Retrospective Theses and Dissertations
1952
The mechanism of the benzidine rearrangement Mendel David Cohen Iowa State College
Follow this and additional works at: http://lib.dr.iastate.edu/rtd Part of the Physical Chemistry Commons Recommended Citation Cohen, Mendel David, "The mechanism of the benzidine rearrangement " (1952). Retrospective Theses and Dissertations. 13398. http://lib.dr.iastate.edu/rtd/13398
This Dissertation is brought to you for free and open access by Iowa State University Digital Repository. It has been accepted for inclusion in Retrospective Theses and Dissertations by an authorized administrator of Iowa State University Digital Repository. For more information, please contact
[email protected]
THE MiCiiMlSM OF THE BMZICIKE REABRMOEinf
by
Mendel David Cohen
A rissortation Submitted to the Graduate Faculty in Partial Fulfillment of The Requiremeats for the Degree of BOCTOE OF PHILOSOPHY
Major Subjeoti
Physical CheiBietry
Approved:
Signature was redacted for privacy. In tih^ge of Major Work
Signature was redacted for privacy. Head of Major Departiasntl
Signature was redacted for privacy. of Gradate ^Colleie
Iowa State College 1952
UMI Number: DP12649
INFORMATION TO USERS
The quality of this reproduction is dependent upon the quality of the copy submitted. Broken or indistinct print, colored or poor quality illustrations and photographs, print bleed-through, substandard margins, and improper alignment can adversely affect reproduction. In the unlikely event that the author did not send a complete manuscript and there are missing pages, these will be noted. Also, if unauthorized copyright material had to be removed, a note will indicate the deletion.
UMI UMI Microform DP12649 Copyright 2005 by ProQuest Information and Learning Company. All rights reserved. This microform edition is protected against unauthorized copying under Title 17, United States Code.
ProQuest Information and Learning Company 300 North Zeeb Road P.O. Box 1346 Ann Arbor, Ml 48106-1346
ii C (ptrm TABLE OP COHTSiTS
Paga iii
A0KHOWLEDaM]^fS INTRODUCTIC®
...»
1
EISTOEiCAl..
4
SXPlRIiffltAL MBTHODS AKD HESULTS Praparatioa of Matsrials BiiadsohodXer'e Green. Iilthloffl perohlorate (anhydrous) %draBobensene Styphnlo acid Titanoue chloride solution Sodium ehloroaoetate» styphnate, glyoolate and eyazioaoetate Attei^ted preparation of hydrasobenKesie dihydroiodide... Beusidiae dihydroiodide Other aaterials
23 23 23 24 24 25 25 26 2g SO 30
Kiaetio Prooedure aad Analytioal Teohaique
30
Variation of Reaction Rate with Change of Acid Catalyst.....
32
Variation of Kate with Concentration of Undiasooiated Acid..
36
AtteB^ted Application of the Broneted Kelation.
43
AtteBgjted Isolation of the Kearraagemant Step...
44
DISGOSSlGi
48
Limitations Inherent in the B3Eperimsntal Method.
48
The InforiMition Gained from this Study
60
Xs^lioations of the Kinetics
61
The Mature of the Transition-state.
S4
The Mature of the Rearrangement Step
68
The Therimdyaamios of Pre-equilibration.
59
The Bronsted Belation
62
SUMtUHY.....
6& TlOt^l
iii
ACKNOWLEDCMEiiTS
The author wishes to express his deep appreoiation to Dr. G.S. EaisQond for his friendly and enthusiastic guidaaoe throughout the course of this work. Acknowledgment is also made of a United States Public Bealth Service PredootoraX Research Fellowehip, which enabled the author to undertake this study.
1
INTRODUCTION
The bentidine rearrangement is the rearrangement of substituted and unsubstituted hydraaobenzenes to the corresponding benzidines.
For the
unsubstituted cofflpouiids th© reaction is formulated!
0 o This change oan soraetimes be effected by application of heat alone« but is generally produced by use of acids as catalysts.
In the latter
case the formation of benzidine is aoconpanied by formation of vasrying amounts of diphenylin9 (l), o-seraidino (II) and £-se!nidine (III),
0
"™"0
NHg
Itiig
1
II
III
The relative proportions of these prouuots are determined by tdie particular hydraaobenEene used and by th© experimental conditions. It h&a been known for some years that the acid catalysed rearrange ment is intrait»lecular.
It is this fact which lends special interest to
a study of the mechanism involved.
Intramoleoularity suggests a continuous
binding together of the two sections of the molecule.
Thus while the new
carbon - carbon linkage is being formed there must be partial bonding of the two residues, either directly through the nitrogen atoms, or by some secondary valence involving the aromatic Tt^eleotrons.
In particular, the
configuration of the transition-state for the rearrangement has been the subject of much speculation}
for a bensene ring is known to have considerable
2 "thlckaess^, yat it seemed ueoessary for there to be sisiulteiieous p&rti&l bonding between the nitrogen atoms and between the carbon atoms. During the twenty years prior to 1960, a large number of descriptions of the natiure of the transition-state were formulated.
All these were
based on the assiuBption that the reaction was of first order in hydrogen ion.
Thus* most workers in this field postulated that the first conjugate
aoid of hydraxobensene rearranges in the rate>determining step. In 1950 it was shown 1;^ Hammond and Shiae^ that the hydrogen ion dependence was a seoond order one* in fact.
These authors therefore
suggested that the seoond conjugate aoid of hydrasobenxene might be the species which rearranges. (i)
fhere remained the question of whether
the transfer of the seoond proton to the first conjugate acid
of i^drasobensene, or (ii) the rearrangement of the dioonjugate aoid, is the rate-determining step.
The former case would be of particular
interest in that the transfer of l^drogen from an aoid« or a protolytio O solTentt to nitrogen usually takes place extremely rapidly . The rearrangement is catalysed by mineral acids and inoompletely dissociated In solution.
acids which aure
In the latter ease the active
oatalytio speoies could be hydrogwa ion« or molecular aoid, or both. Catalysis
molecular aoid« called general aoid catalysis* is only
kinetioally obseryable when transfer of the proton from aoid to reaotant
^Bsounond and Shine« £. i^. Ch«B. Soo.> 72* 220 (i960}. Ingold and Wilson* Z. Blektroeh«a.* 44* 62 (19S8).
3
takes place in the rate-determining step. It was decided to investigate more closely the type of catalysis inTolved in this rearrangement.
If the catalysis proved to be specific*
ally by l^drogen ion then no choice could be made between the possibilities (!) and (ii)*
If the rearrangement proved to be subject to general acid
catalysis, however, it would indicate that the transfer of a proton to the first conjugate acid of l^rdrasobensene ooours in the rate-determining step. For the TOSt general type of acid catalysis, the specific rate constant may be expressed in the form k - kj[b*]
[ail] . i
fhe first term on the right represents the contribution to the overall rate by the hydrogm ion catalysed reaction.
The terms in the summation
are the contributions of the various molecular acids, where the kj^'s are the catalytic constants of the acids.
It was decided to employ two
methods of testing for general acid catalysis.
Both require constant
ionic strength and constant hydrogen ion concentration*
Then the acid
oatalysis is general If there is a variation in the specific rate on re placing the aoid
by acid Eky or if there is a change in rate with
variation in the concentration of undissooiated acid* Additional evidence on the identity of the rate-controlling step oould be obtained if the individual steps were isolable and their specific nites measurable.
It was therefore decided to attes^t to isolate the
dioonjugate aoid of hydrasobensene and to determine its rate of rearrange ment in a suitable solvent.
4 HISTORICAL In 1846 Zioin studied the reduction of aeobenxene in aoid media* Almost tnrenl^ years later Fittig** showed that the product, bensidine, is a diaffiinodlphenyl.
Hoffman^, soon after, disoovered %draxobens(me and
showed that it was the precursor of beneidine in the acid reduction of asobensene.
Sehults and his coworkers were the first to isolate some
5 diphenyline from the rearrangement products*
They were also able to
6 7 establish the structures of benxidine and of diphenyliae . Since that time the rearrangements of a considerable number of asoand hydraso»b«mxeae8 have been studied*
8
work of Jaoobsen
Particularly as a result of the
and his students the scope and limitations of the re
arrangement hare been fairly clearly established*
fhe main purpose of
these early investigations was to establish the effect of substituents in positions in the aromatic nuclei at which carbon - carbon linking might otherwise have been expected to occur. fhe knowledge thus accumulated may be summarised, as follows. all cases a certain amount of dlsproportionatlon ooourst aCgHgKi-HBDgHg
'Pittig,
CgHgll.«5gBg*2C^HgHI^.
12^ 282 (1862).
^Hoffman, Proo* Roy* Soo., 12, 576 (1863), Schmidt and Schults, Ber., jUL, 1764 (1678). ®Schultx,
174, 227 (1874).
^Schmidt* Schults and Strasser, ibid., 207, 548 (1881). Sjacobsen, ibid., 428, 76 (1922).
In
6
When both p>r> positions are free the main product is the bensidine, with smaller amounts of the diphenyline and of the ortho and para semidines, provided the appropriate positions are free*
XX as the main product, where X «
Thus
XX -SOgH or -CMg.
This is also
true if the hydrasobensene is both ortho- and insta~sub8tituted.
From
]:^draxonaphthaleae^ (lV)t for exaii|)le, the main product is th9 oorreiponding £,£'-disubstituted confound (V).
Howe-ver, in this case there is also
formed the £,£' product (VI), or the corresponding carbazole (VIl), while no o,£' product analogous to diphenyline is isolated.
The case of l^drasobenzenes with one para position substituted is of sonw interest*
Usually the principal product is the diphenyline.
HoweTer, when the subtituent is -Gl«
-CO^H or -OCOCIig the rearrange'
ment often gives benzidine, with elimination of the substitueat*
^Yesely, Ber., 58, 1S6 <1906).
Thus
6
hydrasob«]:usene-4>oarbo3e3rlio acid (Vlll) yields mainly unsubstituted
10
beneidine .
-HH-NH-
-COOH
VIII Serersl exan^les of H-substituted hydrasobensenes have been studied. SoBie of their reaotioas are illustrated below* the products shown being the main ones but not being formed exolusively. (CgHg)2»-H(CgHg)2 ^CgHg.HB.CgHg.
in SBtsdl yields^^. (CgHg )2S-K^Cgiig
CgHg-HB.
-
*
X2 An interesting oase is the reftrrcuagement
N—N13 Hughes and Ingold"^ have sunsaurised Jaoobsen'a results in tabular form. 14 Dewar has developed general rules whieh describe the various direetive influences.
Considering the system
where A is the laore
^%oeltiag and Werner* Ber., Zg* 2366 (1890). ^%ieland and Gasbarjan, ibid., 59, 1499 (1906). ^ieland and Baas* ibid.,
1336 (1920).
1.3 Hughes and Ingold* Quart. Sev., <6* 34 (1952). ^^Dewar, Satire,
784 (1945).
7 14
basic aromatic ring, he formulated the rules as follows* s (i) The ease of reaction is in the order ben«idine>diph6ii.yliiie> seioidind. (ii) Only the para position in A can take part in diphenyline formation. Thus in 4-chlorohydrazobenaene, where the unsubstitutea ring is A, benzidine formation does aot take place unless elimination ooours. The main product is 2,4'»diamino5-ohlorobiphenyl (IX), (iii) la semidine formation A carries the oiiH2 in the product. Thus, for 4-metho3^hydra«obenzene, benzidine cannot be formed and diphenyline is not formed as the p^a position of A is substituted. The product found is 2-»amino-5-methoaydiphanylRraine (X). (iv) Disproportionation probably occurs in all oases but is independejat of the rearrangement.
NHg
NH2
- O - O
o-™-o
CI
"OCH3
IX
X
In reoent years the search for new synthetic applicaticms of the raarrangesisnt has decreased in intensity.
Instead en|>hasis has been
placed on the question of the nwohanism - a question which has aroused considerable interest and controversy. An early suggestion was that semidines. often isolated from the re» 15 aetlon mixtures, are the precursors of the cozresponding bensidines . Howerer, Robinson
showed that semidines do not form benzidines under the
conditions of "Uie rearrangement.
Thus semidine and bensidine foxnnation
appear to be independent and simultaneous. It is conceivable that the rearrangement proceeds by initial fission of the hydrasobensene into fragments which then recombine to give the
^^Lapworth, £. Chem» Soc,, 73^, 461 (1898). ^^obinson and Robinson, ibid., 115, 639 (1918)*
8
products.
18
Boleoules
17
It has bo«n suggested, in tum« that radicals , ions and are the reactire fragments.
suoh fissi
There is strong evidenoe that no O Jaeobsen° studied the re*
arrangeaent of a large number of moleoules C-MS-HB-S, where C and D are different aromatio tfystens.
In no oase was any product eontaining two
C nuclei, or two D nuclei* isolated. This failure to find cross-products is not valid proof against initial heteropolar fission*
If C-NB-KB«D — + (0101)",
the absence of cross-products nay be due to electrostatic repulsion between ions of like charge.
For this reason Ingold and Kidd^® studied
the rearrangement of 2,2'-dioethoxy- and 2,2'-diethcxy-hydrajsobenxenes, which undergo the reaction at comparable rates and with forsMition of bensidines* mainly*
They found no erideaee for cross-products.
Similarly,
20 , using a tracer method* concluded that if there are
IQieland and Schmurts
any cross-products they constitute less than one percent of the total products of the rearrangement. Further evidence against initial fragmentation is obtained from studies of moleoules suoh as tetraphenyll^drasine, which are known to be partially dissociated into free radicals under certain conditions.
17Jones and Kenner* £* Chem. Soc., 1848 (1931)} 711 (1932). ^®SteiglitB, £•
Chem. Soo«, 26, 62 (190S).
^^Ingold and Kidd* £• Chem. Soc*. 984 (19S3). ^^heland and Schwarts,
Chem. Phys.« 17, 426 (1949).
l^er
9 the BABie conditions no rearrangemextb ooours^^.
Eridenoe against iater-
21
moleoularity was olitained hy Bloink and Pausaoker * in their studies of the rearrangesients of 2> and 3<>earboxyhydrasobensenes. pairs like
Finally, laoleoular
Cgi^NfiOH, which might be postulated as intermediates,
do not react to give benzidine. Identifioation of the products of the rearrangeraeats of various substituted hydrasobensenes thus led to the oonolueion that the rearrangeawnt is an intramolecular one*
In subsequent work the centres of interest
have been the questions of the nature of the transition-state and of the driving force for the rearrangement. During the initial period of study of the mechanism kinetic evidence was scanty.
In 1904, van Loon^^ found that the rate is of first order in
b^drasobensene and suggested that the hydrogen ion dependence was a seoond order one*
However, the latter aspect of the work is of doubtful value
since ionic strengths were not controlled.
Twenty years later Biilman
and Bloi^® measured the rate electrometrically and suggested that there ms a one-half order dependence on hydrogen ion. Some thirty-five years ago Eobinson^^ was interested in the synthesis of highly substituted pyrroles*
He arrived at a synthesis through
application of a series of analogies and by postulating a mechanism for the benzidine rearrangement.
It is of interest that his meohanism
allows for intramolecularity.
^^Blolnk and Pausaoker, £• Chem. Soo«, 950 (1950). ^^van Loon, Rec. Trav* Chim., 23, 162 (1904). ^^Biilman and Blom, J* Chem. Soo., 125, 1719 (1924).
10
Boblnson first eonsidsred the Fischer indole synthesis which, he suggested, aight take place through some such path ast
0.ra.*<(cH,)
i
Q.»B.KB.?^H^-L(3-OH^^ NHg
c
H the whole reaction proceeding under the influence of l^droohlorio acid and heat.
Here step b utilises an o*b(»isidine type rearrKigeaont, which
Robinson e«q^ared with the formation of oarbasoles from hydrasonaphthalenss• He treated this as a iq^eoifio ease of the rearrangements of the systen A«b»c-D, where A and D are unsaturated groves and either b, or both b and o« can attain a higher malence state by salt formation*
fbis system* on
activation Iqr aoid« ean undergo the following series of changest (i) form A<»i««»o<»D, whwre dotted lines represent ''partial bonds**, (ii) these partial boosds conjugate with A and D, to give which (iii) can close to form a ring* Shis trealaMmt is reminiscent of that of fhiele for ceoijugated unsaturated systems.
fhus Robinson*s mechanism for the indola synthesis was
formulated asi
11
-CH-C-Ih.•Jfii-C-CH-
CH—^ -CJi
-ca.
CH-
E l l
I
-c-nh hh-c-
1 -C—C
-CH
-C
-C-lilH HN-C-
I
0-
-CH-
I
It is notsd th&t the partial bonds allow for the oontinuous binding together of the two sections of the noleoule*
Finally* by analogy* Bobinson suggest
ed that the following reaction sequence might be feasible for the pyrrole (^thesis t -CHg-C-0
»
«C—0- —
-CIM-KS-MH-i-Cl-
HgK-G-C-C-^-K^
-C^ /CM
H In faot« he obtained tetraphex^lpyrrole, in alsost quantitatlre yield, treating bisophesylbensylketasine (Xl) with dry hydrogen chloride*
XI Sobin8on*8 synthetic procedure has recently beoa extended by 24 fheilacker .
Ibis worker showed that aeetophenonephex^lhydrasone
yields 2-ph«nylindole on heating with strong acid.
Beasophenone-
phenyl]:^drasone, however* apparently undergoes the following series of
^^Huckelii **American F.1.A«T» Reriew of German Science* 1939-1946**, Vol* 34, Leibiger Research Laboratories, Petersburg, H.I., 1951, p. 182.
12 changes t
'%en the appropriate para position is substituted the products are mainly ammonia and tar. 2B 26 27 other work * * in this direction has produced an unusual variation. It was shown that rearrangement of a di-ortho-substituted phenylhydracone may lead to elinination of an ortho substituent*
Thus» 2,6--dichloro-
phenyll^drazones (XII) have been found to give 5,7-dichloroindoles (XIII), on treatment with tine chloride. CI <^^NH-H=G(R')CH2R 01 XII
XIII
That non-bensenoid unsaturated systems can participate in the rearrangement has been long established.
28
A glyoxaline nucleus , for
•xaiaqple, can replace a benxene ring, and rearrangement still occurs:
n—KK VN-N-CeHg " W
„
—[i-\
HCl
"
It has been suggested that for the rearrangement to take place the beneene rings must become parallel in the course of thermal notion.
^®Carlin and Fisher, £. Am. Chem. Soo», 70* 3421 (1948)* fiames« Pftusaoker aad Badrook, £• Cham# Soo»» 730 (1951}* ^^Carlin, Wallace and Fisher, £. Aa. Chem* Soc., 74, 990 (1952). ^®Pyman and Ravald, £. Chem. Soc., 117, 1428 (1920).
xs This is considered very is^robable tmless allovranoe is made for the participation of at least ona proton in the trausition-state*
It is
knovm from studies on the structures of crystalline benxene ooi^lexes, for exanplei that it is difficult for two aromatic nuclei to approach faoe to faoe to a distance less than three times the normal carbon carbon single bond distance.
Further information is derived from cases
in whioh the benzene rings are constrained to lie in such a position. fhus* in £,£*-di»etl^lene-l«2-dipheiq7lethane (XIV), in whioh the two o aromatic "planes" are 3.09A apart, the overlap of TT-electrons introduces such a strain as to cause puckering of the bensene rings^^*^^.
In
addition, it is probable that if the rearrangement proceeded by such a mechanism the actiration energy would be Tery high» whereas the experiB»nSI tal Talue^*^ is only about 18 Kcals* per mole*
2
Z
XIV A series of workers have formulated hypothetical meehanisBM based on the assumption that the species which rearranges is the first conjugate acid of l^drasobensene. S2 Robinson suggested that in this reaetion, as in others* a series of
^^Brown and Farthing* Nature, 164, 915 (1949). ^%ran and Steinberg, £.
Chen* Soo*, 78, 6691 (1961)*
Slflewar, J. Ghem. Soe., 777 (1946). ®^Robinson, ibid.. 220 (1941).
14
partially ooB^ansating electron oseillations is involved.
In sii^lified
form his theory ean be repres«ated as follonrsi
(S +)h (^+)h Bobinson's theory* as stated, is both oon^lioated and obscure*
It tends
to overeB^hasise the location of specific electrons and the ohr(»xology of eleetron-B»Tement*
It has the advantage that it allows for partial
quinoid structure in the transltioo-state.
A suitable buckling of the
rings might allow the steric problem involved to be partially overcone* lughes and Ingold^^ objected to the fact that Bobinson's meohanisn required a recession of electrons from positively charged nitrogen* They therefore suggested that the prooess is a concerted ones Cg^Ha-HBSgHg • H* — CgBgNH.**1135585..•fi*—bensidine • H*.
fhe dotted lines indicate that the ionic residues are continuously held together
eleetrostatic forces.
The cationic portion thwa. attaches to
the negatively charged nucleus in an electrophilic aromatic substitutiim reaction. Such a B»ohanism cannot be kinetically differentiated from that in which the first conjugate acid of hydrasobenxene rearranges in the ratedetermining step.
Both of these mechanisms require a mono-positively-
charged transition-state in which, it is ia^lied, there is participation
35Hughes and Ingold, £. Chem. Soc., 606 (1941).
15 by Btruotur«8 such as
Both the meohanism of Kobinaon and the mechanism of Hughes and Ingold postulate simultaneous nitrogen - nitrogen and carbon - carbon bonding. Hamraick and Mason
34
transition-states.
pointed out the unattractive steric aspects of such These authors suggested that it is the first conjugate
acid of hydrazobensene which rearranges* uid that this acid is a reso
Such a charge distribution would allow for a weakened nitrogen - nitrogen bond while there is an electrostatic attraction between the para carbon atoms, which can be operative over a distance large coo^ared to the oo*valent bond length.
It is noteworthy that such a charge distribution now
appears inore naturally, as a result of recent kinetic studies.
Contri
butions from quinoid structures, as shown, also allow for closer approach of the terminal carbon atoms. Attractive as such a scheme is, for the above rsasons, it must be considered highly in^jrobable insofar as energy factors are concerned. Dewar
proposed to overcome the stereochemical difficulties by an
^^Hanuiiiok and Mason* £. Chom. Soc., 638 (1946). Dewar, ibid., 406 (1946).
16
altematlre approach.
As in tha fiughas-Xngold treatment* heteropolar
fission occurs with neutralisation of the charge on the negative fragment by a proton*
fhe difference between the two treatments lies in the t^e
of partial bonding envisaged to hold the segments together.
Dewar
suggested that the bonding is not localised between the nitrogen atoms but that the aromatic systems oan be linked along their lengths.
This type
of bonding, yielding a "TT-coa^lex", involves the TT-electrons of the neutral frapa«it and the vaoant tt-orbital of the cation.
The concept of
TToooB^lexes has been developed extensively by Ce««r^^ and these coB|>lexes have been postulated, by him, as intermediates in electrophilic aromatic substitution reactions.
Thus the nature and proportions of the products
formed in the rearrangement can be predicted, by this token, on the basis of the rules for normal electrophilic substitution. Dewar *s qiumtum mchanical treatment of systeias involving ^-oon^lexes 37 is, perforce, based on a large nuniber of simplifying assumptions * Nevertheless, the concept is a useful one on which oan be based formula tions of UKU^ reaction paths.
It seems surprising that workers of the
calibre of Hughes and Ingold, Robinson and Dewar should have pondered at such length en the mechanina of the benzidine rearrangement without establishing the correclaiess, or otherwise, of the assuiqption of first order hydrogen ion dependence. That this assumption might be incorrect, in fact, was suggested by
®®D«imr, "The Electronic Theory of Organic Chemistry", Chtford Ihiversity Press, London, 1949. ^^Coulson and Dewar, Disc. Farad. Soc., No. 2, *The lAbile i/blecule", 64 (1947).
17 38 39 the work of Pongrats and his oo-worlcers * .
fhese workers heated
azobenxene and, separately, hydrasobensene with methyl iodide in a sealed tube, with or without methanol as solrent. mett^lated benxidines as products*
In all oases they obtained
They formulated the reaction with
hydrasobensene as: CgHgHB-NI^eHg CHgl
6 6
5
3
6 6
•• O.H (CH,LM-H(CH ) C H (I ) 6 5^ 5'2 5'2 6 6 S
(I"),
I
(GH LS-C H-C H-il(CH ) 1 8 Si 6 4 6 4 1 3 3 4
1
1
B
(OH ) i; H -C H S(CH ) d"). 3'3 6 4 6 4 S 8 2 Cos^iound B oould probably be ntore aoourately represented as
fhe authors stated that ooi^ound A can actually be isolated from solution after a short while at a oosparatiTely low ten^eratvire*
Similarly they
obtained ciystals of the dihydrobromide of l^drasobensene from methyl bromide solution at room tesqperature.
It was suggested that these salts
®®Pongrate and Wustaer* Ber», 75, 423 (1940). ®®PongratE and Soholtis, ibid., 75, 2369 (1942).
18 of the diaoid are fttziGtioaal intermediates in the rearrangmnent, sinoe on solution in a suitable solvent they are oonyerted into fully metl^lated bwasidine.
P<»igrats suggested that the course taken by the rearrangMsent
might be
Finally, in 1950, the hydrogen ion depsndenoe of the aoid catalysed beaeidine rearrangement was established unequlTooally Shine^.
^unamd and
These workers found that the rate varies with the square of the
hydrogen ion omoentration, under conditions of constant ionic strength. They tiwrefore suggested that the overall mechanisa is: CgHgM-KHCgHg ^
- CgHglHg-lWgHg
If the diaoid is the rearranging speoies it allows for a stuch more reasonable picture of the transition state.
The adjacent positive
oharges cause a weakening of the nitrogen - nitrogen bond, the electrons of which oen conjugate with the aroaaitie nuclei t
+
There could thus be a continuous separation of like charges while the unlike oharges coae together^.
This formulation allows of a satisfactory
solution of the steric problem.
^%ughe8 and Ingold, J« Chem. Soc., 1638 (1950).
41
Subsequently, Carlln, Selb and Odio«o ion dependenoa is a second order one.
confirBied that the Ij^drogcaa
These -workers measured the rates
of forioatiai of the Individual products of tha resuTangemsnt.
They showed
that for the l^droohlorio aoid oatalysed rearrangement of unsubstituted li^drasobmzene at 26°C. the amount of semidine formed is negligible. They demonstrated, further, that the activation energy is the same for the formation of benEidine as for the formation of diphenyline.
In terms of
the mechanism of Haionond and Shine, this suggests that the transitionstate approximates the structure of the diconjugate aoid more closely than it does the structures of the products.
Ceirlin and co-workers
presented some evidence which supports the belief that the benzidine re arrangement may occur in the solid state.
They therefore suggested that
the intervention of solvent molecules is not necessary either to bring about the reaction or to hold portions of the molecule together while intra molecular rearrangement occurs.
Finally, they gave some results which
indicate that the rearrangement of the diconjugate aoid occurs very rapidly in coaipai'ison to the rate of the overall reaction* Hughes and lugold^® i^iave recently published a review on aromatic rearrangements which includes an extensive section on the mechanism of the bensidiue rearrangement.
These authors stated that Hanraond and Shine
had proved that the catalysis was specifically
hydrogen ion.
On this
basis they discussed the forces which stabilise the transition-state and
^^Carlin, Helb and Odioso, J. Am. Chem. Soc., 73, 1002 (1951).
20
sdlonr the occurreuoe of a reaction which seems so improbable stereochemically.
The transition-state nhich they described is interiaediate
in structure between the dioonjugate acid of hydrasobenzene and the products of reaction.
They pointed out that there osm be considerable
interaction between the electrons originally comprising the nitrogen nitrogen bond and the aromatic if-electrons*
Thus the transition-state
they pictured is a resonance hybrid of such structure a«<
bgh x
mhg
%h
*
D
h n
mh
h n
2||
II 2
21
hh I 2
h''
• fil
fi \J
+
K 0+ 0 2II
Hughes and Ingold's statement of the principle of resonance* stabilisation is refreshing: These are the octet-preserving routes by which any one of these structures can be converted into any othert some involve circulation of the electrons one way around, sonte the other and SOIM either or both ways. The essence of this theory is that the existence of all this free interc^nmunioation betweeia the different conventional electron distributions, and, of course, between the infinitude of inter* mediate unconventional ones, determines, on account of the un certainly principle, a very strong transition-state, one able, and even prone, to form itself, despite its great difference of shape from that of any normal molecule. The authors interpreted the preference for bensidine over diph«a.yline fox^tion to the greater eleetron disturbance, and therefore greater
•cKeference 13, p.60.
21
imeartalnl^f assoeiated with the establishment of the transition-states inrolved para than with ortho rearrangements. It has long boon realised that the rearrangement of hydrasobenzene into benzidine is energetically favorable, in that it involves a change from a weak to a comparatively strong base under acid conditions^.
In
addition, it entails a transformation from two practically isolated benzene rings to a conjugated biphenyl system.
On the basis of these considera
tions alone, a heat of rearrangement of the order of 35 Keals. per mole might be expected^®. The published value for the heats of combustion of the various com pounds involved in the rearrangement hare be«i somewhat inconsistent, until recently.
Thus "Landolt-Bornstein Tabellen* have values varying
from 1539.0 to 1605.6 Koals. per mole, for hydrazobenzene, and from 1557,8 to 1565.1 Koals. per mole, for benzidine. These thermochemical properties have been redetermined by Pcaxgratz and Roth
.
Soao of their results are presented in Tables 1 and 2.
Despite the large energy difference, N,H*-diaoetyli^drazobenzene does not rearrange to the corresponding benzidine*
Instead, hydrolysis
to the N-aoetylhydrazobenzene takes place, followed by rearrangement to M-acel^lbenzidine.
It is interesting to note that the salts of the di-
oonjugate acid, which are apparently intermediates in the rearrangement,
^^Tiohminsky, J. Kuss. Phys. Chem. Soo., 35, €67 (1903), Chem* Centr., II, lETO^dSOSj.
43
Pongratz, Bohmert-Suss and Soholtis, Ber., 77, 651 (1944).
22a
Table 1 Heats of Combustion of Soma Hydrasobeusenes and Benzidines (Koals. per !m>le)
CoB^ovmd
Heat of Combustion
%draEobenKene (HB)
1590.4 + O.S
Beasidine (B)
1555.4 ± O.S
K-acetyl HB
1792.9 • 2.8
Hoacetyl B
1756.1 *2.4
H,K'-diaeetyl HB
2008.3 + 1.1
Difference
35.0 * 0.4
36.8 + 3.7
67.9 • 1.8 **diaoetyl S
1938.4 4-1.4
HB. 2HI
1644.8 + 6.0
B. 2H1
1577.0 +6.0
67,8 + 8.4
Table 2 Heats of Some Reactions Related to the Bensidine Rearrangemant (Koals. per SK>le)
Reaotant
Product
Heat
HB
B. 2HC1
50.00 + 0.17
N-aoetyl HB
N-aoetyl B. 2EC1
51.74 * 0.19
B
B. 2BC1
25.43 * 0.05
S-aeetyl B
N-aoetyl B. ICl
16.55 * 0.10
22b
are relatively very high energy apeoios.
Unfortunately, the heats of
rearrangement and the heats of neutraliaatioa are not directly coaparable, since they were not determined under identical conditions.
23
EIPmMMfAL METHODS AND RESULTS Preparation of Materials Blndsohedler* 8 Qresn A modlfloatlon of the method of Wleland^ was us«d*
Thirty grams of
dliMthylanlllae, frse of the monometl^l coi^ound* was measured into a 600 ml> beaker containing 100 ml* oonoentrated l^droohlorlo aold, and cooled below 5*'c.
A solution of 18 grams of sodivun nitrite In 30 ml.
of water was added slowly, with stirring, from a funnel with the stem dipping below the liquid surface*
The ten|>eratur« was kept below 8^C.
The product was kept stirred for a further hour* flltored, iuid the residue washed with 40 ml* of 1:1 hydrochloric acid, drained well, and washed with a little ethanol. The resultant £-nltro8odlmotl^lanlllne hydrochloride was transferred, with the aid of 70 ml* of water. Into a one litre beaker containing 275 Bil* of oonoentrated ijQrdroohlorlc acid and 55 ml* of water* cooled below O^C*
This was
(tee hundred and fifty grams of sine dust was added
slowly, with stirring, so that the teo^^erature remained below 5^C* Stirring was maintained for a further hour. tered, the sine being kept da^*
The product was then fil
The filtrate was cooled to O^C*
Thirty grams of dlmethylanlllne was dlssolred In 30 ml* of concen trated hydrochloric acid and 20 ml* water, and the solution cooled to Then the solution and the above filtrate were mixed In a one litre beaker. To this was added, dropwlse, a solution of twenl^-elght grams of sodium bichromate In a minimum volume of water.
^fleland, Ber., 48, 1087 (1916).
The rate of addition was
24
oontrolI«d to keep the tes^erature below 15*^0. for ten minutes.
Stirring was continued
The resultant suspension was filtered at a puB^, using
a large Buchner funnel.
As much water as possible was reii>07ed by
applying suotion in a closed system.
The green dye was then washed
repeatedly with anhydrous ethanol until the filtrate oaiae through color less.
Finally* the dye was washed twice with ant^drous ether.
The
product was •aeuum-desicoated over silica gel and anhydrous barium perohlorate.
The yield of the dye was 60 grams.
It was found to be of some importance to thoroughly remove all traces of water in the final stages.
The product thus prepared appears to be
stable indefinitely in the solid state and deconposes very slowly in aqueous solution.
lAthiuffi perohlorate (anhydrous) AQ%drou8 lithiiUB perohlorate was prepared from lithium carbonate AC
and perchloric aeid
the method of Roberts and llarr«a^.
SydraKoben«
It was
recrystallised from petroleim ether (fraction boiling 66®-100®C.) immediately before use. ll.p. 129.0®-129.6®C. (uncorr.)
^%obert8 and Warren, £.
Chem. Soc., 72, 4869 (i960).
^®Hioklnbottom, "Reactions of Organic Cos^ounds", Longmans Green and Co., London, 1948, p. 119i p. 3S2.
25
Stypbnlo aoid Styphnie aoid was prepared according to Hickinbottom^®.
Titanottg chloride solution Twenty five ml. of 20^, analytical grade, titanous chloride (Fiaher Scientific Co*} iras measured into etn oxygen-free solution of 150 ml. concentrated hydrochloric acid in 2750 ml. water.
A nitrogen atmosphere
was maintained in the stook bottle and in the attached automatic burette*
Sodixam ehloroacetate, stypbnate, glyoolate and cyanoacetate A solution of sodium ethoxide in ethanol was added, dropwise, to a stirred solution of the aoid in ethanol, in an atmosphere free of carbon dioxide.
Excess acid was used.
Stirring was continued for thirty
minutes after all the base had been added. filtration. ethanol.
The salt was separated
It was purified by recrystallisation from methanol, or Sodium chloroacetate, prepared in this wa^, was foimd to be
99^ pure by titration against perchloric acid in glacial acetic acid^^.
Attainted preparation of hydrasobensene dihydroiodide An attei^t was made to prepare hydrazobenzene dihydroiodide
the
method of Pongrats^^, the system being maintained under a nitrogen atmosphere.
A yellow solid was separated and unchanged l^drazobenzene
extracted with ether.
It was not possible to establish tmequivocally
the nature of the product.
It decomposes, on heating, over a five-
degree range, 170®-175°C* (uncorr.), whereas Pongratz's product decoaaposes
^^Pritz, Anal. Ghem., 22, 1028 (19S0).
26
at 210°C.
Mo «uitabl« solvent was found for reorystallisation,
hydroj^lic solvents being avoided for fear of inducing rearrangement. The product li soluble in water and in ethanol, but is insoluble in acetone.
Potentiometrio titration with silver nitrate^® gave the figure
of 53.9^ for the iodide content of the confound, as oon^ared with the calculated value of 57.^.
By titration with base in aqueous solution,
the equivalent weight as an a:oid was found to be 254, while the calculated equivalent weight is 220* titration.
Only one wadopoint was observed in the latter
This suggests that, if the coR|>ound is a dibasic acid, either
the second end-point is not observable in water or the two p& values are of similar magnitudes. frost hydrasobensene.
This is certainly not the case for the diaoid However, the two amino-groups in benzidine are of
similar basicioity, the base pK's being 9*03 and 10.26.
It would seem,
then, that at least at the end of the titration the species in solution is bensidine.
Fiurther, an approximately 0.002 molar solution of the
oonpound in water has a pH of 4.90.
If the acid present in solution
were the diconjugate acid of hydrasobenxttae, which must be a strong acid, the pH would be expected to be considerably lower.
An attesqpt was made to titrate the con^ovind, dissolved in ethanol, with aqueous Bindschedler's Green.
The initial rate of decolorisatim
was fairly rapid, but this rate soon decreased markedly*
After only
one quarter of the theoretical amoixat of the dye had been added, no fading
48 w a Kolthoff and Sandell, "Textbook of Quantitative Inorganic Analysis®, The Maamillan Company, New York, 1946, p. 507.
:ry .T.
27
was ob8«rT9
fhe spectra were obtained,
mulls in lujol or hexaohlorobutadiene, with a Baird infrared re^
cording spectrophotometer.
Runs
v/ere
isade by Mr.
Hedges.
The long
wave length portion of the speotrum of the coi^ouad shows strong similari ties with that of hydrasobensene, indicating a similar type of substitution of the aromatic rings*
The weak absorption peak at 3*9 microns suggests
a quaternary iiitrogen.
Beyond this there is no clear evidence for the
diconjugate acid structure since the spectral characteristics of the bond between two positively charged nitrogen atoms are not known.
The absence
of an absorption maximum in the region of 7.0-7.6 microns, when a hexaohlorobutadiene mull is used« indicates the absence of terminal metl^l groups.
Thus there is nothing to indicate contamination of the diacid
by the corresponding tetramethyl iodide. The product was examined under a polarizing microscope by Mr. S. Flikkema.
lith all mounting media evidence was found for non-uniformity
in the san|>le.
However, crystals of one typ* form the bulk of the
material and these crystals were found to be internally homogeneous. Their measured properties arei elongation - negative}
habit « prismatic;
extinction - parallel|
refractive indices - 1.75 + 0.01, 1.65 + 0.01.
The crystalline system is probably tetragonal.
lichen the oos^ound was
moimted in methylene iodide a chemical reaction apparently took place, as a change of crystal type with time was observed.
28
WAVi mts M o*> IIM I4M IN* IW
n«
WA«« IMi—U M CkH
t«
t ^
i n
y^.
7*^
/
\ / 1
/
I/
il M
1 1
1 li
WAV! IM»TH M MOONS
I4« IMO (MO
n A
1J a1
,, /i
1 li i ! if WAVE IfiHftTH M hlCltONS
T!
M
\
i
%
\
: / M
;
i
1 ^
<
I I
A
WAVEtlL»WrH M IIMOONS
-
A
/ 1
^
; f 10
/i
•
WAVE IMim M CkH
l«
'w^
h! h A1
• V
ll<0
u
I'
1WAVE kMm M MCKONS Figure 1.
Infrared Speotra of Benzidine Dii^droiodide, Ij^ydrazobenzene and Benzidine in Nujol Mulls.
29
Figure 2.
Infrared Spectra of ^drasobenzene Dihydroiodide (?) in Hexaohlorobutadiene and Nujol Mulls.
30
Benaidine dihydrolodide Tdohnloal grade bsasidine (Matheson Co.) was racrystalliccd three times from water, twice with deoolorieing carbon.
The resultant
benzidine, which melts at 124.7®-125.0°C. (uncorr.), was dissolved in ether into which was then passed hydrogen iodide vapor. flocculent, light yellow, precipitate was formed. by filtration and dried. 300®C.
An iiranediate
This was separated
The salt does not melt or decon^jose below
The ooiis>ouad was found to be opaque in any mount uader the
microscope, and to disintegrate to a granular powder.
Other materials Foinaio aoid (90^, B. and A.); phoric acid (8S^, Merck))
sodium fortc^te (B. and A.)}
chloroaoetic aoid (Eastinaxm)i
phos
glycolic acid
(Matheson) and cyanoaoetic aoid (id&theson) were used as supplied, without further pwifioation#
Kinetic Prooediu-e and Analytloal Technique All rates were measured for rearrangement at 56.1®C. In the first section of the work the reacticm vessel consisted of two flasks joined below their necks by a tube of one-half inch diameter. The acid and salt were dissolved In ethanol-water and placed in one flask. In the second flask was placed an appropriate volume of a solution of hydrazobenzene.
The air in the vessel was flushed out with nitrogen cuad
the two flasks closed.
The flask containing the hydrasobenzene was
•quipped with an outlet tube which was kept closed by a sbkiII rubber stopper.
Aliquot8 could be removed through this stopper by use of an
hypodermic syringe.
61
7his closed vossel was immersed in a conventional-type thermostat for an hour.
The acid solution
by tilting the vessel.
was
then transferred to the second flask
The soluticaas were mixed thoroughly.
The total
transfer time was of the order of one-half minute, the time of tilt being taken
as
zero time for the
kinetic run.
Ten millilitre aliquots, of a
and
total initial volume of 110 ml., were iwithdrawn
added to en excess of
Bindschedler's Green solution., at appropriate intervals.
The excess of
the dye was detercdned by titration with titanous chloride solution in a nitrogen atmosphere.
The dye solution was standardized against titanous
chloride before and after each run to allow for the gradual decrease in the concentration of active dye.
Thus the volvune of titanous chloride
equivalent to the l^rdrasobensene in the aliquot was obtained. In later work the reaction vessel used consisted of a single flask with a small indentation below the neok, into which solid hydrasobenseae ira.8 placed*
The acid solution was put into the bottom of the flask.
Kinetic results were obtained as before*
It required about one minute
for coB^lete solution of the hydrazobenzene.
For this latter stage of
the study a new lot of Bindschedler's Green was prepared by the modified method described on page 23.
This
preparation
was
found to be stable in
solution during the period of a kinetic run, within the limits of experimental aocuraoy, so that no correction was required. For each rate determination a plot was made of the logarithm of the volume of titanous chloride (equivalent to the hydrasobenaene in the aliquot) against time after mixing-
The best straight line through the
points was calculated by the method of least squares.
The pseudo-first
32
order rate ccmsteuat was calculated from the slope of this line.
Between
six and ten points were obtained in each run. Since disproportionation of hydrasobenzene into asobenzene and aniline occurs under certain oonditicms, several blank rtms were made. A solution of hydraeobensene in ethanol^water, with added salt but in the absence of acid, was placed in the thermostat.
After an hour aliquots
were withdrawn at intervals and treated as in the usual rate deteroinations.
io change in the concentration of faydrasobentene
was observed
over a fturther one hour period.
Variation of Reaction Rate with Change of Acid Catalyst The first test for general acid catalysis consisted of an inves tigation of whether or not the rate depends on the nature of the aoid present, under conditions of constant ionic strength and constant hydrogen ion concentration. Solutions of aoid catalyst, sodium ethoxide and lithium perohlorate, all in the same solvent, were prepared. of the same concentration.
fho latter two solutions were
The aoid was partially neutralised by
addition of sodium ethoxide.
Lithium perohlorate was then added and the
pH measwed with a Beckmann model-6 pH meter. aoid was taken for all runs.
The same initial volume of
The total volume of sodium ethoxide and
lithium perohlorate added was kept the same throughout, thus maintaining constsotit ionic strength.
A series of runs was made with each acid, the
catalyst differing from run to run in a series in that different amounts of sodium ethoxide were added.
Thus the pH-meter reading is a measure
33
of the degree of neutralization of the acid, and each run corresponds to a different hydrogen ion concentration. The catalytic acids used in this study were phosphoric, chloroaoetio, styphnic and formic acids.
The runs with styphnic acid gave non-linear
plots, possibly due to conplexation with the hydranobensene, so rate constants were not obtained. The results of this study are presented in Tables 3,4 and 5, and are represented graphically in Figure 3.
66^, by weight, aqueous ethanol.
The solvent used throughout was
The ten|)erature was 55.1® + 0.06®g.
The total salt concentration was 0.013 molar, maintained by use of lithium perohlorate.
Table 3 Catalysis by Chloroaoetio Acid at Various pH*s [chloroaoetio aoid^ +
jsodium chloroaoetate^ « 1.12 molar
10®k
A(10®k)
Rua Ho*
pH
(ffiia';^)
(mia.^)
8
2.29
476
4.7
9
2.33
322
2.9
11
2.60
216
16.9
12
2.69
197
6.3
13
2.60
206
2,5
34 In the inTestlgatlon of oatalyeis by ohloroaoetio acid some diffi* oulty was caused by the slow oonversion of the acid, in stock solution, to glycollo acid and chloride ion.
For this reason the acid solution was
utilized as soon after preparation as possible.
The pH of an aliquot of
this freshly prepared solution was found to remain constant when the aliquot was held at 5&^C. for an hour, and the resultant solution gave no turbidity with silver nitrate. The quantities A(io®k) were derived from the probable errors involved in the slopes of the lines obtained by the method of least squares.
If
Table 4
Catalysis by Formic Acid at Various pfi's [Formic aoid] + [sodium formate] «•
lo^'k
1.61 molar
ado^k)
Run Ho.
pH
(min."^)
(mia.'^)
17
2.25
290
9.0
18
2,30
228
7.8
19
2.41
143
2.S
20
2.46
108
1.4
21
2.72
ISl
5.1
35
the calculated equation for the line representing the eaqperimental data is> y • fflx + b, and the i^^ point has co-ordinates (Xj, in Bt is given by:
then the probable error^® (r^^)
_ 0.6746 n (n-2 )53
)2-( n-2 ) (
Here n is the number of points in the plot and
is the deviation, in the
y direction, of the i^^ point from the calculated line. During the runs with phosphoric aoid a brown solid separated out. This solid ms Identified as bensidine phosphate.
For this reason
duplicate runs* nitmbers 25 end 27, were made* with and without additi<»i of solid benzidine phosphate,
lo appreciable difference in rate was
found.
^^Gibbs» "The Adjustment of Errors in Practical Science**, Oxford tSaiversity Press, London, 1929, p. 89.
36
Table 5 Catalysis by Phosphoric Acid at Various pH's J^Phosphoric acid]
+ [sodium phosphat^
= 0.57 molar
(I0®k)
lO^k* Sun Ho.
pH
(minTl)
22
2.20
2,425
47
23
2.30
1,972
0
24
2.39
1,652
0
25
2.41
1,634
1&
26
2.48
1,401
4
27
2.41
1,590
6
23
2.48
1,266
5
(mint!)
"'In calculation of the amount of lithium parohlorate to be added it was assuBted that, at the low degrees of neutralization invol-ved,
only uniTalent ions are present.
Variation of Rate with Concentration of Undissooiated Acid The second test used in investigating whether the acid catalysis is
60
general is the test usually applied use of em aoid and its sodium salt.
.
The solutions were buffered by
The ratio of acid to salt was
constant throughout a series* but the absolute amount of aoid varied from rune to run.
Constant ionic strength was maintained by adding the
®®Hammett, "Physical Organic Chemistry", McGraw-Hill Book Co., Inc., New York, 1940, p. 221.
37
oaloulftted amount of potassium chloride.
The effeotiyeness of the buffer
action ms oheoked by measuring the pH of each solution before use.
A
Tariation in rate from run to run in a series* under these conditions, indicates a dependence on the concentration of undissociated acid. fhe acids used in this section of the work were forraio, glycolici ohloroacetiof salicylic and cyanoaoetic acids.
fhe rate corresponding
to zero coneimtrati^ of undissociated acid« and, therefore* due to catalysis by hydrogen ion alone, was obtained from runs using hydrochloric aoid > potassium chloride solutions of the same pH and ionic strwQgth. Some of the results for rearrangement catalysed by glycolic aoid are reproduced in Figure 4. in the diagram.
The concentrations of the aoid are indicated
fhe results are presented in fables 6 to 10, and are
collected in graphical form in Figure 5.
38
Table 6 Catalysis by Formlo Acid at Fixed pE
Aoid
loS
Run Ko.
Molarity
(minT^)
61
2.35
92
62
1.68
62
63
2.06
76
65
1.40
56
66
2.24
83
67
2.14
74
68
1.96
67
59
Tabl® 7 Catalysis by Glyoolio Acid at Fixed pH
io\ Rua. Ho.
Aoid Molarity
(loint^)
80
2.64
309
81
2.40
244
82
2.16
196
83
2.06
141
84
2.26
214
85
2 .52
272
66
2.44
279
87
1.97
175
88
2.11
189
89
2.36
210
90
2.30
205
40
Table 6 Catalysis
Chloroaootlo Acid at Flxsd pH
lo^k Run Ho.
Acid Molarity
(niaT^)
37
1.31
201
38
1.26
201
39
1.20
180
40
1.16
168
41
1.10
169
42
1.04
160
43
0«99
148
44
0.94
127
49
1.66
214
SO
1.63
208
61
1.33
196
62
1.69
200
63
1.26
191
64
1.39
201
66
1.66
268
6?
1.46
196
68
1.19
160
69
1.00
171
41
Table 9 Catalysis by Salioylio Aoid at Fixed pH
io\ Run He*
Acid Molarity
(miaT^)
69
1.38
264
70
1.33
286
71
0.96
166
72
1.28
208
73
1.12
187
74
1.01
157
75
1.17
208
76
1.07
189
77
1.22
240
42 Table 10 Catalysis by Cyaaoaoetlo Aold at Fixed pH
10*k Rtm No.
Aoid Molarity
(min."^)
91
0.404
161
93
0.275
90
94
0.220
111
96
0,350
159
96
0.366
162
98
0.294
106
99
0.312
137
100
0.386
164
101
0.257
106
102
0.276
106
105
0.147*
55
*pH 2.49
The oatalytio effect of each aoid oan be represeated, approxiisately, by a llnesur increase of rate with ooaoeatration of aoid with a positiTe deyiation at high concentrations.
The slopes of the linear portions
have been taken as a measvire of the catalytic oonstants of the varioas aoids.
Thus, for a point on the linear section of the rate vs. aoid
43
oono«ntratlon eurro, the speoifio rate constant k is glTen hyt k « 0.0014 •
.
where the k^'s are the oatalytio ocKastants*
These eonstants are listed
in fable 11.
Table 11 Catalytic Constants of Various Aeids
Aoid BA, i
Catalytic Constant k^^
Fornic
0.0027 t 0.0003*
Glycolic
0.0081 + 0.0006
Chleroaoetic
0.0135 • 0.0010
Salicylic
0.0162 * 0.0010
Cyanoacetic
0.0368 + 0.0064
^Concentrations are esqpressed in moles per 1,000 gm. of solution, and rates in min.
Atten^ted Application of the ^onsted Relation Maoy reactions whioh are subject to general aoid catalysis hare been found to satisfy the Bronsted relation^^t - g * »pkhai '
^^Ironsted and Pedersen, Z. phys. Cheia.» 108, 18S (1924).
44 «ker9 ftpplleation.
is confined to one reaoti<»i of one substrate, and the
•arious acids
are all of the same charge type.
A plot was made of the logarithms of the catalytic constants, listed in Table 11, against the pE values of the acids in water at
A
similar plot was made against the approximate values of the pi values of the acids in S(^ aqueous ethanol at 26°U., which are reported in the literature®^'®®.
In both cases no linear relation was foxmd,
Attetnpted Isolation of the Rearrangement Step Sinoe the dioonjugate acid of hydrasobensene might be the species which actually rearranges, an attei^t was made to prepare that acid and to compare its rate of rearrangem^t with that of the overall reaetim. It has not been clearly established whether the dieonjugate acid is, indeed, the product of the reaction of methyl iodide and i^drasobenzene at room tei^erature •.
In an ethanolic solution of the confound there
is an initial decolorising of added aqueous Bindsohedler's Green. Such deoolorixation is no longer observed after a few seconds.
Thus
it would seem that there is some hydrazo- ucaipound present in the etlmnolic solution, but that rearrangement occurs very rapidly on the addition of a trace of water.
^^Minniek and Kilpatrick, £. Phys« Chem., 4S, 269 (1939). ®®Hamed and Bone, Chem, Soc,, 63, 2579 (1941), and other references cited therein.
45
2]
^
^
^
'
"Zl
pH-Meter Reading Figure 3 .
L o g a r i t h m of
Phosphoric
0,
Rate
vs.
Apparent
Chl&roacetic • ,
Formic
pH.
o.
46
rO
2.06 M 0.5 2.26 M
2.64 M.
50
100
150
Time (min.) Figure 4 . First for c a t a l y s i s by
Order R a t e Glycolic A c i d
Plots of vorious
concentrations.
47
C E
0
)
Molarity Figure 5 .
Rate
C y a n o a c e t i c ®,
vs.
of
Molarity
Salicylic • ,
Gly colic
Acid
CD,
of
Acid
at
Chloroacetic o,
Formic 0 .
constant
pH.
48 DISCUSSION
Lisaitatloas Inherent in the Bxperimsntal Method
An investigation of the type of acid catalysis to which the bensidine rearrangement is subject has to contend with several conflicting require ments.
The conditions have to be so adjusted that the rate of the
hydrogen ion catalysed reaction does not msk rearrangement catalysed by undissociated acid.
Since the catalytic constant of hydrogen ion is
relatively large, the hydrogen ion concentration must be kept low and constant*
This entails buffering of the solutions and requires use of
fairly weak acids* correspondingly low.
The catalytic constants of such acids eu*e, however, The rates of rearrangement catalysed by these
acids are, therefore, inappreciable, unloss both high tes^eratures and high concentrations of the acids are ea^loyed* The use of elevated teD|>erature8 is limited by two factors. Disproportionation of hydrazobenxene is likely to become of ii^ortance at high teisperatures.
The second litnitation is introduced by the use of
aqueous ethanol as solvent* suggested.
The utilisation of another solvent is thus
Water, for exaaple, would be ileal, since it is higher
boiling and since pE readings for aqueous solutions are interpretable. However, hydrazobentene is insoluble in water, and soluble hydrazobensene derivatives introduce coi^lications as regards ionic strength.
Other
higher boiling solvents are unacceptable because pH meter reaaings for such solvents are meaningless, at our present state of knowledge. The conditions en^loyed were considered to be the most acceptable
49
coBprosise ones.
There remain the questions of the meaning of constant
pH meter reading in these solutions, and of the mag^itued of the effects produced by variations in the nature of the solvent at high acid concen trations. It has been shown by Dole
fil
that the negative error exhibited by the
glass electrode in non-aqueous solvents can be explained on the basis of solvation of the proton.
Aooording to this author the electrode will
always be in error when the activity of water in the solution differs from unity.
This theoiry predicts that the error will be constant for
solutions in which the activity of water is a constant.
Although this
theory has not been unequivocally established, it suggests that, in the absence of large disturbing effects by dissolved acids, the variation in hydrogen ion concentration is not large. Little can be said about the perturbing effects of the acids on pH readings in such mixed solvents.
There are, however, sovurces of comfort.
Under the conditions ea^loyed in this investigation the observed variation in rates was large compared to the rate of the hydrogen ion catalysed reaction.
The variation is not, therefore, attributable to relatively
minor changes in hydrogen ion concentration.
In addition, the large
changes in rate found with the stronger acids, where smaller concentrations were eaployed, suggest that the variation in hydrogen ion concentration at constant pH meter reading is not a determining factor. Equally io^ortant is the question of variation in the reaction rate^
®^Dole, J. Am. Qhem. Soc., 64, 3095 (1932).
50
at oonstaat hydrogen ion concentration, due to changes in the nature of the medium.
It is knotsn., for exai^le, that the rate of rearrangemaat is
sensitire to the dielectric constant of the solvent. magnitude of the variation that the effect observed
is
with
Here again the
stronger acids supports the contention
due mainly to changes in the concentration of
undissooiated acid. It is felt, however, that the positive deviations from linesurity observed with all acids at high concentrations (Figure 5) nay be due to medium effects. found
no more
This explanation is offered only since the writer has
satisfactory one*
Because of the nature of the analytical technique, the rates studied were those of disappearance of hydrasobeneene and, probably, of its first conjugate acid.
The results are, nevertheless, interpretable in terms of
41 the rates of formation of both bensidine and diphenyllne, since Oarlin has shown that the rates of formation of both these confounds have the same hydrogen ion dependence and the same aotivation energies.
The Inforauition Gained from this Study
It can be seen from Figures 3 «tnd 5 that the rate is dependent on the particular acid catalyst, at constant pE meter reading.
Figure 5
also shows that the rate varies with the concentration of added acid, under the sacts conditions.
V^ithin the limitations is^osed by medium
effects on reaction rates and on the potential differences established at a glass electrode, these conditions may be taken as corresponding to
51
coastancy of hydrogen ion concentration*
Thus, it has been established
that the rata of rearrangement varies with the concentration of undis> sociated acid present. The studies on the oou^ound which has been reported to be hydrasobensene dihydroiodide are less definitive in that no unequivocal evidence concerning the structure of the confound was obtained.
3Q Pongrats
showed that solution of this conpound led to the formation of benzidine. Carlin
42
obtained a solid confound which he believes to be hydrazobenzrae
dihydrochloride, but in aqueous solution only benzidine was found.
The
least that can be said, then, is that no indication has been obtained for the final step being slow.
In fact, all tests applied to the above com
pound suggest that it rearranges very rapidly at room teaperature, in the prssenoe of water.
Iiq)lications of the Kinetics
Consider the acid catalysed reaction of a confound X. X • HA
^1^ XH"^ + A',
XH"^ + A" XH*
X + HA, ^2^
Product +
If XH"*^ is a stronger acid than HA, probably than
(1)
is correspondingly larger
Under these conditions the concentration of XH^ soon attains a
steady state.
The rate of formation of Xh"^ is given byj d[XH^] dt
kl[X][fiAj - k^^[XH''][A"] - kgCXH^ .
52
This gives for the oonoentration of
under steady state conditions.
[3CH*]. "-JA-] - ICj
Then the rate of fornaition of products is: Rate = k2[xii'*']
kjklWCiu] S
Kil*^
Ihen
^
3^^
IIIII—IWI ••••IIIIIIIIII Iiimnii
(2) •
there are two limiting oases which can be considered.
If kg ^ k_^j^, the rate becomes s
k^kJX]^ Rate -»
' ^-l[^ ]
Since [
HA]
and [
A~]
are effectively constant in any nm, and are related
by the equation
[hajkj^ « [h^][a"] .
(s)
the measured rate is Rate » ic^k2%i[H"^][A'] .
The second limiting case is
^ ^2*
measured rate is then
Rate - ki[x][HA] ,
the addition of a proton to X then being rate-determiniag.
(4)
Between these
63
two limits ths dapanuence on acid concentration is a fractional one. Specifically, if the first stop in the rearrangement is the rapid equilibraticm CglgNH - MHCgHg + HA - CgHgHHg - NiiCgHg + A"
,
(6)
where tne first conjugate acid follows a sequence such as outlined for ooi^ovmd X in (1)« then the kinetics can be treated similarly.
In
assigning relative magnituaes to the various rate oonetants it must be borne in mind that the rearrangement is of second order in hydrogen ion, when the catalyst is a mineral acid.
The rate is a function ol' the
concentration of undissociated acid, when weaker acids are en^loyed. Since the addition and removal of the first proton roust
be fast, the
initial step, as written in (6), canniot be kinetioally differentiated from
CgHgKH -
*
Because the rats vaaries in a linear maimer with the ooncentration of undis sociated aoid, at moderate concentrations, the addition of the second proton leads to an equation such as (4) and must be rate-determining. If the deviation from linearity et high concentrations is not due to the change in the nature oi the medium, then the two molecules of undis sociated acid must participate in the slow steps +
+
CgHgNH - HaCgHg + 2HA « CgHgHHg -
*
Such a termoleoular process is not considered probable. Because of the existence of the aoid equilibrium represented in (S), a depenaence on the ooncentration of undissociatea aoid is not kinetioally
54 distingui3hal)le from a hydrogen ion
dependence on the product of the conoeatrations of
and anion A".
• urther, tho system of process
(1) will give
the saiae kinetic results as the concerted processi M. + X
—»• products + HA.
The idea of a concerted process is considered with disiavor for other reasons.
All the work with salts of the diconjugate acid of hydrazo-
benzene suggests that that acid rearranges rapidly*
This cannot be
explained in terms of re-equilibration with the first conjugate acid prior to rearrangement, as such a sequence of steps would not proceed faster than the overall process.
Further, if the process is concerted then the
energies of activation woula, in all probability, be different for the formation of bensidlne and diphenyline. found to be the same.
These energies, however, are
An experimental test of whether or not the process
is concerted could be made by coaparing the rates of rearrangement of hydfazobaneene and of hyarasobenzene labelled with deuterium in the 4 and 4' positions.
The Nature of the Transition-State
The kinetic results show that the benzidine rearrangement, when catalysed by undissooiated acid, proceeds through a transition-state involving an hydrazobenzene molecule, two protons and an acid anion.
It
has not been established whether anions participate in the catalysis by mineral acids.
Ai^ discussion of the relative arrangement of the groups
in, or configviration of, the transition-state must be largely conjectural.
65
The sooat obvious fortoulation of th« rat9*detarmlai &s
that of
a proton transfer froa a neutral acid aiolecule to the neutral nitrogen in the first conjugate acid of hydrasobensene.
If this is the case* the
geometrical relationships between the various participating entities, in this step, can be closely approximated.
The concerted process, on the
other hand, is forimilated ae the transfer of a solvated proton to neutral nitrogen, ifith simultaneous removal of a proton from a para earbon by an anion. Because of the adjacent positively charged nitrogen it is conceivable that the second proton enters the oos^lex at an alternative position, and does not bond to nitrogen in this step.
The proton isig^t, for exatsple,
enter into a ir-coi^lex with the bensene ring to which the neutral nitrogen is attached.
It is believed, howover, that these structures are energet
ically less favorable than the structxire involving two adjacent positive charges, for they involve the partial loss of the resonance energy of one beneene ring, while the energy due to electrostatic repulsion is still appreciable.
The same would be true of structures such as XV and XVI,
which are of the type postulated by Brown oon^lexes*
52
as more probable than TT-
In addition, the course of the actual rearrangement step,
proceeding from such intermediates, would be of oonsiderable oon¥)lexity.
XV
52
XVI
Brown and Pearaall, J. Am. Chem« Soo., 74, li^l (1962).
56
It is therefore suggested that the path of lowest aotiTation energy corresponds to the transfer of a proton from undissociated acid to neutral nitrogen*
It remains to enquire whether my inforcmtion can be obtained
about the location of the transition-state in the reaction oo-ordiz^te. The method which has been used most for gaining information concerning coiifigurations and energies of trsuasition-states is the aemiCIS ea^irioal one of Byring and Polanyi
.
In general* a four dimensional
space is required for the potential energy figure of a three oon^onent system.
Proton transfer reactions allow of a considerable simplification
in that two of the three entities can be treated as stationary during the reaction.
At a first glance it would seem that the system of interest
here could be treated particularly singly by this method.
Thus, because
of the charge distribution in the reactants, a linear ooxifiguration seems probable, in which case a two •> dimensional potential energy diagram suffices.
In addition, in a reaction of this sort, which involves
extensive electron redistribution, the perturbation, or resonance stabil-
C* isation, of the transition*state should be small
.
Information about
such a system should, therefore, be derivable from consideration of the dissociation courves of -0-H and
f-H.
Bell*^^ has made such calculations for systems of the type CH + 0"
—»• C"" + 01.
Unfortunately, in the bengidine rearrangement
we are iaterestad in
®®lyriag and Polanyi,
physik. Chem., B«,
27S (1931).
^^Bvans and Polanyi, Trans. Faraday Society, 34, 14 (1938). ®®Ball and Udwell, Proc. Roy. Soc.. A., m, 114 (1940).
67
= fi -
+ ho
—^
- nh= + o" .
Here the ©leetrostatic field changes during reaction, pjad a large solvation effect is expected.
The niagnitude of this effect is not cal
culable. For this reason only ft haa been undertaken.
simplified and coii^iaratively crude oalculatioa
For the prooese -oa +
-0" • iMH ,
using the valuesi •0 -OH -ii* *nd
+ e
KA
«• —»- -0
+71 Koals*,
-0* • fi* - 110 Kcals., —* e
- 311 £.eals.,
+ —•-SHg
• 206 Kcals
67
the overall energy ohangs is calculated to be -144 Kcals. per mole. increase of solvation energy of a Exodlar magnitude is probable
.
An In
this case the bases of the two potential energy curves woulu be at similar levels.
However, because of the positive charge initially present in the
hydrajBo-coBpound, an additional electrostatic energy is involved.
IJsing
a linear model for the two nitrogens, hydrogen and oxygen, with soiaewhat arbitrary values for the dielectric constants and internucleaur distances, a value of 16 K-oals. per saole was calculated for the added energy of the
®%ier and Mayer, £. Ghem. Pby&,, 12, 28 (1944). e n
Syrkin and Dyatkina, "Structure of Molecules and the Chemical Bond*, Interseience Publishers, Inc., Kew York, 1960, p. 116. ®%ortum and Bockris, "fextbook of Klectrocheaiistry'*, Elsevier Publishing Co>« Amsterdam, 1961, p* 120.
68 transfer prooess.
This would elevate the base of the nitrogen - hydrogen
curve to a point elose to the intersection of that curve with the oxygen hydrogen curve.
The configuration of the transitlon»state« by this token,
is close to that of the dioonjugate aoid.
The Nature of the Rearrangement Step
As the reaotion proceeds beyond the transition-state the potential energy decreases, probably because of the formation of the hydrogen » nitrogen bond.
Beyond this the energy must go through a minisiusi and a
Biaxijausi before formation of tiie products.
The height of the maxiiitutn
2sust be different for tlie formation of bensidine ana of diphenyline, being sioaller for the latter. The faot that the dioonjugate add can rearrange indicates tl:iat there is resonance stabilization of tirie intermediate structures.
40
of Hughes and Ingold
are applicable to this step.
The arguaonts
The initial decrease
in ener^ indioates that interactions between the para carbons is effee> tive over a considerable distance.
The maxista corresponds to the
decreased nitrogen - nitrogen bonding and, possibly, to repulsion between non-bonded oarbon and hydrogen.
59
The fhermodyn&ffiios of Pre-equillbration
It Is interesting to speculate as to whether it oan be deterioined, by a quantitative study of the variation in rate with dielectric constant or ionie strength* if the addition of the second proton or the actual rearrangeinent step is rate-determining*
In order to be able to discuss
this problem it is necessary to investigate the theriBodynamic basis of pre-equilibration• Consider the equilibria bB + oC •
"jaM + nH + ...•
f f l M + n H + . . , . « w W + x X + . . . .
0
and the oorrei^onding thermodynamic equilibrium constants
• ••
.p ./Lj_
(ag)^(aQ)®....
We see that there is a valid equilibrium constant
K -
•aFo,2)/RT
for the equilibrium bB + oC*»...«»iiflif + x X + . . . .
•
Aceording to the theory of absolute reaction rates, the rate* £* of
60
a r«ft9tion is gl-ran by®®
R «
«X— o . h
^
(7)
where o* is the concentration of the activated complex, or transitioiistate, and barrier.
the average frequency of passage over the potential According to this theory the population of the transition-
state can be treated as a molecular species which is in equilibrium with the reactants.
the corresponding equilibrium constant is written
K
»
, TIar r
where the a2.*8 are the activities of the reactants.
fhen
TJar c* - r .
Substituting this value in (7), the rate expression becomes
R -\
Id"?" ^ eK h
. h
$9
t
Glasstone, Laidler and Eyring, "The Theory of Rate Processes'*, HcGraw^Bill Book Co», Inc., New York, 1941.
(6)
61
Squation (8) is a restricted form of the Bronsted rate-equation®®. There has been some speculation as to whether there can be an effectiTe mintenanoe of equilibrium, either before the rate-determining step, or in that stop.
It seems, however, that the maintenance of both
these equilibria can, at least, be closely approximated.
On this basis,
consider the proeess
bB+cC • .... "idi+aa + adi+nN + ...» « transition-state . By appl;^ng equations (6) and (8), the rate is found to be
h
and we oan write
"•o.l *^''0 - «)* • Bowever, if the transit ion-state is intemediate in configuration between and
the rate equation would be
K -
UB)Vsf....
^
h
In "tiie beneidine rearrangement -tee steps in question are C-H-MH- • 6 6 "Z
6 5
CgHgNHg - MHgCgHg
6
5
2
»-produots
- HH^C H 2 6 6 .
^^Bronsted, Z* physilc. Chem*, 102, 169 (1922).
62
The dominaat ohaage in, activity ooefficiente with ohariging ionio strength or dielectric oonstant is due to the charge ana charge-type of the speoies*
Time, oompexiag equations (9) and (lO), it is seen that as ^ applies to a doubly-positive charged species in both oases, both would give the same rat® dependence on, say, ionic strength.
'
Thus, ionic strength and dielectric constant studies can support or discredit the hypothesis that the transition-state bears two positive charges* state*
They can give no information as to the relative location of that This brief analyaes points out that kinetic results which can be
taken to indicate that a certain step is the slow one, can also, often, be interpreted in tersis of a pre-equilibration.
The Bronsted Relation
The Bronsted relation has never been "derived" but Its in^licatlons are clear*
Coitsider rate equation (8),
isS r R. X
r K
$
h as applied to the prooess B +
BE* + Ek « where B is a base such as hydrasobenxene.
a' ,
(11)
If the activities of B and of
hydrogen ion are snaintalned constant, a_„4. will also be constant. BH F\irther, when all acids HA are of the saste charge type,
\/f
o will be
63
constantf too.
Under tnese conditions equation (8) sia^lifies to
R - CK^ .
(12)
where C is constant. The ftronBted
relation Buggests that it is the "Intrinsic acidities"
of the aeids HA nhieh oause the variations in
and, therefore* in H.
fhe stanaaird mathod of measuring aoidities utilises the i^stem HgO + HA - HgO"^ + A" . fhe corresponding equilibrium constants are
the standard ionitation
constants
Kj^ «
" •
'
(is)
*HA
which must be oon^iared with
K
t
A"
» ft
.
A ilA
,4.
BH
Because of th© different charge types involved, Kj^ and K* probably do not vary linearly with one another*
fhis can be seen more readily if
we modify the transition-state so that (11) becomes • HA -
...
• A* .
For this case
^ r
E'*' It
^
a HA
%-
64
and
—
«
. . . H"^/
I .'
..
0
61
Hamaett '
has shown that activity ooeffioi«nt ratios of the form
are constants in media of reasonably high dieleotrio constant,
^m*/h when S and g are bases of the same charge type. no longer holds when the charge types differ.
However, this constancy fhus, non-linearity is to
be expected when the rearrangemmt rate constants are coiapared with the standard ionisation constants of the acid catalysts. MaifflMtt
6X
has defined a function,
which is a measure of the
strength of a solution in transferring a proton to a positively charged base.
This function can be determined colorimetrically, with the aid of
suitable indicators.
It would be of interest to compare the rates of
rearrangement catalysed by mineral acid solutions with the
values of
these solutions. It should be pointed out that too few acids were ea^loyed in the study here reported to enable an analysis of the deviation from linearity. It is possible* for exas^le* that some of the acids used introduce special catalytic influences or specific madiuA effects about which we have no iniormtiou.
®^Hammett, Chem. Rev., 16, 67 (1935).
66
SUMMARY It is shomi that the rate of the acid catalysed benaidlne rearrange ment varies with the nature and the oonoentration o£ undissociated aoid, under conditions of constant ionic strength and constant pH meter reading. The meaolng of these conditions is discussed euad interpreted as indicating approxismtely constant activity of hydrogen ion, at moderate acid concentrations*
It is shown that the kinetic results suggest a
transition-state oon{>osed of an hydrasobeusene molecule, two protons and 8tn aoid anion*
From consideration of the energetics of the reaetion, it
seems that the oonliguration of the transition-state is not far removed from that of the dioonjugate aoid of hydraaobensene. the actual rearrangement step is relatively rapid.
It is proposed that This contention is
supported by the properties of a compound believed to be hydrasobenzene dihydroiodide. The thermodynamics of pre-equilibration is en^jloyed to demonstrate that studies of the ionic strength dependence of the rate cannot show which of the two possible steps is the rateodetermining one*
The
oatalytio constants of the acids do not obey the Bronsted relation. is suggested that a more valid correlation would be with Hammett's Ef. fxudotion*
It
