Cis -Decalin oxidation as a stereochemical probe of in-MOF versus on-MOF catalysis

Article abstract of DOI:10.1039/c7cc02570j

Development of catalyst-controlled C-H hydroxylation could provide direct access to valuable synthetic targets, such as primary metabolites. Here, we report a new family of porous materials, comprised of 2-dimensional metalloporphyrin layers and flexible aliphatic linkers, and demonstrate C-H hydroxylation activity. We demonstrate that the stereochemistry of cis-decalin oxidation provides a useful tool for differentiating catalysis in from catalysis on porous materials, which is critical to leveraging the potential of porous materials for catalyst-controlled oxidation chemistry.

Full text of DOI:10.1039/c7cc02570j

View Article Online
View Journal
ChemComm
Accepted Manuscript
This article can be cited before page numbers have been issued, to do this please use: A. D. Cardenal, H.
J. Park, C. J. Chalker, K. G. Ortiz and D. C. Powers, Chem. Commun., 2017, DOI: 10.1039/C7CC02570J.
This is an Accepted Manuscript, which has been through the
Volume 52 Number
1 4 January 2016 Pages 1–216
Royal Society of Chemistry peer review process and has been
accepted for publication.
ChemComm
Accepted Manuscripts are published online shortly after
Chemical Communications
www.rsc.org/chemcomm
acceptance, before technical editing, formatting and proof reading.
Using this free service, authors can make their results available
to the community, in citable form, before we publish the edited
article. We will replace this Accepted Manuscript with the edited
and formatted Advance Article as soon as it is available.
You can find more information about Accepted Manuscripts in the
author guidelines.
Please note that technical editing may introduce minor changes
to the text and/or graphics, which may alter content. The journal’s
standard Terms & Conditions and the ethical guidelines, outlined
in our author and reviewer resource centre, still apply. In no
event shall the Royal Society of Chemistry be held responsible
for any errors or omissions in this Accepted Manuscript or any
consequences arising from the use of any information it contains.
ISSN 1359-7345
COMMUNICATION
Marilyn M. Olmstead, Alan L. Balch, Josep M. Poblet, Luis Echegoyenet al.
Reactivity differences of Sc3N@C2n (2n 68 and 80). Synthesis of the
first methanofullerene derivatives of Sc N@D5h-C80
=
3
rsc.li/chemcomm

Please do not adjust margins
ChemComm
Page 1 of 4
View Article Online
DOI: 10.1039/C7CC02570J
Journal Name
COMMUNICATION
Cis-decalin oxidation as a stereochemical probe of in-MOF versus
on-MOF catalysis
Ashley D. Cardenal,† Hye Jeong Park,† Cody J. Chalker, Kacey G. Ortiz and David C. Powers*
Received 00th January 20xx,
Accepted 00th January 20xx
DOI: 10.1039/x0xx00000x
www.rsc.org/
Development of catalyst-controlled C–H hydroxylation could a substantial challenge to differentiate catalysis on materials
provide direct access to valuable synthetic targets, such as primary (i.e. catalyst sites on the exterior of crystallites) from catalysis
metabolites. Here, we report a new family of porous materials, in materials. We are interested in porphyrinic materials that
comprised of 2-dimensional metalloporphyrin layers and flexible feature flexible organic linkers as platforms for catalyst-
aliphatic linkers, and demonstrate C–H hydroxylation activity. We controlled oxidation of organic molecules based on the
demonstrate that the stereochemistry of cis-decalin oxidation contention that network flexibility9 may provide substrate-
provides a useful tool for differentiating catalysis in from catalysis specific reaction environments in which to accomplish
on porous materials, which is critical to leveraging the potential of catalysis. Critical to development of catalysts for these
porous materials for catalyst-controlled oxidation chemistry.
applications is determining if catalysis is accomplished inside
catalyst pores or on the surface of catalyst particles. Here we
report the synthesis of a new family of porphyrinic materials
Enzymatic C–H oxidation during xenobiotic metabolism and
biosynthesis is accomplished with catalyst-controlled
selectivity within substrate-adaptive enzyme active sites.1 In
contrast, most synthetic C–H hydroxylation methods proceed
with substrate-controlled selectivity at either the weakest or
the most sterically unencumbered C–H bond.2 The
development of strategies for catalyst-controlled C–H
hydroxylation could provide direct entry into complex
functional molecules, such as metabolites.3 We are interested
in utilizing flexible metal-organic frameworks (MOFs) as
catalyst platforms to achieve oxidation selectivities that are
challenging to access in solution-phase chemistry.
and present
a stereochemical test based on cis-decalin
oxidation to differentiate catalysis in materials from catalysis
on materials.
Solvothermal combination of 5,10,15,20-tetrakis(4-
carboxyphenyl)porphyrin iron chloride (Fe(tcpp)Cl), 1,3-di(1H-
1,2,4-triazol-1-yl)propane (btp), and Zn(NO3)2·6H2O in N,N’-
diethylacetamide (DEA) and ethanol (EtOH) at 80 °C affords a
dark-purple crystalline material with the empirical formula
[Zn2(Fe(tcpp))(btp)(NO3)]n (Fe(btp), Figure 1). The chemical
composition of Fe(btp) was evaluated by: 1) elemental
analysis; 2) acid digestion followed by UV-vis and 1H NMR
analyses to define the Fe(tcpp) and btp content, respectively
(Figures S1 and S2); 3) ICP-MS analysis of an acid-digested
sample to establish the Fe:Zn ratio (Table S1); and 4) diffuse
reflectance UV-vis spectroscopy (Figure S1). Elemental analysis
and IR spectroscopy (Figure S3) indicate that the chloride
counterions of the Fe(tcpp)Cl starting material are replaced by
nitrate ions during solvothermal synthesis of Fe(btp). The
average Fe(tcpp):btp ratio of acid-digested samples is 0.8,
which indicates the presence of some ligand absences from
the ideal [Zn2(Fe(tcpp))(btp)(NO3)]n empirical formula.
MOFs are a class of materials comprised of metal nodes
and organic linking elements that typically display rigid 3-
dimensional structure and network porosity.4 Porphyrinic
MOFs, in which metallopophyrins are incorporated as linking
elements, have garnered intense interest due to the potential
use of these materials as biomimetic heterogeneous oxidation
catalysts.5 Among the challenges to developing efficient
oxidation catalysis based on porphyrinic materials are: 1)
generating large-pore, non-interpenetrating materials that
support substrate diffusion,5a,6 2) controlling M–L binding to
access materials with open metal sites,7 and 3) accessing
robust frameworks that maintain structure during guest
removal or exchange.8 Even with these criteria met, it remains
A wide variety of network topologies are available for
layered porphyrinic materials.7b,10 For example, while PPF-5,
which is a layered material comprised of Ni(tcpp), 4,4’-
bypyridine (bpy), and dizinc tetracarboxylate nodes features
an AA stacking motif in which dizinc sites are connected by
bpy, PPF-3, which is a layered material comprised of Fe(tcpp),
bpy, and dizinc tetracarboxylate nodes features an AB stacking
This journal is © The Royal Society of Chemistry 20xx
J. Name., 2013, 00, 1-3 | 1
Please do not adjust margins

Please do not adjust margins
ChemComm
Page 2 of 4
View Article Online
DOI: 10.1039/C7CC02570J
COMMUNICATION
Journal Name
Figure 1. Solvothermal synthesis of Fe(btp). (a) PXRD pattern of as-synthesized Fe(btp), (b) simulated PXRD pattern
for Fe(btp), and (c) simulated PXRD pattern for Fe(btp) with application of a [00l] preferred orientation. Structure
model of Fe(btp) along the (d) crystallographic c-axis, and (e) crystallographic a-axis, which was generated using both
PXRD and SCXRD data.
motif, in which dizinc sites are bound to metalloporphyrin indefinitely stable in DEA solutions of btp. Soaking Fe(btp) in
sites.10 We have used a combination of single-crystal X-ray solutions of bpy resulted in the isolation of an isostructural
diffraction (SCXRD), powder X-ray diffraction (PXRD), and layered material Fe(bpy). Similar to Fe(btp), interlayer
modeling to confirm that Fe(btp) features AA-stacking of connectivity of Fe(bpy) could not be resolved by SCXRD (Figure
porphyrinic layers and is isostructural with PPF-5. SCXRD S6), but the PXRD pattern measured for Fe(bpy) indicates that
analysis confirmed the 2-D porphyrinic layers in Fe(btp) are Fe(bpy) is an AA-stacked material isostructural with PPF-5
connected by dizinc tetracarboxylates and indicated AA (Figure 2).10 Additionally, the unit cell parameters of Fe(bpy) (a
stacking of the 2-D layers with alignment of the dizinc units = b = 16.70 Å, c = 14.05 Å; P4/m) are similar to those of PPF-5
along the crystallographic c-axis. Interlayer connectivity, (a = b = 16.65 Å, c = 14.05 Å; P4/mmm) but dissimilar to those
however, could not be established by SCXRD due to severe of PPF-3 (a = b = 16.63 Å, c = 25.08 Å; I4/mmm). We have
disordering in the crystallographic c-axis (see Figure S4 for extended similar post-synthetic linker exchange chemistry to
SCXRD data). PXRD was collected for Fe(btp) and is displayed the preparation of two additional isoreticular frameworks –
in Figure 1. We constructed a model based on the 2-D layers [Zn2(Fe(tcpp))(bpe)(NO3)]n
determined by SCXRD and the layer-to-layer spacing set at (NO3)]n Fe(bpp) which are based on 1,2-bis(4-
15.28 Å, which was calculated from the first main peak at 2 pyridyl)ethane (bpe) and 1,3-bis(4-pyridyl)propane (bpp)
(Fe(bpe)) and [Zn2(Fe(tcpp))(bpp)-
(
)
–
Θ
=
5.781° in the measured PXRD pattern. The measured PXRD linkers, respectively. The obtained isoreticular family of
data are in good agreement with the structure model materials displays linker-dependent layer-to-layer spacing
presented in Figure 1d and 1e. The PXRD data for Fe(btp) (14.04 Å (Fe(bpy)), 15.28 Å (Fe(btp)), 16.39 Å (Fe(bpe)), and
display a strong [00l] preferred orientation, which is common 16.45 Å (Fe(bpp); Figure S7). Characterization data is collected
for layered porphyrinic materials10 and consistent with the in Figures S8–S10 and Tables S1 and S2.
plate morphology of Fe(btp) crystallites. We also simulated the
The original report of PPF-5 explored the solvothermal
PXRD pattern expected of AB stacking (Figure S5) and the chemistry of Zn(NO3)2, Fe(tcpp), and bpy and showed the
observed PXRD is most well-matched to the presented AA evolution of PPF-3, which features an AB stacking motif.10 We
stacking model.
The assignment of AA stacking is further bolstered by Fe(btp) provides access to an AA stacking motif (PPF-5
conversion of Fe(btp) to [Zn2(Fe(tcpp))(bpy)(NO3)]n Fe(bpy) topology), and specifically were interested in addressing
were interested in the observation that linker exchange from
(
)
via post-synthetic linker exchange11 chemistry (Figure 2). whether templating provided access to network topologies not
Fe(btp) is unstable in fresh DEA; 1H NMR analysis of the accessible by direct synthesis. To address this issue, we have
solvent indicated leaching of btp from the material and PXRD carefully examined the impact of bpy loading on the reaction
indicated amorphization of the network. In contrast, Fe(btp) is outcome and have found that Fe(bpy) is obtained from
2 | J. Name., 2012, 00, 1-3
This journal is © The Royal Society of Chemistry 20xx
Please do not adjust margins

Please do not adjust margins
ChemComm
Page 3 of 4
View Article Online
DOI: 10.1039/C7CC02570J
Journal Name
COMMUNICATION
Figure 3. Epimerization of cis-decalin-derived radicals
provides a mechanism to generate both cis- and trans-1
during radical hydroxylation.
abstraction (HAA) from C9 generates a tertiary radical that can
undergo epimerization (Figure 3).14 The ratio of cis- and trans-
Figure 2. Comparison of the PXRD pattern of Fe(bpy)
with the simulated and experimental PXRD patterns for
decahydronaphthalen-9-ol (1) provides information regarding
PPF-3 and PPF-5 materials, which feature AB- and AA- the rate of hydroxylation relative to epimerization. The
stacking motifs, respectively.
modest preference for stereoretention during oxidation
catalyzed by soluble metalloporphyrins has been rationalized
as arising from radical rebound within a solvent cage.15
Treatment of Fe(btp) with cis-decalin and 2-tert-
solvothermal synthesis at low bpy loading, while PPF-3 is
obtained at higher loading (Figure S11). Similar stoichiometry-
dependent phase isolation was observed when btp loading
was varied in Fe(btp) syntheses (Figure S12).
butylsulfonyliodosylbenzene,16
atom transfer reagent, results in a >100:1 mixture of cis- and
trans- (Figure 4, entry 1). Fe(btp) could be recycled without
a common soluble oxygen-
Fe(btp) undergoes amorphization upon removal of volatile
solvent guest molecules (Figure S13). Thermogravimetric
analysis (TGA) measured under N2 flow reveals ~50% weight
loss between 23–150 °C (Figure S14) and IR spectroscopy
shows that this mass loss is associated with removal of DEA
(Figure S3). Gas sorption isotherms (up to 1 bar of pressure)
were measured for the desolvated solid (Fe(btp)act) with N2
and CO2 at 77 K and 195 K, respectively, and indicate that
Fe(btp)act has very low accessible internal surface area (i.e. 89
m2/g based on CO2 isotherm as compared to theoretical
surface area of 3580 m2/g, Figure S15). Network structure, as
probed by PXRD analysis, did not return after soaking
Fe(btp)act in DEA and EtOH (Figure S13). Additionally, in situ
PXRD experiments carried out at the Advanced Photon Source
(APS) while sequentially subjecting Fe(btp)act to vacuum and
then to high pressure gas (up to 100 bar CH4 or 30 bar CO2
pressure) indicate that the vacuum-induced phase transition of
Fe(btp) to Fe(btp)act is irreversible (Figures S16 and S17).
For application to oxidation catalysis, porosity in the
absence of solvent is not required, only maintenance of
porosity during guest exchange is necessary.12 The layered
materials presented here readily participate in guest exchange.
Soaking Fe(btp) in a CH3CN solution of fluorescent dye
rhodamine 6G13 results in the intercalation of dye into the
material (30 wt%), as evidenced by UV-vis analysis of both the
soaking liquid and the digested solids (Table S3 and Figure
S18). The ability for solution-phase guest exchange reactions
to proceed demonstrates that the porosity of our materials is
maintained in the presence of suitable guests. In contrast, no
1
loss of activity or selectivity. Removal of Fe(btp) and addition
of oxidant to the filtrate does not result in cis-decalin oxidation
indicating that Fe(btp) functions as a heterogeneous catalyst.
Molecular catalysts Fe(tpp)Cl and Fe(tmp)Cl afforded 3.3:1 and
8.5:1 mixtures, respectively (entries 7 and 8; tpp: 5,10,15,20-
tetraphenylporphyrin; tmp: 5,10,15,20-tetrakis(4-methoxy-
carbonylphenyl)porphyrin). Similarly, catalysis with Fe(btp)act
,
which does not exhibit dye-uptake activity (Table S3), results in
a 4:1 mixture of cis- and trans-1 (Entry 3). The availability of
both porous (Fe(btp)) and non-porous (Fe(btp)act) materials
with the same chemical composition allows direct examination
of the selectivity of oxidation inside the porous material as
compared with at exposed particle surfaces.
Based on the hypothesis that metalloporphyrin sites
housed within Fe(btp) accomplish highly stereospecific
oxidation while metalloporphyrin sites on the surface of
catalyst particles accomplish oxidation in low stereospecificity,
we evaluated the selectivity that was achieved using ground
samples of our porous materials. Grinding Fe(btp) reduces the
particle size and thus, for a given catalyst loading, increases
the proportion of catalyst sites on the exterior surface of
particles. Examination of the impact of grinding on catalyst
performance has been pursued previously to probe the activity
of the exterior surface of catalyst materials.17
As-synthesized samples of Fe(btp) exhibit a broad particle
size distribution, with many crystallites exhibiting larger than
100 μm edges. Ground samples of Fe(btp) showed narrower
particle size distribution centered at ~10 μm (SEM images are
collected in Figures S19 and S20). Whereas as-synthesized
dye uptake is observed with Fe(btp)act
.
We selected cis-decalin as a substrate for hydroxylation
catalysis using our layered materials for two reasons. First, cis-
decalin presents multiple chemically and stereochemically
Fe(btp) affords a >100:1 mixture of cis- and trans-1, oxidation
with a ground sample of Fe(btp) provides a 50:1 mixture (Entry
2). Importantly, grinding has no impact on dye uptake capacity
indicating that while grinding diminishes the average crystallite
distinct C–H bonds. Second, cis-decalin provides
a
stereochemical probe during oxidation: Hydrogen-atom
This journal is © The Royal Society of Chemistry 20xx
J. Name., 2013, 00, 1-3 | 3
Please do not adjust margins

Please do not adjust margins
ChemComm
Page 4 of 4
View Article Online
DOI: 10.1039/C7CC02570J
COMMUNICATION
Journal Name
Notes and references
1
(a) J. Clardy and C. Walsh, Nature, 2004, 432, 829. (b) B.
Meunier, S. P. de Visser and S. Shaik, Chem. Rev., 2004, 104
3947. (c) L. Que, Jr. and W. B. Tolman, Nature, 2008, 455
333.
,
,
2
3
(a) B. A. Arndtsen, R. G. Bergman, T. A. Mobley and T. H.
Peterson, Acc. Chem. Res., 1995, 28, 154. (b) T. Newhouse
and P. S. Baran, Angew. Chem. Int. Ed., 2011, 50, 3362.
(a) K. Zhang, B. M. Shafer, M. D. Demarsil, H. A. Stern and R.
Fasan, J. Am. Chem. Soc., 2012, 134, 18695. (b) J. Genovino,
S. Lütz, D. Sames and B. B. Touré, J. Am. Chem. Soc., 2013,
135, 12346.
4
5
6
(a) B. F. Abrahams, B. F. Hoskins and R. Robson, J. Am. Chem.
Soc., 1991, 113, 3606. (b) O. M. Yaghi, G. Li and H. Li, Nature,
1995, 378, 703. (c) H. Li, M. Eddaoudi, M. O’Keefe and O. M.
Yaghi, Nature, 1999, 402, 276.
(a) K. S. Suslick, P. Bhyrappa, J. H. Chou, M. E. Kosal, S.
Nakagaki, D. W. Smithenry and S. R. Wilson, Acc. Chem. Res.,
2005, 38, 283. (b) W.-Y. Gao, M. Chrzanowski and S. Ma,
Chem. Soc. Rev., 2014, 43, 5841.
Figure 4. Oxidation of cis-decalin affords a catalyst-
dependent mixture of cis- and trans-1.
(a) M. E. Kosal, J.-H. Chou, S. R. Wilson and K. S. Suslick, Nat.
Mater., 2002, 1, 118. (b) A. M. Shultz, O. K. Farha, J. T. Hupp
size, it has no impact on material porosity (Table S3). A similar
trend is observed when Fe(bpy) is used as an oxidation
catalyst. As synthesized samples afford a 53:1 mixture of cis-
and S. T. Nguyen, J. Am. Chem. Soc., 2009, 131, 4204. (c) O.
K. Farha, A. M. Shutlz, A. A. Sarjeant, S. T. Nguyen and J. T.
Hupp, J. Am. Chem. Soc., 2011, 133, 5652. (d) X.-L. Lv, K.
Wang, B. Wang, J. Su, X. Zou, Y. Xie, J.-R. Li and H.-C. Zhou, J.
Am. Chem. Soc., 2017, 139, 211.
(a) E. Y. Choi, P. M. Barron, R. W. Novotny, C. H. Hu, Y. U.
Kwon and W. Y. Choe, CrystEngComm, 2008, 10, 824. (b) E.-
Y. Choi, P. M. Barron, R. W. Novotny, H.T. Son, C. Hu and W.
Choe, Inorg. Chem., 2009, 48, 426. (c) E. Y. Choi, C. A. Wray,
C. H. Hu and W. Choe, CrystEngComm, 2009, 11, 535. (d) P.
M. Barron, C. A. Wray, C. Hu, Z. Guo and W. Choe, Inorg.
Chem., 2010, 49, 10217.
and trans-1 while ground samples afford a 37:1 ratio (Entries 4
and 5).
7
The collected data is consistent with rapid hydroxylation of
radical intermediates inside our porous materials. Relatively
slow hydroxylation of radical intermediates by surface-bound
catalysts provides opportunity for loss of stereospecificity
during oxidation. These observations suggest that our porous
materials effectively function as tight solvent cages and
prevent stereochemical scrambling during hydroxylation
chemistry. Further, the observed relationship between the
outer surface area of catalyst particles and reaction
8
9
B. F. Abrahams, B. F. Hoskins, D. M. Michail and R. Robson,
Nature, 1994, 369, 727.
(a) C. Serre, F. Millange, C. Thouvenot, M. Noguès, G.
Marsolier, D. Louër and G. Ferey, J. Am. Chem. Soc., 2002,
124, 13519. (b) S. Kitagawa, R. Kitaura and S.-I. Noro, Angew.
Chem. Int. Ed., 2004, 43, 2334. (c) H. Sato, W. Kosaka, R.
Matsuda, A. Hori, Y. Hijikata, R. V. Belosludov, S. Sakaki, M.
Takata and S. Kitagawa, Science, 2014, 343, 167. (d) J. A.
Mason, J. Oktawiec, M. K. Taylor, M. R. Hudson, J. Rodriguez,
J. E. Bachman, M. I. Gonzalez, A. Cervellino, A. Guagliardi, C.
M. Brown, P. L. Llewellyn, N. Masciocchi and J. R. Long,
Nature, 2015, 527, 357.
stereospecificity provides
a useful tool to differentiate
catalysis in from catalysis on these porous materials.
In summary, we have prepared a new isoreticular family of
layered porphyrinic materials with substrate-accessible
unsaturated Fe porphyrin sites. We have shown that these
materials affect highly stereospecific oxidation of cis-decalin
while hydroxylation with homogeneous and non-porous
analogs proceeds with low stereospecificity. Oxidation
stereospecificity provides a chemical probe for differentiating
catalysis in versus catalysis on these porous materials.
Accomplishing catalysis in materials is a critical prerequisite to
utilizing pore structure to control chemical selectivity. We
anticipate that use of cis-decalin oxidation specifically, and
other stereochemical probes generally, may provide useful
tools for evaluating the locus of catalysis with porous
materials.
10 P. M. Barron, H.-T. Son, C. Hu and W. Choe, Cryst. Growth
Des., 2009, , 1960.
11 (a) B. J. Burnett, P. M. Barron, C. Hu and W. Choe, J. Am.
Chem. Soc., 2011, 133, 9984. (b) S. Takaishi, E. J. DeMarco,
M. J. Pellin, O. K. Farha and J. T. Hupp, Chem. Sci., 2013,
1509.
12 L. J. Barbour, Chem. Commun. 2006, 1163.
13 L. Ma, J. M. Falkowski, C. Abney and W. Lin, Nat. Chem.,
2010, , 838.
14 (a) P. D. Bartlett, R. E. Pincock, J. H. Rolston, W. G. Schindel
and L. A. Singer, J. Am. Chem. Soc., 1965, 87, 2590. (b) M.
Zhou, D. Balcells, A. R. Parent, R. H. Crabtree and O.
Eisenstein, ACS Catal. 2012, 2, 208–218.
15 J. T. Groves and T. E. Nemo, J. Am. Chem. Soc., 1983, 105
9
4,
2
We thank Texas A&M and the Welch Foundation (A-1907)
for financial support, John Gladysz for use of
a gas
,
chromatograph, and Yu-Sheng Chen and Wenqian Xu for
assistance with SCXRD and in situ PXRD, carried out at Sectors
15 (NSF/CHE-1346572) and 17 of the APS. Use of the APS, an
Office of Science User Facility operated for the U.S.
Department of Energy (DOE) Office of Science by Argonne
National Laboratory, was supported by the U.S. DOE under
Contract No. DE-AC02-06CH11357.
6243.
16 D. Macikenas, E. Skrzypczak-Jankun and J. D. Protasiewicz, J.
Am. Chem. Soc., 1999, 121, 7164.
17 (a) K. S. Suslick, P. Bhyrappa, J.-H. Chou, M. E. Kosal, S.
Nakagaki, D. W. Smithenry and S. R. Wilson, Acc. Chem. Res.
2005, 38, 283. (b) C. Zou, T. Zhang, M.-H. Xie, L. Yan, G.-Q.
Kong, X.-L. Yang, A. Ma and C.-D. Wu, Inorg. Chem. 2013, 52
3620.
,
4 | J. Name., 2012, 00, 1-3
This journal is © The Royal Society of Chemistry 20xx
Please do not adjust margins