Systematic Interpretation of Raman Spectra of
`Organic Compounds
`IV-Nitrogen Compounds
`
`T. Visser and J. H. van der Maast
`Laboratory for Analytical Chemistry, University of Utrecht, Croesestraat 77A, 3522AD Utrecht, The Netherlands
`
`The Raman spectra of 79 nitrogen containing organic compounds have been rec6rded. The frequencies of
`amines, pyridines, cyanides, amides and nitro-compounds are reported and correlations with the structure have
`been established by means of computer searching. The collected data have been incorporated into an
`interpretation system for C, H and 0 compounds, and difficulties in the interpretation of the data are reviewed.
`A further extension of the system is discussed.
`
`INTRODUCTION
`
`Systematic interpretation of vibrational spectra has
`proved to be very useful, especially for a beginner.'
`Interpretation procedures for Raman spectra of organic
`P
`compounds have been develo ed up till now by
`Schrader and Meier,' and by us.
`The most recent system3 is suitable for a variety of C,
`H and 0 compounds, and as the results are quite satis-
`fying we decided to extend it to nitrogen compounds.
`As a start a limited number of classes of compounds
`has been selected, viz amine-, pyridine-, cyanide-,
`amide-, and nitro compounds.
`It was expected'that the nitrogen compounds would
`not require too many modifications to the CHO system.
`Depending on the results a furthzr extension would be
`considered.
`
`EXPERIMENTAL
`
`Raman spectra were recorded on a Spectra-Physics
`model 700 spectrophotometer equipped with a Spec-
`tra-Physics model 165 argon ion laser as a lightsource.
`The 488 nm line was used as excitation wavelength with
`a continuous output power of 200mW. The spectro-
`photometer was calibrated daily on indene, resulting in
`an overall wavenumber accuracy of *2 cm-' (readabil-
`ity included). Spectra were run in two parts, i.e. from
`0-2000 cm-' and from 2000-4000 cm-'. Sensitivity
`was adjusted in such a way that the most intense peak
`in the spectral region had an intensity between 85 and
`95% (scale divisions). Intensity is indicated as weak
`(02-20%), medium (20-50'/0)
`and strong (50-95%).
`Further.details can be found in Ref. 1.
`
`SPECTRAL DATA
`
`The nitrogen atom can be present in a number of
`structural elements such as amine, amide etc., which
`
`t Author to whom correspondence should be addressed.
`
`may be correlated with certain bands in the Raman
`spectra.
`To find the correlations a computer program INTVA
`has been developed. This programme searches in any
`preselected file of spectral data for frequency intervals
`that all compounds have in common. A file must be
`composed of compounds having (at least) one common
`structural element. Limits for (minimum/maximum)
`intensity and (maximum) interval width can be set at
`will.
`As all possible intervals are found, those that are not
`clearly correlated with the structural element should be
`omitted.
`The results are discussed successively below:
`
`Amines
`
`In the Raman spectra of all primary and secondary
`amines at least one weak band shows up in the region
`3442-3211 cm-',
`caused by
`the N-H
`stretching
` vibration(^).^ A second weak band is observed for all
`primary amines in the same region. Dollish et d4 refer
`to these bands as being strong. The discrepancy is due
`to the fact that these authors use an indistinct intensity
`criterion. This once more emphasizes that one should
`use well defined scanning conditions.
`The C-N stretching vibrations in tertiary amines6 do
`have bands in the region 1070-1031 cm-' and 871-
`841 cm-'. However these intervals are hardly useful for
`interpretation purposes as other structural elements
`have peaks in these regions as well.
`In the class of cyclic aliphatic amines, piperidines can
`be detected from v(N-H) around 3330 cm-' and two
`ring skeletal vibrations between 990 and 730 cm-'.'
`Pyrrolidines show a strong skeletal breathing vibration,
`already mentioned by Tschamler and Voetter,' in the
`range 903-872 cm-'. The ring breathing vibrations of
`normal aliphatic five membered ring is found in about
`the same region. For primary pyrrolidines v(N-H)
`lies
`between 3303 and 3297 cm-'.
`the so called
`In
`the
`region 2804-2735cm-'
`Bohlmann bands are characteristic for the majority of
`the aliphatic and alicyclic amines. The appearance of
`these bands is ascribed as a combined effect of Fermi
`resonance and at least one C-H bond trans to the lone
`
`CCC-0377-0486/78/0067-0278 $02.00
`278 JOURNAL OF RAMAN SPECTROSCOPY, VOL. 7, NO. 5, 1978
`
`@ Heyden & Son Ltd, 1978
`
`Merck Exhibit 2235, Page 1
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`

`

`SYSTEMATIC INTERPRETATION OF RAMAN SPECTRA OF ORGANIC COMPOUNDS. IV: NITROGEN COMPOUNDS
`
`pair of the nitrogen atom. No bands show up in case the
`lone pair is conjugated.'
`A methyl group attached to the nitrogen atom can be
`recognized from a band in the range 2815-2775 cm-'.
`Methylene to nitrogen shows a useful band around
`1465 cm-'.
`In /3-olefinic amines neither v(N-H) nor v(C=C)
`has been shifted compared with aliphatic v(N-H) and
`olefinic v(C=C). The same holds for the amine group
`and the triple bond.
`Primary anilines display very characteristic bands in
`the regions 3387-3348 cm-', vas(N-H), and 3229-
`(less intensive). For secondary
`321 1 cm-', v,(N-H),
`anilines" only one band is observed in the region
`3402-3398 cm-'.
`
`Pyridines
`
`Pyridines show two rather characteristic v(C=C) bands
`in the interval 1599-1560 cm-', somewhat lower than
`benzene derivatives." They are useful for identification
`purposes. Detection of substitution of
`the pyridine
`nucleus is very difficult although at least one intensive
`band
`is present
`in one of
`the regions 1073-
`1030 cm-', 1003-988 cm-' and 786-715 cm-'. Too
`few compounds have been studied so far to draw a
`reliable conclusion.
`
`Amides
`
`All aliphatic amides are characterized by a carbonyl
`attached to a nitrogen atom. These compounds show
`two peaks which are useful for identification, v(C=O)
`in the region 1669-1623 cm-' (medium intensity) and a
`v(C-N)
`in
`the range 1431-1405cm-'
`(weak to
`m e d i ~ m ) ~ . The v(N-H) of primary and secondary
`amides is found as a weak band(s) in the interval 3348-
`3200 cm-'.
`
`Nitro compounds
`
`Much work on aliphatic nitro compounds has been
`done by Geiseler et a1.12*13. The nitro group is charac-
`terized by its asymmetric N-0
`stretching vibration
`between 1560 and 1-54! cm-' (weak) and its symmetric
`one at 1386-1358 cm- (strong to medium). Difference
`between primary and secondary compounds can be
`made through the more precise values in Table 1.
`Some naphthyl compounds show a strong v(C=C)
`band in the same region but can be detected from other
`v(C=C) bands around 1600 cm-'.
`Aromatic nitro compounds show a weak va,(N-0)
`in the interval 1522-1512 cm-' and a strong v,(N-0)
`in the range 1348-1336 cm-'. The data compiled from
`the recorded spectra by means of
`the programme
`INTVA are presented in more detail in Table 1.
`
`Cyanides
`
`Cyanides show a very intensive v ( C r N ) in the range
`(Ref. 15) where also non terminal
`2246-2219cm-'
`acetylenes might show a band. In aliphatic nitriles
`is found in the region 2246-2234 cm-',
`v(C=N)
`whereas in aromatic nitriles it is found between 2234
`and 2219 cm-'.
`
`THE INTERPRETATION SYSTEM
`
`In extending an interpretation system to a new class of
`compounds two stages can be distinguished:
`(A) adaptation of the existing system such that the new
`compounds do not produce wrong answers,
`(B) addition of the new I(nformation)-elements.
`
`Table 1. Nitrogen correlations. The frequency intervals include the given wavenumbers
`aliphatic -NH2
`3392-3360102
`3342-3302102
`aromatic -NHz
`3387-3348102
`3229-321 1 102
`aliphatic -NH-
`3345-3300102
`aromatic -NH-
`3404-3398102
`piperidine NH
`3346-3321102
`pyrrolidine NH
`3303-3297102
`-C-N-C-
`2788-2706120
`CH3-N-
`281 5-2775102
`-CH2-N-
`1484-1476102
`-CHz-N-CHz-(or
`2805-2794102
`formamide
`1669-1656130
`pivalamide
`1635-1623110
`prim. amide
`3348-3300102
`sac. amide
`3342-3200102
`tert. amide
`1647-1635110
`pyridine
`1599-1 560102
`
`CH3)
`
`1606-1 600130
`2684-2656102
`973-965102
`
`1 130-1 109102
`1480-1454102
`1 41 7-1 41 01 1 5
`1413-1407102
`31 72-31 50102
`1669-1 62311 0
`1413-1409110
`1276-1 210115
`
`990-9 1 21 1 0
`903-872150
`
`1068-1 0301 15
`
`1674-1650110
`1419-1405102
`
`1073-1 036130
`1579-1 558102
`
`881-730140
`
`143 1-1 41 01 1 0
`1252-1 246125
`
`786-715150
`1003-988140
`
`aliphatic - C r N
`aromatic - C r N
`prim. aliphatic NOz
`sec. aliphatic NOz
`aromatic -NOz
`
`2246-2234150
`2234-2219150
`1560-1 551 102
`1555-1 544102
`1522-1 51 2102
`
`1602-1 597120
`13a6-13ao125
`1376-1 358125
`1348-1336150
`
`91 6-896102
`865-841 102
`813-787125
`
`@ Heyden & Son Ltd, 1978
`
`JOURNAL OF RAMAN SPECTROSCOPY, VOL. 7, NO. 5, 1978 279
`
`Merck Exhibit 2235, Page 2
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`

`

`T. VISSER AND J. H. VAN DER MAAS
`
`In stage (A) Q(uestion)-elements are modified and
`added to overcome wrong answers. It is the aim to
`preserve all existing information in the system, but it
`may be necessary to sacrifice an I-element.
`The 79 nitrogen compounds have been passed
`through the CHO system; 16 wrong reports were
`obtained. Addition of 6 Q- and the removal of 1 I-
`element proved to be sufficient to overcome these
`mistakes. Then the system was suitable for the N-
`compounds though it did not yet establish any specific
`nitrogen information.
`The aim of stage(B) is to add as much as possible new
`structural elements. This can be reached by adding
`Q-elements based upon specific intervals (selecting all
`compounds with a certain functionality and no others),
`all
`pseudo-specific
`intervals
`(selecting
`almost
`compounds and no others), and selective intervals
`(selecting all compounds with some others) in that
`order.
`As all useful intervals for the N-functionalities have
`been found already (Table 1) only those which are
`(pseudo)specific have to be traced. A programme
`SELEC has been developed to do so.
`From the intervals of the 21 N-functionalities only
`one proved to be specific (aromatic nitro compounds)
`and two pseudo-specific (primary aliphatic amines and
`primary anilines), regarding the total file of 606
`compounds. Hence the addition of selective intervals
`was inevitable. However this brought about wrong
`answers for some CHO compounds.
`There are two possibilities to overcome this: (1) omit
`information, and (2) making use of pre-information to
`by-pass certain Q-elements (and thus I-elements). On
`applying both possibilities it appeared that the system
`could be extended with 20 I-elements at the cost of 49
`new Q’s.
`In Table 2 the values of the number of different 1 ’ s
`(XI) and Q’s (XQ) and the ratio ZQ/XI have been
`summarized in the different stages of incorporation.
`Although the ratio W / Z I (being a measure for the
`efficiency of a system) did not rise very much, 2.0 to
`2.1, the (sub)value 2.9 for the added Q’s and 1 ’ s is
`rather high.
`To get insight in the reliability of the system two tests
`have been done. In the first test all compounds have
`been passed through the system in which each interval
`was broadened with 2cm-‘ to both sides. As can be
`
`Table 2. Values of the total number of different information-
`elements (XI), question-elements (ZQ) and the
`ratio ZQ/ZI, in the three stages of the incorpora-
`tion
`zam
`2.0
`2.1
`2.1
`2.9
`
`I0
`321
`327
`376
`55
`
`I I
`159
`158
`178
`19
`
`CHO system
`Stage (A)
`Stage (B)
`Increase
`
`seen in Table 3 the total number of collected I-ele-
`ments (Ito,) increased with 93 and the percentage
`wrong answers with 1.4%. In the second test the
`minimum intensity limit was set at 10% for all intervals.
`I,,, decreased from 3259 to 2655 whereas the percen-
`tage wrong answers increased from 0.5 to 2.5%. These
`figures indicate that the system is reliable.
`
`GENERAL REMARKS
`
`The extension of an interpretation system with a new
`class of compounds is usually attended with the intro-
`duction of new functionalities. As each functionality is
`allied with at least one frequency interval, the chance
`on
`interference
`increases progressively with
`the
`number of functionalities. There are two possibilities to
`prevent such interference, both coupled with some
`disadvantages;
`(1) addition of still more Q-elements to by-pass, on
`condition that they do not interfere,
`(2) removal of I-elements (and the allied Q’s).
`The disadvantage of (1) is that with the growth of XQ,
`the chance of mistakes increases, and thus the reli-
`ability decreases. At the same time the system becomes
`less surveyable and consequently harder to modify. The
`disadvantage of possibility (2) is that the information
`content of the system decreases.
`We intend to extend the systemt but in view of the
`above mentioned remarks it seems useful to set out
`rules and procedures first. The development of INTVA
`and SELEC should be regarded as a step in that direc-
`tion.
`t Up till now the system contains 376 Q- and 178 I-elements. A
`FORTRAN IV programme of the system is available on request.
`
`Table 3. Influence of intensity and wavenumber variation. Values of the numbers of collected I-elements (I,,,,), the average
`numbers of I,,, per compound (Ilot/n) and the percentage wrong answers, split up for the different chemical classes.
`(a) The CHON system, (b) the system with intervals broadened with 2 cm:’ to both sides and (c) the system with the
`minimum intensity limit of each interval set at 10%
`CH
`1 67
`88 1
`5.3
`0.6
`898
`5.4
`2.2
`702
`4.2
`3.1
`
`n
`
`1tat
`
`ltatJn
`% fault
`
`Itat
`ItotJn
`% fault
`
`Itat
`ItotJn
`% fault
`
`OH
`122
`656
`5.4
`0.2
`680
`5.6
`1.5
`509
`4.2
`1.2
`
`-0-
`67
`364
`5.4
`0.3
`379
`5.7
`2.4
`295
`4.4
`2.4
`
`C=O
`49
`281
`5.7
`0.7
`282
`5.8
`2.1
`243
`5.0
`1.7
`
`COOH
`22
`78
`3.5
`2.6
`79
`3.6
`3.8
`77
`3.5
`5.2
`
`COOC
`100
`521
`5.2
`0.8
`537
`5.4
`1.5
`465
`4.7
`1.7
`
`N
`79
`478
`6.0
`0.4
`493
`6.2
`1.8
`364
`4.6
`3.8
`
`total
`606
`3259
`5.4
`0.5
`3352
`5.5
`1.9
`2655
`4.4
`2.5
`
`Number of compounds
`
`Standard system
`
`intervals broadened
`
`Minimum intensity 10%
`
`280 JOURNAL OF RAMAN SPECTROSCOPY, VOL. 7, NO. 5, 1978
`
`@ Heyden & Son Ltd. 1978
`
`Merck Exhibit 2235, Page 3
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`

`

`SYSTEMATIC INTERPRETATION OF RAMAN SPECTRA OF ORGANIC COMPOUNDS. IV: NITROGEN COMPOUNDS
`
`REFERENCES
`
`1. J. H. v. d. Maas and T. Visser, J. Raman Specfrosc. 2, 563
`( 1974).
`2. 8. Schrader and W. Meier, Z. Anal. Chem. 275, 177 (1975).
`3. T. Visser and J. H. v. d. Maas, J. Raman Specfrosc. 3, 125
`(1978).
`4. F. R. Dollish, W. G. Fateley and F. F. Bentley, Chefacferisfic
`Raman Frequencies of Organic Compounds, Wiley, New
`York (1974).
`5. H. Wolff and H. Ludwig, Eer. Bunsenges. Phys. Chem. 71,
`1107 (1967).
`6. P. H. Clippard and R. C. Taylor, J. Chem. Phys. 50, 1472
`(1969).
`7. D. Vedal, 0. H. Ellestad, P. Klaboe and G. Hagen, Spec-
`frochim. Acfa 32,877 (1976).
`8. H. Tschamler and H. Voetter, Monafsch. Chem. 83, 302
`(1952).
`9. E. E. Ernstbrunner and J. Hudec, J. Mol. Sfruct. 17, 249
`(1973).
`
`10. A. Perrier-Datin and J. M. Lebas, J. Chem. Phys. Physi-
`cochim. Biol. 69, 591 (1 972).
`17. J. H. S. Green, D. J. Harrison, W. Kynaston and H. M. Paisley,
`Specfrochim. Acfa 26,2139 (1 970).
`12. G. Geiseler and H. Kessler, Ber. Bunsenges. Phys. Chem. 68,
`571 (1964).
`13. G. Geiseler, H. Kessler and J. Fruwert, Ber. Bunsenges.
`Phys. Chem. 70,918 (1966).
`14. J. V. Shukla and K. N. Upadhya, Indian J. Pure Appl. Phys. 7.
`830 (1969).
`15. K. Kumar, Specfrochim. Acfa 28,459 (1972).
`
`Received 7 May 1978
`@ Heyden & Son Ltd, 1978
`
`@ Heyden & Son Ltd, 1978
`
`JOURNAL OF RAMAN SPECTROSCOPY, VOL. 7, NO. 5, 1978 281
`
`Merck Exhibit 2235, Page 4
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`