Effect of porosity on deuterium retention in titanium thin film

The absorption process of deuterium in titanium was studied in titanium film produced in two different types of copper substrate, one was a polished copper substrate and the other one was chemically etched copper substrate. Titanium film was produced by thermal evaporation method. It was activated at a temperature of 500 °C followed by deuteration at room temperature. Titanium film was characterized by XRD for crystallographic information, SEM for surface morphology, RGA for deuterium desorption studies and weight measurement for D/Ti ratio. The difference in porosity of both the samples is confirmed from XRD analysis and SEM images. Different diffusion process is observed in two different substrates from the RGA spectra. Presence of multiple trap sites in the thin film of both the substrates is observed from the RGA spectra. From the weight measurement, D/Ti ratio in polished substrate is found to be 1.03 whereas in case of chemically etched substrate it is 1.54.


Introduction
Solid deuterated targets are used in accelerator based neutron generators for generation of neutrons [1][2][3].Quantity of deuterium in the solid target is one of the important factors for high yield of neutrons.These targets are usually made up of titanium thin film coated on copper substrate.Group IV and V elements of the periodic table have a distinct property to retain deuterium at room temperature through a process called gettering [4].Titanium being group IV element of periodic table has the property of gettering.Gettering is a process in which deuterium absorption takes place in three step processes.First the deuterium molecule is adsorbed after dissociation at the surface of the titanium film, then the dissociated atom penetrates to the near surface region of the film and at the end, diffuses into the bulk of the film.The quantity of deuterium diffused in the titanium thin film depends upon the deuterium concentration at the surface, temperature of substrate,voids, defects, grain size and grain boundary in the titanium thin film.The deuterium concentration at the near surface region depends up on the concentration of deuterium adsorbed at the surface after dissociation.The dissociative adsorption at the surface depends on the dissociation energy, sticking probability, surface area and surface temperature whereas penetration in to the near surface region depends on voids, grain size and grain boundaries in the thin film and concentration gradient of deuterium.Required dissociation energy can be provided either by charging the film at higher deuterium pressure or heating the surface.The effect of higher pressure for higher retention of deuterium in titanium thin film is experimentally demonstrated in [5].Surface morphology plays an important role for absorption of deuterium in the titanium thin film.Effective surface area of the titanium thin film increases with roughness of the film.If the substrate surface area is rough, the film produced on this rough surface also carries the roughness while coating.Effective surface area increases with repetitive peak and valley formation on the film.With increased surface area, the probability of interaction of deuterium molecule with surface increases and this leads to higher rate of adsorption of deuterium.
The molecules at bulk of a solid experience zero resultant force due to the surrounding molecules.But the molecule at the surface experiences a force towards the bulk of the solid.Thus a surface tension is created.Hence the molecule at the surface has affinity to adsorb gas molecules to reduce the resultant force.The van der Waals force is involved in the interaction of a gas with a solid and this physical process is known as physisorption.In physisorption, at the equilibrium of charge distribution there is no sharing or transfer of electrons takes place.In this case the properties of adsorbent and adsorbate do not change.The interaction which involves exchange of electrons is known as chemisorption and in this process the properties of the adsorbent and the adsorbate change.The Lennard-Jones potential function [6,7] as mentioned below describes the process of physisorption.
Where ε is the depth of the potential well, σ is the distance at which the inter particle potential is zero and r is the distance between the particles.The first term in above equation ( ) is the van der Waals force at long ranges.The exponents used in this equation can be modified for different surface conditions.For interaction of gas molecules with flat surface the exponents would be 10 and 4 whereas for the interaction of gas molecules and pores it would be 9 and 3 respectively.Depending on the equation (1) above, plots have been shown for Lennard-Jones potential function versus the pore size [7].It is demonstrated that the pores with smaller size have enhanced interactions of the gas molecule and the atoms of the adsorbent.
The van der Waals forces between gas molecules and the flat surface or porous surface is described in [8].The adsorption mechanism between gas molecules and a flat surface is not significant.In this case the gas molecules are partially interactive with surface atoms from one side.In case of porous surface the gas molecule to be adsorbed is surrounded by atoms of the adsorbent.Due to additive properties of van der Waals force, the gas molecule inside the pore, capillary, crack or cavities is surrounded by many neighbouring atoms of the adsorbent.When a molecule is adsorbed, it executes attractive force and contributes to the adsorption energy, so further adsorption takes place.From the theory, it is assumed that the adsorption of gases would be better in porous surface rather than in the flat surface.Better adsorption leads to better penetration of gaseous atoms into the near surface region which leads to better diffusion due to building up of higher concentration gradient of deuterium.Because of this reason the porous surface of the adsorbent could be used for higher absorption of gases.A gas molecule goes through three different processes before adsorption on a solid surface.These are free gas diffusion, Knudsen diffusion and surface diffusion [9].The porous surface favours the Knudsen diffusion.When the mean free path of the gas molecules is higher than the dimension of the pore, the gas molecule collides more frequently with the wall of the pore rather than with each other.So probability of getting adsorbed for the gas molecule increases.
In this work, a comparative study was carried out for retention of deuterium in uniformly flat titanium thin film and porous titanium thin film on copper substrate.Porosity is observed by the micrograph of the thin film and crystallographic measurements by the x-ray diffraction.Effect of porosity on the diffusion of deuterium in titanium thin film is observed by residual gas analyzer.Quantitative measurement of the deuterium retention in porous as well as flat surfaced samples is done by measuring the mass of deuterium absorbed in the sample.Experimental details and results are described in the following sections.

Methods and materials
Titanium thin films were prepared on two types of copper substrates; mechanically polished and chemically etched by thermal evaporation method.Copper substrate of 30 mm diameter and 1 mm thick was used.Mechanical polishing of the sample was done in the automatic grinding/polishing machine made by Metatech Industries with model no.Autopol-II.Polishing of the sample was carried out on the emery with varying grit sizes from 200 up to 1500 followed by diamond polishing.Diamond polishing was carried out with diamond particle sizes of 1 and 0.25 μm subsequently.Chemical etching was performed in aqueous solution consisting of ammonium hydroxide NH 4 OH 50% in volume, hydrogen peroxide (H 2 O 2 ) 3% in volume, and distilled water 47% in volume.Chemical attack was performed at room temperature for 20 min with a quick stirring of the solution [10].The etched copper substrate was coated with titanium by thermal evaporation method in a vacuum system described in earlier work [5].Titanium wire of 0.5 mm diameter was heated to its vaporization temperature in a tungsten spiral coil in a vacuum system pumped by turbo molecular pump.Coating was carried out at a vacuum of the order of 10 −7 mbar.During coating, one interesting event was observed; instead of pressure rise, vacuum improved.As the titanium vapors were coated throughout the chamber, the vacuum during evaporation was improved to the order of 10 −8 mbar.This was due to the fact that when the titanium film was coated on the chamber wall the freshly coated surface pumped the residual gases by gettering.The thickness of the film was estimated volumetrically as 2 micron by weight difference method.The weight of the substrate before and after coating was measured by precision microbalance.Coating thickness depends on evaporation rate which is decided by the current in the tungsten spiral coil.In our case the current in the coil was 85 A which corresponds to a power of 425 W. The substrate was heated by the radiation due to the hot filament during deposition of the film.The deposition procedure was same for both types of samples.All the coated samples were annealed at a temperature of 500 °C under vacuum of the order of 2 × 10 −7 mbar for 1 h before sending them for different analysis.
After coating, the titanium thin films were deuterated in a vacuum system dedicated for vacuum and high pressure applications.Detail description of the vacuum system was published earlier [5].The vacuum system has two chambers separated by one 35 CF ball valve.One residual gas analyzer (RGA) was mounted on main vacuum chamber and one substrate heater was mounted on the chamber dedicated for deuteration at elevated deuterium pressure.One turbo molecular pumping station was used to create vacuum.At a vacuum of the order of 10 −7 mbar the titanium thin film was activated at a temperature of 500 °C for one hour and allowed to cool down to room temperature.When the substrate temperature attained room temperature after natural cooling, the chamber was filled with deuterium.At a deuterium filled pressure of 10 −4 mbar, the 35 CF ball valve was closed and the substrate chamber was segregated from the turbo molecular pump followed by filling with deuterium at a pressure of 3 bar and left for 120 h for absorption.Effect of higher pressure and higher duration on quantity of deuterium in titanium thin film was studied earlier [5] and was observed that charging with higher pressure and higher duration leads to higher quantity of deuterium in the titanium thin film.
The morphology of both type of the copper substrates without coating and with coating were observed in scanning electron microscope (SEM).The crystallographic information was studied in x-ray diffraction in a diffractometer of Rigaku make and Smart Lab model, effect of porosity on thermal desorption was studied by residual gas analyzer and the quantitative measurement of absorbed amount of deuterium was carried out in microbalance.The weight of the substrate before coating, after coating and before deuteration and after deuteration was taken.Taking the weight difference the deuterium to titanium ratio was estimated.The SEM image is analysed by image J software for particle size and the average particle size is found to be 500 nm.The morphology of Ti film on chemically etched copper substrate as shown in the figure 1(b) shows contrast character to that of Ti film on polished copper substrate.The morphology looks like a mesh type of structure having pores of slit geometry surrounded by boundaries made up of particles.The particles are grown on different planes which exhibits the roughness of the film.The average particle size is found to be 1.6 μm and the average pore size is 1 μm.  Figure 3 shows the spectra of energy dispersive x-ray analysis (EDAX) of both types of samples.The analysis was performed on a 3 × 2 μm surface of the sample.It is a common practice to coat the sample for 2-20 nm with gold or gold-palladium alloy for SEM analysis.This coating prevents the surface from charging, conducts evenly, facilitates the emission of secondary electrons and provides a homogeneous surface for imaging [11].In our case, as the sample was coated with gold for SEM imaging and EDAX analysis was followed by SEM imaging, hence the gold peak is observed in the EDAX spectrum.In the Ti film on mechanically polished copper substrate the titanium weight percentage is 100 whereas in the chemically etched sample it is 99.6.As the range of electron beam in EDAX analysis is ∼2 μm and also the thickness of the Ti films under investigation is ∼2 μm, the presence of copper peak in the spectra is negligible.

Crystallographic analysis
The spectra shown in figure 4 represent the XRD spectra with and without annealing of the chemically etched copper sample.From the spectrum of without annealing in figure 4, there is no diffracted x-ray peak that indicates the Ti film as amorphous.As the titanium film on the copper surface is amorphous, and not wellordered, the reflected x-rays from the copper surface that is beneath the disordered titanium film, scattered in many directions.So could not form diffracted beam.After annealing at a temperature of 500 °C for 1 h the Ti film is transformed to crystalline.Figure 5 shows the XRD spectra of the Ti thin film on polished copper as well as on chemically etched copper substrate taken after annealing of the substrate.The Ti peak is identified from the JCPDS card no.44-1288 and the Cu peaks are identified from the JCPDS card no.85-1326.The FWHM of the Ti peak on chemically etched copper substrate is 0.363°and that on polished surface it is 0.321°.The width of the x-ray diffraction curve increases as the grain size decreases [12].The grain size is calculated from the Debye-Scherrer's equation [12].The size of the Ti grain on chemically etched copper is 23.24 nm whereas on polished copper it is 26.19 nm.The grain size is smaller in the film produced on chemically etched copper than that of polished copper.Decrease in grain size in case of chemically etched substrate indicates increase in porosity [13,14].So the Ti film produced on chemically etched copper substrate is porous than the Ti film produced on the polished copper substrate.
Figure 6 represents a comparative study of the spectra of titanium peak for titanium thin film on chemically etched copper samples with and without deuteration.The shifting of Ti peak of the deuterated sample towards lower angle is due to increasing of lattice parameter upon occupation of interstitial sites by deuterium atoms [15].The FWHM of the deuterated sample is 0.568°whereas that of only Ti film sample it is 0.363°.The FWHM of the deuterated sample is higher than that of without deuterium sample.The grain size is reduced to 14.74 nm from 23.14 nm after deuteration.Broadening of the peak width is attributed to the occupation of the interstitial sites among the titanium atoms by deuterium.Thus absorption of deuterium leads to expansion of titanium lattice [16,17].The reduced diffracted peak intensity in case of deuterated sample is because of the tensile stress developed due to expansion of the crystal lattice after deuterium absorption [18].With reduced grain size, the structure factor reduces after deuteration.As the intensity of the diffracted x-ray beam depends proportionately on the structure factor [12], with reduced grain size, the intensity decreases after deuteration.
From XRD analysis, it was observed that after coating, the samples required annealing, so that the film becomes crystalline.Estimation of grain size was obtained from the peaks obtained in the XRD spectrum.In case of chemically etched sample, the grain size was decreased than that of polished sample.Decrease in grain size is a result of the porosity as reported in [13] and [14].The deuterium absorption in the Ti film is observed on the XRD spectrum in the figure 6 due to change in characteristics of Ti peak after deuteration.

Thermal desorption analysis
Figure 7 represents the substrate temperature versus time.The rise time of the temperature to 500 °C is 3 min and 25 s.The rise time is 5.7% of the total duration of the thermal desorption experiment.Figure 8 represents the residual gas analyzer spectra for thermal desorption of molecular deuterium from Ti film on polished copper and chemically etched copper substrate.The residual gas analyzer spectrum was obtained by heating the sample at a temperature of 500 °C under vacuum.The heating started mark on the spectrum is the time when the heating was started.As the temperature rises at the substrate, the deuterium atoms start moving towards the surface.At the surface a deuterium atom recombines with other deuterium atom and is released to vacuum.During heating, a deuterium atom which is absorbed at the surface recombines with other deuterium atom and released quickly.This quick release of deuterium gives to a sharp peak of the deuterium partial pressure.The deuterium atom which is diffused to the bulk of the thin film during pump down phase diffuses in reverse direction to the surface during heating.The diffused deuterium atom takes several jumps between nearest neighbours, diffuses through the voids, grain boundaries and thus takes time to pass through different locations on varying thickness before reaching at the surface.As the bulk deuterium atom takes longer time than the surface atom, the partial pressure due to the diffused deuterium atom has a wider distribution of time.In the RGA spectrum for polished sample as shown in the figure 8, two peaks are found, the sharp peak corresponds to the deuterium desorption from the surface whereas the wider peak is due to desorption from the bulk.The peak  height due to surface desorption is 7.6 × 10 -6 mbar whereas the peak height due to bulk desorption is 3.5 × 10 -7 mbar.The RGA spectrum for deuterium desorption from the chemically etched copper substrate is shown above that of polished sample.The peak due to desorption from the bulk is higher than that of surface desorption.The peak height due to surface desorption is 2.9 × 10 -5 mbar whereas the peak height due to bulk desorption is 3.6 × 10 -4 mbar.Higher amount of deuterium desorption from the surface as well as bulk of the chemically etched sample is the evident for higher retention of deuterium during charging.Thus the higher diffusion rate of deuterium in the chemically etched sample is due to changed microstructure with higher porosity, reduced grain size and different grain boundaries than that of the polished sample.There are limitations for quantification of the amount of desorbed gas due to outgassing in the pump down stage, so the RGA spectrum due to thermal desorption could give only qualitative results.It is useful for comparative studies.

Determination of D/Ti ratio
The mass of Ti coating was estimated by the difference in the weight of before and after coating.The mass of deuterium that was absorbed in the Ti film was estimated by the weight difference of before and after deuteration.The average mass was taken out of 10 measurements for each sample.The average mass of mechanically polished copper substrate (MP_Cu), with Ti coating (MP_Cu_Ti) and with deuteration (MP_Cu_Ti_D) is 6.206631 g, 6.211781 g and 6.212221 g respectively.The amount of deuterium absorbed in mechanically polished sample is 440 μg.The average mass of chemically etched copper substrate (CE_Cu), with Ti coating (CE_Cu_Ti) and with deuteration (CE_Cu_Ti_D) is 6.315126 g, 6.320486 g and 6.321171 g respectively.The amount of deuterium absorbed in chemically etched sample is 685 μg.From the mass of deuterium and titanium the atomic ratio was calculated.In case of Ti film on polished copper, the D/Ti ratio is found to be 1.03 whereas in case of Ti film on chemically etched copper, the ratio is 1.54.The D/Ti ratio found in chemically etched sample is comparable with the previous results [19].The increased deuterium concentration in chemically etched sample is attributed to the overall contribution from increased porosity, reduced grain size and different grain boundaries.The measurement uncertainty is in the range of 10 -5 .The samples were handled under argon purging while shifting from coating chamber to deuteration chamber.Only during measurement in the balance these were exposed to air.The air exposure condition for all samples was same.During exposure to air, passivation layers of oxide, carbides are formed.Once passivation layer of oxides/carbides of few nm thickness is formed over the titanium thin film, there would not be any further adsorption at the surface at room temperature.Mass of oxygen/carbon in few nm thick is in the order of less than pico gram.So the effect of air exposure of samples during weight measurement is neglected for calculation of deuterium mass.

Conclusion
Though there are many articles regarding deuterated targets available in public domain and also it is commercially available, deuterated titanium target in a chemically etched copper substrate is not reported elsewhere.As the deuterium absorption is better in titanium film on chemically etched copper substrate, requirement for higher charging pressure of deuterium will be reduced.Thus deuterated targets can be made with easy instrumentation and safe operating pressure of deuterium.The result of one sample each is discussed in this report.But in our observations in the laboratory it is found that there is same trend of RGA spectrum for multiple samples.The D/Ti ratio is observed within a range of ±6% for multiple samples of both types.

Figure 1
shows the morphology of the copper substrate before titanium coating on (a) mechanically polished, (b) chemically etched.Both the figures are of 50 KX magnifications.Effect of etching is observed on the figure 1(b).As per the visuals, the roughness is increased due to different planes.Figure 2 shows the morphology of titanium thin film of two different types of copper substrate i.e.(a) mechanically polished and (b) chemically etched taken after annealing.Figures with same magnification are shown beside one another.The magnification of the images from top is 10 KX, 25 KX, 50 KX and 100 KX respectively.The morphology of the Ti film on polished copper shows uniformly distributed similar particles.The film exhibits smooth surface with closely packed particles.

Figure 1 .
Figure 1.SEM images of copper substrate before titanium coating on (a) polished (b) chemically etched with 50 KX magnifications.

Figure 2 .
Figure 2. SEM images of titanium thin film produced on (a) polished copper and on (b) chemically etched copper with different magnifications.

Figure 3 .
Figure 3. EDAX spectrum for Ti film on (a) mechanically polished and (b) chemically etched copper substrate.

Figure 4 .
Figure 4. XRD spectra of Ti thin film on chemically etched copper sample before and after annealing.

Figure 5 .
Figure 5. XRD spectra for Ti thin film on polished copper and chemically etched copper.

Figure 6 .
Figure 6.XRD spectrum of Ti peak with and without deuterium absorption in Ti film on chemically etched copper substrate.

Figure 7 .
Figure 7. Substrate temperature versus Time for thermal desorption analysis.

Figure 8 .
Figure 8. RGA spectrum of deuterium obtained during thermal desorption at a temperature of 500 °C of Ti film on polished copper chemically etched copper.