3.7.1.3.3 Comments on the results 

On the basis of the realized measurements, it’s possible to state that the wood samples, impregnated with starch and differently treated and dried, remains also in any case porous. The medium porosity, taken as ratio between the samples (2 – 5 – 6 – and 7) that show better the dimensional stabilization, is about 44%. We have eliminate the sample that had marked cracking. This medium value corresponds practically to the one obtained for the sample (7) dried by DSS (43,76% for a dry mass equal to 2,3301 g). If we assume that the water quantity of the waterlogged samples is equal to their maximal porosity from one side, and knowing that the mean water content of this wood is about 88,21% (humid base) from the other side, it’s possible to remark that the treatment allowed to reduce to an half the initial wood porosity. 

Moreover the wood samples dried by hot air (2), by continuous vacuum (5) and by DDS (7) have been better preserved. The deformation for the sample (2) was reduced at least of one half and it has a very slight cracking compared to the sample (1). So, finally, the positive effect due to the thermal treatment on the properties of the samples (1) and (2) is demonstrated. It’s the same result for the sample (5) and (3) dried by continuous vacuum; the case of sample (4) must be examined apart because it has not been impregnated. Moreover it’s possible to notice the difference between (4) and (5) related to starch impregnation.

Figure 1: Obtained results for the porosity measurements compared to the samples of wood called ECBL impregnated with starch, after thermal treatment by warm air (SAC), under continuous vacuum (SVC) and DDS, without thermal treatment in the case of the freeze dried sample (LYOPH). 

Finally we notice from these results that the freeze dried samples have been less deformed than the ones dried by DDS. The difference is not very considerable: 49,29 % against 43,76% that means a difference of 5,53 % between the two procedures (relative gap of 11,21%). These results confirm the ones obtained by measuring the ASE in which, the freeze drying is still the best from the dimensional point of view. So, as shown by the results of mechanical analysis, the wood freeze dried specimen are more fragile than those dried by DDS. 

Finally, the porosity results showed that the there is a coherence among the different measurements: mechanical, structural (RMN in particular) and retraction (ASE). These results have confirmed the best dimensional stabilization both for the freeze drying (at the beginning of the classification) and by DDS (in second position) as concerns the drying process. The results of porosity measurements confirmed our model for which the shape maintenance, through the dimensional stability is obtained with detriment of the mechanical resistance. 

3.7.1.4 Analysis of the colour of the treated wood 

3.7.1.4.1 Evaluation of the adopted procedure 

On the contrary of A.S.E defined to characterise the dimensional stability, any indicator has been used to standardize the colour of the waterlogged wood objects. So it’s convenient to know which were the real colours that the wood had before its immersion (probably unknown) and which the one eventually acquired after a long immersion time and after the loss in coloration substances. It’s more intelligent, for us, that the reference colour should be considered the one obtained by natural drying of the wood object coming from its site. 

The equipment set up at laboratory level to control the colour uses the system CIELab (Commission Internationale de l’Eclairage), internationally normalized CIE -1931. We remind that the system (L, a, b) - often report simply the Lab write- is a system coordinated by the norm is un DIN 6174. 

The equipment is constituted by: a camera of type Fujinon CV-M 90, fixed to a vertical support and linked to a micro-computer that give the visualization and the acquisition of the data, a graduated ruler on which it’s put the support of the camera and that is the guide; the ruler is at the same time support of the camera and allows to maintain an identical position to this last one for all the measurements; it is fixed vertically on a base on which is connected the measuring cell, where is put the sample and where specific lamps give the necessary light to expose the sample.

The acquisition of the data in the device last about 5 – 10 seconds but the parameters of measured colours are many. The software for colour analysis MFC (Main Frame C++) exploits the system (L a b) on the sample colours and give us some parameters as: 

✔    The mean of each colour and their nearer radiation: medium (R), medium (V), medium (B);
✔    The gap – types on each colour: gap-type (R), gap-type (V), gap-type (B); 
✔    The entropies on their combination: entropy (R, V), entropy (V, B), entropy (R,B);
✔    The colour - H, the clarity – L, the saturation – C;
✔    The medium values on L, a and b; 
✔    The gap-type on L, a and b; 
✔    The entropies on La, ab and Lb; 
✔    The saturation variant; 
✔    The medium value on A, c1 and c2; 
✔    The gap-type on A, c1 and c2; 
✔    The entropy on c1 and c2;
✔    And finally the variation of saturation on these last ones. 

The parameter considered as most representative and adapt because considers all the other parameters measured in this colour analysis is the «variation of saturation of the measured colour» on the wood samples. 

If the colour criteria and the absence of cracking are important concerning the quality of the treated waterlogged wood samples, the basic criteria remains the dimensional stabilization one. Moreover, a treated object maintaining its original colour and shape, shall also exhibit a certain mechanical resistance after consolidation. For this reason we tried to evaluate the resistance of the treated wood samples by mechanical analysis. 

3.7.1.4.2 Obtained results 

To compare the loss or the stability of colour on samples of waterlogged wood treated with the conservation method developed, nothing is better than consider the initial colour of the wood (i.e. coming out from the water). But the wood colour when waterlogged doesn’t reflect its natural colour. It happens an obscuration of the colour of the samples when they come out from the water. 

The equipment that analyse this colour has been already presented . This obscuration disappears to leave the natural colour to the sample when the wood starts to dry. Moreover, the darkening occurring under waterlogged condition doesn’t’ allow to appreciate the aesthetic quality of the wood samples. The sample analysis when the wood is wet, with the above mentioned equipment, show a colour that looks like the one of wood impregnated by PEG 400 and freeze dried. This is the reason why the reference colours adopted are the ones of the raw wood simply dried at air. 

The data found during these colour analysis especially the parameter «Variance of colour Saturation», have been statistically compared with the program STATGRAPHICS Plus, adapted to this kind of analysis. The analysis of the variance (ANOVA) for the «Variance of colour Saturation», the test of Krustkall - Wallis and of Kolmorghorv – Smirov, have been developed to compare the colour of the treated samples with some reference samples. The results of these analysis to find the VARIANCE of SATURATION of the colours, that has the 95% of reliability of results, showed that: 

✔    Our defined process allow a very good preservation of the colour for the treated samples; any significant changes haven’t been noted as shown in Figure 1, Figure 2 and Figure 3. Starch impregnated, submitted to thermal treatment and dried by DDS, the sample maintains their colour. These measured objectives allowed to confirm the sensorial observation noted during the different treatments. 

✔    The freeze drying induces a significant change of colour on the samples. The figures below show that the wood samples, often treated with starch impregnation or with PEG 400 and 4000, exhibits a great colour variation between the freeze dried samples and the one treated with our procedure The colour variation in the impregnated wood with PEG 400 and PEG 4000 don’t seems too important when the wood has been degraded (in the case of wood GJF-G) (Figure 2 and Figure 3). 

✔    As concern the different drying procedures tested, the DDS doesn’t cause any colour change compared to the hot air drying, the continuous vacuum and the freeze drying systems. So, we could classify the drying processes by decreasing order in accordance to their effects on the colour of the waterlogged wood. The freeze drying is the highest with its whitening effects on all the varieties of treated wood. The hot air procedure induce a change that remind the oxidation reactions, for any adopted temperature level (generally lower than 40°C, excluding some preliminary tests in which the temperature reached 60 to 105°C). The drying system by continuous vacuum gives a quality of the wood colour similar to the one found for DDS treatments. The drying tests by continuous vacuum, realized at temperatures of 20°C, 25, 30, 40, 50, and 60°C, did not reveal any change in the colour of the wood samples. 

Finally, the new drying method, the DDS, offer an exceptional quality of colours, maintaining them in the wood samples. 

These results corroborate the laboratory ones on the alimentary food as the fish (Juhel, 2000) by a continuous vacuum process, slightly comparable to DDS that allowed to dry the wood of green oak between 60 and 70°C without any change in the colour. 

From this point of view we can conclude that the DDS has an absolute superiority to freeze drying. 

Figure 1: Comparison of the colours for the samples of wood ECBL dried with different methods.

Figure 2: Results of analysis of colours for the sample bois 15 E:

- BRUT: waterlogged sample without any drying treatment. 

- DDS: dry sample by DDS (I = impregnated; NI: not impregnated; NT: without thermal treatment) 

- SAL: sample simply dried at open air 

- PEG400 (et PEG400L): impregnated samples with PEG 400 not dried (and dried with freeze drying). 

- LYOPH: starch impregnated samples, without thermal treatment and freeze dried. 

- IANTNS: samples impregnated with starch, without thermal treatment and not dried.

Figure 3: Results of the colour analysis for the samples of the boat from Grado, Julia Felix Grande 

(GJF-G) treated in different ways as shown below:

IANSEC: starch impregnated samples, not submitted to thermal treatment and not dried. 
IATDDS: starch impregnated samples, soumis au traitement thermique, séchés par DDS. 
NIDDS: not impregnated samples, non soumis au traitement thermique, séchés par DDS. 
PEG4000L: samples impegnated with PEG 4000, freeze dried. 
PEG400L: samples impegnated with PEG 400, freeze dried. 
NISVC: not impregnated samples, without thermal treatment, dried by continuous vacuum. 
SAL: not impregnated samples, without thermal treatment, dried at open air 10 – 20°C. 

In the two last figures we have observed the same variety of impregnated wood with PEG 400 and PEG 4000. We can noted that the impregnation with PEG (400 et 4000) cause a colour change quite important to the wood sample deviating them from their original colour. The obscuration of the samples induced by the PEG remain both in the dry and in the wet samples 

3.7.1.4.3 Evaluation of the obtained results 

In conclusion, the results of the colour analysis we can state that the colour criteria , in the researched quality for the conservation procedures of the waterlogged wood, gave a great satisfaction for our developed method. The objectives measured are in perfect accordance with the sensorial observation, especially regarding the quality for exposition of the treated objects in Museums. 

The following images allow the aesthetic comparison between the two treatments (waterlogged wood treated with Arkè method and with the PEG 4000 + freeze drying method). 

Note that the photos of the samples have been executed with the same camera and with the same exposition level.

     

Wood sample GJF – Treated with starch + DDS

Wood sample 15E – Treated with starch + DDS

Figure 4: Comparison of wood samples treaded with DDS and other drying methods

Wood sample 15E –  Treated with PEG 4000 + freeze drying	Wood sample GJF –  Treated with PEG 4000 + freeze drying

 

Figure 5: Comparison of wood samples treaded with DDS and other drying methods

It’s clear that the chromatic alteration of the freeze drying treated samples penalizes the good technical results of the consolidation, while with the Arkè method the good consolidation results are coupled with the very satisfactory aesthetic ones. 

Moreover the samples treated with Arkè method have a density, texture and consistence very similar to the one of the natural wood while the products treated with PEG are heavier (higher density) and have a waxy aspect completely different from the natural wood. 

At this moment we can declare that the results obtained in laboratory are surely satisfactory and meet the project requirements.

3.8 DEFINITION OF THE OPTIMISED ARKÈ PROCESS

The preliminary phase of the new consolidation process for the restoration of archaeological wood finds developed in this project is wood characterization by samples analysis;  this phase has a strong influence on the other operative phases of the process that are : the impregnation with starch based products, the intermediate thermal treatment and the final drying. 

This chapter is devoted to the presentation of the procedure developed for the selection of the optimised consolidation process for a standard waterlogged wood find. In the previous chapters we have defined several possible valid treatment conditions that can be choiced according to the intrinsic properties of the find concerned. 

These tretament conditions can be combined in several manner in order to guarantee the best performances for the final product and the criteria for this combination will constitute the most important part of the procedure developed.

3.8.1 Treatment process proposed for waterlogged wood 

As previously reported, the main phases of the consolidation process are three: the impregnation with starch based products, the intermediate thermal treatment and the final drying. But in order to obtain optimised results some preparatory task has to be carried out. 

3.8.1.1 Cleaning of the wood finding 

This operation is necessary to obtain a good level of impregnation by diffusion of the starch based impregnating agent in the wood structure. A simple external washing with water will be enough to remove mud, sludge and sands from the sample. The use of stronger washing treatments will be possible only for less degraded samples: the choice of the washing system should be taken case by case. 

3.8.1.2 Determination of the wood degradation level 

In this preparatory activity, one of the main problem for the consolidation of waterlogged wood is concerned: the characterization of the find to be treated. It’s necessary to classify the wood element as concerns mainly level of degradation and structural homogeneity. The sampling activity plays a very important role in this phase but the problem is strictly linked to the destructive character of this operation: The sample should be more representative as possible but the excessive damage of the object to be treated should be avoided carefully. To solve this problem we suggest to: 

✔  use the characterization by «pilodyn», instrument proposed and developed by Clarke § Squirell (1985) to evaluate the degradation level of waterlogged wood without any supplementary deterioration; 

✔  use a method based on micro penetration, applying, with constant speed, trough a nail a load on waterlogged wood; the characterization will be possible by comparison with the results obtained testing some fragile sample previously identified (Kazanskaya § Nikitina, 1989);

✔  the use of sensorial analysis to evaluate the degradation.

The main objective of the characterization phase, for any kind of analytical solution adopted, is to orientate the consolidation treatment and to allow the selection of the optimal treatment conditions.

3.8.1.3 Starch impregnation 

3.8.1.3.1 Selection of starch based products mixtures 

The impregnation process can be carried out under atmospheric pressure or under a pressure slightly higher: For both cases it’s necessary that the objects to be treated are completely submerged in the impregnating solution. The project work allowed us to individuate a non exhaustive list of starch based products useful for the Arkè process. 

3.8.1.3.2 Duration of the impregnation process 

We have demonstrated that the optimal duration of the impregnation process is depending from the kind and from the structure of the waterlogged wood concerned but also from the size of the object to be treated. The phenomenon is mainly based on diffusion, with a duration increasing whit the square of the thickness for a defined level of impregnation. But on practical level, also if seems to be difficult to evaluate in advance the evolution of the degradation level of the waterlogged wood according to its thickness, (this is one of the main factor in the wood characterization phase) we can normally observe that, in the internal zones of the samples (heart), the cellulose content is still high, so it’s not needed to reach an high level of impregnation in these areas. 

The effective duration of the impregnation process can usually be lower than the duration determinable on the basis of the theoretical diffusion. We have to mention that, for a complex object, constituted by several wood species and sizes, it will be necessary to select the longer impregnation time in order to assure homogeneous and optimal results. The objects to be treated can have radial and tangential dimensions very variable and various geometrical shapes. In these cases, a characteristic thickness (e) could be calculated on the basis of the estimation of the volume (V) and of the exchange surface (S) as e= V/S in any waterlogged wood sample. 

From the practical point of view and for a defined impregnation level, the duration of the impregnation process could be correlated to the thickness of the wood object (real or characteristic) by a function as (Y = a x2 + b. x + c). In both the cases, we have obtained with quite all the testes starch mixtures: (a = 0,001; b = - 0,0861 and c = 7,5987), with the square of the correlation coefficient equal to R2 = 1 (where x represents the characteristic thickness x=e=(V/S)2]. These results allowed us to estimate that, for a specific kind of samples, when the thickness increases from 12mm to 20mm the duration of the impregnation process increases from 6 days to 17 days. Finally we underline that is possible also to carry out a consolidation process without the impregnation step: to this purpose, the level of degradation of the wood sample should be exceptionally low. 

3.8.1.4 Intermediate thermal treatment 

On the basis of a similar phenomenon based on diffusion, the heat is transferred, with a sped considerably higher than matter (i.e. starch), inside to the treated element, because heat is characterized by a higher diffusion coefficient. Thus, also this treatment will have globally a shorter duration if the thickness (real or characteristic) of the object to be treated is lower. It’s important to underline that only the duration of the phase of temperature increasing has to be decided according to the thickness of the object to be treated, while it’s important that the phase of temperature/pressure reduction remains very slow, in order to avoid structural tensions inside the wood. The main item to be enhanced as concerns this treatment is the great difference between the heat diffusion speed in waterlogged wood and in dry wood. In the first case, the heat diffusion speed is higher, probably because of some convection phenomena that facilitate the kinetic of the operation. The fact is that the reaching of a defined uniform temperature of treatment within a waterlogged wood object needs considerably less time than within a dry wood object. 

3.8.1.5. Final drying of the wood 

Like the wood impregnation phase, also the final step of the treatment represents a matter transfer, but in this case the matter (water) that is contained in the object has to be extracted from the wood structure and not introduced as done with starch in the in initial step of the process. If in the impregnation phase the target conditions are short duration with highest level of starch introduction, in the drying phase we are aiming to combine a short duration with cracks elimination and perfect control of shrinkage, in order to be able to give back to the wood elements is original dimension and aesthetical aspect. The trials carried out in this project has demonstrated the advantages of DDS applied under light conditions (Ph £ 0,80 bar; Tc £ 40°C); This treatment allow us to couple a good drying speed with high quality final products both as concerns restoration and conservation purposes of waterlogged wood. Also under operative conditions that seems not be adapt to its kinetics, DDS results the most rapid drying technique among the process at present on the market including, hot air, vacuum and freeze drying. It’s important to underline that, operating on some (rare) samples strongly damaged and reduced to a waterlogged spongy mass, could be necessary to use freeze drying instead of DDS. But in this case will probably be preferable to use the same DDS treatment under lower pressure conditions (1mbar of final pressure) to obtain the best results. Unfortunately was not possible to carry out the optimisation of this new DDS treatment within this project. 

3.8.1.6 Possible variations of the proposed process 

The following flow chart (Figure 1) shows the possible variation of the whole consolidation process, expressed here to simplify the exposition, as function of the Maximum Water Content (MWC) of the waterlogged wood element to be treated. We have already underlined that the MWC parameter, even considering its limits, remains the simplest and probably the cheapest to use. The proposed process is characterized by its adapting capacity to the several categories of waterlogged wood and thus by the presence of several treatment combination. 

Some problems can appear when the object to be treated is constituted by various wood elements, that differ for kind of wood and level of degradation. In this case it’s necessary to identify the most deteriorated element and choice the optimal treatment conditions for all the object according to the characteristic of this element. Effectively, one of the most important advantages of the starch impregnation, already underlined in chapter 5.3, is represented by the capacity of each kind of wood to absorb starch till an optimal degree of impregnation, as required by the level of degradation of its structure; this will avoid saturation and will strongly increase the uniformity of objects constituted by several kinds of wood. On this basis, whatever is the level of degradation of the wood to be treated, a good level of impregnation will be reached, even taking more time than required by the less deteriorated elements. 

Another variation can be defined for the objects constituted by relatively homogeneous wood and for the objects that can be easily disassembled. The process can be better adapted for this specific kind of application but these are not so common. The flow diagram in Figure 1 gives a good summary of the main possibilities and allows the choice of the optimal treatment. 

As conclusion, we have to underline that, taking into account the flow diagram indications, the process should be finally optimized according to the specific characteristics of the wood object to be treated. The diagram takes into account the initial waterlogged wood classification (i.e. on the basis of MWC parameter) and the level of impregnation reached with starch solutions. We have enhanced the fact that the more deteriorated samples show a starch absorption higher than the less deteriorated. That means that the more deteriorated wood exposes an higher water elimination in the drying phase also if the final water content in the treated wood is similar. Moreover the gelatinization of starch granules introduced in the object requires a certain amount of water. In the chapter related to the intermediate thermal treatment we stressed the importance of the moisture content during the operation. Without the introduction of humidity during the thermal treatment the water contained in the wood will evaporate and will increase the shrinkage of the treated sample. 

Figure 1: Flow diagram of the proposed consolidation/restoration process for waterlogged wood objects, based on starch impregnation coupled with thermal treatments. The variations are expressed as function of the degree of degradation of the wood. 

According to this, higher is the water content in the wood after the impregnation phase, more rapidly will start the evaporation during the thermal treatment, if the humidity content on the treatment chamber will be below the saturation point. This explain why, in the diagram, the steam saturation is preferred for the wood sample showing the higher MWC value. The same treatment is non dangerous for the less deteriorated wood samples. But the water elimination from this kind of samples is strongly influenced by the density of the structure, that will slow this process more than if applied on more porous and deteriorated samples. So it’s possible to reduce the energy consumption of the treatment, introducing directly liquid water in the treatment chamber before the heating start. 

The flow diagram identifies two wood categories as concerns the final drying process. We suggest that, until the conclusion of a research on low pressure DDS, freeze drying should be used for the ultra deteriorated wood, while DDS should be preferred for the other wood categories. The phases of impregnation and intermediate thermal treatment will remain the same in the two cases. 

The choice of the separation level between the two drying system depends from a compromise. We have to decide if is preferable to allow a shrinkage slightly higher for ultra-deteriorated wood samples to obtain a treated sample with higher mechanical stability or to maintain the dimensions of the sample obtaining a lower mechanical stability (and a little degradation of the color). 

Among this kind of samples that we have studied, (i.e. sample AF9, ECBL, DN5-BC, CpIII-CpVI), the DDS drying generated a shrinkage level reduced but systematically slightly higher than freeze drying. 

The Arkè process developed in the project shows a lot of advantages if compared with the state of art but some limitations are still present. For ultra deteriorated waterlogged wood elements, the combination of a very satisfying sensorial aspect, in terms of color, density, texture and presence of fissures with good mechanical properties and with a shrinkage level relatively high (can reach 12-20% for some kind of wood) is really at present the best available solution? In our opinion, this can be surely a good point of compromise. 


↳  TREATMENT PLANTS