The Choice of Solvents for Cleaning Metal Optics in Order to Increase the Threshold of Optical Damage

Metal optics is actively used in IR optics, which makes it necessary to clean it from dust and other operational contaminants. This increases the life of these notably expensive products. Usually their mechanical wiping is excluded, since it leads to the appearance of scratches and other defects on the surface, markedly increasing the scattering of radiation. Methods for the selection of solvents have been analyzed and experiments have been carried out that make it possible to accept efficient their use in the purification of metal optics from contaminations. Contaminations have different physicochemical properties and are present, as a rule, together on the surface of metal optics (taking into account the corrosion properties of the latter). Solvent compositions based on azeotrope mixes with freon-114B2, having a solubility parameter (δ) corresponding or close to the solubility parameter (δ) of the main (by weight) contamination or a mixture of contaminants which are present on the optical surface, were used.


Introduction
Metal mirrors are widely used in IR optics. During operation of metal mirrors, they must be regularly cleaned from dust and other operational contaminants. Particular attention in this case should be applied to cleaning the mirrors of high-power lasers [1].
The group of compounds which are presently used as solvents covers > 300 compounds, without taking into account a virtually unlimited number of solvent mixtures. The differences in their physicochemical properties are large. Already for these reasons it is difficult to classify and select solvents for the purposes of chemical cleaning of metal optics. Taking into account the corrosion properties of metal optics, especially those made from aluminum, copper and their alloys, it is advisable to use dehydrated solvents and their mixtures in the cleaning process. The latter are characterized by a variety of properties associated with different types of intra-and intermolecular interactions, which determine many of their intra-and intermolecular properties. In particular, the role of association processes, specific solvation and formation of complexes is important.
Usually solvents are classified according to the classes of chemical compounds -alcohols, ketones, carboxylic acids, esters, nitrogen-containing compounds, etc. But this approach does not allow one to understand their similarities and differences when choosing the solvents for the purification of various contaminants whose solubility and miscibility in different solvents reflect individual features in the nature of the interaction between particles in a given system. Therefore, solvents are often classified based on other physical and chemical properties.
In the model of electrostatic solvation [3][4][5], the effect of constant dipole moment (D) and dielectric constant (μ) for the classification and selection of solvents is highlighted. The dipole-dipole interaction between constant fields is observed, and what is more the magnitude of the dipole of the molecule depends primarily on the size of the molecule and the nature of the functional group present in the molecule (table 1 [2]). This model treats the ion as a homogeneous conducting ball and the solvent as a continuous medium characterized by a macroscopic dielectric constant (dielectric continuum) and makes it possible to estimate the ion solvation, understood as the difference between the works required charging the model sphere in a vacuum and in a given solvent. In the case of a continuum [3,6], having a dielectric constant D < 0, this work (W) can be expressed as follows where R -is the radius of the ions. Hence, for the case of thermodynamic solvation [7] of one mole of ions, the change in Gibbs energy (ΔGi) .
However, D is not constant and depends on the electric field strength, and, therefore, on the distance "molecule-ion", the concentration of ions in solution, the structure of solvent molecules, etc. This can be taken into account using the theory of the dielectric nonlinear effect [8], which predicts changes in D of a solvent in the direct vicinity of an ion (E ≈ 10 9 V / cm 2 ) using the formula where Dd, D -dielectric constant differential and at zero field strength (pure solvent), respectively; n is the refractive index of light for pure solvent; b is a constant equal to 1.08 • 10 -8 units. SGS 2 . The introduction of the expression (4) into the theory of Born's solvation gives the best interpretation [9] of the energy of the processes in formula (2) for the Gibbs energy (ΔGi) of ion solvation.
A complicated version of the electrostatic theory of solvation is possible, where the latter has consistent stages: 1) the formation of the coordination sphere of an ion in the gas phase by the interaction of ions with solvent molecules; 2) introduction of the ion having the solvation shell into the solution. The energy effect of solvation (E) is a complex function of the "ion -dipole" distance f (r), ionized potentials, polarizability, dipole and quadrupole moments, and D of solvent. The sum of energy effects of these stages: 1) is the sum of the energies of the "ion -dipole" and "ion -quadrupole" interactions with solvent molecules (Е1), induction interaction of the ion with the induced moment of the solvent If we neglect the difference between the Gibbs energy at solvation ( ) of 1/D should be observed. Difficulties are associated with the use of the Born model [3] and its modifications to describe the solvation of "composite" organic ions, due to the need to determine ionic radii, the charge localization place (z), and the way to account of the nonlinear effect.
However, (D) and (μ) often change and are not very suitable as a measure of polarity, since the total number of all interactions between solvent molecules and dissolved contamination is much more extensive and includes non-specific Coulomb and specific (hydrogen, "donor -acceptor" and solvophobic connections) interactions.
Since the reactivity of the solvent is determined by the measure of its strength as a donor or acceptor,   DN and AN). The main disadvantages of chemical classification by acid-base properties: (DN) and (AN) are identified for a limited number of solvents due to the difficulties of experimental determination, and the uncertainty of the prospects for the distribution of these data to solvent mixtures for analyzing their properties.

Experimental method
From the review [10] of the impurities present on the surface of metal optics, it can be assumed that their removal is most effective with a mixture of solvents. Therefore, it is advisable to analyze the parametric theory of solubility [11] as the most promising for choosing both individual solvents to remove certain types of contamination from the surface and predicting the properties of solutions formed  where Since at the temperature of the liquid phase the cohesion energy can be considered approximately equal to the evaporation energy ( 1 V E  ), then (7) can be rewritten as where V -is the molecular volume of the solvent.
Since (c) is the cohesion energy per unit volume, it can be called the "specific density of the cohesion energy". There is a strong interaction between the two components in the solution. At the same time, if the molecular volumes (V) of both components are significantly different, and the energy of thermal motion is significantly greater than the cohesion energy, we can assume that the entropy of mixing (  (11) where V M is the molar volume of the component.
An example of the effectiveness of the application (δ) is the interaction between solvents and polymers: since on the surface of metal optics, contaminations with different properties are usually present together, in order to increase the affinity of the solvent medium to contamination (Afch) and in accordance with economic and technological requirements, it is necessary to use solvent mixtures when cleaning. To calculate (δ) of the mixture of solvents we used the expression [ where M i x -is the mole fraction of the component, calculated as interest. Other freon's are either too volatile, or scarce and expensive [13], or contain chlorine atoms, as a result of which they are not stable to decomposition. In addition, the azeotrope mix "freon-114B2acetone" (2.4% by weight of acetone (composition 3)) contains the maximum number of fluorine atoms compared with other azeotrope mixes (both in the freon itself -4 atoms, and in the azeotrope mixes as a whole, due to the relatively small content of the other component). Fluorine compounds adsorbed on the metal during cleaning, due to the high ionization energy, increase the optical damage threshold of metal optics.
This occurs because of the extremely high ionization potential of the fluorine atom (Table 4) and the high ionization energy of compounds containing fluorine (Table 5). During adsorption on the optical surface, fluorine compounds prevent the growth of electron concentration in the space above the surface to the values corresponding to ionized gas, when exposed to laser radiation. Even in the presence of a significant emission of electrons in the space directly adjoining to the optical surface of metals, which may occur under certain conditions due to the low work function of electrons from metals, these electrons will be absorbed by freon-114B2, pairs of which will be present in the space above the optical surface as a result of its evaporation under the influence of laser radiation. This is due to the significant electron affinity of both the fluorine and bromine atoms themselves, of which freon-114B2 consists, and of compounds containing these atoms. Data on the electron affinity are presented in tables 6, 7. A comparison of the parametric theory of solubility with experiment has been carried out for the thermodynamic values of non-polar and low-polar solvents at room temperature: heat capacity at constant pressure ( E p C  ) is consistent with experimental data, coinciding with an accuracy of 10-20% of thermal energy RT (2.5 kJ / mol at room temperature) for most of the solutions of non-polar solvents.
Usually the results do not agree when the maximum value ( E p C  ) is close to the data for ideal solutions and < 0.2 kJ / mol, which is obviously due to the inapplicability of using the Berthelot approximation.