1. Introduction
Carbon atoms can be cracked down from hydrocarbons by photo-electrons emitted from irradiated optical surface in light source beamline, and the carbon atoms will constitute a film [1-3] on the optical surface. Carbon contamination is a serious problem to the beamline, especially in soft X-ray beamline. Carbon film deposits on a reflecting surface result in a substantial loss of flux and severe efficiency degradation at or above K edge, amounting to orders of magnitude in highly contaminated cases.
The Phase-II beamlines of Shanghai Synchrotron Radiation Facility (SSRF) will employ more wigglers and undulators, producing photon beams of much higher brightness than those from bending-magnet beamlines. This in turn results in more carbon contamination, as the deposition rate of contaminant film is proportional to photon flux on the optical components [4,5], especially for a soft X-ray beamline[6]. It’s essential to remove carbon contamination from these optical components.
The methods proposed to clean the beamline optics include RF discharge cleaning, synchrotron radiation cleaning[7-9], UV lamp cleaning in oxygen atmosphere, DC discharge cleaning, glow discharge cleaning[10,11], laser pulse cleaning[12,13], etc. RF discharge and synchrotron radiation cleaning are good choices for in situ beamline optics cleaning. In this work, RF discharge cleaning was used during the shutdown period of SSRF. The RF plasma (a mixture of oxygen and argon) produces a large quantity of oxygen radicals[14,15], which oxidizes carbon contaminate into CO and CO2. Then the reflection of mirrors increases [16] and beamline photon flux becomes higher.
Before cleaning the contaminated optics, key cleaning parameters were determined, such as the optimal RF output power, mixing rates of O2 and Ar, and chamber vacuum. Sample mirrors of Si substrate coated with carbon film were made and beamline optics cleaning experiments performed to determine the cleaning parameters, and to ensure that possible adverse effects will not happen while cleaning the real beamline optics. Roughness of the samples before and after cleaning was measured to check gentleness of the RF plasma removal of the carbon films. Also, possible detrimental kinetic effects of RF plasma on the Kapton films covering the wires were examined by checking conductivity of the Kapton film.
Using the optimal parameters, the RF plasma cleaning of SSRF beamline optics was carried out. Photodiodes were used to detect the beamline photon flux before and after cleaning. According to the photon flux improvement ratio, thickness of the removed carbon film was estimated.
2. Experimental
2.1 Experimental setup
As shown in Fig. 1, the experimental cleaning equipment mainly consists of the vacuum chamber, gas flow controller, liquid nitrogen tank, RF generator, vacuum pump and gauge, and residual gas analyzer (RGA). The vacuum chamber, sized at 1200 mm ×400 mm×400 mm (in similar size of a beamline optics chamber, is made of 304L stainless. It has different kinds of flange interface to install RF power, vacuum pump, vacuum gauge and RGA. Oxygen and argon release from the bottles are controlled by the gas flow controllers, so as to obtain the optimal mixing rates and the chamber pressure. The water and oil contents in the gas mixture passes are frozen off by liquid nitrogen, and the purified gas mixture arrives at the RF generator, which is a GV10x plasma source, in maximum RF power of 100 W [17,18], installed on the vacuum chamber. The chamber is pumped at the opposite end to establish a constant flux of oxygen plasma along its longitudinal axis.
Residual gas analyzer (RGA) was used to detect elemental ratio of the residual gases. The cleaning process should be accompanied by increasing carbon ratio. If other elements ratios increased, the feed stock gases would not be pure enough.
Experiments with carbon-coated samples were performed to determine the optimal O2-Ar rate, gas pressure and RF power, and to check safety of the cleaning process. These optimized parameters were the basis for installing the vacuum pump, vacuum gauge and RGA on the beamline optics chamber, and cleaning the beamline optics.
2.2 Sample preparation and characterization
Samples mirrors were prepared with Φ30 mm Si wafers. They were coated with 50 nm Au film and then 10, 30 or 50 nm carbon film, through chemical vapor deposition (CVD) at Tongji University. [19,20] As shown in Fig.2, the carbon films of different thicknesses are different colors (golden, red and blue for 10 nm, 30 and 50 nm films, respectively, due to the difference in visual light refraction of the carbon films) [21,22].
Zygo NewView600 white color interferometer was used to measure roughness of the samples. Most samples did not change in roughness before and after carbon-coating, with the RMS of <3Å (Fig.3). A few carbon-coated samples became worse in roughness, due to the non-uniformity of carbon film CVD.
2.3 Cable insulation experiments
Cables for adjusting the beam optics mirrors are insulated by Kapton films, which should not be damaged by the beam optics cleaning process. To confirm this, a sample cable was placed near the sample mirror, so as to suffer the same amount of RF plasma as the mirror. This cable resistance was measured before and after cleaning process.
3. Results and discussion
3.1. Cleaning experiments
3.1.1 Optimized cleaning parameters
The RF plasma cleaning parameters optimized through the experiments are summarized in Table1. The data almost coincide with those given in Ref. [1], except the RF power of 40–50W, due to probably the difference in chamber sizes. Ref. [17] suggested RF power of 150W. In our experiments, however, the RF generator always stopped work at ≥100W. Therefore, we reduced the RF power to 75W.
| Parameters | Values |
|---|---|
| RF power /W | 75 |
| Ratio of O2:Ar | 10:1 |
| Vacuum / mbar | 0.005–0.03 |
| Flow rate of O2-Ar gas mixture /sccm | ≤5 |
3.1.2 Roughness of samples before and after cleaning
RF plasma can remove carbon film, but violent removal of the carbon contamination film would damage the Au layer on the mirrors surface. Three samples (x13, C1 and d30-1) were used to measure the surface roughness before and after cleaning. The results are given in Table.2. Before and after RF plasma cleaning, Sample x13 and C1 showed no obvious changes in surface roughness, indicating that the beamline optics could stay safety in the RF plasma cleaning process. Sample d30-1 changed in carbon coating roughness, but its greater roughness before cleaning suggested non-uniform deposition of carbon film. After RF plasma cleaning, roughness of d30-1 became better again.
| Samples | Before (Å) | After (Å) | Carbon thickness (nm) |
|---|---|---|---|
| x13 | 2.3 | 2.8 | 30 |
| C1 | 2.3 | 2.6 | 10 |
| d30-1 | 4.3 | 2.5 | 10 |
In Ref. [14], RF plasma effects on roughness and reflectivity of vacuum ultraviolet-optics were studied. The sample roughness was measured by interferometer; scanning tunneling microscope (STM) and mechanical profile meter (Talystep) before and after cleaning. No changes in roughness were found. So RF plasma does not damage the mirror surfaces.
3.1.3 Measurement of cable resistance
The original cable resistance was 99.9 GΩ. After RF plasma cleaning for five times, in an accumulated time of ~200 hours, its resistance was still 99.9 GΩ, hence no RF plasma effect on cable insulation.
3.2. Cleaning of beamline optics
According to the optimal cleaning parameters in Table.2, M1 mirror of the soft X-ray beamline was cleaned. Beamline photon fluxes before and after cleaning were detected, and thickness of the removed carbon film could be estimated from the improvements in photon flux ratio.
3.2.1 Improvement in photon flux
The M1 mirror of the soft X-ray beamline was cleaned, twice during two shut down periods of times in an interval of about six months. Before and after shut down of the beamline, photon flux was measured with the same beam intensity, gap size of the undulator magnet pole, slit size, and grating. In the first shutdown, the mirror was cleaned for 20 h at the RF output power of 75 W, and vacuum pressure of 0.008 bar. The electron beam current before and after the shutdown was 240 mA. The photon flux after cleaning increased obviously (Table 3). In the second shut down, the mirror was cleaned for 50 h at the RF power of 75 W and vacuum pressure of 0.02 bar. At the beam current of 210 mA, the improvement of photon flux ratio was more obvious (Table 3).
| Photon energy(eV) | Gap (mm) | Φ B (μA) | Φ A(μA) | (ΦA−ΦB)/ΦB | ||||
|---|---|---|---|---|---|---|---|---|
| 1st | 2nd | 1st | 2nd | 1st | 2nd | |||
| Fe | ||||||||
| 723 | 82.5 | 2.20 | 0.52 | 3.92 | 3.33 | 78% | 540% | |
| 670 | 79.7 | 2.10 | 0.51 | 3.27 | 3.01 | 56% | 490% | |
| O | ||||||||
| 543 | 72.9 | 1.67 | 0.17 | 2.98 | 1.49 | 79% | 776% | |
| N | ||||||||
| 410 | 65.8 | 0.83 | 0.08 | 1.97 | 0.52 | 138% | 550% | |
| C | ||||||||
| 296 | 58.8 | 0.08 | 0.01 | 0.22 | 0.08 | 175% | 700% | |
| 284 | 57.8 | 0.05 | 0.01 | 0.08 | 0.05 | 60% | 400% | |
3.2.2 Evaluation of removed carbon film
Figure 4 shows the beamline photon flux as function of photon energy, after the second cleaning process. Before cleaning, the flux ratio of carbon edge to the flux without contamination was 3%, while the ratio after cleaning increased to 13%. According to the ratios, the thickness of removed carbon film can be estimated. The photon flux ratio can be calculated by I/I0 = e−ut, where I and I0 are flux of carbon edge from the mirrors with and without carbon contamination, respectively; u=1.1×10−5 cm−1 is the carbon edge absorbance coefficient; and t is the carbon film thickness.
Assuming that contamination of every mirror is proportional to the thermal power, from the grazing angle of M1 mirror, its carbon film thicknesses are 6 nm before cleaning and 2.8 nm after cleaning, i.e. the removal of 3.2-nm carbon film.
The second cleaning process continued for 50 h, thus the carbon removal rate is 0.011Å /min. The result is far smaller than that of Ref. [17], probably due to the fact that the RF power in this work is only half of that in Ref. [17]. Another reason is that the M1 mirror chamber differs in structure from the experimental cleaning system, on which the GV10x plasma source and the vacuum pump are installed at the same side of chamber. Thus, the oxygen plasma could not reach the mirror surface efficiently, hence the reduced carbon removal rate.
4. Conclusion
RF plasma cleaning of beamline optics was conducted on the experimental system and the M1 mirror chamber of a soft X-ray beamline at SSRF. Possible detrimental kinetic effects due to the RF plasma were studied. Under the RF power of 75 W, the O2/Ar mixing ratio of 10:1 and flow rate of ≤5 sccm, and the chamber pressure of 0.005–0.05 mbar, the SSRF beamline optics cleaning can be efficient, and safe, without obvious changes in roughness of the cleaned mirror surface, and in insulation of the cable for adjustment of the mirrors. Further efforts shall be made to improve the removal rate of carbon contamination.
Cleaning of contaminated XUV-optics at BESSY II
. Nucl. Instrum. Methods A 467-468(2001). doi: 10.1016/SO168-9002(01)00312-6Resuscitation of carbon-contaminated mirrors and gratings by oxgen-discharge cleaning. 2: Efficiency recovery in the 100-1000-eV range
. Appl. Opt. 27, 4305-4313(1988). doi: 10.1364/AO.27.004305Ultraviolet/ozone cleaning of carbon-contaminated optics
. Appl. Opt. 32, 4114-4118(1993). doi: 10.1364/AO.32.004114Resuscitation of carbon-contaminated mirrors and gratings by oxgen-discharge cleaning 1: Efficiency recovery in the 4-40-eV range
. Appl. Opt. 26, 3884-3894(1987). doi: 10.1364/AO.26.003884Evaluation of airborne molecular contamination inside the LIL
. Nucl. Instrum. Methods A 557, 676-683(2006). doi: 10.1016/j.nima.2005.11.16HERMES: a soft X-ray beamline dedicated to X-ray microscopy
. J. Synchrotron Rad. 22, 968-7979(2015). doi: 10.1107/S1600577515007778In situ removal of carbon contaminated from a chromium-coated mirror: ideal optics to suppress higher-order harmonics in carbon K-edge region
. J. Synchrotron Rad. 22, 1359-1363(2015). doi: 10.1107/S1600577515015040Cleaning of carbon contamination on Si wafer with activated oxygen by Synchrotron Radiation
. Acta Optica Sinica. 30, 907-910(2010). doi: 10.3788/AOS20103003.0907In situ removal of carbon contaminated from optics in a vacuum ultraviolet and soft X-ray undulator beamline using oxygen activated by Zeroth-order synchrotron radiation
. J. Synchrotron Rad. 19, 722-727(2012). doi: 10.1107/S0909049512024971In-situ optical components cleaning by glow discharge
. Vacu. Sci. Tech. 21, 147-149(2001). doi: 10.13922/j.cnki.cjovst.2001.02.016Self-enhanced plasma discharge effect in the deposition of diamond-like carbon films on the inner surface of slender tube
. Appl. Surf. Sci. 393, 467-473(2017). doi: 10.1016/j.apsusc.2016.10.049Cleaning of carbon layer from the gold films using a pulsed Nd:YAG laser
. Appl. Surf. Sci. 283, 612-616(2013). doi: 10.1016/j.apsusc.2013.06.157Evaluation of laser cleaning for the restoration of tarnished silver artifacts
. Appl. Surf. Sci. 387, 118-127(2016). doi: 10.1016/j.apsusc.2016.06.017Cleaning of carbon contaminated vacuum ultraviolet-optics: influence on surface roughness and reflectively
. Rev. Sci. Instrument. 63, 1428-1431(1992). doi: 10.1063/1.1143034Electrical breakdown characteristics of an atmospheric pressure RF capacity plasma source
. Phys. Plasmas. 12, 093504(2005). doi: 10.1063/1.2035447Carbon contamination of soft X-ray beamlines: dramatic anti-reflection coating effects observed in the 1 keV photon energy region
. J. Synchrotron Rad. 18, 761-764(2011). doi: 10.1107/SO0909049511023119Remote plasma cleaning of optical surface: Cleaning rates of different carbon allotropes as a function of RF powers and distances
. Appl. Surf. Sci. 362, 448-458(2016). doi: 10.1016/j.apsusc.2015.11.117Characterization, optimization and surface physics aspects of in situ plasma mirror cleaning
. J. Synchrotron Rad. 21, 300-314(2014). doi: 10.1107/S1600577513032402PECVD synthesis of 2D nanostructured carbon material
. Surf. Coat. Technol. 211, 2-8(2012). doi: 10.1016/j.surfcoat.2011.07.092Selective growth of diamond by hot filament CVD using patterned carbon film at mask
. Nucl. Sci. Tech. 19,83-87(2008). doi: 10.1016/S1001-8042(08)60028-7Influence of deposition time on chromatics during nitrogen-doped diameter like carbon coating on pure titanium
. West China J Stomatol. 28, 543-546(2010). doi: 10.3969/j.issn.1000-1182.2010.05.023Direct evaluation of the sp3 content in diamond-like carbon films by XPS
. Appl. Surf. Sci. 136, 105-110(2010). doi: 10.1016/S0169-4332(98)00319-5