Introduction

Fusion energy is believed to be the potential energy source for the future. Two isotopes of hydrogen, namely deuterium and tritium are identified as feasible fuels for fusion reaction owing to their high reaction cross-section with the least energy requirement [1]. Deuterium and tritium fuse in the presence of plasma and yield 17.6 MeV of energy according to the reaction in Eq. 1.

$$D+T\to H{e}^{4}+n+17.6MeV$$
(1)

Deuterium is available naturally in large quantities but due to the radioactive nature of tritium, it is not abundant in nature, thus it has to be synthesized. Generally, lithium is used for the production of tritium as per the reactions 2 & 3. Therefore, lithium-based ceramics are used as breeder material in nuclear fusion reactors.

$$L{i}^{6}+n\to T+H{e}^{4}+4.8MeV$$
(2)
$$L{i}^{7}+n\to T+H{e}^{4}+n-2.5MeV$$
(3)

Lithium titanate is a promising material for breeder application in nuclear fusion reactor, due to its superior properties such as minimal activation characteristics and good tritium release properties as compared with other ceramics [1, 2]. High Li density is one of the essential criteria for tritium breeding application [3]. Lithium content in the breeder material will decrease during operation due to the transmutation of lithium [2]. Additionally, lithium being volatile in nature, is lost from the ceramic during the high temperature synthesis process [4,5,6]. To compensate for lithium losses during the reactor operation and during the synthesis stage, extra lithium is added to lithium titanate. Thus, hyper-stoichiometric lithium titanate [Li(2+x)TiO(3+y)] with high Li/Ti ratio have been proposed [7,8,9]. Lithium titanate ceramics can be produced through different processing techniques. The conventional production route of lithium titanate includes solid state synthesis followed by high temperature sintering process to achieve the desired shape (e.g. pebbles) with the required density [3]. During production of lithium titanate, high temperature sintering step is associated with the loss of lithium due to its volatile nature. The lithium loss has been found to be a function of sintering temperature and time. Thus, it is important to quantify lithium content in the ceramic breeder material. Conventional chemical analysis techniques like Atomic absorption spectroscopy (AAS), Inductively coupled plasma mass spectrometry (ICP-MS) and Inductively coupled plasma optical emission spectroscopy (ICP-OES) require the ceramic sample to be in the solution form, which needs strong acids for dissolution for exact quantification of the element of interest. Although these methods have good accuracy and precision but are destructive in nature and the sample integrity is lost after the analysis. Any method which is capable of analysing the solid sample without dissolution is preferred over the conventional wet chemical techniques. On the other hand, analytical techniques namely particle induced X-ray emission (PIXE), X-ray fluorescence (XRF) and neutron activation analysis (NAA) are non-destructive in nature but not suitable for the quantification of low Z elements including lithium due to their unfavourable atomic and nuclear properties [10]. Nuclear analytical technique namely particle induced gamma-ray emission (PIGE) technique is very sensitive to low Z elements including lithium [10,11,12,13].

In the present study, stoichiometric (Li2TiO3) and hyper-stoichiometric (Li2+xTiO3+y) lithium titanate ceramic samples have been synthesized at various temperatures ranging from 900–1100 °C and characterized by various techniques namely scanning electron microscopy and X-ray diffraction. Further, the external PIGE technique using nitrogen (from air) as an external current normalizer has been utilized to evaluate lithium concentration in stoichiometric and hyper-stoichiometric Li2TiO3 samples. Compositional information obtained in the present study will be useful for the measurement of the net lithium concentration in lithium-based breeder ceramics for nuclear fusion reactors.

Principle of the PIGE technique

Particle induced gamma-ray emission (PIGE) is an isotope-specific nuclear analytical technique (NAT). It involves the interaction of energetically charged particles (p, d and α) with the target nuclei and results in the emission of prompt gamma rays from reactions such as inelastic-scattering (p, p'γ) or from nuclear reactions like (p, αγ), (p, nγ) and (p, γ)[14,15,16,17]. PIGE method is very sensitive to the low Z elements (usually 3 ≤ Z ≤ 16) including Li in the wide spectrum of solid materials. The gamma-rays thus emitted are the characteristic signature of the isotopes of elements. Figure 1 represents the schematic diagram showing interaction between energetic proton beam and the target nucleus.

Fig. 1
figure 1

Schematic diagram showing the interaction of proton beam with the nucleus of the target atom (Li)

Full size image

PIGE involves online measurement of prompt gamma rays emitted during de-excitation of excited nuclei of the isotope. The activity (A) produced in the PIGE method is given by Eq. 4 [18].

$$A= N\times I\times {\int }_{{E}_{0}}^{0}\frac{\sigma (E)}{S(E)}dE$$
(4)

where, N is the number of target atoms per cm3, I is the beam current, σ(E) and S(E) are the cross-section and the stopping power of the isotope of analyte in the target at a particular particle energy, respectively. The cross section of the relevant isotope for a given analyte will be the same for both the sample and the standard at a given proton beam energy. If the sample is thin and conducting, a Faraday cup can be placed behind it to measure the current directly. But if the sample is thick and insulator, then the beam is stopped by the sample and it is difficult to measure the beam current variation by the conventional approaches. Under these circumstances, current normalisation is carried out through the Rutherford Backscattering Spectroscopy (RBS) method [19, 20]. Since it is not necessary to measure absolute current in relative PIGE method, in-situ current normalization approach is an advanced method for monitoring the fluctuations of the beam current during the experiment. In this method, the sample is supplemented with a known amount of an element that is highly sensitive to PIGE and is not present in the sample matrix. The count rate (counts per second, CPS) of the sample is normalized with sensitivity (CPS per unit mass) of the current normalizer to account for the current variation. This normalization process makes the analysis independent of beam fluctuation during irradiation [17]. In the in-situ current normalised PIGE method, the sample and the standard are pelletized in a similar matrix using a hydraulic press, resulting in a similar stopping power S(E) in the sample and the standard. Thus, the ratio of count rate for an analyte ‘x’ present in sample and standard is given by Eq. 5 [19]

$$\frac{{CPS}_{x, sam}}{{CPS}_{x, std}}= \frac{{C}_{x, sam}}{{C}_{x, std}}\times \frac{{I}_{sam}}{{I}_{std}}$$
(5)

where, Cx is the concentration of the analyte ‘x’ in sample and standard and I is the corresponding beam current. In in-situ current normalisation process, the variation in the current is monitored through count rate of current normalizer (y). The ratio of absolute current in sample and the standard is proportional to the ratio of the sensitivity of the current normalizer ‘y’ in sample and standard as given by Eq. 6.

$$\frac{{I}_{sam}}{{I}_{std}}=\frac{{S}_{y,sam}}{{S}_{y,std}}= \frac{{\left({CPS}_{y,sam}/{C}_{y}\right)}_{sam}}{{\left({CPS}_{y,sam}/{C}_{y}\right)}_{std}}$$
(6)

where, Sy is the sensitivity of (y) and Cy is the concentration (mg kg−1 or % mass fraction) of the in situ current normalizer (y) in sample and standard and the Eq. 5 is modified as Eq. 7 [19].

$${C}_{x,sam}= \frac{{{(CPS}_{x,sam}/{S}_{y,sam})}_{CN}}{{{(CPS}_{x,std}/{S}_{y,std})}_{CN}}\times {C}_{x,std}$$
(7)

In the present work, beam current was monitored through the count rate of gamma peak at 2313 keV from the14N(p,p’γ)14N reaction from the atmospheric nitrogen. Table 1 provides the relevant nuclear data of elements of interest and corresponding thick target gamma ray yield at 4 MeV proton energy.

Table 1 Nuclear data of the elements [21]
Full size table

Experimental

Stoichiometric Li2TiO3 powder was produced through solid state reaction of Li2CO3 (> 99%, Sigma-Aldrich) and TiO2 (AR Grade, M/s Travankore Titanium Products Ltd., India). Li2CO3 powder was mixed with TiO2 powder using wet ball milling under ethanol media for 24 h. The wet mixture was dried in an oven for 2 h at 90 °C. The dried mixed powder was calcined at 700 °C for 2 h for phase formation. Particle size of the calcined powder was measured by particle size analyser in ethanol medium (Cilas 1090). Using similar processing technique, hyper-stoichiometric lithium titanate with different Li/Ti ratio (Li/Ti = 2.06, 2.16, 2.20) was prepared and calcined at 700 °C for 2 h. The calcined powder phase purity was confirmed through X-ray diffraction (XRD; Inel XRG 3000). Stoichiometric and hyper-stoichiometric lithium titanate powders were compacted into green compacts (10 mm diameter and 1–2 mm height) using a uniaxial hydraulic press at ~ 150 MPa pressure. The stoichiometric Li2TiO3 pellets were sintered at different temperatures varying from 900 to 1100 °C for 4 h to investigate sinterability and lithium loss behaviour. Based on the sintering results of stoichiometric lithium titanate, hyper stoichiometric lithium titanate pellets were also sintered in air atmosphere at 1050 °C for 4 h. The phase purity of the sintered pellets was analysed through XRD. Mass loss during sintering was measured and the relative density of the pellets was determined by Archimedes’ method. Microstructural characterisation of the sintered specimens was carried out using scanning electron microscopy (SEM Zeiss Evo 18). Lithium content in sintered ceramic specimens was measured using external PIGE.

In the present investigation, external (in-air) PIGE method using nitrogen as an external current normalizer was utilized for the quantification of Li in ceramic samples. A facility with 5 MeV proton beam from the FOlded Tandem Ion Accelerator (FOTIA), Bhabha Atomic Research Centre (BARC), Mumbai, was already set up and optimised for external PIGE equipment. The collimated proton beam of 3 mm diameter and 5 MeV energy was extracted through a vacuum chamber through a window made of tantalum (Ta) metal with a thickness of 25 μm. Ta has been chosen as a window material because the other conventional window materials used such as Al, Ti, Si3N4 might contain one of the elements to be measured. This Ta window helps in sustaining the vacuum in the chamber however proton energy decreases by 1.25 MeV while passing through it which was calculated by the SRIM software [18,19,20]. Consequently, 3.75 MeV proton beam has been used in all the external PIGE experiment. Direct sintered pellets of samples and elemental standards were irradiated with 3.75 MeV proton beam. The prompt gamma-rays at 429 and 478 keV from 7Li(p, nγ)7Be and 7Li(p, p’γ)7Li nuclear reactions, respectively, were measured using a P-type high purity germanium (HPGe) portable detector with 50% relative efficiency, having 1.9 keV resolution at 1332 keV of 60Co. The detector's energy calibration was carried out with reliable sources, such as 152Eu. Irradiation cum acquisition time per sample was 15 min or more depending upon the current which was maintained ~ 10 nA (approx. 9 µC charge). The counts under the peaks of 478 keV of Li and 2313 keV of N were more than 1,50,000 and 20,000, respectively. The acquired spectra were analysed by peak-fit method using Pulse Height Analysis Software (PHAST) developed at BARC. The samples and standards were irradiated in similar positions and geometrical configurations with regard to the detector in the current work, which used a relative approach of concentration calculation. Lithium titanate from Sigma-Aldrich (99.9% pure, AR grade) has been used for method validation and accuracy verification. The nitrogen (78%) contained in the air is employed for beam current normalisation using the relative technique of calculations, which quantifies the element in the sample using standards of known composition [20]. In the present studies, count rate under the prompt γ-ray peak (2313 keV) from nuclear reaction (14N(p,p′γ)14N) has been used for current normalization. In the relative method, the final Eq. 7 is reduced to Eq. 8

$${C}_{Li,sample}=\frac{{{(CPS}_{Li}/{CPS}_{N, 2313})}_{CN, sample}}{{{(CPS}_{Li}/{CPS}_{N,2313})}_{CN, standard}}\times {C}_{Li,standard}$$
(8)

where CPS is the count rate of Li and N of sample and standard and C is used for the concentration of Li in sample and standard and CN is the current normalizer[19].

Results and discussions

Phase purity of calcined stoichiometric lithium titanate (Li2TiO3) was confirmed through XRD (Fig. 2). All the observed peaks match with the monoclinic structure of β-Li2TiO3 phase (ICDD PDF 033–0831). The mean particle size β-Li2TiO3was measured as ~ 3.7 μm.

Fig. 2
figure 2

XRD pattern of lithium titanate calcined at 700 °C for 2 h

Full size image

The calcined samples ware sintered at various temperatures, during sintering of lithium compounds, weight loss due to volatilisation of lithium was observed. In order to confirm this, mass of the sample pellets was measured before and after sintering and the corresponding calculated mass loss % values (difference between the mass taken before and after sintering) are tabulated in Table 2. It is observed that with increase in sintering temperature, the sintered density of the pellets increases with concomitant increase in mass loss. Since optimized pebble density for breeder application is between 85 and 90%, sintering of all hyper-stoichiometric compositions was done at 1050 °C for 4 h.

Table 2 Mass loss and relative density of lithium titanate at different sintering temperatures
Full size table

Figure 3 shows a representative gamma-ray spectrum of lithium titanite sample from external PIGE using 3.75 MeV proton beam and 10 nA beam current. Peaks at 429 keV and 478 keV represent high yield of lithium from the sample. Apart from lithium peaks, other peaks are due to titanium (48Ti, 889, 983, 1312 keV), nitrogen (14N, 2313 keV), back ground due to tantalum window (181Ta, 605, 705, 1047 and 1815 keV), iron structural material of beam tube (56F, 844, 1811 keV) and aluminium impurity (27Al, 1014 keV) [15, 16, 21]. As seen in the Fig. 3, there are no interference of these mentioned gamma-rays mainly due to Ta and Fe with gamma-rays of interest from Li and N.

Fig. 3
figure 3

Prompt gamma-ray spectrum of a lithium titanate sample irradiated with 3.75 MeV proton beam at ~ 10 nA current

Full size image

Concentration of lithium in all the lithium titanate samples sintered at different temperatures was calculated by relative PIGE method as per Eq. 8 and the corresponding calculated mean concentrations (wt%) of Li are given in Table 3. The mean Li concentration was calculated from three replicate measurements with the individual propagated uncertainties within ± 0.73% and the corresponding standard deviation (N = 3) values within 0.09 wt%. Previous studies on quantification of lithium using PIGE in Li2TiO3 synthesised with different starting material and varying sintering temperatures reported lithium concentration in the range of 11.2–12.68% with ± 2% uncertainty [22]. Other investigation on lithium-based ceramics such as Li2TiO3, LiAlO2 and Li2SiO3 using ~ 4 MeV proton beam reported lithium concentration in the range of 10.1–15.8wt% with ± 5% propagated uncertainty [20]. The results obtained in this investigation are in good agreement with previous results for lithium measurement using PIGE.

Table 3 Calculated Li concentration (wt%) in lithium titanate ceramic sintered at different temperatures for 4 h by external PIGE method
Full size table

The deviation of calculated lithium concentrations with the theoretical Li content in lithium titanate is in the range of − 0.9% to − 4.7% as can be seen from Table 3. This deviation is due to increase in mass loss with increase in sintering temperature due to the volatile nature of lithium at high temperature (Table 3).

XRD pattern of hyper-stoichiometric lithium titanate (Li/Ti = 2, 2.06, 2.16, 2.20) calcined at 700 °C for 2 h are shown in Fig. 4(a). It is observed from the patterns that β-Li2TiO3 is the major phase in all the compositions along with the presence of residual Li2CO3 for samples with Li/Ti ratio as 2.16 and 2.20. Presence of Li2CO3 in higher stoichiometric lithium titanate is also reported in earlier studies [5, 23, 24]. Figure 4(b) shows XRD pattern of hyper-stoichiometric lithium titanate (Li/Ti = 2, 2.06, 2.16, 2.20) sintered at 1050 ͦ C for 4 h. It is observed that all compositions up to Li2.2TiO3+y have single phase β-Li2TiO3 and there are no additional peaks observed.

Fig. 4
figure 4

XRD patterns of (a) 700 °C calcined and (b) 1050 °C sintered lithium titanate with different stoichiometries

Full size image

Figure 5 shows SEM images of hyper-stoichiometric lithium titanate with varying Li/Ti ratio sintered 1050 °C for 4h. The microstructures indicate liquid phase sintering caused due to the presence of excess Li2CO3 in the calcined powders of hyperstoichiometric lithium titanate samples.

Fig. 5
figure 5

SEM micrographs of Li2+xTiO3+y sintered at 1050 °C for 4 h

Full size image

The mean Li concentration was calculated by external PIGE method using nitrogen as external current normalizer from three replicate sample analysis with the individual propagated uncertainties within ± 0.73% and the corresponding standard deviation (SD) (N = 3) values within ± 0.12%. Unlike the stoichiometric Li2TiO3 sample, the lithium loss is found to be more in hyper-stoichiometric sample, which is evident by mass loss calculation as well as lithium weight percent data obtained by external PIGE (Table 4). With the increase in lithium concentration in lithium titanate, lithium loss also increases significantly. At Li/Ti = 2.16, the lithium losses are around twice than that of stoichiometric compound. The theoretical weight of non-stoichiometric composition (Li2+xTiO3+y) has been calculated taking y = x/2 to balance the overall charge of the compound.

Table 4 Calculated and measured Li concentration (wt%) in the hyper-stoichiometric lithium titanate ceramics sintered at 1050 °C for 4 h by external PIGE method
Full size table

Conclusions

Stoichiometric and hyper-stoichiometric Li2TiO3, which are potential tritium breeder material in nuclear fusion reactors, have been prepared by solid state reaction and sintered at various temperatures. Non-destructive quantification of lithium present in these ceramic samples have been successfully carried out by relative external (in air) PIGE method using 3.75 MeV proton beam and 14N as an external current normalizer. Losses incurred during different sintering temperatures in stoichiometric and hyperstochiometric lithium titanates are quantified by comparing the results with the theoretical and calculated Li concentration values. Present study suggests that it is required to add 6–8% of excess lithium in the precursors during synthesis in order to avoid secondary phase formation and to increase the breeder lifetime. External PIGE method was found to be simple, rapid and non-destructive method for quantification of Li in such lithium based ceramics.