Organic Polymers as Porogenic Structure Matrices for Mesoporous Alumina and Magnesia

Mesoporous alumina and magnesia were prepared using various polymers, poly(ethylene glycol) (PEG), poly(vinyl alcohol) (PVA), poly(N-(2-hydroxypropyl) methacrylamide) (PHPMA), and poly(dimethylacrylamide) (PDMAAm), as porogenic structure matrices. Mesoporous alumina exhibits large Brunauer–Emmett–Teller (BET) surface areas up to 365 m2 g−1, while mesoporous magnesium oxide possesses BET surface areas around 111 m2 g−1. Variation of the polymers has little impact on the structural properties of the products. The calcination of the polymer/metal oxide composite materials benefits from the fact that the polymer decomposition is catalyzed by the freshly formed metal oxide.


Introduction
Mesoporous metal oxides with large specific surface areas and uniform pore sizes have recently attained great interest, particularly regarding potential applications in such areas as catalysis [1], energy conversion and storage [2], and gas sensing [3,4].By definition, mesopore widths range from 2 to 50 nm [5].For metal oxides with uniform and ordered mesopores, a variety of synthesis methods have been established, mostly by utilization of porogens; said porogens may be supramolecular entities of amphiphilic species dispersed in liquid media ('soft templates' [6]) or solid structure matrices such as porous silica ('hard templates') in the so-called 'nanocasting' process [6][7][8][9].
Polymer chains begin to overlap and form entanglements when the polymer solution is above a critical concentration.Hence, a physical network is formed between different polymer chains in a concentrated solution [36,37].Therefore, a concentrated polymer solution could also theoretically work as a structure matrix to prepare mesoporous metal oxides.In this paper, we describe the synthesis of mesoporous alumina and magnesium oxide using simple polymers, such as poly(ethylene glycol) (PEG), poly(vinyl alcohol) (PVA), poly(N-(2-hydroxypropyl) methacrylamide) (PHPMA), and poly(dimethylacrylamide) (PDMAAm), as matrices.The strategy proposed here to prepare porous alumina and magnesium oxide is based on a one-pot synthesis approach using saturated aluminum/magnesium nitrate as precursor solutions and introducing direct polymers.
Gel permeation chromatography (GPC) was performed in chloroform for PEG and PDMAAm at 30 • C and at a flow rate of 0.75 mL min −1 on a Jasco (Groß-Umstadt, Germany) 880-PU Liquid Chromatograph connected to a Shodex (Yokohama, Japan) RI-101 detector.The instrument was equipped with four consecutive columns (PSS-SDV columns filled with 5 µm gel particles with a defined porosity of 10 6 Å, 10 5 Å, 10 3 Å, and 10 2 Å, respectively), and both samples were calibrated by poly(methyl methacrylate) standards.GPC was performed in hexafluoroisopropanol for PVA at 0 • C and at a flow rate of 1 mL min −1 on a Merck (Darmstadt, Germany) LC-6200 liquid chromatograph connected to a Shodex (Yokohama, Japan) RI-101 detector.The instrument was equipped with a PSS-PFG 10 3 Å and PSS-PFG 10 2 Å column, and the sample was calibrated by poly(methyl methacrylate) standards.GPC was performed in N,N-dimethylacetamide for PHPMA at 50 • C and at a flow rate of 0.5 mL min −1 on a Merck (Darmstadt, Germany) LC 655A-11 liquid chromatograph connected to a Waters (Milford, Massachusetts, United States) RI 2410 detector.The instrument was equipped with PSS-GRAM 10 4 Å, PSS-GRAM 10 3 Å, and PSS-GRAM 10 2 Å columns, and the sample was calibrated by poly(methyl methacrylate) standards.Thermogravimetric analysis (TGA) was conducted under synthetic air at a heating rate of 10 • C min −1 using a Mettler Toledo (Columbus, Ohio, USA) TGA/SDTA851.N 2 physisorption analysis was performed at 77 K on a Quantachrome (Boynton Beach, Florida, United States) Autosorb 6B instrument; samples were degassed at 120 • C for 12 h prior to measurement.Specific surface areas were assessed via multi-point Brunauer-Emmett-Teller (BET) analysis [38] in the range of 0.1 ≤ p/p 0 ≤ 0.3.Pore volumes were calculated at p/p 0 = 0.99.Pore size distributions were calculated via Barrett-Joyner-Halenda (BJH) analysis [39] from the desorption branches of the isotherms.Powder X-ray diffraction was performed with a Bruker (Billerica, Massachusetts, USA) AXS D8 Advance diffractometer with Cu Kα radiation (40 kV, 40 mA) with a step size of 0.02 • and a counting time of 3 s per step.
Homopolymer PHPMA was synthesized by free radical polymerization initiated with AIBN as described in the literature [42].Monomer HPMA (1.5 g, 10.5 mmol) and AIBN (1.7 mg, 0.010 mmol) were dissolved in 1,4-dioxane (20 mL) in a 50 mL nitrogen flask and was degassed three times by freeze/thaw cycles.The HPMA was polymerized at 65 • C for 8 h under an argon atmosphere.The mixture was poured into acetone to get a white solid, which was collected and washed with acetone repeatedly.Further purification was carried out by dissolving the polymer in methanol and precipitating into acetone.The product was collected and dried under vacuum to obtain the homopolymer as a white powder and characterized by NMR spectroscopy and GPC (see Table 1).

Preparation of Mesoporous Metal Oxides
One hundred ninety six milligrams of polymer were dissolved in 800 µL of a saturated aqueous solution of aluminum nitrate (1.9 mol L −1 ) or magnesium nitrate (4.9 mol L −1 ).The Al(NO 3 ) 3 -containing solution was treated at 60 • C with a vapor of an aqueous ammonia solution (12.5%) for 3 h to convert Al(NO 3 ) 3 to Al(OH) 3 /AlO(OH); the resulting material was dried overnight at 60 • C and then calcined in a tube furnace for 4 h at 500 • C (heating rate 1 • C min −1 ) to form Al 2 O 3 and to combust the polymer.The Mg(NO 3 ) 2 -containing solution was dried overnight at 120 • C; the resulting material was calcined in a tube furnace for 2 h at 300 • C and for 2 h at 500 • C (heating rate 1 • C min −1 ) to convert Mg(NO 3 ) 2 to MgO and to combust the polymer.

Results and Discussion
A variety of four simple water soluble polymers were used as porogenic structure directors for mesoporous Al 2 O 3 and MgO.The polymers possess different hydrophilicity and distinct ability to coordinate to Al 3+ and Mg 2+ metal cations: (i) poly(ethylene glycol) (PEG; ether groups), (ii) poly(vinyl alcohol) (PVA; hydroxyl groups), (iii) poly(dimethylacrylamide) (PDMAAm; tertiary amido groups), and (iv) poly(N-(2-hydroxypropyl) methacrylamide) (PHPMA; secondary amido with hydroxyl groups).The latter two polymers were synthesized by free-radical polymerization, as shown in Figure 1b,c.Their properties are summarized in Table 2; molecular weights, dispersities, and yields are typical of free-radical polymerization synthesis.

Preparation of Mesoporous Metal Oxides
One hundred ninety six milligrams of polymer were dissolved in 800 μL of a saturated aqueous solution of aluminum nitrate (1.9 mol•L −1 ) or magnesium nitrate (4.9 mol•L −1 ).The Al(NO3)3containing solution was treated at 60 °C with a vapor of an aqueous ammonia solution (12.5%) for 3 h to convert Al(NO3)3 to Al(OH)3/AlO(OH); the resulting material was dried overnight at 60 °C and then calcined in a tube furnace for 4 h at 500 °C (heating rate 1 °C•min −1 ) to form Al2O3 and to combust the polymer.The Mg(NO3)2-containing solution was dried overnight at 120 °C; the resulting material was calcined in a tube furnace for 2 h at 300 °C and for 2 h at 500 °C (heating rate 1 °C•min −1 ) to convert Mg(NO3)2 to MgO and to combust the polymer.

Results and Discussion
A variety of four simple water soluble polymers were used as porogenic structure directors for mesoporous Al2O3 and MgO.The polymers possess different hydrophilicity and distinct ability to coordinate to Al 3+ and Mg 2+ metal cations: (i) poly(ethylene glycol) (PEG; ether groups), (ii) poly(vinyl alcohol) (PVA; hydroxyl groups), (iii) poly(dimethylacrylamide) (PDMAAm; tertiary amido groups), and (iv) poly(N-(2-hydroxypropyl) methacrylamide) (PHPMA; secondary amido with hydroxyl groups).The latter two polymers were synthesized by free-radical polymerization, as shown in Figure 1b,c.Their properties are summarized in Table 2; molecular weights, dispersities, and yields are typical of free-radical polymerization synthesis.The aim of this study was to investigate the impact of the polymers on the porosity of the metal oxides.For this purpose, the respective polymer was dissolved in a saturated aqueous solution of aluminum nitrate, followed by treatment in ammonia vapor at 60 °C to convert Al(NO3)3 to Al(OH)3/AlO(OH), as described in the experimental section.After evaporation of the water, the material was then calcined at 500 °C to turn Al(OH)3/AlO(OH) into Al2O3 and simultaneously combust the polymer.For MgO, the same procedure was applied, but without the ammonia treatment step; magnesium nitrate was directly converted to magnesium oxide by calcination.By this procedure, a composite of the metal oxide precursor (Al(OH)3/AlO(OH) or Mg(NO3)2, respectively) and the polymer was formed first, with the polymer being entangled within the inorganic phase.The aim of this study was to investigate the impact of the polymers on the porosity of the metal oxides.For this purpose, the respective polymer was dissolved in a saturated aqueous solution of aluminum nitrate, followed by treatment in ammonia vapor at 60 • C to convert Al(NO 3 ) 3 to Al(OH) 3 /AlO(OH), as described in the experimental section.After evaporation of the water, the material was then calcined at 500 • C to turn Al(OH) 3 /AlO(OH) into Al 2 O 3 and simultaneously combust the polymer.For MgO, the same procedure was applied, but without the ammonia treatment step; magnesium nitrate was directly converted to magnesium oxide by calcination.By this procedure, a composite of the metal oxide precursor (Al(OH) 3 /AlO(OH) or Mg(NO 3 ) 2 , respectively) and the polymer was formed first, with the polymer being entangled within the inorganic phase.Then, simultaneous conversion of the precursor into the metal oxide and thermal combustion of the polymer led to a mesoporous product.
To study the calcination/polymer combustion step in some detail, thermogravimetric analysis (TGA) was carried out.As an example, the TGA curves of the Al(OH) 3 /PDMAAm composite and of the pure PDMAAm polymer were compared and are shown in Figure 2. A mass loss of ca.72% can be observed for the composite material in the temperature range up to 230 • C, which can be attributed to both the dehydration of Al(OH) 3 /AlO(OH) (i.e., Al 2 O 3 formation) and the combustion of the polymer.mass loss of ca.14% can be observed between 230 and 570 • C. By comparison, the pure polymer shows an initial mass loss of 6% below 200 • C, probably due to loss of residual water, then a mass loss of about 74% between 300 to 400 • C, followed by another 18% up to ca. 600 • C. Obviously, the presence of the aluminum hydroxide/oxide led to a combustion of the polymer at lower temperature; this effect has already been observed for the combustion of amorphous carbon [18,26] and organic hydrogel matrices [35].Very similar results were obtained for Al 2 O 3 prepared using the other polymers (see Figures S1-S3).
Then, simultaneous conversion of the precursor into the metal oxide and thermal combustion of the polymer led to a mesoporous product.
To study the calcination/polymer combustion step in some detail, thermogravimetric analysis (TGA) was carried out.As an example, the TGA curves of the Al(OH)3/PDMAAm composite and of the pure PDMAAm polymer were compared and are shown in Figure 2. A mass loss of ca.72% can be observed for the composite material in the temperature range up to 230 °C, which can be attributed to both the dehydration of Al(OH)3/AlO(OH) (i.e., Al2O3 formation) and the combustion of the polymer.Further mass loss of ca.14% can be observed between 230 and 570 °C.By comparison, the pure polymer shows an initial mass loss of 6% below 200 °C, probably due to loss of residual water, then a mass loss of about 74% between 300 to 400 °C, followed by another 18% up to ca. 600 °C.Obviously, the presence of the aluminum hydroxide/oxide led to a combustion of the polymer at lower temperature; this effect has already been observed for the combustion of amorphous carbon [18,26] and organic hydrogel matrices [35].Very similar results were obtained for Al2O3 prepared using the other polymers (see Figures S1-S3).The porogenic impact of the polymers on the polymer-free metal oxides was confirmed by N2 physisorption analysis.Figure 3 (left) shows the sorption isotherms of four Al2O3 materials prepared with different polymers.All isotherms exhibit a faint type-IV(a) behavior [43] with a more or less well-pronounced hysteresis.This indicates mesopores with an ill-defined shape, but with a fairly uniform size, as confirmed in the BJH pore size distribution curves [41] derived from the isotherms (Figure 3, right).Pore widths from 3 to 8 nm can be observed, with a clear peak occurring at 3.6 nm in all materials.The pore size distribution is somewhat narrower in the two samples prepared with PVA and PEG, respectively.The specific pore volumes and BET surface areas are shown in Table 2, confirming that a reproducible synthesis of porous alumina with a large surface area up to 365 m 2 g −1 is possible by the utilization of these polymers as porogens.Comparison of all prepared Al2O3 materials reveals similar mesopore sizes, mesopore volumes, and specific BET areas.The choice of the porogenic polymer matrix has little impact on the porosity.Although polymers with different binding sites were used, the appearance of a polymer rich phase due to physical network formation in concentrated solution can be considered the sole reason for pore formation.The porogenic impact of the polymers on the polymer-free metal oxides was confirmed by N 2 physisorption analysis.Figure 3 (left) shows the sorption isotherms of four Al 2 O 3 materials prepared with different polymers.All isotherms exhibit a faint type-IV(a) behavior [43] with a more or less well-pronounced hysteresis.This indicates mesopores with an ill-defined shape, but with a fairly uniform size, as confirmed in the BJH pore size distribution curves [41] derived from the isotherms (Figure 3, right).Pore widths from 3 to 8 nm can be observed, with a clear peak occurring at 3.6 nm in all materials.The pore size distribution is somewhat narrower in the two samples prepared with PVA and PEG, respectively.The specific pore volumes and BET surface areas are shown in Table 2, confirming that a reproducible synthesis of porous alumina with a large surface area up to 365 m 2 g −1 is possible by the utilization of these polymers as porogens.Comparison of all prepared Al 2 O 3 materials reveals similar mesopore sizes, mesopore volumes, and specific BET areas.The choice of the porogenic polymer matrix has little impact on the porosity.Although polymers with different binding sites were used, the appearance of a polymer rich phase due to physical network formation in concentrated solution can be considered the sole reason for pore formation.Table 2. Specific Brunauer-Emmett-Teller (BET) surface areas ABET, pore volumes V, and mean pore widths r obtained from N2 physisorption of mesoporous alumina synthesized using various polymers.
Polymer Used ABET/m 2 g −1 V/cm Figure 4 shows the powder X-ray diffraction patterns of the alumina materials.Again, the differences between the materials are rather low.All samples exhibit only a few broad reflections, two of which can be attributed to the cubic defect spinel structure of γ-Al2O3.(JCPDS card number 75-0921).The formation of this phase with low crystallinity is commonly observed for Al2O3 syntheses under these conditions [17,18].The crystallite sizes calculated by the Scherrer method are between 5 and 6 nm.Table 2. Specific Brunauer-Emmett-Teller (BET) surface areas A BET , pore volumes V, and mean pore widths r obtained from N 2 physisorption of mesoporous alumina synthesized using various polymers.

Polymer Used
A BET /m 2 g −1 V/cm Figure 4 shows the powder X-ray diffraction patterns of the alumina materials.Again, the differences between the materials are rather low.All samples exhibit only a few broad reflections, two of which can be attributed to the cubic defect spinel structure of γ-Al 2 O 3 .(JCPDS card number 75-0921).The formation of this phase with low crystallinity is commonly observed for Al 2 O 3 syntheses under these conditions [17,18].The crystallite sizes calculated by the Scherrer method are between 5 and 6 nm.Table 2. Specific Brunauer-Emmett-Teller (BET) surface areas ABET, pore volumes V, and mean pore widths r obtained from N2 physisorption of mesoporous alumina synthesized using various polymers.
Polymer Used ABET/m 2 g −1 V/cm Figure 4 shows the powder X-ray diffraction patterns of the alumina materials.Again, the differences between the materials are rather low.All samples exhibit only a few broad reflections, two of which can be attributed to the cubic defect spinel structure of γ-Al2O3.(JCPDS card number 75-0921).The formation of this phase with low crystallinity is commonly observed for Al2O3 syntheses under these conditions [17,18].The crystallite sizes calculated by the Scherrer method are between 5 and 6 nm.Since the choice of polymer turned out not to have any significant impact on the Al 2 O 3 synthesis, only one polymer, PDMAAm, was chosen for the preparation of porous MgO.The TGA curves of the Mg(NO 3 ) 2 /PDMAAm composite and of the pure PDMAAm polymer are shown in Figure 5.For the composite, the mass loss occurs in two distinct steps: by ca.54% up to a temperature of 265 • C and by another 32% between 265 and 500 • C. It is fair to assume that the first step is mainly attributable to the conversion of magnesium nitrate into magnesium oxide, while the second step corresponds mostly to the polymer decomposition.This seems likely because the pure polymer starts to combust only above ca.300 • C (after some initial mass loss of 6% below 200 • C, presumably due to loss of residual water); a steep reduction in mass by ca.74% occurs between 300 and 400 • C, followed by another 18% between 400 and 600 • C. Again, the presence of the magnesium species results in a polymer decomposition at a slightly lower temperature, although this effect is less pronounced than in the case of the aluminum species.
Since the choice of polymer turned out not to have any significant impact on the Al2O3 synthesis, only one polymer, PDMAAm, was chosen for the preparation of porous MgO.The TGA curves of the Mg(NO3)2/PDMAAm composite and of the pure PDMAAm polymer are shown in Figure 5.For the composite, the mass loss occurs in two distinct steps: by ca.54% up to a temperature of 265 °C and by another 32% between 265 and 500 °C.It is fair to assume that the first step is mainly attributable to the conversion of magnesium nitrate into magnesium oxide, while the second step corresponds mostly to the polymer decomposition.This seems likely because the pure polymer starts to combust only above ca.300 °C (after some initial mass loss of 6% below 200 °C, presumably due to loss of residual water); a steep reduction in mass by ca.74% occurs between 300 and 400 °C, followed by another 18% between 400 and 600 °C.Again, the presence of the magnesium species results in a polymer decomposition at a slightly lower temperature, although this effect is less pronounced than in the case of the aluminum species.Figure 6 shows the N2 physisorption data of the porous MgO sample.The isotherm shape is mostly type II, with a slight type-IV character and little hysteresis, indicating a fairly low degree of porosity.Accordingly, the pore size distribution peak is very low in intensity.The specific BET surface area and pore volume are 111 m 2 g −1 and 0.37 cm 3 g −1 , respectively.Obviously, the polymer failed to have a pronounced porogenic impact in the case of MgO, which may be explained by the sintering of MgO particles during calcination upon combustion of the polymer.During the Al2O3 synthesis, by contrast, a solid network of Al(OH)3/AlO(OH) was formed before the combustion of the polymer.Figure 6 shows the N 2 physisorption data of the porous MgO sample.The isotherm shape is mostly type II, with a slight type-IV character and little hysteresis, indicating a fairly low degree of porosity.Accordingly, the pore size distribution peak is very low in intensity.The specific BET surface area and pore volume are 111 m 2 g −1 and 0.37 cm 3 g −1 , respectively.Obviously, the polymer failed to have a pronounced porogenic impact in the case of MgO, which may be explained by the sintering of MgO particles during calcination upon combustion of the polymer.During the Al 2 O 3 synthesis, by contrast, a solid network of Al(OH) 3 /AlO(OH) was formed before the combustion of the polymer.
Since the choice of polymer turned out not to have any significant impact on the Al2O3 synthesis, only one polymer, PDMAAm, was chosen for the preparation of porous MgO.The TGA curves of the Mg(NO3)2/PDMAAm composite and of the pure PDMAAm polymer are shown in Figure 5.For the composite, the mass loss occurs in two distinct steps: by ca.54% up to a temperature of 265 °C and by another 32% between 265 and 500 °C.It is fair to assume that the first step is mainly attributable to the conversion of magnesium nitrate into magnesium oxide, while the second step corresponds mostly to the polymer decomposition.This seems likely because the pure polymer starts to combust only above ca.300 °C (after some initial mass loss of 6% below 200 °C, presumably due to loss of residual water); a steep reduction in mass by ca.74% occurs between 300 and 400 °C, followed by another 18% between 400 and 600 °C.Again, the presence of the magnesium species results in a polymer decomposition at a slightly lower temperature, although this effect is less pronounced than in the case of the aluminum species.Figure 6 shows the N2 physisorption data of the porous MgO sample.The isotherm shape is mostly type II, with a slight type-IV character and little hysteresis, indicating a fairly low degree of porosity.Accordingly, the pore size distribution peak is very low in intensity.The specific BET surface area and pore volume are 111 m 2 g −1 and 0.37 cm 3 g −1 , respectively.Obviously, the polymer failed to have a pronounced porogenic impact in the case of MgO, which may be explained by the sintering of MgO particles during calcination upon combustion of the polymer.During the Al2O3 synthesis, by contrast, a solid network of Al(OH)3/AlO(OH) was formed before the combustion of the polymer.The powder X-ray diffraction diagram of MgO is shown in Figure 7, confirming the cubic rock salt structure of MgO (JCPDS card number 77-2179) with a substantially higher degree of crystallinity than in case of Al 2 O 3 .This is consistent with the above-made assumption of strong sintering upon polymer combustion.The crystallite size calculated by the Scherrer method is ca.12 nm.

Conclusions
Mesoporous γ-Al 2 O 3 and mesoporous MgO with large specific BET surface areas were successfully synthesized using simple polymers (PEG, PVA, PDMAAm, and PHPMA) as porogenic matrices under relatively mild conditions.The polymers were mixed with a metal nitrate solution.The polymer matrices were removed by thermal combustion, while the metal oxides were formed at the same time.The mesoporous alumina products exhibit mesopore sizes in the range from 3.6 to 6.4 nm, large specific BET surface areas up to 365 m 2 g −1 , and specific pore volumes up to 0.54 cm 3 g −1 .Variation of the polymer has little impact on the structural properties of the products.The mesoporous magnesium oxide product has a mesopore size of 3.6 nm, a specific BET surface area of 111 m 2 g −1 , and a specific pore volume of 0.37 cm 3 g −1 .The calcination of the polymer/metal oxides composite materials benefits from the fact that polymer decomposition is catalyzed by the freshly formed metal oxides.

Figure 3 .
Figure 3. N2 physisorption isotherm (left) and pore size distribution (right) of mesoporous γ-Al2O3 prepared using various polymers as the porogenic structure matrices as indicated.(Data are vertically shifted for clarity).

Figure 4 .
Figure 4. Powder XRD patterns of mesoporous γ-Al2O3 prepared using various polymers as the porogenic structure matrices as indicated.(Data are vertically shifted for clarity).

Figure 3 .
Figure 3. N 2 physisorption isotherm (left) and pore size distribution (right) of mesoporous γ-Al 2 O 3 prepared using various polymers as the porogenic structure matrices as indicated.(Data are vertically shifted for clarity).

Figure 3 .
Figure 3. N2 physisorption isotherm (left) and pore size distribution (right) of mesoporous γ-Al2O3 prepared using various polymers as the porogenic structure matrices as indicated.(Data are vertically shifted for clarity).

Figure 4 .
Figure 4. Powder XRD patterns of mesoporous γ-Al2O3 prepared using various polymers as the porogenic structure matrices as indicated.(Data are vertically shifted for clarity).

4 .
Powder XRD patterns of mesoporous γ-Al 2 O 3 prepared using various polymers as the porogenic structure matrices as indicated.(Data are vertically shifted for clarity).

Figure 6 .
Figure 6.N2 physisorption isotherm (left) and pore size distribution (right) of MgO prepared using PDMAAm polymer as the porogenic structure matrix.

Figure 6 .
Figure 6.N2 physisorption isotherm (left) and pore size distribution (right) of MgO prepared using PDMAAm polymer as the porogenic structure matrix.

Figure 6 .
Figure 6.N 2 physisorption isotherm (left) and pore size distribution (right) of MgO prepared using PDMAAm polymer as the porogenic structure matrix.

Figure 7 .
Figure 7. Powder XRD pattern of MgO prepared by using PDMAAm polymer as the porogenic structure matrix.

Figure S1 :
TGA of the Al(OH) 3 -PVA composite and of the pure polymer PVA; Figure S2: TGA of the Al(OH) 3 -PHPMA composite and of the pure polymer PHPMA; Figure S3: TGA of the Al(OH) 3 -PEG composite and of the pure PEG.

Table 1 .
Characterization of the used homopolymers.