**3. Issue of possible decomposition in radiation‐damaged zircon**

In the study of radiation effect and metamict state, what happens at the atomic level is an unclear and important question. Researchers have focused on issues such as possible changes in the coordination number of Zr [27], radiation‐induced disordering rather than amorphiza‐ tion [31], damage‐related distorted and disoriented isolated silica tetrahedra [16], the fraction of amorphized phase [14, 46], as well as whether metamictization leads to phase separation or damaged minerals decompose into their oxides, and what is the structural state of the decom‐ posed phases [47, 48]. This issue of radiation‐induced decomposition was, in fact, debated over decades [30]. Based on infrared data on metamict zircon, a two‐stage damage process was proposed [49]. It was suggested that the first stage produces, throughout the lattice, highly stressed and expanded zircon with distorted SiO4 tetrahedra, while the second stage was suggested to result in the decomposition of ZrSiO4 to ZrO2 and SiO2 , probably together with some aperiodic ZrSiO4 [49]. However, the decomposition of damaged zircon was often reported in zircon samples only annealed experimentally at high temperatures, in which the decomposition commonly leads to different polymorphs of ZrO2 and glassy silica. Monoclinic ZrO2 was found in heavily damaged samples heated to 1373 K [50]. In a high‐temperature study, Vance and Anderson [15] observed cubic and tetragonal ZrO2 at 1073 and 1373 K, respectively. It was reported that highly metamict zircon contained randomly orientated ZrO2 when annealed at 1173 K, and further annealing at 1523 K resulted in monoclinic ZrO2, as well as a silica glass phase [51]. Ellsworth et al. [52] suggested that decomposition of metamict zircon into ZrO2 and glassy SiO2 could be one possible path for recrystallization. In contrast, a high‐temperature neutron work by [53] suggested that zircon decomposes into crystalline β‐cristobalite (rather than silica glass) and tetragonal ZrO2 . An X‐ray powder diffraction study at high temperatures reported the appearance of pseudo‐cubic ZrO2 [54]. Meldrum et al. [55] observed a decomposition of zircon into tetragonal ZrO2 when irradiating zircon with heavy ions at around 950 K. As discussed by different works [56, 57], some of these previous works based on X‐ray diffraction measurements might have experienced difficulties in the determi‐ nation of cubic and tetragonal ZrO2 .

Vibrational (Raman and infrared) spectroscopy is a good analytical tool for resolving these problems related to decomposition of ZrSiO4 into ZrO2 and SiO2, as pointed out by Ref. [30], because Raman and infrared spectra of zirconia (ZrO2 ) have been well studied previously and also because vibrational spectroscopy has short length scales. ZrO2 has three common poly‐ morphs at different temperatures. The room‐temperature phase is monoclinic, while the tetragonal and cubic phases occur at high temperatures [58]. The theoretical calculations [59, 60] have given the optical phonon modes (for zero wave vector) for each polymorph of zirco‐ nia ZrO2 . The monoclinic ZrO2 has space group <sup>5</sup> *C*2*<sup>h</sup>* /*P*21 /*c* and *Z* = 4, and it has eighteen Raman modes and fifteen infrared modes [9A<sup>g</sup> (R) + 9B<sup>g</sup> (R) + 8Au (IR) + 7Bu(IR)] (R indicating Raman‐ active and IR indicating infrared‐active). In tetragonal ZrO2 (with space group <sup>15</sup> *D*4*<sup>h</sup>* /*P*42 /*nmc* and *Z* = 2), the phase has six Raman modes and three infrared modes [A1g(R) + 2B1g(R) + 3E<sup>g</sup> (R) + A2u(IR) + 2Eu(IR)]. For cubic ZrO2 (with space group <sup>5</sup> *Oh* /*Fm*3*m* and *Z* = 1), its vibrational

spectra have only one Raman and one infrared modes [F2g(R) + F1u(IR)].

The data in **Figure 1** indicate that it is apparent that there is a lack of signals of ZrO2 in highly damaged zircons. This shows that ZrO2 and SiO2 are not the final products of metamictiza‐ tion in zircon. Results from infrared spectroscopy of radiation‐damaged zircon [14, 47] also support this observation. As mentioned early, although decomposed zircons were commonly reported in lab‐treated samples, tetragonal ZrO2 was recorded in only one natural sample [8], with an unknown thermal history, among a large number of natural zircon samples with different degrees of damage analyzed in Refs. [14, 47, 48]. In order to explore and examine high‐temperature behavior of damaged zircon and the possible causes for decomposition in zircon, systematic works were carried out by different groups [47, 36, 61, 62]. These works show that thermal annealing of heavily damaged zircon at high‐temperature experiments may lead to the decomposition of metamict zircon into tetragonal ZrO2 and glassy SiO2 at 1200 K, and upon further heating tetragonal ZrO2 transforms to monoclinic ZrO2 near 1400 K. Undamaged and weakly damaged zircons are less likely to show decomposition during high‐temperature annealing. It was reported that the decomposition‐induced silica tended to evaporate on further heating [61, 63]. As naturally damaged samples might experience high‐ temperature processes, some reported decomposed metamict zircon could be due to natural thermal annealing prior to experiments. In contrast to high‐temperature annealing, the pres‐ ence of ZrO2 in some natural zircons might be due to the reaction of fluids with metamict zircons, because radiation damage may alternate the chemical stability of zircon. The obser‐ vation of ZrO2 in dissolution experiments was reported in highly damaged zircon samples [64]. Raman data of hydrothermally leached metamict zircon [65] showed the formation of monoclinic ZrO2 . In general, the decomposition of zircon into ZrO2 and SiO2 is related to radiation‐damaged zircons. Their crystal lattice is heavily damaged and has more defects. As a result, the durability of these heavily damaged samples is expected to be affected, and they are vulnerable to the impact of external physical and chemical conditions (e.g., water, solu‐ tions, high temperature, and even high pressure).

Raman data of metamict titanite (**Figure 3**) also show there is a lack of formation of oxides in highly damaged samples [40]. The findings are supported by X‐ray measurements and TEM [11] and infrared data [43]. The observation further shows that phase decomposition into oxides is not the final state of alpha‐decay damage in these materials, and the materials are safe to be used as nuclear waste forms. It has been found that decomposition of radia‐ tion‐damaged zircon is commonly related to high‐temperature heating (see a below section).

#### **4. Metamict state** *versus* **glass state of CaTiSiO<sup>5</sup> and ZrSiO<sup>4</sup>**

a high‐temperature neutron work by [53] suggested that zircon decomposes into crystalline

ions at around 950 K. As discussed by different works [56, 57], some of these previous works based on X‐ray diffraction measurements might have experienced difficulties in the determi‐

Vibrational (Raman and infrared) spectroscopy is a good analytical tool for resolving these

morphs at different temperatures. The room‐temperature phase is monoclinic, while the tetragonal and cubic phases occur at high temperatures [58]. The theoretical calculations [59, 60] have given the optical phonon modes (for zero wave vector) for each polymorph of zirco‐

has space group <sup>5</sup> *C*2*<sup>h</sup>* /*P*21

The data in **Figure 1** indicate that it is apparent that there is a lack of signals of ZrO2

(R) + 9B<sup>g</sup>

and *Z* = 2), the phase has six Raman modes and three infrared modes [A1g(R) + 2B1g(R) + 3E<sup>g</sup>

and SiO2

tion in zircon. Results from infrared spectroscopy of radiation‐damaged zircon [14, 47] also support this observation. As mentioned early, although decomposed zircons were commonly

[8], with an unknown thermal history, among a large number of natural zircon samples with different degrees of damage analyzed in Refs. [14, 47, 48]. In order to explore and examine high‐temperature behavior of damaged zircon and the possible causes for decomposition in zircon, systematic works were carried out by different groups [47, 36, 61, 62]. These works show that thermal annealing of heavily damaged zircon at high‐temperature experiments

Undamaged and weakly damaged zircons are less likely to show decomposition during high‐temperature annealing. It was reported that the decomposition‐induced silica tended to evaporate on further heating [61, 63]. As naturally damaged samples might experience high‐ temperature processes, some reported decomposed metamict zircon could be due to natural thermal annealing prior to experiments. In contrast to high‐temperature annealing, the pres‐

zircons, because radiation damage may alternate the chemical stability of zircon. The obser‐

[64]. Raman data of hydrothermally leached metamict zircon [65] showed the formation of

in some natural zircons might be due to the reaction of fluids with metamict

in dissolution experiments was reported in highly damaged zircon samples

into ZrO2

. An X‐ray powder diffraction study

when irradiating zircon with heavy

and SiO2, as pointed out by Ref. [30],

) have been well studied previously and

/*c* and *Z* = 4, and it has eighteen Raman

(with space group <sup>15</sup> *D*4*<sup>h</sup>* /*P*42

(R) + 8Au (IR) + 7Bu(IR)] (R indicating Raman‐

are not the final products of metamictiza‐

was recorded in only one natural sample

(with space group <sup>5</sup> *Oh* /*Fm*3*m* and *Z* = 1), its vibrational

transforms to monoclinic ZrO2

[54]. Meldrum et al. [55]

has three common poly‐

/*nmc*

in highly

and glassy SiO2

near 1400 K.

(R)

at

β‐cristobalite (rather than silica glass) and tetragonal ZrO2

observed a decomposition of zircon into tetragonal ZrO2

nation of cubic and tetragonal ZrO2

110 Raman Spectroscopy and Applications

. The monoclinic ZrO2

modes and fifteen infrared modes [9A<sup>g</sup>

+ A2u(IR) + 2Eu(IR)]. For cubic ZrO2

damaged zircons. This shows that ZrO2

reported in lab‐treated samples, tetragonal ZrO2

1200 K, and upon further heating tetragonal ZrO2

nia ZrO2

ence of ZrO2

vation of ZrO2

problems related to decomposition of ZrSiO4

because Raman and infrared spectra of zirconia (ZrO2

at high temperatures reported the appearance of pseudo‐cubic ZrO2

.

also because vibrational spectroscopy has short length scales. ZrO2

active and IR indicating infrared‐active). In tetragonal ZrO2

spectra have only one Raman and one infrared modes [F2g(R) + F1u(IR)].

may lead to the decomposition of metamict zircon into tetragonal ZrO2

One of the interesting issues related to radiation damage and metamictization is the similari‐ ties or differences between the amorphous phases produced by alpha‐decay radiation dam‐ age and those produced through thermally quenched glass melts. Previous studies [47] have led to unanswered questions and concerns: for example, What is the relationship between the structural characteristics of disordered materials and the type of irradiation or physical process used to produce them?

As a type of disordered or amorphous materials, metamict minerals were commonly con‐ sidered as "glass‐like" materials in early studies on naturally occurring radiation effect and metamictization, and the metamict state was referred as a glassy state, which is somehow similar to that obtained by rapid quenching of high‐temperature melts [30]. With experimen‐ tal data gathering, evidences have emerged that indicate their important differences between the aperiodic states. The dispersion depends on the physical processes, which produce the amorphous states. Vibrational spectroscopy (Raman and infrared) is very useful to study this issue because of its sensitivity to local structures.

Raman data of metamict titanite (CaTiSiO<sup>5</sup> ) and its glass analogue (with the same chemical compositions) produced by quenching melts show that the two types of materials have dif‐ ferent vibrational features (especially in the Ti**‐**O and Si**‐**O stretching regions) (**Figure 4**) [40], although they both are almost amorphous in terms of electron microscopy and X‐ray diffrac‐ tion analysis. The CaTiSiO<sup>5</sup> glasses produced from melts show a Si**‐**O stretching band near 827 cm−1, while metamict titanite has a relatively weak band in a higher wavenumber, 844 cm−1. What is more, the quenched melts of CaTiSiO<sup>5</sup> have a relatively strong band peaked near 709 cm−1, but this feature is almost absent in metamict titanite. The Raman observations are supported by the results for the same samples analyzed by infrared reflection and absorption spectroscopy [43]. These results suggest structural differences associated with Ti**‐**O and Si**‐**O bonds in the glasses and metamict phases. Interestingly, the glassy and metamict titanites also exhibit some similarities in a couple of band positions in the far infrared region (below 450 cm−1). For example, they both have bands around 170, 330 and 430 cm−1.

**Figure 4.** CaTiSiO<sup>5</sup> glasses produced by quenching titanite melts (thermal glass‐1 and thermal glass‐2 are prepared by quenching CaTiSiO<sup>5</sup> melts) show overall spectra patterns different from those of metamict titanite, especially in the region of 600 and 1100 cm–1 [Ref, [40], modified). This indicates the spectral dispersion related to Si‐O and Ti‐O vibrations in the two types of disordered materials. However, the two materials exhibit some local maximums with similar or close peak positions (indicated by dash lines) in the far infrared region (100 and 500 cm–1).

Zircon has a high melting temperature (above 2700–2800 K), and crystalline ZrO2 and a liq‐ uid of SiO2 may coexist for a composition of ZrSiO4 at a temperature region of ~1960 K and ~2670 K [66]. It is difficult to quench zircon melts without decomposition into ZrO2 and SiO2 ; however, glass‐like zircon (ZrSiO4 ) has been recorded with Raman spectroscopy in laser‐ treated zircon in a study with laser melting, near the boundaries between the unmolten and molten regions (where a relatively large temperature gradient could exist) [67]. The large tem‐ perature gradient is expected to increase the quench rate and facilitate a "freeze" of the local configuration of the ZrSiO4 melts before the decomposition takes place. These experimental results indicate that the metamict state is different from the glassy state obtained by quench‐ ing melts. Apparently, the processes and structural states associated with metamictization and irradiation amorphization are more complex than those in common thermal glasses. The formation of the metamict state involves not only amorphization, but also defect accumula‐ tion caused by alpha‐particle damage and further radiation or irradiation may lead to damage as well as recrystallization.

So far, although the issue remains under debates, there have been substantial evidences indicating characteristic discrepancies between two types of amorphous states (metamict and glass states) [68], for example: (i) the two types of materials commonly have spectral and structural discrepancies [40, 43, 69, 70]; (ii) high‐energy heavy ion irradiation may lead to significant modifications in local structures of glasses [71–74]; (iii) upon heating, radiation‐damaged minerals tend to recrystallize epitaxially and recover to their original cryptographic orientations [48, 51, 75], while during high‐temperature treatments, glasses commonly undergo a glass transition; and (iv) for common glasses, their glass transition tem‐ peratures are roughly defined, while various responses at different temperatures are seen in metamict minerals. For example, radiation‐induced defects in metamict zircon may be annealed or healed at temperatures as low as 600 K accompanied by changes in the oxidation state of U ions; partial decomposition of ZrSiO4 into SiO2 and SiO2 in heavily damaged zircon may take place at 1050 K; diffusion and conversions of hydrogen‐related species together with dehydroxylation may occur between 1200 and 1600 K (e.g., [16, 18, 52, 76, 77], see for the transition point [6]).
