PROTON MAGNETIC RESONANCE AND INFRARED SPECTRA OF

May 1, 2002 - Pickup, Frank D. Blum, Warren T. Ford, M. Periyasamy. Transport of small molecules in ... Joseph J. Pesek and Jack H. Frost. Adsorption ...
1 downloads 0 Views 1MB Size
1150

JOHN E.GORDOK

diffusion controlled if the process is not significantly endothermic. As mentioned previously, azulene y a s observed to be the best of the tested quenchers. Calculations based on three measurements yielded a value of k,/k, = 1100. Azulene presumably has more than one triplet lower in energy than the lowest triplet of benzophenone. The lowest azulene triplet probably lies in the infrared, a few thousand wave numbers below the lowest excited singlet state which is responsible for absorption a t 15,000 cm.-l (6000-7000 8.).16Two higher triplets, corresponding to the singlets obtained by the transitions at 3500 and 2800 A . , I 6 should lie below or near to the benzophenone triplet. It is presumed that benzophenone is quenched by excitation of azulene to an upper triplet, since this process best conserves electronic energy. It was of interest to determine experimentally if azulene also could quench by energy transfer to its lou7est triplet. 9-Anthraldehyde was selected as the donor, since the energy of its lowest triplet must be a t least as low as that of anthracene (14,700 cm.-l or 42 kcal.13) and probably slightly lower. This value would be well below the expected energy

Vol. 66

of an upper triplet of azulene, hence energy transfer to an upper triplet would not be anticipated. With no azulene present, 9-anthraldehyde was photoreduced by 0.15 M tributylstannane with a quantum yield of 0.045 to give the corresponding glycol, 1,2-di-(9-anthryl)-ethan-1,2-diol. Under similar conditions, in the presence of 5 X 10-4 M azulene, the quantum yield was reduced to 0.025 (Table 11). Thus, quenching again was found to calculated to be 5300. be significant, with IC& Since the systems are different, this number cannot be compared directly with the data from the benzophenone-benzhydrol system. The conclusion t o be drawn is that quenching by energy transfer to the lowest triplet of azulene does occur. TABLE PHOTOREDUCTION

OF

STANNANE IN Bnthraldehyde Stannane

1.0 X 1.0 X 10-8

11

g-ASTHR-4LDEHYDE BY TRIBUTYLTHE PRESEXCE OF AXULEXE

0.148 0.148

Aeulene

., ,.. 5 X

Q

k,/lcr

0.045 0.025

5300

,

.

Acknowledgment.-We are indebted to Mr. Harold Thomas, who carried out the experiments with cyclooctatetraene.

PROTOY MAGNETIC RESONASCE AND ISFRARED SPECTRA OF SOME ION-EXCHAKGE RESIN-SOLVENT SYSTEMS BY JOHN E. GORDON Mellon Institute, Pittsburgh, Pennsyloania Recsived January 4, 1968

The proton magnetic resonance spectra of a variety of ion-exchange resins immersed in several liquids and liquid mixtures have been recorded, and these are found to be remarkably well defined. Due to the close similarity of the diamagnetic sus~ beceptibilities of Kater and the resins, rather narrow lines are observed in these systems; due to the fact of s 1 0 exchange tween water mithin and without the resin beads, separate lines appear for water in the two environments. These circumstances allow a number of phenomena of importance in the physical chemistry of the exchangers t o be observed directly; among these are the exchange of proton-containing counterions, the change of counterion molality with cross-linking, the distribution of mixed solvents between the phases, and the loss of rotational freedom of counterions in the resin phase as a function of the swelling. In addition the spectra provide a powerful tool for the investigation of resin heterogeneity; several interesting cases of thr existence of two discrete, though closely similar, fractions in commercial resins are described. Infrared spectra and swelling measurements indicate that hydrogen resinate in anhydrous acetonitrile exists predominantly as the undissociated sulfonic acid, in equilibrium 11 ith a small amount of lyonium sulfonate ion pairs, while in dioxane the converse is true.

Introduction The description of chemical structure within a particle of ion-exchange resin in contact mith solvent is a problem of considerable interest and importance. For example, the nature of ion association in mater-swollen exchangers, heterogeneity of exchange sites within and between individual resin particles, interaction between counterions and between fixed exchange sites, and the state of the sorbed liquid are incompletely understood topics which have been approached primarily by interpretation of thermodynamic, diffusion, and conductance data.l On going to mixed and non-aqueous solvents, these problems are compounded, and new ones emerge, e.g., the degree of solvolysis of -S03H groups. One might. expect spectroscopic methods to contribute rather more (1) See J A Iiitchenei in J O ’ V Bookri; ‘ Nodein Aspects of LlcLtioclieinistrr,” Academic R e s - Ncu Yolk K T 1959 11 87

directly to the elucidation of some of these questions; however, only a single report on the infrared spectra of ion exchangers2 appears to have been made. The present work consists of attempts to (i) discover what phenomena may be observed in the proton magnetic resonance (p.m.r.) spectra of resin-solvent systems, and (ii) to employ the infrared spectra as a criterion of the chemical state of the resin exchange functions 111 non-aqueous media. The p.m.r. section contains descriptions of the spectra along with attempts to identify the molecular origins of the effects observed and generally to understand the spectra in terms of known principles of single-phase systems, followed by a discussion of what contribution the spectra can make to understanding the physical chemistry of the exchangers ( 2 ) h S t i a ~ l i e i i na n d IC I:UIJ’

Spa fiorhm

l i f a , 17, 388 (1961)

June, 1962

PROTON M A G N E T I C

RESOSANCE OF ION-EXCHANGE RESIN-SOLVENT SYSTEMS

Experimental Materials .--Resins m r e principally purified grades of Dowex 50W and Dowex 1 obtained from Bio-Rad Laboratories, Richmond, California as AG 50W and AG 1. They r e r e converted to the desired ionic states by standard column procedures, and when anhydrous materials were desired, were first air dried t o 10-20~0water, then dried to constant weight a t 75-80’ and a final pressure of about mm. (6-8 hr ,). Direct potentiometric Karl Fischer titration under methanol showed that the materials so prepared contained no more than 0.04 mmole of water per meq. of resin. Acet,onitrile and dioxane were reagent materials distilled from p.hosphorus pentoxide and sodium, respectively, through a 61-cm. glass helices-packed column; constantboiling center cuts possessing properties agreeing with published values were used. Water was deionized with a mixed-bed ion-exchange column. Spectroscopic Measurements.-Infrared spectra were determined on a Beckman IR-4 spectrophotometer equipped with a rocksalt prism. Air-dried resin samples nfere repeatedly ground in a stainless steel ball mill on a Wig-L-Bug amalgamator, then dried to constant weight as above, and sealed in vacuo; opening, mixing with Sujol or the appropriate solvent, and pressing between rocksalt plates was done in a drybox. P.m.r. spectra were recorded on a Varitn 9-60 spectrometer whose probe temperature mas 32 j, 2 . SUSpensions of resin beads in t’he appropriate solvent (pr’epared in the drybox where necessary) were observed in Wilniad Glass Co. semi-precision, 5 mm. X 15 em. cells. Swelling Measurements.-Selected 50-100 mesh beads were dried as above, and the diameters were measured, before and after addition of solvent, with a microscope equipped with an eyepiece micrometer which allowed estimation to about 0.004 mm., the beads being protected from the atmosphere by enclosure between two greased cavity slides.

Results and Discussion Proton Magnetic Resonance Spectra.-Most of the observations were made on the Bio-Rad Dowex resins and these are summarized in Table I; results with some other materials are described in a later section. Some examples of these spect’ra, which are seen to be remarkably well defined, are given in Fig. 1 and 2. Several features are common t o most of the spectra: (1) Separate signals arising from liquid within and without the resin particles are observed3; the sign and magnitude of the shift’ of interior relative to ext’erior liquid depends both on the nature of the proton under observation and the nat’ure of the resin salt.. (2) Suspension of the resin in the liquid broadens the exterior liquid resonance and shifts it by ea. 2-10 c.p.s. (3) Lines due to interior liquid and to prot,on-containing counterions are broadened to an ext’ent dependent on structure. Lines DLE to E:xterior Liquid.-The shift produced in the position of the exterior liquid line by the presence of the solid can be accounted for to a good a,pproximation as the result of the change in bulk diamagnrtic susceptibility of t,he sample. This is best shown using spheres of pure crosslinked polystyrene, for which the volume suscepti(3) Identification of interior a n d exterior liquid resonances is accomplished by manipulation of the sample cell to place a more dense or less dense suspension of the solid within the probe; the intensities of the two signals then are observed t o vary in a continuous, complirnentary fashion. Further evidence t h a t production of two lines is not a n artifact of the heterogeneous nature of the sample IS obtained: (1) using spheres of pure crosslinked polystyrene which are not penetrated by water-in which case water suspensions display only the exterior water line. while suspensions in toluene, which is distributed into the polymer, again give two lines for both methyl a n d aromatic protons; ( 2 ) b y observing continuous intensity changes, with changing solvent composition, of Ihe four expected lines in resin-binary liquid systems; and ( 3 ) by observing c l i ~ i i x e swith time in the number and positions of lines in niiutures of two different resin salts and water (see below).

1151 (0

in Aqaeous Res-, H1C- b u t a n o l

‘ u i

j

x-8 x-12

H2°

((

Res’, Anisate-

In/

-

x-L X-L

/I

H-

Fig. 1.-Proton magnetic resonance spectra of ionexchange resin-liquid systems. The scale is the same for a, b, and e. Resins are Bio-Rad ACT 1 and AG bOW.

I

I

I

I

I

I

H-

Fig. 2.-Proton magnetic resonance spectra of ionexchange resin-liquid systems. Resins are Bio-Rad AG 1 and AG 5OW.

bility is easier to estimate. Table I1 shows how the observed shifts for four liquids do indeed fall in the order of their susceptibilities; using these

JOHN E. GORDON

1152 DATAFROM 60 M C . P.M.R.

Na +

%

DVB

Partide sizeb

C B

H+

1

c

HzO

s

c c

HzO HzO H2O Hz0 HzO Hz0-t-BuOII HzO HzO HzO-t-BuOH HZO-t-BuOH HzO HzO-t-BuOH

H+ H+ H' HT H+ H+ NHI +(I1')I +(K H , ) ~ ?;H( +(

12 8

8 8 8 8 8 8

c1-

4

"202-

HCOZOH OH Anisate -

4 4 2 4 2

Anisate-

-

Anisate

C

HzO-t-BuOH

B C

Hz0 HzO

c

A C D

C C C

C C c

(C.P.8.)

Width ofQ interior liquid resonance (C.P.S.)

Width ofa exterior liquid resonance (C.P.S.)

E20

HzO HzO-t-BuOH

HzO

17 -2 6 9 8 -24 -27 -43 -47 - 79 - 102 -86 -79 - 73 -89 - dioxane > aqueous dioxane > water, is indeed the order of solid-liquid susceptibility differences as inferred above, avid the exterior liquid line width increases with decreasing particle diameter (decreasing mean distance of liquid molecules from the solid) as expected. The sharpness of the spectra in water is thus a result of fortunate near identity of the liquid and solid susceptibilities. A quantitative estimate of the line broadening from this source can be made for the simplified model obtained by replacing the real sample by a non-interacting collection of individual, spherical solid particles each surrounded by a spherical shell of liquid, the diameters being chosen to maintain the observed volume fraction of liquid in the sample. This was done by computing the component, in the direction of the applied field, of the field due to the magnetized sphere a t radial intervals across the liquid region by numerical integration over the solid sphere of

sv

H = HOAX

(1 - 3 cos2 0 1 ) ~ - sdV

both in the direction of the applied field and perpendicular to it. Here Ho is the applied field, Ax the difference in susceptibility of liquid and solid, a: the angle between the induced dipole in the volume element dV of the solid sphere and the line connecting the dipole with the point in the liquid region a t which the field is to be calculated, and T is the distance of this point from the dipole. From these results the field at intermediate angular orientations can be obtained and the entire (4) J. A. Pople, W. G. Sohneider, a n d H. J. Bernstein, “High-Resolution Nuclear Magnetic Resonance.” MoGraw-Hill Book Co., New York, N. Y., 1959, p. 408.

1153

liquid region thus mapped. The results of this procedure for a liquid fraction of about 0.4, the measured value for several of the resins in water, gave a distribution of local field strengths corresponding to a line width a t half-height, AH = 7.2AxHo; for the susceptibilities employed above, this is about 17 and 86 C.P.S. a t 60 Me. for water and acetonitrile, respectively (observed : ea. 6 and 40 C.P.S.).~ Reverting from this simplified case to the actual sample where each sphere is in contact with several neighbor spheres, the line is expected to be narrower than the above prediction for two reasons: First, the mean distance of the proton from the sphere will increase; and second, the presence of neighboring spheres will decrease the inhomogeneity of the field (e.g., a t the center of a void in close-packed uniform magnetized spheres the field due to the spheres is zero). The observed broadening thus can be said to be of the correct order of magnitude for the effect suggested. The Line Positions of Interior Liquid and Counterion Resonances.-Since, to a good approximation, the resin particles are perfect spheres, there will be no contribution to the observed shifts of interior from exterior liquid resonances due t o the difference in diamagnetic susceptibility of liquid and solid.6 Two factors which are important are shifts produced by solvation of fixed- and counterions, and anomalous diamagnetic shielding effects produced by the aromatic nuclei‘ of the resin matrix. As a first approximation, the first of these factors might be neglected for carbon-bound protons, and we can hope to account for the shifts of the methyl resonance of t-butyl alcohol on the basis of the second factor alone. lirom Table I the observed shifts for t-butyl alcohol on entering the resin phase are indeed always positive, Le., to higher applied fields, as required by the effect under consideration. From the magnitude of the high-field shift observed by Bothner-By and Glick for aliphatic solutes dissolved in aromatic liquids, the densities of the resin^,^.^ swelling measurements, and the capacity of the resins, one can estimate the following values for the expected shifts due to the diamagnetic anisotropy of the polystyrene matrix: for cation exchangers, 14-18 c.p.s.; for anion exchaiigers, $9 c.P.s.; for anion exchangers chsrged with aromatic counterions, $23 c.p.s These estimates are for the case of random orientation of the molecule with respect to the aromatic ring. The observed shifts for f-butyl alcohol in the cation exchangers (average = +16 c.P.s.) agree well with the prediction, while those for the anion exchangers (4 and 12 c.P.s., without slid with an aromatic counterion) are in the correct direction but have about half the

+

( 6 ) The range of a/ao ( a = distance of liquid molecule from the center of the solid sphere of radius aa) included was 1.0185 t o 12035: neglect of the innermost region affects the intensity of the line a t its far extremities b u t has little influence on the width a t half height. (6) Reference 4, p. 81. (7) 4. A. Bothner-By and R. E. Ghck, J . Chem. Phw., 26, 1651 (1957). ( 8 ) H. P. Gregor, J. Belle, and R . A. Marcus, J . Am. Chem. SOC., 76, 1984 (1964). (Q) H. P. Gregor, B. Gutoff,and J. I. Bregman, J . Collozd Sea., 6,245 (1951).

1154

JOHK E. GORDON

1’01. 66

predicted magnitude. Some analogous information for homogeneous solution in 5% butanol is obtainable from Table 111; here also the effect of the aromatic nucleus is smaller for the aromatic cation than for the aromatic anion, though the average shift is x7ery close to that predicted for the anomalous diamagnetic shielding due to 5 m aromatic nuclei (22 c.P.s.). These results probably reflect orientation effects in the ion-solvent interaction; the result for sodium chloride shows further that the effect of inorganic ions on methyl proton line positions can be sizeable. These and other effects (e.g., inter-chain orientation effects, exclusion of liquid from regions of high cross-link density) doubtless contribute something to the resin results, though the anomalous shielding of the aromatic matrix, in random orientation, appears to account for the bulk of the shift.

12% DVB resins, in reasonable agreement with swelling measurementsll on similar materials of 1950-1951 vintage which gave 1.07, 2.45, and 4.80 m for the first three cases. I n 80% dioxane, the apparent general increase in the shift to lower field of interior relatke to exterior water, as compared with the pure water suspensions, is in large part artificial, as it includes a shift of +55 C.P.S. in the position of the exterior water line on going from pure water to 80% aqueous dioxane, a result of the change in hydrogen bonding equilibria. This semi-quantitatively accounts for all of the observed shifts except that in the hydrogen resinate, where replacement of some of the interior water by dioxane has increased the fraction of HaO+in the H30+-HzO rapid exchange system, leading to an additional shift to lower field. The only case for which the position of the TABLE I11 counterion resonance has been measured is the 2% SOLVENT LINE POSITIOXS FOR 5.00 m SoLumoss OF SOME DVB tetramethylammonium resinate (Fig. 2e). ELECTROLYTES I N 5% AQUEOUS&BUTYL ,~LCOIIOL RESince salt effects on the tetramethylammonium ion FERRED TO PURE EXTERNAL SOLVENT MIXTURE resonance in homogeneous solution are small, shift (c.p.s.)the line position yarying no more than A 2 C.P.S. Solute Hz0 ( CHd a COH with concentration over the range 0.5-3 m, the NaCl 8.5 -12.2 shift of the interior line from that displayed by a XMe4C1 2 3 - 2.9 surrounding tetramethylammonium chloride soluCsH6NMe&1 8.3 12.8 tion of approximately the same concentration HC02Na - 0.9 - 1.2 may be expected to result primarily from the anomC6H,C02Ka 7 2 31.2 alous shielding of the resin matrix. The expected HNOI or HC1 -83 5a shift from this source, estimated from the observeds NaOH -26b resin volumes as before (assuming the deswelling a Interpolated from reference 10; for pure v-ater and corrected for bulk susceptibility. Interpolated from H. S. due to exterior tetramethylammonium chloride to be the same as that observedQfor an equal conGutowsky and A. Saika, J . Chem. Phys., 21, 1688 (1953). centration of ammonium chloride), is +9 c.P.s., The shifts of interior from exterior water con- in good agreement with the obserred +8 c.p.s. trast markedly with those for t-butyl alcohol, de- shift. pending strongly upon the resin structure, as Line Widths of Interior Liquid and Counterion expected for a direct chemical interaction with the Resonances.-In contrast to the situation for exelectrolyte functions present ; the observed effects terior liquid molecules near resin beads considered are in all respects analogous to those produced by as magnetized spheres, niolecules within the electrolytes in homogeneous solution. The order spheres should experience a uniform field12 and of interior water line positions for the various their resonance lines should remain unbroadened, counterions in the resin spectra (H+ < nl(CH3)4+< except that in a dense suspension of spheres iSa+ and OH- < HC02- < C1-) is the same as that broadening due to neighboring spheres may bemeasured (Table 111) in homogeneous solution, come important. The systems at hand indeed though there is probably some aromatic high- show some broadening (about 4 C.P.S. us, 2 C.P.S. field shift, superimposed upon this variation. for the homogeneous liquid) of the interior water Further, taking phenyltrimethylammonium chlo- line and this is less than that for exterior water, ride solution, a t the concentration prevailing in as expected. Since the interior mater line width the resin, as a homogeneous model for the resinium is nearly independent of the nature of the electrochloride, the expected water line shift, f3.6 c.P.s.. lyte functions present and of the cross-linking, but is in good agreement with the observed shift of depends more markedly 011 the particle size, about $4 c.p.s. Finally, the dependence of the approaching the line width of pure water at large interior water shift, for a given counterion, upon diameters, it is reasonable to attribute all of the the degree of cross-linking of the resin follows the detected broadening (except in the case of chemiexpected pattern for changes in electrolyte con- cally heterogeneous resins treated below) to the centration paralleling swelling changes dependent field inhomogeneity arising from the particulate on the cross-linking. This can be quantitatively nature of the sample. tested for the hydrogen resinates by combining Except for formate, the broadening of the organic the observed shifts with the relation between counterion lines is generally greater than that of chemical shift and molality establishedI0for strong interior water in the same resin, and it increases acids in homogeneous aqueous solution; the very rapidly with increased cross-linking (Table molality of hydrogen counterions obtained in this (11) G. E. 13ojd and B. A , Soldano, J . A m Chem. Soc , 76, 6092 way is 1.34, 2.50, 4.36, and 5.70 for the 2 , 4 , 8 , and (10) G C . Hood, 0. Redlicli, and C. A Reilly, J. Chem. Phus., 22, a067 (1964).

(1953). (12) L. Page and N. I. Adanis, Sr., “Prinoiples of Electricity.” D. Vau Nostiand Co , Princeton, N.J., 3rd Ed. 1928, y. 123.

June, 1962

PROTOK &IilGi\iETIC

RESOSANCE O F ION-EXCHANGE RESIS-SOLVEKT SYSTEMS

I ; Fig. 1). This is presumably the result of increasing viscosity of the gel as the volume fraction of the hydrocarbon network increases accompanying increased cross-linking and concurrent decreased swelling, leading to loss of rotational freedom in the ion and enhanced dipolar broadening13 of its resonance. These effects correspond directly to the behavior of self-diffusion coefficierW4 of resin components, where diffusion of counterions is found to be more drastically reduced, relative to homogeneous solution, and more dependent upon the cross-linking, than is diffusion of maher. The depression of the equivalent conductance of counterions, which also increases with decreasing swelling of the resin, is attributed to the same source. l5 Figure 3 shows how the counterion line widths extrapolate smoothly to the line widths observed in homogeneous aqueous solution at zero crosslinking, both for the tetramethylammonium ion and for the anisate ion. This is of interest because, though the tetramethylammonium resinate appears to be unassociated in l6 exchange thermodynamicsi7 and swelling measurements8 have been interpreted as indicating extensive ion association in anion exchangers charged with large organic ions. 'The absence of a discontinuity, in the width of the anisate ion line, between the resins of low cross-linking and homogeneous aqueous solution indicates that (since the ion association itself does not go to zero at lorn cro~s-linking'~) formation of ion pairs between counterions and the charged resin matrix produces no further dipolar broadening of the counterion resonance. It appears that both of the deficiencies which might be anticipated in the investigation of counterion line broadening as an index of ion association in these systems are operative-that is, bound ions retain enough motion i o give rather sharp lines while at the same time free and associated ions suffer broadening induced by the physical nature of the medium. Anhydrous Dioxane and Acetonitrile.-Spectra in these liquids were measured for hydrogen, arnmonium and tetramethylammonium resinates, and for resinium formate and anisate; all of the spectra are very similar and have the following characteristics: (a) The exterior liquid line is greatly broadened, as previously discussed, the average widths a t half height being ea. 25 and 40 c.p.s., respectively. (b) X o lines due to protons in countexions were detected. (c) KOlines plausibly attributable to interior liquid were observed. Shoulders on the high-field side of the exterior liquid line were present in some cases, though it was difficult to separate these completely from the spinning side bands which are virtually impossible to eliminate from these spectra. It seems probable (13) Reference 4, p. 29. (14) ( a ) G. E. Boyd and B. A. Soldano, J . Am. Chem. Soc.. 76, 6091, 6099,6105, 6107 (1953); (b) M. Tetenbaum and H. P. Gregor, J . P h s . Chew , 58, 116: (1954). (15) A. Dsspi6 and G J. Hills, Trans. Faraday Soc., 6 1 , 1260 (1955). (16) H. P Gregor, D. Nobel, and AI. H. Gottlieb, J . Phys. Chen., 59, 10 (1955). (17) H. P. Gregor, J. Belle, snd R. -4.Marcus, J . Am. Chem. Soc., 77, 2713 (1955).

4

a

1155

12

Nominal Crosslinking , %DVS. Fig. 3 --Suclear magnetic resonance line widths for counterions in ion-exchange resin-water suspensions us. the degree Of cross-linking of the resin: 0 , tetramethylammonium ion; 0, anisate ion (methyl): A, anisate ion (aromatic).

that interior liquid lines, where present a t are buried under the exterior line; this was shown to be the case for dioxane in SO% aqueous dioxane by direct examination of the solid phase. The absence of counterion resonances doubtless reflects the very low fluidity of the resins in contact with these liquids which would be expected from the swelling measurements. Resin Homogeneity.-In all of the resins so far described, the interior water line in the spectra of the water suspensions is only slightly broadened compared to the pure liquid, and this broadening is reasonably attributed to the particulate nature of the sample. The presence of a single, narrow line for interior water implies a high degree of homogeneity, both intra- and inter-particulate, in these materials. In the first sense this is an independent confirmation of the internal structure of the swollen resin1* as a homogeneous phase of considerable fluidity in which the sorbed liquid molecules all have the same time-average environment.19 This does not mean that cross-linking or other gradients do not exist within a single particle, but only that there is rapid exchange between water molecules in the various physicochemical situations present. On the other hand, since there is no rapid exchange between particles, and since there is a strong dependence of the interior water line position upon the counterion molality, especially in the case of the hydrogen resinates and resinium hydroxides, the existence of a single, sharp interior water line places narrow limits on the inter-particle heterogeneity of the sample (18) Reference 1, p 93. (19) It is impossible t o rule out the presence of a fraction of inteiior a a t e r molecules so tightly bound t o the polymer network or so icelike t h a t its resonance line is dipole-broadened t o the point of disappearance; a sizable fiaotion of this type could not be undergoing rapid exchange with the visible interior liquid, however

JOHN E. GORDON P.1t.R.

DATAFOR

Vol. 66

TABLE IV 10s EXCHANGERS IN IJ'ATER

VARIOUS

Particle size

Cross-linking

Bio-Rad AG 50

100-200"

2

Hf

Bio-Rad AG 50 Bio-Rad AG 50

100-200" 200-400"

2

9 a+

4

"

Rexyn AG SOb

16-5OC

Medium porosity

I3 +

Rexyn RG SOb

16-50'

Medium porosity

H

Dowex 50 W d

50- 100a

2

H'

Resin

Amberlite I R 120° Amberlite I R 120" after cycling Dowex 50' Dowex 50' after cycling Dowex 50'

100-200 100-200 200-400

Amberlite I R 120 CPo Amberlite IR 120 CPg Amberlite CG-120h Amberlite CG-120' Perinutit Qi Rexyn RG 1*

2O-5Oc 20-5OG 100-200" 100-200" 16-50" 16-50'

% DVB

2 2

4 High porosity High porosity

Counter ion

1-

0.p.s.

-27.5 -31 +7

-88 -100(sh) -88 - loo(sh) - 27 -30

-84 -84 -28 -28

Na

f 2 -11 -t16 -87 -87 - 15

+

H+ H+ OH-

Width of interior water lint?,

Width of exterior water line,

C.P.8.

C.P.S.

4

-4.5 2.5

6

4 6

8 6

>6 8 ?

6

-6

3

-43 -47

H+ H+ H+ H+ H+

Na + H+ Medium porosity

Shift of interior from exterior water.

15 9 15 3 28

-44

-3 2.5 I1 9 8 3

6 6 3 3 6 -3 2 11 8 6 6

Amberlite I R 45 Free base i i 10 Amberlite IR C 50 Free acid 8 i 19 Dry mesh range prior t o sulfonation, etc. b Fisher Scientific Go. Actual wet mesh range. J. T. Baker Chemical Co. Cat. KO. 1920. e Old stock. ',Purchased in 1956. 0 Mallinckrodt Chemical 3347. Mallinckrodt Chemical 3337. Matheson Coleman & Bell 9055. 1 No interior line.

with respect to cross-linking, degrees of sulfonation, etc. With the hydrogen resinates for example, the presence of an appreciable fraction differing by 0.1 in counterion molality (0.2-0.4 of a nominal %DVB unit in cross-linking) should be detectable. Using the present data, this criterion can be applied to a number of different resins. Most of these are the purified Dowex resins supplied by BioRad Laboratories; with two exceptions discussed below these give evidence of a high order of homogeneity when examined directly as supplied; the narrow interior water lines observed also rule out the presence of paramagnetic counterions. Table IV summarizes the information on heterogeneity and on the resins from other sources. The table is divided into four parts; the first lists those resin samples which give evidence of containing t w o or more distinct fractions. The second part includes resins which displayed an unusually broad interior water line when observed directly as received; in the two cases where it was tried, cycling the resin with sodium hydroxide-hydrochloric acid sharpened the line, suggesting that the broadening was due to the presence of paramagnetic counterions. The third part of Table IV contains materials from other sources whose behavior was normal, and the fourth part consists of resins which show no interior water line. The type of heterogeneity shown by the materials in the first section of Table IV is unexpected; the re>qolutionof the interior water resonance for

the Bio-Rad AG SOW X2, 100-200 mesh hydrogen resinate into two rather sharp lines is shown in Fig. 2b. Therc is, however, some previous evidence for the same sort of behavior. HogfeldtZ0found that the selectivity coefficients for silver-hydrogen exchange measured by tracer techniques on individual resin beads not only varied from bead l o bead, but, with one resin sample, tended to cluster around two distinct values of the selectivity coefficient, probably reflecting clustering about two cross-linkings. The heterogeneous X2 100-200 mesh resinate was further examined. On cycling, the splitting of the interior water line disappeared in the sodium state as expected (the separation would be less than 1 c.P.s.) but reappeared unchanged in the hydrogen state, ruling out any explanation based on foreign counterions. Table V shows the correlation found between the particle size and the p.m.r. data obtained on fractionation of the sample on standard sieves ; this demonstrates the inter-particulate nature of the heterogeneity and shows that the low-field fraction is concentrated among the largest particles. The most likely structural differences involved in producing the different line positions are in the degrees of crosslinking and sulfonation. It is difficult to conceive a chemical process capable of producing, in a single preparation, such a discrete binary mixture of beads with respect to either of these variables. (20) E. I-Idgfeldt, Aikiv Kern;, 13, 491 (1059).

June, 1962

PROTON MAGNETIC RESOSANCE OF ION-EXCHAKGE RESIN-SOLVENT SYSTEMS

A more likely source would seem to be combination, during or after polymerization, of similar but distinct lots of nominally identical material. TABLE V P.M.R.SPECTRA OF A HETEROGENEOUS RESINAS FUNCTION OF PARTICLE SIZE ---Wet screen analysis-? Retained by %

45 mesh 50

Trace 15

60

16

70

22

80 100 120 140

29 15 3 Trace

Interior water line, C.P.S. from exterior water line

A

Relative intensity

.. -27.5 -31 -27.5 -31 -27.5 -31 -27.5 -27.5 -27 (broad)

1 2.5 1.2 1 2.3 1

..

Kinetic Processes.--Observation of separate lines for interior and exterior water sets a lower limit for the mean lifetime of water molecules in either phase; this rangcs from 0.004-0.1 sec. for the resins in Table I, Exchange of interior water with exterior deuterium oxide proved too rapid to measure under the present conditions; the half-time for this reaction in a shallow-bed flow system was of the order of 1 see. in the measurements of Tetenbaum and Greg0r.1~~Seither has it been possible to observe exchange of ions between the resin and an electrolyte solution. The exchange occurring between resin particles charged with different counterions immersed in deionized water is observable, however, and provides a good illustration of the correctiiess of our analysis of the line identities in the simple spectra (Fig. 2d). Other Applications.-Several further uses of spectra of this sort suggest themselves. Figure la implies that it should be possible to measure distribution coefficients between solution and resin phases for binary liquid mixtures. The possibility of measuring selectivity coefficients for protoncontaining counterions is suggested by Fig. 2e. The molality of proton-containing counterions can be obtained by integration of the counterion and interior liquid lines. Determination of counterion molality for hydrogen resinates from the position of the interior water line already has been illustrated. Infrared Spectra.-Observed infrared bands for various ionic states of the cross-linked polystyrenesulfonic acid exchanger are given in Table VI. It is clear that the sodium, ammonium, and hydrated hydrogen resinates have closely similar spectra which are readily distinguished from that of the anhydrous hydrogen resinate. These differences agree well with the results of Detoni and Hadii21, who find for simple arylsulfonic acids, characteristic bands for the anhydrous acids (ArSOzOH) near 896,221092, 1165, 1342, 2350, and 2900 em.-1, (21) S Detoni and D. Hadti, Spectrochim. Acta. 11, 601 (1957). (22) T h e frequencies quoted here are for p-toluenesulfonic acid; a band near 650 om.-' also is apparent in the reproduoed speotrum.

1'157

while their salts and the hydrated acids (ArS03-) show, instead, absorption near 113OZ3and 1200 em.-', together with changes in bands near 1050 em. -I. TABLE VI BANDS O F THE S U L F O N I C ACID TYPEC A T I O N EXCHANGER AG 50W-X8 IN VARIOUSIONIC STATE#

INFR.4RED

H (dry)

H (hydrated) b

Na (dry)

NH4 (dry)

680s 776w 836m

680s 776w 834m

680s 777w 835m