Phenomenological modeling and intensification of texturing/grinding-assisted solvent oil extraction: case of date seeds (Phoenix dactylifera L.)

1.1 State of art

Date palm, with a botanical name that probably derived from Phoenician as Phoenix dactylifera L., belongs to the family of Palmaceae with about 235 genres and 400 species (Munier, 1973). Date palm is a tree of arid and semi-arid native hot and humid countries but it has a wide capacity of adaptation. It is mainly located in northern Africa and the Middle East as well as west Asia and United States. According to the Food and Agriculture Organization of the United Nations (FAO, 2010), the annual world production of date is estimated at 7 million tons (MT), from which seeds represent about 1 million tons.

Tunisia is one of the main date producers with about 125 000 tons/years with 60% of “Deglet Nour” variety (Rhouma, 1993). Hence around 17 5000 tons of date seeds could be recovered and used as by-products thanks to their content of fatty acid, protein and high dietary fiber (Besbes et al., 2004a).

Nowadays date seeds are barely used for animal feeding and generally thrown away. The losses usually are higher than 30% of the whole production, which represents a huge tonnage near to 8700 tons/year of date seeds in Tunisia (Borchani et al., 2010). It is important to note that date seed plays an important part in date advantages. It contains more oil than date flesh, with some nutrient compounds (Al-Qarawi et al., 2003; Briones et al., 2011; Aris et al., 2013). Indeed, date seeds contain between 5 and 12 g of oil/100 g db (dry basis) whereas date flesh does not exceed 0.5% db of oil depending on the raw material composition and processing conditions (Al-Hooti et al., 1997; Hamada et al., 2002; Besbes et al., 2005; Habib and Ibrahim, 2008).

It has been reported that date seed oil is composed of about 44% of saturated fatty acids, 41% of monounsaturated fatty acids and 14% of polyunsaturated fatty acid (Besbes et al., 2004a). The common composition of fatty acids is characterized by the presence of four main compounds including oleic acid (41–50%), linoleic acid (12–19%), lauric acid (10–15%), and palmitic acid (10–11%) (Besbes et al., 2004b; Rahman et al., 2007; Nehdi et al., 2010; Aris et al., 2013); Oleic acid is reported to be used as a good anti-inflammatory drug and has fundamental role in cardiovascular disease prevention (Larrucea et al., 2001), while lauric acid has various medicinal, therapeutic (antibiotic and antiviral effects), and nutritional virtues (Desbois, 2012). Furthermore, Devshony et al., (1992) have proved that linoleic acid of date seed oil had good potential as cosmetic products. Since date seeds are rich in extractible components, they are regarded as real potentially economic source once they are valorized.

The main issue in developing date seed valuing is strictly correlated to oil extraction limits, in the main part from the hardness and compactness of its matrix. Thus press process for oil extraction is difficult to apply. Moreover, since the oil content is quite low, solvent extraction remains the only appropriate tool for this type of material. However, the most conventional solvent extraction technologies such as maceration require high solvent consumption and long extraction time (Besbes et al., 2005). The main issue in this case is the slow diffusion of solvent and solute through the solid and the core successively (Allaf et al., 2011). One of the most famous solutions was the use of super or sub-critical fluids as non-conventional solvents (Aris et al., 2013). Other very effective solutions have concerned some pretreatments such as grinding or texturing. Besides, the efficiency of solvent processes can be enhanced with ultrasound (Pan et al., 2011), microwaves (Chemat and Cravotto, 2013), accelerated solvent extraction, etc. However these treatments trigger technical difficulties at industrial level and/or economic constraints. Thus, these operations were only successfully conducted at laboratory or pilot scales. Thus, scientists are attempting to adapt extraction processes to meet these strategic requirements.

Instant Controlled drop pressure (DIC) treatment is used for texturing. Since DIC treatment involves a high temperature short-time treatment of the product followed by an ‘instant’ decompression toward a vacuum (Ben Amor and Allaf, 2009; Allaf and Allaf, 2014), it frequently results in controlled expansion, and allows the modification of the matrix microstructure. Thus technological abilities of DIC-treated plants are usually improved vis-à-vis the solid–liquid interaction. It can also preserve the molecular profile, even for heat-sensitive compounds.

1.2 Fundamental (kinetic modeling)

The fundamental phenomenological study of solvent extraction process was achieved through the two-stage kinetic model, which implies both “washing stage” (solvent/exchange surface interaction) and “effective diffusion” within the matrix. This allows identifying the limiting phenomenon and advising adequate technological solutions. Enhancements can then be divulged through the starting accessibility and effective diffusivity (Allaf et al., 2014).

The intensification of solvent extraction is usually firstly performed by increasing the external interaction between solvent and solid surface (washing) triggering the solute on the surface removal. The amount of solute able to be extracted m_s depends on the nature of solvent and desired solute, the exchange surface, and the temperature. The extraction rate of solute extracted from the surface can be revealed by the following equation:

• m_s: is the mass of solute (kg of solute);

• K: is the solvent mass convection rate (kg of solvent m⁻² s⁻¹) depending on the stirring velocity and temperature.

• S: is the exchange surface area (m²) between the solvent and the solid, normally affected by grinding;

• w_sat: is the dissolving coefficient at equilibrium (kg of solute per kg of solvent);

• w_solv: is the solute dissolved in the bulk solvent (kg of solute per kg of solvent).

The density of solute from the surface per unit of dry matter is

This stage of extraction process can be enhanced by the following:

• Agitation/stirring implying more solvent convection, which increases the value of K.

• Grinding/crushing of material, to increase the interaction surface.

• Heating, because higher temperature normally increases the dissolution ability of solute in the solvent.

Once this external intensification is performed, internal diffusion becomes the limiting phenomenon. Most of solvent extraction operations are then regimented by the complex phenomena of solvent penetration in the solid, dissolution of solute in the solvent, and diffusion of solute in the internal solvent within the solid matrix. Thus, the driving force of these transfer phenomena can be assumed as a gradient of density ratio (Allaf et al., 2011).

• $({\overrightarrow{v}}_e-{\overrightarrow{v}}_{\mathit{matrix}})$ is the relative velocity of the solute (m s⁻¹) to solid dry material, (m s⁻¹).

• ρ_e is the apparent density of the solute within the porous solid, (kg m⁻³).

Because of the absence of expansion, shrinkage … , v_matrix = 0 and ρ_matrix = constant.

1. “washing”; disclosed by the starting accessibility δY_s (expressed in kg of extract per kg of dry material): it is the amount of extract removed in very short time (t near 0) from the interaction surface. The interaction between the solvent and the exchanging surface takes place over this short time frame. The external solvent dissolves the superficial solute. The intensification of this stage should imply a dynamic convection (Amor et al., 2008; Ben Amor and Allaf, 2009) instead of external diffusion.

2. and an internal diffusion. Since the “washing” stage removes some part of solute from the sample surface, a solute gradient takes place between the core and surface of the matrix. This acts as a driving force of the internal effective diffusion of solute in the solvent within the porous matrix solids (Allaf, 2009). D_eff (m² s⁻¹) acts as the effective diffusivity of a similar Fick’s law (Allaf et al., 2011).

The whole extraction process involved in the first stage (solvent interacting with the exchange surface) does not have to be involved in the diffusion model part.

By adding continuity to Fick’s-type law (Allaf et al., 2011) and by assuming the homogeneity of both temperature and structure within the matrix, (Mounir and Allaf, 2008) proposed effective diffusivity D_eff constant; Eq. (4) becomes

• Y: The quantity of extracted solute (oil) at time t (kg/kg ddb),

• Y_∞: The maximal quantity of extracted solute (oil) Y when t → ∞,

• Y_t=to: The quantity of solute (oil) at the beginning of the stage only concerned by the internal diffusion stage.

The correspondence between experimental data and model-diffusion is performed to calculate the effective diffusivity D_eff using Crank polynomial expression:

2.1 RM and chemicals

Date seeds were acquired from the National Institute of Arid Zone (Degache, Tunisia). Initial water content W_i was measured 8% ± 0.2 dry basis (db). Dried seeds were isolated from full ripeness date fruit “Tamr stage”.

n-Hexane was the only solvent we used for oil extraction. It was purchased from Carlo Erba (Val de Reuil, France) with 99.99% purity.

2.2 Treatment operations

The protocol of the treatment operations including assessments is illustrated in Fig. 1.

Close

Figure 1 Protocol for the extraction of oil from Tunisian date seeds.

Export to PPT

2.2.3

2.2.3 Instant controlled pressure drop

2.2.3.1

2.2.3.1 Equipment and laboratory scale unit

The experimental DIC setup is composed of three main elements: (1) A high-temperature; high-pressure treatment vessel, with a suitable heating gas; (2) A vacuum system with a vacuum tank with a volume 100 times greater than the processing vessel, and an adequate vacuum pump for maintaining an initial vacuum level of about 5 kPa in all the experiments; and (3) An instant opening pneumatic valve between the vacuum tank and the processing vessel; it can be opened in less than 0.2 s; this ensures the abrupt pressure drop within the treatment vessel.

A data acquisition and automation system takes into account pressure, temperature, and time parameters. It is connected to a personal computer, establishing combination with the manual control of the system.

2.2.3.2

2.2.3.2 Treatment operating parameters

Dried powder of date seeds (50 g) at 8 ± 0.2% db (dry basis), was firstly placed in the DIC treatment vessel (between 20 and 70 s). A first vacuum stage was established in order to reduce the resistance between the exchange surface and the saturated steam, which acts instantaneously (in less than 1 s) as a heating fluid by condensation. Heat transfer inside the raw material is performed by effective conduction, strictly correlated with condensed water diffusion.

An abrupt pressure drop toward a vacuum systematically follows the thermal treatment. It results in an instant autovaporization inducing an “instant” cooling of the solid material. After their DIC texturing, date seeds were recovered and ready for extraction (Fig. 2).

Close

Figure 2 Temperature and pressure history of a DIC processing cycle.

Export to PPT

2.2.3.3

2.2.3.3 Experimental design

We first carried out some preliminary experiments to identify the global impact of the texturing treatment, the most important operative parameters, and their ranges. The main operative parameters were the steam pressure P and the thermal treatment time t. P was ranged between 0.1 and 0.7 MPa and t from 20 to 70 s. This time should assure the homogeneity of both temperature and water content inside the particle. Optimization of DIC operating parameters was performed using Response Surface Methodology (RSM) with a 2-parameter, 5-level central composite experimental design. Thus we could reduce the experimental trials to 13 implying 2^k = 4 factorial points; 2k = 4 star points with five repetitions for the central point; k is the number of operating parameters used in the experimental design (Table 1).

Table 1 Independent variables and ranges used in Response Surface Methodology (RSM) and Design of Experiments (DoE).

Coded level	−α	−1	0	+1	+α
Saturated steam pressure (MPa)	0.2	0.29	0. 5	0.71	0.8
Processing time (s)	20	27	45	63	70

Since we used an orthogonal factorial design, α is the axial distance: α = 1.41421.

The experiments were run in random in order to minimize the effects of unexpected errors due to extraneous factors.

(11)

$$\alpha =\sqrt[4]{n}=\sqrt[4]{2^k}$$ In the present case, k = 2, the number of factorial trials is n = 4 and α = 1.414213562.

Various extraction process parameters such as yields and kinetic model parameters, were assessed and considered as the responses (dependent variables). Their values were introduced in the analysis design procedure of Statgraphics Software (MANUGISTICS Inc., Rockville, USA). It enables the interpretation of the results, optimization of the treatment by multidimensional ways through recognizing the measured response variables Y by fitted second order polynomial models versus operating parameters X_i as factors:

(12)

$$Y={\beta }_0+{\displaystyle \sum _{i=1}^k}{\beta }_i{\chi }_i+{\displaystyle \sum _{i=1}^k}{\beta }_{\mathit{ii}}{\chi }_i^2+{\displaystyle \sum _{i=1}^{k-1}}{\displaystyle \sum _{j=2}^k}{\beta }_{\mathit{ij}}{\chi }_i{\chi }_j+\mathcal{E}$$ where β_ii, β_ij, β_i: are regression coefficients, ɛ: is the random error, and i and j: are the indices of factors.

RSM empirical polynomial model is used to optimize the factors. Experimental results allow determining the ANalyses Of VAriance (ANOVA), which are performed to determine significant differences between independent variables. Thus, the adequacy and the significance of the model are analyzed by estimating the lack of correspondence, Fisher test value (F-value) and R² from the evaluation of ANOVA. The independent variables have a statistical significance here proved at 5% probability level (p < 0.05) and revealed through Pareto chart. Then to build response surface, the software uses the quadratic model equation.

2.3 Assessments and characterization

2.4 The main Responses for identification of extraction operation

The quantification of the intensification effects was performed through yields Y_∞ (g/g ddb) and kinetics extraction, which were the starting accessibility δY_s (g/g ddb), the effective diffusivity D_eff (m² s⁻¹), and the Relative Extraction Time (RET) based on untextured material (%). Yields were issued from ASE measurements, while kinetic parameters were gathered from the DM data through the phenomenological model of surface-interaction (washing)/internal diffusion.

Optimization of solvent extraction process aimed at getting the highest values of Y_∞, δY_s, and D_eff, and the lowest value of RET.

In our study of process modeling, we used at first ASE experimental data to determine the final oil yield Y_∞ in the relative diffusion extraction:

The difference between the value Y_o calculated by extrapolating the diffusion model until the initial time t = 0 and the experimental value Y_i = 0 corresponds to the starting accessibility δY_s on dry–dry basis (Fig. 3):

Close

Figure 3 Determination of the effective diffusivity D_eff and the starting accessibility δY_s from experimental data of DM extraction kinetics and diffusion model.

Export to PPT

3.1 Oil extraction from coarsely ground date seeds

3.1.1

3.1.1 Effect of DIC parameters on ASE oil extraction yields

Let us firstly note that DIC saturated steam pressure in the treatment vessel reveals the thermal level of treatment. The optimization of DIC treatment conditions was based on ASE yields (Y_ASE) as the main response parameter (Table 2). Pareto chart, General Trends, Response Surface and Iso-Response for Y_ASE oil yields from date seed powder were determined and are shown in Fig. 4(A, B, C, D). In Pareto chart histogram, the effect is considered as statistically significant if the corresponding factor crosses the vertical 5% significance line.

Close

Figure 4 Effect of DIC parameters on the total ASE oil extraction yields of date seed powder: (A) Standardized Pareto Chart; (B) Main Effects Plot; (C) Estimated Response Surface, and (D) Iso-Response.

Export to PPT

In these statistical limits, and the ranges of operating parameters, the linear steam pressure P value was the most significant effect followed by linear effect of the treatment time t. The favor effect of P and t on the operation is illustrated by their positive values. Thus, in the considered ranges, the higher the values of P and t, the higher the ASE extracted oil. The second-order empirical regression model of ASE yields versus DIC parameters had a regression coefficient R² = 0.88, which means a good fit with experiments.

(18)

$$Y_{\text{ASE}}=13.59+0.62P+0.012t+0.20P^2-0.009\mathit{Pt}-0.00006t^2$$ where Y_ASE is expressed in % ddb; P in MPa; and t in s.

From Eq. (18) in regression model, it was possible to optimize DIC operating conditions and the highest value of the oil yields of Y_ASE = 14.29% ddb was achieved at P = 0.8 MPa and t = 38 s.

3.1.2

3.1.2 Effect of DIC parameters on Dynamic Maceration (DM) extraction kinetics

General results

Experimental results obtained from Dynamic Maceration (DM) extraction of oil from date seeds are presented in Fig. 5. As expected, DIC texturing pretreatment implied increasing of oil quantity in comparison with untreated particles.

Close

Figure 5 Kinetics of oil extraction from various DIC textured date seeds powder.

Export to PPT

The yield obtained after 2 h extraction reaches a maximum value of 11.28–11.42 ± 0.08% ddb for DIC treatments against 7.5 ± 0.05% ddb for controlled raw material. Oil yield of raw material could increase versus time and the maximum value of 8% ddb were obtained after 8 h while a value of 11% was reached after 24 h of extraction, whereas, the oil yield obtained in 8 h of extraction of DIC textured seeds was 12% ddb. It was clear that whatever DIC treatments, we could obtain in 30 min the same oil yield as that occurred by raw material after 8 h of dynamic maceration DM. We can deduce the efficiency of DIC treatment in intensifying extraction kinetics from date seed powder.

Starting accessibility and effective diffusivity

Table 2 shows that the effective diffusivity D_eff of the treated DIC textured samples reached values of 8.9–10.1 · 10⁻¹² m² s⁻¹, while it was about 8.7 · 10⁻¹² m² s⁻¹ for untreated sample (RM), which means a great increase of kinetics.

DIC texturing treatment significantly improved the starting accessibility δY_s by more than 71% to be up to 10.13% instead of 6.14% ddb for untreated seeds.

Fig. 6 shows the significant effect of DIC saturated steam pressure and thermal treatment time. General trends and Response surface were also relevant to illustrate the significant effect of the operating parameters. The second-order empirical regression model of starting accessibility versus DIC parameters was established with a regression coefficient R² equal to 0.78%, which is an acceptable fit with experiments.

(19)

$$\delta Y_s=8.67+1.918P+0.0228t-1.35204P^2-0.00198\mathit{Pt}-0.000016t^2$$ The impact of DIC texturing as a specific pretreatment prior to solvent extraction was very important compared to untreated raw material. Thus, D_eff values were 10.12 ⋅ 10⁻¹² against 8.7 ⋅ 10⁻¹² m² s⁻¹ for DIC treated and untreated raw material, respectively. However, Pareto chart histograms for effective diffusivity D_eff show slowest effect of DIC parameters. This means that the range of operative parameters was too focused.

Close

Figure 6 Effect of DIC parameters on starting accessibility: (A) Standardized Pareto Chart; (B) Main Effects Plot; (C) Estimated Response Surface, and (D) Iso-Response.

Export to PPT

Oil extraction yields

RSM study was performed with various kinetic parameters versus DIC operating parameters P and t. Pareto Chart obtained from experimental data of DM oil extraction yields in Fig. 7, showed the significant effect of saturated steam pressure P and the treatment time t through its interaction with P, whereas, t did not have a major influence. General trends and Response surface were also relevant to illustrate these significant effects. Here too, the second-order empirical regression model of DM yields versus DIC parameters had a regression coefficient R² = 0.94%, revealing a proper fitting to experiments.

(20)

$$Y_{\text{DM;}\infty }=12.27+1.3P-0.0005t-1.17914P^2+0.01\mathit{Pt}-0.00004t^2$$ where Y_DM;∞ is expressed in % ddb; P in MPa; and t in s.

Close

Figure 7 Effect of DIC parameters on total DM oil extraction yield: (A) Standardized Pareto Chart; (B) Main Effects Plot; (C) Estimated Response Surface, and (D) Iso-Responses.

Export to PPT

The same optimized DIC parameters P = 0.8 MPa and t = 70 s were obtained with maximum oil yields of DM Y_DM;∞ = 13.01% ddb.

3.2 Impact of particle size on oil yields and kinetics

The second part of this study concerned the impact of grinding and particle size on the ability of the oil extraction. ASE and DM were used with different particle sizes (0.2–1.4 mm) of powder from raw material and DIC textured samples. Measurements of ASE oil yields Y_ASE and DM Oil yields versus time were carried out and compared.

3.2.2

3.2.2 Kinetics of total extraction

All experimental results obtained from extraction kinetics of DIC treated date seed powders (Fig. 8), show that decreasing sample size implied increasing of oil quantity, which systematically were higher than untreated considered powders.

Close

Figure 8 Extraction kinetics of date seed oil.

Export to PPT

Thus, 2 h DM extraction allowed obtaining oil yields of 3.47% ddb for 1.4 mm particle. This amount was increasing with decreasing particle size to become 9.54% ddb for the smallest untreated seeds (0.2 mm). The same behavior was observed with DIC treated seed powders. And when DM extraction was performed for the same time, this value increased from 4.67% to 11.62% ddb, respectively. The highest improvement issued from DIC treatment was revealed by RIE = 34.58 ± 0.05% for powders of 1.4 mm particle size.

Obviously, oil yield was remarkably as higher as the particle size was smaller and much more when powders were DIC textured. This can be attributed to the fact that the smaller the particle size, the larger the specific surface area and the shorter the mass transfer distance. The oil is being isolated much more efficiently by reducing the particle size of the powder. Furthermore, DIC texturing generates reduction of treatment time, increases yields and improves extraction kinetics. Higher extraction availability thus triggered should be due to its possible disruption of cell walls where oil is located in date seeds. So, larger contact area between solvent and material was created and more oil extraction was appeared at the surface as starting accessibility.

The kinetic parameters of starting accessibility δY_s (% ddb), effective diffusivity D_eff (m² s⁻¹), and DM yield Y_t=2h (% ddb) are summarized in Table 4 for different cases.

Table 4 Starting accessibility δY_s, effective diffusivity D_eff, and oil yield after 2 h DM extraction Y_t=2h for 1.4 mm and 0.2 mm particle sizes.

D (mm)	1.4	1	0.8	0.6	0.4	0.2
RM δYs (% ddb)	2.53%	4.06%	5.15%	7.21%	8.16%	8.77%
DIC δYs (% ddb)	3.33%	3.62%	6.01%	8.73%	9.01%	10.29%
RM Deff (10−12 m2 s−1)	47	11.1	7.9	3.6	1.5	0.4
DIC Deff (10−12 m2 s−1)	62.4	20.8	10.6	5.8	2.9	0.7

Experimental data should be utilized to identify the possible models revealing (1) the whole improvement resulted from DIC versus RM in terms of δY_s and D_eff; (2) the whole evolution of δY_s versus the granule size D. Indeed, while D_eff should depend on various parameters such as D, availability of oil inside the residual cells, and permeability and tortuosity of the grain; a simplified empirical model of the starting accessibility versus D would exist to make global link with both surface and diameter D. Indeed, since it is evident that exchange area strictly depends on D, such model can get more the following form:

(21)

$$\delta Y_s=f(D)$$ The relative expression of DIC versus RM powders at corresponding granulometry for both D_eff and δY_s offered about 15% improvement factor with R² of 0.99 and 0.95, respectively (Figs. 9 and 10) (Eqs. (22) and (23)).

(22)

$$D_{\text{eff}}(\text{DIC})=1.1584D_{\text{eff}}(\text{RM})+1.301\text{with}{R}^2=0.98904$$

(23)

$$\delta Y_s(\text{DIC})=1.1517\delta Y_s(\text{RM})-0.0017\text{with}{R}^2=0.9469$$ These values of R² translate the good relevance of both these empirical models. The global impact of DIC versus RM would be revealed by the fact that DM extraction time was 264 min and 85 min for 1.4 and 0.2 mm diameter particle, respectively, while for DIC samples extraction time was 122.5 for 1.4 mm and 55 min for 0.2 mm particle diameter. The evolution of starting accessibility δYs versus particle diameter was calculated as linear correlation models for both RM and DIC-textured powders, respectively.

(24)

$$\delta Y_s(\text{RM})=9.96(1-171\pi D)\%\text{ddb}\text{with}{R}^2=0.88274$$

(25)

$$\delta Y_s(\text{DIC})=8.86(1-171\pi D)\%\text{ddb}\text{with}{R}^2=0.9716$$ where D is expressed in m and δY_s in % ddb.

Close

Figure 9 Empirical models of relative evolution of extraction kinetics of DIC versus RM powders.

Export to PPT

Close

Figure 10 Empirical models of starting accessibility δY_s versus the particle diameter of DIC and RM powders.

Export to PPT

Both RM and DIC textured powders presented similar behaviors in terms of starting accessibility evolution versus diameter.

Well-optimized DIC treatment coupled with the most realistic grinding operation should always present a very effective and relevant intensification way for the solvent oil extraction.

3.3 Fatty acid composition

The fatty acid methyl esters (FAMEs) composition of the oils of the species is shown in Table 5. The most abundant fatty acids of date seed oil were oleic (C18:1), linoleic (C18:2), palmitic (C16:0), myristic (C14:0), and lauric (C12:0) which together composed about 90–95% of the total fatty acids.

The major fatty acids found in those cultivars for DIC treated and untreated samples were similar. They were oleic acid (47.81–53.14%, respectively), for Deglet Nour seed oil and from 44.88% to 49.40% for Ftimi seed oil, followed by lauric acid (21.03–25.66%), myristic acid (10.28–11.66%), palmitic acid (9.11–10.53%), linoleic acid (7.05–7.80%), and Stearic acid (3.10–3.63%). Behenic acid (0.42–0.51%) and Arachidic acid (0.38–0.54%) were present in low amounts. These results are in general agreement with those done by Besbes et al. (2004a), Nehdi et al. (2010), and Aris et al. (2013). This similarity in fatty acid profiles of oils issued from RM untreated and DIC textured treated date seed powders should reflect the absence of any significant degradation trigged by DIC. Indeed, since DIC is a high-temperature short-time process with an abrupt pressure drop toward a vacuum resulting in instant cooling, optimized DIC treatment avoids any discernible thermal degradation.