Adsorption-driven self-sorting of dynamic imine libraries

Article abstract of DOI:10.1002/anie.201409741

Differences in adsorption among the components of complex mixtures play a role in separations, surface-based sensing, and heterogeneous catalysis, and have been implicated in theories of the origin of life. Herein, we consider mixtures of imines and we show that if such complex mixtures are also dynamic - that is, if their components equilibrate among themselves - then they can dramatically simplify in composition during the course of column chromatography. As they travel down the column, imines continuously trade their aldehyde and amine constituents, favoring the formation of molecules with extremes of polarity at the expense of species with intermediate polarities. Iterative application of this principle leads to simplification of imine libraries containing up to 16 members into 4 major products.

Full text of DOI:10.1002/anie.201409741

Angewandte
Chemie
DOI: 10.1002/anie.201409741
Dynamic Combinatorial Chemistry
Adsorption-Driven Self-Sorting of Dynamic Imine Libraries**
ˇ
´
Chia-Wei Hsu and Ognjen S. Miljanic*
Abstract: Differences in adsorption among the components of
complex mixtures play a role in separations, surface-based
sensing, and heterogeneous catalysis, and have been implicated
in theories of the origin of life. Herein, we consider mixtures of
imines and we show that if such complex mixtures are also
dynamic—that is, if their components equilibrate among
themselves—then they can dramatically simplify in composi-
tion during the course of column chromatography. As they
travel down the column, imines continuously trade their
aldehyde and amine constituents, favoring the formation of
molecules with extremes of polarity at the expense of species
with intermediate polarities. Iterative application of this
principle leads to simplification of imine libraries containing
up to 16 members into 4 major products.
reached. This key benefit of DCC is also its key drawback as
working in a closed system at equilibrium does not allow the
production of kinetically favored products.
By switching instead to an open system with a continuous
flow of material, kinetic and thermodynamic factors can work
together, so that kinetically controlled products can be
produced in high yields enabled by continuous recycling of
precursors before they enter an irreversible removal pro-
cess.[3] We and others have explored this behavior to achieve
self-sorting,[3,4] dynamic kinetic resolution,[5] and self-replica-
tion.[6] Our work demonstrated that iterative application of an
irreversible kinetic stimulus in an open system results in the
cascade of disproportionation processes, simplifying a library
containing n2 members into just n products. The stimulus for
these self-sorting processes can be a chemical reaction[7] or
a physical transformation, such as distillation,[8] precipita-
tion,[9] or transfer of materials between two liquid phases.[10]
Herein, we show that DCL members can be amplified based
on the difference in their partition coefficients between the
mobile and the stationary phase during the course of column
chromatography.[11]
T
he differential adsorption of compounds from a liquid or
gaseous mixture on solid materials is of critical importance in
the fields of heterogeneous catalysis, surface sensing, separa-
tions, and even in theories of the origin of life.[1] Adsorption of
components of dynamic combinatorial libraries (DCLs)[2]
should additionally result in substantial redistribution of the
material among the equilibrating species so that the best-
adsorbed species are amplified at the expense of their less-
well-adsorbed counterparts. Further benefits can come from
the use of open systems which do not establish a thermody-
namic equilibrium and can amplify even small differences in
equilibrium partition coefficients to yield synthetically useful
product distributions. Herein, we present such a case: simpli-
fication of dynamic libraries containing as many as 16 com-
ponents into just 4 final products during the course of column
chromatography on silica gel.
We hypothesized that slow elution of a dynamic combi-
natorial library constructed from some of the imines A1–E5
(Figure 1) would amplify the least polar component that
Dynamic combinatorial chemistry (DCC) studies equili-
brating compound mixtures which can respond to external
stimuli by increasing the proportion of (i.e. amplifying) those
library components that best adopt to the disturbing stimulus,
at the expense of those that do not. This error-correcting
mechanism allows production of thermodynamically stabi-
lized species in yields that are often quantitative, as the
material can be continuously recycled across shallow poten-
tial-energy surfaces until the thermodynamic minimum is
Figure 1. Structures of imines employed in this study.
elutes first. As this occurred, its more polar counterparts that
share either the same aldehyde or amine starting components
as this least polar imine, would disproportionate to replenish
the least polar imine and to amplify the most polar imine at
the expense of species of intermediate polarity (illustrated on
a mixture of imines B2–C3 in Scheme 1).[12]
ˇ
´
[*] C.-W. Hsu, Prof. O. S. Miljanic
Department of Chemistry, University of Houston
112 Fleming Building, Houston, TX 77204-5003 (USA)
E-mail: miljanic@uh.edu
Each of the examined imine libraries was formed by
mixing equimolar amounts of the requisite aldehydes and
anilines in toluene as the solvent, followed by heating the
mixture at reflux for 12 hours. During the heating, H2O was
removed using a Dean–Stark trap. Once dehydration was
complete, the toluene was evaporated and the crude imine
library mixture was pre-adsorbed onto oven-dried silica with
an approximate mass 2.5 times that of the crude reaction
mixture. This solid mixture was placed on top of a chroma-
[**] This research was supported by the University of Houston and its
Grant to Advance and Enhance Research, the National Science
Foundation (award CHE-1151292), and the Welch Foundation
ˇ
(award E-1768). O.S.M. is a Cottrell Scholar of the Research
Corporation for Science Advancement.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201409741.
Angew. Chem. Int. Ed. 2014, 53, 1 – 5
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1
These are not the final page numbers!

.
Angewandte
Communications
In the first experiment, a mixture of imines B2, B3, C2,
and C3 was prepared and analyzed by 1H NMR spectroscopy.
The relative ratios of the four imines were calculated to be
1.00:0.65:0.66:0.99 (Figure 2a, white vertical bars), respec-
tively. Column chromatography of this mixture was initially
performed with a 100:1 v/v mixture and then with a 20:1 v/v
mixture of hexane and EtOAc. Two major fractions were
isolated. In the first fraction, the dominant product was B2
(85 Æ 1%), as determined by integration of the resonance
1
signals in the H NMR spectrum of this fraction by compar-
ison with an internal standard 1,3,5-trimethoxybenzene.
Similar analysis of the second fraction revealed C3 as the
major component (89 Æ 1%) and a small amount of C2 (8 Æ
0.2%). In essence, fast elution of the least polar imine B2
disturbed the equilibrium distribution, forcing B3 and C2 to
react and replenish the removed B2—until their constituents
were completely consumed in the process. The most polar
imine C3—which did not share either of the components of
B2—was also amplified in the mixture at the expense of its
counterparts of intermediate polarity (Figure 2a, black bars).
As this experiment required the addition of a significant
amount of silica gel to the imine mixture, we next confirmed
that this addition does not significantly change the equilib-
rium. The initial mixture was combined with an approxi-
mately 2.5 times greater mass of silica gel and then heated at
reflux in toluene overnight. After the removal of silica gel by
filtration, integration of the resonance signals in the 1H NMR
spectrum of the resultant solution revealed a B2:B3:C2:C3
ratio of 1:0.96:0.88:0.80, suggesting that silica gel does not
dramatically change the ratio of the library members (Fig-
Scheme 1. Simplification of a [2ꢀ2] imine library during the course of
column chromatography. The least polar imine B2 combines the two
black components and travels fastest through the column. As it is
being depleted, the two crossover components (B3 and C2) react to
produce more of B2, while simultaneously amplifying the most polar
fraction C3, which combines the two gray “sticky” components.
tography column (inner diameter 150 mm) which was pre-
loaded with a layer of oven-dried silica gel (10–20 cm high).
Finally, the part of the column containing the silica gel was
wrapped in a heating tape, which was used to keep the column
at a temperature of approximately 508C.
Figure 2. Relative normalized distributions of imine components in dynamic libraries before (white bars) and after (gray bars) addition of silica
gel. The black bars show the total yields of individual imines from the isolated fractions, which were determined using the integrals of resonance
signals in the 1H NMR spectra with an internal standard.
2
www.angewandte.org
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2014, 53, 1 – 5
These are not the final page numbers!

Angewandte
Chemie
ure 2a, gray bars). As only the solution composition was
monitored, the measured changes in library composition can
be interpreted as a consequence of differences in adsorption
on silica gel among the various imines in the mixture.
fraction composed of C1 (5 Æ 0.1%) and C2 (8 Æ 0.1%);
e) the fifth fraction made up mostly of C3 (80 Æ 1%) and
some C1 (5 Æ 0.1%); and f) a final fraction as a mixture of D4
(76 Æ 3%), D2 (2 Æ 0.1%), D1 (2 Æ 0.1%), and A4 (1 Æ
0.1%). Conceptually, rapid elution of A1 forced imines
A2–A4 to react with B1, C1, and D1 to replenish A1. After
elution of A1 is completed, the original 16-imine library lost
these seven members and was reduced in size to form a
[3 ꢀ 3] library which now had B2 as the least polar member. In
the next step, its elution consumed B3, B4, C2, and D2,
decreasing the size of the leftover library to just 4 members,
and so on.
Two additional [2 ꢀ 2] experiments were performed. In the
first (Figure 2b), imines B2, B4, D2, and D4 were similarly
analyzed with and without silica gel, before being subjected to
column chromatography. This mixture was biased from the
outset: the equilibrated library shows a high preference for
the formation of the least polar imine B2 (relative abundance
1.00) and the most polar D4 (0.88), relative to their counter-
parts of intermediate polarity (B4: 0.18, D2: 0.20). This ratio
changed minimally upon addition of silica gel. Ultimately,
chromatography—eluting first with a mixture of hexane/
EtOAc (20:1 v/v) and then with pure THF—led to the
isolation of B2 (82 Æ 1%) as the dominant component of the
first fraction, followed by D4 (79 Æ 3%) as the major
component in the second fraction. Small amounts of B4
(3 Æ 0.1%) and D2 (8 Æ 0.3%) were also detected.
This procedure is limited in resolution. Each of the
isolated imines is eluted in several fractions and that is
a necessary feature of this procedure. For example, in the
experiment shown in Figure 2d, imine B2 is eluted as three
separate fractions, although they may not be collected as such.
The amount of B2 that is present in the initial mixture is
eluted quickly. The remainder of B2 has to be produced on
the column by disproportionation of partners that a) contain
its constituents and b) elute at approximately the same rate.
Thus, B3 and C2—which are of comparable polarity—
produce the second batch of B2, while B4 and D2 elute
later and are responsible for the production of the third batch
of B2. With larger libraries, it is likely that there will be
overlap of multiple fractions, and thus the amplification loses
fidelity as the library increases in complexity.
In the final [2 ꢀ 2] system (Figure 2c), the initial distribu-
tion was between the first two experiments—biased towards
the least polar imine C3 and the most polar imine D4, but not
as dramatically as in the previous experiment. Column
chromatography significantly amplified this bias, producing
C3 in 96 Æ 1% yield and D4 in 93 Æ 3% yield.
In a more complex [3 ꢀ 3] system, three aldehyde and
three amine starting materials were reacted together to yield
a mixture of nine imines, the distribution of which is also
represented by white vertical bars (Figure 2d). The propor-
tion of the most abundant imine in the mixture, imine C3, is
approximately six times greater than that of the least
abundant member of the library, imine B4. Upon addition
of silica, this distribution equalizes somewhat, with the C3/B4
ratio decreasing to approximately 3.5. Column chromatog-
raphy of the system initially employed pure hexane as the
eluent. The polarity of the eluent was then increased to
hexane/EtOAc 100:1 v/v and then to a 10:1 ratio, and was
finally changed to pure THF. Imine B2 was isolated in the first
fraction in 80 Æ 1% yield, essentially consuming (almost) all
of benzaldehyde and aniline constituents in the process. The
second small fraction contained 12 Æ 0.2% of C2, while the
third fraction carried C3 (81 Æ 1%). In the final fraction,
imine D4 dominated (89 Æ 3%), and small amounts of D2
(4 Æ 0.1%) were also detected.
Based on the same logic, an even more curious trans-
mutation experiment was performed (Scheme 2). Equimolar
amounts of pure D5 and E4 (which both contain one highly
Scheme 2. Transmutation of imines during column chromatography.
polar and one highly nonpolar component) were loaded onto
a silica gel column, and eluted first with hexane/EtOAc (40:1
v/v) and subsequently with pure THF. The first isolated
fraction contained D4 (92%) and the second E5 (94%). This
is a particularly unusual column chromatography experiment,
given that two compounds eluted from the column are
different to the two compounds loaded onto it.
In conclusion, we have shown that complex libraries of
equilibrating compounds can simplify in composition during
the course of column chromatography on silica gel. This
process could easily be combined with chemical reactions—
by, for example, performing a chromatographic separation on
a column impregnated with a catalyst—resulting in multi-
dimensional simplification of complex libraries.
In the future, we will attempt to automate and monitor
this procedure using an HPLC instrument. Such an extension
would allow the use of this dynamic procedure as a physical
chemistry tool to, for example, determine polarity indica-
tors[13] in a direct competition experiment. This method will
also be expanded to other adsorbents and dynamic compound
classes.[14]
The final [4 ꢀ 4] experiment started with a mixture of 16
imines in which the most abundant imine B2 was present in
approximately threefold excess relative to the least abundant
imine B4 (Figure 2e, white bars). Upon addition of silica gel
the relative abundances changed, leaving C2 as the imine with
the highest proportion and D1 as the least concentrated
member of the library ([C2]/[D1] = 2.44). Column chroma-
tography—eluting first with hexane/toluene (10:1 v/v, then
8:1), then with hexane/EtOAc (40:1, then 10:1, then 2:1), and
finally with pure THF—enabled us to isolate six distinct
fractions: a) the first fraction composed mostly of A1 (64 Æ
1%); b) the second fraction with small amounts of A2 (1 Æ
0.1%) and B1 (3 Æ 0.1%); c) the third fraction, which was
a mixture of B2 (78 Æ 1%), A2 (17 Æ 0.1%), B1 (13 Æ 0.1%),
and an additional amount of A1 (3 Æ 0.02%); d) the fourth
Angew. Chem. Int. Ed. 2014, 53, 1 – 5
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
3
These are not the final page numbers!

.
Angewandte
Communications
K.-W. Chi, P. J. Stang, Acc. Chem. Res. 2009, 42, 1554 – 1563;
e) J. R. Nitschke, Acc. Chem. Res. 2007, 40, 103 – 112.
Experimental Section
Representative self-sorting experiment (Figure 2a): Equimolar
amounts of 4-methylbenzaldehyde (110 mL, 0.933 mmol), 4-methoxy-
benzaldehyde (114 mL, 0.933 mmol), 4-methylaniline (100 mg,
0.933 mmol), and 4-methoxyaniline (115 mg, 0.933 mmol) were
[5] a) F. F. Huerta, A. B. E. Minidis, J. E. Bꢂckvall, Chem. Soc. Rev.
2001, 30, 321 – 331; b) P. Vongvilai, M. Angelin, R. Larsson, O.
Ramstrçm, Angew. Chem. Int. Ed. 2007, 46, 948 – 950; Angew.
Chem. 2007, 119, 966 – 968; c) P. Vongvilai, R. Larsson, O.
Ramstrçm, Adv. Synth. Catal. 2008, 350, 448 – 452; d) P. Vongvi-
lai, O. Ramstrçm, J. Am. Chem. Soc. 2009, 131, 14419 – 14425;
e) M. Sakulsombat, P. Vongvilai, O. Ramstrçm, Org. Biomol.
Chem. 2011, 9, 1112 – 1117; f) M. Sakulsombat, Y. Zhang, O.
Ramstrçm, Chem. Eur. J. 2012, 18, 6129 – 6132; g) Y. Zhang, L.
Hu, O. Ramstrçm, Chem. Commun. 2013, 49, 1805 – 1807.
[6] a) V. Patzke, G. von Kiedrowski, ARKIVOC 2007, 293 – 310;
b) V. del Amo, D. Philp, Chem. Eur. J. 2010, 16, 13304 – 13318;
c) V. del Amo, A. M. Z. Slawin, D. Philp, Org. Lett. 2008, 10,
4589 – 4592; d) J. W. Sadownik, D. Philp, Angew. Chem. Int. Ed.
2008, 47, 9965 – 9970; Angew. Chem. 2008, 120, 10113 – 10118;
e) V. C. Allen, C. C. Robertson, S. M. Turega, D. Philp, Org. Lett.
2010, 12, 1920 – 1923.
added into
a 50 mL round-bottom flask, along with toluene
(20 mL). The mixture was heated at reflux for 12 h, and a Dean–
Stark trap was used to remove H2O. Toluene was then removed under
vacuum to afford a mixture of four imines as a light brown solid. This
mixture was preadsorbed onto silica gel (approx. 1 g) and then added
on top of a chromatography column, which was pre-loaded with oven-
dried silica gel (approximately a 10 cm high layer). The column was
then wrapped with a heating tape and heated to approximately 508C.
The elution was performed first with hexane (300 mL), followed by
a 100:1 hexane/EtOAc (v/v, 200 mL), and 10:1 hexane/EtOAc (v/v,
110 mL) mixtures, and finally with pure EtOAc (100 mL). Three
fractions were isolated. In the first fraction, imine B2 was the
dominant product (85% yield, based on an internal standard
calculation with 171 mg of the crude product and 22.3 mg of the
internal standard, see Figure S1). The second fraction contained
imine C2 as the major component (8% yield, based on an internal
standard calculation with 27 mg of the crude compound and 9.7 mg of
the internal standard, see Figure S2). The final fraction contained
imine C3 (89% yield, based on an internal standard calculation with
224 mg of the crude compound and 21 mg of the internal standard, see
Figure S3).
ˇ
´
[7] K. Osowska, O. S. Miljanic, J. Am. Chem. Soc. 2011, 133, 724 –
726.
ˇ
´
[8] a) Q. Ji, N. El-Hamdi, O. S. Miljanic, J. Chem. Educ. 2014, 91,
ˇ
´
830 – 833; b) Q. Ji, O. S. Miljanic, J. Org. Chem. 2013, 78, 12710 –
ˇ
´
12716; c) K. Osowska, O. S. Miljanic, Angew. Chem. Int. Ed.
2011, 50, 8345 – 8349; Angew. Chem. 2011, 123, 8495 – 8499; See
also: d) A. Herrmann, Chem. Eur. J. 2012, 18, 8568 – 8577; e) B.
Buchs (nꢃe Levrand), W. Fieber, F. Vigouroux-Elie, N. Sreeni-
vasachary, J.-M. Lehn, A. Herrmann, Org. Biomol. Chem. 2011,
9, 2906 – 2919; f) B. Buchs (nꢃe Levrand), G. Godin, A. Trachsel,
J.-Y. de Saint Laumer, J.-M. Lehn, A. Herrmann, Eur. J. Org.
Chem. 2011, 681 – 695.
Full experimental data, including the procedures for the calcu-
lations of yields using an internal standard, can be found in the
Supporting Information.
Received: October 3, 2014
Revised: November 22, 2014
Published online: && &&, &&&&
ˇ
´
[9] a) R. C. Lirag, K. Osowska, O. S. Miljanic, Org. Biomol. Chem.
2012, 10, 4847 – 4850; b) M. Hutin, C. J. Cramer, L. Gagliardi,
A. R. M. Shahi, G. Bernardinelli, R. Cerny, J. R. Nitschke, J. Am.
Chem. Soc. 2007, 129, 8774 – 8780; c) S. J. Cantrill, K. S. Chichak,
A. J. Peters, J. F. Stoddart, Acc. Chem. Res. 2005, 38, 1 – 9;
d) C. D. Meyer, R. S. Forgan, K. S. Chichak, A. J. Peters, N.
Tangchaivang, G. W. V. Cave, S. I. Khan, S. J. Cantrill, J. F.
Stoddart, Chem. Eur. J. 2010, 16, 12570 – 12581.
Keywords: chemoselectivity · column chromatography ·
.
dynamic combinatorial chemistry · self-sorting ·
systems chemistry
[10] a) G. Vantomme, N. Hafezia, J.-M. Lehn, Chem. Sci. 2014, 5,
1475 – 1483; b) N. Hafezi, J.-M. Lehn, J. Am. Chem. Soc. 2012,
134, 12861 – 12868; c) R. Pꢃrez-Fernꢄndez, M. Pittelkow, A. M.
Belenguer, J. K. M. Sanders, Chem. Commun. 2008, 1738 – 1740.
[1] For relevance of differential adsorption for the origin of life, see:
a) D. A. M. Zaia, Amino Acids 2004, 27, 113 – 118; b) S. J.
Sowerby, C. A. Cohn, W. M. Heckl, N. G. Holm, Proc. Natl.
Acad. Sci. USA 2001, 98, 820 – 822.
ˇ ˇ
[11] a) A. Ciesielski, M. El Garah, S. Haar, P. Kovarꢅcek, J.-M. Lehn,
P. Samorꢆ, Nat. Chem. 2014, 5, 1017 – 1023; For resin-bound
DCLs, see: b) A. V. Gromova, J. M. Ciszewski, B. L. Miller,
Chem. Commun. 2012, 48, 2131 – 2133; c) B. R. McNaughton,
B. L. Miller, Org. Lett. 2006, 8, 1803 – 1806; See also: d) A. M.
[2] a) J. N. H. Reek, S. Otto, Dynamic Combinatorial Chemistry,
Wiley-VCH, Weinheim, 2010; b) B. L. Miller, Dynamic Combi-
natorial Chemistry In Drug Discovery, Bioorganic Chemistry,
and Materials Science, Wiley, Hoboken, 2009; c) P. T. Corbett, J.
Leclaire, L. Vial, K. R. West, J.-L. Wietor, J. K. M. Sanders, S.
Otto, Chem. Rev. 2006, 106, 3652 – 3711; d) S. J. Rowan, S. J.
Cantrill, G. R. L. Cousins, J. K. M. Sanders, J. F. Stoddart,
Angew. Chem. Int. Ed. 2002, 41, 898 – 952; Angew. Chem. 2002,
114, 938 – 993; e) R. F. Ludlow, S. Otto, Chem. Soc. Rev. 2008, 37,
101 – 108.
ˇˇ ´
Belenguer, T. Friscic, G. M. Daya, J. K. M. Sanders, Chem. Sci.
2011, 2, 696 – 700.
[12] Strictly speaking, column chromatography sorts compounds not
based on their polarity, but on their affinity for silica gel.
Although the two concepts are not the same, there is a strong
correlation between them and—in the jargon of preparative
organic chemistry—the two terms are used interchangeably with
polar compounds being defined as having a large affinity for
silica gel. For the sake of brevity, we will continue this tradition
and use the term “polar” to signify compounds which we
experimentally found to elute later.
[13] R. M. Smith, Retention and Selectivity in Liquid Chromatogra-
phy, Elsevier, Amsterdam, 1995.
[14] Y. Jin, C. Yu, R. J. Denman, W. Zhang, Chem. Soc. Rev. 2013, 42,
6634 – 6654.
ˇ
´
[3] Q. Ji, R. C. Lirag, O. S. Miljanic, Chem. Soc. Rev. 2014, 43, 1873 –
1884.
[4] a) M. M. Safont-Sempere, G. Fernandez, F. Wꢁrthner, Chem.
ˇ
´
Rev. 2011, 111, 5784 – 5814; b) K. Osowska, O. S. Miljanic,
Synlett 2011, 1643 – 1648; c) S. Ghosh, L. Isaacs in Dynamic
Combinatorial Chemistry in Drug Discovery, Bioorganic
Chemistry, and Materials Science (Ed.: B. L. Miller), Wiley,
Hoboken, 2010, pp. 155 – 168; d) B. H. Northrop, Y.-R. Zheng,
4
www.angewandte.org
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2014, 53, 1 – 5
These are not the final page numbers!

Angewandte
Chemie
Communications
Dynamic Combinatorial Chemistry
C.-W. Hsu,
ˇ
´
O. S. Miljanic*
&&&& — &&&&
Adsorption-Driven Self-Sorting of
Dynamic Imine Libraries
Mix and match: Complex and dynamic
imine libraries with n2 members can be
sorted into just n final products during
the course of column chromatography on
silica gel. Through a cascade of dispro-
portionation reactions, the most and
least polar imines in the mixture are
amplified at the expense of their coun-
terparts of intermediate polarity.
Angew. Chem. Int. Ed. 2014, 53, 1 – 5
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
5
These are not the final page numbers!