Theoretical Considerations and Control Experiments for
Isotope Effects. No reference compounds with defined DID exist.
During the sample preparation, the DID could be influenced by
isotope effects and hydrogen exchange with the solvent. A
straightforward way to test for such undesired effects would be
to follow the δD of a glucose sample through the whole protocol.
However, no established protocols exist for δD measurements of
the glucose derivatives, and the variable number of hydroxyl and
methyl groups is difficult to account for. Therefore, three control
experiments were performed to rule out isotope effects and isotope
exchange.
Csleads to a discrimination of less than 1%, which is below the
precision of DID measurements by NMR. The mechanism of
formation of derivative 3 is a series of trans-esterifications, initiated
by deprotonation of glucose hydroxyl groups. Secondary KIEs
have not been described for such reactions and are most likely
small. The very high yield of ether formation (>95%) further
minimizes any D discrimination. In summary, D discrimination
by isotope effects is negligible during sample preparation. Note
that for reproducible yields (Table 1), even a substantial secondary
KIE cannot reduce the precision of DID measurements, because
the induced D discrimination would be constant. However, the
constant D discrimination would degrade the accuracy of the
measurements.
Primary kinetic isotope effects (KIE) occur when an atom is
directly involved in a chemical reaction and can strongly influence
isotopomer abundances. In the sample preparation, primary KIEs
could occur if C-H bonds were broken, which would be detect-
able as hydrogen exchange with the solvent and epimer formation.
In control experiment 1, the reaction mixtures were analyzed after
In control experiment 3, independent δD measurements on
cellulose samples were compared with NMR measurements on
derivative 3. From the δD values of cellulose (-52‰ vs VSMOW)
and derivative 1 (-87.7‰), δD of the methyl groups of derivative
1 was calculated via isotope mass balance to be -108‰ vs
VSMOW. Because the methyl groups of derivative 3 originate from
derivative 1, this value was used as the δD value of the methyl
groups of derivative 3. The cellulose samples of group II (with
1
each step by H NMR for epimer formation. Epimers were only
found after wood hydrolysis (sample groups III and IV), but these
are naturally present in wood (mannose, arabinose, galactose, and
xylose). This agrees with previous results that acid hydrolysis of
2
9
19
cellulose occurs without epimer formation. Breakage of C-H
bonds could occur in acid-catalyzed reactions (steps A-C) and
would give rise to hydrogen exchange with the solvent. In control
experiment 2, steps A-C were carried out with D-enriched solvent,
to test for hydrogen exchange. Hydrogen exchange was below
the 0.2% detection limit (data not shown). From the absence of
epimers and of hydrogen exchange, it follows that C-H bonds
are not broken during the sample preparation. Therefore, D
isotopomer abundances cannot be affected by primary KIEs.
Secondary kinetic isotope effects occur when a hydrogen atom
is in close proximity to an atom implied in a chemical reaction.
They are generally small (<1.1) but could affect any reaction step.
For cellulose hydrolysis, the greatest expected secondary KIE is
of R type (during C1-O bond breakage) and should be 1.09 at
published δDs which varied by 45‰) were converted into
derivative 3 and measured by deuterium NMR. From the NMR
measurements and the δD value of the methyl groups, δD of the
glucose moiety of derivative 3 for each sample was calculated
using eq 2 (see Experimental Section). The δD values calculated
in this way correlated linearly with the published δD values
measured by IRMS (R ) 0.95, P < 0.0001, n ) 9). The slope and
intercept of this relationship were not significantly different from
1 and 0 (values within 90% confidence interval), respectively,
indicating that the δD values calculated from NMR measurements
and measured by IRMS agreed within the measurement accura-
cies.
Influence of Purifying R-Cellulose on the Measured DID.
The standard method for δD measurements of tree rings requires
purification of R-cellulose for nitration. Because our protocol yields
a pure glucose derivative, prior purification of cellulose is not
necessary. This assumes that the glucose units incorporated into
cellulose and other wood polymers come from a common glucose
pool with a common DID. To verify this, we compared the DID
of derivative 3 obtained from whole wood (sample group III) and
from R-cellulose extracted from the same wood. There was no
significant difference in the relative abundance of each isotopomer
3
0
the C1-H position. For 92% yield (Table 1), this KIE could
3
1
deplete the product by 2% in the D1 isotopomer. However, we
did not observe any byproducts (such as oligomers of glucose or
its decomposition product, hydroxymethyl furfural)32 in the
reaction. This indicates that the hydrolysis of dissolved cellulose
was complete and that the 92% yield simply reflected incomplete
physical dissolution of cellulose, which is not subject to isotope
effects. This agrees with the lower yield for wood hydrolysis as
compared to cellulose, because other polymers present in wood
hinder the access of acid to cellulose fibers. Thus, no significant
secondary KIE during the hydrolysis of cellulose is expected.
During the formation of derivative 2, the isopropylidene groups
are formed and hydrolyzed via protonation of the carbonyl oxygen
atom of acetone and attack on OH groups of glucose. Secondary
KIEs on the C-H groups of the glucose would therefore be of â
type (concerning an H bound to the carbon next to the reacting
OH). However, â secondary KIEs range only up to 1.04 and have
only been observed in reactions involving carbonium ions.33
Secondary KIEs on the C-H groups of glucose should therefore
be smaller than 1.04, whichsfor 90% overall yield of steps B and
x
(RA ), expressed relative to the methyl groups, between the two
samples (unpaired t-test, n ) 4 spectra, P > 0.1; Table S-1 of the
Supporting Information). The minimum difference between whole
wood and cellulose that was detectable was around 0.06, based
on an average standard error of 0.02. Thus, the tedious purification
of R-cellulose can be omitted in analyses of DIDs of tree rings,
especially if samples to be compared are processed identically.
Starting the sample preparation from whole wood results in
increased glucose recovery as it includes all glucose units present
in wood and avoids loss of material during the purification of
R-cellulose.
Variation in Tree-Ring DID and Interpretation. DIDs of
all compounds analyzed to date were nonrandom; therefore, it is
likely that the difficulties in interpreting tree-ring δD arise from
nonrandom DIDs. Measurements of DIDs of tree-ring cellulose
are presented in Figure 3, which compares DIDs of early and late
wood of the same tree ring of a Douglas fir (group III sample) as
(
(
30) Bennet, A. J.; Sinnott, M. L. J. Am. Chem. Soc. 1986, 108, 7287-7294.
31) Melander, L.; Saunders, W. H. Reaction Rates of Isotopic Molecule; Wiley:
New York, 1980.
(
32) Saeman, J. F. Ind. Eng. Chem. 1945, 37, 43-52.
(
33) Westaway, K. C. In Secondary and Solvent Isotope Effects; Buncel, E., Lee,
C. C., Eds.; Elsevier: Amsterdam, 1987; pp 275-392.
8410 Analytical Chemistry, Vol. 78, No. 24, December 15, 2006