EFFECT OF CARBON CATALYSTS ON GLYPHOSATE SYNTHESIS
517
spect to active carbons, one may assume that these species,
particularly transitions metal cations such as Fe2 or Fe
+
3+
,
were the active sites for the oxidative decarboxylation of
PMIDA. This hypothesis can be ruled out since we have
verified that the addition of metallic salts such as FeCl2
to catalyst AC40 produced the opposite effect, namely a
rate decrease. Furthermore, SNK carbon issued from the
carbonization of a synthetic polymer and thus containing
no metal impurities was one of the most active catalysts
(
Tables 1 and 3). Therefore, one can conclude that min-
eral impurities including transitions metals ions were not
FIG. 1. Products distribution versus time for the oxidation with air at involved in the mechanism of PMIDA oxidation.
�
� 1
9
5 C of 300 mL of aqueous solution of PMIDA (0.18 mol L ) over 3 g of
Active carbons contain variable amounts of oxygenated
functional groups such as carboxylic and phenolic groups
�
active carbon 4S treated at 900 C under nitrogen.
(
10). It may be assumed that these groups play a role in
the adsorption of PMIDA as well as in the mechanism of
oxidation. Catalyst 4S was submitted to different oxidiz-
ing treatments with NaOCl or with HNO3 to increase the
number ofoxygen-containinggroups(Table 2). The amount
of acidic functional groups titrated with sodium hydroxide
which means that the conversion of PMIDA followed a zero
order rate law. Then, at total PMIDA conversion, the yield
of glyphosate started to decrease because it was converted
into aminomethyl phosphonic acid (AMP) by a second ox-
idative decarboxylation (Scheme 1). These data indicate
that PMIDA molecules are strongly adsorbed on the carbon
surface. The glyphosate concentration, measured by analy-
sis of solutions at total PMIDA conversion, was in the range
�
1
increased from 0.2 mmol g for the untreated carbon to
�
1
1
.3 and 2.9 mmol g for the samples treated with HNO3
and NaOCl, respectively. The specific activities of the lat-
�
1
� 1
for HNO3 and
ters were lower (1.5 and 1.1 mmol h
g
�
1
0
.155–0.16 mol L , i.e., smaller than the initial amount of
NaOCl, respectively) than that of the untreated carbon
�
1
PMIDA (0.17 mol L ). This was attributed to the reten-
tion in the catalyst pores of up to 10% of the glyphosate
formed. It was verified with blank adsorption experiments
carried out under nitrogen on 3 g of carbon suspended in
� � 1
1
(
2.8 mmol h g ); the larger the amount of functional
groups was, the lower the activity. The negative effect on the
reaction rate of acidic functional groups was corroborated
by experiments designed to eliminate these groups from ac-
�
1
�
0
.17 mol L solution of glyphosate at 95 C, that the amount
tive carbon 4S by thermal treatment under nitrogen at 450,
of glyphosate adsorbed on the carbon corresponded well to
the deficit in glyphosate yield. The influence of the initial
pH of the reaction medium on the reaction rate was inves-
�
7
00, and 900 C. A TPD analysis indicated that oxygenated
�
groups were eliminated at temperatures higher than 700 C.
This was in accordance with the NaOH titration indicating
that no acidic functions remained on the carbon treated
at 900 C (Table 2). The specific rate measured on the ac-
tive carbon treated at 900 C was 40% higher than those
�
1
tigated. Reactions carried out with 0.17 mol L of PMIDA
dissolved in 300 mL of 0.1 M solution of H3PO4 (pH 1)
were 14% slower than those conducted in water (initial pH
�
�
1
.5). In contrast, addition of even small amounts of KOH
of samples still containing oxygenated functional groups.
Such thermal treatment was known to create basic groups
resulted in a large drop of activity, e.g., at pH 2.1 the catalyst
experienced a four-fold rate decrease, and it was inactive
at basic pH. The negative effect of alkali addition could be
attributed to the formation of phosphonomethylimido di-
acetate ions which remained in the water solution rather
than being adsorbed on the catalyst.
(
pyrone-type basic groups) (11) which could be responsible
for the catalytic activity.
The structure of active carbons is based on disori-
ented small graphite planes (12) which behave as large
The specific and areal rates of activated carbons, car-
bon blacks, and graphite used without any pretreatment
were compared in Table 1. Carbons with small surface ar-
eas (HSAG 300 graphite, Vulcan XC72R carbon black, and
CBP activated carbon) had low specific activities. However,
the comparison of areal rates indicates that there is no cor-
relation between the activity and the surface area of cata-
lysts which means that the catalytic activity in PMIDA ox-
idation depended on factors other than the surface area of none
the catalyst.
Since low active catalysts (graphite and carbon blacks)
contained negligible amounts of mineral impurities with re-
TABLE 2
Influence of Acidic Groups on the Activity of Carbon 4S
Acidic
groups
Specific
activity rS
(mmol h
Catalyst
treatments (m g
SBET
� 1
Areal activity rA
2
� 1
� 1 � 1
� 1 � 2
� 3
10 )
)
(mmol g
)
g
)
(mmol h
m
1164
1243
0.2
1.3
2.9
� 0
2.8
1.5
1.1
4.0
2.4
1.2
NaOCl/RT
HNO3/RT
�
N2/900 C
940
4.2