Ultrasmall Angle N - American

Nov 8, 2011 - Department of Pharmacology, Georgetown University, Washington, D.C. 20057, United States. bS Supporting Information. Amonolith is a rigi...
0 downloads 0 Views 3MB Size
LETTER pubs.acs.org/ac

A More Informative Approach for Characterization of Polymer Monolithic Phases: Small Angle Neutron Scattering/Ultrasmall Angle Neutron Scattering Kathleen M. Ford,† Brian G. Konzman,† and Judith F. Rubinson*,†,‡ † ‡

Department of Chemistry, Georgetown University, Washington, D.C. 20057, United States Department of Pharmacology, Georgetown University, Washington, D.C. 20057, United States

bS Supporting Information ABSTRACT: Neutron scattering techniques have been used frequently to characterize geological specimens and to determine the structures of glasses and of polymers as solutions, suspensions, or melts. Little work has been reported on their application in determining polymers' structural properties relevant to separations. Here, we present a comparison of characterization results from nitrogen porosimetry and from combined small angle neutron scattering (SANS) and ultrasmall angle neutron scattering (USANS) experiments. We show that SANS is extremely sensitive to the pore characteristics. Both approaches can provide information about porosity and pore characteristics, but the neutron scattering techniques provide additional information in the form of the surface characteristics of the pores and their length scales. Fits of the scattering data show that cylindrical pores are present with diameters down to 0.6 μm and that, for length scales down to approxmately 20 Å, the material shows self-similar (fractal) slopes of 3.4 to 3.6. Comparison of these characteristics with other examples from the scattering literature indicate that further investigation of their meaning for chromatographic media is required.

A

monolith is a rigid porous structure consisting of microglobules that precipitate during polymerization from solutions of monomer, porogen, cross-linker, and a small amount of solvent.1 These microglobules eventually grow together yielding a continuous porous solid. The fused microglobules have been termed “porons”.2 Polymer monoliths offer a number of advantages over the popular silica bead option for separations. The presence of a rigid structure with a distribution of pore sizes allows for increased permeability as well as improved mass transfer both parallel and perpendicular to the column axis. This renders these phases useful in separations ranging from high pressure liquid chromatography to batch separations.13 It has generally been accepted that there are roughly two groups of pores present: micro- and macropores.2 It is important to realize that, because of the mechanism of formation of the monolith, the fixed relationship between particle size and pore size that exists for packed bed columns no longer needs to hold for the polymer monolith. In other words, for a polymer monolith, the pore size and pore distribution can be tuned by adjusting reaction conditions, e.g., reaction temperature, concentrations of initiator, monomer, and cross-linking monomer, and identity of the porogen.1,4 Guiochon has proposed that optimization of the separations utility of the monolith relies on the following: the solid of the monolith should consist of small porons; the total volume of the porons must be a large fraction of the total monolith’s volume (to yield a high total surface area); and the average size of the larger pores must be significantly larger than the micropores so that the permeability is high.2 Any decrease in overall surface area r 2011 American Chemical Society

compared to packed bed columns is typically addressed by incorporating a reactive group onto the monolith for subsequent attachment of groups to provide specificity.2 For both bare and modified monoliths, optimization of structure for separations depends on learning about the effects of a wide range of reaction conditions on pore size and the surface characteristics. Porosity of monolithic media (as for other porous media) is generally carried out using either nitrogen or mercury porosimetry.2,5 The utilization of small-angle neutron scattering (SANS), small-angle X-ray scattering (SAXS), and ultrasmallangle neutron scattering (USANS) techniques for characterization of porosity has not been widespread for organic polymeric materials6,7 However, these techniques have found use for materials such as porous glasses,810 aerogels,11 polysiloxanes,12 molecular sieves,12 silica microparticles,13 geological specimens,9,1417 metallurgical samples,18,19 graphite, activated carbon,20 and zeolites21 (Table 1). The large atom percentage of hydrogen in polymer monoliths leads to neutron scattering densities that provide large signals, thus enhancing signal-to-noise in the data. Second, the data from the experiment is plotted as I(q) vs q. The variable q, is defined as q = (2π/λ) sin θ, where 2θ is the scattering angle. This, in turn, translates into a length l = 2π/q for the features probed at a given angle. Thus, SANS/USANS results provide detailed information about the heterogeneity of the surface and the shapes of the Received: August 24, 2011 Accepted: November 8, 2011 Published: November 08, 2011 9201

dx.doi.org/10.1021/ac202238r | Anal. Chem. 2011, 83, 9201–9205

Analytical Chemistry

LETTER

Table 1. Materials Characterized by Neutron and X-ray Scatteringa type of sample organo modified mesoporous molecular sieves silica aerogels

-slope sample 1

4.62

0.0327

0.0838

2.65

0.0838

0.16

most dense

3.59

0.004

0.06

least dense

3.59

0.06

0.2

3.6

0.02

0.2

4.1

0.01

0.25

shale

coal

ref 12 11 9

3.04

0.002

0.1

9

sample 1 sample 2

3.45 3.26

0.002 0.002

0.088 0.065

9

sample 3

4

0.002

0.017

sample 4

4

0.002

0.023

sample 1

3

0.00003

0.001

sample 2

3

0.00001

0.01

1416

basal plan oriented graphite

3.5

0.013

0.16

29

activated carbon

3.08

0.005

0.04

20

3 3.5

0.0005 0.0006

0.04 0.12

17 17

carbonates carbonates porous glass carbon filled polymer

skarn marble unmetamorphosed sample 1

3.7

0.07

0.18

sample 2

3.98

0.07

0.18

44.5%

2.98

0.13

0.25

1.1%

3.68

0.13

0.25

annealed nanocrystalline Cr dry Vycor

silica microparticles

a

high q limit

sample 2

porous glass

limestone

low q limit

3.76

0.005

0.02

SANS

2.6

0.06

0.1

USANS slit-smeared SANS

2.6 2.6

0.00003 0.08

0.0001 0.4

33 nm pore

4

0.035

0.07

8.5 nm pore

4

0.07

0.185

6 nm pore

4

0.185

1.12

10 6 19 8

13

Values for q are expressed in Å1. The ranges given are those over which the indicated slope was found.

pores.21 The length scale probed (1010 000 Å) using a combination of SANS and USANS is ideal for the characterization of pore dimensions for monoliths. In addition, it is possible to match the scattering density of the monolith with a mixture of the deuterated and nondeuterated forms of a mobile phase (contrast matching) and to make it “disappear” anywhere the solvent penetrates. This means that it would be possible not only to detect any areas into which the solvent cannot penetrate (with data in the USANS and low-q SANS region)22 but also to detect any thin film or adsorbate on its surfaces vs unadsorbed fluid in the pores (with data in the high-q SANS region).13 The detection of areas that are not accessible to solvent is especially important when the solvent is very different in polarity from the monolith. Further, the possibility of spatial ordering of the pores can be probed by varying the position of a single monolithic sample oriented with respect to the beam. As such, scattering techniques provide confirmatory and complementary information to that of nitrogen and mercury porosimetry for the investigation of polymer monoliths. We describe here a comparison for a poly(4vinylaniline)/divinylbenzene monolith between the surface area and pore diameter measurements from nitrogen porosimetry and SANS/USANS characterization as well as how SANS/USANS can be used to provide information about some structural features of the porous media.

Scheme 1. Preparation of Monolith

’ EXPERIMENTAL SECTION Synthesis of the Monoliths. Two different types of monoliths were characterized. The polymers were intended for linkage of a ligand to the surface, and thus, 4-vinylaniline was used in combination with the divinylbenzene cross-linker, providing an amine group for further reaction. (All numbers in parentheses indicate purity of the reagent as used from the supplier.) The initiators for polymerization, either azobis(isobutyronitrile), AIBN, or 4,40 -azobis(4-cyanovaleric acid), ACVA (both from Fluka, >98%), were stored at 4° under nitrogen, as their activity 9202

dx.doi.org/10.1021/ac202238r |Anal. Chem. 2011, 83, 9201–9205

Analytical Chemistry

LETTER

measurements were carried out on the Perfect Crystal Diffractometer of beam port BT5 at the NCNR.26 Data reduction was accomplished using Igor Pro software (WaveMetrics, Lake Oswego, OR) with SANS macros developed at the NCNR.27 The data is desmeared, background subtracted, and converted to an absolute scale. Raw counts were normalized to a common monitor count and corrected for empty cell counts, ambient room background counts, and nonuniform detector response. Data were placed on an absolute scale by normalizing the scattered intensity to the incident beam flux. Finally, the data were radially averaged to produce the scattering intensity I(q) to plot as I(q) vs q curves.

Figure 1. FESEM image of poly(4-vinyaniline)/polydivinylbenzene monolith.

decreases significantly in air at room temperature. Styrene (Sigma, 99%), 4-vinylaniline (4-VA, Sigma, 90%), divinylbenzene (DVB, Sigma, 80%, mixture of isomers), and 1-dodecanol (Sigma, 98%) were refrigerated until used. Toluene (Sigma, 99%) and tetrahydrofuran (g99.9%) also were used as purchased from Sigma Aldrich. Preparation of the monolith was carried out using the scheme shown in Scheme 1 and was based on methods used by Svec and Frechet4 and by Tripp et al.23 More details of the syntheses are given in Supporting Information. Monolith Characterization. Scanning electron microscopy was performed using a Zeiss Ultra Plus field emission scanning electron microscope (FESEM). Nitrogen porosimetry data were acquired using a Quantachrome Autosorb-1 MP; monolith samples were outgassed at 343K for 45 h before measurements were performed. Specific surface area values were obtained using the BrunauerEmmetTeller (BET) equation, and average pore diameters were calculated by the ratio of pore volume to surface area and by the BarrettJoynerHalenda24 (BJH) method based on the porosimetry data. Neutron scattering was carried out for a 20 min period for a 0.5 mm thick, 1.25 cm diameter disk of unactivated 4-vinylaniline/ divinylbenzene copolymer. Data acquisition for the activated monolith was carried out for 5 min on a 1 mm path length cell with fused silica windows filled with a coarsely ground sample (particle diameter approximately 0.1 mm to 0.5 mm). Care was taken for the ground sample to make sure that the particle size was much larger than the length scales probed by the experiments. The use of two different methods for preparing the polymer for the neutron scattering experiments served two purposes. First, gently crushing the monolith ensured that the orientation of any pores with respect to the sampledetector angle was random. Second, it allowed for a more valid comparison of the pore diameter measurements between the neutron scattering and porosimetry measurements. SANS measurements were performed on the NG7 and NG3 30-m SANS instruments at the NIST Center for Neutron Research (NCNR) in Gaithersburg, MD.25 Neutron wavelengths, λ, between 5.2 Å and 5.5 Å were used with wavelength spreads Δλ/λ of between 0.11 and 0.15. Scattered neutrons were detected with a 64 cm  64 cm two-dimensional position sensitive detector with 128  128 pixels and 0.5 cm resolution per pixel. Ultrahigh resolution small-angle neutron scattering (USANS)

’ RESULTS AND DISCUSSION Features on a variety of scales can be observed in the SEM images of a 10/90 4-vinylaniline/divinylbenzene monolith. Figure 1 shows larger channels as well as smaller ones in the monolith. Although some features appear to have diameters as large as 20 μm or more; on average, the larger features range from 1 to 4 μm. The more common, smaller, ones appear to have diameters in the nanometer range. Samples of the AVCCl-activated form of this monolith showed similar features. The latter monolith was part of our efforts to produce an activated surface for attachment of ligating groups. Values for surface area, pore volume, average pore diameter (BET method), and pore diameter (BJH method) for three batches of the activated copolymer are given in Table 2. The model used for treatment of the data was a Type II isotherm containing macropores. The calculated figures of merit for these monoliths are similar to those found by other authors.28 Figure 2a depicts the USANS and SANS data across the entire range measured for a 4-vinylaniline/divinylbenzene copolymer prepared with an excess of initiator. After smearing (required to account for the neutron optics of the measurement), the power law fit for the SANS data over the range 0.004 Å1 < q < 0.05 Å1 corresponds to a slope of 3.61. Fitting of the USANS data (once again after smearing to account for the optics) yields a slope of 3.4 for the range 0.0005 Å1 < q < 0.005 Å1. Error bars are included on the plot for each point but do not appear (other than in a small region in the high-q USANS data) since the uncertainties were smaller than the points, even for the short irradiation times employed. It is interesting to note that the slope is nearly the same over 3 orders of magnitude in q, far wider than that seen for most of the materials of Table 1. The break seen between the SANS and USANS data in this data set is a result of the difference in the source-scatterer-detector geometry. Although there are computational methods that can be used to address the discontinuity, we have chosen not to use them here in order to show the similarity of the slopes in the SANS and USANS regions. Also, application of these corrections introduces artifacts into the combined scattering curve. For q > 0.09 Å1, the incoherent background from hydrogen provides a background level for I(q). On the other end of the range, at q < 0.0003 Å1 where the curve flattens, the power law fit yields a slope of 2.01. This slope is characteristic for scattering from randomly oriented cylindrical pores penetrating a homogeneous solid. A structural model fitting this region shows that the pores have a minimum diameter of about 0.6 μm and are several orders of magnitude longer in length. The meaning of the slopes for the regions where 0.0005 Å1 < q < 0.5 Å1 is less clear. A slope of 4.00 would result from 9203

dx.doi.org/10.1021/ac202238r |Anal. Chem. 2011, 83, 9201–9205

Analytical Chemistry

LETTER

Table 2. Nitrogen Absorption Results for 10/90 DVB/4VA Monoliths surface area, SA (m2/g)

pore volume (cc/g)

monolith (a)

14.4

7.96  102

22.11

1.359

monolith (b)

14.06

6.14  102

17.46

1.885

monolith (c)

16.99

9.16  102

21.56

1.399

average

15.15

7.75  102

20.38 (RSD = 12%)

1.548 (RSD = 26%)

average pore diameter (nm) (4*Vp)/SA

pore diameter (nm) BJH method

fractal slope for the material monotonically decreased from 3.68 to 2.98. This raises the questions of what kind of particles are present, the spacings between them, and the surface qualities of the particles and their aggregates that give rise to a decrease in slope away from 4.00 down toward 3.00, particularly in light of the similarity in slope for many of the materials in Table 1. As mentioned above, optimization of monolithic media requires a method for seeing changes introduced by variation of a large number of synthetic variables. To verify the ability to see changes in the characteristics of the monoliths as a function of one of these synthetic conditions (the variation in ratio of monomer to cross-linker), a series of three activated monoliths were compared. The results are shown in Figure 3. The change in shape of the resulting curves in the range 0.01 Å1 < q < 0.1 Å1 shows there is a significant change in population of features in the length range of 60 Å < r < 600 Å. The relationship of this change to the pore shapes and surface features warrants further investigation.

Figure 2. SANS/USANS data for SANS/USANS data and corresponding slopes of power-law fits for monolith prepared with excess of initiator.

Figure 3. Comparison of SANS data acquired for three different 4-vinylaniline/divinylbenzene ratios in the reaction mixtures. The ratios which appear in the legend indicate the 4VA/DVB ratio.

scattering from an interface between volumes that are semiinfinite on the length scale of q and that have different scattering properties. If the slope is 3.00 or less, the scattering occurs within a bulk region.9 The slopes for the monoliths characterized here are intermediate between 4 and 3. It is instructive to look at these intermediate values in the context of those for the series of carbon-filled polymers studied by Wignall.6 He found that, as carbon black became a greater part of the composite, the

’ CONCLUSIONS SANS and USANS are shown here to be valuable tools for investigation of the internal structures of porous monoliths. The drawback of the requirement of a cold neutron source is balanced by the greater amount of information which can be obtained and the speed with which it is possible to acquire data for multiple sample sets. This is particularly interesting for all types of monolithic media in view of Guiochon’s assertion that a tool is needed which will allow estimation of the average size of the domains (a quantity which at present relies on SEM measurements of bed sections) for media prepared under a large variety of conditions2 His definition of domain is the average sizes of the throughpores and porons. The diameters, tortuosities, lengths, and spacings between the channel all contribute to such an estimation, and we might be able to approach these variations using the fractal slope and the range over which these appear in these scattering experiments. In addition, contrast matching should provide a means to look at adsorbed species or to detect solvent-inaccessible regions. In general, the data collected in Table 1 suggest that a closer look at the physical meaning of the fractal slope of the I(q) vs q data is warranted. In addition, the variations seen suggest that it might be possible to develop physical (or in some cases, chemical structural) explanations for the trends seen. ’ ASSOCIATED CONTENT

bS

Supporting Information. Additional information as noted in text. This material is available free of charge via the Internet at http://pubs.acs.org.

’ AUTHOR INFORMATION Corresponding Author

*Address: Department of Pharmacology, SE402 Med-Dent, Georgetown University, 3900 Reservoir Rd, NW, Washington, 9204

dx.doi.org/10.1021/ac202238r |Anal. Chem. 2011, 83, 9201–9205

Analytical Chemistry DC, 20057. Phone: 202-687-3627. Fax: 202-687-8825. E-mail: [email protected].

’ ACKNOWLEDGMENT The authors are grateful to Dr. Kenneth Rubinson and Dr. Paul Butler for their assistance in acquiring the SANS and USANS data and for helpful discussions on their interpretation. B.G.K. acknowledges the support of the National Science Foundation through the Research Experiences for Undergraduates Program (CHE-0552586). The authors also wish to acknowledge the support of the National Science Foundation for partial support of the equipment at the NIST Center for Neutron Research under agreement DMR-0454672 and for acquisition of the FESEM instrument used for these studies (DMR-521170).

LETTER

(27) Kline, S. R. J. Appl. Crystallogr. 2006, 39, 895–900. (28) Wang, Q. C.; Svec, F.; Frechet, J. M. Anal. Chem. 1993, 65, 2243–2248. (29) Gilbert, E. P.; Reynolds, P. A.; White, J. W. J. Chem. Soc., Faraday Trans. 1998, 94, 1861–1868.

’ REFERENCES (1) Buchmeiser, M. R. Polymer 2007, 48, 2187–2198. (2) Guiochon, G. J. Chromatogr., A 2007, 1168, 101–168. (3) Lammerhoffer, M.; Svec, F.; Frechet, J.; Lindner, W. Anal. Chem. 2000, 72. (4) Svec, F.; Frechet, J. J. J. Anal. Chem. 1992, 64, 820–822. (5) Urban, J.; Eeltink, S.; Jandera, J.; Schoenmakers, J. J. J. Chromatogr., A 2008, 1182, 161–168. (6) Wignall, G. D.; Farrar, N. R.; Morris, S. J. Mater. Sci. 1990, 25, 69–75. (7) Rogante, M.; Lebedev, V. T. Polym. Eng. Sci. 2007, 47, 1213–1219. (8) Kim, M.-H.; Glinka, C. J. Microporous Mesoporous Mater. 2006, 91, 305–311. (9) Schmidt, P. W. J. Appl. Crystallogr. 1990, 24, 414–435. (10) Hoinkis, E. Adv. Colloid Interface Sci. 1998, 7677, 39–55. (11) Vacher, R.; Woignier, T.; Pelous, J. Phys. Rev. B 1988, 37, 6500–6503. (12) Li, Z. H.; Gong, Y. J.; Wu, D.; Sun, Y. H.; Wang, J.; Liu, Y; Dong, B. Z. Microporous Mesoporous Mater. 2001, 46, 75–80. (13) Glinka, C. J.; Sander, L. C.; Wise, S. A.; Hunnicutt, M. L.; Lochmuller, C. H. Anal. Chem. 1985, 57, 2079–2084. (14) Bale, H. D.; Schmidt, P. W. Phys. Rev. Lett. 1984, 53, 596–599. (15) Radlinski, A. P.; Mastalerz, M.; Hinde, A. L.; Hainbuchner, M.; Rauch, H.; Baron, M.; Lin, J. S.; Fan, L.; Thiyaagarajan, P. Int. J. Coal Geol. 2004, 59. (16) Radlinski, A. P.; Busbridge, T. L.; Gray, E. M.; Blach, T. P.; Cheng, G.; Melnichenko, Y. B.; Cookson, D. J.; Masstalerz, M.; Esterle, J. Langmuir 2009, 25, 2385–2389. (17) Anovitz, L. M.; Lynn, G. W.; Cole, D. R.; Rother, G.; Allard, L. F.; Hamilton, W. A.; Porcar, L.; Kim, M-H Goechim. Cosmochim. Acta 2009, 73, 7303–7324. (18) Inui, M.; Takeda, S.; Maruyama, K.; Kawakita, Y.; Tamaki, S.; Imai, M. Phys. B 1995, 213-214, 552–554. (19) Przenioslo, R.; Wagner, J.; Natter, H.; Hempelmann, R.; Wagner, W. J. Alloys Compd. 2001, 328, 259–263. (20) Avdeev, M. A.; Blagoveshchenskii, N. M.; Martynov, P. N.; Mel’nikov, V. P.; Novikov, A. G.; Puchkov, A. V. Phys. Solid State 2010, 52, 985–987. (21) Brown, R. J. C.; Evans, M. J. B.; Tun, Z. Can. J. Phys. 2010, 88, 707–713. (22) Melnichenko, Y. B.; Wignall, G. D. J. Appl. Phys. 2007, 102, 021101/021101–021101/021124. (23) Tripp, J. A.; Svec, F.; Frechet, J. M. J. J. Comb. Chem. 2001, 3, 216–223. (24) Barrett, E. P.; Joyner, L. G.; Halenda, P. P. J. Am. Chem. Soc. 1951, 51, 373–380. (25) Glinka, C. J.; Barker, J. G.; Hammouda, G.; Krueger, S.; Moyer, J. J.; Orts, W. J. Appl. Cryst. 1998, 31, 430–445. (26) Barker, J. G.; Glinka, C. J.; Moyer, J. J.; Kim, M. H.; Drews, A. R.; Agamalian, M. J. Appl. Crystallogr. 2005, 1004–1011. 9205

dx.doi.org/10.1021/ac202238r |Anal. Chem. 2011, 83, 9201–9205