X-ray has the ability to interact with electrons in the atoms to ionize targeted materials and shows a good dose-dependent feature [1,2]. Therefore, X-ray detection is of great significance for medical imaging, security screening, industrial nondestructive inspection, scientific research, etc. [3–5]. To meet these applications, high detection sensitivity is crucial for achieving high-contrast images and reducing radiation-related health risks [6,7].

Typically, two different methods are adopted for X-ray detection. The dominant indirect-conversion X-ray detectors use phosphors or scintillators to transform an X-ray into visible light so the spatial resolution and conversion efficiency are limited [8,9]. In contrast, the direct-conversion X-ray detectors use a semiconductor layer to convert an X-ray directly into electric signals and show the advantages of high spatial resolution, fast response time, and wide linear dynamic range [10,11]. However, the commercial a-Se photoconductor used in direct-conversion X-ray detectors has the drawback of low X-ray absorption coefficient for general radiology applications, resulting in a limited sensitivity of $22.5{\mathrm{\mu CGy}}_{\mathrm{air}}^{-1}{\mathrm{cm}}^{-2}$ for 20–60 keV X-ray [4,5,11]. In addition, the ${\mathrm{HgI}}_2$ photoconductors are widely applied in X-ray detection owing to the heavy atoms, low electron-hole pair (EHP) production energy (5 eV), and high charge carrier mobility lifetime ($\mathit{\mu }\tau$) product value (${10}^{-5}$ to ${10}^{-3}{\mathrm{cm}}^2{\mathrm{V}}^{-1}$); further, the detectors exhibit a high sensitivity of $\sim 2.2\times {10}^3{\mathrm{\mu CGy}}_{\mathrm{air}}^{-1}{\mathrm{cm}}^{-2}$ for 26–72 keV X-ray [12,13]. However, the poor device compatibility and thermal stability of ${\mathrm{HgI}}_2$ photoconductors must be addressed before their practical application [14]. Recently, perovskite photoconductors prepared by solution process were introduced and exhibited a high X-ray absorption coefficient that enabled the low-dose X-ray detection [5,15]. The sensitivity of $1.1\times {10}^4{\mathrm{\mu CGy}}_{\mathrm{air}}^{-1}{\mathrm{cm}}^{-2}$ (100 keV X-ray) at $0.24\mathrm{V}{\mathrm{\mu m}}^{-1}$ and the response time of $\sim 10\mathrm{ms}$ were reported in polycrystalline ${\mathrm{CH}}_3{\mathrm{NH}}_3{\mathrm{PbI}}_3$-based X-ray detectors consisting of $1428\times 1428\mathrm{pixels}$ with a pixel pitch of 70 μm [5]. However, the ion migration and long-term operation stability problems of perovskite X-ray detectors still limit detector performance [16].

In addition, the X-ray detection sensitivity can be tremendously improved using the photoelectron multiplication mechanism. There are three main photoelectron multiplication mechanisms for realizing highly sensitive X-ray detection: photoemission and secondary emission in vacuum photomultiplier tubes (PMTs) [17], avalanche multiplication mechanism in avalanche photodiodes (APDs) and pixelated silicon photomultipliers (SiPMs) [18,19], and photoconductive gain in X-ray detectors with electronic injection at high electric field [20]. However, commercial PMTs have a bulky size and high cost; the APDs and SiPMs have drawbacks of temperature-dependent gain and low X-ray absorption coefficient due to material limitation; the X-ray detectors with photoconductor gain suffer from the issue of a current runaway effect that occurs at a high electric field so the gain is limited. Therefore, it is challenging to develop a flat-panel X-ray detector with simple preparation method, high-stability, high X-ray absorption coefficient, and high internal gain.

In order to develop a flat-panel X-ray detector for highly sensitive and low-dose X-ray detection, vacuum cold cathode flat-panel X-ray detectors have been fabricated successfully in recent years and achieve high internal gain [21,22]. The detectors consist of a-Se high-gain avalanche rushing photoconductors (HARPs) for photoelectric conversion and Spindt-type field emission arrays (FEAs) for signal readout [21]. Those vacuum X-ray detectors achieved a pixel number of $640\times 480$, a spatial resolution of 20 μm, and an avalanche multiplication factor of 200 [22]. However, the realized X-ray sensitivity and effective device area were limited due to material limitation of HARPs and complicated fabrication method of Spindt-type FEAs, making those X-ray detectors far from actual applications. Recently, the vacuum cold cathode flat-panel X-ray detectors consisting of ZnS photoconductor and ZnO nanowire (NW) field emitters with a photoelectron multiplication mechanism called electron-bombardment-induced photoconductivity (EBIPC) were proposed to achieve large-area ($4.8\mathrm{cm}\times 8\mathrm{cm}$) and high-gain ($\sim {10}^4$) indirect-conversion X-ray detection [23,24]. Nevertheless, the direct-conversion vacuum cold cathode X-ray detectors operating with an EBIPC mechanism have not been realized, and the corresponding charge transport mechanism requires further investigation.

Here, the direct-conversion vacuum cold cathode X-ray detectors consisting of ZnO NW field emitters and $\mathrm{\beta }\text{-}{\mathrm{Ga}}_2{\mathrm{O}}_3$ photoconductor targets were proposed to realize sensitive and low-dose X-ray detection. The ZnO NW field emitters show advantages of simple preparation method, good controllability, large area, and excellent field emission performance [25]. The $\mathrm{\beta }\text{-}{\mathrm{Ga}}_2{\mathrm{O}}_3$ photoconductors have a wide bandgap of $\sim 4.9\mathrm{eV}$, a breakdown voltage of $8\mathrm{MV}{\mathrm{cm}}^{-1}$, a relatively high density ($6.44\mathrm{g}{\mathrm{cm}}^{-3}$), and a high absorption coefficient ($17.2{\mathrm{mm}}^{-1}$ for 20 keV X-ray) for low-energy X-ray [26]. Those metrics meet the requirement of vacuum cold cathode X-ray detectors with a low dark current and an effective EBIPC effect for achieving high internal gain. In addition, the charge carrier transport mechanism of EBIPC effect in X-ray detectors was investigated to achieve sensitive X-ray detection and low detection limit. These results demonstrate that the direct-conversion vacuum cold cathode X-ray detectors with an EBIPC mechanism can potentially be applied in various X-ray detecting applications.

Conventional direct-conversion X-ray detectors with thin film transistor (TFT) readout are illustrated in Fig. 1(a) [4]. The EHPs are generated in a photoconductor through the absorption of X-ray photons. Nowadays, a-Se photoconductors used in commercial X-ray detectors have drawbacks of low X-ray absorption efficiency and charge collection efficiency. The developing X-ray detectors based on CdTe [27], PbO [28], ${\mathrm{HgI}}_2$ [12], and perovskite [1] photoconductors have high X-ray absorption coefficient and low EHP production energy to generate adequate EHPs, resulting in a significant improvement of X-ray detection sensitivity. Nevertheless, the reported X-ray detectors lacked sufficient internal gain to overcome the recombination or trapping of EHPs. The architectures of 3D photosensitive TFT based X-ray detectors were proposed to achieve high internal gain of $\sim {10}^4$; however, those devices raised fabrication complexity and cost [29].

The proposed device structure formed by gated FEAs and photoconductor target for achieving low dose X-ray imaging is illustrated in Fig. 1(b). These vacuum cold cathode X-ray detectors with high-resolution imaging (20 μm) and high-radiation tolerance ($1{\mathrm{MGy}}_{\mathrm{air}}\mathrm{\gamma }$-ray) have been demonstrated [30]. In addition, the EHPs in photoconductor can be modulated by both incident photons and accelerated electron emission from a cold cathode, leading to high internal gain. Owing to such high gain effect, the ability of vacuum cold cathode photodetectors to detect ultraweak visible light ($100\mathrm{nW}{\mathrm{cm}}^{-2}$) has been well demonstrated [23]. Therefore, the direct-conversion X-ray detection with high gain can be expected by using vacuum cold cathode detectors. The corresponding X-ray response process of designed X-ray detectors can be divided into four steps: (1) EHPs are generated by incident X-ray photons in the photoconductor layer; (2) the generated EHPs drift across the photoconductor under the electric field; (3) the conductivity of the photoconductor is decreased and causes the electron emission of a cold cathode; (4) the photoconductor is bombarded by the field emission electrons of a cold cathode, and abundant EHPs are generated by the EBIPC effect. The EBIPC effect and field emission of a cold cathode form a positive feedback process until reaching a balance when the voltage dropped in the photoconductor increases at high current. Through this high gain mechanism, internal signal amplification and a highly sensitive X-ray detection can be realized in vacuum cold cathode X-ray detectors. More importantly, the vacuum gap in the proposed X-ray detector can realize a low dark current at high electric field, enabling a low-dose X-ray detection.

Our device adopts a vacuum photodiode formed by a ${\mathrm{Ga}}_2{\mathrm{O}}_3$ photoconductor and ZnO NW field emitters to realize the direct-conversion X-ray detection with an EBIPC mechanism, as illustrated in Fig. 2(a). Figure 2(b) shows an actual photograph of the X-ray detector. The generation rate of EHPs inside the ${\mathrm{Ga}}_2{\mathrm{O}}_3$ photoconductor caused by the EBIPC effect strongly depends on the electron beam accelerated voltage and accelerated current emission from the ZnO NWs. Therefore, the photoelectron multiplication factor can be tuned in an X-ray detector using cold cathodes made in forms of addressable FEAs [25].

In addition, the charge collection efficiency of the X-ray detector is related to the field emission characteristics of ZnO NWs and the performance metrics of ${\mathrm{Ga}}_2{\mathrm{O}}_3$ photoconductors. The top image of Fig. 2(c) shows the actual arrays of ZnO NWs grown on an indium tin oxide (ITO)-coated glass substrate; the bottom image of Fig. 2(c) shows the cross-sectional view of ZnO NWs. The ZnO NWs grown directly on the substrate had a distribution density of $\sim 5\times {10}^8{\mathrm{cm}}^{-2}$, a height of 3–4 μm, and a tip diameter of $\sim 20\mathrm{nm}$. The grown ZnO NWs are reported to have a uniform morphology and are easy to be integrated in gated FEAs, which is beneficial for X-ray imaging applications [25]. Figure 2(d) shows the field emission current density-electric field (J-E) characteristics of ZnO NWs. The grown ZnO NWs with a turn-on field (which corresponds to a current density of $10\mathrm{\mu A}{\mathrm{cm}}^{-2}$) of $4.4\mathrm{V}{\mathrm{\mu m}}^{-1}$ showed excellent field emission characteristics compared with previously reported ZnO NWs [31]. The Fowler–Nordheim (F-N) plot [shown in the inset of Fig. 2(d)] was approximately a straight line, indicating that the field emission of ZnO NWs followed the classic vacuum electron tunneling mechanism [32].

In order to obtain the crystallinity of ${\mathrm{Ga}}_2{\mathrm{O}}_3$ photoconductors, the typical X-ray diffraction (XRD) pattern was measured [Fig. 2(e)]. The peaks fitted well with the monoclinic ${\mathrm{Ga}}_2{\mathrm{O}}_3$ Joint Committee on Power Diffraction Standards (JCPDS) card No. 43-1012 with the lattice constants of $a=12.23$, $b=3.04$, and $c=5.8$. The peaks located at 18.1°, 37.7°, and 58.5° correspond to the ($-201$), (311), and ($-603$) planes. The corresponding full width at half maximum of the XRD peaks of current ${\mathrm{Ga}}_2{\mathrm{O}}_3$ was 576 arcsec, which was nearly the same as that of reported ${\mathrm{Ga}}_2{\mathrm{O}}_3$ crystal (540 arcsec) prepared by metal–organic chemical vapor deposition method [33]. The XRD results demonstrate the high crystal quality of the ${\mathrm{Ga}}_2{\mathrm{O}}_3$ photoconductor, which is beneficial for realizing low dark current, fast response time, and high charge collection efficiency. Figure 2(f) shows the X-ray mass attenuation coefficients of different photoconductors [obtained from the National Institute of Standards and Technology (NIST) database] [34]. On the one hand, the photoelectric effect dominates the total mass attenuation coefficient of the ${\mathrm{Ga}}_2{\mathrm{O}}_3$ photoconductor throughout the low-energy X-ray range; on the other hand, the total mass attenuation coefficient of the ${\mathrm{Ga}}_2{\mathrm{O}}_3$ photoconductor is close to that of a-Se and ${\mathrm{CsPbBr}}_3$ photoconductors throughout the low-energy X-ray range ($<50\mathrm{keV}$). Those metrics demonstrate that the ${\mathrm{Ga}}_2{\mathrm{O}}_3$ photoconductor is an ideal material for low-energy X-ray detection.

In addition to exploring the charge transport mechanism of the detector to achieve maximum X-ray detection sensitivity, the dark current and photocurrent were measured at low and high electric fields [Figs. 3(a) and 3(b)]. The nonzero dark current and photocurrent at zero electric field shown in Fig. 3(a) might be caused by the trapped electron in the ${\mathrm{Ga}}_2{\mathrm{O}}_3$ crystal due to the electron bombardment. The negatively charged ${\mathrm{Ga}}_2{\mathrm{O}}_3$ crystal would release electrons forming a reversed leakage current when we repetitively measured the device. The current-electric field (I-E) characteristics of X-ray detector shown in Fig. 3(b) exhibited three processes: (1) the device current increased gradually until reaching saturation at $1.5\mathrm{V}{\mathrm{\mu m}}^{-1}$; (2) the device current increased slightly at the electric field between 1.5 and $4\mathrm{V}{\mathrm{\mu m}}^{-1}$; (3) the device current increased dramatically after $4\mathrm{V}{\mathrm{\mu m}}^{-1}$. In process (3), the device current (I) can be described using the F-N theory [32]: $$I=A\frac{V^2}{E_{\phi }{t}^2(y)}\exp (\frac{B{E}_{\phi }^{\frac{3}{2}}}{V}v(y)),$$(1)where $A$ and $B$ are constants, $V$ is the equivalent voltage in the tip of ZnO NWs, $E_{\varphi }$ is the work function of ZnO NWs, and $t(y)$ and $v(y)$ are the Nordheim elliptic functions. Figure 3(c) shows the F-N plots of dark current and photocurrent in process (3), suggesting that the device current followed the classic electron tunneling mechanism first and then exhibited a saturation feature due to the increasing voltage drop in ${\mathrm{Ga}}_2{\mathrm{O}}_3$ photoconductor induced by the high current [23].

The operation principle of the X-ray detector is illustrated to further demonstrate the I-E characteristics and X-ray response process at different electric fields. As shown in Figs. 3(d)–3(f), the charge transport mechanism of the X-ray detector can be divided into three regions. In Region I, the internal photoelectric effect of X-ray photons in the ${\mathrm{Ga}}_2{\mathrm{O}}_3$ photoconductor dominates at the small electric field. The collected charge increases before reaching saturation at the saturated electric field ($E_{\mathrm{sat}}$), owing to the limited photogenerated carriers in the ${\mathrm{Ga}}_2{\mathrm{O}}_3$ photoconductor and high vacuum barrier of the ZnO NWs. In Region II, the charges are injected from the electrodes when the charge recombination lifetime is larger than the charge transit time at a higher electric field, which triggers a photoconductive gain [20]. The device current increases slightly in this region, but the total collected charge is still limited by the large vacuum barrier of ZnO NWs. In Region III, the EBIPC effect occurs when the ${\mathrm{Ga}}_2{\mathrm{O}}_3$ photoconductor is bombarded by the high-energy accelerated electrons emission from ZnO NWs at a high electric field ($E_b$). Under the EBIPC effect, the conductivity of a ${\mathrm{Ga}}_2{\mathrm{O}}_3$ photoconductor becomes smaller, and the vacuum barrier of ZnO NWs becomes lower and narrower, making the photocurrent increase dramatically; further, the collected photogenerated charge is amplified eventually.

Under the EBIPC mechanism, the X-ray detector achieved high detection sensitivity. Calculated from the X-ray responsive J-E curves of the metal-semiconductor-metal-structured ${\mathrm{Ga}}_2{\mathrm{O}}_3$ X-ray detector, the resistivity of the ${\mathrm{Ga}}_2{\mathrm{O}}_3$ photoconductor decreased from $8.6\times {10}^{11}\mathrm{\Omega }\mathrm{cm}$ to $3.4\times {10}^{11}\mathrm{\Omega }\mathrm{cm}$ after the illumination of 50 keV X-ray. According to the field emission J-E curve of the ZnO NWs [Fig. 2(d)] and X-ray responsive J-E curve of the proposed vacuum cold cathode X-ray detector, the bias variation of ZnO NWs was $\sim 0.08\mathrm{V}{\mathrm{\mu m}}^{-1}$ after the illumination of 50 keV X-ray, making resistivity of the ${\mathrm{Ga}}_2{\mathrm{O}}_3$ photoconductor become $6.2\times {10}^{10}\mathrm{\Omega }\mathrm{cm}$, owing to the bombardment of accelerated electrons with higher energy. Figure 3(g) shows the X-ray detection sensitivities as a function of an applied electric field under different X-ray tube voltages. The sensitivity increased quickly with the increasing applied electric field of the detector owing to the high generation rates of EHPs in ${\mathrm{Ga}}_2{\mathrm{O}}_3$ photoconductor at high electric field, and then the sensitivity saturated at a higher applied electric field due to a balance between the accelerated voltage of field emission and bias voltage of the ${\mathrm{Ga}}_2{\mathrm{O}}_3$ photoconductor. The maximum sensitivity of the X-ray detector was $3.0\times {10}^3{\mathrm{\mu CGy}}_{\mathrm{air}}^{-1}{\mathrm{cm}}^{-2}$ (the applied electric field was $22.5\mathrm{V}{\mathrm{\mu m}}^{-1}$) under the X-ray tube voltage of 6 kV. The realized X-ray detection sensitivity was much higher than that of the metal-semiconductor-metal-structured ${\mathrm{Ga}}_2{\mathrm{O}}_3$ X-ray detector with the maximum sensitivity of $63{\mathrm{\mu CGy}}_{\mathrm{air}}^{-1}{\mathrm{cm}}^{-2}$ (6 keV X-ray), exhibiting the high internal gain of vacuum cold cathode X-ray detectors with EBIPC mechanism. The internal gain ($G$) of the X-ray detector can be described as [10] $$G=\frac{I_R}{I_P}=\frac{I_R}{\phi \beta e},$$(2)where $I_R$ and $I_P=\phi \beta e$ stand for the experimental current and theoretical photogenerated current of detector under X-ray illumination, in which $\phi =\frac{\epsilon D{m}_s}{E_{\mathrm{ph}}}$ is the photon absorption rate and $\beta =\frac{E_{\mathrm{ph}}}{W_{\pm }}$ is the number of carriers excited by an X-ray photon. Notably, $e$ is an electronic charge, $\epsilon$ is the fraction of absorbed photons [this value is nearly 100% in 0.68 mm ${\mathrm{Ga}}_2{\mathrm{O}}_3$ photoconductor for low-energy (1–20 keV) X-ray calculated from the NIST database] [34], $D$ is the X-ray dose rate, $m_s$ is the mass of ${\mathrm{Ga}}_2{\mathrm{O}}_3$ photoconductor, $E_{\mathrm{ph}}$ is the X-ray energy, and $W_{\pm }=2.7E_g+E_{\mathrm{phonon}}$ is the EHP production energy [11], where $E_g=4.9\mathrm{eV}$ is the bandgap of the ${\mathrm{Ga}}_2{\mathrm{O}}_3$ photoconductor and $E_{\mathrm{phonon}}=29\mathrm{meV}$ is the phonon energy of ${\mathrm{Ga}}_2{\mathrm{O}}_3$ photoconductor [35]. The calculated internal gain of an X-ray detector up to $2.9\times {10}^2$ at $22.5\mathrm{V}{\mathrm{\mu m}}^{-1}$ electric field is achieved [Fig. 3(h)], demonstrating that the EBIPC mechanism is promising for use in sensitive X-ray detectors with high internal gain.

Figure 3(i) shows the X-ray detection sensitivity under different X-ray tube voltages measured by a DC X-ray tube without beam filters. The sensitivity decreased quickly from $1.2\times {10}^3$ to $0.3{\mathrm{\mu CGy}}_{\mathrm{air}}^{-1}{\mathrm{cm}}^{-2}$ when the X-ray tube voltage increased from 6 to 50 kV at $10\mathrm{V}{\mathrm{\mu m}}^{-1}$, indicating that the ${\mathrm{Ga}}_2{\mathrm{O}}_3$-based vacuum cold cathode X-ray detectors are suitable for low-energy X-ray detection owing to the high absorption efficiency of the ${\mathrm{Ga}}_2{\mathrm{O}}_3$ photoconductor for low-energy X-ray. The detection sensitivity (S) can be theoretically described as [36,37] $$S=GC\frac{e\mathrm{\Psi }{\alpha }_{\mathrm{en}}(1-\exp (-\alpha T))}{W_{\pm }},$$(3)where $C$ is the charge collection efficiency, $\mathrm{\Psi }$ is the incident radiant energy flux density, ${\alpha }_{\mathrm{en}}$ is the energy absorption coefficient of ${\mathrm{Ga}}_2{\mathrm{O}}_3$, $\alpha$ is the linear attenuation coefficient of ${\mathrm{Ga}}_2{\mathrm{O}}_3$, and $T$ is the thickness of ${\mathrm{Ga}}_2{\mathrm{O}}_3$. According to Eq. (3), the sensitivity is mainly dominated by the energy absorption coefficient and linear attenuation coefficient of the ${\mathrm{Ga}}_2{\mathrm{O}}_3$ photoconductor under different X-ray energies. The simulated detection sensitivity as a function of X-ray energy using Eq. (3) is shown in the inset of Fig. 3(i), showing good agreement with the experimental results.

Typical temporal responses of the X-ray detectors at the electric field of $5\mathrm{V}{\mathrm{\mu m}}^{-1}$ were tested with X-ray square-waves (the dose rate is $17.9{\mathrm{mGy}}_{\mathrm{air}}{\mathrm{s}}^{-1}$), as displayed in Figs. 4(a)–4(c). The fluctuation ($F$) of the photocurrent is calculated using $F=(I_2-I_1)/I_1$, where $I_1$ is the average photocurrent of the first pulse and $I_2$ is the average photocurrent of the last pulse. The fluctuations in the photocurrent were 0.4% over 170 s [Fig. 4(a)] and 0.6% over a 1 h period [Fig. 4(c)], indicating that the proposed X-ray detectors have an excellent stability owing to the high crystal quality of the ${\mathrm{Ga}}_2{\mathrm{O}}_3$ photoconductor and current limitation of the vacuum gap. The rise and fall times (the time for the transit between 10% and 90% of the photocurrent) were calculated to be 140 and 40 ms, respectively. The response time of proposed X-ray detectors is much faster than that of the previously reported ${\mathrm{Ga}}_2{\mathrm{O}}_3$-based X-ray detector (1.8 s) [38], which is mainly caused by the fast charge transport speed in the high-crystallinity ${\mathrm{Ga}}_2{\mathrm{O}}_3$ photoconductor and vacuum gap. The realized response speed is fully adequate for radiography applications requiring a readout time of less than 5 s [1]. In addition, Fig. 4(d) shows the signal-to-noise ratios (SNRs) [39] under different X-ray dose rates at the electric field of $5\mathrm{V}{\mathrm{\mu m}}^{-1}$; the inset of Fig. 4(d) shows the corresponding temporal responses under different X-ray dose rates. The SNR of X-ray detector increased from 0.9 to 64.8 when the radiation dose rate increased from $1.2\times {10}^{-3}$ to $38.1{\mathrm{mGy}}_{\mathrm{air}}{\mathrm{s}}^{-1}$. The SNR showed a saturation feature under the illumination of a low-dose X-ray ($<0.1{\mathrm{mGy}}_{\mathrm{air}}{\mathrm{s}}^{-1}$) due to the measured current noise [as shown in the inset of Fig. 4(d)], which could be increased by adopting more sophisticated electrostatic shielding and cable. The corresponding detection limit (according to the IUPAC standard of SNR of 3) [39] was as low as $0.28{\mathrm{mGy}}_{\mathrm{air}}{\mathrm{s}}^{-1}$, exhibiting a large improvement compared with that of the reported ${\mathrm{Ga}}_2{\mathrm{O}}_3$ based X-ray detector ($\sim 140{\mathrm{mGy}}_{\mathrm{air}}{\mathrm{s}}^{-1}$) [38].

The performance metrics of the proposed X-ray detector and literature-reported direct-conversion X-ray detectors using different photoconductors and photoelectron multiplication mechanisms are compared in Table 1. The proposed X-ray detectors formed by photoconductor targets and cold cathode field emitters can achieve high internal gain using the EBIPC mechanism. The ${\mathrm{Ga}}_2{\mathrm{O}}_3$ photoconductor is an ideal material for a vacuum cold cathode X-ray detector owing to its high absorption efficiency for a low-energy X-ray and high breakdown electric field to operate at a high electric field. Under the EBIPC mechanism, the proposed X-ray detector realized a high internal gain of $2.9\times {10}^2$, a high X-ray detection sensitivity of $3.0\times {10}^3{\mathrm{\mu CGy}}_{\mathrm{air}}^{-1}{\mathrm{cm}}^{-2}$ for $6\text{−}50\mathrm{keV}$ X-ray, and a low detection limit of $0.28{\mathrm{mGy}}_{\mathrm{air}}{\mathrm{s}}^{-1}$. Furthermore, the detector achieved a fast response time of 40 ms owing to the high crystal quality of the ${\mathrm{Ga}}_2{\mathrm{O}}_3$ photoconductor and fast charge transport speed in the vacuum gap. The performance metrics of the proposed X-ray detectors are much better than those of the literature-reported ${\mathrm{Ga}}_2{\mathrm{O}}_3$-based X-ray detectors without a photoelectron multiplication mechanism [40]. In contrast, the vacuum cold cathode X-ray detectors formed by HARPs and Spindt-type FEAs had a high internal gain ($2.0\times {10}^2$) and were expected to realize low-dose X-ray imaging using the avalanche effect [41,42]. However, those detectors have limited X-ray detection sensitivity due to the material limitation of HARPs and are not eligible for large area applications due to the complicated fabrication method of Spindt-type FEAs. Furthermore, the a-Se-based X-ray detectors without photoelectron multiplication mechanism had the drawback of low X-ray absorption coefficient for general radiology applications, resulting in a limited sensitivity of $22.5{\mathrm{\mu CGy}}_{\mathrm{air}}^{-1}{\mathrm{cm}}^{-2}$ for 20–60 keV X-ray [11,43,44]. In addition, the perovskite photoconductors had high X-ray absorption coefficient and high charge collection efficiency, benefiting to realize high X-ray detection sensitivity ($4.1\times {10}^3{\mathrm{\mu CGy}}_{\mathrm{air}}^{-1}{\mathrm{cm}}^{-2}$ for 20–60 keV X-ray) and low detection limit ($0.22{\mathrm{\mu Gy}}_{\mathrm{air}}{\mathrm{s}}^{-1}$) [45]. Furthermore, the perovskite based X-ray detectors could realize a higher detection sensitivity ($5.6\times {10}^4{\mathrm{\mu CGy}}_{\mathrm{air}}^{-1}{\mathrm{cm}}^{-2}$ for 30–50 keV X-ray) using the photoconductive gain effect and a fast response time ($<0.3\mathrm{ms}$) by applying a photoconductor with high crystal quality [6,45]. Nevertheless, the issues of ion migration and long-term operation stability make the perovskite-based X-ray detectors far from practical applications [16].Table 1.

Comparison of Performance Metrics of Direct-Conversion X-ray Detectors Based on Different Photoconductors and Photoelectron Multiplication Mechanisms

In summary, a direct-conversion vacuum cold cathode X-ray detector based on ZnO NW field emitters and $\mathrm{\beta }\text{-}{\mathrm{Ga}}_2{\mathrm{O}}_3$ photoconductor targets was proposed to realize sensitive and low-dose X-ray detection. The charge carrier transport mechanism of the X-ray detectors was investigated, finding that the X-ray response process could be divided into internal photoelectric effect, photoconductive gain effect, and EBIPC effect when the applied electric field increased accordingly. The proposed X-ray detectors operating with an EBIPC mechanism achieved a high internal gain of $2.9\times {10}^2$ and high detection sensitivity up to $3.0\times {10}^3{\mathrm{\mu CGy}}_{\mathrm{air}}^{-1}{\mathrm{cm}}^{-2}$ for 6 keV X-ray at the $22.5\mathrm{V}{\mathrm{\mu m}}^{-1}$ electric field. Moreover, the direct-conversion vacuum cold cathode X-ray detector also had the advantages of fast response time, long-term stability, and low detection limit owing to the high crystal quality of the ${\mathrm{Ga}}_2{\mathrm{O}}_3$ photoconductor and current limitation effect of the vacuum gap. The proposed direct-conversion vacuum cold cathode X-ray detectors demonstrate huge potential for use in actual X-ray detection applications.