AUGUST 1967
`
`HYDROPHOBIC BONDING
`INTRAMOLECULAR
`
`2583
`
`TABLE I11
`
`DIMETHYLNITROPHENYLSULFONIUM AND -SELENONIUM METHYL SULFATE SALTS
`%-
`Yield,
`MP,
`---Carbon,
`%
`OC
`Calod
`Found
`. . .
`85
`160-162a
`36.60
`
`Formula
`CoHisNSiOs
`
`78
`
`92
`
`63
`
`83
`
`93
`
`142-143b
`
`C~H13NS206
`
`36.60
`
`157-158'
`
`C ~ H I ~ N S ~ O ~ 36.60
`
`156-158
`
`153-157
`
`163-165
`
`C 9H i3X S SeOe
`
`31.58
`
`C9H~3NSSe06
`
`31.58
`
`CeH13NSSeOs
`
`31.58
`
`Compound
`Dimethyl-o-nitrophenylsulfonium
`methyl sulfate
`Dimethyl-m-nitrophenylsulfonium
`methyl sulfate
`Dimethyl-pnitrophenylsulfonium
`methyl sulfate
`Dimethyl-o-nitrophenylselenonium
`methyl sulfate
`Dimethyl-m-nitrophenylselenonium
`methyl sulfate
`Dimethyl-p-nitrophenylselenoninm
`methyl sulfate
`a Lit. mp 155-157', K. Brand and 0. Stallmann, Ber., B54, 1578 (1921).
`b Lit. mp 140-141', K. Brand and 13. W. Leyerzapf,
`ibid., B70, 284 (1937). <Lit. mp 157-158.5', F. G. Bordwell and P. J. Boutan, J . Am. Chem. SOC., 78, 87 (1956).
`
`-Hydrogen,
`Calcd
`4.44
`
`4.44
`
`4.44
`
`3.83
`
`3.83
`
`3.83
`
`%-
`Found
`. . .
`
`4.62
`. . .
`
`3.83
`
`4.00
`
`4.00
`
`36.85
`. . .
`
`31.38
`
`31.56
`
`31.72
`
`concentrated H2S04 and 2 ml of concentrated "01
`a t room tem-
`perature. As soon as a homogeneous solution was obtained, the
`solution was analyzed in the normal fashion. Gas chromato-
`graphic analysis indicated methyl phenyl sulfide and a trace of
`methyl m-nitrophenyl sulfide (<0.57,). The selenonium salt 2
`was treated in a similar manner except that the reaction mixture
`had to be cooled in Dry Ice since, at room temperature, a con-
`siderable amount of substitution occurred even with a short
`reaction time. Only methyl phenyl selenide was observed with
`a trace of methyl m-nitrophenyl selenide (<0.57,). This in-
`dicated that there was esaentially no substitution occurring dur-
`ing work-up of the reaction mixture.
`The nitration of 1 and 2 wab carried out in the normal fashion
`except that 0.3 g of methyl phenyl sulfide and methyl phenyl
`selenide, rspectively. were added to the reaction mixture. Analy-
`sis of the reaction mistlire indicated no change in isomer distribu-
`tion. Kitration of methyl phenyl sulfide and selenide under the
`
`nitration conditions of 1 and 2 yielded no methyl nitrophenyl
`sulfides or methyl nitrophenyl selenides.
`Each individual dimethylnitrophenylsulfonium methyl sulfate
`(3, 4, 5 ) (0.5 g) and dimethylnitrophenylselenonium methyl sul-
`fate (6, 7, 8 ) (0.5 g) was subjected to the normal nitration and
`analysis procedure.
`In all cases, only the corresponding methyl
`nitrophenyl sulfide or selenide was observed, indicating that there
`was no rearrangement under the conditions of the reaction and
`analysis. Two different known concentrations of the dimethyl-
`nitrophenylsulfonium methyl sulfate salts (3, 4, 5 ) were subjected
`to the normal nitration and analysis procedure. The results given
`in Table I1 indicate good agreement between the calculated and
`the actural percentages found. No dinitration was observed.
`Registry No.-1,
`6203-16-3; 2, 13118-29-1; 3,
`13118-30-4; 4, 13118-31-5; 5 , 13118-32-6; 6 , 13118-
`33-7; 7, 13118-34-8; 8, 13118-35-9.
`
`The Effect of Intramolecular Hydrophobic Bonding on Partition Coefficients
`CORWIX HANSCHI AND SUSAN hl. AKDERSON~
`Department of Chemistry, Pomona College, Claremont, California
`Received February 5, 1967
`
`The 1-octanol-water partition coefficients are presented for 54 organic compounds. The additive-constitutive
`nature of the logarithm of partition coefficients is considered.
`I t is postulated that intramolecular hydrophobic
`bonding can result in lower than expected values for partition coefficients in certain types of compounds.
`
`Relatively few partition coefficients have been de-
`termined for simple neutral aliphatic compounds.
`While many studies have been made on simple aliphatic
`acids and bases, the difficulties involved in the problems
`of analysis of small concentrations of neutral molecules
`has not invited work in this area. The largest' collec-
`tion of such constants, although admitt'edly not very
`accurate, is that of C ~ l l a n d e r . ~
`In extending4-' the classical studies of lleyer and
`Overton on the use of partition coefficients for struc-
`ture-activity correlations, we have used a relat,ive
`constant x defined as x = log Px - log PH. P H is the
`the parent compound in a
`partitmion coefficient of
`congeneric series and P x is that of a derivative. We
`(1) John Simon Guggenheim Fellon.
`(2) Smith Kline and French research associate.
`(3) R. Collander, Physiol. Plantarum., 7, 420 (1954).
`(4) C. Hansch, R. RI. Xluir, T. Fujita, P. P. Maloney, F. Geiger, and
`M. Streich, J . Am. Chem. Soc., 86, 2817 (1963).
`( 5 ) K. Kiehs C. Hansch, and L. Moore, Biochemistry, 6, 2602 (1966).
`(6) C. Hansch and E. W. Deutech, Biochim. Biophya. Acta, 196, 117
`(1966).
`(7) E. W. Deutsch and C. Hansch, A'ature, 911, 75 (1966).
`
`have shown that T and log P are additive-constitutive
` constant^*^^ and this property has now been confirmed
`by others.IO
`In our first studies8 using aromatic compounds, we
`found the ultraviolet spectrophotometer to be a con-
`venient tool for the determination of concentrations of
`the partitioned compounds. For the sake of analytical
`convenience, we decided to take advantage of the addi-
`tive character of log P and T to obtain log P for aliphatic
`compounds. The approach given in eq 1 was used.
`
`( 8 ) T. Fujita, J. Iwasa, and C. Hansch, J . Am. Chem. Soc., 86, 5175 (1964).
`(9) J. Iwasa, T. Fujita, and C. Hansch, J . M e d . Chem., 8, 150 (1965).
`(10) D. J. Currie, C. E. Lough, R. F. Silver, and H. L. Holmes, Can. J .
`Chsm., 44, 1035 (1966).
`
`Mylan Exhibit 1034, Page 1
`
`

`
`2584
`
`HANSCH AND ANDERSON
`
`VOL. 32
`
`TABLE I
`LOG P FOR ~-OCTANOL-~ATER PARTITION COEFFICIENTS
`
`No. of
`determi-
`nation8
`
`Registry
`no.
`75-52-5
`79-24-3
`108-03-2
`75-05-8
`107-12-0
`141-78-6
`10537-3
`67-56-1
`71-23-8
`78-92-2
`75-65-0
`75-85-4
`75-84-3
`108-93-0
`592-50-7
`109-69-3
`106-94-5
`75-03-6
`67-66-3
`78-93-3
`628-81-9
`352-93-2
`627-19-0
`
`Log Pa
`
`-0.33 + 0.02
`0.18 i 0.02
`0.65 f 0.03
`- 0 . 3 4 f O . 0 3
`0.16 f 0.02
`0.73 f 0.03
`1.21 f 0.02
`- 0 . 6 6 f 0.02
`0.34 f 0.02
`0.61 f 0.01
`0.37 f 0.03
`0.89 f 0.01
`1.36 i 0.03
`1.23 f 0.03
`2.33 f 0.01
`2.39 f 0.03
`2.10 f 0.06
`2.00 i 0.10
`1.97 f 0.04
`0.29 i. 0.02
`2.03 i 0.03
`l , 9 5 f 0.02
`1.98 f 0.03
`
`Registry
`no.
`88-89-1
`63-74-1
`
`Log Pa
`1.34 f 0.06
`-0.78 f 0 . 0 2
`
`No. of
`determi-
`nation$
`8c
`76
`
`1.05 f 0.01
`1.75 f 0.05
`1.25 f 0.02
`1.52 f 0.01
`
`4c
`14c
`4"
`4"
`
`3.56 f 0.01
`2.98 f 0.01
`2.84 i 0.01
`2.01 i 0.01
`0.23 f 0.01
`0.47 =t 0.01
`1.26 f 0.01
`1.79 f 0.01
`1.23 f 0.01
`1.58 f 0.01
`3.30 f 0.01
`0.76 f 0.02
`2.40 f 0.01
`1.34 f 0.01
`0.42 f 0.02
`
`3"
`2c
`2=
`3c
`4c
`2c
`2C
`2c
`4=
`2c
`20
`4"
`2c
`2c
`3"
`
`Compound
`Compound
`1. Nitromethane
`29. Picric acid
`2. Nitroethane
`30. Sulfanilamide
`3. 1-Nitropropane
`31. 3,CMethylenedioxy-
`4. Acetonitrile
`495-76-1
`benzyl alcohol
`5. Propionitrile
`329-71-5
`32. 2,5-Dinitrophenol
`6. Ethyl acetate
`573-56-8
`33. 2,6-Dinitrophenol
`7. Ethyl propionate
`34. Dbcamphorquinone 10373-78-1
`35. 4-Trimethylsilyl-
`8. Methanol
`13 132-25-7
`phenol
`9. 1-Propanol
`90-15-3
`36. 1-Naphthol
`10. %Butanol
`135- 19-3
`37. %Naphthol
`11. &Butyl alcohol
`12. &Amyl alcohol
`95-16-9
`38. Benzothiazole
`60-80-0
`39. Antipyrin
`13. Xeopentyl alcohol
`40. Methyl phenyl sulfone 311285-4
`14. Cyclohexrmol
`41. Salicylamide
`65-45-2
`15. 1-Fluoropentane
`42. ZXitroaniline
`88-74-4
`16. 1-Chlorobutane
`43. Acetylsalicylic acid
`50-78-2
`17. 1-Bromopropane
`44. 4-Ethoxyacetanilide
`62-44-2
`18. Iodoetharie
`19. Chloroform
`45. Thymol
`89-83-8
`46. Morphine
`5 7- 27- 2
`20. 2-Butanonee
`47. Diphenyl sulfone
`127-63-9
`21. Butyl ethyl ether
`48. Benzotriazole
`95-14-7
`22. Diethyl sulfide
`49. 1,3-Methylphenylurea 1007-36-9
`23. 1-Pentyni?
`24. N-Phenyl ethyl
`50. 1,3-Methylphenyl-
`thiourea
`carbamate
`25. K-Methyl phenyl
`51. N,N-Dimethyl-Y'-
`2C
`0.98 f 0.02
`1.24 i 0.02
`2c
`101-42-8
`phen ylurea
`1943-79-9
`carbamate
`2.14 f 0.01
`2c
`120-72-9
`52. Indole'
`109-97-7
`0.75 i 0.01
`26. Pyrrole
`1.73 f 0.02
`2.26 i 0.03
`69-72-7
`53. Salicylic acid
`130-95-0
`27. Quinine
`6c
`2.08 f 0.02
`49
`119-68-3
`- 0 . 2 1 f 0 . 0 2
`541-35-5
`54. Butyramide
`28. Isoquinoline
`a Log P values are given with the standard derivation for the indicated numbers of determinations. The partition coefficient was
`b Log P is the average of the indicated number of determi-
`calculated foi the undissociated molecule* in the case of acids and bases.
`nations.
`c Analysis made using a Cary Model 14 spectrophotometer.
`d Analysis made using vapor phase chromatography.
`our previous report should be negative. The discovery of this error triggered the work reported in this paper. ' Our previously re-
`e This value for 2-butanone is more reliable than our previously reported9 figure of 0.32. The sign of value for 2-butanone in eq 4 of
`ported value for this compound is in error because of a typographical mistake. 0 Determined using Nessler's method of nitrogen de-
`termination.
`
`101-99-5
`
`2.30 f 0.03
`
`2724-69-8
`
`0.85 f 0.01
`
`In this way the benzene ring served as a useful analytical
`marker. The ir values obtained for functions such as
`OH, C S , halogen, etc., could then be added to ir (0.5)
`for methyl and methylene groups to obtain log P for a
`great variety of aliphatic compounds. For example,
`log P for ethyl alcohol could be calculated according to
`eq 2 . While this procedure gives a good self-consistent
`
`set of T and log P values of aliphatic compounds which
`yielded9 an excellent structureactivity correlation with
`Overton's data, further work has now revealed an inter-
`action between the aromatic ring and the side chain
`of the phenylpropyl derivatives which affects the abso-
`lute value of x or log P. We have now measured the
`1-octanol-water partition coefficients for a variety of
`aliphatic molecules directly, using vapor phase chro-
`matography as the analytical tool. These values are
`shown in Table I. Table I also contains a group of
`miscellaneous log P values which we have measured
`for various structure-activity studies. Table I1 shows
`the difference in ir values for aliphatic functions ob-
`tained in these different ways. The values for irl were
`obtainedg cia eq 1 and those for 7 2 were obtained
`similarly using CeHECH,X and CeH&HS. Values for
`
`* 8 C
`
`-0.86
`-1.25
`-1.13
`
`ai - aa
`0.43
`
`0.59
`0.54
`0.29
`
`TABLE I1
`COMPARISON OF DIRECTLY B N D INDIRECTLY
`DETERMINED R VALUES
`ai - aa
`Function
`a26
`ala
`0.64
`-1.59
`-1.16
`-1.80
`OH
`-0.17
`F
`0.56
`-0.73
`c1
`0.52
`0.39
`-0.13
`0.56
`0.60
`0.04
`Br
`0.64
`-0.27
`-0.91
`COOCHI
`0.55
`-0.71
`-1.26
`COCHI
`0.63
`-0.84
`-1.47
`CN
`0.51
`-0.47
`-0.98
`OCHI
`0.53
`0.57
`-1.71
`-2.24
`-2.28
`CONH2
`Founds using eq 1.
`Found analogously to r1 using CsHs-
`CH2X and C6H5CH3. Found from the data in Table I. The
`reference molecules for ~3 are 1-propanol, 1-fluoropentane,
`1-chlorobutane, 1-bromopropane, ethyl acetate, 2-butanone, pro-
`pionitrile, butyl ethyl ether, and butyramide. For each methyl-
`ene or methyl group a value of 0.5 is assigned.
`aa were obtained from the data in Table I, as illustrated
`for F in eq 3.
`
`
`log Pclall~ - T C ~ H ~ ~ = TF = 2.33 - 2.50
`
`-0.17
`
`(3)
`
`Experimental Section
`To determine partition coefficients for compounds not absorb-
`ing in the range covered by the Cary spectrophotometer, analysis
`was carried out using vapor phase chromatography. The Loenco
`
`Mylan Exhibit 1034, Page 2
`
`

`
`AUGUST 1967
`
`HYDROPHOBIC RONDIXG
`INTRAMOLECULAR
`
`2585
`
`Model 70 gas chromatograph with hydrogen flame detector was
`used. The general method for purifying the octanol and parti-
`t,ioning has been described!
`In our previous work, partitioning
`was done in centrifuge bottles with rubber stoppers. For our
`present work, partitioning was done in specially made centrifuge
`bottles having ground-glass stoppers. For the' more volatile
`compounds, care was taken to use enough of the solvents so that
`the bottles were almost full; in this way partitioning with air
`could be neglected! Care must be excercised with acidic or basic
`compounds to exclude carbon dioxide. Carbon dioxide free
`water and a nit'rogen atmosphere were employed in these ex-
`periments.
`The concentration was measured in either the octanol or water
`phase, most commonly the water phase. The concentrations
`were usually about 10-3 X. The amount of octanol used varied
`from 5 to 175 ml and the amount of water varied from 200 to
`25 ml depending on the solubility characteristics of the compound
`being examined. These variations were made in order to keep
`the errors resulting from dividing small numbers into large num-
`bers to a minimum.
`A weighed portion of compound was usually dissolved in a
`specific amount of water to form a standard solution. Quantities
`(1 pl) of this standard were injected into the chromatograph.
`Usually five repetitions were made for each concentration of the
`st,andard. The desired amount of octanol was then added to each
`standard and the solutions were shaken and centrifuged. A 1-p1
`portion of the water (or octanol) phase was then chromato-
`graphed. The area of each peak was obtained by Xeroxing the
`graph (t,o get heavier paper), cutting out the peak, and weighing
`this piece of paper. The weight was proportional to the concen-
`tration of t,he compound in the solution which was proportional
`to the weight of sample used. The weight in the octanol layer
`was obtained by subtracting the weight of
`the peak of
`the
`partitioned solrition from the weight of the standard peak.
`
`wt of standard peak - wt of partitioned peak
`P =
`wt of partitioned peak
`
`X
`V O ~ . of HzO
`vol. of octanol
`
`Discussion
`The most striking point in Table I1 is the essentially
`constant difference between x1 and x3. The agreement
`is especially good when one considers that x1 values
`were obtained by one worker using the ultraviolet
`spectrophotometer for analysis and 57-3 values were
`obtained 1 year later by another worker using vapor-
`phase chromatography as the analytical tool. The
`average standard deviation on the values of x2 - is
`approximately 0.05. Our first thought was that this
`difference must be a result of an error in the determina-
`tion of log P c ~ H ~ c H ~ C H ~ C H ~
`since this is a constant factor
`in the determination of xl. However, log P for propyl-
`benzene and benzene was determined by three different
`workers, each separated by about 1 year in time, and
`good agreement was obtained. Log P for benzene was
`measured using both the ultraviolet and vpc techniques
`and found to be the same by each method. The dif-
`ference between log P for benzene and log P for propyl-
`benzene is 1.56, which gives an average value of 0.52
`for each CH2 unit. This is in good agreement with what
`we have found in many other instances. Thus, the
`value for propylbenzene must be reasonably good. By
`no means can it be off anything near the amount of x1 -
`x3. This leaves us with the conclusion that there must
`be an interaction between the phenyl ring and the side
`chain in C6H5CH2CH2CH2X not present in propyl-
`benzene. Also, this interaction appears to be essen-
`tially independent of the nature of X. It seems likely
`to us that an interaction of the side-chain dipole with
`the x electrons of the aromatic ring could result in a
`
`folding together of these two protions of the molecules.
`The electronic force is postulated because the effect is
`not present in propylbenxene. We would expect the
`electronic force to be weak but reinforced by intra-
`molecular hydrophobic bonding. The net result is a
`more compact structure for CaHsCH2CH2CH2X and
`hence when it is placed in water, less perturbation of
`the water structure results and therefore greater than
`expected water solubility might be observed. This is
`what we find comparing TI and as; the greater negative
`value for x1 means higher water solubility. Such effects
`are to be expected" and in fact have been observed.
`Currie, et u1.,l0 have suggested that such folding results
`in a lower than expected log P for vitamin K1. This,
`it would seem that the study of the difference between
`additively calculated partition coefficients and experi-
`mentally determined values could shed light on the
`conformations of complex molecules in aqueous solu-
`tion. The study of such hydrophobic interactions aided
`by hydrogen bonding and/or dipolar interactions has
`been shown to play an important role in protein
`structure. l 2 t 1
`Shortening the length of the side chain in eq 1 be-
`tween the aromatic ring and X, one would expect the
`effect on x to decrease and even to disappear. This
`effect can be observed in Table I1 in the column x z -
`8 3 . However, even in the benzyl derivatives there
`appears to be enough interaction between the side chain
`and the ring to seriously disrupt the normal envelope
`of water which one would expect" to be loosely held
`around the ring in aqueous solution. The values for
`x2 - x3 are smaller and more varied than x1 - x3.
`Presumably in the case of the benzyl derivatives less
`than complete shielding of one side of the aromatic
`ring occurs, while in the case of the phenylpropyl deriva-
`tives one side of the benzene ring is likely to be com-
`pletely shielded from water interaction. In calculating
`partition coefficients for mixed aliphatic-aromatic
`compounds, our previously reportedg values for x will
`be quite useful. However, when folding can occur and
`especially when it can be promoted by dipolar interac-
`tions or hydrogen bonding, caution must be used in
`additively corllbining x and log P values.
`A point of concern to us has been the possible effect
`of a very polar group on x for an apolar moiety such as
`a methyl or methylene unit. Early works showed that
`the electronic effect of substituents on a lone pair of
`electrons could have a pronounced effect on log P .
`Such an effect seems to be small for methyl and methyl-
`ene groups. For example, the Alog P increment in
`going from nitromethane to nitroethane is 0.52 and in
`going from nitroethane to nitropropane it is 0.47.
`The difference between acetonitrile and propionitrile is
`0.50. The difference between ethyl acetate and ethyl
`propionate is 0.48. The difference between t-butyl and
`t-amyl alcohols is 0.52. The average value for CH3
`found on 15 different aromatic nucleis is 0.505. It thus
`appears that polarizing effects of neighboring electro-
`negative atoms on the carbon-hydrogen bond do not
`greatly affect log P.
`In summary, one can say that x and log P values
`appear to be additive whenever there are no new effects
`
`(11) W. Kauzmann, Aduan. Protein Chem., 14, 37 (1959).
`(12) C. Tanford, J . A m . Chem. Soc., 84, 4240 (1962).
`(13) G. NBmenthp and H. A. Scheraga, J . P h y s . Chem., 66, l i 7 3 (1962).
`
`Myaln Exhibit 1034, Page 3
`
`

`
`2586
`
`SMITH AND TAN
`
`VOL. 32
`
`in the summation not present in the constituent parts.
`Such intramolecular interactions which we have so far
`observed are electronic, hydrogen bonding, and the
`shielding effects considered in this report. The shield-
`ing may be of two kinds. For example, when two
`apolar groups are adjacent (e.g., ortho) to each other,
`they will not have the same number of structured
`water molecules around
`them as when separated
`(e.g., para). The other type of shielding occurs from
`folding in nonrigid molecules.
`In this type one can
`
`expect an important role for intramolecular hydro-
`phobic bonding.
`
`Acknowledgment.-This work was supported under
`Research Grant GM-07492 froin the Xational In-
`stitutes of Health. We are also indebted to Smith
`Kline and French for financial assistance. We wish to
`thank Professor C. Freeman Allen for advice on chro-
`matography technique and for supplying us with several
`of the compounds in Table I.
`
`The Reaction of Some Quaternary Hydrazones with Grignard Reagents1
`PETER A. S. SMITH AND H. H. TAN
`Department of Chemistry, University of Michigan, Ann Arbor, hf ichigan 48104
`Received Sovember 9, 1966
`Quaternary hydrazones of the type ArtC=?iN +RJ - were prepared by treating various benzophenone dialkyl-
`hydrazones with methyl iodide. They reacted exothermically with Grignard reagents to form substances from
`whirh the quaternary hydrazone could be recovered on mild hydrolysis. Upon refluxing in tetrahydrofuran,
`further reaction took place, involving K-N cleavage, leading to tertiary amine, benzophenonimine, X-substi-
`tuted benzophenonimine, and biaryl (or bialkyl) (eq 3 and 4). A small part of the N-substituted imines corre-
`sponded to C to N migration of a C-aryl group, to account for which an amine N-imide intermediate is pro-
`posed (eq 5 ) . However, the ratios of the X-attached groups from experiments in which the group in the Grignard
`reagent and those on the azomethine carbon were different are not consistent with known migration aptitudes
`and indicate that the major proportion of K-substituted imine arises by direct substitution on the nitrogen.
`p-Chlorobenzaldehyde quaternary hydrazone reacted with p-tolylmagnesium bromide to give p-methy1-p’-
`chlorobenzophenonimine, which may have arisen through initial base-catalyzed elimination to p-chlorobenzo-
`nitrile.
`
`-f
`
`It is well known2 that the azomethine system is less
`altogether and the highest yield reported was only
`reactive than the carbonyl system toward addition of
`32%. The reaction of quaternary hydrazonium salts
`RZC=K\’N+R3 X- with organometallic reagents has not
`organometallic reagents, a consequence presumably of
`greater electronic symmetry. Although aldimines
`been reported before and is the subject of this paper.
`undergo conventional addition of Grignard reagents
`The sluggishness of the carbon-nitrogen double bond of
`more or less readily, ketimines may undergo a compet-
`imines toward addition of Grignard reagents is largely
`overcome in immonium salts, R2C=N+R2 X-, which
`ing base-catalyzed aldol-type condensation i n ~ t e a d , ~
`even to the complete exclusion of addition.
`In such
`give good yields of tertiary amines.6 In view of this,
`cases, the ketimine acts as a source of active hydrogen
`we expected that the inductive effect of the positive
`and destroys the Grignard reagent (eq 1). Benzo-
`charge of quaternary hydrazonium salts would activate
`PhC-NPh + RMgX
`RH + PhC=CHC=NPh + PhNHMgX
`the azomethine system toward addition reactions.
`However, this expectation was not realized; this explora-
`I
`I
`I
`
`tory paper charts the principal types of reaction that
`(1)
`CH,
`Ph
`CH3
`actually take place.
`phenone anil reacts exothermically with phenyl-
`magnesium bromide to form a complex, from which the
`anil may be recovered unchanged by hydr~lysis.~ When
`forcing conditions are used, 1,4 addition occurs, giving
`N-o-phenylben~hydrylaniline.~ Phenyllithium, how-
`ever, adds conventionally to give K-triphenylmethyl-
`anilines5
`Hydrazones behave somewhat analogously to imines.
`Some aldehyde phenylhydrazones add Grignard re-
`agents, but the NH proton destroys 1 mole of reagent
`and the reactions are accompanied by N-K cleavage
`and are not clean.2 Benzophenone phenylhydrazone
`does not undergo addition at aIL4 Some dimethyl-
`hydrazones have been reported to undergo normal
`addition of butyllithium or phenyllithium to give tri-
`substituted hydrazines12 but many examples failed
`(1) Taken from the doctoral thesis of H. H. Tan, University of Michigan,
`1962.
`( 2 ) A. Rfarxer and At. Horvath, H e h . Chim. A h , 47, 1101 (1964).
`( 3 ) TV. F. Short and J. S. Watt, J . Chem. Soc., 2293 (1930).
`(4) H. Gilman, J. E. Kirby, and C. R . Kinney. J . A m . Chem. Soc., 61,
`2252 (1929).
`( 5 ) H. Gilman and J. Morton, abad., 70, 2514 (1948).
`
`Results
`Quaternary hydrazonium iodides of benzophenone,
`p,p‘-dichlorobenzophenone, p,p‘-dimethoxybenzophe-
`none, and p-chlorobenzaldehyde were prepared by
`treating the corresponding dimethylhydrazones or
`pentamethylenehydrazones with methyl iodide.’
`Treatment of benzophenone trimethylhydrazonium
`iodide with phenylmagnesium bromide, methylmag-
`nesium iodide, or methyllithium in ether resulted in
`marked heat evolution, but quaternary hydrazonium
`salt could be recovered upon hydrolysis, nearly quan-
`titatively in the case of the Grignard reagents. Only
`when a threefold excess of Grignard reagent and a re-
`action time of 10 hr in refluxing ether were used was
`extensive further reaction observed with phenyl-
`magnesium bromide. Hydrolysis led to the recovery
`
`(6) D. Craig, ibid., 60, 1458 (1938): E. Bergmann and W. Rosenthal,
`J . Prakt. Chem., 121 196, 267 (1932); H. G. Reiher and T. D. Stewart, J . A m .
`Chem. Soc., 62, 3026 (1940).
`(7) P. A . S. Smith and E. E. Moat, Jr. J . Ora. Chem., 2’d, 359 (1957).
`
`Mylan Exhibit 1034, Page 4