(C26O8H12)(H2O)2]·6DMF·
4H2O. Cu(NO3)2 ·2.5H2O (0.50 g, 1.92 mmol) was dissolved in
DMF (40 mL) containing H4L10 (0.228 g, 0.50 mmol) to afford a
blue solution. To the solution were added distilled water (20 mL)
and HCl (37%, 0.05 mL). The resulting green solution was heated
to 85 °C in a programmed oil bath with a heating rate of 0.1 °C
min-1, and kept at 85 °C. After 24 h, a green crystalline product
(Yield: 0.438 g, 77.8%) was separated by filtration. Anal. Calcd
(Found) for C44H66Cu2N6O20: C, 46.93 (43.06); H, 5.91 (6.62); N,
7.46 (7.74)%. Selected IR(KBr): ν (cm-1) ) 1645(s), 1557(vs),
1450(s), 1370(vs), 1275(m), 1228(s), 1113(w), 1091(w), 917(m),
843(w), 774(s), 752(s), 727(vs), 662(m).
Single Crystal X-ray Diffraction Studies. Single crystal dif-
fraction data on NOTT-103 and NOTT-105 were collected at 150(2)
K on a Bruker SMART APEX CCD area detector using graphite-
monochromated Mo KR radiation. Data for NOTT-104, -106, -108,
and -109 were collected at station 9.8 or station 16.2SMX at the
Daresbury Synchrotron Radiation Source (Table 1). Data for NOTT-
100, -101, and -102 have been reported previously.20 The details
for data collection are included in the CIF file in the Supporting
Information. Structures were solved by direct methods and devel-
oped by difference Fourier techniques using the SHELXTL software
package.56 Hydrogen atoms on the ligands were placed geo-
metrically and refined using a riding model, and the hydrogen atoms
of the coordinated water molecules could not be located but are
included in the formula. The unit cell volume includes a large region
of disordered solvent which could not be modeled as discrete atomic
sites. We employed PLATON/SQUEEZE57 to calculate the con-
tribution to the diffraction from the solvent region and thereby
produced a set of solvent-free diffraction intensities. The final
formula was calculated from the SQUEEZE results combined with
elemental analysis data.
Nitrogen, Hydrogen, and Deuterium Isotherms. Nitrogen and
hydrogen isotherms were determined at 78 K using an IGA
gravimetric adsorption apparatus at the University of Nottingham,
in a clean ultrahigh vacuum system with a diaphragm and turbo
pumping system. Ultrapure plus grade (99.9995%) H2 and high-
purity (99.95%) D2 were purified further using calcium alumino-
silicate and activated carbon adsorbents to remove trace amounts
of water and other impurities before introduction into the IGA
system. The densities of desolvated samples used in buoyancy
corrections were determined by running helium isotherms at room
temperature, and the measured densities of NOTT-100, NOTT-
101, NOTT-102, and NOTT-103 were determined as 1.75, 1.70,
1.63, and 1.65 g cm-3, respectively. The use of different sample
densities results in very small differences (<0.1% in all cases) in
the buoyancy correction at 20 bar. Therefore, we used 1.75 g cm-3
for the buoyancy correction for all other materials. The density of
H2 at its boiling point (0.0708 g cm-3) was used for the adsorbate
buoyancy correction. The density of bulk H2 in the buoyancy
correction was calculated by the Redlich-Kwong-Soave equation
of state of H2 which is built into the software IGASWIN in the
IGA system.
High Pressure Hydrogen Isotherms. Volumetric H2 sorption
measurements were performed at General Motors over a pressure
range of 0-60 bar using an automatically controlled Sievert’s
apparatus. The mass of adsorbent that was used to determine the
adsorption/desorption isotherms ranged from 150 to 500 mg. All
measurements were made with ultrahigh purity grade (99.999%)
H2 and He, the latter used for dead volume measurements. The
volumetric measurements were carried out at 77 K by submerging
the sample holder in a liquid nitrogen bath, the level of which was
maintained constant throughout each experiment. Our volumetric
analysis can be viewed as a precise application of the real gas law
taking into account deviation from nonideal behavior found at high
pressures and/or low temperatures. The nonideal gas behavior is
taken into account by determining the H2 compressibility factor,
Z, from National Institute of Standards and Technology (NIST)
data.58 The volume of the sample holder was determined using
helium at 298 K and H2 at 298 and 77 K. H2 was used to determine
the dead space volume correction for a nonporous inert insert of a
known geometrical volume (typically steel or aluminum); this
correction accounts for the change in effective sample volume
observed when cooling the sample holder from room temperature
to 77 K. Also, helium was used to determine the volume occupied
(58) Lemmon, E. W. Peskin, A. P. McLinden, M. O. Friend, D. G. NIST12
Thermodynamic and Transport Properties of Pure Fluids - NIST
Standard Reference Database 23, Version 8.0; U.S. Secretary of
Commerce: Washington, DC, 2007.
(56) Sheldrick, G. M. Acta Crystallogr., Sect. A 2008, 64, 112–122.
(57) Spek, A. L. J. Appl. Crystallogr. 2003, 36, 7–13.
9
J. AM. CHEM. SOC. VOL. 131, NO. 6, 2009 2163

A R T I C L E S
Lin et al.
Figure 1. Views of (a) the node structures in NOTT-101; (b) simplified 4-connected node; (c) NbO type framework assembled by two 4-connected nodes;
(d and e) structures of the two types of cages.
by a sample in the sample holder at 298 K because of its negligibly
small adsorption on solid surfaces.
and 6 ligands, and the [Cu12(L)12] cage comprises 6 {Cu2}
paddle-wheel units and 12 ligands.
Neutron Powder Diffraction. Neutron powder diffraction (NPD)
experiments were performed on NOTT-101 as a function of D2
loadings using the high resolution powder diffractometer BT1 at
the National Institute of Standards and Technology Center for
Neutron Research (NCNR) in Gaithersburg, MD, USA. Prior to
the NPD experiments the sample was thermally activated at 140
°C for 20 h to remove the solvent from the nanopores and the water
molecules from the Cu(II) sites leading to unsaturated metal
coordination sites. A dehydrated sample (1.7451 g) was loaded into
a vanadium cell with D2 dosed at 50 K. All data collection occurred
with the sample at 4 K and using a Ge(311) monochromator with
a 75° takeoff angle, λ ) 2.0787(2) Å, and in-pile collimation of
15 min of arc. Data were collected over the range of 1.3-166.3°
2θ with a step size of 0.05° on samples containing D2:Cu ratios of
0.46, 0.91, 1.82, and 2.73. The NPD data were analyzed by Rietveld
refinement and difference Fourier map analysis as implemented
within GSAS-EXPGUI.59,60
It was observed that changing the length of the ligand in these
porous materials can dramatically influence the size and shape
of the cages within the structure. In NOTT-100, the two kinds
of cage are approximately spherical (Figure 2a) with the
[Cu12(L)12] moiety having a diameter of 8.0 Å (after considering
the van der Waals radii of the atoms), while the diameter of
the [Cu24(L)6] cage is 10.0 Å. The largest cavities were found
in NOTT-102 in which the [Cu12(L)12] cage has a spherical
cavity of diameter 13.4 Å, while the [Cu24(L)6] cage in NOTT-
102 is elongated with the same 10.0 Å diameter in the equatorial
plane as for NOTT-100, but with a diameter along the c axis of
32.0 Å. Figure 2 illustrates how the size and shape of these
cages evolve with different lengths of ligands.
There are two types of window at the interconnections of
the [Cu24(L)6] and [Cu12(L)12] cages, and the shape of these
windows can be described by considering the connections
between Cu centers. The first type of window (TYPE-I) is
defined by the equilateral triangular channels running along the
c axis (Figure 3a), which contains three {Cu2} paddle wheels
bridged by carboxylate-phenyl-carboxylate spans. For NOTT-
100, the distance between two {Cu2} paddle-wheel centers is
7.91 Å, and the diameter of the windows is ∼5.0 Å. In all
structures, the size of the TYPE-I window is very similar. The
other type of window (TYPE-II) is based upon isosceles
triangles which can be viewed along the a or b axes (Figure
3b). The length of the two equal sides of the isosceles triangles
is defined by the length of ligands. In NOTT-100, this gives a
Cu...Cu separation within the window of 7.60 Å, and in NOTT-
102 the distance is increased significantly to 15.96 Å. The
TYPE-II windows are also defined by the functional groups on
the spacers. Thus, introduction of substituted phenyl groups as
in H4L6, H4L7, H4L8, and H4L9 (Scheme 1) narrows the TYPE-
II window, but the TYPE-I window is not affected by this
substitution. When we consider the potential energy overlaps
inside the cavity, the smaller the window the more efficient the
overlaps will be. This is confirmed by gas adsorption results
discussed below.
Results and Discussion
Single Crystal X-ray Structures. The frameworks NOTT-
100-NOTT-108 are isostructural but differ due to the lengths
and nature of the bridging ligands employed. In all these crystal
structures, two Cu(II) cations are bridged by four carboxylate
groups to form a [Cu2(RCO2)4] paddle-wheel configuration with
four carboxylate groups distributed around the Cu-Cu axis thus
defining a square-planar 4-connected node. The ligands are also
square-planar 4-connected nodes since the four carboxylate
groups on the isophthalate in all ligands lie in the same plane
(Figure 1a and 1b). The assembly of the two types of planar
4-connected nodes results in NbO type networks of 64 ·82
topology with the complexes having the general formula
[Cu2(L)] (Figure 1a). The NbO network can be viewed as the
packing and combination of two types of metal-ligand cages
having the same molecular D3d symmetry (Figure 1c and 1d)
with stoichiometries of [Cu24(L)6] and [Cu12(L)12] (Figure 1e).
The [Cu24(L)6] node comprises 12 {Cu2} paddle-wheel units
(59) Larson, A. C.; von Dreele, R. B. General Structure Analysis System
(GSAS), Los Alamos National Laboratory Report LAUR 1994, 86–
748.
NOTT-104 has the same network topology as the other
systems discussed herein, except that the extended nature of
(60) Toby, B. H. J. Appl. Crystallogr. 2001, 34, 210–213.
9
2164 J. AM. CHEM. SOC. VOL. 131, NO. 6, 2009

H2 Adsorption in Cu(II) Tetracarboxylate Materials
A R T I C L E S
Figure 2. Views of the internal cages, the dimensions of which are tuned by the ligands: (a) NOTT-100; (b) NOTT-101; (c) NOTT-102. Top row: [Cu24(L)6]
cages. Bottom row: [Cu12(L)12] cages.
Figure 4. View of the structure of NOTT-104: (top) the individual [Cu2(L5)]
network; (bottom) double interpenetrating frameworks.
NOTT-109 are constructed by four {Cu2} paddle wheels as
opposed to three in the above examples. The large cage in
NOTT-109 has D4h symmetry and an approximate spherical
Figure 3. Space-filling views of some [Cu2(L)] frameworks, in which water
molecules have been removed from Cu(II) sites to show the two types of
window in the frameworks: (a) view along the c axis; (b) view along the a
axis.
cavity diameter of 13 Å with a stoichiometry for the cage of
[Cu16(L10)16] (Figure 5b). The smaller cage is comprised of
[Cu16(L10)4] moieties (Figure 5a) which have a channel of 4 Å
in diameter. A change of framework symmetry in the [Cu2(L)]
material PCN-14 has also been reported recently upon incor-
poration of 5,5′-(9,10-anthracenediyl) diisophthalate moieties
into the bridging ligand positions.37 Although the framework
topology of PCN-14 (64 ·82) is the same as that of NOTT-
the pentaphenyl ligand leads to a doubly interpenetrated network
structure (Figure 4). This interpenetration leads to reduced
porosity and limited thermal stability (discussed below), making
NOTT-104 unsuitable for sorption applications.
NOTT-109 has a different network topology than the other
systems described above and crystallizes in the tetragonal space
group I4/mmm. The building units are also based upon 4-con-
nected {Cu2} paddle-wheel centers and 4-connected tetracar-
boxylate ligands with the resultant 4,4-connected network having
a 42 · 84 or PtS topology. However, due to the bulky central
aromatic groups in the [L10]4- group, the triangular TYPE-I
windows cannot be formed and the expanded windows in
j
100-NOTT-108, it crystallizes in the trigonal space group R3c
with the bulky anthracene group shaping two types of cuboc-
taheral and nanoscopic cages within the framework.37
Structural Thermal Stability. With the exception of NOTT-
104, all of the framework complexes were stable to the removal
ofsolventandcoordinatedwater.FrameworksNOTT-100-NOTT-
103 and NOTT-105-NOTT-109 showed similar behavior by
9
J. AM. CHEM. SOC. VOL. 131, NO. 6, 2009 2165

A R T I C L E S
Lin et al.
Figure 5. Views of the structure of NOTT-109: (a) [Cu16(L10)4] cage; (b) [Cu16(L10)16] cage; (c) overall framework structure.
Table 2. Physical Data and Gas Adsorption Data for [Cu2(L)]∞ Frameworks
compound
NOTT-100
NOTT-101
NOTT-102
NOTT-103
NOTT-105
NOTT-106
NOTT-107
NOTT-109
BET surface areaa (m2 g-1
pore diameterb (Å)
)
1640
6.5
63.3
0.680
0.683
2316
7.3
70.4
0.886
1.083
2942
8.3
75.5
1.138
1.284
2929
8.0
71.6
1.142
1.120
2387
7.3
68.2
0.898
0.934
1855
7.3
64.0
0.798
0.889
1822
7.0
n/a
0.767
n/a
1718
6.9
68.6
0.705
0.868
calcd free spacec (%)
pore volumea (cm3 g-1
)
pore volumec (cm3 g-1
)
total H2 adsorption in wt% 2.59/4.02/- 2.52/6.06/6.60 2.24/6.07/7.20 2.63/6.51/7.78 2.52/5.40/- 2.29/4.50/- 2.26/4.46/- 2.33/4.15/-
(1bar/20bar/60bar)d,e
total H2 adsorption in wt% 2.52/3.86/- 2.46/5.71/6.19 2.19/5.72/6.72 2.56/6.11/7.22 2.46/5.12/- 2.24/4.31/- 2.21/4.27/- 2.28/3.98/-
(1bar/20bar/60bar)d,f
a Calculated from N2 isotherms. b Pore diameters estimated from Dubinin-Astakhov analysis. c Calculated from single crystal structures with
PLATON/VOID.57 d 1 bar and 20 bar H2 adsorption data were obtained by gravimetric method, and 60 bar data were obtained by volumetric method.
e Hydrogen uptake quoted in terms of wt% most commonly defined in the literature as 100(weight of adsorbed H2)/(weight of host). f Hydrogen uptake
quoted as a wt% in terms of 100(weight of adsorbed H2)/(weight of host + weight of H2 adsorbed).15 See Supporting Information for further details of
this analysis and methods of presentation and calculation of wt%.
TGA with rapid loss of solvent under a N2 atmosphere below
100 °C and high stability up to 300 °C, above which the
materials start to decompose. Under a dry air atmosphere, the
decomposition temperature is ∼250 °C which corresponds to
the combustion temperature of the organic ligand. In situ powder
X-ray diffraction indicates that the solvent-free frameworks
retain their original structure and unit cell (see Supporting
Information). In the case of NOTT-104, the framework is very
unstable after removal of solvent by heat or vacuum and
degassing at room temperature which can turn the crystalline
NOTT-104 into an amorphous phase. This is an interesting
observation given that interpenetrated networks might be
expected to show higher structural stability than noninterpen-
etrated analogues. Because of this framework instability, NOTT-
104 will be excluded from all subsequent discussions.
two methyl groups (NOTT-106) results in a reduction of pore
volume, BET surface area, and pore size; four methyl groups
(NOTT-107) reduce the pore volume and size further. Introduc-
ing fluorine substituents, as in NOTT-105, seems to have little
impact on the BET surface area, pore volume, and size measured
by N2 isotherms compared with NOTT-101. This is not
surprising because the pore geometry changes only slightly on
replacement of H with fluorine substituents on the aromatic ring.
A methyl group, however, is significantly more bulky than a
fluorine atom and therefore reduces the pore volume and BET
surface area.
Hydrogen Adsorption. Gravimetric measurements obtained
using the IGA system for H2 adsorption on framework materials
are given in Figure 6. At 78 K and 20 bar, the highest total
uptake was observed for NOTT-103 with 65.1 mg g-1 H2. From
NOTT-100 to NOTT-102, the increasing ligand length results
in an increase in BET surface area and pore volume, and this
accounts for the increase in overall H2 capacities on going from
NOTT-100 to NOTT-102. Thus, the increasing pore volume
provides more space to accommodate H2; however, the large
pore size in NOTT-102 gives a weaker overlap potential from
the walls, and a higher pressure is thus required to store the
same amount of H2 in the pores compared to NOTT-100 and
NOTT-101. This can be clearly identified from the adsorption
data at 1 bar. Thus, increasing the pore size from 6.5 Å in
NOTT-100 to 7.3 Å in NOTT-101 leads to a small reduction
of total H2 uptake at 1 bar from 25.9 to 25.2 mg g-1, while
NOTT-102 with the largest pore size of 8.3 Å adsorbs only
22.4 mg g-1 H2, the lowest value of all samples at 1 bar.
Although NOTT-102 has a much larger pore size and volume
than NOTT-101, its H2 uptake does not exceed NOTT-101 until
the pressure reaches 20 bar. This demonstrates that larger pore
Nitrogen Isotherms. The desolvated framework structures
show very high porosity. Table 2 summarizes BET surface areas,
pore volumes, and pore sizes estimated from N2 adsorption
experiments at 77 K. The framework NOTT-100 based on the
shortest ligand [L1]4- shows the lowest BET surface area, pore
volume, and pore size. The highest pore volumes and BET
surface areas were observed for NOTT-102 and NOTT-103, the
former incorporating [L3]4- which is the longest ligand used in
these porous frameworks. Interestingly, however, the BET
surface area and pore volume of NOTT-102 and NOTT-103
are very similar. Only the pore size estimated from Dubinin-
Astakhov (D-A) analysis revealed that NOTT-102 has a larger
pore size than NOTT-103, in agreement with their crystal
structures.
[L6]4-, [L7]4-, and [L8]4- are derived from [L2]4- (Scheme
1) and incorporate electron-withdrawing fluorine or electron-
releasing methyl groups on the central phenyl ring. Introducing
9
2166 J. AM. CHEM. SOC. VOL. 131, NO. 6, 2009

H2 Adsorption in Cu(II) Tetracarboxylate Materials
A R T I C L E S
Figure 7. Excess (points) and total (lines) hydrogen uptakes for NOTT-
101 (blue), NOTT-102 (cyan), and NOTT-103 (purple).
Figure 6. Total hydrogen adsorption at 78 K.
size materials require higher pressure to confine the H2 into the
channels and suggests that there may well be an optimum pore
size and dimension to maximize H2 uptake. Based upon our
observations this optimum pore size lies between 7.3 and 8.3
Å.20 This is further supported by the successful synthesis of
NOTT-103 which has a pore size of 8.0 Å and its H2 adsorption
at 20 bar and 77 K is the highest in the series of complexes. So
far only two metal-organic framework compounds (MOF-5 and
MOF-177) are known to have such high H2 uptake at 20 bar.11,15
More interestingly, at 1 bar NOTT-103 adsorbed 26.3 mg g-1
H2, higher than any other framework showing a total H2 uptake
of more than 6 wt% at 20 bar. Thus, at 1 bar MOF-58 adsorbs
14.2 mg g-1, MOF-17711 13.2 mg g-1, Mn-BTT14 22.8 mg g-1,
and Cu-BTT46 23.0 mg g-1. This suggests that, in addition to
using optimum ligand lengths to form intermediate pore size
materials, the shape and nature of the central functional group
(i.e., the napthyl in NOTT-103) can give a stronger potential
overlap to improve adsorption of H2 at low pressures.
Sievert’s apparatus) and the gravimetric method (using the IGA
system) are in very good agreement. The total H2 adsorption
data in Figure 7 show that NOTT-101 is nearly saturated at 30
bar. The larger pores of NOTT-102 reach saturation at higher
pressures, 40 bar, and the higher pore volume allows the total
uptake of NOTT-102 to reach 72.0 mg g-1 at 60 bar. These
experiments confirm the importance of high pore volume for
high H2 adsorption and are in line with previous theoretical
modeling analyses.61,62 NOTT-103 has almost the same pore
volume and BET surface as NOTT-102 according to their
N2 isotherms, but the total H2 adsorption on NOTT-103 is
5.8 mg g-1 higher at 60 bar, reaching 77.8 mg g-1. This derives
from and correlates with the higher adsorption heat of NOTT-
103 (see below).
Using an extended To´th equation63 (eq 1) to analyze the high
pressure excess H2 isotherms at 77 K, we can estimate the pore
volume V (and hence maximum total uptake) independently of
N2 data.
NOTT-105, -106, and -107 incorporate fluorine or methyl
substituents on the central phenyl rings of the ligand and are
isostructural with NOTT-101. Such modifications afford materi-
als with smaller pore volume and pore size than NOTT-101
due to the additional atoms occupying space within the
framework, thus creating a higher overlap of potential curves
and stronger H2-framework interactions. The stronger interac-
tions lead to increased uptake at low coverage and are observed
below 0.5 bar for NOTT-106, -107 and below 1 bar for NOTT-
105 although the enhancements are small (<1 mg g-1) compared
to NOTT-101. Although modification to the ligands does not
improve the overall H2 adsorption capacity at higher pressure
due to the reduced surface area, it does give increased
H2-framework interaction potentials, which can be seen in the
analysis of H2 adsorption enthalpies (see below).
The high pressure H2 adsorption measurements revealed that
NOTT-103 has the highest excess and total uptake of all the
materials studied. The total uptake is given by nT ) nE + FBV
where nE is the experimental excess uptake, FB is the bulk
density of H2, and V is the pore volume which at this stage we
take to be the volume obtained from the 78 K N2 isotherms
(see Supporting Information). We used the Redlich-Kwong-
Soave equation of state to calculate FB. Up to 20 bar excess H2
adsorptions obtained by the volumetric method (using the
mbp
nE )
-
c 1⁄c
[
]
1 + (bp)
1.489 × 10-7
V - 1.838 × 10-5
p +
(1)
(
)
(
-5
)
RT
(
)
V V + 1.838 × 10
In eq 1, m is the maximum total gas uptake, b is the Henry
constant, and c is a constant describing the surface energetic
heterogeneity. This is a development of an earlier, less flexible
approach63,64 where the Langmuir isotherm was used for total
adsorption. The second term on the right-hand side of eq 1 treats
bulk H2 with the Redlich-Kwong-Soave equation of state.
Fitting the excess high pressure isotherm data with eq 1 where
V (as well as m, b, and c) is taken as an adjustable parameter
gives pore volumes of 0.91, 1.37, and 0.97 cm3 g-1 for NOTT-
101, NOTT-102, and NOTT-103, respectively. The maximum
total uptakes, m, estimated from eq 1 are 75.0, 98.5, and 88.2
mg g-1 for NOTT-101, NOTT-102, and NOTT-103, respec-
(61) Frost, H.; Snurr, R. Q. J. Phys. Chem. C 2007, 111, 18794–18803.
(62) Frost, H.; Du¨ren, T.; Snurr, R. Q. J. Phys. Chem. B 2006, 110, 9565–
9570.
(63) To´th, J. Adsorption: Theory, Modeling, and Analysis; CRC Press/
Marcel Dekker: New York, 2002; pp 1-104.
(64) Myers, A. L.; Monson, P. A. Langmuir 2002, 18, 10261–10273.
9
J. AM. CHEM. SOC. VOL. 131, NO. 6, 2009 2167

A R T I C L E S
Lin et al.
Figure 8. (a) NPD patterns of NOTT-101 with different D2 loading at 4 K. (b) Hydrogen adsorption sites in NOTT-101 from Rietveld refinements.
tively. These results are reasonable as NOTT-102 has the longest
ligand and the crystal structure confirms that NOTT-102 has
the largest void. The To´th analysis shows the effect of the length
of ligand on the maximum uptakes, although such effects cannot
be observed directly from H2 isotherms up to 60 bar. However,
the To´th analysis also suggests that the maximum total uptakes
appear at pressures above 5000 bar (the To´th equation saturates
quite slowly with increasing pressure), which is of course not
practical for any application.
The hydrogen storage capacities for the NOTT series
complexes are also very impressive in volumetric units and are
among the highest obtained for physisorption materials. At 60
bar and 77 K the total volumetric capacities are 43.1 g L-1 for
NOTT-101, 42.3 g L-1 for NOTT-102, and 50.0 g L-1 for
NOTT-103, the latter being comparable to MOF-177.11 Only
MOF-5, when prepared and handled under anhydrous condi-
tions,15 shows a higher volumetric storage capacity (60 g L-1
at 60 bar and 77 K).
Neutron Powder Diffraction. Neutron powder diffraction was
used to determine and define the adsorption sites in the
framework NOTT-101. The structure of the unloaded material
was refined, and this was used as the starting host structure for
refinement of data from subsequent loading sequences. The
refinement of the unloaded and pore empty sample was
characterized by a ꢀ2 ) 1.198, which is in good agreement with
the single crystal analysis. We neither observe any major
changes to cell parameters nor detect any extra features in the
NPD patterns, which suggest that there is no structural phase
change during the experiment (Figure 8a). The differential
nuclear scattering density Fourier map of the first deuterium
loading (0.46 D2 per Cu, equivalent to 0.33 wt% H2 adsorption)
confirms unequivocally that the exposed Cu(II) sites are the first
site in the framework where the D2 molecules bind with a
distance of Cu-D2 (centroid) of 2.50(3) Å slightly longer than
that observed in HKUST-1 (2.40 Å),65 but clearly not of the
“Kubas” type σ-bond binding.66 The Fourier map has a fairly
smooth background which suggests that the majority of the D2
molecules bind to the exposed Cu(II) sites (Site I in Figure 8b).
Similar behavior has previously been observed in the case of
HKUST-1 using a similar experimental procedure.65 At higher
loadings (0.91 and 1.82 D2 per Cu), we were able to identify
both a second and third adsorption site. One is located in the
middle of the TYPE-1 window, while the other is in the cusp
of three phenyl rings. Both of these sites incorporate a 3-fold
symmetry axis (Site II and III in Figure 8b). The distances
between pairs of sites (Site I · · ·Site II ) 3.8 Å, Site III · · ·Site
II ) 3.8 Å, Site I · · ·Site III ) 7.04 Å) are physically acceptable
and consistent with the minimum distance allowed between two
D2 molecules (3.4
Å
in solid D2).67 At 1.82
D2/Cu loading, the occupancy of D2 at the Cu site increases to
0.85 suggesting that D2 is almost saturating the Cu sites, but
the Cu-D2 binding energy is not significantly higher than that
for Site II and III, so D2 occupies both Site II and III before
Site I is fully saturated. We also observe higher residues in the
spherical [Cu12(L)12] cage than in the ellipsoid-shaped [Cu24(L)6]
cavity. In conclusion, at 4 K and at low loading of D2, most D2
molecules populate at the top and bottom of [Cu24(L)6] cavity
around the TYPE-I window. As shown in Figure 8, these
adsorption sites are (i) the exposed Cu(II) sites; (ii) the small
TYPE-I window formed by three {Cu2} paddle wheels and
ligands; and (iii) the cusp formed by the phenyl rings. Interest-
ingly, the binding energy of D2 at the Cu(II) site appears not to
be significantly higher than that at the other two identified sites.
Isosteric Heat of Hydrogen Adsorption. H2 isotherms up to
20 bar were recorded by the gravimetric method at 78 and 88
K and from 0-20 bar, and the data were analyzed using virial
methods. Virial Method 168-72 is based upon
ln(n ⁄ p) ) A0 + A1n + A2n2 + ...
(2)
where p is the pressure, n is total amount adsorbed and A0, A1,
etc. are virial coefficients. A0 is related to adsorbate-adsorbent
(67) Silvera, I. F. ReV. Mod. Phys. 1980, 52, 393–452.
(68) Cole, J. H.; Everett, D. H.; Marshall, C. T.; Paniego, A. R.; Powl,
J. C.; Rodriguez-Reinoso, F. J. Chem. Soc., Faraday Trans. 1974,
70, 2154–2169.
(69) O’Koye, I. P.; Benham, M.; Thomas, K. M. Langmuir 1997, 13, 4054–
4059.
(65) Peterson, V. K.; Liu, Y.; Brown, C. M.; Kepert, C. J. J. Am. Chem.
Soc. 2006, 128, 15578–15579.
(70) Reid, C. R.; O’Koye, I. P.; Thomas, K. M. Langmuir 1998, 14, 2415–
2425.
(66) Kubas, G. J. Acc. Chem. Res. 1988, 21, 120–128.
(71) Reid, C. R.; Thomas, K. M. Langmuir 1999, 15, 3206–3218.
9
2168 J. AM. CHEM. SOC. VOL. 131, NO. 6, 2009

H2 Adsorption in Cu(II) Tetracarboxylate Materials
A R T I C L E S
Table 3. Comparison of Virial Parameters (Method 1) for Adsorption of H2 on the Series of Compounds NOTT-100 through NOTT-107
uptake range
(mol g-1
compound
T/K
KH/mol g-1 Pa-1
A0/ln (mol g-1 Pa-1
)
A1/g mol-1
)
NOTT-100
78
88
78
88
78
88
78
88
78
88
78
88
78
88
9.051 × 10-7
2.920 × 10-7
4.626 × 10-7
1.748 × 10-7
5.164 × 10-7
1.978 × 10-7
5.288 × 10-7
1.898 × 10-7
5.330 × 10-7
1.892 × 10-7
9.297 × 10-7
2.976 × 10-7
8.114 × 10-7
2.436 × 10-7
-13.915 ( 0.007
-15.046 ( 0.007
-14.586 ( 0.004
-15.560 ( 0.006
-14.476 ( 0.013
-15.436 ( 0.008
-14.453 ( 0.017
-15.478 ( 0.005
-14.445 ( 0.008
-15.481 ( 0.003
-13.888 ( 0.024
-15.028 ( 0.014
-14.024 ( 0.013
-15.228 ( 0.002
-150.232 ( 1.038
-130.865 ( 1.201
-111.895 ( 0.639
-91.072 ( 0.858
-150.258 ( 1.750
-132.938 ( 1.167
-118.001 ( 2.873
-104.978 ( 1.001
-113.257 ( 1.601
-100.113 ( 0.721
-191.032 ( 4.082
-171.138 ( 2.648
-169.735 ( 2.193
-140.602 ( 0.385
<0.008
<0.008
<0.008
<0.008
<0.009
<0.009
<0.007
<0.007
<0.007
<0.006
<0.008
<0.007
<0.007
<0.006
NOTT-101
NOTT-102
NOTT-103
NOTT-105
NOTT-106
NOTT-107
interactions, whereas A1 describes adsorbate-adsorbate interac-
tions. Thus, the Henry’s Law constant is given by KH ) exp(A0).
At low surface coverage A2 and higher terms can be neglected.
has A0 and A1 values of -13.416 ln(mol g-1 Pa-1) and -227.7
g mol-1, and this compound has a similar pore size to that of
the NOTT series. In contrast, the values for the materials in
this study are lower than the values for the ethanol-templated
phase of [Ni2(bpy)3(NO3)4] (A0 ) -12.931 ln(mol g-1 Pa-1);
A1 ) -1032.8 g mol-1)35 which has a narrower porosity. Figure
9a illustrates the relationship of adsorption enthalpies to H2
uptake at low coverage: the values at zero coverage are
extrapolated from the linear equation (eq 2).
76-78
In Virial Method 2
m
n
1
T
ln(p) ) ln(n) +
a ni+ b n j
(3)
∑
∑
i
j
i)0
j)0
where p is pressure, n is amount adsorbed, T is temperature,
and ai and bj are temperature independent empirical parameters.
The isosteric heat of adsorption (Qst) as a function of H2 uptake
can be obtained over a wide loading range from the following
equation:
Comparisons of NOTT-100 with NOTT-101, -102, and -103
can give us an insight to the effects of pore size and the presence
of additional aromatic rings on the enthalpy of H2 binding. As
shown in Figure 9a, the adsorption heat of NOTT-100 at zero
coverage is 0.9 kJ mol-1 higher than those of NOTT-101 and
NOTT-102, and the difference is the same in most loading
ranges. Pore size appears to affect the adsorption heat signifi-
cantly, even at zero surface coverage when the exposed Cu(II)
coordination sites are readily accessible to H2 molecules. The
[Cu12(L)12] and [Cu24(L)6] cavities in NOTT-100 have ap-
proximate spherical diameters of 8.0 and 10.0 Å, respectively.
Cavities in this size range can potentially provide more effective
profile overlap than the corresponding larger cavities in NOTT-
101 and NOTT-102. The spherical cavities [Cu12(L)12] in NOTT-
101 and NOTT-102 have diameters of 12.7 and 13.4 Å,
respectively, and the long axes in the ellipsoidal [Cu24(L)6]
cavities of NOTT-101 and NOTT-102 are 19.6 and 32.0 Å,
which are too long to allow the potential profiles from opposite
ends to overlap efficiently. The adsorption enthalpies obtained
are, therefore, strongly correlated to the local structures, i.e.,
the binding energies of the three adsorption sites which are
located at the bottom of the ellipsoid [Cu24(L)6] cavity as
revealed by NPD results. For NOTT-101 and NOTT-102, the
local structures are very similar. As shown in Figure 9a, the
adsorption enthalpies of NOTT-101 and NOTT-102 at zero
loading are essentially the same taking into account the
uncertainties in the measurement, which is consistent with the
similarity in their local structure, and their different pore sizes
do not noticeably affect the adsorption heat at low H2 loading.
This also suggests that the introduction of phenyl (NOTT-101)
or biphenyl groups (NOTT-102) and the significant difference
(19.6 vs 32.0 Å) in the shape of their ellipsoid [Cu24(L)6] cavities
has very little impact on the adsorption heat at low coverage.
The adsorption heat of NOTT-101 becomes evidently higher
than that of NOTT-102 at higher H2 loading (Figure 9b). Such
a difference (0.5 kJ mol-1 at 0.025 mol g-1 H2 loading)
corresponds to the filling of the middle of the cavities and the
m
Q ) -R a ni
(4)
∑
st
i
i)0
Virial method 1 can be readily used to estimate of the isosteric
heat of adsorption at zero coverage by applying the van’t Hoff
isochore to the Henry constant, KH, while Virial Method 2 allows
a series of Qst values to be obtained as a function of n. Minimal
deviations are observed for all experimental points measured
over a wide range of pressures and temperatures.
Table 3 lists the virial parameters obtained by fitting the low
coverage H2 data for porous NOTT materials using virial
Method 1. At very low coverage, pore volume does not limit
adsorption. Therefore, a less negative A0 and larger KH suggest
a stronger H2-framework interaction and a more negative A1
suggests a stronger H2-H2 interaction inside the framework.
The A1 values in Table 3 are less negative than those observed
for H2 adsorption on carbons (e.g., -1063.9 and -962.5 g mol-1
for carbons73 and -588.3 and -315.9 g mol-1 for oxidized
activated carbons4) and M’MOF (-1069 g mol-1 74 due to the
)
larger pore size in the NOTT series of materials. HKUST-175
(72) Reid, C. R.; Thomas, K. M. J. Phys. Chem. B 2001, 105, 10619–
10629.
(73) Zhao, X.; Villar-Rodil, S.; Fletcher, A. J.; Thomas, K. M. J. Phys.
Chem. B 2006, 110, 9947–9955.
(74) Chen, B.; Zhao, X.; Putkham, A.; Hong, K.; Lobkovsky, E. B.;
Hurtado, E. J.; Fletcher, A. J.; Thomas, K. M. J. Am. Chem. Soc.
2008, 130, 6411–6423.
(75) Xiao, B.; Wheatley, P. S.; Zhao, X.; Fletcher, A. J.; Fox, S.; Rossi,
A. G.; Megson, I. L.; Bordiga, S.; Regli, L.; Thomas, K. M.; Morris,
R. E. J. Am. Chem. Soc. 2007, 129, 1203–1209.
(76) Czepirski, L.; Jagiello, J. Chem. Eng. Sci. 1989, 44, 797–801.
(77) Jagiello, J.; Bandosz, T. J.; Schwarz, J. A. Langmuir 1996, 12, 2837–
2842.
(78) Anson, A.; Jagiello, J.; Parra, J. B.; Sanjuan, M. L.; Benito, A. M.;
Maser, W. K.; Martinez, M. T. J. Phys. Chem. B 2004, 108, 15820–
15826.
9
J. AM. CHEM. SOC. VOL. 131, NO. 6, 2009 2169

A R T I C L E S
Lin et al.
smaller with the increasing H2 loading, and above 0.0075 mol
g-1, their adsorption enthalpies converge (Figure 9b). The
adsorption enthalpies of NOTT-105, NOTT-106, and NOTT-
107 decrease at the same rate from the loading of 0.0075 mol
g-1 indicating that H2-pore interactions at higher loading are
similar and not affected significantly by the introduction of
fluorine or methyl substituents. At high loading, H2-H2 interac-
tions may also contribute to changes in enthalpies of adsorption.
Compared to NOTT-100, NOTT-101, and NOTT-102, NOTT-
103 has improved performance in both heat of adsorption and
total H2 uptake. The heat of adsorption for NOTT-103 lies in
the range 5.1-4.0 kJ mol-1 between 0.0075 and 0.029 mol g-1
loading, with the initial heat of 5.71 kJ mol-1 at zero loading
approaching that observed for the narrower porosity of NOTT-
105. Interestingly in the high loading region, the adsorption heat
of NOTT-103 decreases more slowly than that of NOTT-105,
NOTT-106, and NOTT-107, which have smaller pores. NOTT-
103, therefore, incorporates a high surface area and pore volume
and provides excellent H2 uptake at all pressures by enhancing
both heat of adsorption and pore volume within the same
framework material.
The initial enthalpies for H2 adsorption for all the complexes
lie in the range 5.34-6.70 kJ mol-1. The adsorption enthalpies
decrease with similar gradients at low coverage, except NOTT-
107 which decreases more rapidly. The NPD experiments on
NOTT-101 revealed that, at 1.72 D2/Cu loading (equivalent to
0.0065 mol g-1), all the adsorption sites are partially occupied,
even at 4 K, indicating that the H2 binding energies of the
exposed Cu(II) sites are not significantly higher than those at
other adsorption sites. If we consider that the initial adsorption
heat is due entirely to the adsorption of H2 molecules on the
exposed Cu(II) sites, the binding energy of H2-Cu is only 5.42
kJ mol-1 for NOTT-101. At a 0.025 mol g-1 loading, all the
exposed Cu(II) sites should be saturated, but NOTT-101 still
has an adsorption heat of 3.51 kJ mol-1; thus the difference
between the initial adsorption heat and the adsorption heat at
20 bar is ∼1.9 kJ mol-1. This value varies for the different Cu
materials reported herein. NOTT-107 has the largest difference
of 2.76 kJ mol-1, NOTT-106 2.37 kJ mol-1, NOTT-105 2.43
kJ mol-1, NOTT-103 1.92 kJ mol-1, NOTT-102 2.34 kJ mol-1,
and NOTT-100 1.35 kJ mol-1. Thus, we can conclude that the
binding energies of H2 molecules to the exposed Cu(II) sites
are similar to those for other, nonmetal, adsorption sites. Another
factor that requires consideration is that the enthalpies of
adsorption on the exposed Cu(II) sites are affected significantly
by the pore size and pore geometry. In NOTT-100, the close
proximity of the opposite wall creates a high H2-framework
potential. In NOTT-105, -106, and -107, the fluorine or methyl
substituents create a small pocket in the ellipsoidal [Cu24(L)6]
cage (Figure 10b and 10c). Significantly, these pocket structures
can result in apparent high adsorption enthalpies. The NPD
experiments on NOTT-101 gave the D2 occupancy of 0.85 at
the exposed Cu(II) site at 1.72 D2/Cu (0.0065 mol g-1).
Therefore, we reasonably consider the differences in the heats
of adsorption (Figure 9a) between zero loading and 0.0075
mol g-1 as the difference of binding energy between the exposed
Cu(II) site and the nonmetal site.
Figure 9. Variation of isosteric heat of adsorption with amount of adsorbed
H2 in [Cu2(L)] frameworks: (a) at low H2 loading; (b) at high H2 loading.
differences between the H2 interactions with the cavities in
NOTT-101 and NOTT-102. The adsorption heat of NOTT-100
at high H2 loading is 1.2 kJ mol-1 higher than that of NOTT-
101 and 1.7 kJ mol-1 higher than that of NOTT-102, and this
reflects the smaller cavity in NOTT-100 (diameters below 10
Å) and the greater potential overlap in the narrower pore.
Comparison of NOTT-101 with its structural derivatives
NOTT-105, -106, -107 reveals how the heat of adsorption at
low loading can be modified by introducing functional groups
into the cavities. As shown in Figure 9a, the heat of adsorption
for NOTT-105-107 is significantly higher than that for NOTT-
101 and ranges from 5.77 to 6.70 kJ mol-1 at zero coverage
and decreases much more rapidly with increasing H2 loading
(Figure 9b). The highest initial adsorption heat (6.70 kJ mol-1
)
was observed for NOTT-107, followed by NOTT-106, suggest-
ing that the four methyl groups contribute toward the increased
isosteric heat. It decreases rapidly with increasing loading due
to the saturation of the strong adsorption sites. NOTT-106 has
an initial adsorption heat of 6.34 kJ mol-1, 0.57 kJ mol-1 higher
than that for NOTT-105. The difference of adsorption heat
between NOTT-105, NOTT-106, and NOTT-107 becomes
The initial heat of adsorption for NOTT-107 is 1.36 kJ mol-1
higher than that of NOTT-101 (Figure 9a). This value is larger
than the difference between the initial adsorption heat between
the exposed metal site and nonmetal sites. This clearly indi-
cates the importance of pore geometry and functionality in
enhancing the heat of adsorption of H2. A similar pocket
9
2170 J. AM. CHEM. SOC. VOL. 131, NO. 6, 2009

H2 Adsorption in Cu(II) Tetracarboxylate Materials
A R T I C L E S
Figure 10. Comparison of local structures of [Cu24(L)6] cage in (a) NOTT-101, (b) NOTT-105, and (c) NOTT-106. NOTT-101 has no pocket structure.
Methyl groups in NOTT-106 create a closed pocket. For clarity, only those groups which project into the cavity are shown as space-filling (i.e., with van
der Waals radii). Green, F; dark gray, C; light gray, H.
structure is also observed in NOTT-103 where the kinked
topology of naphthalene groups can be compared to the linear
biphenyl groups in NOTT-102, which results in an ∼0.8 kJ
mol-1 increase for all H2 loadings measured. The naphthalene
group in [L4]4- in NOTT-103 appears to form an optimized
pocket while the length of [L3]4- can still create appropriate
and large voids for storing H2 at high pressure; NOTT-103,
however, maintains a relatively high adsorption heat at high
loading.
However, the reduced pore volume leads to lower overall H2
capacities. The naphthalene ring in NOTT-104 achieves a key
balance of enhancing H2-framework interaction within a large
pore volume, thus affording effective H2 uptake at low and high
pressures. This can be rationalized by the bent 2,6-naphthalene
group and fused aromatic ring. Highly compacted H2 adsorptions
have been reported in metal-organic frameworks;47 therefore,
it is possible to achieve higher H2 adsorption by incorporating
a sufficiently large pore such that the local structure can enhance
the H2-framework interaction.
Conclusions and Perspectives
Acknowledgment. This paper is dedicated to Professor Jack
Lewis, the Lord Lewis of Newnham, on the occasion of his 80th
birthday. We thank the EPSRC (UK Sustainable Hydrogen Energy
Consortium, http://www.uk-shec.org.uk/) and the University of
Nottingham for support and funding. M.S. gratefully acknowledges
receipt of a Royal Society Wolfson Merit Award and of a Royal
Society Leverhulme Trust Senior Research Fellowship. The work
at NCNR was partially supported by the U.S. Department of
Energy’s Office of Energy Efficiency and Renewable Energy within
the Hydrogen Sorption Center of Excellence. We are grateful to
STFC Daresbury Laboratory for the award of beam time on SRS
Stations 9.8 and 16.2SMX and to Drs. J. E. Warren and T. J. Prior
for experimental assistance. I.T., M.Z., and G.S.W. gratefully
acknowledge support from the EU-FP6 HYTRAIN RTN project.
We have synthesized a series of [Cu2(L)] framework com-
plexes and their H2 adsorption properties have been investigated.
NOTT-103 has achieved 77.8 mg g-1 (7.22 wt%) total H2
adsorption at 77 K and 60 bar. We have investigated the
correlation of the H2 uptake capacity and heats of adsorption
with pore metrics and the influences of functional groups on
the ligands. Exposed Cu(II) sites and two sites close to the center
of three {Cu2} paddle wheels were identified as strong adsorp-
tion sites at low H2 loading by Rietveld refinement of NPD data.
The exposed Cu(II) site is the strongest adsorption site in the
pore, but its adsorption heat is only slightly higher than those
for other sites. Furthermore, it is significantly affected by the
ligand structure near the {Cu2} paddle wheels. Therefore, we
conclude that the higher adsorption heat observed in NOTT-
106 and NOTT-107 at low loading represents the collective
contribution from the exposed Cu(II) site and the dimensions
and functionality of the pore defined by the methyl groups on
the ligands.
Supporting Information Available: Crystallographic informa-
tion files (CIFs), views of crystal structures, PXRD data for
as-prepared and evacuated materials, and isotherm analyses. This
material is available free of charge via the Internet at http://
pubs.acs.org.
In NOTT-106 and -107 the methyl groups project into the
pore forming a small pocket where overlap of potential energy
fields occurs. This provides a strong interaction between H2 and
the framework and results in high H2 adsorption enthalpies.
JA806624J
9
J. AM. CHEM. SOC. VOL. 131, NO. 6, 2009 2171