Nature of the water-epoxy interaction

zyxwvu zyxw zy Macromolecules 1985, 18, 1091-1095 1091 Nature of the Water-Epoxy Interaction zyxwvuts L. W. Jelinski,* J. J. Dumais, A. L. Cholli,t T. S. Ellis,$and F. E. Karaszt AT&T Bell Laboratories, Murray Hill, New Jersey 07974. Received September 6, 1984 ABSTRACT Absorption of small amounts (1-3 wt %) of water by epoxy resins effects a substantial plasticizing action and degradation of their mechanical properties. We present results that provide new insight into the molecular details of the interaction of epoxy resins with water. These results show (1)that the water in epoxy resins is impeded in its movement, with the molecules hopping from site to site with an approximate residence s, (2) that there is no "free" water, (3) that there is no evidence for tightly bound time per site of 7 X water, and (4) that it is unlikely that the water disrupts the hydrogen-bond network in the epoxy resin. These results are consistent with the notion that water acts simply as a plasticizer for these hydroxyl-containing systems, just as it does for polymers which contain no exchangeable protons. Water is known to plasticize certain amorphous,1t2semi~rystalline,~ and cross-linked polymer^.^^ In particular absorption of small amounts (1-3% by weight) of water causes a large reduction in the glass transition temperature (T,)of epoxy resin^,^^^ often considered to reduce the temperature threshold of utility and general mechanical integrity of this class of materials by a considerable margin. This feature of water plasticization in this and other systems has led to some speculation with regard to the nature of the interaction of water molecules with polar polymeric substrates. Here we present results which not only clarify this situation but improve the fundamental understanding of the molecular details concerning the interaction of water with these materials. Some early spectroscopic and calorimetric data have been interpreted to indicate a highly specific "binding" between the water molecules and the polar groups of the epoxy resin,'i8 proposing that the water disrupts the hydrogen-bonding n e t ~ o r k . However, ~ recent work4Js12 shows that the plasticization of epoxy resins by water can be explained simply by treating the T,depression as arising from the expected composition dependence of miscible polymer-diluent systems. The latter treatment has the virtue that it successfully predicts the depression of Tgfor polymer systems that contain polar groups such as OH residues,12as well as for those that are considered relatively inert. In a preliminary communication on the interaction of water (D,O) with epoxy resins,13 quadrupole echo deuterium NMR spectroscopy was used to show (1)that although the absorbed water is very mobile, it is not identical with bulk water in this respect, (2) that the water is distributed homogeneously throughout the sample, (3) that approximately one water molecule is entrained for every six OH/OD sites on the polymer backbone, and (4)that the water exchanges with the OH/OD groups on the polymer backbone, but the time scale of this exchange was not determined. The communication also established that solid-state deuterium NMR spectroscopy shows particular promise for the further investigation of plasticization phenomena in these and other systems. We report here an extension of the early study.14 The present work is designed to clarify and further establish the molecular details of the interaction between water and epoxy resins. In particular, the experiments described here are intended to (1)obtain in residence time for the mobile water molecules, (2) examine the effect of water plasticization on the backbone mobility of the epoxy resin, (3) determine whether any fraction of the water can be considered to be tightly bound to the epoxy resin, and (4) Current address: BOC Corp., Murray Hill, NJ 07974. * University of Massachusetts, Polymer Science and Engineering Department, Amherst, MA 01003. 0024-9297/85/2218-1091$01.50/0 zyxwvu establish upper limits on the rate of exchange between the water molecules and the OH/OD groups on the epoxy backbone. Materials and Methods Epoxy Resins. The epoxy samples used in this study are based zyxwvu on the diglycidyl ether of bisphenol A. The sample we designate as EX4 was prepared as a stoichiometric cure of Epon 825 (Shell Co.) (epoxy equivalent weight (eew) 175; theoretical eew of pure monomer is 170) with m-phenylenediamine. The calorimetrically determined onset T , of this completely dry, fully cured sample is 424.2 K and the midpoint is 432.0 K (heating rate 20K/min). The depression of Tgwith water plasticization4 shows no discontinuity in the depressed Tgup to a water content of approximately 3.5% w/w. Sample NX2 was made according to literature methods15using Epon 825 and nadic methyl anhydride (Aldrich Chemical Co.). This sample has an approximate stoichiometry of 1.00.75 epoxide to anhydride groups. An IR spectrum shows no evidence of residual anhydride or epoxide moieties, although a small number of OH groups are present in this sample. The onset Tgis 386 K and the midpoint is 392.6 K (heating rate of 20°/min). Poly(N-vinylpyrrolidone)(M, 360 000) was obtained from Aldrich Chemical Co. This material was dried in a vacuum oven at 180 "C for 24 h prior to use. Figure 1 shows the chemical structures of the materials used in this study. Exchange with Deuterium. The epoxy samples were exchanged with DzO in one of two ways. In the first method, an ampule of D 2 0 was placed in a glass tube containing the epoxy. After evacuation and sealing, the ampule was broken by shaking and the whole assembly was placed in an oven at 160 "C. This process was repeated 5 times, each time with a fresh ampule of DzO, to ensure complete exchange. After the final cycle a given amount of D,O was introduced into the system and. the final weight percent DzO was determined gravimetrically. The samples plasticized in this manner contained between 1and 3 wt % D,O, with the exception of NX2, which absorbed less than 0.5 wt % D20. In the second method, 2 X 2 X 2 mm cubes of the dried epoxy samples were soaked in D 2 0 a t room temperature for a period of 1 week. At the end of the soaking period, the samples were blotted dry and allowed to equilibrate in a 100% relative humidity D 2 0 atmosphere16 for 2 days prior to NMR measurement. This method produced samples containing approximately 2 wt % DzO. Samples prepared by either method produced equivalent NMR results. Nondeuterated epoxy resins were plasticized with Me$O-d6 by soaking the resin in this material for 48 h. The samples were blotted dry prior to NMR measurement. They contained 5% by weight Me2SO-d6. The poly(N-vinylpyrrolidone)was plasticized with D 2 0 by exposing the dried polymer to DzO in the 100% relative humidity chamber for 24 h. This produced a sample containing -40 wt % DzO. Several hours of drying under a slight vacuum a t room temperature reduced this to 21% by weight of DzO. Exchanged and Dried Samples. Some of the samples that had been previously exchanged with DzO were dried to provide reference samples. The dried samples were produced by heating 0 1985 American Chemical Society zyxw zy zyxwvutsrqpon zyxwvutsrq zyxwvutsrqponm zyxwvutsrqponmlk Macromolecules, Vol. 18, No. 6,1985 1092 Jelinski et al. (a) ,O, CHZ-CH-CH2-0- (C) 0 f"b 0 -C- 0 /I 0 -0-CH2-CH-CH2 zyxwvutsrqponm cHgo .**-N-CHz-CH-CH2-OI CH3 b (d) N 1w OH N1 0 (e) 4 CH-CH,+, I O m zyxwvutsr zyxwvut Figure 1. Chemical formulaa of materials used in this study: (a) diglycidyl ether of bisphenol A; (b) m-phenylenediamine;(c) epoxy resin EX4; (d) nadic methyl anhydride; (e) poly(N-vinylpyrrolidone). the deuterium-exchanged resin to 170 "C in vacuo for 2 days in the presence of phosphorus pentoxide. These sampleswere sealed under dry nitrogen until the NMR experiments were performed. NMR Measurements. Each NMR measurement was performed on a fresh sample (approximately 100 mg) of the appropriate polymer unless noted otherwise. NMR experiments were performed by using a home-built solid-state deuterium NMR spectrometer operating at 55.26 MHz for deuterium. The spectrometer" and its sample temperature controlls have been described previously. The quadrupole echo pulse s e q u e n ~ e 'was ~*~ used for most experiments. Relaxation data were obtained by using an inversion-recovery pulse sequenceprior to the quadrupole echo sequence. Tl values were measured from the null points in the inversion-recovery spectra. The 90"pulse width was 3.2 ps and data were obtained in quadrature by using 2K points per channel. For most experiments the digitization rate was 100 ns/point (10 MHz). Results General Description of N M R Spectra. Figure 2 shows a typical quadrupole echo solid-state deuterium NMR spectrum of the exchanged epoxy resin EX4 when it contains 2 w t % DzO (a) and when this same sample is dried (b). We have previously shown that the sharp central line arises from the sorbed DzO, at this peak disappears when the sample is heated or dried.13 The outer, broad part of the NMR spectrum is attributed to OH residues that have undergone exchange with DzO and have become OD groups on the polymer backbone. The quadrupolar splitting between the outer peaks (125 kHz) indicates that the OD groups are not undergoing any large-amplitude motions on the NMR time scale. The line width of the central sharp peak is somewhat broader than that of free water. This broadening may be attributed to contributions from sample inhomogeneities. In addition, the line width decreases as the weight percent of DzO increases. Because sample EX4 contains OD residues on the backbone (i.e., OH groups that have exchanged with DzO) which produce a static-like pattern, such a broad pattern would obscure the signal from a small amount of tightly bound water, if it were present. Two experiments were designed to circumvent this problem. In the first, sample NX2 (which contains very few OH groups) was exchanged with D20. It absorbed less than 0.5 wt % DzO, and this - 200 - 100 0 IO0 200 kHz Figure 2. Quadrupole echo deuterium NMR spectra of epoxy resin EX4 (a) containing 2 w t % D20; and (b) the EX4 sample in a after it has been dried. The spectra were obtained at 20 "C and 55.26 MHz for deuterium. Figure 3. Inversion-recovery quadrupole echo NMR spectra for epoxy resin EX4 containing 2 w t % D,O. From top to bottom, the inversion-recovery delay times are 5000, 500, 200, 100, 50, 10,1,0.5,and 0.1 ms. The enlargement on the right shows data obtained at inversion-recovery delay times of 9 (top), 7, and 3 ms. AU data were obtained at 20 O C and 55.26 MHz for deuterium. level of deuterium was not detected in the NMR experiment. A model polymer, poly(N-vinylpyrrolidone), was used in the second experiment. Poly(N-vinylpyrrolidone) is a hydrophilic polymer that does not contain exchangeable OH groups. The deuterium NMR spectrum of a sample of poly(N-vinylpyrrolidone)containing 21 wt % DzO shows no evidence for bound or static-like water. At this content of deuterium, a broad component arising from bound water would be clearly observed. Relaxation Data. Typical inversion-recovery relaxation data for an EX4 resin containing 2 wt % D20 are shown in Figure 3. It is clear that the OD groups and the signal due to DzO relax at significantly different rates. Relaxation data for EX4 (both "wet" and dried) at various temperatures are reported in Table I and are zyxwvutsrqpo zyxwvutsrq zy Nature of Water-Epoxy Interaction 1093 Macromolecules, Vol. 18, No. 6, 1985 Table I Deuterium NMR Relaxation Data” for Epoxy-D20 and Related Systems zyxwvutsrqponmlk zyxwvutsrq measurement temp, sample EX4-backbone OD dry EX4 wet EX4b DZO *g - measurement “C temp, O C T I ,ms 20 60 20 65 139 99 116 71 203 44 63 36 in EX4 in poly(vinylpyrro1idone) neat plus Fe(AC)z 20 20 20 20 12c 12 362 362 in EX4d 20 20 73 306 MezSO-d6 neat zyxwvutsrq “All data were obtained at 55.26 MHz for *H.*Sample contained 2% DzO by weight. This value obtained by using a quadrupole-echo inversion-recovery pulse sequence. TI of 12 ms also obtained by using standard high-resolutioninversion-recovery sequence. d Sample contained 5% MezSO-dsby weight. compared to the relaxation times of pure and paramagnetically doped DzO. Several points are apparent from inspection of the data in Table I. First, the Tl of the mobile water in sample EX4 is substantially different from that of pure DzO. It is unlikely that this difference is due to small amounts of paramagnetic impurities, as paramagnetic doping of D20 (Table I) does not change its relaxation time. Second, the relaxation times for the static component of the wet and dried EX4 samples are significantly different when measured at 20 OC. The glass transition temperatures for the wet and dried EX4 sample are different, which could account for the differences in the T1 of the static component. The relaxation times for the static-like component of these two samples become similar when measured at a constant AT below Tg(Le., at 20 OC for the wet EX4 and at 60 “C for the dried sample). Figure 4 shows a plot of the T1data for the backbone OD group. Data for both the dry (triangles) and wet (circles) EX4 samples are shown. The abscissa corresponds to the difference in temperature between Tgand the measurement temperature. Data from both the wet and dry samples fall on this line, suggesting that the T1differences between the wet and dry epoxy samples (Table I) are due primarily to plasticization effects. 60 I I I 1 00 100 120 140 p - ( measurement temperature 1 Figure 4. Plot of TIdata for the OD backbone group in epoxy sample EX4 vs. the difference between T and the temperature of measurement. The triangles represent t i e dry samples; circles represent the sample containing 2 wt % D,O. by a single exponential, consistent with there being only one species contributing to this signal. Finally, to eliminate the possibility that the signal from free water is suppressed by the quadrupole echo NMR experiment, the T1of this sharp component was measured by using a standard inversion-recovery Bloch decay experiment. This experiment also produces a 12-ms T1for the sharp component (Table I). Calorimetric studies down to -80 OC on differently prepared samples have also failed to reveal thermal phenomena associated with free water. Taken together, the data from the low-temperatureexperiments, the relaxation data, and the line width data support the conclusion that there is no isotropically free water in this epoxy-water system. This obviates the presence of the water in macroscopic voids, a situation which has been seen to occur in a different epoxy systema8Instead, the water appears to be distributed homogeneously throughout the system. These data also show that the motion of the water is impeded. When the standard relationship between correlation time and relaxation times is used,21the 12-ms T1 translates into a correlation time of 7 X s. (The correlation time of free water is 2 X lo-” s at this temperature.) This correlation time can be interpreted as the residence time of the water as it hops from site to site. It is likely that the OD or residual ND groups act as the impeding sites. Question of Tightly Bound Water. Although it is clear that essentially all of the water is very mobile, we are unable to directly eliminate the possibility that there is a small amount of bound water. This situation arises because the signals from the OD of the epoxy resin backbone would obscure any static water, if it were present. However, the following experiments and considerations indicate that it is highly unlikely that there is a significant amount of bound water. (We define “bound” water as water with a residence time of s or longer, or water whose only motion consists of a flip about the molecular symmetry axis.) Firstly, the depression of Tgin the EX4 system by water exhibits a monotonic decrease from 0% moisture content: If tightly bound water does not exercise the same plasti- zyxwvut Discussion Question of “Free”Water. Several lines of evidence argue against the presence of “free” water. (We define “free” to mean water that is isotropically mobile and thus has the same relaxation time as pure water.) First, in our previous work13we showed that the water in resin EX4 did not freeze at temperatures down to -20 “C. This experiment is not unequivocal, however, because the signal for the water resonance did broaden, and a small amount (less than 10%) of free water could have been obscured by the OD signal from the polymer backbone. In contrast to the above experiment, relaxation data provide a clear indication that there is no free water in this system. The spin-lattice relaxation time of pure D20 under our conditions is 362 ms, whereas the corresponding relaxation time of the sharp signal in the spectrum of EX4 is 12 ms (Table I). This finding provides strong evidence that the motion of the water molecules in the epoxy resin is impeded. In addition, the relaxation data are described zyxwvutsrqponm zyxwvutsr 1094 Jelinski et al. cizing action as normal “dissolved” water, then some discontinuity may be anticipated. However, evaluation of the depression of Tgis obviously determined at the respective temperature of the transition and not at ambient conditions. We know from our experiments here that Tl times of water molecules in the epoxy matrix vary significantly with temperature and hence the water molecules have different mobilities at different temperatures. This in fact forms the very basis of the great utility of the deuterium NMR technique used here since molecular interactions are determined at the temperature in question and are not inferred from property measurements at a temperature far removed from the point of interest. Further evidence against the presence of bound water a t ambient conditions comes from the fact that the composite line shape of Figure 2a can be simulated satisfactorily by assuming that there are only two components-a static pattern and a Lorentzian line.I3 Thirdly, evidence comes from experiments on a model polymer, poly(N-vinylpyrrolidone). This material is known to be exceptionally hydrophilic, yet it contains no exchangeable NH or OH groups. When plasticized with DzO, the deuterium NMR spectrum of this polymer shows only a single line and no evidence of bound water. A final line of reasoning comes from examination of the crystal structures of inorganic molecules that contain water of crystallization. Gypsum, or CaS04.2H20,provides a relevant example. Although the water in CaS04.2Hz0 forms part of the structure in the solid state, NMR experiments show that it is actually undergoing two-site flips about the c p axis.22 The NMR pattern for water that is involved in two-site flips consists of a tentlike pattern that is approximately 120 kHz in breadth. This line shape would not be obscured by the signals from the OD groups, and hence we can rule out the presence of bound, but flipping, water molecules. These experiments and lines of reasoning suggest that it is highly unlikely that the epoxy-water system contains tightly bound water. Question of D 2 0 Disruption to the HydrogenBonding Network. In this section we obtain an upper limit for the OD/DzO exchange rate and show that the water molecules are hopping from site to site at least 6 orders of magnitude faster than this. The two-component nature of the spectrum of EX4 plasticized with water (Figure 2a) indicates that the OD/DzO exchange rate is slow compared to the frequency separation between the OD signal and the D20 signal. If the exchange rate were faster than this separation in frequency units (ca. 100 kHz), we would not observe these two components. Instead, the signals would be averaged together, producing only a sharp signal in the center of the spectrum. This observation sets an upper limit on the exchange rate of lo5 s-1.23 This upper limit can be lowered to ca. lo3 if we can rule out the possibility of intermediate exchange. The T, data indicate that the OD and DzO are not exchanging a t an intermediate rate on the NMR time scale. If intermediate exchange were the case, we would expect that the hopping water molecules could communicate their efficient relaxation to the OD groups. We would predict that the relaxation time of the OD groups would be shorter in the EX4 sample that contains water than in the one that is dry. T I measurements performed a t fixed temperatures below the respective Tgof the dry and wet samples (Table I) show that the static components have identical relaxation times. This finding provides strong evidence that the OD/D20 exchange rate is much slower than 100 kHz, al- zyxwv zyxw Macromolecules, Vol. 18, No. 6, 1985 lowing us to rule out the possibility of exchange at an intermediate rate. We therefore have an upper limit for chemical exchange of lo3 s-,, and a hop rate of the water molecules of lo9 s-,. Because the water molecules hop from site to site at least 6 orders of magnitude faster than they undergo chemical exchange, it is unlikely that the water molecules serve to disrupt the hydrogen-bonding network in the epoxy resin. These conclusions are supported by water relaxation measurements on poly(N-vinylpyrrolidone). The T1 of water is the same in this system as it is in the epoxy-water system (Table I). As poly(N-vinylpyrrolidone) does not have the possibility of forming a hydrogen-bonded network, the similarity in relaxation times suggests that the water is acting merely as a plasticizer in both of these systems. Summary The results presented here provide new insight into the molecular details of the interaction of epoxy resins with water. In particular, these results establish (1) that the water is impeded in its movement, with the molecules hopping from site to site with an approximate residence time per site of 7 X 10-los, (2) that there is no free water, (3) that there is no evidence for tightly bound water, and (4) that it is unlikely that the water disrupts the hydrogen-bond network in the epoxy resin. These results are consistent with the notion that water is simply a plasticizer for these systems, just as it is for polymers that contain no exchangeable protons. However, these results clearly demonstrate that the water does interact with specific sites or traps on the epoxy backbone, insomuch as the water molecules hop from site to site, rather than behaving as a gas in an inert matrix. However, these results do not connect the hopping water molecule with specific hydrogen-exchange mechanisms. zy zyxwvutsrqp Acknowledgment. T.S.E. and F.E.K. acknowledge support for part of this work from AFOSR 84-0100 and CUMIRP. Registry No. Epon 825, 25068-38-6; water, 7732-18-5. References and Notes Sung, Y. K.; Gregonis, D. E.; Russell, G. A.; Andrade, J. D. Polymer 1978, 19, 1362. Tan, Y. Y.; Challa, G. Polymer 1976, 17, 739. Jin, X.; Ellis, T. S.; Karasz, F. E. J. Polym. Sei., Polym. Phys. Ed., in press. Ellis, T. S.; Karasz, F. E. Polymer 1984, 25, 664. McKague, E. L.; Reynolds, J. D.; Halkias, J. E. J . Appl. Polym. Sei. 1978, 22, 1643. Browning, C. E. Polym. Eng. Sei. 1978, 18, 16. Kong, E. S. W.; Adamson, M. J. Polymer 1983, 24, 171. Moy, P.; Karasz, F. E. Polym. Eng. Sci. 1980, 20, 315. Maxwell, I. D.; Pethrick, R. A. J. Appl. Polym. Sci. 1983,28, 2363. Kong, E. S. W. Proc. Org. Coat. Appl. Polym. Sci. 1983, 48, 727. Banks, L.; Ellis, B. Polym. Bull. 1979, I , 377. ten Brinke, G.; Karasz, F. E.; Ellis, T. S. Macromolecules 1983, 16, 244. Ellis, T. S.; ten Brinke, G.; Karasz, F. E. J . Appl. Polym. Sei. 1983, 28, 23. Ellis, T. S.; Karasz, F. E. Proc. Org. Coat. Appl. Polym. Sei. 1983, 48, 721. Jelinski, L. W.; Dumais, J. J.; Stark, R. E.; Ellis, T. S.; Karasz, F. E. Macromolecules 1983, 16, 1019. A preliminary account of some of these experiments has been published in preprint form: Jelinski, L. W.; Dumais, J. J.; Cholli, A. L. Polym. Prepr. (Am. Chem. Soc., Diu.Poly. Chem.) 1984,25, 348. Antoon, M. K.; Koenig, J. Lo; Serafini, T. J.Polym. Sci., Polym. Phys. Ed. 1981,19, 1667. Young, J. F. J . Appl. Chem. (London) 1967, 17, 241. Jelinski, L. W.; Dumais, J. J.; Engel, A. K. Macromolecules 1983, 16, 492. zyxwvutsrqpo zyxwvut zyxw zyxwvuts zyxwv 1095 Macromolecules 1985, 18, 1095-1100 (18) Cholli, A. L.; Dumais, J. J.; Engel, A. K.; Jelinski, L. W. (22) Spiess, H. W.; Mahnke, H. 2.Naturforsch 1972, 27A, 1536. Macromolecules 1984, 17, 2399. ~ ~ *discussed ~~ the situation of magnetization (23) E i s e n ~ t a d thas transfer between spin nuclei in the presence and absence (19) Jeffrey, K. R.; Bloom, M.; Valic, M. I.; Higgs, T. P. Chem. of exchange. Such a treatment would be premature in the Phys. Lett. 1976, 42, 390. V. Chem. Phvs. Dresent case. as we observe sinele-exDonentialrelaxation within (201 Blinc. R.: Rutar. V.: Selieer. J.: Slak. J.: Smolii. zyxwvutsrqpo Lett.’1977, 48, 576: the uncertainty of the data. (21) Mehrine. M. “Hieh Resolution NMR in Solids”:. SDrinaer(24) Eisenstadt. M. J. Mam. Reson. 1980. 38. 507. Ver1ag:-New Yo&, 1983. (25) Eisenstadt; M. J. Main. Reson. 1980; 39;263. , I I - I . I Exposure of Hydroxyl Groups in Phenol-Acetaldehyde Oligomers, As Investigated by Photo-CIDNP lH NMR and Infrared Spectroscopy zyxwvutsr zyxwv Lucia Zetta,*2aAntonio De Marco,2aGiovanni Casiraghi,2bMara and Robert KapteinZc Zstituto di Chimica delle Macromolecole del C.N.R., 1-20133 Milano, Italy, Zstituto di Chimica Organica dell’Uniuersitd, 43100 Parma, Italy, and Physical Chemistry Laboratory, University of Groningen, 9747 AG-Groningen, The Netherlands. Received J u n e 18, 1984 ABSTRACT: A number of oligomers related to phenol-acetaldehyde novolac resins possessing different stereoregularity are investigated by conventional and photochemically induced dynamic nuclear polarization (photo-CIDNP) lH NMR and IR spectroscopy. Complete assignment of NMR spectra is achieved by double-resonance techniques, spin-echo multiplet selection, use of shift reagents, and comparison with spectra of two monomeric model compounds. For each oligomer, the spectral parameters can be directly related to the tacticityof the molecule (Le., to the sequence of meso and racemic units). The photo-CIDNP technique provides information on the accessibility of the phenolic moieties. In both monomers access to the OH’S is hindered. From the dimer to the pentamer, a gradual reduction of the exposure of the OH groups from the e x t e n d rings is observed, whereas those from the inner rings are always accessible, irrespective of the number of units. This indicates the absence of the isodromic intramolecular H bonds which were observed in the X-ray structure. In the IR experiments the OH. .OH absorption band shifts to lower frequencies and broadens by increasing the molecular weight. Combined with the photo-CIDNP results, this suggests a progressive strengthening of the H bonding of the outer functional groups. Two empirical rules are proposed, which correlate the intensity of the aromatic signals with the number of internal phenolic units. Introduction All-ortho alkylidene-bridged oligophenols (novolacs) have been the subject of several investigation^.^-^ IR studies have shown that in carbon tetrachloride these molecules adopt well-defined conformations, involving OH---OH intra- and intermolecular hydrogen bonds.”’ The X-ray structure analysis of a number of oligomers indicates that the conformation in the crystal is mainly determined by an isodromic, intramolecular H-bonding system, formed by the phenolic functional groups.6 The H-bond stabilization defines two domains, one highly polar, containing all OH groups, the other strongly hydrophobic, in which the aromatic rings form a lipophilic matrix. The present communication is focused on the aromatic resonances of a number of oligophenols (number of internal rings 0 < n < 3) having different steric regularity. The accessibility of the phenolic groups is investigated by photochemically induced dynamic nuclear polarization (photo-CIDNP) and IR spectroscopy, in order to obtain some evidence on the molecular structure, in particular on the stabilizing role of the hydrogen bonds. Materials and Methods The two monomers and the oligomers in this study were obtained as described elsewherea6 The methyl ester of 3-N-(carboxymethyl)lumiflavinwas the kind gift of Dr. F. Muller (Wageningen). C2HC13was from Merck Isotopes. The conventional NMR spectra were recorded with a Bruker HX-270 spectrometer, controlled by an Aspect 2000 computer. Chemical shifts are quoted in ppm from internal tetramethylsilane (Me,Si). Typically 64 scans were 0024-9297/85/2218-1095$01.50/0 zyxwvu zyxw accumulated for each spectrum. The spin-echo experimenta shown in Figure 6 were performed according to the pulse sequence (90°-tl-1800-tl-f900-t2-900-acquisition), where tl is 1 / 2 J and t 2 allows for partial relaxation.8 Photo-CIDNP spectra were obtained with a Bruker HX-360 spectrometer, also equipped with an Aspect 2000 computer. The sample concentrations ranged between 6 and 1 2 mM in C2HC13. Difference spectra were obtained by f sking “light” and “dark” free induction decays (see text) and subtracting the spectra after Fourier transformation. A Spectra Physics Model 171 argon ion laser was employed as the light source. A 0.6-s light pulse (4 W, multiline) was used, with a 0.05s delay before the 90° rf pulse. Four scans were accumulated for each spectrum. All spectra were resolution enhanced via the Gaussian convol~tion.~ The line at 7.23 ppm in the photo-CIDNP difference spectra results from incomplete cancellation of the CHC13 signal. IR spectra were obtained with a Perkin-Elmer 457 spectrometer, the sample concentrations varying between 0.2 and 2.0 mM in C2HC13. The cell paths ranged between 75 and 500 p. Results and Discussion Resonance Assignment. The phenol-aldehyde oligomers of this study are listed in Chart I. The monomers I and I’ are models for inner and outer rings in the oligomers, respectively. Figure 1 shows the aromatic region of the lH NMR spectra of I and 1’, after resolution enhancement. The axial symmetry of compound I prevents the observation of J couplings across more than three 0 1985 American Chemical Society