Weidman 2016

Subscriber access provided by UQ Library
Article
Unusually Stable Hysteresis in the pH-Response of Poly(Acrylic Acid)
Brushes Confined within Nanoporous Block Polymer Thin Films
Jacob Logan Weidman, Ryan A Mulvenna, Bryan W. Boudouris, and William A. Phillip
J. Am. Chem. Soc., Just Accepted Manuscript • DOI: 10.1021/jacs.6b01618 • Publication Date (Web): 12 May 2016
Downloaded from http://pubs.acs.org on May 13, 2016
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Journal of the American Chemical Society is published by the American Chemical

Society. 1155 Sixteenth Street N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.
Page 1 of 39 Journal of the American Chemical Society
1
2
3
4 Unusually Stable Hysteresis in the pH-Response of Poly(Acrylic Acid)
5 Brushes Confined within Nanoporous Block Polymer Thin Films
6
7 Jacob L. Weidman,† Ryan A. Mulvenna,‡ Bryan W. Boudouris,‡ and William A. Phillip†,*
8
9 †
10 Department of Chemical and Biomolecular Engineering, University of Notre Dame, Notre
11 Dame, IN 46556-5637, United States
12
13 ‡
School of Chemical Engineering, Purdue University, West Lafayette, Indiana 47907, United
14 States
15
16
17 * To whom correspondence should be addressed: wphillip@nd.edu
18
19 Abstract
20
21 Stimuli-responsive soft materials are a highly-studied field due to their wide-ranging
22
23 applications; however, only a small group of these materials display hysteretic responses to
24
25 stimuli. Moreover, previous reports of this behavior have typically shown it to be short-lived. In
26
this work, poly(acrylic acid) (PAA) chains at extremely high grafting densities and confined in
27
28 nanoscale pores displayed a unique long-lived hysteretic behavior caused by their ability to form
29
30 a metastable hydrogen bond network. Hydraulic permeability measurements demonstrated that
31
32 the conformation of the PAA chains exhibited a hysteretic dependence on pH, where different
33
34
effective pore diameters arose in a pH range of 3 to 8, as determined by the pH of the previous
35 environment. Further studies using Fourier transform infrared (FTIR) spectroscopy demonstrated
36
37 that the fraction of ionized PAA moieties depended on the thin film history; this was
38
39 corroborated by metal adsorption capacity, which demonstrated the same pH dependence. This
40
41 hysteresis was shown to be persistent, enduring for days, in a manner unlike most other systems.
42 The hypothesis that hydrogen bonding among PAA units contributed to the hysteretic behavior
43
44 was supported by experiments with a urea solution, which disrupted the metastable hydrogen
45
46 bonded state of PAA toward its ionized state. The ability of PAA to hydrogen bond within these
47
48 confined pores results in a stable and tunable hysteresis not previously observed in homopolymer
49
50
materials. An enhanced understanding of the polymer chemistry and physics governing this
51 hysteresis gives insight into the design and manipulation of next-generation sensors and gating
52
53 materials in nanoscale applications.
54
55 Keywords. stimuli-responsive soft materials; nanoscale confinement of polymer chains;
56 hysteresis; block polymer self-assembly; pH-responsive nanoporous thin films; hydrogen
57 bonding; poly(acrylic acid)
58
59
60
1
ACS Paragon Plus Environment
Journal of the American Chemical Society Page 2 of 39
1
2
3
4 Introduction
5
6 Stimuli-responsive materials are a major field of materials science and their impact
7
8 continues to burgeon as their ability to provide a real-time response to changes in their
9
10
11
environment (and, thus, provide on-demand sensing information) can be used to control
12
13 processes in a ready manner.1 For instance, environmental stimuli that have been shown to
14
15 produce responses in materials include changes to the relative basicity or acidity (i.e., pH),2–4
16
17
18 temperature,5–7 or the presence of specific chemicals.8,9 These stimuli-responsive materials
19
20 generally follow a reversible, one-to-one relationship between the environment quality and the
21
22 material property. For example, for a given pH, the material will correspondingly have a given
23
24
25 property, regardless of whether the material was at a higher or lower pH previously. However, in
26
27 some cases, which may be extremely useful, materials display hysteretic properties, where
28
29
30
multiple states of the material can exist in a single environment depending on the history of the
31
32 material.10–15 Polymeric materials that display a hysteresis in response to external stimuli may
33
34 allow for the further development of ionic circuits for use in numerous applications, including
35
36
37 smart drug delivery systems and point-of-care diagnostics.16 This type of hysteretic response to
38
39 environmental stimuli is rarely reported in polymeric membranes, which tend to display
40
41 reversible responses to external stimuli. The ability to remember the previous conditions is
42
43
44 typically a short-term effect that fades as time passes and the material reaches its equilibrium
45
46 state,11–15 and it is exceedingly rare for a material to display a hysteresis that does not decay with
47
48 sufficiently long periods of time.10 Conversely, we demonstrate here that when poly(acrylic acid)
49
50
51 (PAA) moieties are confined at a high density within the nanopores of a self-assembled block
52
53 polymer thin film, they exhibit an unusually stable and persistent hysteresis in their response to
54
55
56
pH.
57
58
59
60
2
1
2
3
In more traditional cases, the nature of the material response depends on the complex
4
5
6 interactions between the material and it surroundings. For many weak polymer electrolytes,
7
8 which have reversible pH-responses, the equilibrium between the ionized and neutral forms of
9
10
11
the polymer is determined by the pH of the solution.17–23 The induced electrostatic interactions
12
13 along the chain then play a major role in the final conformation of the polymer, which also has to
14
15 balance the entropic forces that can inhibit the chains from attaining the most extreme extended
16
17
18 conformations. The interplay between these two energies results in polymeric materials that have
19
20 relatively small average sizes for neutral chain conditions and relatively large average size in
21
22 conditions that promote ionization.24 However, in the presence of additional attractive enthalpic
23
24
25 interactions, the material response can be impacted in a complex manner that may yield multiple
26
27 stable states at a single environmental condition. One such interaction is the formation of ionic
28
29
crosslinks for materials where multiple oppositely-charged chemical moieties are present.10
30
31
32 Another is the formation of hydrogen bonds, which can happen in homopolymers containing
33
34 acidic hydrogens as part of a carboxylic or amide group.25,26 For instance, we demonstrate here
35
36
37 that PAA can be manipulated into high density configurations that display this complex behavior
38
39 through the self-assembly and non-solvent induced phase separation (SNIPS) method of
40
41 membrane casting of block polymers.
42
43
44 Casting via the SNIPS method generates thin films with self-assembled nanopores from
45
46 dissolved block polymers that contain moieties with significant chemical dissimilarities.27–30
47
48 During the casting, sufficient evaporation from the top layer of the drying thin film concentrates
49
50
51 the block polymer at the film-air interface, increasing the interactions between the unique
52
53 moieties of the block polymer chains and eventually favors the formation of distinct domains of
54
55
56
separate chemistries on the nanometer scale.27–29 The final structure of these domains is highly
57
58
59
60
3
1
2
3
dependent on the block polymer chemistry, composition of the block polymer, and the conditions
4
5
6 used during the casting process.31,32 For many materials that are studied for water filtration
7
8 applications, the preferred configuration is that of cylindrical domains aligned perpendicular to
9
10
11
the membrane surface, resulting in pores that allow flow though the membrane.3,33–35 This
12
13 nanoporous thin film has been used as a membrane when a polyisoprene-b-polystyrene-b-
14
15 poly(acrylic acid) (PI-PS-PAA) triblock polymer was cast from solution.36 The pore-lining PAA
16
17
18 moiety has been shown to be pH-responsive and have a high enough density that it competes
19
20 with other copper binding materials as a metal adsorber.37 This high density of PAA chains
21
22 confined within the nanoscale features of the membrane allows for hydrogen bonding under
23
24
25 favorable pH conditions, as well as limited conformational freedom for the polymer chains.
26
27 These factors, in combination with the pH-dependent ionization, allow it to have a hysteretic
28
29
response to changes in environment.
30
31
32 In this current study, PI-PS-PAA-based thin films containing nanoscale pores lined with
33
34 the poly(acrylic acid) (PAA) moiety were fabricated. These thin films were used as membranes,
35
36
37 and they displayed a pH-response that changed the size of the nanoscale pores when they were
38
39 challenged with a variety of aqueous solutions, as measured in permeability experiments. Unlike
40
41 most materials that have a reversible pH response, the membranes were shown to have hysteretic
42
43
44 permeabilities, with the value of the hydraulic permeability under DI water conditions differing
45
46 by a factor of 6-8 times difference depending on whether the membrane had previously been in a
47
48 basic or acidic solution. The hysteresis was consistently repeatable and a long-lived
49
50
51 phenomenon. The electrostatic properties of the materials also displayed this unique dependence
52
53 on environmental history, as did their performance as metal adsorbers. The deviation from a
54
55
56
reversible pH response was attributed to the metastable hydrogen bond network of PAA formed
57
58
59
60
4
1
2
3
under acidic conditions, which allowed the membrane material to retain information about the
4
5
6 history of the membrane conditions. The elucidation of this unusually-long hysteretic
7
8 phenomenon provides significantly deeper insights into the physics of these materials and how to
9
10
11
manipulate the chemistry, geometry, and environmental conditions of nanoporous polymer thin
12
13 films in order to generate next-generation sensors38 and gating materials.39
14
15
16 Results and Discussion
17
18 Hysteresis in the Physical Properties of Nanoporous Block Polymer Membranes
19
20 The self-assembled, nanoporous membranes used here were generated following the
21
22 procedure described previously.36 In brief, the polyisoprene-b-polystyrene-b-poly(N,N-
23
24
25 dimethylacrylamide) (PI-PS-PDMA) block polymer was synthesized via a reversible addition-
26
27 fragmentation chain transfer (RAFT) polymerization.40,41 Subsequently, the isolated PI-PS-
28
29
30
PDMA was dissolved in organic solvents and cast into a nanoporous thin film through the SNIPS
31
32 method. A subsequent reaction of the membrane in a 6 M hydrochloric acid bath converted the
33
34 PDMA block to a PAA functionality, which resulted in the PI-PS-PAA structure shown in Figure
35
36
37 S1a. The PDMA block that had lined the pore walls was fully converted to PAA, resulting in a
38
39 surface like that shown in Figure S1b, where the PI and PS blocks (represented with red in the
40
41 cartoon) are the matrix of the membrane, while the PAA (shown by the blue entities) is the
42
43
44 moiety that lines the pore walls. Following the functionalization reaction, the membrane surface
45
46 retained its high density of consistently-sized pores, as shown by the SEM micrograph in Figure
47
48 S1c. Because the PDMA moieties, which are subsequently converted to PAA moieties, are
49
50
51 introduced during the synthesis of the block polymer precursor we can confine PAA brushes
52
53 within the nanopores of thin films cast using the SNIPS method at higher densities than can be
54
55
56
achieved using standard grafting-to or grafting-from methodologies. Estimates of this density
57
58
59
60
5
1
2
3
suggest it is two to three times larger than the densities reported for other methods of introducing
4
5
6 polymers onto surfaces.42,43 It is likely that this high surface density of polymer chains is critical
7
8 to the hysteretic behavior discussed in this work.
9
10
11 In addition to directing the assembly of nanoscale pores on the membrane surface, the
12
13
14 SNIPS method generates an asymmetric structure with large open channels beneath the
15
16 nanoporous active layer. The large openings in this underlying support layer provide little
17
18 resistance to flow compared to the active layer, which provides the majority of the resistance due
19
20
21 to its small pores. The flux of solution through the membrane is a function of the applied
22
23 pressure and the hydraulic permeability, as shown in Equation 1.
24
25
26 (1)
27
28
29 Here, Jw is the volumetric flux of the permeate through the membrane, ∆P is the applied pressure,
30
31 and Lp represents the hydraulic permeability of the membrane. According to the scaling of the
32
33
34 Navier-Stokes equation in the low Reynolds number limit, the permeability depends on a
35
36 characteristic size, in this case pore diameter at the upper surface (D).44 For a membrane with a
37
38
collection of cylindrical pores, the Hagen-Poiseuille equation can be used to calculate the
39
40
41 hydraulic permeability, as shown as Equation 2, which has a dependence on the pore diameter
42
43 (D) to the fourth power.
44
45
46
47 επ D 4 (2)
48 Lp =
49 128µ L
50
51 Here, ε is the porosity of the thin film, µ represents the viscosity of the solution, and L represents
52
53
54 the length of the pore. Thus, changes in the pore diameter modify the overall permeability of the
55
56
57
58
59
60
6
1
2
3
membrane, or alternatively, observed changes in membrane permeability suggest corresponding
4
5
6 changes in the pore diameter.
7
8 The hydraulic permeabilities of the PI-PS-PAA membranes were measured in stirred cell
9
10
11
experiments (see the Supporting Information for the full experimental procedure). Solutions of
12
13 varying pH were added to a cell containing a 1-inch diameter circular membrane sample. The
14
15 solutions were prepared at different pHs by dilution of the following acids or bases with DI
16
17
18 water: hydrochloric acid (pH 1.0-3.0), citric acid (pH 3.3-5.0),
19
20 tris(hydroxymethyl)aminomethane (pH 6.5-9.5), and sodium hydroxide (pH 10.5-13.0). A
21
22 nitrogen gas line was used to apply pressure to the cell in order to push the solution through the
23
24
25 membrane. The output of the membrane testing device led to a vial on a scale so that the mass of
26
27 solution permeated was readily measured as a function of time. The volumetric flux of water
28
29
through the membrane was determined by dividing the slope of the accumulated mass versus
30
31
32 time data by the area of the membrane and the density of water. Dividing this measured flux by
33
34 the pressure applied (i.e., ∆P) yielded the permeabilities displayed in Figure 1a.
35
36
37 At the beginning of the experiment, with the membrane exposed to a pH 2.0 solution
38
39 (corresponding to Point 1 in Figure 1) for two hours, the membrane permeability was 10.9 L m-2
40
41 h-1 bar-1. Following this, solutions of incrementally higher pH were added to the stirred cell, and
42
43
44 left to equilibrate with the membrane for ~20 minutes. The experiments began and permeated a
45
46 total of ~2 mL for each data point over a timespan ranging from 0.5 to 12 hours, and the
47
48 corresponding permeabilities were measured. The results of these measurements, which are
49
50
51 denoted by the red triangles in Figure 1, were executed with the solution pH changing as shown
52
53 (i.e., moving from left to right in Figure 1). The permeability decreased between pH 3.0 and pH
54
55
56
4.0, then remained constant at ~6.1 L m-2 h-1 bar-1 in DI water at pH 5.5 (Point 2, Figure 1).
57
58
59
60
7
1
2
3
Further increasing the pH resulted in a gentle decrease in permeability until the solution pH was
4
5
6 around pH 9.5 where the permeability neared its lowest value of 0.9 L m-2 h-1 bar-1. The
7
8 permeability remained constant at this value through a pH of 13.0.
9
10
11 Beginning from a pH 13.0 solution (Point 3, Figure 1) and decreasing the pH, the
12
13
14 permeability values matched that of the increasing pH experiments through a pH of 9.5.
15
16 However, the permeability values showed clear hysteresis beyond this point, as the
17
18 permeabilities below this pH remained near 1.3 L m-2 h-1 bar-1 through to a pH of 4.0, including
19
20
21 DI water at pH 5.5 (Point 4). Below pH 4.0, the permeability began to rise sharply. At a solution
22
23 acidity of pH ~3.2, the permeability value for the experiments executed with solutions of
24
25 decreasing pH rejoined the permeability values from the curve for experiments conducted with
26
27
28 increasing pH. For solutions with a pH below this value through a solution pH of 2.0, the
29
30 measured permeabilities matched. It is noted that between the pH values of 3.2 and 8.3, the
31
32
33
permeabilities did not match between the two curves. This clear hysteresis in the permeability
34
35 with respect to pH is of significant interest as controlling the previous pH conditions of the
36
37 solution to which the membrane was exposed will impact the permeability in the near DI water
38
39
40 (i.e., typical operating) conditions. Continued experiments in both directions showed the trend
41
42 was repeatable and consistent. Furthermore, this phenomenon was observed across membranes
43
44 fabricated from several block polymer samples. That is, the PI-PS-PAA material consistently
45
46
47 produced nanoporous thin films that can have different permeabilities at a single pH value
48
49 depending on whether it was previously in a high or low pH solution, and this phenomenon is
50
51
unique relative to other reports of PAA in the literature.17,45–47
52
53
54
55
56
57
58
59
60
8
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31 Figure 1. Membrane Permeability Displays a Hysteretic Dependence on pH. (a) The
32
33 hydraulic permeability of a PI-PS-PAA block polymer-based membrane as a function of the
34 solution pH. The membrane was fabricated from the PI-PS-PDMA-44 block polymer sample.
35 Two series of experiments were implemented. First, a membrane began in a pH = 1.0 solution
36 and its permeability was measured as the membrane was exposed to solutions of incrementally
37 higher pH (red symbols); this is referred to as increasing pH. Second, a membrane began in a pH
38
39
= 13.0 solution and the permeability was measured as the membrane was exposed to solutions of
40 incrementally lower pH (blue symbols); this is referred to as decreasing pH. (b) Schematic
41 diagrams of the pore structure and PAA conformation corresponding to the four points noted in
42 Figure 1a. The PI and PS block segments form the matrix of the membrane, and are represented
43 as the red region surrounding the pore in the illustration. The PAA chains that line the pore walls
44 are represented in blue. (1) In a pH 2.0 solution, the PAA chains are not extended, resulting in a
45
46 pore with a relatively large diameter. The carboxylic acid groups of the PAA are protonated and
47 are capable of forming hydrogen bonds with another PAA moiety. (2) After addition of a base to
48 increase to a pH 5.5 solution, some of the PAA repeat units are deprotonated resulting in the
49 PAA chains extending slightly and decreasing the pore size. Some of the PAA repeat units
50 remain hydrogen bonded to their counterparts, which prevents the full extension of the PAA
51
52
brushes. (3) In a pH 13.0 solution, most of the PAA repeat units are deprotonated and
53 electrostatically repel one another, which results in highly extended chain conformation that
54 reduces the pore diameters. (4) In DI water (pH = 5.5) after exposure to a basic solution, most of
55 the PAA groups remain deprotonated, and the pores retain a very small diameter. The process is
56 repeatable and reversible.
57
58
59
60
9
1
2
3
Therefore, this unusual phenomenon must be related to the exclusive chemistry of the triblock
4
5
6 polymer and the confined geometry afforded by membranes fabricated using the SNIPS casting
7
8 procedure.
9
10
11
The observed changes in permeability result from changes in the diameters of the pores
12
13 within the active layer of the membrane, as we have demonstrated previously.36 These are, in
14
15 turn, determined by the physical and chemical interactions between the repeat units of the PAA
16
17
18 chains that line the pore walls. Because these repeat units contain a carboxylic acid moiety, their
19
20 interactions with the surrounding solution may cause them to protonate or deprotonate,
21
22 depending on the solution pH. At pH 1, the chains are highly protonated, as depicted in Figure
23
24
25 1b. Protonated carboxylic acids, including PAA, have the ability to form hydrogen bonds with
26
27 other carboxylic acids (e.g., repeat units along the same chain or with units on neighboring PAA
28
29
chains).10 This results in the possibility of some PAA units forming a hydrogen-bonded network,
30
31
32 as shown schematically in the top left panel of Figure 1b. In turn, the resulting pore has a
33
34 relatively large diameter due to the collapsed state of the uncharged PAA chains, which
35
36
37 corresponds to the highest permeabilities observed in Figure 1a. Increasing the solution pH by
38
39 adding a DI water solution deprotonates some PAA repeat units that are not involved in
40
41 hydrogen bonds, as shown in the top right panel of Figure 1b. Many hydrogen bonds remain
42
43
44 intact, while the partial charging of the free (i.e., repeat units not associated with hydrogen
45
46 bonds) PAA units results in the extension of the PAA chains toward the center of the pore,
47
48 decreasing the observed permeability to nearly half of that of the acidic state.
49
50
51 Adding a basic solution (pH 13.0) results in the deprotonation of almost all of the PAA
52
53 units, inducing a greater extension of the PAA chains toward the center of the pore due to
54
55
56
electrostatic repulsion between the repeat units of like charges. The electrostatic repulsion
57
58
59
60
10
1
2
3
between PAA units is strong enough to disrupt the hydrogen bonded network. The ionized PAA
4
5
6 chains swell to the point that the pores are nearly closed off in the base state, as shown by the
7
8 bottom right panel of Figure 1b, which is plausible given the radius of gyration of the PAA block
9
10
11
is ~25 nm and the dry state radius of the pores is ~20nm. This, combined with the high surface
12
13 density of PAA (between 0.5 to 1 chains nm-2) along the pore walls, results in the pore volume
14
15 being nearly completely occupied by PAA chains and a near zero permeability. Adding a more
16
17
18 acidic DI water solution to the membrane results in the re-protonation of some PAA units,
19
20 though still only a small fraction. The electrostatic interactions that are still present dominate the
21
22 PAA chain response and cause the pore to retain its nearly closed state, as displayed in the
23
24
25 bottom left of Figure 1b. Addition of the original acidic solution re-protonates the chains and
26
27 induces hydrogen bonding, returning the system to the state shown schematically in the top left
28
29
panel of Figure 1b. This cycle can be repeated with the same states achieved in a consistently
30
31
32 reproducible manner.
33
34 The effects of solution pH on the permeability of membranes with pores lined by PAA
35
36
37 brushes have been investigated in several prior studies.17,45–47 However, in these prior
38
39 investigations no hysteretic behavior was noted for the PAA-lined membranes. Rather, the
40
41 permeability of the membrane displayed a reversible response to pH with no dependence on the
42
43
44 previous state of the membrane. Above the pKa of PAA, the chains were deprotonated and
45
46 swelled. Below the pKa, the PAA units became neutral (i.e., protonated) and assumed a less
47
48 extended conformation. While the pKa value of acrylic acid monomers is 4.3, the pKa value of
49
50
51 homopolymer PAA dissolved in a bulk solution is higher (6.5) due to charge regulation (i.e., the
52
53 repulsive electrostatic interactions along the polymer chain driving the acid-base equilibrium
54
55
56
toward protonated PAA repeat units).18,48 Additionally, the pKa value of polymers in confined
57
58
59
60
11
1
2
3
spaces may deviate slightly, in the case of a weak acid to an even higher value, as the excluded
4
5
6 volume interactions increase the unfavorable interaction energies of the ionized units.18,49 It is
7
8 assumed that the effective pKa of the PAA in this system lies between the sharp changes in
9
10
11
permeability observed at pHs of 3.4 and 7.5, though the actual value remains unclear due to the
12
13 hysteretic behavior.
14
15 Hysteretic events have been observed in the pH response of some membranes and other
16
17
18 polymeric materials.10,12–15 However, it is extremely rare for a material to display long-lived
19
20 hysteretic swelling; for instance, this has been reported in a gel that contained both positively and
21
22 negatively charged moieties, but not in a porous membrane.10 Rather, almost all experiments
23
24
25 with hysteretic films or membranes demonstrated that the hysteresis was kinetic in nature and,
26
27 given sufficient time, would disappear.12–15 The kinetic limitations could be associated with
28
29
polymer chain relaxation or lack of solvent access to all of the chains within the membrane
30
31
32 pores. By allowing sufficient time for polymer chain relaxation, the chains were able to reach
33
34 equilibrium conformations that are not achieved immediately after changing pH conditions.12,14
35
36
37 Alternatively, the residual behavior could be a result of needing sufficient permeation of a
38
39 solution to allow the system to return to equilibrium.15 The return to equilibrium could be
40
41 achieved in these cases within seconds to hours of being left to equilibrate. For this reason,
42
43
44 permeability experiments on long time scales with DI water solutions were performed on the
45
46 membranes.
47
48 Specifically, membranes were exposed to a pH 1.0 solution, 5 mL of which was
49
50
51 permeated through the membrane to ensure that the pore walls throughout the active layer of the
52
53 membrane achieved equilibrium with the solution. Then the cell was washed several times with
54
55
56
DI water and left to soak for 1 h before adding a fresh 10 mL of DI water, and the membrane
57
58
59
60
12
1
2
3
permeability was measured as a function of time. After 4 mL had permeated, the cell was
4
5
6 emptied and a fresh 10 mL of DI water was added. This was repeated until a total of 12 mL had
7
8 permeated through the membrane. The measured permeabilities are displayed by the red squares
9
10
11
in Figure 2. Initially, the permeability decreased over the first 3 mL in the presence of the DI
12
13 water solution. The pore-lining PAA groups swelled partially during this time as the DI water
14
15 deprotonated more of the carboxylic acid units as the water permeated though the pores. This
16
17
18 value then leveled after 3 mL and remained at the relatively high value of 8.5 L m-2 h-1 bar-1
19
20 during the rest of the experiment, which lasted several hours. A similar experiment was
21
22 performed after the membrane had been exposed to a pH 13.0 solution, with the measured
23
24
25 permeabilities shown by the blue circles in Figure 2. The permeability did not change
26
27 significantly throughout the permeation of 12 mL of DI water over a period of 10 hours. This
28
29
suggests that the hysteresis is neither kinetically-trapped by insufficient relaxation time, nor
30
31
32 limited by lack of access of the bulk solution to PAA repeat units within the pores. Rather the
33
34 membrane remains in long-lived metastable states that have different pore sizes at the same pH.
35
36
37 The approximated pore diameters shown by the cartoons in Figure 1b are based on the
38
39
40 observed permeabilities in Figure 1a. These are assumed to be affected greatly by the differences
41
42 in electrostatic repulsion in the different pH conditions, which is determined by how many of the
43
44 PAA units are charged. One straightforward way to establish the degree of ionization is through
45
46
47 FTIR spectroscopy, which has been shown to display different absorption peaks for the different
48
49 states of PAA.18,50,51 Protonated carboxylic acid units absorb at wavenumbers of ~1720 cm-1.
50
51
This is true for both hydrogen-bonded PAA (1729 cm-1) and free protonated PAA (1712 cm-1),
52
53
54 though the observed peaks may be difficult to deconvolute.51 The deprotonated form of PAA has
55
56 several absorption peaks, one of which occurs around 1550 cm-1.18 The relative areas of these
57
58
59
60
13
1
2
3
characteristic peaks can be used to determine the relative ionization of PAA in membranes under
4
5
6 different experimental conditions. Thus, membrane pieces were prepared by soaking them in pH
7
8 1.0 or pH 13.0 solutions for at least 1 h, followed by soaking the membranes in a solution
9
10
11
prepared at another pH. The solutions used were prepared in the same manner as described above
12
13 for the permeability experiments. The membranes were left to soak in the solutions of various pH
14
15 values for 4 h, then pulled from the solution, dabbed with a porous cloth to visual dryness, and
16
17
18 vacuum-dried for 20 m.
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
Figure 2. Hysteretic States are Long-lived and are affected by the Addition of Urea. The
41 hydraulic permeability of a PI-PS-PAA block polymer membrane as a function of the volume of
42 DI water permeated. The membrane was fabricated from the PI-PS-PDMA-97 block polymer
43 sample. All measurements were collected using a solution at pH 5.5. Data points represent the
44 average value for each 1 mL of permeate collected. The history of the solution passing across the
45 membrane is summarized in the legend with the terminal entry being the condition used during
46
47 permeability measurements. A membrane was rinsed with DI water after being exposed to a pH
48 1.0 solution, then the DI water permeability was measured (red squares). A 1 M urea solution
49 was then added to the stirred cell and the permeability was evaluated (magenta triangles). The
50 cell was then filled with DI water again and the permeability was measured (purple triangles). A
51 membrane that was exposed to a pH 13.0 solution was rinsed with DI water and measured for its
52
53
DI water permeability (blue circles). This was followed immediately by permeability
54 measurements of a 1 M urea solution (green triangles).
55
56
57
58
59
60
14
1
2
3
The samples were vacuum-dried in order to remove water that was present within the hydrophilic
4
5
6 pores that would absorb in the range of 1650 cm-1, disrupting the readings of the protonated and
7
8 deprotonated peaks. A sample Fourier transform infrared (FTIR) spectra of a membrane soaked
9
10
11
in a pH 1.0 solution is shown in Figure S2, with the most prominent peak in the plotted range
12
13 occurring between 1700 and 1750 cm-1.
14
15
16 Four representative FTIR spectra are shown in Figure 3a, and these are labeled 1–4 in
17
18 order to correspond with the preparation conditions shown in Figure 1. After exposure to a pH
19
20
21 1.0 solution, the resulting FTIR spectrum shown in the red curve and labeled with a 1 displayed a
22
23 broad peak near the 1720 cm-1 wavenumber, suggesting the presence of protonated PAA groups.
24
25 A membrane that was soaked in pH 1.0 solution followed by DI water, shown by the magenta
26
27
28 curve (Trace 2), provided a similar peak in the 1720 cm-1 region. This suggests that the PAA
29
30 chains in these two states are mostly protonated, as was depicted in Figure 1b. Note that it is
31
32
33
difficult to determine if the PAA forms hydrogen bonds using this experimental technique due to
34
35 the potential for overlap of the hydrogen-bonded and free PAA signals. Upon exposure to pH
36
37 13.0 solution, the resulting FTIR spectrum shown in blue (Trace 3) displayed a peak near 1550
38
39
40 cm-1, and the acquired FTIR spectrum of a membrane soaked in DI water following the soak in
41
42 basic solution showed similar absorption, as shown by the green curve with the label 4. This
43
44 suggests that the PAA groups were mostly deprotonated, as was depicted in Figure 1b. Similar to
45
46
47 the permeability measurements in DI water, there was a significant difference in the FTIR
48
49 spectra of the membranes soaked in DI water after exposure to highly acidic and basic solutions,
50
51
(magenta and green curves), respectively. This difference demonstrates that the ionization of
52
53
54 PAA units in DI water is strongly dependent on the previous solution conditions surrounding the
55
56 membrane.
57
58
59
60
15
1
2
3
Additional FTIR spectra were collected for pieces of membranes over a range of pH
4
5
6 values, using solutions prepared in the same manner as those described for the permeability
7
8 experiments. The degree of ionization was observed as the solution pH changed incrementally
9
10
11
between the four states shown in Figure 3a. The differences in ionization were quantified using
12
13 area calculations for the peaks at 1720 cm-1 and 1550 cm-1, where the fraction of the area under
14
15 the 1550 cm-1 peak divided by the area under both peaks gave the value plotted on the vertical
16
17
18 axis for the data in Figure 3b. Samples that were initially soaked in pH 1.0 solution, followed by
19
20 a solution of a higher pH are shown by the red triangles, which showed a small increase in
21
22 percent ionization from near 10% at pH 1.0 to around 25% under DI water conditions. The
23
24
25 ionization then increased sharply between pH 6.5 and pH 7.5 to near 75% ionization, as more
26
27 PAA chains were able to deprotonate and attain their equilibrium conformation as per the
28
29
mechanism proposed above. The blue triangles display the percent ionization for samples that
30
31
32 were initially soaked in a pH 13.0 solution before soaking in a lower pH solution. The percent
33
34 ionization began near 85%, then remained near this value, dropping only to around 75% under
35
36
37 DI water conditions. The percent ionization showed a sharp decrease between the pH values of
38
39 4.5 and 3.5, before the ionization levels of the membranes agreed with samples initially soaked
40
41 in acid below a pH of 2.5. The region where the two curves do not lie on top of one another show
42
43
44 that a hysteresis in the ionization levels occurs, which matches well with the permeability
45
46 hysteresis shown in Figure 1. While somewhat counterintuitive, the relatively gradual change
47
48 over multiple pHs from low to high ionization levels (and vice versa) is consistent with theory
49
50
51 for densely packed polymers grafted to a surface.52 A comparison between the permeability and
52
53 the percent ionization is also displayed as Figure S3, and it shows a strong correlation between
54
55
56
the two metrics. This backs the hypothesis that the observed hysteresis that arises in the physical
57
58
59
60
16
1
2
3
properties of the block polymer membrane is correlated with the discontinuous change in the
4
5
6 electrostatic interactions between the repeat units of the polymer chains. This discontinuous
7
8 change in the electrostatic interactions is caused by hydrogen bonds between the PAA repeat
9
10
11
units preventing some of the repeat units from deprotonating.
12
13 Hysteresis in Chemical Properties of Membrane
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
17
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39 Figure 3. Degree of Ionization of PAA Shows a Hysteretic Dependence on pH. (a) FTIR
40 spectra of vacuum-dried membranes after exposure to the four pH conditions labeled in Figure
41 1a. These conditions are the membrane: (1) exposed to a pH 1.0 solution (red line); (2) soaked in
42 a pH 5.5 solution (magenta line), after exposure to a pH 1.0 solution; (3) exposed to a pH 13.0
43
44 solution (blue line); and (4) soaked in a pH 5.5 solution, after exposure to a pH 13.0 solution
45 (green line). The peak at ~1712 cm-1 in the red and magenta curves corresponds to protonated
46 carboxylic acids, and the peak at ~1560 cm-1 in the blue and green curves corresponds to
47 deprotonated carboxylic acids. (b) The relative degree of ionization of the PAA chains as a
48 function of solution pH. The pH was increased following exposure to a pH 1.0 solution (red
49
50
triangles) or decreased following exposure to a pH 13.0 solution (blue triangles). The values of
51 percent ionization were determined from the peak areas in the normalized FTIR spectra.
52
53 In addition to controlling the physical properties of the membrane, the ionization of the
54
55 PAA groups controls the chemical properties of the membranes. One property that relies on
56
57
58 ionized PAA repeat units is the adsorption of copper ions, which has been discussed in previous
59
60
18
1
2
3
literature.37,53,54 In our prior studies, membranes that were exposed to a basic solution, and thus
4
5
6 had mostly ionized PAA groups, were soaked in CuCl2 solutions in DI water. Copper ions were
7
8 reversibly adsorbed by the negatively-charged PAA units, and the membranes were regenerated
9
10
11
with the addition of a pH 1 solution, which shifted the equilibrium to favor protonated PAA
12
13 repeat units and released the copper ions back into solution. The concentration of copper in the
14
15 retentate solution and the pH 1 copper regeneration solution were determined via ultraviolet-
16
17
18 visible (UV-Vis) light spectroscopy, resulting in measured copper uptakes for the membrane
19
20 under different conditions. In this study, the hysteretic behavior of the copper uptake was
21
22 assessed in a similar manner. Several 8 mM CuCl2 solutions were prepared at pH values between
23
24
25 1.0 and 7.5 via dilution of hydrochloric acid, citric acid, and tris(hydroxymethyl)aminomethane
26
27 (above 8.0 the copper hydroxide may precipitate). Pieces of membrane were soaked in either a
28
29
pH 1.0 or pH 13.0 solution for at least 30 m, then removed and dipped quickly in a large DI
30
31
32 water bath. This was done to remove any residual solution on the membrane before placing them
33
34 into the 8 mM CuCl2 solutions. The nanoporous thin films were left for 8 hours to adsorb copper,
35
36
37 then removed and placed in the pH 1.0 solutions to release the bound copper.
38
39
40 The amount of copper bound to the membrane in these experiments was determined from
41
42 the depletion of copper from the retentate solution and the concentration of the acid wash
43
44 solution by the characteristic absorption of copper ions at λ = 930 nm. These data are plotted in
45
46
47 Figure 4a as a function of the pH of the copper solution. The red data points indicate the
48
49 membrane pieces that were initially soaked in pH 1 solutions, while the blue data points are
50
51
membranes initially in pH 13.0 solutions. For the membranes that had soaked in a pH 1 solution,
52
53
54 the amount of copper adsorbed was very low, even in the DI water solution, before increasing to
55
56 near 0.45 mmol g-1 at a pH of 7.5. For the membranes soaked in pH 13.0 solutions, the uptake
57
58
59
60
19
1
2
3
was significantly higher at 7.5, and remained high in DI water, showing a large difference from
4
5
6 the uptake after exposure to acid.
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25 Figure 4. Copper Binding in Membranes is Hysteretic and Tunable. (a) The amount of
26 copper adsorbed by membranes exposed to solutions of varying pH. After the membrane was
27 soaked in a copper-free solution of pH-adjusted water [pH 1.0 (red arrows), pH 13.0 (blue
28
29
arrows), or pH 13.0 followed by pH 3.5 (orange circles)], it was placed in a 8 mM CuCl2 solution
30 at the pH values indicated on the horizontal axis. (b) Photographs of the membrane samples after
31 soaking in an 8 mM CuCl2 solution at pH 5.5. Prior to submerging the membrane in the copper
32 chloride solution, they were (i) soaked in a pH 13.0 solution and (ii) soaked in pH 1.0 solution.
33 Sample (C) is a control membrane that was not exposed to a copper-containing solution.
34
35
36 This can also be observed visually, as shown in Figure 4b, as two pieces of membrane that were
37
38 soaked in a (i) basic or (ii) acidic solution were added to an 8 mM CuCl2 solution at pH 5.5, then
39
40 removed and photographed after adsorbing copper. The membrane after exposure to the basic
41
42
43 solution appeared blue, while the membrane that was soaked in acidic solution had a similar hue
44
45 to the membrane not placed in a copper solution (c). Continuing on to lower pH values, for the
46
47
48
pieces initially soaked in basic solution, the membrane uptake decreased at pH 3.5, while still at
49
50 a higher value than the acidic membranes at this pH, then finally matched the pH 1 uptake. The
51
52 hysteresis in copper adsorption followed a trend similar to that which was observed for values of
53
54
55 PAA ionization percentage as a function of pH. A comparison is shown in Figure S4, where the
56
57 PAA percent ionization in the membrane, as measured by FTIR spectroscopy, was compared to
58
59
60
20
1
2
3
the PAA ionization level based on copper uptake at different pH values. These values were
4
5
6 calculated using the assumption that a Langmuir isotherm, which is defined by Equation 3, holds
7
8 true for the system,
9
10
11 QKC
12 q= (3)
13 1 + KC
14
15 where q is the amount of copper adsorbed in mmol g-1 membrane, Q is the maximum capacity of
16
17
copper binding for a given membrane state in mmol g-1 membrane, C is the concentration of
18
19
20 copper in solution, and K is an equilibrium constant. The concentrations of copper in solution
21
22 and copper adsorbed on the membrane were known experimentally. Solving the equation at a
23
24
25 given condition allowed the value of K to be determined. By assuming that the number of
26
27 available binding sites arises from the number of negatively charged PAA units, the value of Q
28
29 determined at any given condition and the fraction of PAA repeat units that are ionized are
30
31
32 proportional, allowing for the estimation of the PAA percent ionization based on the copper
33
34 uptake under each pH condition. These calculations are described in more detail in Section S6 of
35
36 the Supporting Information. These data are plotted, in addition to the FTIR-measured percent
37
38
39 ionizations, in Figure S4. A comparison of the percent ionization levels determined using the two
40
41 techniques demonstrated a similar trend. This is consistent with Figure S5 where the percent
42
43
44
ionization determined from FTIR was compared directly to the copper uptake values. From these
45
46 data, it can be seen that the observed hysteresis in the electronic state of the membrane also
47
48 occurs very similarly in the chemical properties of the membrane.
49
50
51 Further studies attempted to determine if the pH values, which showed intermediate
52
53 levels of PAA ionization, would display corresponding intermediate levels of copper uptake.
54
55 This was successfully demonstrated by the experiments detailed in Section S7 of the Supporting
56
57
58 Information. Whereby, using careful conditioning of the membrane, the pH-dependent properties
59
60
21
1
2
3
could be tuned by controlling the exposure of the membrane to solutions of varying pH. In one
4
5
6 specific example, the copper uptake can be trapped at intermediate values, as shown in Figure 4.
7
8 That is, the orange data points in Figure 4 demonstrated this point and support the assertion that
9
10
11
the observed pore sizes strongly correlate with chemical properties.
12
13 A final copper uptake experiment was performed, in which membranes that had
14
15 previously soaked in acidic (pH 1.0) or basic (pH 13.0) solutions were added to scintillation vials
16
17
18 containing DI water as before, but left for ~150 h instead of ~3 h. The membranes were then
19
20 removed from the DI water solutions and added to vials of 8 mM CuCl2 to determine the copper
21
22 uptake. Based on both the depletion of the copper from the feed solution and the amount released
23
24
25 in the acid wash, values for the copper uptake were determined. The membrane that was initially
26
27 soaked in basic solution had a resulting uptake of 0.72 ± 0.05 mmol g-1 membrane, while the
28
29
membrane initially soaked in acidic solution had a resulting uptake of 0.35 ± 0.06 mmol g-1.
30
31
32 These values fall just outside the range of the error of the short soak time experiments, but
33
34 remain distinct from one another. This suggests that unlike other materials that show a short-
35
36
37 lived hysteresis on the order of minutes to hours, the PI-PS-PAA membrane displays a hysteresis
38
39 that endures up to a week in solution, and possibly longer.12,14
40
41 Hysteresis Arises from the Hydrogen Bonding of PAA
42
43
44 The pores of the nanoporous block polymer thin film show hysteretic and tunable
45
46 behavior in both physical (i.e., pore size) and chemical (i.e., copper binding) properties, but the
47
48 cause of the hysteresis is not elucidated merely by the presence of these properties. If the state of
49
50
51 the PAA chains that line the pore walls in DI water after exposure to either acidic or basic
52
53 solution is the local minimum in free energy, a disturbance of the PAA chains from the
54
55
56
metastable state may drive the PAA conformation toward to the global minimum in free energy.
57
58
59
60
22
1
2
3
It has been noted that PAA has the ability to form hydrogen bonds with itself when it is placed in
4
5
6 acidic conditions that cause the carboxylic acid moieties to protonate. It is possible that a
7
8 network of hydrogen bonds may persist in DI water after the membrane was initially exposed to
9
10
11
acidic conditions, as shown in Figure 1b. In this case, the DI water does not deprotonate enough
12
13 PAA repeat units to perturb the system sufficiently to break the network of hydrogen bonds. This
14
15 results in a metastable state, where the ionization level would be higher if all of the PAA units
16
17
18 were free rather than some of them being locked in hydrogen bonds. This keeps the PAA in a
19
20 more collapsed conformation, resulting in larger pores with less ionization of the PAA units.
21
22 Because the difference between hydrogen-bonded and free PAA in FTIR spectroscopy is not
23
24
25 easily measured, experiments were performed that introduced urea, which can be used as a
26
27 hydrogen bond disruptor in solution,55,56 to the nanopores. That is, two membrane pieces were
28
29
placed in DI water following exposure to an acidic (red) or basic (blue) solution, and they were
30
31
32 then evaluated using FTIR spectroscopy (Figure 5a). Similarly, two more pieces were placed in
33
34 the same conditions followed by soaking in 300 mM urea for 2 h, and then they were evaluated
35
36
37 using FTIR spectroscopy. The spectra for the urea-soaked membranes after initially soaking in
38
39 an acidic solution are displayed in Figure 5a as the magenta curve, with membranes initially
40
41 soaked in a basic solution shown by the green curve. For membranes initially soaked in a basic
42
43
44 solution, the change of the spectra before and after the addition of urea is minimal. However, the
45
46 change in spectrum for the membrane that was initially soaked in an acidic solution is
47
48 significant, with a large increase in the degree of ionization. This was further supported by
49
50
51 copper uptake experiments shown in Figure 6b, where the amount of copper uptake is plotted
52
53 against the concentration of urea in the soaking solution before being added to the copper
54
55
56
solution. The membranes that had been exposed to urea after exposure to acid showed an
57
58
59
60
23
1
2
3
increase in copper uptake above 100 mM of urea. This demonstrates that the presence of urea
4
5
6 disrupts hydrogen bonds between PAA repeat units, and results in an increase in the degree of
7
8 ionization of the PAA repeat units.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45 Figure 5. Urea Disrupts Hydrogen-Bonded PAA Network. (a) FTIR spectra of membranes
46 exposed to a solution containing urea. After exposure to either an acidic (pH 1.0) or a basic (pH
47
48
13.0) solution and subsequent soaking in DI water or DI water then a 300 mM aqueous solution
49 of urea, the membranes were dried under vacuum and analyzed. (b) Copper uptake of
50 membranes after soaking the membranes in solutions of varying urea concentrations (10 mM,
51 100 mM, 300 mM) after an initial exposure to pH 1.0 (red squares) or pH 13.0 (blue circles). The
52 membranes were removed from the urea solutions and placed in copper solutions at pH 5.5 and
53
left to adsorb copper overnight. A subsequent soak in pH 1 acid released the copper ions, and the
54
55 copper concentrations of both the retentate and acid solutions were used to determine the amount
56 of copper adsorbed in the membrane.
57
58
59
60
24
1
2
3
Urea disrupts the hydrogen bonds between PAA repeat units, which allows more acrylic
4
5
6 acid groups to equilibrate with the surrounding solution, resulting in an increase in the ionized
7
8 form of PAA. This, in turn, increases the magnitude of the repulsive electrostatic interactions
9
10
11
between PAA chains, which is expected to cause chain extension, further decreasing pore size,
12
13 and therefore, the permeability of the membrane. This was established in permeability
14
15 experiments, immediately following the DI water permeability tests in Figure 2. After the
16
17
18 membrane was exposed to acid, followed by DI water, the permeability leveled to a value near
19
20 9.0 L m-2 h-1 bar-1, as shown by the red squares. After the addition of a 1 M urea solution to the
21
22 cell, the permeability decreased and achieved a steady-state value 6.2 L m-2 h-1 bar-1after 2-3 mL
23
24
25 had permeated, as shown by the magenta triangles in Figure 2, which held steady through 8 mL
26
27 of permeate. To test if the urea solution had only temporarily disrupted the state of the pores, the
28
29
cell was washed and an additional 10 mL of DI water was added to the cell. The addition of urea
30
31
32 after the membrane was exposed to DI water after basic solution resulted in no noticeable change
33
34 in permeability, as shown by the green triangles. This suggests that the pore size decreases after
35
36
37 addition of urea, due to both the breaking of hydrogen bonds that limit the extension of the PAA
38
39 and the increase in ionization. Breaking hydrogen bonds perturbs the membrane to a state with
40
41 smaller pores, suggesting that the presence of hydrogen bonds drives the observed hysteretic
42
43
44 behavior in the PI-PS-PAA membranes.
45
46
47 Generally in polyelectrolytes a balance between the conformational free energy from
48
49 entropic considerations and repulsive interactions of the ionized repeat units (which can vary by
50
51
roughly 200 kJ mol-1 for these PAA chains) control the polymer conformation in a reversible
52
53
54 manner.24 However, the energy associated with the formation of a hydrogen bonded network
55
56 provides a source of deviation from more typical reversible conformational changes. The
57
58
59
60
25
1
2
3
stability of the hydrogen bonds, which have typical bond strengths of 10-20 kJ mol-1, provides a
4
5
6 sufficient decrease in free energy in the system to hold the PAA chains in the protonated form
7
8 rather than the deprotonated form.25,26 For an average PAA chain length of 200 units, the
9
10
11
hydrogen bonding of ~10-20% of the units would result in an interaction energy on the same
12
13 order of magnitude as the repulsive electrostatic forces. As such, a hydrogen-bonded network
14
15 could reasonably compete energetically with the electrostatic interactions to produce a local
16
17
18 minimum in the free energy landscape that prevents the deprotonation of PAA repeat units and
19
20 polymer extension in DI water after exposure to an acidic solution. The addition of urea is able to
21
22 partially overcome this barrier due to the breaking of some hydrogen bonds and move the system
23
24
25 toward the equilibrium state. However, it is noted that both the decrease in permeability (Figure
26
27 2) and the increase in copper uptake (Figure 5b), did not result in values equal to those of the
28
29
highly ionized (i.e., base exposed) form in DI water. Experiments with urea binding showed that
30
31
32 at pH 5.5 residual urea was still present, and did not complete vacate the membrane, until a more
33
34 basic solution was added. This suggests that the urea is bound within the membrane in DI water
35
36
37 and that this prevents it from attaining equilibrium, shown by the blue data at pH 5.5 in both
38
39 figures. Still, the addition of urea resulted in the disruption of the hydrogen bonded PAA
40
41 network of the pore walls, which acts as the main driver of the hysteretic response to changes in
42
43
44 pH.
45
46 Heat was also investigated as a means for disrupting the metastable conformation of the
47
48 PAA chains within the membrane pores. Membrane sections were soaked in a pH 1 solution,
49
50
51 then placed into a vial containing ~2mL of DI water. The vial was heated to moderate
52
53 temperatures (i.e., 40 °C, 55 °C, and 70 °C). Experiments above 70 °C were not executed
54
55
56
because the PI-PS matrix material begins to soften at this temperature. After 16 hours at the
57
58
59
60
26
1
2
3
elevated temperature, the vials were cooled in a water bath for ~5 minutes. The copper uptake
4
5
6 capacities of the membrane sections were quantified as described previously. The results of these
7
8 experiments, which are displayed in Figure S8, demonstrate that the increased temperature has a
9
10
11
negligible effect on the uptake of copper in this system, which suggests that the metastable
12
13 hydrogen bonded network is not disrupted upon heating to 70 °C. This result is sensible given a
14
15 comparison of the relative energy within a hydrogen bond (~10 kJ mol-1) to the energy added
16
17
18 upon increasing the temperature of the system (~0.3 kJ mol-1). The relatively mild heating from
19
20 25 °C to 70 °C (the upper temperature limit for this polystyrene-based membrane system) does
21
22 not provide sufficient energy to perturb the metastable hydrogen bonded network to the more
23
24
25 ionized state.
26
27 Conclusions
28
29 The chemistry and confined geometry of the weak electrolyte polymer evaluated in this
30
31
study demonstrates that control over these crucial nanoscale properties leads to tunable behavior
32
33
34 of a nanoscale material that has a memory of the previous environment. This hysteretic behavior,
35
36 which is rarely reported in membrane systems, is possible due to the uniquely high density of
37
38
39
PAA chains that line the pore walls of membranes fabricated using the SNIPS method.
40
41 Nanoporous thin films containing pores with a PAA lining that were exposed to basic solutions
42
43 followed by DI water resulted in extended, highly-charged PAA chains that resulted in low pore
44
45
46 diameters and high copper uptake, while exposure to acidic solutions followed by DI water
47
48 resulted in more relaxed, less charged PAA chains that resulted in larger pore diameters and
49
50 minor copper uptake. This hysteresis was shown to repeatable and long-lived through continued
51
52
53 permeability and copper adsorption experiments. FTIR studies demonstrated that the PAA had
54
55 different levels of ionization in DI water depending on the previous solution in which it had been
56
57
soaked, and these multiple stable ionization levels imparted the hysteretic physio-chemical
58
59
60
27
1
2
3
properties to the polymer. The effect of polymer-polymer hydrogen bonds within the confined
4
5
6 nanopores is the most likely source of the hysteresis, with the bonds forming under acidic
7
8 conditions that held the pores in a stable open state. This is because experiments with urea,
9
10
11
which disrupts hydrogen bonds, showed a shift toward the equilibrium, highly ionized form of
12
13 the confined PAA chains, supporting the mechanism of hydrogen bonding as the cause of the
14
15 hysteresis. This new understanding regarding the underlying phenomena that control hysteretic
16
17
18 behavior in these types of materials provides a unique outlook and a novel approach to
19
20 macromolecular design of potential materials for sensors and gating processes.
21
22
23 ASSOCIATED CONTENT
24
25 Supporting Information. Detailed descriptions of the experimental protocols implemented and
26
27 materials utilized; chemical structure of PI-PS-PAA, pore diagram of self-assembled surface, and
28
29
30 scanning electron micrograph of membrane surface; sample FTIR spectrum showing
31
32 characteristic peaks; plot of hydraulic permeability as a function of observed degree of
33
34 ionization; plot of degree of ionization as a function of solution pH from FTIR and copper uptake
35
36
37 experiments; plot of copper uptake as a function of observed degree of ionization; method of
38
39 calculating percent ionizations plotted in Figure S4 from copper uptake experiments; and
40
41
description and discussion of copper uptake experiments performed for intermediate PAA
42
43
44 ionization levels, as shown in Figure 4a. This material is available free of charge via the Internet
45
46 at http://pubs.acs.org.
47
48
49
50
51 AUTHOR INFORMATION
52
53 Corresponding Author
54
55
56 * William A. Phillip, Email: wphillip@nd.edu
57
58
59
60
28
1
2
3
Author Contributions
4
5
6
7 The manuscript was written through contributions of all authors. All authors have given approval
8
9 to the final version of the manuscript.
10
11
12 Notes
13
14 The authors declare no competing financial interest.
15
16 ACKNOWLEDGMENTS
17
18
19 Portions of this work were made possible with support from the Army Research Office (ARO)
20
21 through the Polymer Chemistry Program (Award Number: W911NF-14-1-0229, Program
22
23 Manager: Dr. Dawanne Poree) and the National Science Foundation (NSF) through the
24
25
26 Interfacial Processes and Thermodynamics Program (Award Number: 1511835, Program
27
28 Manager: Dr. Nora Savage), and we appreciatively acknowledge this support. B.W.B thankfully
29
30
31
acknowledges support from the Ralph W. and Grace M. Showalter Research Trust Award at
32
33 Purdue University. W.A.P. gratefully acknowledges support from the 3M non-Tenured Faculty
34
35 Award. We would like to thank the Notre Dame Integrated Imaging Facility (NDIIF) and the
36
37
38 Center for Environmental Science and Technology (CEST) at Notre Dame; portions of this
39
40 research were performed with instruments at these facilities.
41
42
43
44
45 REFERENCES
46
47 (1) Stuart, M. A. C.; Huck, W. T. S.; Genzer, J.; Muller, M.; Ober, C.; Stamm, M.;
48 Sukhorukov, G. B.; Szleifer, I.; Tsukruk, V. V.; Urban, M.; Winnik, F.; Zauscher, S.;
49 Luzinov, I.; Minko, S. Nat. Mater. 2010, 9, 101–113.
50
51 (2) Silva, J. M.; Caridade, G.; Costa, R. R.; Alves, M.; Groth, T.; Picart, C.; Reis, R. L.;
52 Mano, F. Langmuir 2015, 31, 11318–11328.
53
54 (3) Nunes, S. P.; Behzad, A. R.; Hooghan, B.; Sougrat, R.; Karunakaran, M.; Pradeep, N.;
55 Vainio, U.; Peinemann, K.-V. ACS Nano 2011, 5, 3516–3522.
56
57 (4) Sugnaux, C.; Lavanant, L.; Klok, H. Langmuir 2013, 29, 7325–7333.
58
59
60
29
1
2
3
(5) Hegewald, J.; Schmidt, T.; Eichhorn, K.; Kretschmer, K.; Kuckling, D.; Arndt, K.;
4
5 Dresden, D.-. Langmuir 2006, 22, 5152–5159.
6
7
(6) Beltran, S.; Baker, J. P.; Hooper, H. H.; Blanch, H. W.; Prausnitz, J. M. Macromolecules
8 1991, 24, 549–551.
9
10
(7) Tan, W. S.; Cohen, R. E.; Rubner, M. F.; Sukhishvili, S. A. Macromolecules 2010, 43,
11 1950–1957.
12
13 (8) Liu, H.; Li, Y.; Sun, K.; Fan, J.; Zhang, P.; Meng, J.; Wang, S.; Jiang, L. J. Am. Chem.
14 Soc. 2013, 135, 7603–7609.
15
16 (9) Chen, Y.; Zhao, T.; Wang, B.; Qiu, D. Langmuir 2015, 31, 8138–8145.
17
18
(10) Annaka, M.; Tanaka, T. Nature 1992, 355, 430–432.
19 (11) Itano, K.; Choi, J.; Rubner, M. F. Macromolecules 2005, 38, 3450–3460.
20
21 (12) Hiller, J.; Rubner, M. F. Macromolecules 2003, 36, 4078–4083.
22
23 (13) Lee, D.; Nolte, A. J.; Kunz, A. L.; Rubner, M. F.; Cohen, R. E. J. Am. Chem. Soc. 2006,
24 128, 8521–8529.
25
26 (14) Secrist, K. E.; Nolte, A. J. Macromolecules 2011, 44, 2859–2865.
27
28 (15) Xiang, T.; Tang, M.; Liu, Y.; Li, H.; Li, L.; Cao, W.; Sun, S.; Zhao, C. Desalination 2012,
29 295, 26–34.
30
31 (16) Sun, G.; Senapati, S.; Chang, H. Lab Chip 2016, 16, 1171–1177.
32
33 (17) Zhang, H.; Ito, Y. Langmuir 2001, 17, 8336–8340.
34
35 (18) Choi, J.; Rubner, M. F. Macromolecules 2005, 38, 116–124.
36
37 (19) Phillip, W. A.; Dorin, R. M.; Werner, J.; Hoek, E. M. V; Wiesner, U.; Elimelech, M. Nano
38 Lett. 2011, 11, 2892–2900.
39
40 (20) Geismann, C.; Tomicki, F.; Ulbricht, M. Sep. Sci. Technol. 2009, 44, 3312–3329.
41
42 (21) Cho, Y.; Lim, J.; Char, K. Soft Matter 2012, 8, 10271–10278.
43
44
(22) Shi, Q.; Su, Y.; Ning, X.; Chen, W.; Peng, J.; Jiang, Z. J. Membr. Sci. 2010, 347, 62–68.
45 (23) Himstedt, H. H.; Du, H.; Marshall, K. M.; Wickramasinghe, S. R.; Qian, X. Ind. Eng.
46
47
Chem. Res. 2013, 52, 9259–9269.
48 (24) Dugdale, D. Essentials of Electromagnetism; American Institute of Physics: New York,
49
50 1993.
51
(25) Israelachvili, J. N. Intermolecular and Surface Forces, 3rd ed.; Academic Press:
52
53 Burlington, 2011.
54
55
(26) Ch’ng, L. C.; Samanta, A. K.; Czako, G.; Bowman, J. M.; Reisler, H. J. Am. Chem. Soc.
56 2012, 134, 15430–15435.
57
58
(27) Jung, A.; Filiz, V.; Rangou, S.; Buhr, K.; Merten, P.; Hahn, J.; Clodt, J.; Abetz, C.; Abetz,
59
60
30
1
2
3
V. Macromol. Rapid Commun. 2013, 34, 610–615.
4
5 (28) Pendergast, M. M.; Mika Dorin, R.; Phillip, W. A.; Wiesner, U.; Hoek, E. M. V. J.
6
7
Membr. Sci. 2013, 444, 461–468.
8 (29) Peinemann, K.-V.; Abetz, V.; Simon, P. F. W. Nat. Mater. 2007, 6, 992–996.
9
10 (30) Zhang, Y.; Sargent, J. L.; Boudouris, B. W.; Phillip, W. A. J. Appl. Polym. Sci. 2015, 132,
11 41683–41699.
12
13 (31) Epps, T. H.; Cochran, E. W.; Hardy, C. M.; Bailey, T. S.; Waletzko, R. S.; Bates, F. S.
14
Macromolecules 2004, 37, 7085–7088.
15
16 (32) Bates, F. S.; Fredrickson, G. H.; Bates, F. S.; Fredrickson, G. H. Phys. Today 1999, 52,
17
18
32.
19 (33) Phillip, W. A.; Rzayev, J.; Hillmyer, M. A.; Cussler, E. L. J. Membr. Sci. 2006, 286, 144–
20
21 152.
22
(34) Phillip, W. A.; O’Neill, B.; Rodwogin, M.; Hillmyer, M. A.; Cussler, E. L. ACS Appl.
23
24 Mater. Interfaces 2010, 2, 847–853.
25
26
(35) Karunakaran, M.; Nunes, S. P.; Qiu, X.; Yu, H.; Peinemann, K.-V. J. Membr. Sci. 2014,
27 453, 471–477.
28
29
(36) Mulvenna, R. A.; Weidman, J. L.; Jing, B.; Pople, J. A.; Zhu, Y.; Boudouris, B. W.;
30 Phillip, W. A. J. Membr. Sci. 2014, 470, 246–256.
31
32 (37) Weidman, J. L.; Mulvenna, R. A.; Boudouris, B. W.; Phillip, W. A. Langmuir 2015, 31,
33 11113–11123.
34
35 (38) Mihai, I.; Addiego, F.; Ruch, D.; Ball, V. Sensors Actuators B. Chem. 2014, 192, 769–
36 775.
37
38 (39) Jayant, K.; Auluck, K.; Funke, M.; Anwar, S.; Phelps, J. B.; Gordon, P. H.; Rajwade, S.
39 R.; Kan, E. C. Phys. Rev. E 2013, 88, 1–13.
40
41 (40) Mulvenna, R. A.; Prato, R. A.; Phillip, W. A.; Boudouris, B. W. Macromol. Chem. Phys.
42 2015, 216, 1831–1840.
43
44 (41) Germack, D. S.; Wooley, K. L. J. Polym. Sci., Part A: Polym. Chem. 2007, 45, 4100–
45 4108.
46
47 (42) Bates, F. S. Annu. Rev. Phys. Chem. 1990, 41, 525–557.
48
49 (43) Taylor, W.; Jones, R. A. L. Langmuir 2010, 26, 13954–13958.
50
51 (44) Bird, B. B.; Stewart, W. E.; Lightfoot, E. N. Transport Phenomena, 2nd ed.; John Wiley
52 and Sons: New York, 2002.
53
54 (45) Hautojarvi, J.; Kontturi, K.; Nasman, J. H.; Svarfvar, B. L.; Viinikka, P.; Vuoristo, M. Ind.
55 Eng. Chem. Res. 1996, 35, 450–457.
56
57 (46) Ito, Y.; Kotera, S.; Inaba, M.; Kono, K.; Imanishi, Y. Polymer 1990, 31, 2157–2161.
58
59
60
31
1
2
3
(47) Ito, Y.; Inaba, M.; Chung, D.; Imanishi, Y. Macromolecules 1992, 25, 7313–7316.
4
5 (48) Riddick, J. A.; Bunger, W. B.; Sakano, T. K. Techniques of Chemistry, 4th ed.; John
6
7
Wiley and Sons: New York, 1985.
8 (49) Tagliazucchi, M.; Azzaroni, O.; Szleifer, I. J. Am. Chem. Soc. 2010, 132, 12404–12411.
9
10 (50) Emeis, C. A. J. Catal. 1993, 141, 347–354.
11
12 (51) Rzayev, J.; Hillmyer, M. A. J. Am. Chem. Soc. 2005, 127, 13373–13379.
13
14 (52) Tagliazucchi, M.; Azzaroni, O.; Szleifer, I. J. Am. Chem. Soc. 2010, 132, 12404–12411.
15
16 (53) Wijeratne, S.; Bruening, M. L.; Baker, G. L. Langmuir 2013, 29, 12720–12729.
17
18 (54) Annenkov, V. V.; Danilovtseva, E. N.; Saraev, V. V.; Mikhaleva, A. I. J. Polym. Sci., Part
19 A: Polym. Chem. 2003, 41, 2256–2263.
20
21 (55) Kuntz, I. D.; Brassfield, T. S. Arch. Biochem. Biophys. 1971, 142, 660–664.
22
23 (56) Perry, S. L.; Leon, L.; Hoffman, K. Q.; Kade, M. J.; Priftis, D.; Black, K. A.; Wong, D.;
24 Klein, R. A.; Piercelll, C. F.; Margossian, K. O.; Whitmer, J. K.; Qin, J.; de Pablo, J. J.; Tirrell,
25 M. Nat. Commun. 2015, 6, 6052-6060.
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
32
1
2
3
For Table of Contents Use Only
4
5
6 Unusually Stable Hysteresis in the pH-Response of Poly(Acrylic Acid) Brushes Confined
7 within Nanoporous Block Polymer Thin Films
8
9 Jacob L. Weidman, Ryan A. Mulvenna, Bryan W. Boudouris, and William A. Phillip
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
33
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
Page
Journal
35 ofof
39the American Chemical Society
1
2
3
4
5
6
7
8
9
10
11 ACS Paragon Plus Environment
12
13
Journal of the American Chemical
PageSociety
36 of 39
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
29
30
Page
Journal
37 ofof
39the American Chemical Society
1
2
3
4
5
6
7
9
10
Journal of the American Chemical
PageSociety
38 of 39
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
28
29
Page
Journal
39 ofof39the American Chemical Society
1
3
4

Weidman 2016

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Weidman 2016

Uploaded by

Copyright:

Available Formats

Subscriber access provided by UQ Library

Journal of the American Chemical Society is published by the American Chemical

You might also like