Experiments in Nuclear Magnetic Resonance Citation Bromberg, J. Philip (1959) Experiments in Nuclear Magnetic Resonance. Master's thesis, California Institute of Technology. doi:10.7907/F20V-N746. https://resolver.caltech.edu/CaltechTHESIS:11242009-075328528 Abstract I. Proton and cobalt resonances were observed in single crystals of Co(NH_2CH_2CH_2NH_2)_3Cl_3 ^3H_2_O. The cobalt spectrum consists of a maximum of seven lines whose separation follows a (3cos^2θ – 1) dependence quite closely where θ is the angle between the c-axis of the crystal and H_o. The spectrum is fully explained on the basis of quadrupole interaction. Two sets of doublets were found for the proton resonances; an inner, very narrow set following a dependence similar to the cobalt resonance, and a pair of outer broad peaks which overlap the sharp peaks. The narrowness of the central peaks is attributed to internal rotation of one of the proton bearing groups. II. A summary is given of various high resolution spectra of compounds related to phosphine. Item Type: Thesis (Master's thesis) Subject Keywords: (Chemistry) Degree Grantor: California Institute of Technology Division: Chemistry and Chemical Engineering Major Option: Chemistry Thesis Availability: Public (worldwide access) Research Advisor(s): McConnell, Harden M. Thesis Committee: Unknown, Unknown Defense Date: 1 January 1959 Funders: Funding Agency Grant Number Sloan Foundation fellowship (1958) UNSPECIFIED Record Number: CaltechTHESIS:11242009-075328528 Persistent URL: https://resolver.caltech.edu/CaltechTHESIS:11242009-075328528 DOI: 10.7907/F20V-N746 Default Usage Policy: No commercial reproduction, distribution, display or performance rights in this work are provided. ID Code: 5396 Collection: CaltechTHESIS Deposited By: Tony Diaz Deposited On: 01 Dec 2009 18:32 Last Modified: 18 Oct 2023 16:54 Thesis Files Preview PDF - Final Version See Usage Policy. 1MB Repository Staff Only: item control page

l The | EXPERIMENTS IN NUCLEAR MAGNETIC RESONANCE Thesis by J. Philip Bromberg In Partial Fulfillment of the Requirements For the Degree of Master of Science California Institute of Technology Pasadena, California 1959 | The ] ACKNOWLEDGMENTS The author would like to express his sincere thanks to Professor Harden M. McConnell for suggesting the problem and for his advice during the course of the work. He is also indebted to the various members of the magnetic reson- ance group for their helpful discussions and encouragements. For financlal assistance the author is indebted to the institute for their aid in the form of scholarships and teaching assistantships, and to the Sloan Foundation for a fellowship during the summer of 1958. | Th a ABSTRACT I. Proton and cobalt resonances were observed in single | crystals of Co(NHCH CH NH.) C1, °3H,0. The cobalt spectrum consists of a maximum of seven lines whose separation fol- 25. 1) dependence quite closely where 9 is lows a (3cos the angle between the c-axis of the crystal and Ho: The spectrum is fully explained on the basis of quadrupole interaction. Two sets of doublets were found for the pro- ton resonances; an inner, very narrow set following a de- pendence similar to the cobalt resonance, and a pair of outer broad peaks which overlap the sharp peaks. The nar- rowness of the central peaks is attributed to internal rotation of one of the proton bearing groups. II. A summary is given of various high resolution spectra of compounds related to phosphine. The TABLE OF CONTENTS | PART TITLE PAGE Section I I. Introduction and Previous Work . . . . . + 1 it. Compounds and Crystal Growth . . 6. 1. 2. 2 Tir. Spectra and Spectral Measurements--Data . 5 Iv. Theory, Discussion, and Conclusions ... 17 . Section ITI Vv. High Resolution NMR Spectra of Varicus Phosphorus Compounds . . . + 6 6 6 © © «© 6 23 REFERENCES e ee «© 8& &© &# *® © # # © #© @© © # © @# # @ +6 ey -l- I. ANISOTROPIC EFFECTS IN THE NMR SPECTRA OF SINGLE CRYSTALS OF CORALTIC COMPLEXES INTRODUCTION AND PREVIOUS WORK It was felt that a study of the anisotropic effects in the nuclear magnetic resonance spectra of single crystals of trivalent cobalt complexes might give some insight as to the nature of the crystalline fields existing within the crystals. Hence observations were undertaken on the proton and cobalt resonances in a number of various cobaltic complexes, and in particular Co(NH,CH,CH,NH,),C1,+3HQ0.— The complexes of tri- valent cobalt, containing no unshared electrons, are all diamagnetic thus eliminating any complications due to para- magnetic effects. Previous work has been done on cobaltic complexes, though in the powdered, dehydrated state, by J. 8. Waugh (1) and H, L. Richter (2). For powdered Co(NH, )¢C1., at room tempera~ ture Waugh found a rather narrow proton peak about 3.4 gauss wide and concluded that the Co(NH, )g*9 ion underwent rapid rotation in the lattice. For Co(en) C1, he found a broad line and concluded that there was no rotation of the Co(en),*° ion in the lattice. Richter corroborated these results for the case of Co(NH3)<C1.3 indeed he showed that the group rotated at temperatures down to ~-80 deg.C. For the case of * The ethylenediamine groups, NHoCH,CHONH,, will henceforth be abbreviated to read "en." ~2- the ethylenediamine, however, he could get no consistent results. His first measurements indicated a narrow peak of about 1.3 gauss in width at room temperature which did not begin to broaden until the temperature was reduced to about -60 deg.C. Waugh later obtained a narrow line superimposed upon a broad line with the NMR Spectrometer at MIT (3), how- ever, no explanation has been advanced. These two previous studies covered only the proton resonances, and made no at- tempt to work with the cobalt resonance as was done here, COMPOUNDS AND CRYSTAL GROWTH A number of compounds have been prepared during the course of this work following procedures in the standard reference works (4). These include: Co(NH, ) 6X, (x = Cl, I, NO3, C10, ) (5); racemized Co(en)3C1.+3H,0 and Co(en).Br.: 3H,0 (5); dextro- and levo Co(en)Br.+H,0 prepared from the racemized chloride following the method of Werner (7); and Co((NH3),C1)Cl, (8). The crystal forms included in this list are: cubic (the hexamine fodide and perchlorate); tri- gonal (the racemized ethylenediamine chloride and bromide); tetragonal (the dextro- and levo-~ ethylenediamine bromide, and the hexdmnine nitrate); monoclinic (Co(NH,)<C1,)3 and orthorhombic (Co( (NH, )-C1)C1,) (9). In view of the fact that the growth of larse crystals is a selfpurifying process it was not felt necessary to purify the prepared compounds peyond the levels stipulated in the references mentioned, ~3- To verify the complete resolution of the Co(en) .Brz*H,0 into 1lts respective optically active isomers, the specific rotation was checked against the literature value. The ex- perimental value of 116.5° was in excellent agreement with Werner's value of 115° for the dextro form (10). The levo- form was not checked. Crystals were grown by Slowly lowering the temperature of a seeded aqueous solution saturated initlally at about 15 to 20 degrees above ambient temperature (that is about 4C deg.C. in the laboratory and about 25 des.C. in the 4 der.c. cold room). The seeds, of the order of a millimeter in size, were prepared by saturating a solution at about 15 degrees above room temperature, and allowing the solution to cocl to room temperature over night. The solution was then poured off, and the best seeds extracted with a pair of forceps. To grow the large single crystals, the seeds were suspended in the saturated solutions by gluing them to a length of 4mm pyrex tubing drawn out to a capillary point with duco cement. The tubes were then joined to the glass stoppers of the flasks with shellac. To provide a slow but constant rate of cooling, a “con- stanctly variable” constant temperature bath was constructed. The heart of the bath lay ina large mercury tnermoresulator with an adjustable upper tungsten contact. By means of a serew mechanism the upper electrode was lowered by a clock motor to give a vate of cooling of about one degree centi- The ae grade per day. The erystals were removed and dried with fil- ver paper as soon as they had reached the required size. No noteworthy success has been achieved with this method except for the cases of Co(en).C1,-3H,0 and d-Co(en).Br.+H,0. With these two compounds large erystals of the order of a centimeter in dimensions with well developed faces have been prepared in a period of about a week. It is with the former of these that we shall be primarily concerned. Some reason- able success was obtained with the Co(NH3)¢C1., however, the erystals were not nearly as large as the ethylenediamines nor were the faces as well developed. The Co(en).C1,+3H,0 Porms nexagonal prisms, the hexagonal cross section being somewhat ieresular; the d-Co(en) Br. -H,0 forms a tetragonal prism canped cn either end by tetragonal pyramids composed of (111) planes. No astempt was made to grow erystals of the levo- isomer. ay n all other cases the crystals failed to srow pronerly or even to cap. A possible source of difficulty is slight de- composition of the compounds, in which case a lowered cperat- ims temperature should help greatly. With this in mind the temperature bath was transferred to a 4 deg.c. cold room, Resuits were not much better under these conditions; however, it is felt that time did not permit a proper sustained e?- fort along these lines, and it is quite possible that further avcempts mignt well be successful. Another factor to take into consideration is stirrings. It might well be that stir- ring of the seed erystal might give greatly enhanced results, however, it is also possible that the results might be worse. SPECTRA AND SPECTRAL MEASUREMENTS-~--DATA NMR spectra of both hydrogen and cobalt nuclei have been observed in Co(en) 301, +3H,0. In both cases there is a marked anisotropy. To align the crystal at various angular orienta~- tions a simole goniometer was constructed consisting of a full circle protractor marked off at intervals of ten degrees mounted in the stopper of a standard taper test tube which served as the sample holder, The crystal was alisned by eye with the zero line on the protractor; a length of masking tape stretched across the pole faces of the magnet served as a reference for alignment at various angles, The proton spectra of figure 1 were taken at 15 mega- cycles corresponding to a field of 3760 gauss, while those of ticures 2 and 3 and table I were taken at €.4 megaecyeles cor- resoonding to a field of 1500 gauss, The field scan rate was rather low and assumed linear, The scan rate was calibrated by measuring the chemical shift between m1 203 and 71205 in challous acetate at the same magnetic field value and scan rate at the corresponding frequency for the thallium nucleus, The proton spectra gives rise to two sets of lines vis- idle in flgure 1. The form of each set is dependent upon anzular orientation. The central, sharper peak is in reality a doublet as is clearly obvious from figure 2; under the conditions of modulations for figure 1 the two lines remain Th -7- unresolved. The inner lines have a width of the orders of 0.4 gauss while the outer, much broader peaks are approxi- mately 7 gauss wide. Any attempt to measure the character- istics of the broad outer peaks would be vitiated by the overwhelming influence of the central peaks. Two points may be brought out with regard to these outer peaks. Firstly, the seometry of the derivative curve in figure 1 seems to indicate that it is a doublet; and secondly, spectra taken at other angular orientations clearly indicate that the peaks are devendent upon orientation. At narrower field modulation amplitudes the outer peaks are not observed while the inner peaks form a clearly resolved doublet whose separation is greatly dependent upon orientation, The geometry of the crystal was such that the c-axis and Ho always formed a plane parallel to the floor of the laboratory. To change the angular orientation of the crystal it was ro- taved atout an axis in space nerpendicular to the plane formed by the c-axis and Ho. For the spectra in ficure 2 the axis of rotation coincided with an a-axis of the erystal, now~ ever, other experiments indicated that the orientation of the a-axis with respect to H, is completely irrelevant, the re- sults being unchanced if the erystal is rotated about the e~axis,. The two lines in the doublet have a separation rang- ing from a maximum of 1,7 gauss at zere decrees when the e~axis is parallel to Ho: to Zero gauss at 50-69 decrees where the lines are coincident. Ficure 2 contains traces of Th actual resonance neaks from zero to 180 decrees, Fissure 3 shows a plot of the desendence of sevaration upon anzle. If ——+ the separation follows the theoretical (3c08"e - 1) devend- enee the separation should theoretically so to zero at 55 desrees as indicated by the dotted line. The values of the separations are tabulated in table I. The central column in 2 the table lists the absolute values of (3cos“9 - 1) multiplied by a constant factor of 0.845, 3 A determination of the proton splittine at 15 merna- p ° evyeles yielded a value of 1.65 gauss at zero decrees which f ihe nn ~29 fauss determined TABLE I Separation of the two lines in the proton resonance spectrum of Co{en).C1,+3H,0 at various orientations of the c-axis rela- tive to H). Angle AH(gauss) 0.845(3c0s"0 ~ 1) Angle 4H(gauss) e) 1.67 1,69 180 1,89 1 1.61 1.E1 190 1.63 20 1.40 1,40 200 1.42 30 1,0€ 1.06 210 1.08 AG 0,03 0.64 220 0.65 50 0,00 0.20 230 0,00 &0 G.00 C,2l 240 G.00 70 0.51 0.55 250 0,52 do 0.77 0.77 260 0.78 30 0.86 O85 276 0.34 100 0.75 o.77 280 0.75 110 0.53 0,55 290 0,52 126 c.00 0,21 300 0.00 136 0,00 0.20 320 9.90 140 6.66 0,64 320 0,65 150 1,10 1,06 330 1,08 LOO 1.38 1.40 340 1.41 70 1.62 1.61 350 1.62 180 1.69 1.59 360 1.69 ‘WOHDjUe (sO sojnBUD yo UgsoUNs D sD OTHE 4OHUMOD Jo jossKs0 ofbuys D Ut eBUjLNdS EOYOUOSE UOjOsG JO 10id :2°9/4 °H PUD SIxD-> usemyeq O18ey lJ ~10— y “ssnob ui ‘Burges for) 09"! — OOS | Th -11- The cobalt resonances were measured at 11.976 mecacycles, corresponding to a field of 11,854 gauss for the cobalt nu- — clea@s resonance. The spectrum covered a range of some 1300 gauss, and thus a rather wide scan was necessary. It was found that the scanning unit at this range was quite non- linear and a different technique was used to calibrate and measure the splittings from that used for the protons. In order to expedite matters a simple nomograph was constructed to eliminate the drudgery of carefully measuring each resoh- nance peak with a ruler and interpolating. Marks are placed on the chart paper every revolution of the scan dial at each dial division ending in --95. Utilizing an aqueous cobaltic solution the magnetic field scanning dial was advanced to successive points corresponding to the markers on the chart paper. At each of these points the frequency was changed until the resonance peak was centered on the oscilloscope, and the frequency measured with a Signal Corps Frequency Meter model BC-221-N. Taking the value of 10,11 megacycles for the resonant frequency of cobalt at 10,000 gauss, the magnetic field was then calculated at each marker point on the chart paper by multiplying the measured frequency bz (10,000/10.11). A large piece of graph paper was then marked off at intervals exactly equal to the intervals on the chart paper, and the calculated field plotted at each point; a smooth curve was then drawn through these points. To meas-~ ure the magnetic field at any resonance peak it was then only -~12- necessary to align the chart paper interval markings with the corresponding markings on the abscissa of the nomocraph and tape the chart to a T-square to permit rapid and easy measurements, The T-square was then moved un the nomograph until the center of the resonance peak coincided with the curve on the nomograph. It was then a simple matter to read the magnetic field at that point from the nomograph. With the c-axis of the Co(en).C1,+3H,0 crystal parallel to Hy the cobalt spectrum consists of seven lines spread over a distance of some 1260 gauss with the central peak at about 11,850 gauss. The six outer peaks are symmetrically disposed about the central peak with equal spacing between peaks of 210 gauss. The central peak is some eight gauss wide. From this maximum the spacing decreases to zero at 55 degsrees where only one peak is present. This is clearly visible in the series of traces in figures 4 and 5, The geometry of ‘she erystal orientation is the same as for the proton resonances described on page 7. Once again the position of the a-axis in no way effects the spectra. Figure 6 contains a plot of splittings as a function of angular orientation. The results are tabulated in table II, The numbers at the heads of the columns in table II refer to the various peaks, 4 being the central peak; 1, 2, and 3 the low field peaks in that order; and 5, 6, and 7 the high field peaks. The reference point in tne magnetic field has been shifted so that the zero point coincides with the central peak. Gaps in the table indicate of i 1 i 1 { -635 -422 207 2u 418 624 20° 1 J i 4 i t “520 ~341 -168 7) 343 503 40° L 4 L “82 O 77 FIG.4: Faces of a sequence of cobalt resonances in a single crystal of -Colen)s Cis: 3H20 et various angular orientations. The field scon is non-linear. The scales below the traces denote distances in gauss from the central peak, which is taken as zero, to the outer peaks. Ti ~lLhe 50° 60° 55° a a Sn Cee 36 ) 23 70° “30 O 28 90° i i i i i i i i i i -138 -68 [o) 67 129 “202 -02 0 107 221 FIG.5: Further traces from the sequence of Fig.4. Excegi for the 90 degree trace the field field scan rate is slower than that in fig. 4. I -15~ unresolved or absent peaks. The average separation value in the riext to the last column was arrived at by averagins the Yigures in columns 1 and 7, each divided oy three; figures in columns 2 and © each divided by two; and the figures in columns 3 and 5. The (3cos"e - 1) colum has again been indi- cated, this time multiplied by 105 to indicate the rather close agreement with the expected theoretical dependence. TABLE II Positions of the peaks in the cobalt resonance spectrum of Co(en),C1,+3H,0 as a function of angular orientation * (3c0s"9-1) Angle 1 2 3 4 5 6 7 Averace X105 9° 635 422 e207 oO 211 418 624 210 210 19 508 398 19% 90 200 399 596 198 200 20 520 341 168 Oo 171 343 503 171 173 30 260 125 O 127 256 128 13 ho 82 «OO 77 86 30 50 26 40 23 25 25 55 0 0 re) 60 30 O 2s 29 26 70 138 63 #O C7 129 63 68 30 296 191 93 #0 97 203 °7 96 90 202 102 0 107 221 105 105 * See text An earlier cobalt measurement at 6.4 megacycles yielded a value of approximately 1350 gauss for the total spread at zerc degrees, Tnis value of 1350 was arrived at by a very rough approximation from the ammeter readings on the magnet power supply and is taken as an indication that the splittings in une cobalt resonance are not changed at different frequencies or observation. ‘HE ADf(uqoD 40 1DsksoepBuys BU) SoUDUOSEs yJOQOD Oy, 4O SHULINIES JO 101d 9 ‘Old “OH PUD S}1xD-2 UESMjeq euy 06 7 + 09 OS OY o¢e x4 ol i") ia + a~li= 7 ‘ssno6 ul Guragids 0 Zt — | li -17- Proton resonances were also observed in Co(NH, ) <C1., and dextro-Co(en)Br.+H,0. The spectrum for the hexdmine chloride consists of one single line of the order of two sauss wide. The dextro-ethylenediamine bromide is somewhat anomolous, As in the racemized chloride two types of lines are visible, @ narrow central line superimposed upon a much broader line. However, the narrow central veak is rapidly saturated, and it is impossible to zet a complete line on the chart paper even with the gain of the radio-frequency current turned to its minimum level, The broader peak 1s easily visible, and is arproximately 5 sauss wide. THEORY, DISCUSSION, AND CONCLUSIONS A. Proton resonances in Co(en).C1.+3H,0: In the course of his study of the proton resonance of the waters of hydra~ tion in cypsum single erystals Pake showed (11) that for systems of two identical nuclei one should Pind ¢wo resonance peaks at magnetic field values given by = f 5 3 3 a73 2 we | H = H, + —j- ur-’(3cos*6 ~ 1) (2) where He is the magnetic resonance field strength, u is the nuclear magnetic moment, and 9 the angle between the vector joining the two identical nuclei and Ho Disregarding possible structural differences there are three chemically distinct proton species in the molecule, che CH and NH protons on the ethylenediamine grouns, and | -13- the protons in the waters of crystallization. The spectrum indicates two distinct types of proton lines, It 1s not pos- sible to make any cefinite conclusions at this staze of the research, however, we might indicate certain possibilities, J. ter Berg (12) has studied the structure of Co(en) 301. *3H,0 and has determined the space Group, unit cell, and positions of the cobalt atoms. The crystal belongs to the trigonal system and has space group P3cl with four mole- cules per unit cell. The four cobalt atoms lie at four-fold equivalent lattice positions given by: 1/3,2/3,23 1/3,2/3.(% + 2)s 2/3,1/3,25 and 2/3,1/3,(% - z). In other words, the cobalt atoms lie along two chains each parallel to the c-axis, and situated one third and two thirds the distance along the vector (a+b) respectively. The chains consist of alternate dextro- and levo- groups, and are com- pletely out of phase with each other. Unfortunately this structural determination did not include the positioning of the ethylenediamine groups nor the hydrated waters, The unusual narrowness of the central proton peaks is, no doubt, due to internal rotation of one tyre of proton, and could be easily verified by examining the spectra at fo different temperatures. Since Waugh's work (13) indicated no internal rotation of the chelating groups, it is perhaps most losical to assume that these narrow peaks are due to the water molecules which are rotating. Insofar as this portion of the research is concerned the next logical step -19- is to grow erystals from the anhydrous sait in heavy water. ne HO molecules would all be replaced by DAO, and it seems nighly likely shat che protons on Nip groups would be re- placed by deuterium atoms, the CHy protons remaining un- changed, The proton resonance would then be due solely to one CHo protons and an unambiguous assignment of the reson-~ ance ceaks could be made. ne possibility that the central sharp peaks is due to rovasing ethylenediamine groups is not, however, to be com- pletely negated. Richter's work {14) indicated a great un- certainty as to the validity of Waugh's conclusions regarding the rigidity of the chelating groups. Furthermore, the situ- ation existing in the anhydrous salt might not be duplicated feta * » “ . 3 2 4, « 455 Latte Sm pl oe Y . + n the hydrated single crystal. Figure 7 indicates a Co(en). 3 Pic. 7: co(en),*° fon showing 3-fold rotation axis. Heavy lines are the en bridsing frouns -20~ ion group as it is usually pictured in its simple octahedral structure, It has a three-fold rotation axis as indicated in the diagram, and this could well be an axis of rigid ro- tation. It would still not be clear whether the sharp lines are due to the CH or NH protons, but this too could be determined by crystallization from heavy water, B. Cobalt resonances in Co(en).C1,°3H,0: The seven lines in the cobalt spectrum are easily explainable on the basis of quadrupole splitting. Co°” has a spin of 7/2, thus a quadrupole moment for the cobalt nucleus is expected. For an isolated nucleus of magnetic moment u, spin I, in a mag- netic field Hy along the g-axis, the magnetic energy levels are given by En = -muH/I (2) where m is the magnetic quantum number. In perfectly sym- metrical environments this gives rise to 21 + 1 equally spaced levels, and hence only one resonance peak. For nuclei or spin greater than % if an electric quadrupole moment exists, the nucleus interacts with the electric environment if it has less than cubic symmetry. Using first order per- turbation theory Pound (15) has shown that the energy levels now become NUH By 2 ee + aiSey (am? - 10 1) Fron? (3) the second term representing the displacement of the levels -21- of eq. (2) by the quadrupole interaction. The 2I intervals between the (2I + 1) levels are now no longer equal, Thus in a single crystal containing nuclei disposed at equivalent lattice sites the resonance spectrum is split into 2I com- ponent lines, symmetrically placed on either side of the position the line would take in the absence of a quadrupole interaction. Pound's treatment predicts an angular dependence given by Fv/az" = Zea(3c08°O - 1) (4) where 8 is tne angle between the symmetry axis and Hy (taken as the z-direction), and eq is a scalar description of the electric environment given by eq = ( p(3cos%e ~ 1)r73av (5) where the integral is taken over all charges outside the nucleus, r is the vector joining the nucleus to volume ele- ment dV, and 9 is the angle between r and the symmetry axis. As is quite evident from figure 6 and the last columns of table II, the cobalt resonance splittings follow a (3c08°¢-1) dependence quite closely. In view of the rather larre split- tings involved one would not have been at all surprised to find arpreciable deviation from this law derived from first order perturbation theory. Only one cobalt compound has been examined to any desree ~22- of thoroughness. It would be especially interesting to con- tinue this project with a study of the single crystals of such compounds as Co (Ni) gX, where X is one of many groups. Compounds of this type are assumed to have the ammonia groups octahedrally bonded to the central cobalt atom; if this were true the cobalt resonance could show no quadrupole interaction since the cobalts would lie at the centers of a cubically symmetric field. The existence of splittings would then give an indication of the desree of departure from the idealized octahedral bonding originally envisioned by Werner. -23- II, HIGH RESOLUTION NMR SPECTRA OF VARIOUS PHOSPHORUS COMPOUNDS High resolution nuclear magnetic resonance spectra of a number of phosphorus compounds related to phosphine have been taken. The samples were all supplied by Dr. Walter Mahier of USC. A Varian high resolution NMR spectrometer operating at 40 meracycles was used. The splitt nes and chemical shifts were measured by inducing side bands with an audic oscillator, Fluorine shifts were referred to tri- fluoroacetic acid, proton shifts to benzene. All splittinzs are given in cycles per second; chemical shifts are measured in dimensionless units of (aH/H, ) x 10°, where AH is the distance from the reference peak and Ho the steady magnetic field. The results are tabulated below in table III. Fig- ure G indicates traces of some actual spectra, The results are all quite straightforward except for the cases of (CH ),P5 and (CF3)pPp. These can be explained by postulating a coupling mechanism between the two central phosphorus atoms. One may envision two extreme cases of this mechanism for compounds of the form (CX3) Po. At one extreme is the case of no covpling in which event we would expect each half of the molecule to act independently giving rise to a doublet for the X resonance. At the other extreme is the case of complete coupling in which event the two central phosphorus atoms would act as a single nucleus with Splitting and chemical compounds 3 (CF, )3P: (CF, )PFo: (CF.,),PH: 32 (CF, )PHo: (CF) :PC1: (CF, )PC1p: (CF, )PCly: (CH, ),Po: (CF3) Pot (CH, ) gNP(CH3) 5: H P{CH,)5 ral N(CH) 5 res: chem. shirt CP PF res: res: res: res: ~ oy. Table III shift data for various phosphorus pair of doublets; chem. snift P-F splitting: 89.5 cps P-F splittings: 3,0 eps pair of quintets; chem. shift P-F splitting: 996 eps F-F splitting: 3.2 eps pair of triplets; chem. P-~F splitting: 85.9 eps F-F splitting: 5.8 eps pair of quartets; chem. snift P-F splitting:1250 ers F-F splitting: 5.8 eps pair of doublets; chem, P-F splitting: 68.8 eps F-H splitting: 9.7 eps pair of triplets; chem. shift P-F splitting: 48.3 ers F-H splitting: 12.0 eps pair of quartets; chem. shift P-H splitting: 203 cps F-H splitting: 12.2 cps Yeo fhen snire ln fie ey 4 nift of a oF one P-F one P-F one P-F one P-H ecehnem,. shift 35.3 eps doublet; splittings chem. snift 79.3 eps shift eps doublet; splittings: doublet; Bolittinz: chem. 153 triplet; chem. shift splitting: 7.1 eps “3 of of of Pes: one the Doublet splitting: 84.9 eps Two doublets separated by shirt of 59.5 cps. es: chem. shift of 4.3. PeH Splitting: a,2 eps mn a ft vA OL Dee - P-H solitting: am sae race ornen 2c 4, doubles with an anomolous peax in center; chem. shift of -31. ~25= CF,PH, Fres. H res. | a (CH), P, (CFa)qPe | | FIG. @: Troces of seme typical nigh resolution NMR spectra. Sec table | for splitting Magnitude s. ad ~ ~ ee os = nam & Spin of unity giving rise For the compeunds under ci 5 seconde extreme case givins ris yesonance as is evident from fir ae an intermediave caste with ac a vestigial triplet line between Srirlet in the X resonance. on (CHR) Po aets as the se e@ triplet in the proten ure 2, The (CF,},P.. acts DY learly cefinec coublet and the two main outer veaks, AST ~j REFERENCES J. S. Waugh, Ph.D. Thesis, California Institute cf Tecnnolosy, 1653. H. L. Richter, Ph.D. Tnesis, California Institute cr Technology, 1956. H. L. Richter, ibid. Mellor, Treatise on Inorganic Chemistry, Longman's; Guelin, Handbuch der Anorganischen Chemie; W. Cc. Fernelious, Editor in Chief, Inorganic Syntheses, McGraw Hill, 1946. Fernelious, ibid, Vol, II p. 217-218. The Icdide and perchlorate were precipitated from the chloride by KT and perchloric acid respectively. Fernelious, ibid. Page 221. The bromide was precipi- tated from the chloride by HBr. Werner, Ber, 45, 121 (1912). Mellor, op. cit., Vol. XIV, page 663. Donay and Nowacki, Crystal Data, Geol, Soc. of An,, . E. Pake, J. Chem. Phys., 16, 327 (i948). Franca G J. ter Berg, Rec. Trav. Chim. Pays Bas, 58, 93 (193°). J. S. Waugh, op. cis. - Richter, op. cit. L R. V. Pound, ihys. Rev., 79, 035 (1951).