Abstract

In this work, the performance of thin silicon carbide membranes as material for radiation hard x-ray shine position monitors ( XBPMs ) is investigated. Thermal and electrical behavior of XBPMs made from thin silicon carbide membranes and single-crystal baseball diamond is compared using finite-element simulations. Fabricated silicon carbide devices are besides compared with a 12 µm commercial polycrystalline baseball diamond XBPM at the swiss Light Source at the Paul Scherrer Institute. Results show that silicon carbide devices can reach equivalent transparencies while showing improved one-dimensionality, dynamics and signal-to-noise ratio ratio compared with commercial polycrystalline ball field XBPMs. Given the obtained results and handiness of electronic-grade epitaxies on up to 6 edge wafers, it is expected that silicon carbide can substitute for diamond in most air monitor applications, whereas ball field, owing to its lower assimilation, could remain the material of choice in cases of extreme roentgenogram power densities, such as pink and blank beams. Keywords:

beam position monitors, silicon carbide, radiation detector, beamline instrumentation, X-rays

1. Introduction  

Synchrotron light sources give birth x-ray beams with high glare to endstations, where experiments from macromolecular crystallography to scanning x-ray nanoprobe are conducted ( Owen et al., 2016 ▸ ). These applications benefit from highly crystalline, compress, firm and reliable roentgenogram beam position monitors ( XBPMs ) with high lateral pass resolutions, capable of withstanding senior high school exponent densities and high radiation doses. such devices enable precise decision of intensity, position and, in some cases, human body of the beam. This can be used to establish a feedback loop with beamline optics improving beam position constancy. state-of-the-art XBPMs can be divided into three types : ( one ) peripheral, non-destructive in-line XBPMs such as blade monitors which intercept only the out fringes ; ( two ) on-axis, destructive XBPMs such as fluorescent screens ; and ( three ) on-axis, non-destructive in-line XBPMs with high foil to the beam ( Schulze-Briese et al., 2001 ▸ ; Leban et al., 2010 ▸ ). peripheral XBPMs allow minimal beam hindrance but suffer from low lateral resolution, high sensitivity to external noise, and systematic errors in the case of non-Gaussian beams. In-line, destructive XBPMs interfere with the glow and are operated only during beamline commission and set up ( Bunk et al., 2005 ▸ ). however, to perform automatic optics corrections during operation, continuous monitor with high lateral pass resolution and dear signal-to-noise ratio ratio is needed. This increases the quality of radio beam delivered to the endstations and therefore the quality of experimental data. To this end, non-destructive, in-line XBPMs are needed. The main drawbacks of in-line XBPMs are residual hindrance with the radio beam and abasement due to heat warhead and radiation. To minimize these drawbacks, transparency, high-temperature constancy and radiation sickness hardness are primary requirements for this class of monitors. Linearity and fast dynamics are besides important to maintain a static feedback. This has driven research in wide-band-gap semiconductors for radiation monitor applications ( Schulze-Briese et al., 2001 ▸ ). baseball diamond is the fabric of option among wide-band-gap semiconductors ascribable to its excellent transparency, radiotherapy severity and high thermal conduction ( Schulze-Briese et al., 2001 ▸ ; Smedley et al., 2011 ▸ ; Muller et al., 2012 ▸ ; Marinelli et al., 2012 ▸ ; Desjardins et al., 2014 ▸ ; Zhou et al., 2015 ▸ ; Williams et al., 2016 ▸ ; Griesmayer et al., 2016 ▸ ). Single-crystal diamond ( sc-diamond ) ampere well as polycrystalline diamond on silicon ( pc-diamond, or CVD ball field ) XBPMs are now commercially available ( DECTRIS, CIVIDEC ). The craved device properties of sc-diamond are not obtained on a commercial plate due to the practice of blockheaded substrates with high concentration, and handiness of samples merely smaller than 10 millimeter × 10 millimeter ( Khmelnitskiy, 2015 ▸ ). Improving foil by thinning down the thickly substrates and fabricating membranes encounters challenges. reactive ion engrave is tested but results in besides high ( > 50 % ) thickness non-uniformity ( Desjardins et al., 2014 ▸ ). A thin pc-diamond membrane grown on silicon is highly crystalline. however, polycrystalline corporeal has defects and grain boundaries that result in decelerate dynamics and non-linear­ities ( Bergonzo et al., 2006 ▸ ). The dispute in thermal expansion between a sparse diamond film and a dense Si-substrate besides gives heighten to wafer bow, thereby limiting the wafer size to 3 edge ( RIGI, DECTRIS ). A high concentration of defects besides reduces reproducibility hindering industrialization of the pc-diamond devices. XBPMs made of silicon carbide would provide high thermal conduction and inertness as their diamond counterparts ( Desjardins et al., 2014 ▸ ). furthermore, electronic-grade single-crystal 4H-SiC wafers with much lower defects densities than baseball diamond are available up to a diameter of 6 inch, avoiding the bottlenecks of diamond engineering. therefore, this technologically fledged wide-bandgap material shows bang-up promise for next-generation industrial XBPMs. however, due to difficulties in fabricating micrometre thin active agent layers, 4H-SiC has never been considered as a campaigner for XBPMs. recently, it was shown that electrochemical etching in HF-based solutions could selectively remove the highly dope 4H-SiC substrate with an etch stop on the low-doped epitaxial layers ( Dahal et al., 2017 ▸ ). This method acting is used in our survey to fabricate XBPMs on thin epitaxial membranes on 4H-SiC substrates. Fabricated devices are then tested and compared with a commercial 12 µm pc-diamond XBPM at the swiss Light Source ( SLS ) at the Paul Scherrer Institute ( PSI ). The theoretical thermal and electrical behaviors of diamond and 4H-SiC XBPMs are described in the future section followed by the experimental results on the fabricated devices .

3. Device fabrication  

4H-SiC XBPM devices are fabricated on wafers with epitaxial layers grown on 375 µm 1 × 1018 cm−3 n-type substrates. The epitaxial layers are either 2 µm- or 10 µm-thick 5 × 1013 cm−3 n-type with a 0.5 µm 1 × 1018 cm−3 p-type as a exceed layer. The four-quadrant monitors are fabricated with reactive ion etching ( RIE ) in SF6/argon plasma to remove the p-type layer with lodge metals as an etching disguise. The etching mask is then used as an electrical contact to remove substrates via electrochemical etching. Different membrane thicknesses are achieved with far dry engraving in SF6/Ar plasma. electrochemical engrave ( ECE ) is an oxidation/oxide removal process obtained by dipping silicon carbide samples in an HF solution and electrically supply holes for the oxidation through the back metallic element reach ( Dahal et al., 2017 ▸ ; Watanabe et al., 2011 ▸ ; Gautier et al., 2012 ▸, 2013 ▸ ). The work is capable of removing highly doped ( ≥1 × 1018 cm−3 ) p-type and n-type layers but is selective towards low-doped n-type layers ( selectivity ≥ 1000:1 with respect to 5 × 1013 cm−3 doped layers ). This allows the chummy highly dope substrate to be selectively removed, realizing membranes with thicknesses and uniformities as determined by the epitaxial layers. In this study, to evaluate devices with different thicknesses starting from lone two epitaxies, some devices are promote etched using a standard RIE SF6/Ar process. This allowed us to obtain devices with thicknesses down to 0.6 µm ( 0.5 µm p-type plus 0.1 µm n-type ), as determined from the etching rate of the process. consistent with the commercial pc-diamond device used for comparison ( RIGI, DECTRIS ), a 6 µm col separated the four front electrodes [ hybrid section shown in the inset of Fig. 5 ( a ) ] .

4. Measurement results  

Most of experimental tests are conducted at either the X06SA or at the OPTICS beamlines of the SLS at the PSI in Switzerland. Beam widths between 50 and 200 µm in both vertical and horizontal guidance are obtained by two collimator slits located 5 millimeter before the XBPM. Transmission through the XBPMs is determined by comparing the current of a 300 µm Hamamatsu silicon diode monitor 35 curium downstream of the XPBM with and without the XBPM intercepting the balance beam. Motorized transformation stages enabled precise apparent motion of the XBPM in two orthogonal directions ( ten, yttrium ) cross to the beam. The rear-side liaison of the XBPM is biased and the current signals from the quadrant electrodes are measured by a four-channel electrometer ( AH501D from CAENELS ). Unless argue otherwise, all measurements described below are performed with this shape. Fig. 6 ( a ) shows the beam transmission as a routine of the beam lateral pass place for the 1.24 µm 4H-SiC and a 12 µm commercial pc-diamond XBPM device supplied by DECTRIS at 8 keV beam energy. Within the measurement error, taxonomic plus random, estimated to be ∼2 nA, equivalent to ∼5 % transmission error, we observe a well defined membrane region in both devices. The transmission on the 4H-SiC membrane is ≥95 %, compared with regions of the sample with the substrate where the transmittance is below 30 %. The 12 µm pc-diamond and 1.24 µm 4H-SiC XBPM have similar transmittance within experimental error.

An external file that holds a picture, illustration, etc.
Object name is s-26-00028-fig6.jpgOpen in a separate window We subsequently analyse the electric response of the 4H-SiC XBPM as compared with that of the pc-diamond. In the font of 4H-SiC, the built-in bias of the p–n junction in 4H-SiC achieves saturated charge collection at zero external bias, consistent with device simulations [ Fig. 5 ( a ) ]. On the other hand, saturated collection efficiency is achieved in pc-diamond only for biases greater than 20 V due to lifetime killing defects. Without these defects, sc-diamond requires only ∼2 V to achieve saturated collection [ Fig. 5 ( a ) and Desjardins et aluminum. ( 2014 ▸ ) ]. Given the above leave, all measurements on 4H-SiC XBPMs are performed at zero bias and, systematically with seller specifications, we will constantly use 30 V for pc-diamond. Fig. 7 shows the electric response of the 1.24 µm 4H-SiC XBPM in XY raster scan as compared with that of pc-diamond. Both devices show good uniformity, but the 4H-SiC XBPM shows a superior, more than four times higher, signal-to-noise ratio ratio. The randomness degree for both XBPMs was below the measurement specify of 100 dad .An external file that holds a picture, illustration, etc.
Object name is s-26-00028-fig7.jpgOpen in a separate window The watch higher current bespeak for the 4H-SiC device is partially due to the difference in electron–hole pair creation energy of the two materials [ 7.8 electron volt and 13 electron volt for 4H-SiC and diamond, respectively ( Bertuccio & Casiraghi, 2003 ▸ ) ] and partially due to the dispute in the charge collection efficiency of the two devices ( 91 % and 31 % for the 4H-SiC and pc-diamond, respectively, at 20 V, 12.4 keV, 6 × 1011 photons s−1 ). Fig. 8 shows linear scans of a 100 µm x-ray glow along the x-direction for 4H-SiC devices with different thicknesses. Although all 4H-SiC devices show superior signal uniformity compared with pc-diamond, it shall be mentioned that, in the encase of the very slender 4H-SiC XBPMs ( ≤1.1 µm ), no fully functional four-quadrant device was obtained. The rationality for such low concede can be the very thin low-doped n-layer, which prevents proper rectifying behavior .An external file that holds a picture, illustration, etc.
Object name is s-26-00028-fig8.jpgOpen in a separate window finally, the answer of the 1.24 µm 4H-SiC XBPM to variations in the photon flux is analyzed. The dynamic response of the XBPM is analysed by opening and closing the mechanical shutter of the x ray glow whereas the current response for different photon fluxes is studied by using different filters along the ocular path. The 4H-SiC XBPM shows much faster dynamics ( in the microsecond range, presently limited by the measurement frame-up ) compared with pc-diamond ( millisecond range ) [ see Fig. 9 ( a ) ]. In accession, it shows a analogue reception to photon flux density for more than four orders of magnitude, similar to a reference silicon diode [ see Fig. 9 ( b ) ] .An external file that holds a picture, illustration, etc.
Object name is s-26-00028-fig9.jpgOpen in a separate window

5. Conclusion  

In this newspaper, an drawn-out pretense and experiments-based comparison between 4H-SiC and diamond XBPMs is presented. Device simulations showed that, thanks to the lower electron–hole generation threshold energy, 4H-SiC has a potentially superior placement sensitivity compared with even sc-diamond XBPMs under equivalent transmission. They besides showed that, although 4H-SiC pin diodes can be operated at zero bias, they suffer from subscript care collection efficiency due to higher recombination rates in the p+ area and inferior position sensitivity due to the reduce thickness in the opening between the front collectors.

thermal simulations showed that even when the two materials absorb the lapp energy, with 4H-SiC at one-tenth of the thickness of sc-diamond, the higher absorb energy concentration in 4H-SiC results in a larger increase in temperature. however, this high temperature is found to be below the critical operational temperature of 4H-SiC ( 1200°C ) for the two experimental conditions : a 180 kilowatt cm−2 ( 2.1 % absorbed office ) focused synchrotron pink beam and a 1.4 mJ, 10 degree fahrenheit, 12.4 keV monochromated XFEL balance beam, thus allowing applications of 4H-SiC tied in the character of very gamey magnificence beams. preliminary experimental results based on the inaugural fabricated 4H-SiC XBPMs compared with a commercial 12 µm pc-diamond device ( RIGI, DECTRIS ) showed uniform foil across the device area. This excellent sign homogeneity shows the fabrication process to be suitable for producing highly uniform 4H-SiC membranes. Homogeneous and highly guileless membranes are prerequisites for precise beam-position monitor and enable fast on-line position stabilization procedures. In addition, superior signal-to-noise ratio proportion, one-dimensionality over four orders of magnitude of radio beam flux and fast dynamics ( ≤50 µs ), at the limit of the current measurement apparatus, is achieved. The lend possibility of operating the 4H-SiC XBPM without external biases potentially simplifies the sign process circuit. Given the obtained results and the adulthood of this wide-band-gap semiconductor device, we expect silicon carbide to substitute pc- and sc-diamond XBPMs in most beam-monitoring applications. ferment is presently under room to far characterize 4H-SiC devices in terms of charge collection efficiency, radiation hardness and dynamics down to the nanosecond government .

Leave a Reply

Your email address will not be published.