J. CHEM. SOC. FARADAY TRANS., 1994, 90(4), 559-570 Speciation, Structural Characteristics and Proton Dynamics in the -Systems NH,NO, 1.5H20 and NH,NO, 1.5H20-(HNO, ,NH,F, NH,)-H20 at 5OoC Lars A. Bengtsson, Filip Frostemark and Bertil Holmberg* Inorganic Chemistry I,Chemical Center, University of Lund, P.O. Box 124,S-22100Lund, Sweden In order to obtain a good basis for exploring metal-ion complex formation in molten NH4NO3. 1.5H20, some fundamental characteristics of the pure hydrous melt and a number of compositions in the NH,NO,. 1.5H,0-(HN03, NH,F, NH,)-H,O system, have been investigated at 50"C.Several aspects have been taken into consideration, eg. thermodynamics of solvent autoprotolysis and HF formation, dynamics of proton exchange and structural properties.The acid dissociation constant of NH;, K, , and the equilibrium constant for formation of HF, K,,, were obtained from potentiometric measurements; K, = (2.2 & 0.2) x lo-' (mol kg-'), and K,, = 2160 & 40 (mot kg-I)-'. Results from "F NMR spectroscopy indicate that unprotonated fluoride, F-, probably exists as an H,NH+. * .F-ion pair in the solvent. The change in the "F chemical shift with increasing HNO, content in (NH,NO3-NH4F-HNO3). 1.5H20 verifies the conclusion from potentiometric data that HF is the only protonated fluoride species present. Raman spectroscopy and ' NMR experiments give clear evidence for an increased tendency to NH,+-..NO, ion-pair formation with decreasing water content in the systems NH,NO,-H,O.However, no loss of degeneracy of the internal v, and v, nitrate bands at 1380 and 718 cm-', respectively, was observed. The D,, symmetry of NO, seems to be preserved in the NH,N03. 1.5H20 melt. Results from Raman scattering, 'H NMR and l4N NMR experiments show significant changes in the spectra upon acidification with HNO, . These observations suggest an increase in hydrogen-bonding ability with increas- ing acidity. Results from large-angle X-ray scattering experiments on NH,NO,. 1.5H20 cannot be explained by a model comprising only interactions betwFen water molecules and ions. A residual contribution to the overall radial electron density distribution at 1.8 A is tentatively assigned to remarkably short N(NH,)-O(NO,) distances. 'H NMR spectroscopy shows a strong retardation of the proton exchange between NH; and H20 in the acidic region.The rate constant, k, , for the proton-exchange step H,N .HOH(OH,),-, + H,O -+ H,N . (OH,), + HOH, is estimated at (4.3 1.5) x lo7 s-'. This paper serves as a prologue to a number of investigations on metal fluoride and hydroxide complex formation in hydrous ammonium nitrate media at 50 "C.l Attention will be focused on the formation of polymetal complexes, i.e. com-plexes with formal composition [M, Xy+3, (m> 1, q a 1). Previous investigations in molten equimolar (K, Na)NO, at 240-280 "C provide evidence for the formation of dimetal complexes with a number of nontransition-metal ions, e.g. lead@) halide and hydroxide complexe~,~-~ alkaline-earth-metal fluoride complexes6 and lithium fluoride and hydroxide complexe~.~It has been emphasized that the formation of cationic species with high formal ionic charge should be enhanced by a purely ionic solvent, such as a molten ~alt.~,~ However, the high temperature normally required in work with molten anhydrous salts as solvents may cause crucial experimental problems.The introduction of a small amount of water to form hydrous melts allows lower working tem- peratures, which also enables the study of otherwise ther- mally unstable compounds, e.g. Bi(NO,), . For many so-called hydrous melts with less than 4-6 mol of water per mol of salt, the amount of water may be insufficient to form complete hydration shells around the constituent ions, and short-range Coulombic interactions like those in anhydrous melts prevail.lo Several investigations on metal-ion complex formation with chloride, bromide and iodide in hydrous nitrate melts, have been published over the years (see ref. 11-16 and liter- ature cited therein). Almost exclusively the main goal in these papers has been to explore the formation of anionic species, i.e. MXi-(n 2 1). In some cases the formation of hydrolysis products has been taken into account, at least qualitatively. To the knowledge of the present authors no investigation on complex formation with fluoride in such melts has been reported, probably due to experimental problems associated with hydrogen fluoride equilibria, as well as the more sophis- ticated data evaluation required.Ammonium nitrate holds a unique position in the proper- ties of salts in very concentrated aqueous solutions. The clas- sical works of Vollmar' and Narten' emphasize the striking similarity of the NH,f ion to the water molecule con- cerning mass, partial molar volume, bond angles and inter- atomic distances. NH,f and H20 form hydrogen bonds of about the same strength and the electrostatic forces exerted on their nearest neighbours are likely to be the same. Results from various spectroscopic and thermodynamic investiga- tions also suggest that the NH; ion fits well into the tetra- hedral local structure of liquid water.' Ammonium nitrate has been chosen as the solvent salt because of its very large solubility in water.The temperature of 50°C is appropriate in order to attain a hydrous melt with low water content, 1.5 H20 per NH,NO,, and it also allows the use of experimental techniques mainly based on com-mercially available equipment. The necessary prerequisites for our understanding of complex formation reactions between an ion such as fluoride and metal ions (with the inherent tendency to protolysis in hydrous media) obviously include a proper knowledge of the role of various protolytes of metal-free systems. This includes thermodynamics of solvent autoprotolysis and HF formation, structural charac- teristics and dynamics of proton exchange. These phenomena have been studied in the present work by use of several experimental techniques, such as potentiometry, multinuclear NMR, Raman spectroscopy and large-angle X-ray scattering.Experimental Chemicals Commercially available NH,NO, (Merck) and NH,F (Merck) used were of pA grade. Standard Karl-Fischer analyses showed the water content of NH,NO, to be 0.50 wt.%. The solvent melt, NH,NO,. 1.5H20,was prepared by mixing stoichiometric amounts of NH,NO, and doubly dis- tilled H20 at 50°C.NH,F was analysed according to the titration procedure by Lingane using La(NO,), (Merck, PA)." It was found to consist of 98.44 wt.% NH4F; the rest being H20. KF (Fluka, PA) was dried at 130°C for several days. NH,F, HNO, (65%, Merck, PA) and NH, (25%, Merck, PA) were used in preparing stock melts of various concentrations.NH, was analysed by acid-base titrations and was found to contain 22.5 wt.% NH, and water up to 100.0 wt.%. Fluoride-containing stock melts were stored in plastic containers in order to prevent the formation of fluoro- silicates as a result of hydrogen fluoride attack on glass. In particular, the NH, stock melts were handled with great care in order to avoid NH, losses. The reference melt used in the potentiometric measurements was prepared from AgNO, (Merck, PA) and stored in dark containers. All measurements were performed at 50°C. Concentrations are given in rnol kg-of NH,NO, unless otherwise stated. The melt composi- tion for all investigated systems was such that nH20/XnX = 1.5 (X = F-, NO,). EMF Measurements Apparatus The cell compartment consisted of a 100 ml Plexiglass vessel with thermostatted water encapsulating the inner part of the cell containing the test melt.The melt was stirred with a Teflon-coated magnetic stirring bar. Proton Fluoride Equilibria The change in fluoride-ion activity upon successive additions of (H,NH,)NO, * 1.5H20 to a NH,(F,NO,). 1.5H20 melt, was measured using a fluoride-ion-selective electrode with an LaF, membrane (Orion 94-09). The reference half-cell con- sisted of a Pyrex glass tube with a ceramic plug providing the contact between the internal and external melts. The internal compartment of the reference electrode was filled with an (Ag, NH,)NO,. 1.5H20 solution, 0.1 mol kg-' in Ag'. The cell may be described as NH,NO, * 1.5H20 AgNO, .1.5H20 II(0.1 mol kg-') F-innerNH4F * 1.5H2O (CF) membrane half cell The results of calibration series performed at various total concentrations of ammonia, are shown in Fig. 1. Fig. 1 clearly demonstrates the essential independence of ammonia concentration, indicating negligible formation of hydrogen fluoride species in the test runs. Hence, it holds that [F-] = CF for the calibration experiment. In the concentra- tion range 3 x lo-' < C,/mol kg-' < 0.4, the relation between the emf, EF ,and log[F-] is described by where EF = -588.9 0.2 mV and k, = 64.2 & 0.1 mV. The value of k, coincides well with the theoretical value at 50°C, [RT/F In 10],heof = 64.1 mV. After addition of a known amount of a stock melt of NH,(F,NO,).1.5H20 to the solvent melt, the cell was left for about 2 h in order to allow the temperature to stabilize. When stable emf readings of the cell were obtained, increasing amounts of a stock melt con- taining (H, NH,)NO, * 1.5H20 were added and the change in emf was recorded. A stable emf (within kO.1 mV) was obtained in <20 min after each addition. The total concen- trations used were lo-, < C,/mol kg-' < 5 x lo-, and < CJmol kg-' < 5 x J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 600 I I I I 550 c c\ 300 i350 0 1 2 3 4 5 -log(C,/mol kg-') Fig. 1 Results of calibration series for the fluoride ion-selective elec- trode at 50°C, performed at various total concentrations of ammonia; (0)CNH3= 0, (m) CNH3= 2.6 x mol kg-', (A) CNHJ= 9.2 x lo-' mol kg-' Evaluation of K, ,the Acid Dissociation Constant of NHf The changes in emf, upon addition of a stock melt containing NH, to the solvent test melt, were measured by means of a separate glass electrode (Metrohm 94-06) and an ammonia gas-permeable membrane electrode (Orion 95- 12).For the glass electrode the cell is schematically described as NH,NO, * 1.5H20 AgNO 1.5H20 /I(0.1 mol kg- l) In the concentration range lo-, < CH/mol kg- ' < 0.1, the emf of the cell follows Nernst's law. A typical test run is shown in Fig. 2. The relation between the emf, E, ,and C, is E, = Ei -kH log(CJmo1 kg-') (2) where Ei = 87.8 f 0.3 mV and k, = -65.9 & 0.2 mV. The observed Nernstian behaviour implies that the changes in the activity factors are negligible, and we may safely conclude that C, = [H'], representing the concentration of the hydrated proton H(H20),?.The ammonia electrode utilizes a hydrophobic gas-permeable membrane to separate the sample solution from the internal electrode solution. The internal filling solution I I I I I60 I E.-. LLII -60 > -**I J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 was replaced with a similar solution but more concentrated in NH4N03 (60.4 wt."/,), in order to eliminate the instability due to osmotic effects at the membrane. The cell is described as NH4N03* 1.5H20reference half-cell /I HNO, * 1.5H2O (CH) /INHZ ('NH3) ammonia gas- ammonia permeable membrane inner half-cell NMR Spectroscopy NMR measurements were performed at 50.0"C on a Varian Unity 300 MHz spectrometer, operating at 299.849 MHz for 'H, 21.665 MHz for 14N and 282.203 MHz for 19F.'H NMR spectroscopy was used to investigate the dynamic properties of test melts with varying hydrogen-ion concentration. The proton-exchange rates were studied in a number of melts with varying hydrogen ion and ammonia concentrations. Sample compositions are given in Table 1. The total concentrations given in Table 1 are based on weighed amounts of added HNO, and NH, stock melts, respectively 99.9% D,O (Sigma) contained in a glass capil- lary tube was used as external lock substance. 'H NMR chemical shifts are given relative to an external 0.06 g cmP3 aqueous solution of sodium 3-trimethylsilylpropanesulfonate, C,H, ,NaO,SSi.The proton signal of the reference sample was shifted 0.27 ppm to higher frequencies when the tem- perature was raised from 25 to 50°C. The lineshapes were analysed by the program DNMR520 and experimental rate 60 20 ,-20 E---.n = -60 LU' -1 00 -1 40 0.5 1.0 1.5 2.0 2.5 3.0 3.5 -log(CNH3/rnolkg-I) Fig. 3 Nernst plot of a calibration run at 50°C for the ammonia gas-permeable membrane electrode 56 1 Table 1 Total concentrations of HNO, and NH, used for the investigation of proton dynamics sample C,,,,/mol kg-' sample CNH3/mo1kg- 0.99 8 0 0.10 9 1.4 x 10-4 1.3 x lo-' 10 2.4 x 10-3 7.0 x 10-3 11 2.6 x 10-3 1.7 x 10-3 12 0.19 9.5 x 10-4 13 0.83 1.3 x 10-4 0 constants in the range 0.1 < kexp/s-' < lo5 were evaluated for the proton-exchange reaction by a visual fit of calculated spectra to experimental ones.14N NMR measurements were performed on melts with varying hydrogen-ion activity. The effects on I4N NMR spectra upon addition of F-were investigated. Measure- ments were also performed on aqueous NH4N0, solutions with varying water content. Neat CD,NO, was used as external reference. 19F NMR measurements were undertaken in support of the results obtained from the potentiometric investigation on the hydrogen fluoride complex formation. Samples with a constant total fluoride-ion concentration of mol kg- ' and HNO, concentrations ranging from lop3 to 0.16 mol kg -' were investigated.Measurements were also performed on samples with constant fluoride-ion concentration and increasing amounts of NH,NO,, ranging from dilute solu- tions to the hydrous melt composition, NH,N03 1.5H20. All samples were kept in 5 mm Plexiglass tubes, in order to avoid the formation of fluorosilicates. The resonance fre-quency of the 19F NMR spectra was measured relative to an external solution of neat CF,CO,H. Raman Spectroscopy The Raman spectra were recorded on a Bruker IFS 66-FRA 106 FT Raman spectrometer with a Ge diode detector and a low-power Nd :YAG laser (wavelength 1064 nm) providing the exciting radiation. The resolution was 4 cm-'. Sealed-off 5 mm standard NMR tubes were used as sample containers. A simple furnace consisting of an aluminium block electri- cally heated by internally mounted heating elements was employed.Large-angle X-ray Scattering, LAXS LAXS measurements on NH4N0, . 1.5H20 were performed with a GSD Seifert X-ray 8-8 diffractometer with a curved LiF monochromator of Johanson type, using Mo-Kcr radi- ation (A = 0.7107 A). The melt was kept in a Teflon con- tainer situated inside a closed glass vessel with 0.15 mm glass windows, in order to maintain a controlled atmosphere and to obtain a constant surface level. A slightly modified version of the previously described furnace and temperature control equipment was used.4 The monochromatized radiation scattered from the surface of the melt was measured in the range 0.65 < B/degrees < 58.3 in steps of s = 0.0335 (s = 47cA-l sin 0) by use of a scintillation detector.In the range 0.65 < 8/degrees < 3.2, lo4 counts were collected at each position. In the range 3.2 < B/degrees < 49.1, 4 x lo4 counts and in the range 49.1 < O/degrees < 58.3, 2 x lo4 counts were collected at each position. The observed inten- sities, lobs(8),were corrected for background radiation, Compton scattering, polarization and multiple scattering. After a normalization procedure the reduced intensities, i(s), J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 were obtained and Fourier transformed to give the reduced radial distribution function, D(r)-4nr2p0, where po is the average charge density of the atoms, All calculations were performed using the program KURVLR.2' The theoretical interaction parameters were refined using the least-squares program STEPLR.22 A more extensive description of the data treatment is given Results and Discussion Solvent Equilibrium For the NH,NO,.1.5H20 melt the following principal pro- tolytic equilibrium exists : NH: .(OH,), + (I + s -t)H,O=NH, *(OH,), + H(H,O),? Owing to the low overall water-to-salt-ratio the most likely values of I, s and t are 1 or 2. This reaction is characterized by the following equilibrium cons tan t (4) The ammonium ions and water molecules are present in large excess and their respective activities are therefore assumed to be constant.Hence, eqn. (4) may, to a first approximation, be simplified: K, = [NH,][H+] (5) The potentials of the ammonia and glass electrode were recorded after additions of NH, to the melt and [NH,] and [H'] were determined by use of eqn. (2) and (3). Based on a large amount of data the following value of the equilibrium constant was obtained: K, = (2.2 & 0.2) x (mol kg-')2 (6) Results of preliminary potentiometric measurements on metal-ion hydrolysis in slightly acidic solutions showed that the value of the product [NHJCH'] was equal to K, within the limits of error at each melt composition.' Proton Fluoride Equilibria Addition of a stock melt of (H,NH,)NO, -1.5H,O to a melt containing fluoride, causes a change in emf which may be attributed to formation of species such as HF, HF, or HzF+ according to H+ + F-eHF (11) H+ + 2F-e HF; (111) 2H' + F-e H2F+ (IV) etc. The formation reactions may be summarized xH+ + YF-=H,F;-~ (V) Assuming constant activity coefficients at constant water : anion ratio, the equilibrium constants corresponding to eqn.(V) may be expressed as (7) The total fluoride-ion concentration is given by CF = [F-] + [HF] + 2[HF,] + [H,F+] + Two separate cases can now be distinguished: EF(C, = 0) = E: -RT/F In CF (9) EF(CH)= 15: -RT/F ln[F-] (10) which gives AEF = EF(CH) -E,(C, = 0)= RT/F lIl(c~/[F-]) (1 1) By combining eqn. (7) and (8) we obtain C,/[F-] = 1 -k P1l[Hfl 28,,[H+][F-] + B21[H+I2 + ... (12) Clearly, a plot of CFICF-1 us.[H'] provides some quantita- tive information on the system: (i) Formation of only HF would result in a straight line with the slope /I1 ;(ii) forma- tion of polyfluoric species such as HF; would yield a depen- dence on CF; (iii) existence of HzF+ and other polyprotic species would be indicated by a curvature. A plot of C,/[F-] us. C, for three different values of CF is shown in Fig. 4. The fact that all data can be described by a single straight line clearly demonstrates that HF is the only species formed in significant amounts [case (i) above]. However, note that the approximation C, = [H'] is not a good one at lower values of C,. Such an approximation in the evaluation of the fluoride ion-selective electrode response might introduce errors in the interpretation of the data in that region.For the final computation of the equilibrium constant, the theoretical Nernst slope, total concentrations and emf data were used as input in the least-squares Fortran program EMFALL, which is based on a modified STEPIT s~broutine.~~The program minimizes the sum Z(AEexp -AECalJ2 in the search for the best model. The result of the computations is displayed in Table 2 along with some liter- ature data. Although K,,, the equilibrium constant for formation of HF according to eqn. (11), has been well characterized in dilute aqueous solution at relatively low ionic strength, no data seem to be accessible for ionic strengths larger than 8 mol dm-3 and for temperatures other than room tem-perat~re.~~.~~A smaller value of KHF is obtained in the present investigation as compared to results in concentrated NaClO, solutions (Table 2).24 This fact may be a result of the higher working temperature as well as the stronger solvent cation-fluoride interaction which destabilizes the hydrogen fluoride bond in the present solvent as compared to NaClO, solutions.As a result, a smaller value of KHF would be expected in the present study as compare to results of investi- gations in dilute aqueous NH4N03 solutions at room tem- perature. On the contrary, a larger value of KHF is obtained 300 I I I I I 1 I I A 250 -200 I kl2 150 100 50 0 0 0.02 0.04 0.06 0.08 0.10 0.12 0.14 C,/mol kg -' Fig. 4 C,/[F-] us. C, for different values of C,; (0)1.05 x(m) 3.20 x (A)4.68 x mol kg-' J.CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 Table 2 K,, in NH,NO, -1.5H20and some related systems ionic medium temperature/"C log K,," ref. ~ NH,NO,. l.SH,O 50 3.334 0.008 this work 0.5 rnol dm-3 NH,N03 25 2.84 28 0.5 rnol dm-3 NaCIO, 25 2.922 f0.003 26 2.0 mol dm-3 NaCIO, 25 3.108 f0.005 24 4.0 rnol dm-3 NaCIO, 25 3.539 f0.005 24 6.0 rnol dm-3NaC10, 25 4.058 0.006 24 8.0 rnol dm-j NaClO, 25 4.598 f0.010 24 PKHFexpressed in mol dm-3. (Table 2). This observation suggests that factors other than temperature and straightforward medium concentration effects are of importance for the stability of hydrogen fluoride complexes.This undoubtedly reflects the profound influence of the short-range ordering that becomes predominant as the molten salt characteristics begin to prevail over standard aqueous solution characteristics. The existence of other protonated species in aqueous solu- tion, mainly HF, and H2Ff, has been rep~rted.~~-~~ In an attempt to verify the apparent absence of such species in (NH,NO,-NH,F-HNO,) * 1.5H,O, "F NMR measure-ments were performed on melts with constant total fluoride- ion concentration, CF, and varying hydrogen-ion concentration. Only one single fluorine resonance signal was observed for these melts owing to the rapid exchange of fluo- ride between different sites. To a first assumption the single peak may be considered as emerging from a weighted average of the chemical shifts, bF-and oHF, characteristic of F- and HF, according to (13) a is the ratio [F-]/C,.The concentrations used and the observed "F chemical shifts, bobs, are displayed in Table 3. In order to evaluate the fractions of F- and HF, respec- tively, oF-and cHFwere estimated. The "F chemical shift of a strongly basic melt (NH, added in large excess) was taken as an estimate of c,-. A value of -16.0 ppm relative to CF,CO,H was obtained, which may be considered as the chemical shift of fluoride in an ammonium-ion environment. The other limiting value, bacid, was determined by extrapo- lation of bobs to C" ' = 0. This procedure is shown in Fig. 5. A further analysis of the 19F NMR data was made as follows.If HF is the only protonated fluoride species, it obvi- ously holds that baCid = bHF.Hence, [HF] and [F-] were estimated from eqn. (13), and the quotient QNMR = ([HF]/[F-])NMR was calculated for each value of C,. The free fluoride-ion activity for each sample was measured by Table 3 Total concentrations of HNO, and observed I9F chemical shifts CH/lO-mol kg- 156.2 -88.0 81.44 -87.7 51.17 -87.2 42.40 -87.0 31.18 -86.2 21.64 -84.6 10.43 -67.7 7.45 -54.2 4.86 -40.7 1.15 -22.0 --18.3 ~ ~~~ The fluoride ion concentration was within the range 8.6 d CF/10-, mol kg- < 10.0. ([HF]/[F-]),,, was formed. If the proper equilibrium model involves only the formation of HF, then the ratio, R, between the two quotients QpoJQNMR, should be equal to 1.The value of this ratio was found to be 1.0 & 0.2, which demonstrates the consistence between our analyses of potentiometric and "F NMR data in terms of a simple F--HF model. However, it should be emphasized that small systematic errors in oF- or UHF may cause substantial deviations from R = 1. Table 4 summarizes the results from the present work together with some pertinent literature data. The data displayed in Table 4 clearly demonstrate incon- sistencies in the chemical shifts reported, e.g. the 19F chemical shift of 'free', uncomplexed, fluoride. The word 'free' is some- what misleading since, of course, a 'free' or naked fluoride ion does not exist in a chemical environment owing to its high ba~icity.'~.~' However, the expression will be used throughout the present text to denote unprotonated fluoride.In order to investigate the influence of NH; ions, 19FNMR experiments were performed on samples with increasing con- centration of NH,NO, ranging from 0.1 mol dmP3 aqueous solution to the composition NH,NO,. 1.5H20.The fluoride- ion concentration, expressed as mol Fa (kg solution)- ' was kept constant. The results are shown in Fig. 6. From the diagram the 19F chemical shift of an F-ion in a dilute ammonium nitrate aqueous solution is estimated to be -42.0 ppm relative to CF,CO,H. This value seems to be reasonable compared with the literature data reported for dilute aqueous solutions (Table 4)., The shift of free F-moves towards higher frequency with increasing ionic strength which is in agreement with earlier reports, and may be explained either on the basis of an means of the fluoride ion-selective electrode, [HF] was calcu- increase in hydrogen bonding or on the formation of an ion lated as [HF] = C, -[F-1, and a similar quotient Q,, = pair, H,NHf- .-F-.32*36,37 Formation of protonated species, Table 4 Result of the 19FNMR measurements on the hydrogen fluoride system; literature data are summarized for comparison; chemical shifts are reported relative to neat CF,CO,H medium NH,NO,.1.5H20 aqueous solution aqueous solution aqueous solution aqueous solution aqueous solution Bu,NH:-,(F-/HF;) (n= 1-4y HC0,H OF -OHF ~HF~ ref.~ -16.0" -88.2 -this work -40.7' 33 -40.6' 34 -40.6' -83.2 -73.6 31 -40.6' -84.7 -73.1 35 -40.6' -61.1 -77.0 32 -44.6d --71.9, -79.6 86 -49.8' -83.6, -89.4 -49.8 -101.4 -67.4 51J a This value corresponds to the chemical shift of the fluoride ion probably existing as H,NH+-..F-. Fluoride present as F(H,O);. Molten liquids at room temperature except Bu,NHlHF; which is dissolved in CDCl,. n = 1. n = 4. 19FChemical shifts reported relative to C,F, in formic acid. The data were calculated using the expression a(CF,CO,H) = a(C,F,) + 84.4 ppm.'' 564 -85 t 1 -89l ' ' I I ' I ' 1 0 10 20 30 40 50 Ci'lkg mol-' Fig. 5 Extrapolation of oobs to C,' = 0 in (H,NH,)(F,NO,). 1.5H20 at 50°C 4035 n I I I I I I I 1 0 5 10 15 20 25 30 35 40 ' CNH4No3/mol (kg H,O) -Fig.6 Result of 19F NMR measurements on solutions with con- stant fluoride-ion concentration and increasing amounts of NH,NO,, going from a 0.1 mol dm-3 aqueous solution to NH,N03. 1.5H20. t = 50 "C. i.e. HF, HF;, would cause a shift towards lower frequencies which is in contradiction to what would be expected on the basis of simple first-order shielding arguments. Anion effects on the 19F chemical shift are, of course, much weaker than the cation effects, and an increase in nitrate concentration results in a contribution of <1% to the observed shift.36 Some authors assume that the effects in fluorine shielding are correlated to the strength of the hydrogen bond to the fluo- rine and ab initio calculations for the NH,-HF combination show a minimum-energy state for H,N..HF in which the proton of the hydrogen bond is residing closer to the fluorine than to the nitr~gen.~' In solid NH4F the hydro- gen bonding is between the ammonium cation and the fluo- ride anion, i.e. H,N-H..-F.40 The 3 ppm difference in chemical shift of fluoride in H,O and D20 solution indicates that hydrogen-bonding effects are of importance., 1*33*41 However, the interpretation of the results of "F NMR investigations on solvation of fluoride ions in mixed binary solvents rules out hydrogen bonding as the predominant factor controlling the shifts.42 This observation is consistent with the conclusions from an investigation on hydrogen fluo- ride in aprotic solvents, which suggest that other phenomena, such as a direct overlap interaction, affect the fluorine shield- ing as much as, or even more than, hydrogen bonding.43 In an early study of 19F chemical shifts in aqueous solutions containing F- and NH:, the decrease of fluorine shielding with increase of ammonium fluoride concentration was attributed primarily to an interaction with NH,f.37 An inter- J.CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 esting parallel can be drawn to the solvation of F-in for- mamide, HCONH, ,which represents a compromise between the formation of a hydrogen bond to the HCO-proton and direct N-F interaction with the nitrogen atom of the NH, group. The fluoride ion is located close to the amide A quantum-chemical study in the gas phase of the process NH, + HFeNHZF-shows that the interaction between NH, and HF is of a hydrogen-bonded nature.The system is most stable in the form of the neutral complex H,N. -.HF.45 However, in aqueous solution this reaction equilibrium is shifted towards the right owing to stabilization of the charged components by the medium. The presence of H20 molecules also facilitates proton transfer from HF to NH, It is .45746 therefore suggested that the ion pair, H3NH+...F-, is the most stable intermediate in an aqueous medium. Ab initio calculations carried out on hydrates of NH4F predict that ionic structures, e.g. H,NH+. --F-,are more stable than the neutral H,N. .HF, in the presence of solvating water.47 When more water is included into the calculations, the bridged4' and solvent-separates structures4' shown in Fig.7 become competitive in energy with the hydrogen-bonded ion pair. Rather surprisingly, results from relaxation time measure- ments on fluoride ions in aqueous solution, show that NH: ions do not have any different effect on the "F resonance, as compared to the alkali-metal ions. The dominant factor is a rejection of hydrogen bonding.48 It is not known which configuration of F- prevails in the present solvent melt, whether it is a hydrogen-bonded 'neutral' H,N.. .HF arrangement, an ionic form, H,NH+. .aF-, or a water-separated or bridged entity. It seems likely that a combination of these is the proper one. The neutral arrangement seems unlikely in the present hydrous melt owing to the strong ionic character of the medium. On the other hand, the large change in shift upon addition of NH,f ions to aqueous fluoride solutions is not consistent with a predominantly ionic NH: -F-interaction. The hydrogen-bonded ion pair, H,NH+...F-, or a bridged or water-separated structure are all likely candidates.The former seems most appropriate owing to the large change in 19F chemical shift observed when increasing amounts of NH,NO, are added to a melt containing fluoride. Such a model is consistent with the fact that both NH; and F-ions HI I H (b) Fig. 7 Structural models for the local fluoride environment in aqueous ammonium nitrate solutions; (a)ion pair, (b)bridged and (c) solvent-separated ion pairs J.CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 fit very well into the tetrahedral structure of the liquid water hydrogen-bonded net~ork.~~.~' The inconsistencies in oHFbetween the present medium and dilute aqueous solutions and formic acid may be due to the following facts. In one of the early reports glass tubes were used as container^.^^ Both fluoride and hydrogen fluo- ride solutions readily attack glass to produce SiFg-. The I9F chemical shift of the SiFg- ion occurs at higher frequency than those of HF and HF;. If a significant amount of SiFi-is present in the solutions, a lower value of the chemical shift of the observed average signal will result and errors appear in the evaluation of the chemical shifts of F-, HF and HF;. The chemical shifts of F-, HF and HF; in the investigation performed on formic acid are reported relative to C6F6 .51 This substance is not recommended as an internal standard owing to significant solvent shifts.52 A proper comparison with chemical shifts using CF3C0,H as reference is therefore not possible.Raman Spectroscopy Measurements Raman spectra of an unperturbed nitrate ion of D3h sym-metry ought to exhibit peaks from three fundamental vibra- tional modes: v,(A'J z 1050 cm-', v3(E')z 1380 cm-', v4(Ef)z 716 cm-'.53 The NHZ ion has symmetry, which gives rise to four Raman-active vibrational modes. The fre- quencies of these modes taken from solids are v,(A;) z 3040 cm-', v2(E) z 1680 cm-', ~3(T2) z 3145 cm-' and ~4(T2)x 1400 ~m-'.~~ In aqueous solution, or whenever water is present, 0-H stretching vibrations appear at about 3200- 3600 cm-'.53*55 Table 5 presents some literature data on Raman spectra of NH4N03 in aqueous solution, liquid ammonia solution, melt and the solid state. The interpretation of the Raman spectra may be difficult owing to band overlap. A tendency towards splitting of the asymmetric stretching vibration mode v3 and of the in-plane deformation v4 of the nitrate ion, may be ascribed to a lower- ing of symmetry from D3h to C,, or C,. Such effects are attributed to nitrate-water and nitrate-cation inter-action~.~~,~~,~~A more evident loss of degeneracy is the result of the formation of ion pairs, and monodentate and bidentate metal nitrate c~mplexes.~~,~~For the covalently bonded complexes also a low-frequency metal-oxygen stretching vibration mode Raman spectra of molten NH4N03 and ND4N03 indicate a splitting of the nitrate v3 fundamental mode on the basis of a curve-resolved fit of the broad asymmetric band centred at about 1370 cm-'.58 The two components at 1320 and 1410 565 cm-' were attributed to a splitting of the v3(N03) band. This result is in contradition to neutron diffraction data on molten ND4N03 .59 The interatomic distances found conclu- sively demonstrate that the NDZ ions conform to a tetra- hedral, and the NO, ions possess an equilateral triangular geometry. The NO, ions seem to retain their D,, symmetry in the melt, and consequently no loss of degeneracy of the E' mode of the NO, ion should be expected.Note that only one single band was reported in the 715 cm- ' region,s8 which normally is explained in terms of the absence of long-lived ion pairs. However, splitting of the v3 band was also reported in infrared studies on a 1 : 1 NH4N03 glass.60 That system shows a close resemblance to the ionic liquid. For the anhy- drous molten salt Av3 = 90 cm-' and for the glass Av3 = 82 cm-'. The v3 splitting of molten NH4N03 is interpreted as originating from a specific cation-anion interaction involving some hydrogen bonding.58 These results are in agreement with infrared spectra of solvated H30tN0, and NHZNO; in an argon matrix.60 The distorting ability of H30+ and NH,f equals that of the alkali-metal ions.The large Av3 values are suggested to arise from the effect of the diffusively charged H30+ and NHZ ions which are engaged in hydro-gen bonding with the NO; ion. The ammonium and hydro- xonium ions interact much more specifically with an NO, ion than do the alkali-metal ions, e.g. Li+. This may be an effect of hydrogen bonds being directional and specific, whereas Coulombic forces are not. Liquid ammonia, with a relative permittivity considerably lower than that for water, favours the formation of ion pairs. Some reports indicate contact ion-pairing between NH,f and NO, ions, based on loss of degeneracy of the v3 band Some authors report that the splitting of the v3 band is due to a hydrogen-bonding interaction with the solvent, and since the v4 bending mode is symmetrical in the NH,NO, solutions, contact ion-pairing is not e~ident.~~,~' The splitting of the degenerate v3 mode is ca.17 and 51 cm- less than the values reported in aqueous solution63 and melts, respectively.s8 This result indicates weaker hydrogen bonding in liquid NH, than in aqueous solution and NH,NO, melt. The absence of any measurable splitting of the NO, v3(Ef) or v4(E') modes has also been In this case no attempt was made to resolve the envelope because of overlap with the NHZ vk band. Aqueous ammonium nitrate solutions give little indication of disturbance of the anion symmetry by the cation and it is assumed that the ammonium ion, with a tetrahedral shape, fits well into the hydrogen-bonded network of water."A' Table 5 Observed Raman band positions (cm-') and assignments for some chosen NH,NO, systems reported in the literature 1052 1400 725 88 1048 1665 1420 720 3 100 1680 3220 1400 89 1042 1657 1336 707 2777 1700 3034 1436 61 1375 1040 1657 1325 710 2774 1705 3049 1379 64 1046 1332 712 62 1381 1044 1658 1320 71 5 3150 1677 58 1410 1043 1289 715 1418 90 1415 1461 1043 1655 1289 714 3140 1672 3192 1418 91 1414 715 3220 1445 3262 1462 1042 1655 715 3135 3223 1415 92 ~~ ~ Primed wavenumbers correspond to the NH,f ion. J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 0.10 0.08 0.06 0.04 0.02 0 3500 2700 1900 1100 300 ' wavenumber/cm -Fig. 8 Raman spectrum of NH,NO,.1.5H20at 50 "C Studies of relaxation processes in aqueous solution show that an increased concentration of NH,f has almost no effect on the vibrational relaxation time, zy, or on the rotational reori- entation time, 7or, for nitrate.66 The spectral properties of concentrated aqueous solution, where the immediate environ- ment of an NO; ion must include NH: as well as H20, are not appreciably different from those of dilute solution.55 Raman studies of vibrational dephasing processes suggest an inability of the NO; oscillator to sense the difference between NH:, NO; and H20 in its nearest-neighbour inter- action.67 The following sequence is found for the strength of this interaction: H,O z NH: z NO; < Na+ G Li+. Only at very high concentrations a small disturbance due to ammonium-nitrate interactions appears in the spectra in the form of band broadening.68 Fig.8 shows the Raman spectrum of NH4N0, 1.5H20 at 50°C. The two bands at 718 and 1048 cm-'correspond to the v4 and v1 internal nitrate vibrational modes, respectively. The band centred at about 1380 cm-' is assigned to v3(NO;) and vk(NHf). The appearance of a band at 1680 cm-' is due to the v; vibrational mode of the NHZ ion, but may also contain contributions from the first overtone of the Raman- forbidden fundamental vibration 2v2. The spectrum also shows the appearance of a broad asymmetric band at about 3140 cm-' which is attributed to both N-H and 0-H stretching vibrations.It has not been possible to resolve the components of the bands at 1380, 1680 and 3140 cm-'. The bandwidth at half height for the totally symmetric stretching vibration, v1 is 10 cm- ',which is about the same as reported in aqueous solution.69 However, a slight increase in band- width is observed for the v4 band at 718 cm- ' compared with the results in liquid NH, (17 and 13 cm-', respectively).61 The addition of NH4F and NH, to NH4N0,. 1.SH20 causes no significant change in the spectra. Addition of a 1 mol kg-' HNO, stock melt causes a small but significant 0.03 t 0.02 0.01 I I shift of the totally symmetric v1 vibrational mode of the NO, ion.The band in the 0-H and N-H stretching region is also influenced. These observations may be the result of an increase in hydrogen bond strength to the nitrate ion. For the sake of comparison, Raman spectra were also recorded for melts/solutions of composition NH4N0,. xH20 at 50°C, x = 3, 6 and 11. The spectra are shown in Fig. 9. The effects of increasing amounts of water in NH4N0,.xH20 are clearly displayed in the spectra. An increase of the anion-water and cation-water interactions with increasing water content is evident by the shift of the band in the 0-H and N-H stretching region, and also from the broadening and increased tendency towards split- ting of the internal v4 nitrate band. Considering the small amount of water present in NH4N03 * 1.5H20, this observa- tion implies a substantial increase in ion-ion interactions in the pure hydrous melt.Evidence for the formation of direct NHZ-NO; ion pairs does not exist since no splitting of the internal v4 nitrate band can be observed. This observation verifies the very small change in polarizability of the NO; ion exerted by the nearest-neighbour NH,f ions. The NO; appears to a larger degree to retain D,, symmetry in the hydrous melt than in dilute aqueous solution. Since small changes in the NO; ion symmetry are easily detected, the hydrous melt is a favourable solvent for detection of metal-nitrate interactions by Raman spectroscopy. LAXS Experiments The determination of the local structure of NH4N03 solu-tions by use of LAXS techniques is complicated by the physi- cal and structural resemblance between the ammonium ion and the water molecule, weak intermolecular interactions and also the small scattering power exerted by NH; and NO,.It has been suggested that the effect of ammonium salts on the water structure may be regarded as arising primarily from the anions alone,I8 and the first direct proof of NO;-H20 inter-actions in aqueous solution, provided by diffractometric investigations, was proposed by Caminiti et al. in a study of NH4N0, * 4.4H20.70A peak at 3.5 8, in the electron density correlation function was attributed to an N(N0,)-O(H20) distance. First-and second-order isotopic difference methods of neutron scattering have provided more accurate information about the local structure. A neutron scattering study on 12 and 18 mol kg-' aqueous solutions of ND,NO, demon-strates a direct interaction between the NDZ and NO; ions at the highest c~ncentration.~' An N(ND4)-O(N0,) distance of 2.15 8, is reported.Despite strong peaks due to intramole- cular atom-atom distances at 1.04 8, for the NDZ ion and 1.23 8, for the NO, ion, the rest of the total radial distribu- tion functions are relatively featureless as a result of the weak hydration. Neutron scattering studies of ND,NO, at higher 1 I I I I I I 3400 3200 3000 2800 775 750 725 700 675 650 waven umber/cm -' Fig. 9 Raman spectra of NH,NO,. xH,O at 50 "C,x = 1.5 (a),3 (b),6 (c), 1 1 (d) J. CHEM. SOC. FARADAY TRANS., 1994, VOL.90 concentrations, viz. 50 mol kg-', fail to resolve any N(ND,)-N(N0,) interaction below 3.6 This observation has also been made for the anhydrous molten salt.59 The results suggest a significant change in the structure of NH,NO, solutions in the concentration range 12 < CNH4N03/mo1kg-' < 50. This change in local structure is cor- related to the change in kinetics of the detonation process of ammonium nitrate at an H,O: NH,NO, ratio of 2: 1. In this respect, Adya and Neilson emphasize the need for a structural investigation of liquid NH,NO, .2H20.72To the knowledge of the present authors no such investigation has yet been performed. The nearest-neighbour cation-anion distance is reported to increase in the order NH,NO,(s), (3.82 < NH,NO,(l), (4.5 < NH,N03-D20(1), (6.1 A).', This effect upon melting NH,NO, is opposite to that of the alkali-metal nitrates, in which the cations move closer to the anion when the crystal melts.', Studies with the ND: cation, indicate that no extensive rearrangement seems to occur during the fusion process for molten ammonium nitrate, and hence the nitrate ions retain their D,, symmetry.59 The reduced radial distribution function for NH4NO3.1.5H20, is shown in Fig.lqa). The peaks at 1.18 and 2.14 8, are internal structure features of the nitrate ion. The peak at 3.17 8, and the shoulder at 3.70 8, are assigned to water-nitrate distances, O(H20)-O(NO,) and O(H,O)-N(NO,), respectively. The shoulder at 2.70 8, is assigned to O(H,O)-N(NH,) correlations.The broad band around 5 8, may contain contributions from NHZ-NO, con-tacts. The assignments are based on results of LAXS and neutron scattering measurements cited above. The structure model has been refined by means of a least- squares analysis and a comparison with experimental data is shown in Fig. lqb). The resulting difference curve clearly dis- plays deviations from experimental data at about 1.8 and 2.9 8,. This observation strongly indicates the inadequacy of a model based on only water-ion contacts and implies the exis- tence of direct ammonium-nitrate interactions. Oxygen atoms from water molecules are not likely to interfere since distances in this region are missing in more dilute solutions.70 The shoulder at about 1.8 8, in the reduced radial distribu- tion function may be assigned to O(N0,)-N(NH,) contacts.This distance is remarkably short compared with the value of 2.15 8, obtained in an 18 mol kg-' NH,NO, solution.71 The outcome of this comparison and also the fact that this peak is absent in a 50 mol kg-' aqueous solution of ammonium nitrate,72 definitely reveal the need for a critical assessment of the result of the present investigation. The distance of 1.8 8, is definitely too long for intramolecular interactions but seems on the other hand too short for intermolecular contacts. These facts inevitably suggest that the shoulder at 1.8 8, should be an artefact resulting from the data treatment or from unwanted scattering contributions to the overall inten- sity from the glass windows of the cell used.However, pre- liminary results from LAXS experiments in open containers on NH,NO,. 1.5H20 melts containing foreign ions still reveal the existence of a peak at about 1.8 A.' The data treat- ment was also slightly varied from time to time depending on the system investigated. Hence, we feel safe in concluding that the 1.8 8, peak is a physical reality and we suggest that it arises from a direct N(NH,)-O(N0,) interaction. An obvious preferred interaction geometry would be one with the oxygen atom of the nitrate triangle approaching the nitrogen atom at the centre of one of the triangular faces of the NH; tetrahedron, yielding a linear N(NH,)-O(N0,)-N(N0,) segment with tetrahedral H(NH,)-O(N0,)-H(NH,) angles.In such a configuration the N(NH,)-N(N0,) distance must be in the range 567 61 II I I I & -41Y 0, -61 -8 1 I I 0 2 4 6 8 10 r/A (b) i in QO h Fig. 10 (a) LAXS data: reduced radial distribution function for liquid NH,NO,. 1.5H20at 50°C. (b) D(r)-4nr2p, functions for NH,NO, . 1.5H20. Calculated peak shapes (-- - -), experimental curve(-) and the difference (---) are shown. d[N(NH,)-O(NO,)] + d[N-O(NO,)], i.e. (1.8 + 1.18) 8, = 2.98 A, which may contribute to the deviation at 2.9 8, in the difference curve shown in Fig. lqb). Unfortunately, owing to significant overlap of the peaks in the reduced radial distribu- tion function, no least-squares refinement yielding physically reasonable values of the temperature factor and the number of distances could be performed.It is not possible to derive an unambiguous structure model from LAXS data alone. A neutron scattering investigation of NH,NO, . 1SH20 might add valuable information. 14N NMR Measurements In the discussion, vide supra, on the structural properties of NH,NO, systems, it was noted that only small changes in ion-ion and ion-solvent interactions are observed for a hydrous melt compared with those of an aqueous solution and the anhydrous molten salt. Both the NHZ ion and the NO, ion retain an almost unperturbed tetrahedral and tri- angular geometry. 14N NMR experiments therefore are expected to show only small, if detectable, changes in chemi- cal shifts compared with results from aqueous solutions.Rota- tional correlation times for nitrate salts obtained from 14N NMR measurements indicate that the NHf ion does not associate with NOT even at high concentration^.^'*^^ The cited authors suggest that the weak complex-forming power of NH; is due to its tetrahedral shape, which enables it to fit too well into the hydrogen-bonded network without dis-rupting the water structure around the nitrate ion. This view was thoroughly discussed above, and is in agreement with results reported by others.' 7*1* Results of the present 14N NMR experiments on the pure solvent melt as well as melts containing NH,F, NH, and HNO, ,are shown in Table 6. A slightly more negative shift for the NO; ion is observed than for a 12.3 mol dmP3 saturated aqueous solution, corre- sponding to a molecular ratio of H20 : NH,NO, of 4.5 : l.77 This effect may be due to a combination of an increase in the NHi-NO; interaction, consistent with our Raman scat- tering results, and the higher temperature.Additions of fluoride of ammonia have no effect on the nitrate chemical shift. This is also in agreement with the results of the Raman spectroscopic studies. In the case of fluoride, a small shift of the ammonium 14N signal is observed, conceivably a result of the formation of an NH:-F-ion pair. Note that added NH4F amounts to only about one tenth of the total amount of NH4N03. A significant effect on the Raman spectra was observed upon additions of HNO, . This observation is also consistent with results of the present 14N NMR investigation. The chemical shift of the nitrate ion changes from -4.5 to -5.7 ppm. This trend is furthermore in agreement with results obtained in aqueous solution.77 Despite the strong tendency of H30+ to interact with water molecules, an increase in con- centration of HNO, and a decrease in water content, seem to produce NO;-H(H,O): ion pairs.The results may also reflect the increase in hydrogen bonding between nitrate and water. A similar interpretation may be valid for the results of the 'H NMR experiments on acidic melts in this work. Addi- tion of a 1 mol kg- ' HNO, stock melt causes a change in the shift of 0.55 ppm towards higher frequencies.The results of the 14N NMR experiments are thus in full agreement with the results of the Raman scattering measure- ments discussed above. Dynamics of Proton Exchange The kinetics of proton exchange of amines in water have been studied by Grunwald et u1.78-80 The mechanism of proton transfer may include either the formation of an amine hydrate according to H3NH+ . (OH,), + (r + s -t)H,O + H3N. HOH(OH,),- 1 + H(H,O),? ; k,, acid dissociation (VI) H3N. HOH(OH,),- + H20 + H3N .(OH,), + HOH; k,, proton exchange (VII) H3N * (OH,), + H(H,O),+ 4 H3NHt . (OH,), + (r + s + t)H,O; k-,, acid association (VIII) or for the formation of free amine H3NH+ + OH, + H,N + HOH; (1x1 HOH: + H,O + HOH: + HOH; fast (X) H3N + HOH: + H3NH+ + OH, (XI) J.CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 Owing to the observed dependence of exchange rate on [H '1 Grunwald et al. rejected the second mechanism and the proton transfer was suggested to occur according to the mechanism (V1)-(VII) including the formation of an amine- water complex, the so-called Swain-Grunwald mechanism." This suggestion has been supported by results of ab initio calculations suggesting that the process of proton transfer takes place in solution only if assisted by solvation or by a mechanism involving a third molecule.82 According to the proposed mechanism (V1)-(VIII), the experimental rate constant for proton exchange between NH: and water is given by The constants k,, kH and k-, refer to the three steps of acid dissociation, proton exchange and acid association described above.In the present investigation, variations of hydrogen-ion concentration produce changes in the proton-exchange rates, as monitored by 'H NMR experiments. In acidic melts where exchange rates are slow, a sharp singlet of the protons in the 'H,O site' and a 1 : 1 : 1 triplet of the NH: (14N, I = 1) ion appear. In the intermediate range, the protons move between these environments at a rate which is comparable to the frequency separation which causes successive broadening and merging of the signals. In alkaline melts where exchange rates are fast, only one signal appears. These observations are visualized in Fig. 11. For the present evaluation of rate constants of proton exchange in the rate region 0.1 d kexJs-' d lo5, computed spectra were compared with experimentally obtained ones.The remarkably slow exchange rate observed in an acidic environment is in agreement with earlier reports in aqueous sol~tion.'~-~~The observed dependence on [H'] for the experimental rate constant excludes the formation of free ammonia, and favours the ammonia hydrate model for the dissociation of NHZ ions in NH,NO,. 1.5H20. In the present systems, the acid dissociation constant of NH:, K, , is defined according to eqn. (5) in which K, may be expressed as K, = kJk-, according to eqn. (VI) and (VIII), which results in k, = [H+][NH,]k-, Table 6 Results of I4N NMR measurements obtained in the present study; some literature data are included for comparison; I4N chemical shifts are referred to neat nitromethane ~ medium bN03- 'JNH~+ ref.NH,NO, .ll H ,O -4.90 -360.39 this work NH,NO,. 6H,O -4.78 -360.39 this work NH,NO,. 3H20 -4.64 -360.45 NH,NO,. 1.5H,O -4.53 -360.59 this work NH4N03. 1.5H,0a -4.53 -. 36 1.1 9 this work NH,NO, * 1.5H,0b -4.53 -360.85 this work NH4N0,. 1.5H,O' -4.57 -358.25, -360.53, -362.94 this work (3 : 5 : 3) NH,NO,. 1.5H,0d -4.61 -355.67, -358.08, -360.53, -362.98, -365.44 this work (1 : 3 : 5 : 3 : 1) NH4N03. 1.5H20' -5.67 -355.63, -358.08, -360.53, -362.94, -365.39 this work (1:3:5:3:1) NH,NO, (adJ5.0 mol dm-3 NH,NO, -3.98 -4.64 -359.55 -358.96 77 77 in 2.0 rnol dm-, HNO, (aq) 0.5 < C,/mol kg-' < 1.9.CNH,< 0.95 mol kg-'. C, = 0.01 mol kg-'. C, = 0.1 mol kg-'. C, = 1.0 mol kg-'. 12.3 mol dm-j (satd, 30 "C). J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 1213 lo1'9 Fig. 11 Result of 'H NMR experiments on melts with various hydrogen-ion concentrations. The figures correspond to the concen- trations in Table 1. For the evaluation of the rate law for proton exchange, the rate of formation of the ammonia hydrate and its consump- tion according to eqn. (VII) and (VIII), may be used as a starting point; dCH,N. (OH2)Sl kexp = dt = kH[H3N* HOH(OH2),--k-aCH3N. (OH,)sICH(H,O),+I (16) By use of eqn. (15), this may be written in the equivalent form At lower values of [H+], when k,K,[H+] 9 k,, the relation between log kexp and -log([H+]/mol kg-') may be expressed as log kexp= log(kHK,) -log([H+]/mOl kg-') (18) The upper limit of [H'] for which the condition k, K,[H+] -' 9 k, holds, may be estimated from values of K, obtained in the present study, which is close to the value in aqueous solution, and k, and k, from investigations in dilute aqueous solutions. It is concluded that eqn.(18) may be safely used in the range [H'] < lo-, mol kg-'. A plot of log kexp us. -log([H+]/mol kg- ') for this concentration range results in a straight line of slope 0.96, which is near to the expected value of 1. The plot is shown in Fig. 12. The following value of k, was obtained: k, = (4.3 & 1.5) x lo7 s-'. This value of k, is almost four orders of magnitude smaller than values obtained for dilute aqueous solution.78 This observation may seem fairly reasonable owing to the hindrance of proton transfer by the large amount of nitrate ions present compared with dilute aqueous solution.A more complete investigation of the proton-exchange pro- cesses is out of the scope of the present paper and it should be emphasized that other effects such as temperature, vis- cosity changes and contributions to the observed rate from other proton-transfer reactions, may have a significant influ- ence on kexpand therefore also on k,. Concluding Remarks The solvent properties of NH,N03 1.5H20 have been explored and a good basis has been laid for the investigation of metal-ion complex formation in this medium.The fluoride ion-selective, ammonia and glass electrodes all show Nernst- ian behaviour in a wide concentration range. For systems based on NH,NO,. 1.5H20 the product [H+][NH,] is found to be constant, implying the existence 4.5 n 3.5 P)4 0)-2.5 1.5 3 4 5 -log([H+]/mol kg-') Fig. 12 Estimation of k,, the rate constant of proton exchange, log kcxp0s. -log([H+]/mol kg-') for lop6< [HC]/mol kg-I < of the following principal acid-base equilibrium NHZ *(OH,), + (r + s -t)H,OeNH, *(OH,), + H(H,O),f K, = [NH,][H+] = (2.2 & 0.2) x low9(mol kg-'), Potentiometric measurements demonstrate the formation of HF in the hydrogen fluoride system, K,, = [HF]/[F-][H+] = 2160 & 40(mol kg-')-'.19FNMR mea- surements confirm that HF is the only protonated fluoride species formed in significant amounts. Results from 19F NMR spectroscopy indicate that F- probably exists as an H,NH+...F-ion pair in the NH,(F,NO,). 1.5H20 melt. The triangular geometry of NO, is well preserved in NH,N03 1.5H20 as revealed by Raman scattering investi- gations. In acidic melts, the hydrogen-bonding ability is increased and H+-..NO, ion pairs seem to form. This con- clusion is in agreement with results from 'H and 14N NMR experiments. LAXS experiments reveal a shoulder at 1.8 A in the reduced radial distribution function which may be attributed to a direct N(NHl)-O(NO;) interaction. Results from 'H NMR experiments show a decrease in proton-exchange rate with increasing hydrogen-ion concen- tration.For the proton exchange H,N * HOH(OH,),-+ H,O + H,N. (OH,), + HOH the rate constant, k, was determined to be (4.3 f1.5) x lo7 s-'. Future work will be focused on the formation of polymetal complexes in the present medium. The molten salt-like properties of NH,NO, -1.5H20, which seem to be well pre- served, are of vital importance since the formation of highly charged polymetal complexes is known to be enhanced by the ionic medium. This work has been supported by a grant from the Swedish Natural Science Research Council. The Council is also grate- fully acknowledged for funding the Varian Unity NMR instrumentation. References 1 L. A. Bengtsson, F. Frostemark and B.Holmberg, to be published. 2 L. Bengtsson and B. Holmberg, J. Chem. Soc., Faraday Trans. I, 1989,85, 305. 3 L. Bengtsson and B. Holmberg, J. Chem. Soc., Faraday Trans. I, 1989,85, 317. 4 L. Bengtsson and B. Holmberg, J. Chem. Soc., Faraday Trans. I, 1989,85,2917. 570 J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 L. Bengtsson and B. Holmberg, J. Chem. SOC.,Faraday Trans. 1, 1990, 86,351. 51 C. Coulombeau, C. Beguin and C. Coulombeau, J. Fluorine Chem., 1977,9,483. 6 L. A. Bengtsson, F. Frostemark, B. Holmberg and S. Ulvenlund, 52 C. J. Jameson, in Multinuclear NMR, ed. J. Mason, Plenum 7 J. Chem. SOC.,Faraday Trans., 1991,87, 1141. L. Bengtsson, B. Holmberg and S. Ulvenlund, Inorg. Chem., 53 Press, New York, 1987, p. 437.D. E. Irish and M. H. Brooker, in Advances in Infrared and 1990,29,3615. Raman Spectroscopy, ed. R. J. H. Clark and R. E. Hester, 8 9 L. A. Bengtsson, Thesis, University of Lund, Sweden, 1990. L. A Bengtsson and R. Hoffmann, J. Am. Chem. SOC., 1993, 115, 54 Heyden, London, 1976, vol. 11, p. 212. K. Nakamoto, in Infrared Spectra of Inorganic and Coordination 2666. Compounds, Wiley, New York, 1963. J. Braunstein, in Ionic Interactions, ed. S. Petrucci, Academic 55 D. E. Irish, in Ionic Interactions, ed. S. Petrucci, Academic Press, 11 Press, New York, 1971, vol. I, p. 179. G. G. Bombi, G. A. Sacchetto and C. Macca, J. Electroanal. 56 New York, 1971, vol. 11, p. 187. T. Kato, J. Umemura and T. Takenaka, Mol. Phys., 1978, 36, Chem., 1987,225, 11 1.621. 12 I. J. Zsigrai, J. N. Pilipovic and I. J. Gal, Bull. SOC. Chim. 57 R. L. Frost and D. W. James, J. Chem. SOC.,Faraday Trans. 1, Beograd., 1981,46,579. 1982,78,3263. 13 R. Nikolic and 0.Neskovic, J. Chem. SOC.,Dalton Trans., 1982, 58 J. P. Devlin, P. C. Li and G. Pollard, J. Chem. Phys., 1970, 52, 1417. 2267. 14 D. H. Kerridge, R. Nikolic and D. Stojic, J. Chem. Soc., Dalton 59 A. K. Adya and G. W. Neilson, Mol. Phys., 1990,69,747. Trans., 1986, 1663. 60 G. Ritzhaupt and J. P. Devlin, J. Phys. Chem., 1977,81, 521. I. J. Zsigrai and Z. Novakov, Sci. Znt. (Lahore), 1991,3,203. 61 J. W. Lundeen and R. S. Tobias, J. Chem. Phys., 1975,63,924. 16 H. Braunstein, J. Braunstein and P. T. Hardesty, J. Phys. Chem., 62 A.T. Lemley and J. J. Lagowski, J. Phys. Chem., 1974,78,708. 1973,77, 1907. 63 D. E. Irish and A. R. Davis, Can. J. Chem., 1968,46,943. 17 P. M. Vollmar, J. Chem. Phys., 1963,39,2236. 64 D. J. Gardiner, R. E. Hester and W. E. L. Grossman, J. Chem. 18 A. H. Narten, J. Phys. Chem., 1970,74,765. Phys., 1973, 59, 175. 19 J. J. Lingane, Anal. Chem., 1967,39, 881. 65 R. F. Armishaw and D. W. James, J. Phys. Chem., 1976,80,501. D. S. Stephenson and G. Birsch, DNMRS, QCPE, 1978,11,365. 66 D. James and R. L. Frost, Faraday Discuss. Chem. Soc., 1978,64, 21 G. Johansson and M. Sandstrom, Chem. Scr., 1973,4, 195. 48. 22 M. Molund and I. Person, Chem. Scr., 1985,25, 197. 67 T. Kato and T. Takenaka, Chem. Phys. Lett., 1979,62,77. 23 A. Sandell, Personal communication ; J.P. Chandler, STEPIT, 68 R. L. Frost and D. W. James, J. Chem. SOC.,Faraday Trans. 1, QCPE, 1976,11,307. 1982,78,3235. 24 G. T. Hefter, J. Solution Chem., 1982, 11, 45. 69 H. A. Lauwers, G. P. van der Kelen and 2. Eeckhaut, Inorg. A. M. Bond and G. T. Hefter, in Critical Survey of Stability Con- Chim. Acta, 1969, 3, 612. stants and Related Thermodynamic Data of Fluoride Complexes 70 R. Caminiti, G. Licheri, G. Piccaluga and G. Pinna, J. Chem. in Aqueous Solution, IUPAC Chemical Data Series No. 27, Per- Phys., 1978,68,1967. gamon, Oxford, 1980. 71 P. A. M. Walker, D. G. Lawrence, G. W. Neilson and J. Cooper, 26 G. T. Hefter, J. Solution Chem., 1984, 13,457. J. Chem. SOC.,Faraday Trans. I, 1989,85, 1365.27 G. T. Hefter, Polyhedron, 1984,3, 75. 72 A. K. Adya and G. W. Neilson, J. Chem. SOC.,Faraday Trans., 28 E. W. Baumann, J. Inorg. Nucl. Chem., 1969,31,3155. 1991,87,279. 29 K. Seppelt, Angew. Chem., Int. Ed. Engl., 1992,31,292. 73 R. W. G. Wyckoff, in Crystal Structures, Interscience, New York, C. Rieux, B. Langlois and R. Gallo, C. R. Acad. Sci. Paris, Ser. 2nd edn., 1964, vol. 2. 11, 1990,310, 25. 74 H. Ohno and K. Furukawa, J. Chem. SOC., Faraday Trans. 1, 31 K. Schaumburg and C. Deverell, J. Am. Chem. Soc., 1968, 90, 2495. 75 1978,74, 297. A. M. de P. Nicholas and R.E. Wasylishen, Can. J. Chem., 1987, 32 33 R. Haque and L. W. Reeves, J. Am. Chem. Soc., 1967,89,250. K. Radley and L. W. Reeves, J. Chem. Phys., 1971,54,4509. 76 65, 951.C. L. Perrin and R. K. Gipe, J. Am. Chem. SOC.,1986, 108, 1088. 34 C. Deverell, K. Schaumburg and H. J. Bernstein, J. Chem. Phys., 77 M. Witanowski, L. Stefaniak, S. Szymanski and H. Januszewski, 197 1,54,4509. J. Magn. Reson., 1977,28,217. I. T. Wang and F. I. Skripov, Dokl. Akad. Nauk SSSR, 1961, 136, 78 E. Grunwald and E. K. Ralph, Acc. Chem. Res., 1971,4, 107. 58. 79 M. T. Emerson, E. Grunwald, M. L. Kaplan and R. A. Krom- 36 J. P. K. Tong, C. H. Langford and T. R. Stengle, Can. J. Chem., haut, J. Am. Chem. Soc., 1960,82,6307. 1974,52, 1721. 80 E. Grunwald and A. Y. Ku, J. Am. Chem. Soc., 1968,90,29. 37 I. C. Wang, J. Struct. Chem., 1962, 2, 343. 81 C. G. Swain, J. T. McKnight and V. P. Kreiter, J. Am. Chem. 38 J. Emsley, D. J. Jones, J. M. Miller, R. E. Overill and R. A. Waddilove, J. Am. Chem. SOC.,1981, 103,24. 82 SOC.,1957,79, 1088. 3. J. Delpuech, G. Serratrice, A. Strich and A. Veillard, Mol. 39 M. S. Gordon, D. E. Tallman, C. Monroe, M. Steinbach and Phys., 1975, 29, 849. J. Armburst, J. Am. Chem. Soc., 1975,97, 1326. 83 M. T. Emerson, E. Grunwald and R. A. Kromhout, J. Chem. H. W. W. Adrian and D. Feil, Acta Crystallogr., Sect. A, 1969, 25, 438. 84 Phys., 1960,33,547. C. G. Swain and M. M. Labes, J. Am. Chem. Soc., 1957,79, 1084. 41 42 43 C. Deverell and K. Schaumberg, Anal. Chem., 1967,39,1879. C. Carmona, G. Eaton and M. C. R. Symons, J. Chem. Soc., Chem. Commun., 1987,873. J. S. Martin and F. J. Fujiwara, J. Am. Chem. SOC., 1974, 96, 7632. 85 86 87 E. Grunwald, J. Phys. Chem., 1963,67,2208; 2211. J. Soriano, J. Shamir, A. Netzer and Y. Marcus, Inorg. Nucl. Chem. Lett., 1969,5, 209. K. Jones and E. F. Mooney, Annu. Rep. NMR Spectrosc., 1971, 4, 391 ;R. Fields, Annu. Rep. NMR Spectrosc., 1977, 7, 1. 44 I. Birkel and H. G. Hertz, J. Chem. SOC.,Faraday Trans. I, 1981, 77,2315. 88 89 R. E. Hester and R. A. Plane, Inorg. Chem., 1964,3, 769. R. E. Hester and R. A. Plane, J. Chem. Phys., 1966,45,4588. S. Miertus, and J. Bartos, Coll. Czech. Chem. Commun., 1980, 45, 90 K. Akiyama, Y. Morioka and I. Nakagawa, Bull. Chem. SOC. 2308. Jpn., 1981,54,1662. 46 A. N. Isaev and Yu. I. Khurgin, Izv. Akad. Nauk SSSR, Ser. Khim., 1991, 817. 91 H. C. Tang and B. H. Torrie, J. Phys. Chem. Solids, 1977, 38, 125. 47 P. Kollman and I. Kuntz, J. Am. Chem. SOC.,1976,98,6820. 92 D. M. Adams and S. K. Sharma, J. Chem. SOC.,Faraday Trans. 48 K. A. Alzewel. 2.Phys. Chem., Leipzig, 1974,255, 193. 2, 1981, 77, 1263. 49 A. K. Lyashchenko, G. V. Kokovina and A. S. Lileev, Zh. Strukt. Khim., 1987,28, 88. D. S. Terekhova, A. I. Ryss and I. V. Radchenko, Zh. Strukt. Khim., 1969, 10, 923. Paper 3/05193E; Received 31st August, 1993