Vibrational Behavior of the –NO2 Group in Energetic Compounds ROYCE W. BEAL and THOMAS B. BRILL* Department of Chemistry and Biochemistry, University of Delaware, Newark, Delaware 19716

The vibrational modes of the –NO2 group in more than fifty energetic compounds containing the C-nitro and N-nitro functionalities were observed and then calculated in optimized structures using density functional theory (B3LYP/6-311G*). The trends in the symmetric and asymmetric stretches and scissor and out-of-plane deformations were explained by these calculations. A previously unreported correlation was found between the nitro group internal bonding angle and its asymmetric stretching frequency. The concept of meta and ortho/para directing groups was applicable to the trends in coupled motions in the nitroaromatic compounds. Both the scissor motion of C–NO2 groups and the out-of-plane deformation of N–NO2 groups were found to be virtually insensitive to the remainder of the molecule. These findings may be useful in analytical methods of explosive detection based on their infrared (IR) spectra. Index Headings: Vibrational assignments; –NO2 group modes; Energetic compounds; Explosive detection.

INTRODUCTION The nitro group, –NO2, is prevalent in energetic materials, i.e., explosives and propellants. It can bond to carbon, nitrogen, or oxygen forming nitro-, nitramino-, and nitrate ester compounds, respectively. Commercially important examples are 2,4,6-trinitrotoluene (TNT), 1,3,5-trinitro-1,3,5-triazine (RDX), and nitroglycerine (NG). The nitro functional group including the atom to which it is attached can exhibit six distinct molecular vibrations: the asymmetric stretch, symmetric stretch, scissor or symmetric bend, out-of-plane deformation, pivot or asymmetric bend, and a torsion motion. Typically, the asymmetric and symmetric stretch, which appear in the ranges of 1500–1650 cm21 and 1260–1400 cm21, respectively, are intense and readily identified. The scissor and out-ofplane deformation motions have lower frequencies and are less intense than the stretching modes in the infrared (IR) spectrum. As such, they frequently are unidentified or incorrectly identified in the 600–1200 cm21 region. The pivot and torsion motions are even lower in energy and are rarely observed. The IR spectroscopy of nitro compounds is well summarized by Bellamy1 up to 1968. The nitro group is planar and contains p bonds. As a result, the normal mode frequencies in an X–NO2 unit are sensitive to conjugation or resonance with the X group causing the nitro groups bonded to conjugated systems to absorb at lower frequencies than those of non-conjugated systems. The electron-withdrawing or -donating properties of X affect the –NO2 stretching frequencies. In the case of the asymReceived 26 May 2005; accepted 19 July 2005. * Author to whom correspondence should be [email protected].

1194

Volume 59, Number 10, 2005

sent. E-mail:

metric stretch, this correlation is quite direct. On the other hand, the symmetric stretch often couples with the N–X stretching frequency, making direct correlation with substituent effects less likely. Several spectral patterns have been found with aromatic nitro compounds. For instance, in systems with only a nitro group and one para (1,4) substituent, the asymmetric nitro stretch frequency is directly related to the electron donor or acceptor properties of the substituent. The correlation is less pronounced but still evident when the substituent is meta (1,3). When the substituent is ortho (1,2), the trends become obscured by steric effects that may twist the nitro group out of the ring plane, which reduces conjugation and raises the asymmetric stretching frequency. In some cases, most notably 2-nitrophenol, hydrogen bonding plays a role in determining the stretching frequencies. Bellamy also noted that correlations involving the symmetric stretch are less apparent than those with the asymmetric stretch in aromatic nitro compounds. He proposed that this is at least partly due to coupling with ring vibrations. Furthermore, the range of frequencies for the symmetric stretch is narrower than for the asymmetric stretch. For alkyl nitro compounds, Bellamy indicated that correlations involving the –NO2 group are difficult to establish because the asymmetric stretching frequency always occurs within a few wavenumbers of 1500 cm21. However, substitutions at the a-carbon and, to a lesser extent, at the b-carbon can affect both stretching frequencies. In a more recent compilation, Gunzler and Gremlich 2 noted that increasing the degree of substitution of methyl groups bound to the a-carbon leads to only a very small reduction in the asymmetric stretch frequency (based on mass effects), but a pronounced reduction in the symmetric stretching frequency. This latter effect was based on coupling with C–H bending motions of the adjacent methyl groups. Another correlation was reported by Brill.3 By comparing X-ray crystallographically determined geometries of nitramines and the asymmetric nitro stretching frequencies, a linear correlation between this frequency and the length of the N–N bond was found. He further indicated that a relation exists between the N–N bond length and the amount of gaseous NO2 liberated upon fast thermolysis. The tendency to liberate a specific product, in this case a highly oxidizing NO2 molecule, early in decomposition is important for defining the decomposition mechanisms of nitramines. The frequencies of the normal modes and atomic displacements can be calculated by ab initio methods and the IR spectrum of the molecule predicted. If the correct conformer of the molecule has been modeled, the pre-

0003-7028 / 05 / 5910-1194$2.00 / 0 q 2005 Society for Applied Spectroscopy

APPLIED SPECTROSCOPY

TABLE I. Cyclic and caged nitramines. Symbol 1 2 3 4 5 6 7 8 9 10 11

CL-20 RDX RDX-ketone HMX AZMTTC DPT TNDBN HNDZ DNNC DNP HK-55

12

HK-56

13

K-55

14

K-56

15 16

TNAZ TNCB

Name

Formula

hexanitrohexaazoisowurtzitane 1,3,5-trinitro-1,3,5-triazine 1,3,5-trinitro-1,3,5-triazina-2-one 1,3,5,7-tetranitro-1,3,5,7-tetrazocene 1-(azidomethyl)-3,5,7-tetranitro-1,3,5,7-tetrazocene 3,7-dinitro-1,3,5,7-tetraazabicyclo [3.3.1]nonane 1,3,5,7-tetranitro-3,7-diazabicyclo [3.3.1]nonane 1,3,3,5,7,7-hexanitro-1,5-diazocane 1,3,5,5-tetranitrohexahydropyrimidine 1,4-dinitropiperazine 1,4,6-trinitrohexahydroimidazo [4,5-d]imidazol2(1H)-one 1,4,7-trinitrooctahydro-2H-imidazo [4,5-b]pyrazin2-one 1,3,4,6-tetranitrohexahydroimidazo [4,5-d]imidazol2(1H)-one 1,3,4,7-tetranitrooctahydro-2H-imidazo [4,5b]pyrazin-2-one 1,3,3-trinitroazetidine 1,1,3-trinitrocyclobutane

dicted spectrum matches the measured one quite accurately. Density functional theory (DFT) offers an acceptable trade off between accuracy and speed. Since the ab initio calculations were performed on single molecules in the absence of any external forces or fields, the results are best compared with gas-phase spectra. In many cases gas-phase spectra could not be obtained and so comparison with the condensed-phase spectra were made. This work provides a plausible footing for the detection of explosive materials based on infrared absorption signatures. EXPERIMENTAL Materials. Limits on this study were imposed by the computational power available, the availability of the material to collect IR spectra, or the availability of reliable literature IR spectra. The materials were also preferentially selected so as to fit into related families, e.g., aromatic nitro compounds, nitramines, and geminal-dinitro and trinitromethyl compounds. These categories are somewhat broad so more limited groupings were also made according to cyclic/caged nitramines, other nitramines, geminal-dinitro compounds, monosubstituted nitrobenzenes, dinitrotoluenes, and trinitrotoluenes. The members of these groups are listed and numbered in Tables I through VI. Infrared Spectra and Computations. Samples were prepared as KBr pellets (about 1 part sample to 200 parts KBr). IR spectra were recorded on a Nicolet 20SXC Fourier transform infrared spectrometer with 128 scans at 2 cm21 resolution in the 4000–600 cm21 range. Several spectra were taken from the National Institute of Standards and Technology database.4

C6H6N12O12 C3H6N6O6 C3H4N6O7 C4H8N8O8 C5H10N10O6 C5H10N6O4 C7H10N6O8 C6H8N8O12 C4H6N6O8 C4H8N4O4 C4H5N7O7 C5H7N7O7 C4H4N8O9 C5H6N8O9 C3H4N4O6 C4H5N3O6

Initial geometries were calculated at the PM3 semiempirical level using Arguslab version 2.0.5–10 These PM3-optimized geometries were then used as input geometries for the Gaussian98 software package.11 The B3LYP density functional method was employed with the 6-311G* basis set. The vibrational frequencies from the calculations were multiplied uniformly by a scaling factor of 0.965 to compensate for known overestimation due to incomplete consideration of electron correlation. The factor is empirical but a value of this magnitude is similar to that used with this method and basis set.12 The experimentally determined frequencies in the solid state are still somewhat lower than those calculated, which may be the result of the crystal field effect. Molekel version 4.113 and MOplot version 1.714 were used for visualizing the output files. MOplot was used primarily for observing the motions related to each frequency. The molecular motions for each frequency above 700 cm21 (unadjusted) were examined. The complete list of these motions is available.15 In most cases the correlations were very good and the assignments were made with confidence. The assignments most in doubt are for CL-20 (1), HK-55 (11), HK-56 (12), K-56 (14), and TNAZ (15), all in Table I. In most of these molecules the backbone is strained and can exist in several conformers, which makes the molecules difficult to model computationally and results in poorer spectral correlations. ANALYSIS OF VIBRATIONAL MOTIONS As discussed earlier, six motions are possible for the X–NO2 group. Of these, four are encountered in the midinfrared (MIR) (600–4000 cm21), namely the asymmetric

TABLE II. Chain nitramines. Symbol 19

DATH

20 21

EDNA DMEDNA

Name

Formula

N-(azidomethyl)-N-{[{[(azidomethyl)(nitro)amino] methyl}(nitro)amino]methyl}-N-nitroamine N,N9-dinitroethane-1,2-diamine N,N9-dimethyl-N,N9-dinitroethane-1,2-diamine

C4H8N12O6 C2H6N4O4 C4H10N4O4

APPLIED SPECTROSCOPY

1195

TABLE III. Geminal dinitro compounds. Symbol 22 23 24 25

BDNEC BDNFEC BTNEC FEFO

Name bis(2,2-dinitropropyl) carbonate bis(2-fluoro-2,2-dinitroethyl) carbonate bis(2,2,2-trinitroethyl) carbonate 1-fluoro-2-[(2-fluoro-2,2-dinitroethoxy) methoxy]-1,1-dinitroethane

and symmetric stretch, the bend or scissor, and the outof-plane deformation. Asymmetric Stretch. It is known that, in general, the asymmetric C–NO2 and N–NO2 stretches occur 2 at about 1530–1630 cm21. Furthermore, the electron donor or acceptor properties of the subsitutents on X of the X–NO2 group affect the location of this band.1 Electron-donating X groups decrease the frequency while electron-accepting groups increase it. The results of this study support this notion. The range of frequencies of the symmetric stretch of the nitro groups (non-aromatic) in this study is 1557– 1621 cm21. The extreme of 1621 cm21 occurs with BDNFEC (23) where the a-carbon of each nitro group is substituted with another nitro group and a fluorine atom, both of which are strongly electron withdrawing. The extreme low of 1557 cm21 is attributed to the 3-nitro group on TNCB (16). Other works1,2 find that lower frequencies for the symmetric stretch are possible, but since this study deals mostly with molecules containing multiple nitro groups, lower frequencies were not observed. For the nitramino groups, the range of the asymmetric stretching frequencies is 1508–1627 cm21. In this case, the extreme of 1627 cm21 is observed in K-56 (14) and is attributed to the nitramino group in the 2-position. Because the distance between an oxygen atom of the nitro ˚ according to group and the carbonyl oxygen is 2.84 A the optimized geometry, it appears unlikely that this high frequency might be attributed to interaction between the carbonyl group and the nitro group. However, the calculated geometry of this nitro group indicates strain, ˚ and the other at 1.217 leaving one N–O bond at 1.208 A TABLE IV. Monosubstituted nitrobenzenes.

26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47

1196

Formula

Name

Formula

nitrobenzene 1,2-dinitrobenzene 1,3-dinitrobenzene 1,4-dinitrobenzene 2-nitrochlorobenzene 3-nitrochlorobenzene 4-nitrochlorobenzene 2-nitrofluorobenzene 3-nitrofluorobenzene 4-nitrofluorobenzene 2-nitrophenol 3-nitrophenol 4-nitrophenol 2-nitroaniline 3-nitroaniline 4-nitroaniline 2-nitromethoxybenzene 3-nitromethoxybenzene 4-nitromethoxybenzene 2-nitrotoluene 3-nitrotoluene 4-nitrotoluene

C6H5NO2 C6H4N2O4 C6H4N2O4 C6H4N2O4 C6H4CINO2 C6H4CINO2 C6H4CINO2 C6H4FNO2 C6H4FNO2 C6H4FNO2 C6H5NO3 C6H5NO3 C6H5NO3 C6H6N2O2 C6H6N2O2 C6H6N2O2 C7H7NO3 C7H7NO3 C7H7NO3 C7H7NO2 C7H7NO2 C7H7NO2

Volume 59, Number 10, 2005

C7H10N4O11 C5H4N4O11F2 C5H4N6O15 C5H6N4O10F2

˚ . The O–N–O angle of 129.38 is also very large. The A reason for this distortion is not obvious; however, since the calculated bands agree well with the frequencies, intensities, and pattern of the measured spectrum, it seems reasonable to assume that the distortion also exists in the actual molecule. The extreme low frequency of 1508 cm21 is attributed to the 5-nitramino group of DPT (6), which has somewhat longer than average, but fairly un˚ each remarkable, N–O bonds (calculated to be 1.230 A ˚ and 1.235 A ˚ 16). However, the and found to be 1.232 A calculated angle of 125.58 and observed angle of 123.98 is rather pinched. It seems then that a relation may exist between the O–N–O angle and the asymmetric stretching frequency. Indeed, when the O–N–O angle is plotted vs. asymmetric stretching frequency for the secondary nitramines (Table VII, Fig. 1), a trend is found. This correlation has not previously been reported, although Brill had previously reported the relation between the N–N bond length and asymmetric stretching frequency for secondary nitramines.3 The multiple entries for several compounds in Table VII result from the existence of crystallographically inequivalent –NO2 groups in the compounds. The asymmetric stretch of the nitro group is known to couple with aromatic ring motions.1 It is also known that the frequency is affected by conjugation and, through conjugation, by substitution. This is evident in Table VII. If the nitro group is rotated out of the plane of the ring, then the frequency of the asymmetric stretch increases as a result of reduced conjugation with the ring.1 In general, the donor–acceptor nature of substituents on the ring affects the frequency of the nitro asymmetric stretch. Bellamy suggested that this effect is most pronounced for ortho and para substitution.1 Further details are discussed below in the analysis of aromatic compounds. Symmetric Stretch. Like the asymmetric stretch, the frequency ranges of the symmetric stretch of both nitro groups and nitramino overlap. For the nitro compounds, the extremes from this study are 1390 cm21 for HNDZ (8) and 1299 cm21 for BTNEC (22). In this case the nitro geometries are not extremes. However, the C–N distances for HNDZ appear to be the longest, calculated at 1.558 ˚ . The BTNEC nitro groups, on the other hand, have A relatively short C–N bond lengths of between 1.537 and ˚ . However, both TNAZ (15) and TNCB (16) 1.541 A TABLE V. Dinitrotoluenes.

48 49 50 51 52 53

Symbol

Name

Formula

23-DNT 24-DNT 25-DNT 26-DNT 34-DNT 35-DNT

1-methyl-2,3-dinitrobenzene 1-methyl-2,4-dinitrobenzene 1-methyl-2,5-dinitrobenzene 1-methyl-2,6-dinitrobenzene 1-methyl-3,4-dinitrobenzene 1-methyl-3,5-dinitrobenzene

C7H6N2O4 C7H6N2O4 C7H6N2O4 C7H6N2O4 C7H6N2O4 C7H6N2O4

TABLE VI. Trinitrotoluenes.

54 55 56 57 58 59

Symbol

Name

Formula

234-TNT 235-TNT 236-TNT 245-TNT 246-TNT 135-TNB

1-methyl-2,3,4-trinitrobenzene 1-methyl-2,3,5-trinitrobenzene 1-methyl-2,3,6-trinitrobenzene 1-methyl-2,4,5-trinitrobenzene 1-methyl-2,4,6-trinitrobenzene 1,3,5-trinitrobenzene

C7H5N3O6 C7H5N3O6 C7H5N3O6 C7H5N3O6 C7H5N3O6 C6H3N3O6

show dramatically shorter C–N bond lengths than BTNEC, but not lower frequencies. Therefore, if a relation exits between the C–N bond length and the symmetric stretching frequency, TNAZ and TNCB must be special cases. TNDBN (7) does have a peak at 1399 cm21, which, according to the calculation, is weakly coupled with the nitro symmetric stretch, but this peak can be more definitely attributed to another motion. No clear relation appears to exist between electron donor–acceptor properties of substituents on the a-carbon and the nitro symmetric stretch, as it does with the asymmetric stretch. The nitro group symmetric stretch for the nitramines appears at a slightly lower frequency than for the nitro compounds. The extremes from this study are 1353 cm21 for DNNC (9) and 1233 cm21 for DATH (19). It is not apparent why these compounds produce the spectral extremes because other compounds in this study exhibit more distortion. However, it is noteworthy that the nitro group symmetric stretch couples with a variety of other molecular motions. It has been previously established 1 that coupling almost always occurs with the stretching motion of the N–X bond. This process is discussed below. Furthermore, we find that the symmetric stretch of compounds with both nitro and nitramino moieties can couple. It has been found to couple also with the –N3 symmetric stretch and –CH2– wagging motions. Coupling with N–X helps explain why the symmetric stretches of nitramino groups tend toward lower frequencies than C– NO2 groups because the N–N bond is weaker than the C–N bond. Furthermore, DATH (19) contains the –N3 moiety, which, through this coupling, may draw the nitramino symmetric stretch to lower frequency. DNNC (9) contains both nitro and nitramino groups and, again through coupling, the nitramino frequencies are raised. According to the calculated motions in this study, the nitro group symmetric stretch always couples with the N–X stretch. What has not been clear in previous work is whether the N–O bonds stretch while the N–X bond stretches, the N–O bonds stretch while the N–X bond compresses, or both, and how this can depend upon various factors. Upon extensive review of the calculated motions, it has been determined that the N–O bonds stretch while the N–X bond compresses. This is true for both nitro and nitramino moieties. It therefore seems that, if the symmetric stretch always couples with N–X in exactly the same way, it might be considered to be a single concerted normal mode instead of a coupling of two modes. Although some evidence suggests that the N–X bond strength can affect the symmetric stretching frequency, no general relation between them was found. Since the compounds in this study are primarily energetic materials, multiple nitro groups are frequently present. While the symmetric stretches of multiple nitro

TABLE VII. Nitro group angles and asymmetric stretching frequencies for secondary nitramines. Calculated (scaled frequency) Compound CL-20 CL-20 CL-20 CL-20 CL-20 CL-20 RDX RDX RDX RDX-k RDX-k RDX-k DATH DATH DATH HMX HMX HMX HMX AZMTTC AZMTTC AZMTTC DPT DPT TNDBN TNDBN HNDZ HNDZ DNNC DNNC DNP DNP K-56 K-56 K-56 K-56 TNAZ

O–N–O (degrees) 127.7 127.7 127.7 127.7 127.7 127.7 126.6 126.9 126.9 128.1 128.1 127.4 126.3 126.4 125.9 127.1 126.5 127.1 126.5 126.5 126.0 126.5 126.0 125.5 126.3 126.3 127.0 127.0 126.8 126.8 125.8 125.8 129.3 126.8 126.9 128.7 127.8

Frequency (wavenumber) 1633 1627 1627 1612 1610 1600 1582 1604 1604 1624 1624 1626 1595 1580 1569 1593 1580 1593 1580 1588 1572 1588 1574 1554 1577 1577 1588 1588 1601 1588 1571 1569 1656 1590 1594 1631 1585

Observed O–N–O Frequency (degrees) (wavenumber) 127.7a 127.0 127.3 126.3 126.7 126.8 125.0b 125.8 125.6 127.4c 127.4 126.9 125.5d 125.9 124.7 126.7e 125.9 126.7 125.9 124.2f 124.3 125.5 124.3g 123.9 125.0g 124.6 126.2h 126.2 126.7i 126.0 124.9j 123.9 127.7a 125.7 124.8 127.3 125.3k

1618 1608 1608 1557 1557 1570 1533 1573 1573 1604 1604 1604 1563 1558 1523 1562 1562 1562 1562 1555 1531 1555 1546 1508 1542 1542 1565 1565 1576 1548 1556 1550 1627 1549 1565 1611 1568

a

Ref. 17. Ref. 18. Ref. 19. d Ref. 20. e Ref. 21. f Ref. 22. g Ref. 23. h Ref. 24. i Ref. 25. j Ref. 26. k Ref. 27. b c

groups are often coupled, multiple modes are possible through synchronization. For instance, with two nitro groups, the possibility exists that they are both stretching at the same time or that one is stretching while the other is compressing. Depending on the symmetry of the molecule, this can have a dramatic effect on the intensity of the band. When four nitro groups are present, four bands are theoretically possible, one with all four motions synchronized and three based on the remaining possibilities in which two pairs are in sync with each other. While a similar effect was seen with the asymmetric stretch, it is difficult to assign synchronization to that motion. Furthermore, a similar effect is noticed with the scissor motion. Scissor. The scissor or in-plane deformation involves APPLIED SPECTROSCOPY

1197

FIG. 1.

Nitro group O–N–O angle vs. asymmetric stretching frequency for secondary nitramines.

a change of the interior angle of the O–N–O atoms. Less attention has been paid to this motion compared to the stretching motions. When assignments are made they are frequently to several peaks or an area of the spectrum labeled ‘‘–NO2 skeletal motions’’ or ‘‘–NO2 other motions.’’ Before discussing the representative frequency range of the scissor vibration, mention of possible coupling is necessary. The calculations show that like the symmetric stretch, the scissor tends to couple with the N–X bond stretch. Furthermore, it can couple with C–C and C–N stretches in the backbone of the molecule, especially the symmetric stretching modes centered on the atom X to which the –NO2 group is bound. Upon examination of the characteristic frequencies, however, it appears that two separate ranges are involved because, unlike the symmetric stretch, which couples with the N–X stretch in only one way, the scissor motion usually couples with the N–X stretch in two ways with a distinct difference in energy. In the higher energy band normally assigned to the scissor motion, the interior angle of the nitro group decreases while the N–X bond compresses. This is the case for both C-nitro and N-nitro compounds. In the lower energy band, the interior angle decreases while the N– X bond stretches. However, the degree of coupling with the N–X bond stretching motion varies greatly from compound to compound. The higher range can be reasonably well defined by the current work and occurs at 757–895 cm21, with the majority of modes lying between 800–870 cm21. The lower energy coupling combination extends from 758 cm21 to beyond the lower bound of the spectral measurement of about 600 cm21, although most occur closer to 700 cm21. As such, the lower end of this range is not discussed here. However, as will be shown below, the frequencies for this motion can be very tightly bound1198

Volume 59, Number 10, 2005

ed within sets of compounds related by the nitro group environment. Out-of-Plane Deformation. The nitro group out-ofplane deformation is frequently grouped with the scissor motion when assigning peaks. In the present study, this motion may couple with other motions but frequently does not. When it does couple, it is most often with other nitro group out-of-plane motions or the bending motions involving heavy atoms in the backbone. It is slightly lower in energy than the upper scissor motion range and overlaps the range of the lower scissor motion. Within this study the range is 703–790 cm21. Interestingly, both of these modes appear in the gas-phase spectrum of nitrobenzene as a result of coupling with C–H rocking in the ring. The majority of these peaks occur below 750 cm21 and, in the nitrobenzene case, the peak at 703 cm21 is more intense in both the computation and the observed spectrum. A more typical range for this mode is 710 to 760 cm21. The out-of-plane deformation appears not to be dramatically affected by the electron withdrawing or donating characteristics of nearby moieties or the nitro group geometry. In fact, it seems to occur in a narrow range, as will be discussed further below. However, atoms or groups of atoms near the nitro group that inhibit the outof-plane flexion affect the frequency. As such, it might be expected that this band could be sensitive to the phase. DISCUSSION Patterns Within Similar Categories of Molecules. The largest and most closely related group of compounds in this study is the substituted mononitroaromatics (Table IV). The dinitrobenzenes belong to this group considering one of the nitro groups to be the substituent. The asymmetric stretch of the nitro group tends to couple with the

TABLE VIII. Monosubstituted nitrobenzene nitro asymmetric stretch peaks.a Compound 1,2-dinitrobenzene 1,3-dinitrobenzene 1,4-dinitrobenzene 2-nitrochlorobenzene 3-nitrochlorobenzene 4-nitrochlorobenzene 2-nitrofluorobenzene 3-nitrofluorobenzene 4-nitrofluorobenzene 2-nitrophenol 3-nitrophenol 4-nitrophenol 2-nitroaniline 3-nitroaniline 4-nitroaniline 2-nitromethoxybenzene 3-nitromethoxybenzene 4-nitromethoxybenzene 2-nitrotoluene 3-nitrotoluene 4-nitrotoluene 2-nitrobenzoic acid 3-nitrobenzoic acid 4-nitrobenzoic acid 2-nitroacetophenone 3-nitroacetophenone 4-nitroacetophenone a

Forepeak (observed) Main peak (observed) (cm21) (cm21) — 1601 w — 1590 m, 1582 m — 1602 m, 1583 m 1610 s N/A 1601 m 1620 s — 1602 s 1624 s, 1580 m 1631 w 1604 m 1608 m 1617 w 1597 s 1614 w — 1603 w — 1617 w 1609 w — 1615 w 1604 m

1563 1554 1565 1553 1551 1543 1551 N/A 1544 1546 1547 1539 1522 1544 1530 1548 1550 1538 1546 1546 1538 1556 1552 1547 1546 1549 1546

vs vs vs vs vs vs vs s s vs s vs vs vs vs vs m vs vs s s s s vs vs vs

vw 5 0%–10%, w 5 10%–30%, m 5 30%–60%, s 5 60%–90%, vs 5 90%–100%.

various ring motions. The calculated and observed spectra reveal that this coupling leads in some instances to one combined peak and in others to two peaks separated by no more than 80 cm21. Herein, the (usually less intense) higher energy peak is referred to as the forepeak in Table VIII. As anticipated, more electron withdrawing groups (–NO2, –Cl, –F) on the ring tend to increase the frequency of the –NO2 asymmetric stretching vibration. Another pattern observable in the main peaks is that the frequency shifts to higher energy for the meta- nitro substitution compared to para- and, in general, ortho-nitro substitution. Some of the peaks of ortho-substituted compounds shift to higher energy compared to para-substitution, and this appears to depend on the degree to which the –NO2 group is forced out of the ring plane, as explained by Bellamy.1 This pattern is broken, however, by meta-dinitrobenzene, wherein the main peak is shifted to lower energy. The forepeaks of the dinitrobenzenes also break the pattern in that 1,3-dinitrobenzene has a forepeak and a main peak while the 1,2 and 1,4 isomers have only a main peak. For the rest of the compounds the ortho- and para- substituted compounds have the more pronounced forepeak. The forepeaks in 3-nitroaniline and 3-nitromethoxybenzene are weak compared to those of the 2and 4-substituted compounds. The pattern of absorptions can be explained using the concept of meta or ortho/para directing substituents during nucleophilic attack on the ring, which is based on resonance differences. –NH2, –OH, –OCH3, –alkyl, and –halogen are ortho/para directing groups, while –NO2 is a meta directing group. If meta directing groups produce the separation of the coupled modes just discussed, then they would push the main

peak to a relatively lower frequency and cause the appearance of a forepeak selectively. Conversely, ortho/ para directing groups would cause the splitting selectively when ortho or para to the nitro group. When a meta directing group is ortho or para to the nitro group, it should have little effect. Likewise, when an ortho/para directing group is meta to the nitro, it should have little effect. This notion is consistent with the results in Table VIII. Unfortunately, only one meta directing group was available for comparison; however, additional support is available4 in the spectra for the three isomers of nitrobenzene having two additional meta directing groups, e.g., –C(O)OH and –C(O)CH3. The meta directing groups do not significantly reduce the frequency of the main peak in the compounds having the meta nitro groups. Also, a forepeak is observed for both compounds having para nitro groups. However, the splitting in peaks for the 3-nitro compounds is far more pronounced than the others and the 2-nitro compounds are missing a forepeak, as predicted. The concept in the previous paragraph has not previously been proposed. Although a few contradictions exist, the general pattern is consistent. The double peaks are a result of coupling of the –NO2 asymmetric stretch with the benzene ring motions. Also, the influence of the position-directing groups is based on the effect of different resonance contributions. Therefore, it seems possible that, by altering the role of the resonance contribution, the energies of the various possible benzene ring motions are changed, causing the observed splitting. Inspection of the computer-optimized geometries of the various benzene rings reveals variations in the C–C bond lengths, but they do not follow any direct pattern based on substitution. The symmetric nitro stretch of the nitroarenes also follows a pattern. In general, the symmetric stretch frequency of the 2-nitro derivatives is slightly greater than the 3-nitro derivative, which is, in turn, slightly greater than the 4-nitro derivatives. The differences of 1 and 7 cm21 are small. The position of the absorption seems to be only slightly affected by the electron-withdrawing or -donating characteristics of the substituent. Within this study, the dinitrobenzene absorptions are, on average, the lowest (1347 cm21) and the nitrotoluenes are the highest (1357 cm21), but the difference is small. Exceptions to these patterns include 2-nitrophenol and 2-nitroaniline, where the orientation of the nitro group clearly indicates intramolecular hydrogen bonding. In the case of 2-nitrophenol, this has been previously noted 1,2 and produces a broad and shifted O–H stretching mode in the gas-phase spectrum. A pattern in the nitro scissor frequencies is evident for the monosubstituted nitrobenzenes as well. The 3-nitro category stands out once again. The 2- and 4-nitro groups exhibit a single peak relating to the scissor motion; however, the scissor of the 3-nitro group tends to couple with a higher frequency ring motion (two pairs of oscillating scissor motions) resulting in two peaks. The higher of these is generally of higher frequency than the single peak exhibited by the 2-nitro and 4-nitro compounds. The 3-nitrochlorobenzene is an exception perhaps due to coupling of the C–Cl stretch and the nitro scissor motion. APPLIED SPECTROSCOPY

1199

TABLE IX. DNT nitro asymmetric stretch peak summary.a Compound

Nitro group

Meta directed?

Ortho/Para directed?

Forepeak (cm21)

Main peak (cm21)

2,3-DNT

2 3 2 4 2 5 2 6 3 4 3 5

No No Yes Yes No No Yes Yes No No Yes Yes

Yes No Yes Yes Yes No Yes Yes No Yes No No

— — 1607 w 1602 w — — 1614 w 1614 w — — 1617 (calc) 1617 (calc)

1564 vs 1564 vs 1514 vs 1514 vs 1542 vs 1542 vs 1527 vs 1527 vs 1564 vs 1554 vs 1556 (calc) 1556 (calc)

2,4-DNT 2,5-DNT 2,6-DNT 3,4-DNT 3,5-DNT a

vw 5 0%–10%, w 5 10%–30%, m 5 30%–60%, s 5 60%–90%, vs 5 90%–100%.

No trend in the frequencies based on the characteristics of the substituents themselves is discernable. For the out-of-plane deformation motion within this group of compounds, the 3-nitro compounds again stand out. In general, the 2-nitro and 4-nitro frequencies are similar and the 3-nitro frequency is slightly lower. Variations in the frequency resist explanation based on substituent properties alone. For the dinitrotoluenes (DNT, Table V) and the trinitrotoluenes (TNT, Table VI), only the position of the methyl substituent was considered. Similar to the mononitroaromatics already discussed, some, but not all of the dinitrotoluenes (DNT) and trinitrotoluenes (TNT) show forepeaks slightly higher in frequency than the main nitro asymmetric stretch. The concept explaining the forepeaks in the previous section was applied to these compounds, although the analysis is more complex because each nitro group is potentially affected by other nitro groups and the methyl group. For the DNTs, the concept seems to capture most of the observed trends. That is, the asymmetric stretch of a nitro meta to a meta director and/or ortho or para to an ortho/para director exhibits splitting, resulting in a forepeak and a somewhat reduced frequency main peak. The influences on each of the nitro groups in the DNTs are summarized in Table IX. Clearly 2,4-DNT and 2,6-DNT stand out where both nitro groups are meta to a meta director and either ortho or para to an ortho/para director. In both cases, prominent forepeaks appear and the main peak is somewhat reduced in frequency compared to the other DNTs. In the case of 2,4-DNT, where one nitro is ortho and the other para to the ortho/para director, a distinct separate forepeak for each nitro group is predicted by the calculation and observed. No IR spectrum was available for 3,5DNT, but the calculation predicts a forepeak. The nitro group in the 2-position of 2,3-DNT and 2,5-DNT, plus the nitro group in the 4-position of 3,4-DNT should be influenced by the methyl group, but appear not to be. No explanation is obvious, although perhaps the meta directing influence is more significant than ortho/para directing influences. Another explanation might be that being ortho or para to a strong meta director nullifies the effect of being ortho or para to a weak ortho/para director. The spectral patterns become even cloudier for the TNTs where coupling and steric effects increase. Never1200

Volume 59, Number 10, 2005

theless, data in Table X suggest that the concept just discussed still applies to some degree. The 4-NO2 of 2,4,6TNT contributes to two forepeaks according to the calculation and both are observed. Each nitro group in 1,3,5TNB is meta to two meta directors and forepeaks are again predicted by the calculation and observed. 2,3,4TNT seems to follow the pattern once it is realized that the peak at 1606 cm21, which might be considered to be a forepeak, is actually the asymmetric stretch of the 3nitro group. The significant shift to higher frequency in this case may be explained by the fact that the 3-nitro group is forced significantly out of the plane of the ring. 2,3,5-TNT and 2,3,6-TNT do not, however, show the predicted pattern of forepeaks. Furthermore, the 5-nitro group in 2,4,5-TNT would not, according to the concept, participate significantly in any forepeaks, but, according to the calculation, it does. Both the DNTs and TNTs show narrow ranges of 1348–1358 cm21 and 1343–1353 cm21, respectively, for the nitro symmetric stretch. Patterns within these narrow ranges are not obvious. Patterns within the scissor and out-of-plane deformation modes are equally ambiguous. Comparisons between N–NO2 and non-aromatic C– NO2 compounds reveal only one noteworthy feature. All of the nitramine compounds in this study except RDX and HK-55 display a small peak at 760–767 cm21. This is attributed to the out-of-plane deformation of the nitramino group. Considering the number of compounds involved, this range is extremely narrow. RDX has peaks on either side of this range that may obscure a minor peak and HK-55 has a peak at 755 cm21. The C-nitro groups including those from the mixed nitro/nitramino compounds do not have this distinctive peak. Patterns Within Similar Nitro Group Environments. Informative trends and patterns are found for closely related nitro groups, even when the remainder of the molecule is quite different. The ambiguity in the peak assignment is reduced by correlations between calculated motions and observed peaks. The aromatic nitro compounds were analyzed in the preceding section. The nitro asymmetric stretch of aliphatic nitro compounds was affected by the electron-donating or -accepting properties of other a-carbon substituents, coupling with other bands, and even hydrogen bonding. It was found that a b-carbonyl dramatically increases the frequency of this mode. Examples of this are HK-55 (11), HK-56 (12), K-55 (13), K-56 (14), and RDX-k (3). RDX-k can be directly com-

TABLE X. TNT nitro asymmetric stretch peak summary. Compound 2,3,4-TNT

Nitro group

Meta directed?

Ortho/Para directed?

Forepeak (cm21)

2 3 4 2 3 5 2 3 6

Yes No Yes No Yes Yes Yes No Yes

Yes No Yes Yes No No Yes No Yes

1596 w — 1564 vs — 1632 w 1632 w — — —

2,4,5-TNT

2 4 5

Yes Yes No

Yes Yes No

2,4,6-TNT

2 4 6

Twice Twice Twice

Yes Yes Yes

1,3,5-TNB

1 3 5

Twice Twice Twice

No No No

1614 m 1604 m 1614 m, 1604 m 1617 w 1617 w, 1602 w 1617 w 1623 m 1623 m 1623 m

2,3,5-TNT 2,3,6-TNT

a

Main peak (cm21) 1546 1606 1546 1560 1548 1544 1546 1546 1546 1533 1547 1547 1557

vs w vs s vs vs vs vs vs, s vs vs vs

1539 vs 1539 vs 1539 vs 1546 vs 1546 vs 1546 vs

vw 5 0%–10%, w 5 10%–30%, m 5 30%–60%, s 5 60%–90%, vs 5 90%–100%.

pared with RDX (2) to see this trend, which has not been noted previously. HNDZ (8), DNNC (9), TNAZ (15), and TNCB (16) contain the geminal-dinitro functionality in a ring. The nitro scissor motion lies in the narrow range of 842–846 cm21. If TNCB is left out as the only compound that does not also contain a nitramino group, the range narrows to 842–844 cm21. Examination of the scissor motion of a closely related functionality is revealing. BDNFEC (23) and FEFO (25) both contain the fluorodinitromethyl group. Again, a narrow frequency range of 851–852 cm21 is found for the scissor motion. BTNEC is the only compound with a trinitromethyl group. However, the IR spectra are available for 2-[bis(2,2,2-trinitroethoxy)methoxy]1,1,1-trinitroethane (TNEOF) and 2-[tris(2,2,2-trinitroethoxy)methoxy]-1,1,1-trinitroethane (TNEOC). The IR spectrum of 3,3,3-trinitrobutyric acid,16 bis(2,2,2-trinitroethyl)formal (TEFO), and 1,1,2,2-tetrakis(2,2,2-trinitroethoxy)ethane (DITEFO) 28 are also available. All six compounds have a peak at or near 856 cm21 that is attributable to the nitro scissor. CONCLUSION Density functional calculations at the B3LYP level with a 6-311G* basis set were conducted for more than fifty nitro compounds. This enabled molecular motions to be assigned to observed spectral bands with reasonable surety. The symmetric and asymmetric stretches have received attention before but are better understood by this work, while the bending modes, which have been difficult to assign previously, are discussed in detail. New patterns in the nitro group modes are revealed. First, the influence of the neighboring group, X, molecular motions on the –NO2 group is apparent. The previously noted relation between the electron donor–acceptor properties of a-carbon substitutions and the nitro asymmetric stretch is clearly supported. It was found that the two –NO2 stretching motions coupled the same way in terms of simultaneous N–X bond compressions in all

cases studied. It was found that the strength of the N–X bond influences the frequency of the coupled symmetric stretch. The –NO2 scissor motion was also found to couple with the N–X bond stretch. In this case, both possible combinations of N–X stretch and O–N–O angle compressions were found, and it was determined that they occur in two different regions of the IR spectrum. The upper range was clearly identified. However, the lower bound of the lower range extends below the MIR band pass available such that it could not be specified. Second, the scissor motion frequencies are extremely insensitive to the nitro group environment. Specifically, geminal-dinitro methylene groups, both in cyclic and acyclic environments, trinitromethyl groups, and a-substituted geminal-dinitromethyl groups all showed characteristic peaks at 842–846 cm21. The insensitivity of these peaks to differences in the local chemical environment may enable them to be used as tags for explosive detection and identification. Third, splitting of the bands associated with the coupled motions of the nitro asymmetric stretch and the aromatic ring motions of nitroaromatic compounds was detected. The patterns reflect the resonance concept embodied in meta and ortho/para directing subtituents in mono and dinitro aromatic compounds. The pattern for the isomers of trinitrotoluene is less clearly evident. Fourth, most nitramines have an absorption at 760– 767 cm21 from the out-of-plane deformation of the nitro group. The tightness of this range is useful for identifying the presence of the nitramino group in the explosive detection field. 29 1. L. J. Bellamy, Advances in Infrared Group Frequencies, R. Clay, Ed. (The Chaucer Press, Ltd., London, 1968). 2. H. Gunzler and H.-U. Gremlich, IR Spectroscopy, An Introduction (Wiley-VCH, Weinheim, Germany, 2002). 3. T. B. Brill, Prog. Energy Combust. Sci. 18, 91 (1992). 4. NIST Mass Spec Data Center, S. E. Stein, Director, ‘‘Infrared Spectra,’’ in NIST Chemistry WebBook, NIST Standard Reference Database Number 69, P. J. Linstrom and W. G. Mallard, Eds. (National

APPLIED SPECTROSCOPY

1201

5. 6. 7. 8. 9. 10. 11.

12. 13.

Institute of Standards and Technology, Gaithersburg, MD, 20899, March 2003), http://webbook.nist.gov. M. A. Thompson and M. C. Zerner, J. Am. Chem. Soc. 113, 8210 (1991). M. A. Thompson, E. D. Glendening, and D. Feller, J. Phys. Chem. 98, 10 465 (1994). M. A. Thompson and G. K. Schenter, J. Phys. Chem. 99, 6374 (1995). M. A. Thompson, J. Phys. Chem. 100, 14 492 (1996). J. J. P. Stewart, J. Comp. Chem. 10, 209 (1989). J. J. P. Stewart, J. Comp. Chem. 10, 221 (1989). Gaussian 98, Revision A.7, M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M. A. Robb, J. R. Cheeseman, V. G. Zakrzewski, J. A. Montgomery, Jr., R. E. Stratmann, J. C. Burant, S. Dapprich, J. M. Millam, A. D. Daniels, K. N. Kudin, M. C. Strain, O. Farkas, J. Tomasi, V. Barone, M. Cossi, R. Cammi, B. Mennucci, C. Pomelli, C. Adamo, S. Clifford, J. Ochterski, G. A. Petersson, P. Y. Ayala, Q. Cui, K. Morokuma, D. K. Malick, A. D. Rabuck, K. Raghavachari, J. B. Foresman, J. Cioslowski, J. V. Ortiz, A. G. Baboul, B. B. Stefanov, G. Liu, A. Liashenko, P. Piskorz, I. Komaromi, R. Gomperts, R. L. Martin, D. J. Fox, T. Keith, M. A. AlLaham, C. Y. Peng, A. Nanayakkara, M. Challacombe, P. M. W. Gill, B. Johnson, W. Chen, M. W. Wong, J. L. Andres, C. Gonzalez, M. Head-Gordon, E. S. Replogle, and J. A. Pople (Gaussian, Inc., Pittsburgh, PA, 1998). J. B. Foresman and A. Frisch, Exploring Chemistry with Electronic Structure Methods (Gaussian Inc., Pittsburgh, PA, 1996). S. Portmann and H. P. Lu¨thi, CHIMIA 54, 766 (2000).

1202

Volume 59, Number 10, 2005

14. T. Bally, R. Olkhov, E. Haselbach, and A. Schmelzer, Helv. Chim. Acta 54, 1299 (1971). 15. R. W. Beal, Ph.D. Dissertation, University of Delaware, Newark, Delaware (2004). 16. Y. Oyumi and T. B. Brill, Prop. Explos. Pyrotech. 11, 35 (1986). 17. R. D. Gilardi, NRL, private communication, 1997. 18. C. S. Choi and E. Prince, Acta Crystallogr., Sect. B 28, 2857 (1972). 19. H. Ritter, S. Braun, M. Schafer, H. R. Aerni, H. R. Bircher, B. Berger, J. Mathieu, and A. Gupta, International Annual Conference of ICT (2001), 32nd (Energetic Materials), 91/1–91/14. 20. Y. Oyumi, A. L. Rheingold, and T. B. Brill, J. Phys. Chem. 91, 920 (1987). 21. C. S. Choi and H. P. Boutin, Acta Crystallogr., Sect. B 26, 11235 (1970). 22. T. B. Brill, R. J. Karpowicz, T. M. Haller, and A. L. Rheingold, J. Phys. Chem. 88, 4139 (1984). 23. Y. Oyumi and T. B. Brill, J. Phys. Chem. 90, 2526 (1986). 24. H. L. Ammon, R. D. Gilardi, and S. K. Bhattacharjee, Acta Crystallogr., Sect. C 39, 1680 (1983). 25. Y. Oyumi, T. B. Brill, A. L. Rheingold, and T. M. Haller, J. Phys. Chem. 89, 4317 (1985). 26. M. Pickering, J. Rylance, R. W. H. Small, and D. Stubley, Acta Crystallogr., Sect. B 47, 782 (1991). 27. T. G. Archibald, R. D. Gilardi, K. Baum, and C. George, J. Org. Chem. 55, 2920 (1990). 28. Y. Oyumi and T. B. Brill, Combust. Flame 65, 103 (1986). 29. E. C. Mattos, E. D. Moreira, R. C. L. Dutra, M. F. Diniz, A. P. Ribeiro, and K. Iha, Quim. Nova 27, 540 (2004).