A Chromium Terephthalate-Based Solid with Unusually Large Pore

Volumes and Surface Area

G. Férey, et al.
Science 309, 2040 (2005);
DOI: 10.1126/science.1116275

This copy is for your personal, non-commercial use only.

If you wish to distribute this article to others, you can order high-quality copies for your
colleagues, clients, or customers by clicking here.
Permission to republish or repurpose articles or portions of articles can be obtained by
following the guidelines here.

The following resources related to this article are available online at

www.sciencemag.org (this infomation is current as of October 3, 2011 ):

Downloaded from www.sciencemag.org on October 3, 2011

A correction has been published for this article at:
http://www.sciencemag.org/content/310/5751/1119.full.html
Updated information and services, including high-resolution figures, can be found in the online
version of this article at:
http://www.sciencemag.org/content/309/5743/2040.full.html
Supporting Online Material can be found at:
http://www.sciencemag.org/content/suppl/2005/09/20/309.5743.2040.DC1.html
A list of selected additional articles on the Science Web sites related to this article can be
found at:
http://www.sciencemag.org/content/309/5743/2040.full.html#related
This article has been cited by 531 article(s) on the ISI Web of Science
This article has been cited by 4 articles hosted by HighWire Press; see:
http://www.sciencemag.org/content/309/5743/2040.full.html#related-urls
This article appears in the following subject collections:
Chemistry
http://www.sciencemag.org/cgi/collection/chemistry

Science (print ISSN 0036-8075; online ISSN 1095-9203) is published weekly, except the last week in December, by the
American Association for the Advancement of Science, 1200 New York Avenue NW, Washington, DC 20005. Copyright
2005 by the American Association for the Advancement of Science; all rights reserved. The title Science is a
registered trademark of AAAS.
 REPORTS CORRECTED 18 NOVEMBER 2005; SEE LAST PAGE
building blocks, may connect in 3D space to
A Chromium Terephthalate–Based form periodic lattices. A virtual library of can-
didate frameworks is produced, along with their
crystallographic features (space group, cell pa-
Solid with Unusually Large Pore rameters, atomic coordinates) and their simu-
lated x-ray diffraction (XRD) patterns. The
Volumes and Surface Area comparison of the simulated pattern of each
candidate structure with the experimental one
G. Férey,1,2* C. Mellot-Draznieks,3 C. Serre,1 F. Millange,1 identifies the targeted experimental structure,
J. Dutour,1 S. Surblé,1 I. Margiolaki4 giving direct access to the structural solution
without any recourse to single crystals. The fi-
We combined targeted chemistry and computational design to create a crystal nal structure is refined with the Rietveld method
structure for porous chromium terephthalate, MIL-101, with very large pore sizes from powder data, the guest species being lo-
and surface area. Its zeotype cubic structure has a giant cell volume (È702,000 calized from Fourier difference maps. While
cubic angstroms), a hierarchy of extra-large pore sizes (È30 to 34 angstroms), tackling the underlying issue of polymor-
and a Langmuir surface area for N2 of È5900 T 300 square meters per gram. phism of hybrid materials, our computational
Beside the usual properties of porous compounds, this solid has potential as a approach provides a direct-space tool for solv-
nanomold for monodisperse nanomaterials, as illustrated here by the incorpo- ing structures that may be highly complex.
ration of Keggin polyanions within the cages. To implement our combined method, we first

Downloaded from www.sciencemag.org on October 3, 2011

determined (10) the adequate chemical con-
Porous materials with large, regular, accessible made[ approach is to control the nature of the ditions leading to the existence of trimeric
cages and tunnels are increasingly in demand inorganic cluster and the chemical conditions inorganic building blocks, formed by the assem-
for applications in catalysis (1), separations (2), required for its formation and stability in bly of three octahedra sharing a m3-O common
sensors, electronics, and gas storage (3, 4). De- solution (8). In the second step, we use com- vertex. Simulations were then performed to
pending on their structure and pore size, these putational strategies, typically our global opti- combine these inorganic trimers in 3D space
materials allow only molecules of certain mization AASBU (automated assembly of with 1,3,5-benzene tricarboxylate (BTC)
shapes and sizes to enter the pores. Further- secondary building units) method (9–12), as through the assembly of a computationally
more, giant pores may act as nanoreactors, in recently adapted to hybrids (13), or other designed hybrid building block. Among the
which the confined volume may generate re- closely related methods (14). The AASBU various predicted crystal structures, one candi-
actions that do not occur in the bulk material, method explores how an inorganic cluster and date exhibited the same powder XRD pattern as
or as nanomolds for calibrated and mono- an organic linker, or even predefined hybrid the powdered chromium trimesate MIL-100
disperse nanomaterials (5). In this respect, the
larger the pores, the wider the range of re-
actants that can be combined or stored.
However, the design of materials with
increasingly large pores carries, especially for
metal-organic frameworks, the risk of inter-
penetration of the skeletons within structures.
In addition, although the structural characteri-
zation of such solids with large cells is usually
possible when single crystals are available, the
probability of getting the solutions is known to
drastically decrease or even to become zero
when the cell dimensions increase too much
(6). These problems have restricted the num-
ber of discovered porous solids with extra-
large pores, of which cloverite, with a cell
volume of È125,000 )3 and pore diameters
close to 30 ) (7) is the largest.
We recently developed a strategy to over-
come these limitations based on the combina-
tion of targeted chemistry and computer
simulations. Hybrid porous solids result from
the three-dimensional (3D) covalent connection
of inorganic clusters and organic moieties that
act as linkers, and the first step in our Btailor-

1
Institut Lavoisier, CNRS Unité Mixte de Recherche
8637, Université de Versailles St-Quentin en Yvelines,
45 Avenue des Etats-Unis, 78035 Versailles Cedex,
France. 2Institut universitaire de France, 103, Boulevard
Saint-Michel, 75005 Paris, France. 3Royal Institution, 21
Albemarle Street, London W1S 4BS, UK. 4European
Radiation Synchotron Facility, 38042 Grenoble, France.
*To whom correspondence should be addressed.
E-mail: ferey@chimie.uvsq.fr
Fig. 1. (A) The computationnally designed trimeric building block chelated by three carboxylic func-
tions. The ST was constructed with (B) terephthalic acid, which lies (C) on the edges of the ST. (D)
Ball-and-stick representation of one unit cell, highlighting one ST drawn in a polyhedron mode. (E)
Schematic 3D representation of the MTN zeotype architecture (the vertices represent the centers of
each ST) with the medium (in green, with 20 tetrahedra) and large (in red with 28 tetrahedra) cages
delimited by the vertex sharing of the ST. Chromium octahedra, oxygen, fluorine and carbon atoms
are in green, red, and blue, respectively.

2040 23 SEPTEMBER 2005 VOL 309 SCIENCE www.sciencemag.org


(15) (MIL, Mat2rial Institut Lavoisier), reveal-
ing a giant cell volume (9380,000 )3) and large
pore sizes (È25 to 29 )) consistent with the
very high surface area measurement (SLangmuir 0
3100 m2 gj1). This hybrid crystal structure was
then refined from synchrotron powder data, al-
lowing the further localization of the guest
moities. MIL-100 showed the feasibility of
creating simulation-assisted chemical structures
(16). With that adressed, we investigated other
carboxylates, here terephthalic acid E1,4-benzene
dicarboxylate (1,4-BDC)^ combined with similar
trimers. Compared to other MOFs (metal-
organic frameworks), the resulting solid, MIL-
101, has the best characteristics in terms of
cell dimensions (702,000 )3), pore sizes (29
to 34 )), and surface area (5900 m2 gj1).
The synthesis of MIL-101 consists in the
hydrothermal reaction of H2BDC (166 mg at
1 mmol) with Cr(NO3)3.9H2O (400 mg at 1
mmol), fluorhydric acid (0.2 ml at 1 mmol),
and H2O (4.8 ml at 265 mmol), for 8 hours
at 220-C. This reaction produced a highly
crystallized green powder of the chromium ter-
ephthalate with formula Cr3F(H2O)2OE(O2C)-
C6H4-(CO2)^3.nH2O (where n is È25), based on
chemical analysis. The yield based on chromi-
um is È50%. Analysis of the powder XRD
data indicates a cubic cell (a È 89 )) and a
close relationship with the augmented Mobil
Thirty-Nine (MTN) zeotype structure of
MIL-100 (17). For the simulation process, a
candidate hybrid building block, made on the
sole basis of its compatibility with the exper-
imental metal:organic ratio of the targeted MIL-
Fig. 2. (A and B) Ball-and-stick view and free dimensions (Å) of the pentagonal and hexagonal
windows. (C and D) Ball-and-stick view of the two cages. Chromium octahedra, oxygen, fluorine
and carbon atoms are in green, red, red and blue, respectively.
www.sciencemag.org
101, was computationally designed as a super
tetrahedron (hereafter noted ST) (Fig. 1). It was
made from the linkage of 1,4-BDC anions and
inorganic trimers that consist of three iron atoms
in an octahedral environment with four oxygen
atoms of the bidendate dicarboxylates, one m3O
atom, and one oxygen atom from the terminal
water or fluorine group. Octahedra are related
through the m3O oxygen atom to form the
trimeric building unit. The four vertices of the
ST are occupied by the trimers, and the organic
linkers are located at the six edges of the ST.
Following our previous strategy (14), the ST
building blocks were computationally assem-
bled. The size of the ST requires an expansion of
the cubic unit cell (space group F d-3m) to more
than 85 ) before the construction process. The
connection between the ST was established
through vertices to ensure a 3D network of
Bcorner-sharing[ super tetrahedra with an aug-
mented MTN zeotype architecture (18) (Fig. 1)
and illustrates our concept of scale chemistry
(19). Once the structure construction of MIL-101
was computationally completed (74 atoms per
asymmetric unit without hydrogen atoms), the
model structure of MIL-101 was directly used
for full structural refinement with synchrotron
data (table S1 and fig. S1) (20). Free water
molecules filling the pores were located through
successive Fourier differences.
The STs are microporous (with a È8.6 )
free aperture for the windows), and the result-
ing framework delimits two types of meso-
porous cages filled with guest molecules
(Fig. 2). These two cages, which are present
SCIENCE VOL 309 23 SEPTEMBER 2005
REPORTS
in a 2:1 ratio, are delimited by 20 and 28 ST
(one by 20 and one by 28) with internal free
diameters of È29 ) and 34 ), respectively
(Fig. 2). These values correspond to acces-
sible pore volumes of È12,700 )3 and
È20,600 )3, respectively. The large windows
of both cages make the latter accessible to
very large molecules. The smaller cages ex-
hibit pentagonal windows with a free open-
ing of È12 ), while the larger cages possess
both pentagonal and larger hexagonal win-
dows with a È14.5 ) by 16 ) free aperture
(Fig. 2).
Thermogravimetric analysis in air (fig.
S2) revealed that MIL-101 was stable up to
275-C. X-ray thermodiffractometry showed
that the evacuation of the guest molecules
did not affect the framework. We also found
that MIL-101 exhibited very high uptake of
Downloaded from www.sciencemag.org on October 3, 2011
gases. The N2 sorption isotherm on the de-
hydrated sample (fig. S3A) is of type I with
secondary uptakes at P/P0 È 0.1 and at P/P0 È
0.2, where P is gas pressure and P0 is sat-
uration pressure, characteristic of the presence
of the two kinds of microporous windows.
Using the Dubinin-Raduskhvich equation, we
found a pore volume near 2.0(1) cm3 gj1 for
MIL-101. The apparent Brunauer Emmer
Teller (BET) and SLangmuir surface area are
larger than 4,100(200) and 5,900(300) m2 gj1,
respectively (Fig. 3). The isotherms are proba-
bly overestimating the true internal surface area
of MIL-101, but comparisons to be made with
related materials are possible. To our knowledge,
the highest surface area reported for any
crystalline or amorphous solid E4500 m2 gj1
(Langmuir)^, was obtained with MOF-177, a
porous hybrid solid (20). The large standard
deviations for the surface area of MIL-101
comes both from experimental considerations
(such as error on weight measurements and
purity of the sample) and from the choice of the
points used for the BET or Langmuir calcula-
tion (fig. S3B). The as-synthesized MIL-101
solid exhibits a smaller SLangmuir surface area
within the 4500 to 5500 m2 gj1 range because
of the presence of variable amounts of free
terephthalic acid outside and within the pores.
An activation treatment was thus performed to
reach the maximal surface area and pore
volume (21).
MIL-101 is stable over months under air
atmosphere and was not altered when treated
with various organic solvents at room tem-
perature or under solvothermal conditions.
These properties, together with high adsorp-
tion capacities, make MIL-101 an attractive
candidate for the adsorption of gas or large
molecules. The very large windows easily
allow the introduction of new species into
the cages and possible enhanced reactions
favored by confinement effects (similar to a
pressure) in the cages. Moreover, as soon as
introduced nanometric species fill the vol-
ume, the fixed dimensions of the pores lead
2041
 REPORTS
to monodispersed nanomaterials on the scale
of 1 to 3 nm. The possibility of their intro-
duction depends on the fit between their size
and the accessible dimensions of the windows
of each cage. Large species may therefore
occupy only the large cages (È20,600 )3)
while leaving space for other species with
different properties in the small cages
(È12,700 )3). Such a selective placement
of guests might lead to hitherto unknown
assemblies of monodisperse multifunctional
nanomaterials and the possibility of structural
characterization of these nano-objects when the
host structure is not affected by the introduction
of species.
To illustrate this idea, we explored the in-
corporation of Keggin polyanions in MIL-101.
Such incorporation has been achieved pre-
viously through the encapsulation of a molyb-
denum Keggin anion in a metal-oxygen
system (22). The K7PW11O40.nH2O salt was
selected because of its low acidity, the pres-
ence of a 31P nuclear magnetic resonance
(NMR) nucleus for NMR characterization, and
finally its size (van der Waals radius, È13.1 )),
which rules out the diffusion of Keggin ions
into the small cages. A powdered sample of
MIL-101 was placed in an aqueous solution
Fig. 3. Nitrogen gas
sorption isotherm at 1200
78 K for MIL-101. P/PO
is the ratio of gas pres- 1000
sure (P) to saturation
Vads (cm3.g-1)
pressure (P0 0 750 torr).
800
600
400
200
0 23. Initial hydrogen storage measurements were È0.45
0 0.2
Fig. 4. (A) Schematic view of the insertion
of Keggin anions within the largest pore of
MIL-101. (B) XRD of MIL-101 (1) and MIL-
101(Keggin) (2). q, in degrees. (C) TGA of
MIL-101 (1) and MIL-101(Keggin) (2). T, temperature (K). (D) Nitrogen sorption-desorption
isotherms at 78 K of MIL-101 (1) and MIL-101(Keggin) (2). Vads, volume adsorbed in cm3 gj1. (E)
31P solid-state NMR spectra of the Keggin salt and MIL-101(Keggin). d, chemical shift in ppm.
2042
of the Keggin salt for 2 hours. To probe the
presence of the polynanion within the pores, the
resulting MIL-101–Keggin solid was analyzed
by thermal gravimetric analysis (TGA), N2
sorption measurement, XRD, 31P solid state
NMR (Fig. 4), and infrared spectroscopy (fig.
S3). All of these techniques confirmed the
presence of a large amount of the Keggin ions
within the pores: a strong decrease in the weight
losses ETGA (fig. S2)^ and surface areas
ESLangmuir È 3,750(250) m2 gj1 instead of
5,900(300) m2 gj1 for MIL-101 (fig. S3C)^
due to the higher density of the Keggan
moieties compared to MIL-101. We observed
significant changes in XRD peak intensities
but not of the Bragg peak positions (fig. S4).
Infrared spectroscopy confirmed the presence
of the polyanions within the pores of MIL-101
(fig. S5). 31P NMR also indicated one single
peak at about–12.2(1) parts per million (ppm),
confirming that the integrity of the Keggin
structure was retained within the pores of
MIL-101 (fig. S6).
Quantitative analysis (17) also gave an
estimation of È0.05 Keggin anions per chro-
mium. Considering the size of the K7PW11O40
ion, we assumed that the polyanions could dif-
fuse into the largest cages only, which would
0.4 0.6 0.8 1 and 3.75 weight percent at 293 K and 77 K, re-
P/P0
23 SEPTEMBER 2005 VOL 309 SCIENCE www.sciencemag.org
allow about five highly charged Keggin moities
per large cage (Keggin per cage, È5.3).
Because the volume of a PW11O407– anion is
nearly 2250 A3, five Keggin ions represent
È10,100 A3 in volume, which is lower than the
È20,600 A3 volume of a large cage. The re-
sidual volume is probably occupied by cations
and water molecules. This successful incorpo-
ration of Keggin anions in large amounts
strongly suggests that MIL-101, a crystallized
hybrid solid with a periodical and calibrated
porosity, is an excellent candidate for the intro-
duction of gas (23) nano-objects in a regular
and monodisperse mode with specific physical
properties (24) or for drug delivery (25).
References and Notes
1. J. S. Seo et al., Nature 404, 982 (2000).
2. F. Schüth et al., Handbook of Porous Solids (Wiley-
VCH, Weinheim, Germany, 2002), vol. 3, p. 563.
Downloaded from www.sciencemag.org on October 3, 2011
3. J. L. C. Rosell et al., J. Am. Chem. Soc. 126, 5666 (2004).
4. G. Férey et al., Chem. Commun. 2003, 2276 (2003).
5. B. H. Hong et al., Science 294, 348 (2001).
6. R. S. Batten et al., Angew. Chem. Int. Ed. Engl. 37,
1460 (1998).
7. M. Estermann, L. B. McCusker, C. Baerlocher, A.
Merrouche, H. Kessler, Nature 352, 320 (1991).
8. C. Serre et al., Angew. Chem. Int. Ed. Engl. 43, 6286
(2004).
9. S. Girard et al., Stud. Surf. Sci. Catal. 135, 254 (2001).
10. C. Mellot-Draznieks et al., J. Am. Chem. Soc. 124,
15326 (2002).
11. C. Mellot-Draznieks et al., Chem. Eur. J. 8, 4102 (2002).
12. C. Mellot-Draznieks et al., Angew. Chem. Int. Ed.
Engl. 39, 2271 (2000).
13. C. Mellot-Draznieks et al., Angew. Chem. Int. Ed.
Engl. 43, 6290 (2004).
14. C. Mellot-Draznieks et al., Z. Anorg. Allg. Chem. 630,
2599 (2004).
15. G. Férey et al., Angew. Chem. Int. Ed. Engl. 43, 6296
(2004).
16. G. Férey et al., Acc. Chem. Res. 38, 217 (2005).
17. Materials and methods are available as supporting
material on Science Online.
18. Available at www.iza-structure.org/databases.
19. G. Férey, J. Solid State Chem. 152, 37 (2000).
20. H. K. Chae et al., Nature 427, 523 (2004).
21. Materials and methods are available as supporting
material on Science Online.
22. A. Muller et al., Angew. Chem. Int. Ed. Engl. 39, 3414
(2000).
spectively, at 2 MPa. The values at 77 K seem to be
the highest for MOFs after the contestation of pre-
vious results (3).
24. Initial introduction of semiconducting ZnS nano-
particles in the pores of MIL-101 was successfully
achieved with a ZnS/Cr ratio close to 0.5.
25. Incorporation of ibuprofen within MIL-101 has been
achieved in large amounts (È1 g of drug per gram of
MIL-101) with a total release of the drug within a
few days.
26. We thank F. Taulelle and M. Haouas for collecting solid
state NMR, M. Latroche for hydrogen sorption mea-
surements, P. Mialane for providing the Keggin salt and
for helpful discussions, S. Floquet and E. Cadot for
helpful discussion about the introduction of ZnS nano-
particles, and P. Horcajada and M. Vallet-Regi for the
drug incorporation study.
Supporting Online Material
www.sciencemag.org/cgi/content/full/309/5743/2040/
DC1
Materials and Methods
Figs. S1 to S6
Table S1
References and Notes
17 June 2005; accepted 16 August 2005
10.1126/science.1116275
 C O R R E C T I O N S A N D C L A R I F I C AT I O N S

ERRATUM post date 18 November 2005

Reports: “A chromium terephthalate–based solid with unusually large pore

volumes and surface area” by G. Férey et al. (23 Sept. 2005, p. 2040). The
crystal structure deposition number for MIL-101 was omitted; the CSD num-
ber is 415697. There were errors in note (23); it should read “Initial hydrogen
storage measurement was 4.5 weight percent at 77 K at 3 MPa. This value
seems to be the highest for MOFs after the contestation of previous results
(3).” Finally, the last sentence of the Fig. 1 legend should read “Chromium
octahedra, oxygen, fluorine, and carbon atoms are in green, red, red, and blue,
respectively.”

Downloaded from www.sciencemag.org on October 3, 2011

www.sciencemag.org SCIENCE Erratum post date 18 NOVEMBER 2005 1


New Directions
in Plastic Debris
THE LARGEST EVER MEETING FOCUSING ON
plastic debris in the environment was
recently held in Redondo Beach, California
(1). It is evident that plastic waste presents
major concerns in aquatic habitats world-
wide. However, this meeting differed from
previous efforts/gatherings because repre-
sentatives from industry, government, aca-
demia, and nongovernment organizations
were united in their desire to identify solu-
tions to reducing waste. There has been a
switch in the types of litter recorded, from
shipping- and fishing-related debris to land-
based sources. This was poignantly under-
scored by reports of islands of plastic debris
swept into the sea by Hurricane Katrina.
Polymer scientist A. Andrady explained
that all the plastic introduced into the
oceans remains unmineralized as either
entire objects or fragments, some of which
are less than 20 µm in diameter (2). Large
items of debris cause entanglement,
impaired feeding, and mortality to birds,
turtles, and mammals. Microscopic frag-
ments are also ingested, but the conse-
quences are unknown. H. Takada and C.
Moore presented evidence on the ability of
plastic to accumulate PCBs, DDE, and
nonylphenol (3), and the potential for toxic
chemicals to transfer to the food chain was
identified as a key research direction. It was
also recognized that better understanding of
effects at an organismal level is required
before consequences at population and
ecosystem levels can be examined.
In terms of solutions, much could be
achieved by reductions in packaging.
Keynote speaker W. McDonough made the
case for a “cradle to cradle” (4) strategy to
ensure that plastics are retained in a prod-
uct-specific recycling loop—turning debris
from a waste disposal liability into feed-
stock for production. Although debris can
be removed from drains and rivers by phys-
ical separators, there is also a key role for
education to help reduce littering. The
importance of social research to establish
the public’s willingness to engage with
these solutions was also clearly recognized.
RICHARD THOMPSON,1* CHARLES MOORE,2
CREDIT: DAVID QUINN/AP PHOTO
ANTHONY ANDRADY,3 MURRAY GREGORY,4
HIDESHIGE TAKADA,5 STEPHEN WEISBERG6
1 School of Biological Sciences, University of
Plymouth, Plymouth, Devon PL4 8AA, UK. 2Algalita
Marine Research Foundation, 148 North Marina
Drive, Long Beach, CA 90803, USA. 3 Research
Triangle Institute, 3040 Cornwallis Road, Research
www.sciencemag.org
Toy cars amid debris in New Orleans after Hurricane pare for the worst scenario. A
Katrina.
Triangle Park, NC 27709–2194, USA. 4University of
Auckland, Private Bag 92019, Auckland, New
Zealand. 5 Tokyo University of Agriculture and
Technology, Tokyo 183 8509, Japan. 6Southern
California Coastal Research Project, 7171 Fenwick
Lane, Westminster, CA 92683, USA.
*To whom correspondence should be addressed.
E-mail: r.c.thompson@plymouth.ac.uk
References
1. Plastic Debris Rivers to Seas, organized by the sive care, and even funeral facilities that
California Coastal Commission, 7 to 9 Sept 2005.
2. R. C. Thompson et al., Science 304, 838 (2004).
3. Y. Mato et al., Environ. Sci. Technol. 35, 318 (2001).
4. W. McDonough, M. Braungart, Cradle to Cradle (North
Point Press, New York, 2002).
Preparing for the
Worst-Case Scenario
T HE R EPORT “C ONTAINING PANDEMIC
influenza at the source” by I. M. Longini Jr.
et al. (12 Aug., p. 1083) was encouraging
that an avian pandemic can be contained if
proper intervention is carried out promptly.
N. M. Ferguson et al. published similar
findings (1). However, further investigation
is needed before we can celebrate.
A valid conclusion from a model
requires a careful selection of the parameter
values. Longini’s article assumed that
Tamiflu (oseltamivir) was useful in a pan-
demic. Yet, Tamiflu may not be effective on
all new avian flu viruses, which can have
SCIENCE VOL 310 18 NOVEMBER 2005
Published by AAAS
LETTERS
80% mortality; Tamiflu was inef-
fective in 50% of patients in
Thailand (2). Moreover, although
the basic reproduction number (R0)
below 2 was a reasonable estima-
tion in Longini’s models, previous
flu pandemics have had an R0 up to
3 (3). Apart from these technical
parameters in a hypothetical
model, the logistics in a real situa-
tion can also be fluid. Although
developing countries with pharma-
ceutical factories can issue a com-
pulsory license to make generic
copies of patented drugs in the
Downloaded from www.sciencemag.org on October 3, 2011
event of a medical emergency, in
reality when there is a pandemic
occurring, developing areas with-
out such facilities (perhaps those
that most need Tamiflu) will be in a
difficult position to secure enough
supply.
Handling the ever-changing dis-
ease pattern of pandemic avian flu
requires a contingency plan to pre-
worst-case scenario model should
predict the resources a public
health system needs to cope with a pan-
demic. This model should consider that
Tamiflu may be ineffective in treating the
new strain, a higher-than-expected R 0
value, and the possibility of a paucity of
antiviral drug/vaccines. Such a worst-case-
scenario model provides valuable informa-
tion for resource planning, for example, the
number of ventilators, the amount of inten-
will be required.
PUI HONG ALEX CHUNG
Higher Specialist Trainee in Community Medicine,
10th floor, 147C Argyle Street, Mong Kok, Hong
Kong SAR, China.
References
1. N. M. Ferguson et al., Nature 437, 209 (2005).
2. T. H. Tran et al., N. Engl. J. Med. 350, 1179 (2004).
3. B. Cooper et al., poster presented at Influenza Vaccines
for the World, Lisbon, 24 to 26 May 2005.
Response
CONTAINING A POTENTIAL STRAIN OF PAN-
demic influenza at the source will not be
easy and will require intense surveillance of
influenza-like illness, speedy lab work to
identify isolated viruses, and feasible and
effective control strategies. We showed that
the targeted use of oseltamivir could con-
tain a potential pandemic strain of influenza
if the basic reproductive number (R0) were
below 1.6 and if the intervention took place
2 to 3 weeks after the first case appeared. If
the eff icacy of oseltamivir against the
1117
 LETTERS
antine could effectively uences represent the largest 16S rDNA
1000 contain the spread of the microflora data set reported to date in any
50% virus. Other forms of species.
60% quarantine also could be On the basis of their analysis, the authors
70%
80%
effective (1). Our results suggest that “differences between stool and
100 are highly probabilistic, mucosa community composition” exist. We
90%
and a strategy that works question the validity of this conclusion,
in many simulations does based on our own observation that stool
not necessarily work in all microflora composition can vary signifi-
10
Cases per 1000

simulations. Thus, even if cantly in stool samples collected before

countries are prepared to
try to contain the out-
1
break, the world needs to
be prepared for the event
that containment fails.
It is important to de-
0.1 velop international co-
operation around the main-
tenance of a mobile stock-
pile of oseltamivir that can
and after a colonoscopy (1). The authors
compare microflora composition in
colon biopsy samples obtained during
colonoscopy with a stool sample collected a
month afterwards. The authors acknowl-
edge potential problems with their interpre-
tation because of the delayed stool collec-
tion, but a rationale for collecting delayed
stool samples is not given.
This Report has significantly expanded

Downloaded from www.sciencemag.org on October 3, 2011

0.01
1.1 1.4 1.5 1.6
R0
The effectiveness of the household and
neighborhood cluster quarantine started 14
days after the first case at different values of
R 0. Outbreaks that result in a cumulative
incidence of ≤ 1 case per 1000 are considered
contained (horizontal line). The cumulative
incidence per 1000 does not always decrease
monotonically with increasing intervention
coverage for small differences due to sto-
chastic variability.
emergent virus were as low as 50% against
illness and transmission, the targeted use of
oseltamivir could still effectively contain
the spread (figs. S16 to S18). For a situation
in which oseltamivir is ineffective against
the emergent strain, we modeled the use of
voluntary household and neighborhood
cluster (i.e., one small grouping of about
four households) quarantine alone (figs. 3
and S13). For this intervention, the first
case in a locality (i.e., a region roughly 6 to
9 km in radius) triggers a quarantine policy.
Every case and a certain percentage of sus-
ceptible people restrict their movement to
their household and their neighborhood
cluster. The figure gives the average simu-
lated cumulative influenza case incidence
for different levels of quarantine effective-
ness for different values of R0. For R0 ≤ 1.5,
quarantine levels as low as 50% of people
being restricted to the household and neigh-
borhood cluster could possibly contain out-
breaks. For R0 ≤ 1.6, quarantine levels of
60% or higher could be effective. With an
R 0 as high as 2.4, we would need a 90%
effective quarantine for containment. For
larger values of R0, containment would not
be possible with quarantine measures alone.
If we consider an R0 of 2.4 and total viral
resistance to oseltamivir to be the worst-
case scenario, then only an extremely tight
household and neighborhood cluster quar-
be rapidly deployed to the
1.7 2.1 2.4
location of an emergent
strain. The World Health
Organization (WHO) is currently developing
such a stockpile and recently announced that
Roche is donating 3 million courses of
oseltamivir to its mobile stockpile (2). If a
potential pandemic strain of influenza
emerges, WHO plans to intervene, in concert
with the ministry of health of the affected
country and other public health organiza-
tions, along the lines that we describe above
[see index 1 of (2)]. Such international coop-
eration coupled with a rapid, effective
response will be necessary to have any hope
of containing pandemic influenza before it
spreads throughout the world, or at least slow-
ing the spread while a well-matched vaccine
is developed and sufficient quantities are
deployed.
IRA M. LONGINI JR. AND M. ELIZABETH HALLORAN
Department of Biostatistics, The Rollins School of
Public Health, Emory University, 1518 Clifton Road,
NE, Atlanta, GA 30322, USA.
References
1. N. M. Ferguson et al., Nature 437, 209 (2005).
2. “Donation of three million treatments of oseltamivir
to WHO will help early response to an emerging
influenza pandemic” (available at www.who.int/
mediacentre/news/releases/2005/pr36/en/),
accessed 30 Aug. 2005.
Timing in Collection
of Stool Samples
WE READ WITH GREAT INTEREST THE REPORT
“Diversity of the human intestinal micro-
bial flora” by P. B. Eckburg et al. (10 June,
p. 1635; published online 14 Apr.). We
applaud the authors for advancing this
important field by undertaking comprehen-
sive 16S rDNA sequencing to describe the
composition of the microbiota in stools and
at six sites of the colon in each of three
human volunteers. The analyzed 13,355
prokaryotic ribosomal RNA gene seq-
our knowledge of the diversity of the
intestinal microflora in a few subjects.
However, if we ever want to correlate
microflora composition with health or dis-
ease, we will have to design studies aimed
at understanding the variation in the
microflora composition in a large cohort
of human subjects.
VOLKER MAI,* O. COLIN STINE, J. GLENN MORRIS JR.
Department of Epidemiology and Preventive
Medicine, School of Medicine, University of
Maryland, Baltimore, MD 21201, USA.
*To whom correspondence should be addressed.
E-mail: vmai@epi.umaryland.edu
Reference
1. V. Mai, O. C. Stine, J. G. Morris Jr., unpublished data.
Response
OUR LARGE-SCALE COMPARATIVE ANALYSIS
of bacterial and archaeal 16S rDNA
sequences in the colon and stool revealed
signif icant intersubject variability and
patchy heterogeneity among the colonic
mucosal bacterial populations. Regarding
the statistical differences we reported
between stool and adherent mucosal popu-
lations, subjects B and C harbored differ-
ent bacterial populations in their colonic
mucosa compared with their stool samples
collected 4 weeks later, while the mucosal
populations in subject A were subsets of
the population observed in stool collected
4 weeks after colonoscopy. Each subject’s
stool community was more similar to the
communities of their own mucosal sam-
ples than to any community from a differ-
ent subject.
We acknowledged that the statistically
significant difference between the bacterial
composition of the stool and colonic
mucosa may have been due to the 4-week
delay in stool collection after colonoscopy.
The collection of stool was not originally
planned in the large Canadian population-
based case control study from which the
control subjects were selected. For this

1118 18 NOVEMBER 2005 VOL 310 SCIENCE www.sciencemag.org

Published by AAAS

study of three healthy subjects, from whom
the colonic tissue biopsies had already
been collected, we chose to obtain stool
samples 1 month after colonoscopy when
each subject had full recovery of stable,
baseline bowel function. Despite the origi-
nal study design, we agree with Mai et al.
that stool samples before endoscopy may
provide more reliable representations of
the steady-state stool bacterial population.
However, a controlled comparative study is
needed to reveal the degree to which stools
are microbiologically dissimilar at various
time intervals before and after bowel
cleansing. A small study using denaturing
gradient gel electrophoresis has suggested
that the bacterial composition in colonic
mucosal biopsies differs significantly from
that in stool collected prior to the proce-
dure (1), supporting our conclusions that
significant differences exist between these
microbial communities. Future studies
should address these issues with high-reso-
lution molecular methods and a greater
number of subjects.
PAUL B. ECKBURG,1 ELISABETH M. BIK,2
CHARLES N. BERNSTEIN,3 LES DETHLEFSEN,2
ELIZABETH PURDOM,4 MICHAEL SARGENT,3
STEVEN R. GILL,5 KAREN E. NELSON,5
DAVID A. RELMAN6
GetInfo
science.labvelocity.com
Science announces a new
online life science product
information system, GetInfo,
powered by LabVelocity
• Quickly find and request free information on
products and/or services found in the pages
of Science magazine
• Ask vendors to contact you with information
• View detailed product information
• Link directly to vendors’ websites
Visit GetInfo today at
science.labvelocity.com
www.sciencemag.org
1Division of Infectious Diseases and Geographic
Medicine, Stanford University School of Medicine,
Stanford, CA 94305–5107, USA. 2Department of
Microbiology and Immunology, Stanford University
School of Medicine, Stanford, CA 94305–5124,
USA. 3Section of Gastroenterology, University of
Manitoba, Winnipeg , MB R3A 1R9, Canada.
4 Department of Statistics, Stanford University,
Stanford, CA 94305, USA. 5 The Institute for
Genomic Research, Rockville, MD 20850, USA.
6Departments of Medicine, and of Microbiology
and Immunology, Stanford University School of
Medicine, Stanford, CA 94305–5124, USA, and
Veterans Affairs Palo Alto Health Care System, Palo
Alto, CA 94304, USA.
Reference
1. E. G. Zoetendal et al., Appl. Environ. Microbiol. 68,
3401 (2002).
CORRECTIONS AND CLARIFICATIONS
Editors’ Choice: “Grabbing hydrogen” (4 Nov.,
p. 749). The item should have described the activa-
tion of C-H bonds as typically proceeding by
oxidative addition of a C-H bond to a low-valent
metal complex.
Policy Forum: “Proof of safety at Yucca Mountain”
by L. J. Carter and T. H. Pigford (21 Oct., p. 447). The
figure parts should have been reversed. The original
design is shown on the right and the capillary bar-
rier design is shown on the left.
SCIENCE VOL 310 18 NOVEMBER 2005
Published by AAAS
LETTERS
Brevia: “Bacterial immunity traded for sperm via-
bility in male crickets” by L. W. Simmons and B.
Roberts (23 Sept., p. 2031). The URL for the
Supporting Online Material was omitted. The cor-
rect URL is www.sciencemag.org/cgi/content/
full/309/5743/2031/DC1.
Reports: “A chromium terephthalate–based solid
with unusually large pore volumes and surface area”
by G. Férey et al. (23 Sept., p. 2040).The crystal struc-
ture deposition number for MIL-101 was omitted;
the CSD number is 415697.There were errors in note
(23); it should read “Initial hydrogen storage meas-
urement was 4.5 weight percent at 77 K at 3 MPa.
This value seems to be the highest for MOFs after
the contestation of previous results (3).” Finally, the
last sentence of the Fig. 1 legend should read
“Chromium octahedra, oxygen, fluorine, and carbon
atoms are in green, red, red, and blue, respectively.”
Downloaded from www.sciencemag.org on October 3, 2011
Letters to the Editor
Letters (~300 words) discuss material published
in Science in the previous 6 months or issues of
general interest. They can be submitted
through the Web (www.submit2science.org) or
by regular mail (1200 New York Ave., NW,
Washington, DC 20005, USA). Letters are not
acknowledged upon receipt, nor are authors
generally consulted before publication.
Whether published in full or in part, letters are
subject to editing for clarity and space.
1119