Keywords

Introduction

Diamonds are unique samples for studying the secular evolution of the lithosphere and the asthenosphere (e.g. Shirey et al. 2002; Stachel et al. 2005). Inclusions encapsulated in diamonds provide evidence for conditions of diamond growth and represent the most reliable technique of dating diamonds. Dating of silicate inclusions requires generally multiple inclusion composites from individual to several diamonds (e.g. Richardson et al. 1984, 1990). Sulphide inclusions and Re–Os isotopes enhance the possibility of dating individual (e.g. Pearson et al. 1998; Pearson and Shirey 1999a, b; Richardson et al. 2001; Shirey and Richardson 2011). Diamond growth could occur during discrete growth episodes (Rudnick et al. 1993; Taylor et al. 1995), and dating diamonds is not an easy task. Diamond cannot be directly dated using radiometric methods, and the lack of inclusions in diamonds means that only infrared absorption data on the diamond-hosted nitrogen impurities can provide any age constraints (Evans and Qi 1982; Evans and Harris 1989; Boyd et al. 1994, 1995; Mendelssohn and Milledge 1995; Taylor et al. 1996a, b). The aggregation of nitrogen in diamond is sensitive to both temperature and time and could be used as an indication of the diamond residence time in the mantle (e.g. Evans and Harris 1989; Mendelssohn and Milledge 1995).

Silicate inclusion-bearing diamonds at Finsch (Appleyard et al. 2004) and Premier (Chinn et al. 2003) display complex internal growth patterns in cathodoluminescence and distinct infrared absorption characteristics (FTIR) within individual diamonds. These observations suggest that individual diamonds have experienced slight changes in temperature conditions during prolonged diamond formation or multiple diamond growth episodes over time. Complex diamond growth histories indicative of changes in a growth environment within the mantle have been previously stated from cathodoluminescence images, FTIR characteristics and multiple inclusion chemistry in single diamond (e.g. Bulanova 1995; Sobolev and Yefimova 1998; Taylor and Anand 2004; Taylor et al. 1996a, b, 2000; Spetsius and Taylor 2008). Distinct diamond events of growth, resorption and regrowth separated by billions of years have been suggested in Udachnaya diamonds based on Pb and Re–Os isotopes (Rudnick et al. 1993; Bulanova et al. 1998; Pearson et al. 1999b; Wiggers de Vries et al. 2008).

In order to evaluate the complexity of diamond genesis, we have modelled the infrared characteristics of 21 diamonds from Finsch (South Africa), 1 from Udachnaya (Siberia) and 18 from Premier (South Africa).

Geological Settings and Samples

The Finsch kimberlite is located 37 km east of Postmasburg and 160 km north-west of Kimberley, Cape Province, Republic of South Africa. The kimberlite is of Group II variety with an estimated Rb–Sr emplacement age of 118 Ma (Smith et al. 1985). The geology and petrology of the kimberlite have been described in Ruotsala (1975), Clement (1975) and Skinner et al. (1979). The inclusion suite of diamonds at Finsch is predominantly peridotitic (Harris and Gurney 1979). The genesis ages for the peridotitic-type (P) and eclogitic-type (E) diamonds are quite distinct, yielding a CHUR-based model age of 3.19 ± 0.05 Ga (Richardson et al. 1984) for the former and an isochron age of 1.58 ± 0.05 Ga for the latter (Richardson et al. 1990). Diamonds examined in this study all contained sulphide inclusions. All sulphide inclusions in the diamonds except F215 and F218 have high (>44 wt %) Fe, low-to-moderate (<13 wt %) Ni contents and so may be classified as E-type according to the sulphide Ni contents (Yefimova et al. 1983). Diamonds F215 and F218 are therefore classified as P-type which is consistent with a single purple P-type garnet inclusion recovered in the F215. Cathodoluminescence (CL) mapping reveals internal growth structures characterized by different δ13C and N-abundances (Palot et al. 2013). However, core–rim relationships cannot be derived since the analyses were conducted on diamond fragments. Some diamond fragments exhibit complex growth pattern with non-regular octahedral growth patterns (Fig. 1). Appleyard et al. (2004) also identified CL distinct periods of growth within individual diamonds, resulting in both cubic and octahedral growth layers.

Fig. 1
figure 1

Cathodoluminescence images of Finsch E-type diamond fragment F213-2 which exhibits non-regular octahedral growth patterns

Full size image

The Udachnaya kimberlite (Siberia) with an age of 350 Ma (Kinny et al. 1997) lies within the central Siberian platform (Sobolev and Nixon 1987). A single diamond examined in this study (diamond plate 3648) has been described in detail by Rudnick et al. (1993), Taylor et al. (1995) and Pearson et al. (1999b). The diamond central, cubo-octahedral zone contains a cluster of sulphide inclusions, a wustite inclusion and an Mg-rich olivine. The intermediate zone is octahedral, truncated by the rim zone indicative of a resorption event. The sulphide inclusions located in the rim zone are surrounded by cracks, one of which is re-healed. The high Ni and Os contents of the sulphide inclusions are consistent with the presence of olivine and therefore the P-type paragenesis of the diamond (Pearson et al. 1998).

The Premier kimberlite is located 30 km east–northeast of Pretoria in Republic of South Africa. The kimberlite is Precambrian in age (Allsopp et al. 1967; Barrett and Allsopp 1973). For a detailed review of its geology and petrology, see for example Scott and Skinner (1979). Samples examined here comprised one peridotitic (diamond AP28) and seventeen eclogitic silicate inclusion-bearing diamonds (Chinn et al. 2003). Most of diamonds exhibit octahedral zonations with indications of slight internal resorption (Chinn et al. 2003). E-type diamond AP28 is much more resorbed than the majority of other E-type diamonds at Premier and presents deformation lamellae features. Chinn et al. (2003) reported the complex diamond growth history of one E-type diamond AP33 studied in double-polished central plate (not analysed in this study). This sample comprises a well-defined core region characterized by high nitrogen aggregation state, which is surrounded by a low aggregation state rim overgrowth.

Analytical Technique

Nitrogen abundances and nitrogen aggregation states, the latter measured as a percentage of nitrogen present in the fully aggregated B-defect, were determined in the De Beers Laboratory of Diamond Research at the University of Alberta, using a Thermo Nicolet Nexus 470 FT-IR spectrometer (bench) fitted with a continuum infrared microscope. Finsch diamond fragments were 1.5–4.5 mm in diameter and mainly were bounded by (111) cleavage planes. Spectra (650–4,000 cm−1) were acquired in transmission mode for 200 s with a resolution of 4 cm−1 and an aperture size of 100 μm. Nitrogen concentrations and aggregation states were calculated after spectral decomposition using the Excel program CAXBD97, developed by David Fischer (the Diamond Trading Company, Maidenhead, UK). The absorption coefficients for the A and B centres at 1,282 cm−1 were 16.5 and 79.4 at.ppm/cm, respectively (Boyd et al. 1994, 1995). Detection limits and errors typically range from 5 to 20 ppm and about 10 %, respectively. The error of the aggregation state of nitrogen is estimated to be better than ±5 % (2σ).

Udachnaya sample 3648 and the Premier diamonds were diamond plates, polished mainly in (100) and (110) crystallographic directions. The Udachnaya diamond plate is approximately 7 mm long, and Premier plates range in weight between 60 and 200 mg. Data for 3648 were acquired using the equipment and techniques outlined by Mendelssohn and Milledge (1995). For the Premier samples, spectra were acquired with a Nicolet Magna 760 IR spectrometer, over the infrared range from 4,000 to 650 cm−1. Data were processed using the same parameters and method described for the Finsch samples. The Premier data have been described briefly in abstract form by Chinn et al. (2003).

Results: Nitrogen Abundances and Aggregation States

Nitrogen abundances of Finsch samples were determined on 53 diamond fragments from 21 whole diamonds broken to release the inclusions. Overall, the nitrogen content in the studied diamonds ranges from 21 to 1,093 at.ppm (mean 569 ± 271 at.ppm, 1σ) (Table 1). Internal heterogeneity in N-abundance within single diamond ranges from 5 to 336 at.ppm (mean 141 ± 109 at.ppm, 1σ) (Figs. 2 and 3a). Since the analyses were conducted on diamond fragments, core–rim relationships generally cannot be derived. Overall nitrogen aggregation states (expressed as % of B-defects) vary from 0 to 83 % (mean 31 ± 21 %, 1σ) (Table 1). Internal variations of %B are from 2 to 60 % (mean 14 ± 14 %, 1σ) (Figs. 2 and 3a). Diamonds F203, F211, F213 and F215 display the largest variations in N-abundance and aggregation state (Figs. 2 and 3a).

Table 1 Finsch diamond measurements of nitrogen abundance and aggregation state by infrared spectroscopy
Full size table
Fig. 2
figure 2

Internal variation in nitrogen abundance (atomic.ppm) and aggregation state (% of B-defect) within individual diamonds from Premier (square), Udachnaya (circle) and Finsch (diamond). Internal variation represents the largest measured difference in N-abundance and %B for each diamond

Full size image
Fig. 3
figure 3

Nitrogen abundance and aggregation state (%B) of diamonds from a Finsch, b Udachnaya and c Premier. Isotherms are calculated according to the second-order kinetic law for nitrogen diffusion (Chrenko et al. 1977) assuming an averaged mantle residence time of 1.4, 3.0 Ga and 70 Ma, respectively. Total errors of N-abundance and aggregation state are of 15 % and better than ±3 %, respectively. Errors bars are not represented for Premier diamonds for clarity

Full size image

Nitrogen abundances of Udachnaya diamond 3648 range from 10 to 482 at.ppm (mean 194 ± 143 at.ppm, 1σ) (Table 2). Nitrogen content progressively decreases from the core to the rim. Nitrogen content in the inner core ranges from 482 to 311 at.ppm (mean 376 ± 44 at.ppm), in the outer core from 289 to 135 at.ppm (mean 230 ± 47 at.ppm, 1σ), in the intermediate zone from 113 to 37 at.ppm (mean 57 ± 27 at.ppm, 1σ) and in the rim from 94 to 10 at.ppm (mean 39 ± 25 at.ppm, 1σ) (Fig. 3b). Aggregation states vary from 14 to 86 % (mean of 31 ± 21 %, 1σ). Diamond 3648 also displays a decrease of  %B from the core to the rim. The inner core varies from 86 to 69 % (mean 74 ± 4 %, 1σ), the outer core from 69 to 50 % (mean 62 ± 6 %, 1σ), the intermediate zone from 44 to 27 % (mean 35 ± 6 %, 1σ) and the rim from 44 to 14 % (mean 24 ± 8 %, 1σ) (Fig. 3b). The maximum variation in N-abundance and aggregation state between the inner core and the rim in diamond 3648 is 472 at.ppm and 72 %B, respectively (Fig. 2).

Table 2 Udachnaya diamond 3648 measurements of nitrogen abundance and aggregation state from core to rim by infrared spectroscopy
Full size table

Overall, N-abundances in diamonds from Premier range from 21 to 802 at.ppm (mean 408 ± 166 at.ppm, 1σ) (Chinn et al. 2003, Fig. 3c). Internal variation in N-abundance within individual diamonds ranges from 93 to 654 at.ppm (mean of 320 ± 142 at.ppm, 1σ) (Fig. 2). Some diamonds exhibit spatial relationships in N-characteristics (Chinn et al. 2003). Nitrogen aggregation states vary from 2 to 65 % (mean 28 ± 11 %, 1σ) (Chinn et al. 2003, Fig. 3c). Internal variations of %B are from 7 to 62 % (mean of 24 ± 15 %, 1σ) (Fig. 2). Diamonds AP28, AP31, AP 41, AP45, AP49 and AP50 display the largest variations in N-abundance and %B (Figs. 2 and 3b, Table 3).

Table 3 Premier diamond measurements of nitrogen abundance and aggregation state by infrared spectroscopy
Full size table

Discussion

Nitrogen Geothermometry

The most common atomic impurity in diamond is nitrogen (Kaiser and Bond 1959), and based on this element, diamonds have been classified as Type I, if nitrogen is present, and Type II, if it is absent (<20 ppm detection level). According to micro-FTIR studies of inclusion-bearing diamonds, 70 % is classified as Type I (Deines et al. 1989; Cartigny 2005; Stachel 2007; Stachel et al. 2009). The different nitrogen-bearing defects in diamonds are linked by a diffusion process. The initially single nitrogen atoms (i.e. the C-defect or diamond Type Ib) diffuse fairly rapidly over millions of years due to a low activation energy (Taylor et al. 1996a, b), to form pairs of nitrogen atoms (i.e. the A-defect or diamond Type IaA). Subsequently, the latter diffuses over billions of years, due to a higher activation energy (Evans and Qi 1982; Evans and Harris 1989), to form aggregates of nitrogen atoms tetrahedrally arranged around a vacancy (i.e. the B-defect or diamond Type IaB). The conversions both from Type Ib to Type IaA and then from Type IaA to Type IaB follow second-order kinetics. In the more important second nitrogen sequence, nitrogen abundance ([N] = [A-defect] + [B-defect]), the average residence time (t) and the storage temperature (T) are all linked by the following formula (Chrenko et al. 1977):

$$ \frac{ 1}{{\left[ {{A} - {\text{defect}}} \right]}} - \frac{ 1}{\left[ N \right]} = \left( {A \times { \exp }^{{ - \, \frac{\text{Ea}}{\text{RT}}}} } \right) \times t $$

N-abundance and A-defect are in atomic ppm of nitrogen, t is in seconds and T is in Kelvin, A and R are constants with A = 294,000 ppm−1·s−1 and R = 8.317 J·mol−1·K−1, Ea = 675,510 J·mol−1 is the activation energy (Cooper 1990). Because mantle storage of diamonds is generally over hundreds of Ma, nitrogen aggregation depends chiefly on its time-integrated storage temperature. Nitrogen aggregation is therefore a valuable geothermometer (e.g. Evans and Harris 1989; Palot et al. 2009), which will be used in the next section in order to estimate mantle storage temperatures for the studied diamonds.

Multiple Growth Events for Finsch Sulphide Inclusion-Bearing Diamonds Inferred from Nitrogen Characteristics

In order to investigate the diamond growth histories, we have estimated their probable temperature ranges of formation. To do this calculation, we assume an average mantle residence time of ~1.4 Ga for the Finsch eclogitic sulphide inclusion-bearing diamonds (based on the genesis isochron age of 1.58 Ga for E-type silicate inclusions from Richardson et al. 1990, less than that of the eruption age of 118 Ma from Smith et al. 1985). The estimated mantle storage temperatures for the Finsch diamond fragments range from 1,068 to 1,174 °C with a mean of 1,128 ± 25 °C (1σ) (Fig. 3a). It is noticeable that changes in mantle residence time of the diamonds would affect the absolute storage temperature estimates, but do not alter our conclusions based on relative temperatures between different portions of diamond. Although Finsch diamonds are commonly thought to originate from a relatively homogenous carbon source (Deines et al. 1984), that conclusion was largely based on P-type diamond study. Heterogeneities in N-abundance and aggregation states in some Finsch E-type sulphide inclusion-bearing diamonds suggest that these samples experienced a relatively complex growth and mantle residence history as the data do not plot on a single isotherm (Fig. 3a). For example, diamonds F203, F211, F213, F215 and F218 show internal variations in their integrated mantle residence temperatures of up to ~50 °C (Fig. 3a). These variations reflect either slight changes in temperature conditions during prolonged diamond formation or distinct residence histories for different portions of the diamond, that is, multiple episodes of diamond growth extending over millions to billions of years (e.g. Appleyard et al. 2004; Taylor and Anand 2004). A steady increase in temperatures from the core to the rim is physically unacceptable, since the core must have experienced the same heating event. Therefore, unless these diamonds grew continuously over the course of millions of years while associated with a significant drop in ambient temperature, it is likely that multiple growth episodes occurred at different times. Assuming that the different portions in a single diamond that give the largest heterogeneity in N-abundance and aggregation (Fig. 2) represent distinct growth steps, we modelled the age differences of different portions within the single diamond using the equation of Chrenko et al. (1977) linking temperature, nitrogen abundance and average time residence. We assumed a maximum storage temperature of 1,174 °C based on maximum nitrogen temperature, consistent with the maximum inclusion-based temperature of 1,179 °C reported for E-type silicate-bearing Finsch diamonds (Appleyard et al. 2004). Uncertainties on nitrogen temperatures are of ±12 °C and are based on uncertainties of N-abundance and %B, leading to large uncertainties in time estimates. Our calculations indicate that the time between distinct growth steps within individual diamonds varies from minimum estimates (mantle residence at 1,174 °C) of \( \sim 3_{ - 3}^{ + 88} \) Ma (diamond F208) to 942 ± 368 Ma (diamond F215), to maximum estimates (residence at 1,000–1,140 °C at 4.5 Ga) that are equivalent to the age of the Earth (Fig. 4a). These calculations may be underestimated because the random orientation of fragments may be mixtures of zones with distinct FTIR characteristics. Because of the large uncertainties for some diamonds (F204, F205, F207, F208, F209, F212, F214, F219 and F220, Table 4), it cannot be ruled out that these samples may have experienced a single growth event at relatively constant temperature. The majority of our calculations are, however, compatible with multiple growth events during the formation of sulphide inclusion-bearing diamonds at Finsch. This implies that isochron ages may be flawed as there appears to be no single formation age for a single diamond. This is consistent with distinct carbon isotopic compositions and nitrogen contents observed across some of these diamonds (Palot et al. 2013).

Fig. 4
figure 4

Maximum relative times between distinct growth steps against averaged storage temperatures within individual diamonds from a Finsch, b Udachnaya and c Premier. The maximum averaged storage temperature for a given locality is based on the highest nitrogen (and inclusion)-based temperature estimate reported for each population. The age of the Earth is reported as the upper limit for time difference between distinct diamond portions. Errors bars (dotted lines) are only reported for Udachnaya diamond for clarity

Full size image
Table 4 Relative errors (± %, 2σ) on temperature estimates for Finsch, Udachnaya and Premier diamonds at temperatures from 1,000 to 1,300 °C (50 °C increments)
Full size table

Multiple Growth Events for Udachnaya Peridotitic Sulphide Inclusion-Bearing Diamond Inferred from Nitrogen Characteristics

We assume an average mantle residence time of ~3.0 Ga for the Udachnaya diamond 3648 (based on the model age of 3.1–3.5 Ga on sulphide inclusions in this diamond by Pearson et al. 1999b). The age of inclusions is significantly older than that of the eruption age of 361 Ma (Kinny et al. 1997). The estimated mantle storage temperatures range from 1,137 to 1,200 °C with a mean of 1,166 ± 12 °C (1σ) (Fig. 3b), which are within the range of inclusion-based temperatures of Udachnaya diamonds (e.g. Griffin et al. 1993). The inner and outer cores display a constant estimated storage temperature of 1,163 ± 3 °C (1σ) with data plotting close to the theoretical isotherms at constant time (Fig. 3b). In contrast, the intermediate and rim zones display more scattered time-averaged residence temperatures, with a slightly higher mean of 1,170 ± 17 °C (1σ) (Fig. 3b). A steady increase in integrated temperature from the core to the rim within mantle diamonds is physically implausible, since the former and the latter must have experienced the same temperature during any heating event, and diamonds conduct heat very efficiently. We therefore explore the possibility of multiple growth episodes at different times suggested by Rudnick et al. (1993) based on large variations in Pb isotopic compositions of sulphides included in this diamond and by Bulanova et al. (1996) and Pearson et al. (1999b) based on enriched Os, Re, Pb and Zn and more radiogenic Os and Pb isotopes in one sulphide inclusion in the rim of diamond 3648. Calculations indicate that the maximum relative time between the core and the rim varies from a minimum estimate (mantle residence at 1,200 °C) of ~2,387 ± 931 Ma to maximum estimate (residence at 1,186 °C) that is equivalent to the age of the Earth (Fig. 4b). The minimum estimate equals the estimate of 2 Ga as the age difference between different growth zones of this diamond determined by Pb-isotope analysis of sulphide inclusions by Rudnick et al. (1993).

Multiple Growth Events for Premier Silicate Inclusion-Bearing Diamonds Inferred from Nitrogen Characteristics

Eclogitic silicate inclusion-bearing diamonds at Premier display ages from 1,150 ± 60 to 1,254 Ma (Richardson 1986; Burgess et al. 1989; Phillips et al. 1989) compared to a pipe eruption age of 1,180 ± 30 Ma (Allsopp et al. 1989). We assume a maximum average mantle residence time of ~70 Ma for these diamonds. The estimated mantle storage temperatures range from 1,127 to 1,286 °C with a mean of 1,209 ± 23 °C (1σ) (Fig. 3c) and are slightly lower than estimates from inclusions in Premier diamonds (between 1,199 and 1,363 °C, Gurney et al. 1985). The large variations in N-abundance and aggregation states in some Premier diamonds suggest that these samples also have experienced complex growth histories. Variation in time-integrated storage temperature within individual diamonds ranges from 10 °C (diamond AP26) to 79 °C (diamond AP28) (Fig. 3c). These variations, which are similar to Finsch and Udachnaya diamonds, are interpreted as reflecting multiple episodes of diamond growth separated in time, rather than steady increases or decreases in storage temperature. Calculations indicate that the maximum time between distinct growth steps varies from minimum estimates (mantle residence at 1,286 °C) of 1.5 ± 0.5 Ma (diamond AP37) to 76 ± 23 Ma (diamond AP31), to maximum estimates (residence at 1,076–1,170 °C) that are equivalent to the age of the Earth (Fig. 4c). Marked variation in FTIR data suggests distinct and multiple growth episodes for E-type diamonds at Premier with ages minimal 1 Ma apart, in agreement with some conclusions of Chinn et al. (2003). A single diamond growth event cannot be ruled out for diamonds AP25, AP26 and AP46 due to large uncertainties (Table 4).

Overall, this study therefore suggests multiple episodes of diamond growth based on distinct infrared spectroscopy characteristics within individual macro-diamonds from different localities and parageneses. Similar evidence was found previously for eclogitic diamonds from Siberian kimberlites (Bulanova et al. 1998 and 2012) and by Wiggers de Vries et al. (2008, 2011) in a detailed study of sulphide-bearing diamond plates from Udachnaya, Mir and Mwadui kimberlites. The study by Wiggers de Vries et al. (2008) also revealed complex Re–Os isotope systematics in sulphides derived from different diamond growth zones, clearly indicating the complexity present in single diamond. FTIR study is therefore a valuable diagnostic tool in characterizing samples for radiometric dating. In addition, the technique is an excellent complement to classical CL images in selecting diamonds for in situ carbon isotopic profiles, allowing the ability to determine the likelihood of multiple growth events. Measurements for such work are best taken in situ on polished diamond plates, as for the Udachnaya and Premier diamonds, to preserve the relationships between the data points, using the CL image as a map of the growth history. In this way, any changes in crystal growth or periods of resorption can be integrated with the data measurements for interpretation.

Conclusions

The examination of 21 sulphide inclusion-bearing diamonds from Finsch, 1 sulphide inclusion-bearing diamond from Udachnaya and 18 silicate inclusion-bearing diamonds from Premier revealed distinct N-abundances and aggregation states of nitrogen within individual samples. These variations more likely reflect multiple growth episodes of diamond at different times rather than steady changes in temperature conditions during prolonged diamond formation. Modelling of these heterogeneities indicates that some diamonds have experienced distinct growth episodes over extended time periods, implying that great care is needed in the evaluation of isochron and model ages in radiometric dating studies, as there may be multiple phases of diamond growth in single diamond.

The diamonds investigated in this study were derived from different localities and parageneses, suggesting that such complex growth histories are not rare. Infrared characteristics are a critical tool for constraining diamond growth histories and for selecting diamonds for dating.