Elimination of substituted fluoren-9-ylmethyl benzenesulfonates: Hammett substituent effects at a mechanistic borderline

Article abstract of DOI:10.1135/cccc19991833

Rates of elimination of fourteen substituted fluoren-9-ylmethyl benzenesulfonates have been measured in methanolic sodium methoxide and 90% aqueous ethanolic solutions of triethylamine, trimethylamine and 4-methyl morpholine. For the sodium methoxide, a linear Hammett plot with ρ = 0.74, consistent with reaction by an E2 mechanism, is observed. For the amine bases the Hammett plots are curved, suggesting a transition from an E2 mechanism for electron-withdrawing substituents to an irreversible E1cB mechanism with a smaller value of ρ for electron-donating substituents. The evidence for a change of mechanism is weakened by systematic and random deviations of substituents from correlations which span small changes in reactivity (less than ten-fold), by a surprisingly large value of ρ = 2 implied for the concerted (E2) reaction and by the possible influence of negative hyperconjugation. Nevertheless, it is consistent with independent evidence that the borderline between concerted and stepwise mechanisms is associated with chemically distinguishable reaction paths, even though pronounced carbanion character (and probably a small extent of bond-breaking to the leaving group) ensures a degree of similarity of structure and sensitivity to substituents of their transition states.

Full text of DOI:10.1135/cccc19991833

Fluoren-9-ylmethyl Benzenesulfonates
1833
ELIMINATION OF SUBSTITUTED FLUOREN-9-YLMETHYL
BENZENESULFONATES: HAMMETT SUBSTITUENT EFFECTS
AT A MECHANISTIC BORDERLINE
1
2,
3
Finbar G. LARKIN , Rory A. MORE O’FERRALL * and Donal G. MURPHY
Department of Chemistry, University College Dublin, Belfield, Dublin 4, Ireland;
1
2
e-mail: finbar.larkin@beaufour-ipsen.com, rmof@macollamh.ucd.ie,
3
donal.murphy@yamanouchi.ie
Received July 26, 1999
Accepted October 3, 1999
Dedicated to Professor Otto Exner on the occasion of his 75th birthday.
Rates of elim in ation of fourteen substituted fluoren -9-ylm eth yl ben zen esulfon ates h ave been
m easured in m eth an olic sodium m eth oxide an d 90% aqueous eth an olic solution s of
trieth ylam in e, trim eth ylam in e an d 4-m eth yl m orph olin e. For th e sodium m eth oxide, a lin -
ear Ham m ett plot with ρ = 0.74, con sisten t with reaction by an E2 m ech an ism , is observed.
For th e am in e bases th e Ham m ett plots are curved, suggestin g a tran sition from an E2 m ech -
an ism for electron -with drawin g substituen ts to an irreversible E1cB m ech an ism with
a
sm aller value of ρ for electron -don atin g substituen ts. Th e eviden ce for a ch an ge of m ech a-
n ism is weaken ed by system atic an d ran dom deviation s of substituen ts from correlation s
wh ich span sm all ch an ges in reactivity (less th an ten -fold), by a surprisin gly large value of
ρ = 2 im plied for th e con certed (E2) reaction an d by th e possible in fluen ce of n egative
h ypercon jugation . Neverth eless, it is con sisten t with in depen den t eviden ce th at th e border-
lin e between con certed an d stepwise m ech an ism s is associated with ch em ically distin guish -
able reaction path s, even th ough pron oun ced carban ion ch aracter (an d probably a sm all
exten t of bon d-breakin g to th e leavin g group) en sures a degree of sim ilarity of structure an d
sen sitivity to substituen ts of th eir tran sition states.
Key w ord s: β-Elim in ation s; Mech an istic borderlin e; Ham m ett relation sh ip; Reaction kin et-
ics; Base catalysis; E2 an d E1cB m ech an ism s.
Despite exten sive in vestigation , th e n ature of th e borderlin e between con -
certed an d stepwise reaction s rem ain s an in com pletely solved problem in
m ech an istic organ ic ch em istry1–13. A n um ber of studies of E2 an d E1cB
m ech an ism s of elim in ation poin t to substan tial if n ot com plete con tin uity
between th e two m ech an ism s1. Th us, for base-catalysed reaction s of
β-ph en eth yl h alides, in creasin g th e stability of a poten tial carban ion in ter-
m ediate by substitution of n itro groups in th e ph en eth yl rin g leads to a
gradual loss of sen sitivity of th e reaction rate to th e n ature of th e leavin g
Collect. Czech. Chem. Commun. (Vol. 64) (1999)

1834
Larkin, More O'Ferrall, Murphy:
group2. For un substituted, 4-n itro-, 2,4-din itro- an d 2,4,6-trin itroph en eth yl
substrates, th e ratio of reactivities of brom o an d ch loro leavin g groups de-
creases from 60 to 8.8 to 4.4 to 2.9. Th ese m easurem en ts suggest th at cou-
plin g of carbon –h ydrogen an d carbon -leavin g group bon d-breakin g in th e
E2 tran sition state disappears at th e poin t of m ech an istic ch an ge.
On th e oth er h an d, a com parable study of substituted ph en eth yl
quin uclidin ium ion s im plies a distin ct discon tin uity of leavin g group sen si-
tivity between an E2 elim in ation of 4-cyan o an d 4-acetyl derivatives an d
E1cB reaction of th e 4-n itro com poun d3. A sim ilar con clusion h as been
reach ed for m eth oxide-prom oted elim in ation of 1,1-dim eth yl-2-ph en eth yl
ch loride via E2 an d E1 m ech an ism s4.
Som e years ago, we reported studies of elim in ation an d β-h ydrogen iso-
tope exch an ge of a series of substituted 9-m eth yl fluoren es5,6 (Sch em e 1).
An E2 m ech an ism was assign ed to reaction of brom o an d ch lorom eth yl de-
rivatives (1; X = Br, Cl) on th e groun ds of a sign ifican t leavin g group effect
H
X
MeOH
MeO
CH2
+
+
+
E2
CH2
X
1
2
CH2
X
MeOH
+
SCHEME 1
(kBr/kCl = 7.0) an d positive deviation s from a Taft correlation of rates of iso-
tope exch an ge of n on -elim in atin g substrates (or substrates for wh ich a car-
ban ion in term ediate is form ed reversibly) sh own in Fig. 1. Th is eviden ce is
m itigated by system atic deviation s from th e correlation 6, in particular ac-
celeration of th e rate of exch an ge by γ-alkyl substituen ts7,14 an d th e possi-
ble in fluen ce of β-h alogen h ypercon jugation 1,8,9, wh ich m ake it difficult to
defin e a “n orm al” value of ρ* for th e exch an ge reaction . Th us for th e corre-
lation lin e sh own in Fig. 1 ρ* is 2.6, wh ereas a value of 2.25 h ad been used
previously based on ly on CH3, H, Ph an d OH substituen ts5. Neverth eless,
th e assign m en t of an E2 m ech an ism for th e ch loride, brom ide (an d iodide)
appears to be con firm ed by detailed studies of Brøn sted coefficien ts an d pri-
m ary h ydrogen isotope effects for th e ion isation an d elim in ation reaction s8
Collect. Czech. Chem. Commun. (Vol. 64) (1999)

Fluoren-9-ylmethyl Benzenesulfonates
1835
(alth ough a recen t m easurem en t for fluoride15 is probably con sisten t with
reaction of th at substrate via a carban ion ).
In con trast to th e h alogen s, for carboxylate leavin g groups, an E1cB
m ech an ism with rate-determ in in g form ation of a carban ion in term ediate
h as been assign ed6. Despite som e dispersion between ben zoate, acetate an d
pivalate leavin g groups, perh aps represen tin g an en h an ced polar effect of
carboxylate substituen ts, th ere is n o substan tial deviation from th e correla-
tion lin e in Fig. 1 an d th e sen sitivity of th e reaction to rin g substituen ts of
ben zoate leavin g groups (ρ = 0.42) is sm all com pared with th at of th e ion is-
ation of ben zoic acids in m eth an ol (ρ = 1.63).
Th ese results suggest th at reaction of fluoren -9-ylm eth yl ben zen e-
sulfon ates (1; X = O3SAr), for wh ich th e leavin g groups sh ould be in term e-
diate in reactivity between ch loride an d ben zoate, m igh t react close to th e
m ech an istic borderlin e. In prin ciple, th is is con sisten t with th e sm all (posi-
tive) deviation of th e tosylate group from th e correlation lin e in Fig. 1. De-
tection of a ch an ge from an E2 to an E1cB m ech an ism with in a Ham m ett
correlation of substituted ben zen esulfon ate leavin g groups seem ed an at-
tractive possibility th erefore. Th e observation (or n ot) of a variation in
slope or “break” with in such a correlation would th en seem to offer an ef-
fective way of distin guish in g between a discon tin uity an d “m ergin g” of
tran sition states at th e m ech an istic borderlin e3,10
.
1.0
I
log k
Br
PhSO
2
OTs
Cl
F
0.0
AcO
PhS
MeO
PhCOO
Me N
2
Me CCOO
3
OH
–1.0
FIG. 1
Taft plot of log k versus σ *X for ion isation
(❍) an d elim in ation (❍) of fluoren -9-
ylm eth yl substrates (1) in m eth an olic so-
dium m eth oxide at 25 °C
Ph
H
Me
0.0
0.5
1.0
σ X*
Collect. Czech. Chem. Commun. (Vol. 64) (1999)

1836
Larkin, More O'Ferrall, Murphy:
Fourteen substituted fluoren -9-ylm eth yl ben zen esulfon ates (3) were pre-
pared th erefore an d rates of elim in ation m easured in m eth an olic sodium
m eth oxide. An in itial m easurem en t of a Ham m ett plot of th e rate con stan ts
gave ρ = 0.74. Th is result, coupled with th e positive deviation of tosylate in
H
X
CH2-O3S
3
Fig. 1, seem ed to in dicate reaction by an E2 m ech an ism . However, a report
by Gan dler th at th e E2 m ech an ism is m ore sen sitive to th e stren gth of th e
catalysin g base th an (irreversible) E1cB m ech an ism th en suggested th at use
of a weaker base th an m eth oxide ion m igh t prom ote a sh ift of th e reaction
towards th e E2–E1cB borderlin e. Th e m easurem en ts were exten ded th ere-
fore to reaction s in 90% aqueous eth an olic solution catalysed by th e bases
trieth ylam in e, trim eth ylam in e an d 4-m eth ylm orph olin e.
EXPERIMENTAL
Startin g reagen ts for syn th esis were purch ased from Aldrich Ch em icals or Lan caster Syn th e-
sis. Solution s of sodium m eth oxide for kin etic m easurem en ts were prepared as described pre-
viously6. For buffer m easurem en ts, 4-m eth ylm orph olin e h ydroch loride was prepared by
passin g dry h ydrogen ch loride gas th rough an eth ereal solution of 4-m eth ylm orph olin e. Th e
h ydroch lorides of 4-m eth ylm orph olin e, trieth ylam in e an d trim eth ylam in e were recrys-
tallised from dry eth an ol an d stored in a vacuum desiccator over an h ydrous CaCl2. Buffer
solution s were prepared from doubly distilled water protected from CO2.
Fluoren -9-ylm eth yl aren esulfon ates 3 were prepared from reaction of fluoren -9-yl-
m eth an ol with th e correspon din g aren esulfon yl ch loride. Several of th e sulfon yl ch lorides
were available from Aldrich . However, a n um ber of 3-substituted derivatives were prepared
by a m odification 16,17 of th e San dm eyer reaction developed by Meerwein et al.18
.
Syn th esis of 3-Substituted Ben zen esulfon yl Ch lorides – Gen eral Procedure
To a th ree-n ecked flask fitted with a m ech an ical stirrer were added liquid sulfur dioxide
(30 g), ben zen e (20 m l), cupric ch loride (3 g), potassium ch loride (4 g) an d dioxan (30 m l):
th e flask was cooled in an ice bath . To th e resultin g solution was added a solution of substi-
tuted ben zen ediazon ium ch loride gen erated from th e correspon din g arom atic am in e (0.04
m ol), con cen trated h ydroch loric acid (8.4 m l) an d sodium n itrite (3.1 g). Th e m ixture was
allowed to warm to room tem perature an d th en h eated to 40–50 °C un til evolution of n itro-
gen was com plete (30 m in ). Water (100 m l) was added an d th e aren esulfon yl ch loride was
extracted with ben zen e (3 × 150 m l). Th e com bin ed ben zen e extracts were wash ed with 10%
aqueous sodium h ydroxide (3 × 20 m l) followed by water an d dried over an h ydrous m agn e-
Collect. Czech. Chem. Commun. (Vol. 64) (1999)

Fluoren-9-ylmethyl Benzenesulfonates
1837
sium sulfate. Rem oval of th e solven t gave a liquid wh ich was purified by vacuum distilla-
tion . Ben zen esulfon yl ch lorides prepared by th is m eth od in cluded 3-Cl, 3-CN an d 3-MeO.
For 3-F an d 3-Me derivatives, th e potassium ch loride in th e above preparation was replaced
by m agn esium ch loride, an d a furth er m in or m odification of th e procedure was used for
preparation of th e 3-CF3 derivative19. Th e yields, boilin g poin ts an d NMR spectra of th e
products are recorded in Table I.
Syn th esis of Aren esulfon ates 3
Th e reaction of aren esulfon yl ch lorides with fluoren -9-ylm eth an ol followed th e m eth od de-
scribed below for th e 4-fluoroben zen esulfon ate.
Fluoren-9-ylmethyl 4-fluorobenzenesulfonate. Fluoren -9-ylm eth an ol (3.0 g) an d dry pyridin e
(40 m l) were stirred at –6 °C, an d 4-fluoroben zen esulfon yl ch loride (6.0 g) was added slowly.
Th e reaction m ixture was stirred at 0 °C for 2 h . Water (18 m l) was added slowly an d th e
m ixture extracted with ch loroform (3 × 100 m l). Th e com bin ed ch loroform extracts were
wash ed with sulfuric acid (2 × 100 m l), followed by water (2 × 50 m l) an d 5% sodium
h ydrogen carbon ate (50 m l), an d dried over an h ydrous m agn esium sulfate. Rem oval of th e
solven t un der vacuum gave th e crude solid, wh ich was recrystallised from m eth an ol.
Th e m eltin g poin ts, m icroan alysis an d NMR of th e fluoren -9-ylm eth yl aren esulfon ates are
sum m arised in Table II.
TABLE I
Properties of 3-substituted ben zen esulfon yl ch lorides for preparation of 3
B.p., °C (m m Hg)
(lit.) (m m Hg)
NMR, δ (CDCl3)
X
Yield, %
Cl
F
38
40
115–120 (7.5)
(76–77) (0.8)
77–78 (0.5)
7.5–8.1 (4 H, m , ph en yl)
7.35–8.2 (4 H, m , ph en yl)
(109–110) (3.0)
110–111 (1.5)
86–93 (1.0)
CNa
Me
68
61
7.9–8.6 (4 H, m , ph en yl)
7.5–8.0 (4 H, m , ph en yl)
2.5 (3 H, s)
(93.5–94) (1.0)
86–98 (1.0)
OMe
CF3
58
63
7.2–7.6 (4 H, m , ph en yl)
3.95 (3 H, s, –OCH3)
7.9–8.6 (4 H, m , ph en yl)
75–76 (0.5)
(123–126) (5.0)
a
Solid at room tem perature m .p. 40–41 °C.
Collect. Czech. Chem. Commun. (Vol. 64) (1999)

1838
Larkin, More O'Ferrall, Murphy:
Collect. Czech. Chem. Commun. (Vol. 64) (1999)

Fluoren-9-ylmethyl Benzenesulfonates
1839
Kin etic m easurem en ts were carried out as described previously for elim in ation reaction s
of oth er fluoren -9-ylm eth yl derivatives5–7. In gen eral, a reaction was in itiated by in jection of
15–25 µl of a solution of th e substrate in m eth an ol in to 2 m l of a solution of base in a spec-
troph otom etric cell. For aqueous eth an olic solution s, th e con ten t of eth an ol was adjusted so
th at th e fin al con cen tration of alcoh ol after in jection was 10% (v/v). Reaction s were m on i-
tored usin g a Perkin Elm er 124 spectroph otom eter with a cell com partm en t th erm ostatted at
25 ± 0.1 °C.
RESULTS
Substituted fluoren -9-ylm eth yl ben zen esulfon ates were prepared with th e
followin g substituen ts in th e ben zen e rin g (X in 3): 4-MeO, 4-Me, 3-Me, H,
4-F, 3-MeO, 4-Cl, 4-Br, 3-F, 3-Cl, 3-CF3, 3-CN, 3-NO2, 4-NO2. Th eir reac-
tion s to form diben zofulven e (2) were m on itored spectroph otom etrically
from th e in crease in absorban ce at 258 n m in th e product. First-order rate
con stan ts were m easured at 25 °C for four solven t-base system s com prisin g
m eth an olic sodium m eth oxide an d th e tertiary am in es trieth ylam in e,
trim eth ylam in e an d 4-m eth ylm orph olin e in 90% aqueous eth an ol (v/v).
Secon d-order rate con stan ts were obtain ed from slopes of plots of first-order
rate con stan ts again st base con cen tration . Norm ally five first-order rate
con stan ts were in cluded in each plot but for th e slower reaction s with
4-m eth ylm orph olin e, for som e substituen ts, on ly two first-order rate con -
stan ts were m easured. Th e am in e bases were used in buffer solution s with a
1 : 1 ratio of buffer acid to buffer base an d th e ion ic stren gth was m ain -
tain ed at 0.1 m ol l–1 with added sodium ch loride. In n o case was a “sol-
ven t” reaction detectable in th e absen ce of base. Th e four sets of rate
con stan ts are sum m arised in Table III.
DISCUSSION
Hammett Relationships for Methanolic Sodium Methoxide
Figures 2 an d 3 sh ow Ham m ett plots of rate con stan ts for elim in ation of
fluoren -9-ylm eth yl ben zen esulfon ates in m eth an olic sodium m eth oxide
an d aqueous am in e buffers (con tain in g 10% of eth an ol), respectively. Th e
plot for sodium m eth oxide displays a lin ear correlation of log k with n or-
m al Ham m ett substituen t con stan ts with slope ρ = 0.74. By com parison ,
th e plots for th e am in e bases are subject to con siderable scatter an d are n ot
well correlated by a sin gle straigh t lin e.
Con siderin g first th e reaction in sodium m eth oxide, th e m ost obvious
im plication of th e lin ear depen den ce is th at th e reaction occurs by a sin gle
Collect. Czech. Chem. Commun. (Vol. 64) (1999)

1840
Larkin, More O'Ferrall, Murphy:
m ech an ism , eith er E2 or irreversible E1cB. Th e possibility of reaction by a
reversible E1cB m ech an ism can be excluded because th is is observed on ly
for poorer leavin g groups th an ben zen esulfon ate6.
TABLE III
Rate con stan t for elim in ation of substituted fluoren -9-ylm eth yl ben zen esulfon atesa (102
k/m ol–1 l s–1) in m eth an olic sodium m eth oxide an d 90% aqueous eth an olic solution s of 1 : 1
am in e buffers at 25 °Cb
X
σp
MeO–/MeOH
NEt3
NMe3
NMMc
4-MeO
4-Me
3-Me
H
–0.27
–0.17
–0.07
0.0
148
170
192
214
273
231
342
347
363
397
456
753
–
8.56
8.15
8.51
9.06
10.6
10.3
12.45
–
13.05
13.09
14.20
14.54
18.0
0.502
0.508
0.493
0.620
0.719
0.690
0.590
–
4-F
0.06
0.12
0.23
0.23
0.34
0.37
0.43
0.56
0.71
0.78
3-MeO
4-Cl
(11.21)d
17.1
4-Br
–
3-F
11.9
13.8
16.3
25.1
32.0
30.8
16.8
0.853
(0.580)d
0.704
2.23
3-Cl
16.6
(10.8)d
3-CF3
3-CN
3-NO2
4-NO2
60.7
76.5
3.06
813
62.7
2.85
a
c
b
Cf. structure 3. Ion ic stren gth 0.1; con cen tration ran ge of am in e 0.01–0.05 m ol l–1
.
d
4-Meth ylm orph olin e. Rate con stan ts in paren th esis are n ot sh own in Fig. 3.
1.0
log k
0.5
ρ = 0.74
0.0
0.0
0.4
0.8
σ
X
FIG. 2
Ham m ett plot of log k versus σX for elim in ation of fluoren -9-ylm eth yl ben zen esulfon ates (3)
in m eth an olic sodium m eth oxide at 25 °C
Collect. Czech. Chem. Commun. (Vol. 64) (1999)

Fluoren-9-ylmethyl Benzenesulfonates
1841
To distin guish between th e two m ech an ism s, it is n ecessary to estim ate
th e m agn itudes of ρ to be expected for each . As a guide for th e E1cB m ech a-
n ism , on e m ay take m easurem en ts by Stirlin g for elim in ation of 4-m eth yl
substituted an d 4-n itro substituted 2-(ph en ylsulfon yl)eth yl ben zen e-
sulfon ates (4; X = ArSO3) in eth an olic sodium eth oxide11 (Sch em e 2).
Stirlin g foun d th at th e rate con stan t for elim in ation (kE) of tosylate as a
kE[EtO
]
SO2CH2CH2
X
SO2CH CH2
X
+
k
ex[EtO
]
4
ρ = 4.89
SO2CHCH2
X
SCHEME 2
leavin g group was well correlated by a Taft plot of rate con stan ts for h ydro-
gen isotope exch an ge (kex) of n on -elim in atin g 2-(ph en ylsulfon yl)eth yl sub-
strates (4; X = Ph , H, CH3 e.g.). He reported th at th e differen ce in rates of
tosylate an d 4-n itroben zen esulfon ate leavin g groups correspon ded to a
value of ρ = 0.4 an d suggested th at th is represen ts th e sen sitivity of
rate-determ in in g ion isation to form a ph en ylsulfon yl carban ion towards
–0.5
0.0
Me N
3
log k
–0.5
–1.0
Et N
3
–1.0
–1.5
–2.0
FIG. 3
Ham m ett plots of log k versus σX for elim i-
n ation of fluoren -9-ylm eth yl ben zen esulfo-
n ates (3) in 90% aqueous-eth an olic 1 : 1
am in e buffers at 25 °C; from top to bot-
tom : trim eth ylam in e, trieth ylam in e an d
N-m eth ylm orp h olin e. Th e filled circles
represen t substituen ts lyin g on th e correla-
tion lin e of Fig. 4
O
N-Me
3-CF
3
3-Cl
0.5
–2.5
0.0
σ
1.0
X
Collect. Czech. Chem. Commun. (Vol. 64) (1999)

1842
Larkin, More O'Ferrall, Murphy:
substituen ts in th e ben zen esulfon ate rin g. Th e Taft relation sh ip for th e
ionisation of phenyl sulfones is characterised by ρ* = 4.89 compared with ρ* ≈ 2.6
for fluoren es. Th is suggests th at th e ion isation of fluoren es is about h alf as
sen sitive to substituen ts as th e ph en yl sulfon es an d th at ρ for carban ion
form ation from substituted fluoren -9-ylm eth yl ben zen sulfon ates sh ould be
≈0.2.
For an E2 m ech an ism of elim in ation , Ban ger, Cockerill an d Davis20 h ave
reported ρ = 1.1 for elim in ation of ph en eth yl aren esulfon ates (5) with
t-BuOH/t-BuONa. However, th ese auth ors also n oted th at ρ for substitution
in th e ben zen sulfon ate group depen ded on th e substitution in th e
ph en eth yl group (Y in 5) an d th at for th e substituen ts 4-MeO, H an d 3-Cl, ρ
decreases from 1.24 to 1.08 to 0.94. If th e im portan t factor in th e
ph en eth yl substitution is stabilisation of partial carban ion ch aracter in th e
E2 tran sition state, on e m ay expect ρ for E2 elim in ation of fluoren -9-yl-
m eth yl ben zen esulfon ates to be sm aller again .
Y
X
CH2CH2-O3S
5
From th e m easured value of ρ = 0.74 for elim in ation of fluoren -9-yl-
m eth yl aren esulfon ates th erefore it seem s likely th at th e m ech an ism of re-
action is E2. Th e differen ce in m ech an istic assign m en t from th e 2-(ph en yl-
sulfon yl)eth yl ben zen esulfon ates appears to be reason ably clear cut, but ρ
in th is case was based on on ly two m easurem en ts an d m ay be subject to
som e un certain ty.+
Hammett Relationships for Amine Buffers
Th e Ham m ett plots for th e tertiary am in es presen t a differen t picture from
th at for sodium m eth oxide. In th e first place, scatter from th eir correlation
lin es becom es of in creasin g sign ifican ce as on e passes from th e stron ger to
weaker am in e bases. Th e stron gest base is trieth ylam in e (pKa = 10.65)
wh ich is represen ted by th e m iddle curve in Fig. 3. It sh ould be n oted th at
th is gives lower rate con stan ts th an th e less basic trim eth ylam in e (pKa =
9.76), alth ough th e differen ce appears to be exaggerated in Fig. 3 by th e use
+ Th e rate con stan ts given are 6.7 · 103 an d <5.7 · 104 wh ich yield ρ < 0.9.
Collect. Czech. Chem. Commun. (Vol. 64) (1999)

Fluoren-9-ylmethyl Benzenesulfonates
1843
of a differen t scale for log k for th e two bases to avoid overlap of th e results.
Th e greatest scatter is observed for N-m eth ylm orph olin e (pKa = 7.41) wh ich
is th e lowest curve in th e figure.
Som e of th e scatter in Fig. 3 is system atic in ch aracter an d can be ob-
served also in Fig. 2. Th us Exn er h as drawn atten tion to deviation of th e 4-F
substituen t from Ham m ett relation sh ips21, an d in Fig. 3 th is poin t is gen er-
ally h igh com pared with substituen ts with sim ilar σ values.
An oth er possible reason for th e scatter is th at it reflects a structural differ-
en ce of ben zen esulfon ate leavin g groups from ben zoate an ion s used to de-
fin e Ham m ett σ con stan ts. Th us in a study of elim in ation from fluoren -
9-ylm eth yl ben zoates, a con siderably better correlation with σ was
ach ieved6.
Th e possibility of system atic deviation s from th e latter source can be ex-
am in ed by con siderin g reaction s in wh ich ben zen esulfon ate an ion s are
form ed. Th us Fig. 4 sh ows a Ham m ett plot for reaction of 1-adam an tyl
ben zen esulph on ates in eth an ol22 with ρ = 2.5. Un fortun ately, th e plot in -
cludes on ly a fraction of th e substituen ts used in Figs 2 an d 3, but two dis-
tin ctive features can be recogn ised. Th e first is th at th e 3-NO2 substrate is
m ore reactive th an 4-NO2. Th is is ch aracteristic of solvolyses of ben zen e-
sulfon ates an d is also apparen t in Fig. 2. Kevill h as sh own th at th e relative
m agn itudes of σ for 3-NO2 an d 4-NO2 are solven t-depen den t, an d th at th e
n orm al order of reactivity is restored as th e reaction m edium becom es m ore
aqueous. However, for th e elim in ation in Fig. 3 (in 90% aqueous eth an ol),
th e two substituen ts are eith er of equal reactivity or th e 3-NO2 is th e m ore
reactive, as it is for m eth an olic sodium m eth oxide. Also n oticeable in Figs 2
an d 3 is th e h igh reactivity of th e 3-cyan o substituen t. Un like th e n itro sub-
stituen ts, 3-cyan o was n ot used in Fig. 4. Neverth eless, it seem s reason able
3-NO
–2.5
2
4-NO
2
log k
4-Br
–3.5
–4.5
H
4-Cl
4-Me
FIG. 4
4-MeO
Ham m ett plot of log k versus σX for eth an ol-
ysis of 2-adam an tyl aren esulfon ates at 25 °C
0.0
0.5
1.0
σ
X
Collect. Czech. Chem. Commun. (Vol. 64) (1999)

1844
Larkin, More O'Ferrall, Murphy:
to in fer th at th e deviation s of all th ese substituen ts are system atic rath er
th an ran dom in origin .
A furth er feature of th e correlation of Fig. 4 is th at th e poin ts for CH3O,
CH3, H, 4-Cl an d 4-Br all fit th e lin e well. Th ese poin ts are sh own as filled
circles in Figs 2 an d 3 an d it is apparen t th at in each plot, th ey can be con -
n ected by a satisfactory straigh t lin e. Th e on ly m ajor discrepan cies rem ain -
in g th en are th e poin ts for 3-Cl an d 3-CF3 in th e reaction with 4-m eth yl-
m orph olin e. Th ese substituen ts sh ow a n egative deviation wh ich is n ot ob-
served for th e oth er bases. Th e relevan t poin ts are iden tified by arrows in
Fig. 3.
Havin g assessed th ese deviation s, we m ay ask wh eth er th e correlation s
still h ave ch em ical sign ifican ce. In gen eral, both th e ran dom an d system -
atic deviation s are m agn ified by th e sm all (less th an ten -fold) variation in
rate con stan ts observed, despite th e wide ran ge of substituen ts used. Never-
th eless, we n ote th at for each data set in Fig. 3, substituen ts with low σ val-
ues, in cludin g th e “well-beh aved” substituen ts from Fig. 4 sh own as filled
circles, defin e a Ham m ett plot with a ρ value wh ich approach es zero.
It is apparen t from Fig. 4 th at th e lin e con n ectin g 4-MeO, 4-CH3, H, 4-Cl
an d 4-Br passes between th e 3-NO2 an d 4-NO2 substituen ts. By con trast in
Fig. 3, it is clear th at a straigh t lin e th rough th e filled circles is far from sim -
ilarly accom m odatin g th e 3- an d 4-NO2 groups, wh ich both sh ow stron g
positive deviation s. Such beh aviour im plies a n on -lin ear correlation in
wh ich th e reaction s sh ow a greater sen sitivity to electron -with drawin g th an
electron -don atin g substituen ts.
Such beh aviour is m ost obviously described by com petin g reaction s with
large an d sm all ρ values. Th us if th e experim en tal rate con stan ts in Fig. 3
correspon d to a sum of values k1 an d k2, with differen t reaction con stan ts
ρ1 an d ρ2, th eir depen den ce upon σ would be described by Eq. (1), in wh ich
k1H an d k2H are values for th e un substituted substrate.
log (k1 + k2) = log (k1H 10ρ1σ + k2H 10ρ2σ
)
(1)
Th e curves drawn th rough th e poin ts in Fig. 3 are based on th e ch oices of
ρ1 = 0.2 an d ρ2 = 2.0 for irreversible E1cB an d E2 elim in ation s, respectively.
Values of k1H an d k2H were ch osen to give a best fit to th e filled circles for
low values of σ an d rough ly equal postive an d n egative deviation s for 3-
an d 4-n itro substituen ts for larger values.
Collect. Czech. Chem. Commun. (Vol. 64) (1999)

Fluoren-9-ylmethyl Benzenesulfonates
1845
Th e values of k1H /k2H derived from th is an alysis: 14, 7.3 an d 6.5 for Et3N,
Me3N an d for 4-m eth ylm orph olin e, respectively, im ply an in crease in im -
portan ce of th e con certed m ech an ism as th e stren gth of th e base in-
creases10. In prin ciple th erefore, th e an alysis provides a con sisten t picture
of th e reaction s. In practice, th e value of ρ2 = 2.0 for th e con certed reaction
seem s too large wh en com pared with th e values of Ban ger an d Cockerill for
th e ph en eth yl aren esulfon ates (Sch em e 3). With a sm aller value of ρ2, it is
difficult to fit th e data.
Moreover, in view of th e sm all overall variation in rate con stan ts for th e
am in e buffers (less th an ten -fold), it is h ard to exclude th e possibility th at
th e beh aviour represen ts a variation of substituen t effects with in a sin gle
m ech an ism , perh aps reflectin g th e presen ce of a developin g positive pole
on th e n itrogen base presen t in th e tran sition state. Th ere is am ple evi-
den ce from studies by Wepster an d Hoefn agel th at poles an d dipoles differ
in substituen t beh aviour23. A furth er possibility is th at n egative h yper-
con jugation , iden tified by Kin g as m akin g an im portan t con tribution to
stabilisation of ph en ylsulfon yl carban ion s by oxygen substituen ts, m ay
also play a role in variable stabilisation of fluoren yl carban ion s by aryl-
sulfon ates24. Before a firm m ech an istic con clusion can be drawn , th erefore,
m ore m easurem en ts will be required, e.g. to calibrate substituen t effects in
an auth en tic reaction form in g a carban ion β to substituted aren esulfon ate
groups effected by an am in e base.
Neverth eless, th e results len d som e support to th e view th at th e border-
lin e between E2 an d E1cB m ech an ism s in volves reaction path s wh ich differ
con siderably in sen sitivity to th e n ature of th e leavin g group. In prin ciple,
th is is con sisten t with th e suggestion by Jen cks th at th ere is a sh arp discon -
tin uity of tran sition states at th e E1cB–E2 borderlin e, with a stron g cou-
plin g between carbon –h ydrogen an d carbon –leavin g group bon d-breakin g
in th e E2 but n ot in th e E1cB tran sition state3.
However, Jen cks argued th at th ere is n o eviden ce of competing reaction s
at th e borderlin e an d th at th e reaction path is un ique for each substrate
an d reaction con dition s. As m en tion ed above, rates of reaction of
4-n itroph en eth yl quin uclidin ium ion s sh ow a very weak sen sitivity to th e
n ature of th e on ium ion leavin g group (β1g = –0.15), con sisten t with rate-
determ in in g carban ion form ation with in an E1cB m ech an ism , wh ereas,
rates of reaction of 4-cyan o- or 4-acetylph en eth yl substrates sh ow a m ark-
edly h igh er sen sitivity (e.g., β1g = –0.34), con sisten t with sign ifican t bon d-
breakin g in an E2 tran sition state. Rem arkably, n o eviden ce was foun d of
in term ediate beh aviour between th ese extrem es.
Collect. Czech. Chem. Commun. (Vol. 64) (1999)

1846
Larkin, More O'Ferrall, Murphy:
Th ese results appear to be in con sisten t with a sim ple in terpretation of th e
reaction s in Fig. 2 as com petin g E2 an d E1cB m ech an ism s. Moreover, n ei-
th er our results n or th ose of Jen cks an d Gan dler appear to be con sisten t
with th e sm ooth decrease in kBr/kCl rate ratios as th e carban ion -stabilisin g
effect of a β-substituen t in creases, referred to in th e in troduction to th is pa-
per. As n oted th ere, an in terpretation of th ese results would be th at, at th e
borderlin e, th ere is a “m ergin g” of E2 an d E1cB tran sition states, with th e
sen sitivity to th e n ature of th e leavin g group gradually decreasin g as th e
borderlin e is approach ed.
A sim ilar lack of discon tin uity at th e borderlin e is observed in th e effect
of α-dim eth yl substitution upon elim in ation of fluoren -9-ylm eth yl sub-
strates 6. Th e rate-retardin g effects of two m eth yl substituen ts are sum m a-
rised in Sch em e 3 below for elim in ation s or ion isation s to th e carban ion in
m eth an olic sodium m eth oxide for a series of fluoren -9-ylm eth yl substrates
with differen t leavin g groups. Passin g from left to righ t, th e ease of dis-
placem en t of th e leavin g group decreases in goin g from I, Br an d Cl, wh ich
alm ost certain ly react by an E2 m ech an ism , to OCOCH3 an d OCOPh for
wh ich th e m ech an ism is very probably irreversible E1cB.
H
X
I
Br
Cl
F
OAc OBz OH
6
R
X
k
(R = H)/k(R = Me)
22 111 465 520 920 1120 44
R
SCHEME 3
It is clear th at th ere is a sm ooth tran sition in th e m agn itude of th e ef-
fects, with a value for fluoride, based on a recen t m easurem en t for th e
fluoren -9-ylm eth yl fluoride by Koch 15 an d a m easurem en t for th e
dim eth yl-substituted com poun d by Walsh 25 (k = 3.3 · 10–4 m ol–1 l s–1), fall-
in g close to th at for ch loride.
It is also rem arkable th at th e rate-retardin g effects of m eth yl substitution
are lower for E2 elim in ation th an E1cB (ref.6). On e m igh t h ave expected th e
rate of th e E2 reaction to be retarded both by steric effects to th e plan arity
of th e developin g double bon d in th e tran sition state an d by σ-bon d in ter-
action s between m eth yl an d leavin g groups in th e reactan t. Possibly, th ese
are com pen sated by a favourable electron ic effect on th e relatively dipolar
double bon d of diben zofulven e. However, if on e accepts th e greater m agn i-
tude of th e effect of m eth yl substitution on carban ion form ation , an an om -
aly is presen ted by th e h ydroxy group, for wh ich th is effect is twen ty-fold
Collect. Czech. Chem. Commun. (Vol. 64) (1999)

Fluoren-9-ylmethyl Benzenesulfonates
1847
less th an for oth er carban ion -form in g reaction s6. It was earlier suggested
th at th is arose from n on -additivity of substituen t effects, wh ich m igh t de-
rive from th e differen t size of carboxylate an d h ydroxy substituen ts or from
th e h ydroxy group, providin g a favourable in tram olecular solvation of th e
carban ion 6. Th e possibility th at th e isotope exch an ge m easured for th e al-
coh ol was accelerated by a com petin g retroaldol con den sation to form ace-
ton e an d a m ore reactive isotopically substituted fluoren e appears to be
excluded by TLC an alysis of th e products of th e reaction con ducted on a
preparative scale wh ich failed to detect th e presen ce of fluoren e or ace-
ton e25
.
In con clusion , wh at occurs in th e m ech an istic region between con certed
an d stepwise m ech an ism s of elim in ation seem s still to be un clear. Th us th e
question 10 “How does a reaction ch an ge its m ech an ism ?” m erits furth er in -
vestigation both from an experim en tal poin t of view an d from th e poin t of
view of th eoretical in terpretation .
REFERENCES
1. Gandler J. R.: The Chemistry of Double-Bonded Functional Groups (S. Patai, Ed.), Vol. 2,
p. 733. Wiley, New York 1989.
2. a) Gandler J. R., Storer W. J., Ohlberg D. A.: J. Am. Chem. Soc. 1990, 112, 7756;
b) Gandler J. R.: J. Am. Chem. Soc. 1984, 106, 130.
3. Gandler J. R., Jencks W. P.: J. Am. Chem. Soc. 1982, 104, 1937.
4. a) Thibblin A.: J. Am. Chem. Soc. 1989, 111, 5412; b) Quingshui Meng, Thibblin A.:
J. Am. Chem. Soc. 1995, 117, 1839; c) Quingshui Meng, Thibblin A.: J. Am. Chem. Soc.
1995, 117, 9399.
5. a) More O’Ferrall R. A., Warren P. J.: J. Chem. Soc., Chem. Commun. 1975, 483; b) More
O’Ferrall R. A., Warren P. J., Ward P. M.: Acta Univ. Uppsol., Symp. Univ. Uppsol. 1978, 12,
209.
6. More O’Ferrall R. A., Larkin F., Walsh P.: J. Chem. Soc., Perkin Trans. 2 1982, 1573.
7. Larkin F., More O’Ferrall R. A.: Aust. J. Chem. 1983, 36, 1831.
8. Thibblin A.: J. Am. Chem. Soc. 1988, 110, 4582.
9. Olwegård M., McEwen I., Thibblin A., Ahlberg P.: J. Am. Chem. Soc. 1985, 107, 7494.
10. Jencks W. P.: Chem. Soc. Rev. 1981, 10, 345.
11. Marshall D. R., Thomas P. J., Stirling C. J. M.: J. Chem. Soc., Perkin Trans. 2 1977, 1914.
12. Banait N. S., Jencks W. P.: J. Am. Chem. Soc. 1990, 112, 6950.
13. a) Thibblin A.: Chem. Soc. Rev. 1993, 427; b) Jencks W. P.: Acc. Chem. Res. 1980, 161; c)
Cho B. R., Kim Y. K., Yoon C.-O. M.: J. Am. Chem. Soc. 1997, 119, 691; d) Richard J. P.,
Jencks W. P.: J. Am. Chem. Soc. 1988, 110, 4582; e) Koch H. F., Lodder G., Koch J. G.,
Bogdon D. J., Brown G. H., Carlson C. A., Dean A. B., Hage R., Han P., Hopman J. C. P.,
James L. A., Knape P. M., Roos E. C., Sardina M. L., Dawyer R. A., Scott B. O., Testa C. A.,
III, Wickham S. D.: J. Am. Chem. Soc. 1997, 119, 9965.
14. Bordwell F. G., Drucker G. E., McCollum G. J.: J. Org. Chem. 1976, 41, 2786.
15. Koch H. F.: Personal communication.
Collect. Czech. Chem. Commun. (Vol. 64) (1999)

1848
Larkin, More O'Ferrall, Murphy:
16. Gilbert E. E.: Synthesis 1969, 3.
17. Prinsen A. J., Cerfontain H.: Rec. Trav. Chim. Pays-Bas 1965, 84, 22.
18. Meerwein H., Dittmar G., Gollner R., Hafner K., Mensch F., Steinfort O.: Chem. Ber.
1957, 90, 84.
19. Yale H. L., Sowinski F.: J. Org. Chem. 1960, 25, 1824.
20. Banger J., Cockerill A. F., Davis G. L. O.: J. Chem. Soc. B 1971, 498.
21. Exner O.: Advances in Free Energy Relationships (N. B. Chapman and J. Shorter, Eds), p. 1.
Plenum, London 1972.
22. Kevill D. N., Kolwyck K. C., Shold D. M., Kim C.-B.: J. Am. Chem. Soc. 1973, 95, 6022.
23. Hoefnagel A. J., Hoefnagel M. A., Wepster B. M.: J. Org. Chem. 1976, 41, 2786.
24. a) King J. F., Rathore R.: J. Am. Chem. Soc. 1990, 112, 2001; b) Li M., King J. F.: Fifth
International Conference on Heteroatom Chemistry, London, Canada, July 1998. Abstracts
p. 32.
25. Walsh P. A.: Ph.D. Thesis. National University of Ireland, Dublin 1976.
Collect. Czech. Chem. Commun. (Vol. 64) (1999)