10.1007/s00107-003-0448-8

Abstract Volume changes due to modification with acetic and hexanoic anhydride are due to the volume occupied by the reagent and an associated void volume. The void volume is larger in hexanoic anhydride modified wood. Less weight of water per gm of unmodified wood was accommodated by acetic anhydride modified wood than by hexanoic modified wood, at equivalent WPG. A non-linear relationship was found between weight of water per gm of unmodified Corsican pine wood and bulking, whereas a linear relationship would be predicted. However, this takes no account of void volume. When the value of void volume is deducted from the bulking a linear relationship was indeed obtained with acetic anhydride, but not with hexanoic anhydride modified Corsican pine. It seems therefore that the free volume model offers a reasonable explanation of the differences in swelling recorded for Corsican pine sapwood modifying to varying wpg rsquo

s with acetic anhydride.

Zusammenfassung Volumenänderungen aufgrund einer Modifikation mit Essig- und Capronsäureanhydrid beruhen auf dem Volumen, das die Rektanten einnehmen sowie dazu gehörigen Leerräumen. Letztere sind größer im Fall der Modifikation mit Capronsäure. Das mit Acetanhydrid modifizierte Holz nimmt pro Holzmasse weniger Wasser auf als das mit Capronsäure modifizierte, bezogen auf die prozentuale Massenzunahme (WPG). Eine nicht-lineare Beziehung ergab sich zwischen Wasseraufnahme und Volumenzunahme pro Gramm für nichtmodifiziertes korsiches Kiefernholz, wofür eine lineare Beziehung zu erwarten wäre. Dabei ist allerdings nicht das Volumen der Leerräume berücksichtigt. Wird dieses von der Sperrigkeit abgezogen, so ergibt sich tatsächlich eine lineare Beziehung nach der Modifikation mit Acetanhydrid, nicht aber mit Capronanhydrid. Es scheint also, daß das Modell der Leerräume eine vernünftige Erklärung liefert für das unterschiedliche Quellverhalten, das an korsischen Kiefern-Splintholz beobachtet wurde.

1 Introduction

Various studies of the relationship between the swelling of the material due to modification have been published. In studies of the acylation of spruce, maple and balsa, it has been found that the degree of swelling of the substrate exhibits a proportional relationship with the degree of substitution, up to an acetyl content of 16–18% (Stamm and Tarkow 1947). Furthermore, the volume degree due to modification, has been found to be equal to the volume of the acetyl groups in the wood (Stamm and Tarkow 1947; Rowell and Ellis 1978).

More recently, this latter finding has been disputed, where it has been found that for Corsican pine modified wood, a volume increase larger than theoretically predicted is obtained (Hill and Jones 1996; Papadopoulos 2001).

The volumetric swelling of wood due to esterification, has been studied, and the effect of such modification upon the dynamic mechanical properties of wood determined (Nakano 1994). In this work, it was considered that volume increase of the wood was the sum of volume occupied by chemically bonded reagent, plus a void volume created in the cell wall polymeric network around the adduct.

When wood takes up moisture into the cell wall, the walls swell volumetrically in proportion to the volume of the water absorbed (Skaar 1988). This is based on the assumptions that the cell lumina is constant in size and that there are no voids in the cell wall and therefore, water simply adds its volume to that of dry wood.

The first assumption relies on modification either causing an increase in swelling (due to saturation) into the cell lumina or a reduction in the way in which the lumen size increases due to water swelling. Cell lumina in some wood species have been found to enlarge or shrink when wood is saturated and swells, though where this occurs it is often only by a small amount, and negligible compared to the swelling of the cell wall (Stamm 1964; Siau 1995). It was assumed therefore that no such changed occurred. Any change in dimension due to water swell was considered to be solely due to the swelling of the wood cell wall.

It was the aim of this paper to test the validity of the second assumption. For this reason, wood was modified with two linear chain carboxylic acid anhydrides, namely acetic and hexanoic.

2 Experimental

2.1 Wood modification reactions

Sapwood samples of dimension 20 mm×20 mm×5 mm (radial×tangential×longitudinal) were cut from freshly-felled kiln dried Corsican pine. Samples were carefully smoothed with sandpaper to remove loosely adhering fibres, then placed in a Soxhlet extractor for solvent extraction using toluene/methanol/acetone (4:1:1 by volume) for eight hours. Samples were dried in an oven for 8 h at 105°C. Samples were removed from the oven, transferred to a vacuum desiccator and allowed to cool to ambient temperature over silica gel. Prior to reaction, each sample was weighed on a four figure balance and the dimensions determined using a micrometer (accurate to±0.01 mm). Samples (five replicates) were then vacuum impregnated with dry pyridine (over KOH) for one hour, then transferred to a flask containing pyridine set in an oil bath at 100°C. Samples were allowed to equilibrate in the hot pyridine for one hour. After heating for one hour, the sample batch was transferred to a round bottom flask containing a one molar solution of the anhydride in pyridine set in an oil bath at 100°C. Samples were added at various time intervals so as to give a range of weight percent gains (reaction periods from 15 min to 7 h). At the end of the reaction period, the flask was removed from the oil bath, the hot reagent decanted off and ice cold acetone added to quench the reaction. Samples were kept in the acetone for 1 h, before being transferred to the Soxhlet apparatus for solvent extraction, as previously detailed. Samples were then oven dried at 105°C for 8 h, weight gain due to reaction recorded and dimensions taken.

Weight percent gain (wpg) due to reaction was calculated according to the well known formula:

${\text{wpg}} = {\left[ {{{\left( {w_{{{\text{mod}}}} - w_{{{\text{unmod}}}} } \right)}} \mathord{\left/ {\vphantom {{{\left( {w_{{{\text{mod}}}} - w_{{{\text{unmod}}}} } \right)}} {w_{{{\text{unmod}}}} }}} \right. \kern-\nulldelimiterspace} {w_{{{\text{unmod}}}} }} \right]} \times 100$

where w_mod is the mass of the chemically modified wood and w_unmod is the mass of the unmodified wood.

Molar volume of acyl groups in wood was calculated according to:

${{\left( {V_{{{\text{mod}}}} - V_{{{\text{unmod}}}} } \right)}} \mathord{\left/ {\vphantom {{{\left( {V_{{{\text{mod}}}} - V_{{{\text{unmod}}}} } \right)}} M}} \right. \kern-\nulldelimiterspace} M$

where V_mod is the volume of the modified sample, V_unmod the volume of the sample prior to modification and M is the number of moles of the adduct (Hill and Papadopoulos 2002).

2.2 Swelling measurement

The method for controlling relative humidity is described by Stamm (1964), has been widely used and was selected for being simple, economical and reasonably precise. Test samples were kept above saturated solutions of various salts in containers stored in a controlled temperature room set at 20°C (variation±1°C). Pure water results in the saturated vapour pressure corresponding to 100% relative humidity. The addition of a solute to water reduces its vapour pressure in proportion to its mole fraction in the case of diluted solutions. When a saturated solution at a controlled temperature is used, a constant relative humidity is maintained (Siau 1995).

Six salts were chosen and these are listed in Table 1, along with the relative humidity of the atmosphere above each saturated solution at 20°C (according to Kaye and Laby 1973). They were chosen on the basis of giving minimum relative humidity variation with changes in the temperature (Stamm 1964). Data published by Kaye and Laby (1973) show the equilibrium relative humidity above saturated solutions of these salts to be insensitive to any variation in temperature expected in the controlled temperature room (a variation around 20°C of ±5°C causing a maximum variation of ±1% relative humidity). Excess salt was always present within each solution to ensure saturation was maintained. The solution and air in the container were agitated by bubbling air through the solution.

Table 1 Saturated salt solutions used and their resultant relative humidities at 20Â°C

Tabelle 1 GesÃ¤ttigte SalzlÃ¶sungen und die entsprechenden rel. Feuchten bei 20Â°C

Salt	Relative humidity (%)
Potassium nitrate (KNO₃)	93
Sodium chloride (NaCl)	76
Sodium dichromate (Na₂Cr₂O₇)	55
Potassium carbonate (K₂CO₃)	44
Potassium acetate (CH₃COOK)	23
Lithium chloride (LiCl)	12

Selected wpg rsquo s of oven dry wood samples were placed in the containers above saturated salt solutions. They were left to equilibrate for 4 weeks and then weighed once a week, using a four-place analytical balance until it became obvious that no significant weight change had occurred since the last weight was recorded (and EMC had been attained). Equilibrium moisture content was reached within 6 weeks for all but the two highest humidities, which required longer exposure times. Furthermore, it was observed that at each relative humidity, the time required for the samples to attain EMC increased as the molecular size of the adduct increased, i.e. samples modified with acetic anhydride attained EMC quicker than those modified with hexanoic anhydride, at comparable wpg rsquo s.

Equilibrium moisture content was calculated as follows:

${\text{EMC}}{\left( \% \right)} = \frac{{{\text{Weight}}\;{\text{of}}\;{\text{wet}}\;{\text{modified}}\;{\text{wood}} - {\text{Weight}}\;{\text{of}}\;{\text{dry}}\;{\text{modified}}\;{\text{wood}}}} {{{\text{Weight}}\;{\text{of}}\;{\text{dry}}\;{\text{modified}}\;{\text{wood}}}} \times 100$

Swelling was calculated as follows:

${\text{S}}{\left( \% \right)} = \frac{{{\text{Volume}}\;{\text{of}}\;{\text{wet}}\;{\text{modified}}\;{\text{wood}} - {\text{Volume}}\;{\text{of dry}}\;{\text{modified}}\;{\text{wood}}}} {{{\text{Volume}}\;{\text{of}}\;{\text{dry}}\;{\text{modified}}\;{\text{wood}}}} \times 100$

3 Results and discussion

3.1 Determination of void volume

Modification with reagents leads to an increase in the volume of the wood samples, by determining the volume increase, and dividing this by the number of moles of adduct in the modified samples, a value for the molar volume for each of the reagents at different wpg rsquo

s is obtained. This variation in molar volume in the wood, with extent of modification is illustrated in Fig. 1. The lines represent best fit curves to the data points. Both modificants showed the same behaviour, in that larger molar volumes were found at lower levels of substitution, with an asymptotic decrease to a stable value at higher wpg rsquo

s. Thus, it is apparent that if the volume increase observed in the wood samples has contributions both from the reagent volume (V_i), and a void volume (V_f) (Nakano 1988), then the void volume is larger at lower levels of substitutions. This was commented previously (Hill and Jones 1999).

Fig. 1 Variation in molar volume with weight percent gain (wpg) for wood modified with acetic (circles) and hexanoic anhydride (squares)

Abb. 1 Änderung der Molvolumina mit dem prozentualen Massenzuwachs nach Modifikation mit Acetanhydrid (Kreise) und Capronanhydrid (Quadrate)

The determination of the apparent void volume created by the acyl groups within the wood cell wall polymeric network requires the use of theoretically calculated volumes for the attached acyl groups. These values are reported in the literature (Nakano 1988) and have been also used by Hill and Jones (1999). Those values not reported were obtained by extrapolation.

Void volumes (V_f) have been calculated for anhydride modified Corsican pine wood samples, by substracting the theoretical volumes (V_i) obtained from the literature (for each of the acyl groups) from the stable values of molar volumes (V) obtained from Fig. 1. The results are given in Table 2. From this it can be seen that void volumes created within the cell wall are higher in the hexanoic modified wood.

Table 2 A comparison of measured molar volume (V), calculated molar volume (V_i) and void volume (V_f) for Corsican pine modified with acetic and hexanoic anhydride. (Units are cm³ per mole)

Tabelle 2. Ein Vergleich der gemessenen (V) und experimentell bestimmten molaren Volumina (V_i) und LeerrÃ¤ume (V_f) fÃ&frac;r korsisches Kiefernholz modifiziert mit Acet- und Capronanhydrid (Einheiten in cm³ pro Mol)

R	V	V_i	V_f
CH₃	45	27.2	17.8
C₅H₁₁	115	68.4	46.6

3.2 Relationship between EMC and volumetric swelling

The relationship between EMC and volumetric swelling is illustrated in Figs. 2 and 3. From this, it can be seen that as the weight gain increases, the samples swell less at each EMC attained. At very low wpg rsquo

s (ca. 5–6%) a non-linear relationship is obtained; however at higher wpg rsquo

s the relationship is linear. A linear relationship was also reported by Risi and Arseneau (1957) for acetic and phthalic anhydrides, though no explanation appears to have been offered. This is further discussed later.

Fig. 2 Relationship between EMC and percent volumetric swelling of Corsican pine sapwood modified with acetic anhydride: Control (squares): 5.2 wpg (circles); 15.8 wpg (up triangles); 22.5 wpg (down triangles)

Abb. 2 Beziehung zwischen EMC und prozentualem Quellen von korsischem Kiefernholz nach Modifikation mit Acetanhydrid: Kontrollproben (Quadrate); 5,2 wpg (Kreise); 15,8 wpg (Dreiecke nach oben); 22,5 wpg (Dreiecke nach unten)

Fig. 3 Relationship between EMC and percent volumetric swelling of Corsican pine sapwood modified with hexanoic anhydride: Control (squares): 5.7 wpg (circles); 15.7 wpg (up triangles); 25.4 wpg (down triangles)

Abb. 3 Beziehung zwischen EMC und prozentualem Quellen von korsischem Kiefernholz nach Modifikation mit Capronhydrid: Kontrollproben (Quadrate); 5,7 wpg (Kreise); 15,7 wpg (Dreiecke nach oben); 25,4 wpg (Dreiecke nach unten)

The non-linear relationship between EMC and volumetric swelling obtained at low wpg rsquo s (ca. 5–6%) probably indicates a non-homogenous distribution of adduct at low levels of modification. It maybe also due to the larger void volumes at low wpg rsquo s, which would depress volumetric swelling at low emc rsquo s. As a result, when these voids were filled with water, the expected swelling was observed.

3.3 Analysis of the swelling behaviour of modified softwood

Figure 4 illustrates the relationship between weight of water in the cell wall due to exposure to relative humidity at saturation and bulking. Weight of water and bulking were calculated as follows:

Fig. 4 Relationship between bulking and weight of water in the cell wall due to exposure to relative humidity at saturation. Acetic anhydride (squares); Hexanoic anhydride (circles)

Abb. 4 Beziehung zwischen Volumenzunahme pro Gramm und der Wassermasse in der Zellwand nach Konditionieren über gesättigten Salzlösungen. Acetanhydrid (Quadrate); Capronanhydrid (Kreise)

${\text{Weight}}\;{\text{of}}\;{\text{water}} = \frac{{{\text{Weight}}\;{\text{of}}\;{\text{wet}}\;{\text{modified}}\;{\text{wood}} - {\text{Weight}}\;{\text{of}}\;{\text{dry}}\;{\text{modified}}\;{\text{wood}}}} {{{\text{Weight}}\;{\text{of}}\;{\text{dry}}\;{\text{unmodified}}\;{\text{wood}}}}$

${\text{Bulking}} = \frac{{{\text{Volume}}\;{\text{of}}\;{\text{dry}}\;{\text{modified}}\;{\text{wood}} - {\text{Volume}}\;{\text{of}}\;{\text{dry}}\;{\text{unmodified}}\;{\text{wood}}}} {{{\text{Weight}}\;{\text{of}}\;{\text{dry}}\;{\text{unmodified}}\;{\text{wood}}}}$

From this, it can be seen that for both chemicals a non-linear relationship is obtained. If it is assumed that the volume available in the cell wall for water decreases as a result of volume occupied by adduct, then a linear relationship between the two values would be expected. Thus, as more volume in the cell wall is occupied by adduct, there is less volume available for water.

On closer inspection, the results depicted in Fig. 4 indicate that at low wpg rsquo s (ca. 5–6%) more water is accommodated in the cell wall in acetic anhydride modified wood than in hexanoic anhydride modified wood. This indicates a shielding or masking effect with hexanoic anhydride modified wood (Papadopoulos and Hill). It appears therefore, that the adduct of the larger molecular size hexanoic anhydride, reacts with one hydroxyl group but covers some of the adjacent ones, providing a physical barrier. This barrier prevents the water vapour molecules from reaching some of the unreacted sites, that is, it shields sites. Hence, the effect of the adduct covering reacted and unreacted sites may overshadow the effect of direct chemical bonding to the sites. At intermediate wpg rsquo s (ca. 16%) both acetic and hexanoic anhydride modified wood accommodated the same amount of water, as it can be seen from Fig. 4. At higher wpg rsquo s (ca. 22.5%, although there is no data available for hexanoic anhydride, and therefore an approximate value has been obtained as indicated from the dotted line in Fig. 4) less water was accommodated by acetic anhydride modified wood, the opposite to what was observed at low levels of modification. This may be due to damage of the cell wall in wood modified with hexanoic anhydride. As a consequence, new sorption sites become available and extra space is created in the wood cell wall.

It was shown earlier, that volumetric changes due to modification with acetic and hexanoic anhydrides were to be due to the volume occupied by the reagent and an associated void volume created by the acyl groups within the cell wall. If it is assumed that this void volume can accommodate water without concomitant swelling of the cell wall, then clearly account must be taken of this. This is shown schematically in Fig. 5, where void volume was deducted from the value of bulking (see y-axis in Fig. 4) to give a figure for the apparent volume occupied by acyl groups in the modified cell wall. This apparent volume per gm of unmodified wood, was then plotted against weight of water per gm of unmodified wood in the same way that These recalculated data are given in Fig. 6. These recalculated data points, for acetic anhydride modified wood, were found to lie reasonably close to a linear line (R²=0.99), this was not the case for hexanoic anhydride modified wood. It has to be stressed at this point, that the line for the acetic anhydride data points, goes through the origin, however this is not the case for hexanoic anhydride. An explanation for this deviation is again the masking effect at low levels of modification and the damage of the cell wall at the higher ones. At the intermediate levels, the two lines cross over.

Fig. 5 Schematic representation of the swelling behaviour of unmodified and chemically modified wood, using the concept of void volume. Void volume is filled with water without causing volumetric swell

Abb. 5 Schematische Darstellung des Quellverhaltens von nicht modifiziertem und chemisch modifiziertem Holz unter Berücksichtigung der Leerräume. Letztere werden mit Wasser gefüllt, ohne eine Volumenquellung zu verursachen

Fig. 6 Relationship between weight of water in the cell wall due to exposure to relative humidity at saturation and volume occupied by acyl groups. Acetic anhydride (squares); Hexanoic anhydride (circles)

Abb. 6 Beziehung zwischen der Wassermasse in der Zellwand nach Konditionieren über gesättigten Salzlösungen und dem Volumen, das die Acylgruppen einnehmen: Acetanhydrid (Quadrate); Capronanhydrid (Kreise)

It seems therefore that the free volume model, offers a reasonable explanation of the differences in swelling recorded for Corsican pine sapwood modifying to varying wpg rsquo s with acetic anhydride. A wider range of weight gains, a larger number of replicates, and even swelling measurement during desorption would be required in a comprehensive study of this phenomenon.

ORIGINALARBEITEN · ORIGINALS

References