Defect evolution and color center generation of high energy density femtosecond laser processed IIa diamond

Diamond has unique optical and electronic properties, making it an ideal candidate for quantum information processing and sensor applications. By manipulating the electronic spin states of nitrogen-vacancy (NV) centers, one can achieve high-sensitivity magnetic-field, temperature, and stress sensing,¹ along with the creation of quantum bits for quantum computing and communication purposes;² however, the preparation methods for NV centers often require complex processing steps and prolonged high-temperature heat treatment, limiting their widespread application in practice.

Femtosecond laser processing of diamond can achieve high-precision, non-contact, localized, and controllable modifications, providing an effective way to create diamond materials with specific structures and properties. Femtosecond lasers have extremely short pulse widths, typically between a few femtoseconds and hundreds of femtoseconds, meaning that processing is very fast and can be completed on a nanosecond or sub-nanosecond timescale.³ Femtosecond laser irradiation allows the formation of a dense, cold electron–hole plasma in the focal area while minimizing overheating in the surrounding area. This ultrafast processing can reduce the formation of heat-affected zones, thereby reducing thermal damage and material deformation.⁴ Femtosecond laser processing involves extremely high energy densities and can generate light intensities of up to several TW/cm² on the material surface, thereby achieving high-precision processing and the preparation of micro- and nanoscale structures.^5,6 Combining femtosecond laser direct-writing technology with diamond materials to form NV color centers on the diamond surface has become an increasingly attractive method;⁷ local material modifications and defect introduction can be achieved, thereby directly forming NV color centers in the diamond lattice without going through the traditional complicated preparation process.

During femtosecond laser irradiation, potential laser-induced phenomena include the generation of color centers, changes in refractive index caused by densification and defect formation, the formation of microvoids due to resolidification and shock waves, and the occurrence of cracks induced by destructive breakdown. In this study Raman and fluorescence imaging techniques were used to explore the damage mechanisms and spatial distributions of defects in diamond materials under high-power-density femtosecond laser pulses. Additionally, a method for the formation of low-nitrogen IIa diamond NV centers under non-annealing conditions was investigated.

A type IIa high-purity diamond sample was obtained from Element Six; this was synthesized using chemical vapor deposition (CVD) with an unintentional defect concentration of less than 5 ppb. A femtosecond laser with a wavelength of 1030 nm, a repetition frequency of 100 kHz, a maximum laser-pulse energy of 10 µJ, and a pulse width of 285 fs was used to process the diamond at room temperature. The number of pulses applied to the diamond surface was controlled by a computer. The laser beam distribution was Gaussian, and the focusing objective had a numerical aperture of 0.42 and a spot diameter of about 2 µm. The sample was placed on a motorized triaxial platform, and a distance of 20 µm was maintained between adjacent processing points to avoid overlap of the pulse energies from two neighboring laser pulses. The samples were cleaned with piranha solution before and after the experiment to remove surface impurities and processing residues.

A HORIBA LabRAM Odyssey spectrometer was used for fluorescence and Raman spectroscopy characterization. A laser with a wavelength of 532 nm was used as the excitation source, and a microscope objective with a numerical aperture of 0.95 was used for focusing. The confocal pinhole size was set to 50 µm. Scattered signals on an 1800 gr/mm grating were collected using a CCD camera. For fluorescence spectroscopy, lasers with wavelengths of 405 and 532 nm were used as excitation sources, and the sample temperature was stabilized at 77 K using a THMS600 (Linkam Scientific) heating/cooling stage. Fluorescence lifetime testing was performed using a laser confocal fluorescence lifetime imaging system (ISS Inc.). The machined surface was observed with a scanning electron microscope (SEM; FEI NanoSEM 430).

During the interaction between a femtosecond laser and diamond, the threshold theory^8,9 suggests that the deviation in the ablation threshold for femtosecond laser pulses is minimal. The ablation threshold represents the lowest laser energy density at which the laser can damage the material.^10,11 When femtosecond laser pulses interact with diamond, its surface undergoes ablation, leading to a phase change. The minimum power required for the femtosecond laser during the processing must exceed the ablation threshold. For IIa diamond, the ablation threshold is approximately 4 J/cm².¹² In this experiment, we selected a processing energy of 1.2 µJ (40 J/cm²) per pulse, which significantly exceeds the ablation threshold of diamond. We conducted processing experiments with different numbers of pulses: 10, 10³, 10⁴, 10⁵, 10⁶, 5 × 10⁶, and 10⁷.

An SEM was used to characterize the machined surface, and the results are shown in Fig. 1. It can be seen that after ten pulses, the femtosecond laser creates a shallow pit on the surface, and this is accompanied by numerous periodic nanofringe structures in the surrounding area. These fringes align perpendicular to both the direction of laser polarization and the fringe direction. The excitation of a large number of free electrons by the femtosecond laser leads to the formation of a plasma layer on the material surface. Subsequently, the collective oscillation of the surface plasma induced by the femtosecond laser generates surface plasmon polaritons (SPPs). The excitation of SPPs results in the periodic distribution of the laser field and the deposition of energy in the free electrons. The heating, melting, and even ablation of the crystal lattice occur due to electron–phonon coupling, leading to the formation of laser-induced periodic surface structures.¹³ As the number of pulses is increased to 1000, a heat-affected zone becomes noticeable around the processing area, indicating the presence of significant thermal effects. With further increases in the number of pulses, the depth of the ablation pit gradually increases, and processing debris is ejected from the bottom, with some being redistributed around the surface and accumulating in the ablation-pit area. However, when the number of pulses reaches 10⁶, all internal debris is expelled by the shock wave, and the surrounding recast layer is removed by the impact. When the number of pulses reaches a sufficiently high value, the topography of the crater becomes stable, and the laser no longer directly interacts with the material. Instead, plasma is generated in the air, continuously impacting the material. At this stage, the region surrounding the processing area is also cleaned.

Raman spectroscopy was conducted at a temperature of 77 K to analyze the processing area under various parameters, and the results are illustrated in Fig. 2. The pristine diamond exhibits a characteristic Raman peak at 1332 cm⁻¹. In the regions irradiated by the femtosecond laser, in addition to the diamond Raman peak, two narrow peaks appear at 1550 and 2325 cm⁻¹. Raman spectra of disordered carbon (graphite) typically display two distinct features: a G peak at 1580 cm⁻¹ and a D peak at 1350 cm⁻¹. The G peak is the sole marker peak for the Raman spectrum of highly oriented pyrolytic graphite within the 2000 cm⁻¹ range.¹⁴ The D peak is usually associated with structural disordering.^15,16 Simultaneous observation of Raman peaks corresponding to sp³- and sp²-hybridized carbon indicates an incomplete phase transition from diamond to graphite. It is worth noting that the Raman cross-section values for sp³ and sp² carbon differ, resulting in a significantly lower signal strength for graphite compared to the diamond signal peak.¹⁷ The peak of 2325 cm⁻¹ may be a vibrational pattern associated with nitrogen atoms. This was attributed to the presence of a small amount of sp¹-bonded nitrogen and carbon (C=N and C=C bonds) within the amorphous structure with N=C bond.^18,19 This feature was not present in the pristine sample.

The elemental compositions of the unprocessed and 10⁵-pulse processed regions were analyzed using energy dispersive x-ray spectroscopy (EDS), and the results are shown in Fig. 3. After continuous bombardment with a high-energy-density femtosecond laser, two peaks appeared in the EDS spectrum, at 0.392 and 0.523 eV, and these originated from the elements N and O, respectively. The contents of these elements are significantly higher in the processed region than in the unprocessed region. The deposition of high-energy pulses oxidized a part of the material, and the N atoms may have aggregated, generating C–N-related defects, as evidenced by the Raman test. The increase in N content may have also facilitated the generation of NV color centers.

Raman mapping was used to characterize the cross section of the processing area, and the longitudinal strength obtained is depicted in Fig. 4. When the number of pulses is small, laser–material coupling results in an isotropic expansion effect,²⁰ generating transverse and longitudinal damage areas of equal size within the processing region. Upon reaching 1000 pulses, the transverse damage within the processing area stabilizes; however, the longitudinal damage gradually expands as the number of pulses increases, with the longitudinal diffusion distance being proportional to the logarithm of the number of pulses.

Figure 5 presents statistical results indicating that the transverse diffusion distance of the damaged area progressively slows down after 1000 pulses, while the highly damaged area (shown in blue) expands slowly under the continuous influence of the laser pulses. This phenomenon occurs because once the material is removed, the laser pulse no longer directly interacts with the material, and the generated energy is transferred to the material through shock waves. The anisotropic expansion effect of air-plasma excitation²⁰ causes the propagation to extend further longitudinally than transversely.

Structures formed by previous pulses can interact with subsequent pulses, altering the intensity distribution and influencing the ablation process.^21–23 The enhancement of plasma and shock-wave expansion during femtosecond ablation increases with pulse action.²⁴ This enhanced laser–material coupling increases the energy deposition in the plasma, resulting in the observed stronger plasma and shock-wave expansion. This phenomenon can be theoretically explained by laser-induced air breakdown, which is intensified by pothole-induced laser refocusing. As the number of pulses increases, the morphology of the laser-induced craters becomes rougher. This roughening leads to a loss of laser refocusing capability, significantly altering the properties of the laser-induced plasma and shock-wave expansion. The dominant mechanism is determined by the morphology of the laser-induced crater, and this in turn is highly influenced by the number of pulses.

During femtosecond laser processing, the ablation process occurs through the interaction between the femtosecond laser and the material being processed, starting with the internal electrons of the material.^25,26 Graphitization occurs when internal free electrons absorb energy through inverse bremsstrahlung. As a result of absorbing a significant amount of energy, the sp³ atoms in the diamond lattice undergo a transition and transform into the sp² bonding state.²⁷ In addition to material removal, the graphitization phase transition leads to an increase in the distance between adjacent carbon atoms, a decrease in the density of the phase-transition region, and a change in the material’s properties. At the initial stage of laser-pulse action, the formation of the graphite layer affects the absorption of subsequent laser pulses, influencing the efficiency and behavior of the ablation process as the pulses continue.

The Raman imaging in Fig. 6 shows the distribution of graphite-like layer regions. After ten pulses, a uniform graphite-like surface layer emerges in the processing area. The graphite-like surface layer produced in the diamond ablation area dramatically increases the optical absorption of the diamond. The absorption coefficient of graphitized diamond is approximately 10⁵ cm⁻¹, which is close to the absorption coefficient of a metal surface.²⁸ This transformation not only alters the optical and electrical characteristics but also affects the chemical properties of the surface.²⁹ The micro- and nanostructured pits can enhance laser–material coupling through anti-reflection and incubation effects. As the pulse count increases, the graphitized region begins to diffuse, with a notably shorter longitudinal diffusion distance compared to the horizontal spread.

When the number of pulses is small, the polar plot does not differ much from that obtain from the unprocessed region. When the number of pulses reaches 1000, the torsion of the material begins to increase. Although the material has already transitioned to the graphite phase during this process, most of the processed area retains the diamond structure; however, when processing a material with a femtosecond laser, increasing the number of pulses significantly reduces the material’s threshold for laser interaction. Upon reaching a certain number of pulses, the material’s absorption is significantly enhanced, accumulating a large number of free electrons in an instant, leading to a sharp change in the material. This process is not gradual but exhibits a distinct threshold effect.

As the number of pulses increases by an order of magnitude, the interaction between the femtosecond laser and the diamond material dramatically intensifies, resulting in a large-scale phase transition in the processed region. The symmetry of the structure undergoes significant changes, decreasing and beginning to exhibit isotropic characteristics. With further increases in the number of pulses, the structural changes in the ablated area manifest in the degree of torsion of the material, as the action of the femtosecond laser further increases the torsion angle in the processed area, reaching a maximum of 10°. This results in significant internal stresses in the machined area.

The dense accumulation of vacancies and interstitial atoms triggered by the explosion leads to material densification, creating periodically arranged stress layers within the processing area. As a transition occurs from diamond to graphite, local material expansion induces radial compressive stress and tangential biaxial tensile stress around the diamond. The stress in the processing region can be characterized by changes in the Raman frequency shift. The principle is to provide a quantitative description of stress–strain behavior based on the material’s constitutive relationship, assuming it follows the generalized Hooke’s law during elastic deformation. By determining the characteristic peak frequency shift from the measured Raman spectrum, based on incident and scattered light and the vibrational modes of optical phonons, the relationship between the stress and the Raman-peak frequency shift can be established. Quantitative analysis of strain/stress is achieved by detecting these frequency-shift changes.³² 

After femtosecond laser processing, amorphous carbon and a large number of defects are produced, primarily vacancies and vacancy complexes. To investigate the occurrence of internal defects in diamond, the low-temperature photoluminescence spectra of diamond samples processed with between 10 and 10⁷ pulses were characterized. As shown in Fig. 9, zero-phonon lines (ZPLs) were found at 470 and 741 nm, corresponding to TR12 and GR1 defects in the diamond, respectively. The structure of the TR12 defect is not well defined, and its central part contains a carbon atom bonded in six different directions.³⁴ The signals at 575 and 637 nm correspond to NV⁰ and NV⁻, respectively, indicating uncharged and negatively charged NV color centers, respectively. The graphitized regions produced by the femtosecond laser contain a large number of vacancy defects, which subsequently diffuse outward into the diamond lattice as high-energy laser pulses are deposited, leading to the formation of NV complexes.

Impurity nitrogen in diamond is necessary, and its concentration must be considered to obtain an ideal defect. The higher the nitrogen concentration, the higher the conversion efficiency of nitrogen to NV color centers in diamond. In quantum-grade diamond, the nitrogen impurity content is extremely low, with a focal volume of approximately 1 µm³ for the focused femtosecond laser beam, resulting in a very low probability of vacancy formation next to impurities. The preparation of NV color centers typically involves focusing femtosecond laser pulses with energies below the threshold for amorphization. Under the action of femtosecond laser pulses, diamond generates a large number of free electrons through multiphoton and avalanche ionization, leading to the formation of a strong plasma and the creation of numerous vacancies in the lattice. Subsequently, a long-duration thermal treatment is performed in an annealing furnace, where the vacancies become active at temperatures above 600 °C and are captured by nitrogen impurities, resulting in the formation of NV color centers.³⁵ 

We used high-energy-density femtosecond lasers to continuously irradiate the surface of IIa diamond. Without undergoing any thermal treatment, NV centers were generated after approximately 1000 pulses. As the number of pulses increases, the NV-center luminescence peak tends to become gradually stronger, but after the number of pulses reaches 10⁶, the changes in ZPLs become less noticeable, while the phonon sideband on the right has a more notable enhancement. The cumulative effect of high-energy lasers causes severe damage to the diamond lattice, resulting in the creation of a large number of vacancies and the aggregation of nitrogen atoms in the processing area, leading to the conversion of a significant number of inactive nitrogen atoms into NV defects.

We studied the fluorescence lifetimes and imaging of the GR1 and NV color centers in the 10⁵-pulse region at 300 K. The GR1 color centers were excited by a laser with a wavelength of 633 nm, and the signal was filtered out using a bandpass filter with a central wavelength of 750 nm (bandwidth of 40 nm). The NV color centers were excited by a laser with a wavelength of 532 nm, and the signal was filtered out using a bandpass filter with a central wavelength of 635 nm (bandwidth 38 nm). The GR1 color centers in diamond are attributed to neutral vacancies in the carbon lattice. These color centers are luminescent and remain thermally stable at temperatures of several hundred degrees Celsius.³⁶ 

As shown in Fig. 10, the fluorescence lifetime of GR1 varies between 0.4 and 3.1 ns depending on the excitation source and sample temperature.³⁷ It has also been shown that the luminescence decay time of GR1 color centers is about 2.5 ns at 70 K and 1 ns at 300 K. The radiative lifetime is estimated to be 182 ns at 70 K, suggesting that non-radiative transitions dominate in GR1.³⁸ Our measured GR1 center lifetime of only 0.45 ns is due to the lack of annealing after processing, which resulted in a large number of defects in the processed region. These defects act as trap energy levels, capturing the excited state of the color center and promoting non-radiative transitions instead of allowing energy to be released through radiative transitions. This leads to shorter luminescence lifetimes and reduced luminescence efficiency. This phenomenon is also observed in the measurements of NV color centers. A systematic study of defect-related quenching of NV luminescence in diamond using time-correlated single-photon counting spectroscopy showed that for a single NV color center, the lifetime can be up to 25 ns when no quenching occurs; in the presence of impurities and defects, the lifetime typically ranges from 10 to 20 ns. In areas with graphitization, most NV lifetimes are less than 4 ns.^37,39 The lifetime of the NV color centers that we measured in the processed region was about 1.52 ns. This notable shortening is due to the interaction with defect centers such as GR1 and the quenching of NV luminescence caused by graphitization. In most applications, defects such as GR1 color centers are treated as impurities that can be removed or repaired in the lattice environment of NV color centers by annealing.³⁶ 

We characterized the fluorescence distribution of GR1 and NV color centers in the x–y plane of the processing area layer by layer, as shown in Figs. 11 and 12. The GR1 color center signal is strongest near the processing area, concentrated at the surface, and it diffuses outward with increasing depth. The high-energy femtosecond laser strongly perturbs the electronic system of the diamond during processing, while the accompanying heating induces thermal graphitization. Additionally, ionization triggers reactions between the surface and adsorbed oxygen, leading to nanoscale etching and the formation of vacancy defects on the carbon lattice surface. These vacancy defects are concentrated around the processing area, with some being driven by the high-density energy of the plasma, diffusing into the depths of the crystal and eventually forming dispersed vacancies. Furthermore, as shown in Fig. 12, a large number of NV color centers form within 3 μm of the surface. The distribution range of NV color centers is much broader than that of vacancy defects, and the distribution of nitrogen atoms largely determines the distribution of NV color centers. The femtosecond laser irradiation activates nitrogen atoms, which exist as optically inactive defects during the CVD growth process, and these atoms capture vacancies during the vacancy-diffusion process to form stable NV color centers.

Figure 13 shows the distribution of defects and damage induced by high-power laser pulses on IIa diamond. Figure 13(a) shows an SEM image of a cross section of the processed area; the dashed lines here correspond to the boundaries in Fig. 13(b). Figure 13(b) is plotted based on Raman and PL mapping results. Under the action of laser pulses, the localized pressure on the diamond surface reaches the thermal threshold of the molecular-deposition layer, resulting in an instantaneous high-temperature and high-pressure state and triggering chemical reactions on the diamond surface. The ablative effect of the laser pulse leads to the formation of ablation pits on the diamond surface, surrounded by a covering layer containing oxide particles, and creates a thermally affected zone. With the continuous action of the pulses, the damaged area expands longitudinally. The ablation region generates a graphite layer containing a large number of vacancy defects, and this diffuses into the surrounding area. Additionally, femtosecond laser pulses generate strong nonlinear optical absorption within the diamond, activating a large number of inactive nitrogen atoms near the graphite layer. Simultaneously, vacancies diffuse in the surface region under the influence of shock waves and are captured by active nitrogen atoms in the lattice, forming NV centers.

In summary, we investigated the effects of high-energy-density femtosecond laser pulses on IIa diamond. The structural changes and defect-formation mechanisms of diamond induced by pulsed laser irradiation were studied using Raman and fluorescence spectroscopy. Through Raman spectroscopy, we explored the effects of different numbers of laser pulses on the surface morphology, stress, and structure of diamond. The polarized Raman spectra revealed torsion and deformation of the diamond structure in the processing area, with a maximum tensile stress of 5.2 GPa. The processed areas exhibited higher nitrogen and oxygen contents compared to the unprocessed areas. Continuous exposure to pulsed laser irradiation resulted in the generation of luminescent defects associated with vacancies and interstitials, as observed through low-temperature fluorescence spectroscopy.

Raman mapping and fluorescence imaging were used to establish spatial distribution models for damage and defects. The Raman mapping showed variations in damage and the distribution of graphitized layers under different numbers of pulses, indicating that transverse damage tended to stabilize with an increasing number of pulses; conversely, longitudinal damage increased proportionally to the logarithm of the number of pulses. The graphite layer formed around the ablation pit contained a significant number of vacancy defects, which migrated within the diamond under the deposition of pulse energy and combined with laser-ionized nitrogen atoms to form NV color centers.

In this work, NV color center ensembles were generated by depositing an extremely high pulse energy density on the surface of diamond with very low nitrogen content without annealing. This provides a new approach to the processing of NV color center ensembles. This research delved into the behavior of diamond under the action of high-energy-density pulses, leading to a deeper understanding of the microstructure and properties of the material.

This study was funded by the Henan Key Laboratory of Intelligent Manufacturing Equipment Integration for Superhard Materials (Grant No. JDKJ2022-01), and the Key Lab of Modern Optical Technologies of Education Ministry of China, Soochow University.

The authors have no conflicts to disclose.

The data that support the findings of this study are available from the corresponding author upon reasonable request.

Bing Dong is currently pursuing a Doctoral degree at the State Key Laboratory of Precision Measuring Technology and Instruments, School of Precision Instrument and Opto-Electronics Engineering, Tianjin University, China.

Qingqing Sun is currently pursuing a master’s degree at the State Key Laboratory of Precision Measuring Technology and Instruments, School of Precision Instrument and Opto-Electronics Engineering, Tianjin University, China.

Jianshi Wang is studying for a master’s degree at the State Key Laboratory of Precision Measuring Technology and Instruments, School of Precision Instrument and Opto-Electronics Engineering, Tianjin University, China. His research interests include ultrafast femtosecond laser processing, Raman spectroscopy, and nanoparticle fabrication.

Ying Song is at the State Key Laboratory of Precision Measuring Technology and Instruments, School of Precision Instrument and Opto-Electronics Engineering, Tianjin University, China. Her research interests include the preparation of silicon carbide color centers by ion implantation, three-dimensional Raman and photoluminescence spectral characterization, and the modeling of spectral depth profiling.

Fei Ren received his master’s degree in Nuclear Technology and Applications from Lanzhou University in 2018. He is currently pursuing his doctorate at the State Key Laboratory of Precision Measuring Technology and Instruments, within the School of Precision Instrument and Opto-Electronics Engineering at Tianjin University, China. His research interests include quantum sensing based on spin defects in hexagonal boron nitride and atomistic simulations of point defects in semiconductors.

Zongwei Xu is a Professor at Tianjin University and a Doctoral Supervisor. His research interests include ultrafast energy beam processing (ion and laser), Raman and photoluminescence spectroscopy characterization, wide bandgap semiconductor devices, as well as microcutting and nanocutting technologies.