Subpicosecond hard x-ray streak camera using single-photon counting

Time-resolved x-ray diffraction has become a standard tool for studies of the dynamics of laser excited solids [1]. Synchrotron radiation sources operating in standard mode produce high brilliance x-ray beams, but the time duration of the pulses is in the $50–500\mathrm{ps}$ range. Any experiment requiring better time resolution will need to rely on a fast detector. Streak cameras have been reported to give a time resolution down to $233\mathrm{fs}$ [2] for UV radiation in the accumulating mode and $350\mathrm{fs}$ using $1.5\mathrm{keV}$ x rays [3] in the single-shot mode. This time resolution is, however, not yet reached for hard x rays. The fastest streak cameras rely on limiting the divergence of the photoelectrons [4]. This implies a severe reduction in quantum efficiency.

In time-resolved x-ray diffraction experiments carried out at third-generation electron storage rings, the number of photons per pulse is relatively small and operation in the accumulating mode is necessary. The mechanisms limiting the time resolution of a streak camera are described in detail elsewhere [3, 4, 5]. The main factors are the sweep speed, the size of the input slit, the aberrations of the electron optics, and the energy spread of the photoelectrons. The field generated by the sweep plates also introduces additional dispersion [4, 6].

In this Letter we describe an x-ray streak camera that uses single-photon counting to reduce the effects of dispersion and imperfect imaging. Images are analyzed in real time, and a UV timing reference is used to track and compensate the sweep jitter. The use of a readout camera with built-in image processing capabilities reduces the bandwidth required to transfer data to the host computer. Thus a frame rate of $2\mathrm{kHz}$ can be reached. The concept of a single-photon counting camera was first proposed by Murnane et al. [7], and a proof-of-principle demonstration was published by Larsson [8], who improved the time resolution of a commercial streak camera from 5 to $1.5\mathrm{ps}$.

A set of experiments to demonstrate the performance of the streak camera was conducted at beamline D611 at the MAX-lab synchrotron radiation facility. This bending magnet beamline is designed for laser pump–x-ray probe experiments and produces about 5000 photons per pulse and 0.02% bandwidth. The pulse duration is $500\mathrm{ps}$. A $\mathrm{Ti}\text{:}{\mathrm{Al}}_2{\mathrm{O}}_3$ femtosecond laser system, operating at a repetition rate of $4.25\mathrm{kHz}$, with a $790\mathrm{nm}$ center wavelength and $45\mathrm{fs}$ pulse duration was used for the measurements. The laser beam was split into three arms. Laser pulses with an energy of $100\mathrm{\mu J}$ were used to trigger a photoconducting switch that generates the high voltage sweep ramp for the streak camera. Part of the pulse ($200\mathrm{\mu J}$) was used to generate the third harmonic, which was sent onto the photocathode as a timing reference. Up to $700\mathrm{\mu J}$ was available to pump the sample. The streak camera was mounted with a 100-μm-wide CsI photocathode in a $6°$ grazing incidence configuration [9]. Electrons were accelerated, using a mesh with a $>90\%$ open area ratio, to $8\mathrm{kV}$ over a gap of $2\mathrm{mm}$. Subsequently, electrons were imaged from the cathode to a microchannel plate (MCP) with gain larger than ${10}^8$, enabling a single photoelectron emitted from the cathode to be detected by the readout camera. Thus, the overall quantum efficiency of the streak camera equals that of the photocathode. The quantum efficiency was deduced by measuring the photon flux using a calibrated x-ray diode and comparing it with the single-photon count rate and was found to be larger than 10%. The readout camera (Mikrotron MC1364) uses a complementary metal-oxide semiconductor sensor and has an embedded FPGA image processor that can perform a large part of the image analysis. This enables images to be analyzed in real time at frame rates of several kilohertz.

We will now discuss how single-photon counting in the accumulating mode can increase the temporal resolution compared to the single-shot mode [7, 8]. The x rays are sent onto the photocathode together with a UV pulse derived from the same laser that is used to pump the sample. Each incident x-ray photon generates up to 20 photoelectrons [10]. If the aperture of the streak camera is not constrained, many of the photoelectrons can propagate to the MCP. Because of dispersion and imaging errors, electrons will travel individual trajectories and be imaged as a spot that can be irregular in shape. The idea behind single-photon counting is that the center of each spot can be determined with an accuracy smaller than the spot size. The uncertainty of the center position depends on the number of detected photoelectrons per photon, and, hence, it is important to keep as many as possible to minimize statistical variations. Typically four photoelectrons per x-ray photon are detected. Each sweep generates an image in which all spots corresponding to photon events are found, and their center of gravity is determined. The UV pulse consists of several hundred photoelectrons, and its position can be determined with high precision and is used as a timing fiducial. As this pulse has a known fixed delay relative to the pump pulse, the position of this pulse can be used to remove the shot-to-shot jitter in the processed data.

To test the system, a second UV pulse was generated by splitting off part of the power in the reference beam. This second beam was then used as a simulated x-ray signal at a fixed delay of $8\mathrm{ps}$. Figure 1a shows a gray scale image averaged for 1 s at a repetition rate of $4.25\mathrm{kHz}$. The time resolution in this particular image is $2.5\mathrm{ps}$. When the pulses are accumulated using the single-photon counting mode, the time resolution measured as the FWHM is $280\mathrm{fs}$. The vast improvement can be seen in Fig. 1b.

In the following we demonstrate the effective time resolution of our streak camera for hard x-ray pulses by repeating three previously performed ultrafast diffraction experiments.

The structural rearrangements associated to ultrafast melting of semiconductors, such as InSb, has previously been studied [11, 12]. The disordering can be detected as a fast drop of x-ray diffraction efficiency [12]. An asymmetrically cut InSb sample was illuminated with laser pulses with a fluence of $38\mathrm{mJ}/{\mathrm{cm}}^2$. The incidence angle between laser and sample surface was $15°$. The disordering was probed by x rays with a photon energy of $3.15\mathrm{keV}$ at an incidence angle of $0.9°$ and a bandwidth of 2%. The intensity of the $(111)$ reflection was recorded. The sample was continuously translated to exchange the surface in order to avoid effects from ripple formation [13]. At the translation speed of $1\mathrm{mm}/\mathrm{s}$, the ripples did not reach a high enough amplitude to influence the measurement. This limited the data acquisition time to $30\mathrm{s}$. Figure 2 shows the drop in x-ray diffraction as recorded by the streak camera in the single-photon counting mode. The curve was fitted to an error function, yielding an upper bound of $640\mathrm{fs}$ for the 90% to 10% fall time. The fall time of the $(111)$ reflection has been measured at $430\mathrm{fs}$ [12]. A quadratic deconvolution yielded the time resolution of the streak camera to be $480\mathrm{fs}$.

Laser excitation of bismuth (Bi) leads to the excitation of optical phonons [14]. The $(111)$ reflection in Bi is very close to being forbidden. Excitation of the ${\mathrm{A}}_{1\mathrm{g}}$ phonon mode induces a drop in the x-ray reflectivity followed by an oscillation. The period of the oscillation depends on temperature and excitation strength and is $340\mathrm{fs}$ for weak excitation [15] and $467\mathrm{fs}$ for stronger excitation [14]. A symmetric thin-film Bi sample was illuminated by laser pulses with a fluence of $4\mathrm{mJ}/{\mathrm{cm}}^2$ at an incidence angle of $45°$. The phonons were probed by x rays with a photon energy of $3.2\mathrm{keV}$. Figure 3 shows the drop in x-ray reflectivity associated with the excitation of the optical phonons. The accumulation time was $40\min$. Fitting an error function yields an upper bound for the time resolution of $660\mathrm{fs}$, which is large compared to the expected $170\mathrm{fs}$ fall time.

When a superlattice of ${\mathrm{SrTiO}}_3$ (STO) and ${\mathrm{SrRuO}}_3$ (SRO) is excited by a short laser pulse, large-amplitude coherent acoustic superlattice phonons can be generated. For an excited superlattice consisting of ten bilayers of $17.9\mathrm{nm}$ of STO and $6.3\mathrm{nm}$ of SRO, the (0 0 116) reflection shows a strong oscillating reduction of the reflectivity, with a period of $3\mathrm{ps}$ [16]. The sample was excited at a fluence of $30\mathrm{mJ}/{\mathrm{cm}}^2$, and the phonons were probed by x rays with a photon energy of $5.8\mathrm{keV}$, and a bandwidth of $2\times {10}^{-4}$. Directly after excitation the reflectivity drops, followed by a damped oscillation, as shown in Fig. 4. This oscillation was compared to measurements done using a laser plasma x-ray source at a time resolution of $\sim 200\mathrm{fs}$ [17]. The comparison shows that the amplitude of the oscillation is unaffected by the temporal resolution of the streak camera. This is consistent with a resolution in the $400–600\mathrm{fs}$ range, as found in the studies of nonthermal melting and optical phonons in Bi. For a temporal resolution lower than $2\mathrm{ps}$, the oscillation amplitude of the convoluted data is reduced, as illustrated in Fig. 4.

In conclusion, we have developed a hard x-ray streak camera capable of achieving subpicosecond resolution using single-photon counting. The performance has been demonstrated by reproducing three well-characterized experiments. The temporal spread due to jitter is removed by using a UV reference, and the single-photon counting compensates for smearing due to imperfect imaging and energy spread of the photoelectrons.

Streak cameras with subpicosecond time resolution will have a significant impact on the field of time-resolved science at synchrotron radiation facilities. They will also play a role at x-ray free-electron lasers. When lasers are synchronized to the accelerator it is essential to have a user-controlled diagnostic to track the delay between laser and x rays. By using a streak camera this can be measured directly at the position of the sample.

The authors thank the Swedish Research Council (VR), the Knut and Alice Wallenberg Foundation, the Crafoord Foundation, and the Carl Trygger Foundation for financial support. We also acknowledge the support of the European Commission via the Marie-Curie Program and the IRUVX-PP project.

 

Fig. 1 Images showing two UV pulses separated by $8\mathrm{ps}$. (a) Gray scale image of averaging mode and (b) lineout in averaging mode (dashed curve) and photon counting mode (solid curve). In the photon counting mode, the smearing effects of jitter and imaging are removed, yielding a time resolution of $280\mathrm{fs}$.

Download Full Size | PPT Slide | PDF

 

Fig. 2 Time-resolved drop in x-ray diffraction induced by nonthermal melting of InSb. The dashed line shows the fitted error function.

Download Full Size | PPT Slide | PDF

 

Fig. 3 Time-resolved diffraction of bismuth following laser excitation. The fitted error function is shown as a dashed line.

Download Full Size | PPT Slide | PDF

 

Fig. 4 Oscillations in the x-ray reflectivity of the ${\mathrm{SrTiO}}_3/{\mathrm{SrRuO}}_3$ superlattice as recorded by the streak camera (solid curve) and from [16] (dashed curve). The dotted curve shows data from [16] filtered to simulate a time resolution of $2\mathrm{ps}$.

Download Full Size | PPT Slide | PDF

1. D. Reis, K. Gaffney, G. Gilmer, and B. Torralva, Mat. Res. Bull. 31, 601 (2006). [CrossRef]  

2. J. Feng, H. J. Shin, J. R. Nasiatka, W. Wan, A. T. Young, G. Huang, A. Comin, J. Byrd, and H. A. Padmore, Appl. Phys. Lett. 91, 134102 (2007). [CrossRef]  

3. P. Gallant, P. Forget, F. Dorchies, Z. Jiang, J. C. Kieffer, P. A. Jaanimagi, J. C. Rebuffie, C. Goulmy, J. F. Pelletier, and M. Sutton, Rev. Sci. Instrum. 71, 3627 (2000). [CrossRef]  

4. M. M. Shakya and Z. Chang, Appl. Phys. Lett. 87, 041103 (2005). [CrossRef]  

5. D. Bradley, A. Roddie, W. Sibbett, M. Key, M. Lamb, C. Lewis, and P. Sachsenmaier, Opt. Commun. 15, 231 (1975). [CrossRef]  

6. G. Huang, J. Byrd, J. Feng, H. A. Padmore, J. Qiang, and W. Wan, in Proceedings of EPAC 2006, Edinburgh, Scotland (2006), pp. 1250–1252.

7. M. Murnane, H. Kapteyn, and R. Falcone, Appl. Phys. Lett. 56, 1948 (1990). [CrossRef]  

8. J. Larsson, Opt. Lett. 26, 295 (2001). [CrossRef]  

9. , Rev. Sci. Instrum. 75, 3131 (2004).

10. C. Ortiz and C. Caleman, J. Phys. Chem. C 111, 17442 (2007). [CrossRef]  

11. A. Rousse, C. Rischel, and J. Gauthier, Rev. Modern Phys. 73, 17 (2001). [CrossRef]  

12. A. Lindenberg, J. Larsson, K. Sokolowski-Tinten, K. Gaffney, C. Blome, O. Synnergren, J. Sheppard, C. Caleman, A. MacPhee, D. Weinstein, D. Lowney, T. Allison, T. Matthews, R. Falcone, A. Cavalieri, D. Fritz, S. Lee, P. Bucksbaum, D. Reis, J. Rudati, P. Fuoss, C. Kao, D. Siddons, R. Pahl, J. Als-Nielsen, S. Duesterer, R. Ischebeck, H. Schlarb, H. Schulte-Schrepping, T. Tschentscher, J. Schneider, D. von der Linde, O. Hignette, F. Sette, H. Chapman, R. Lee, T. Hansen, S. Techert, J. Wark, M. Bergh, G. Huldt, D. van der Spoel, N. Timneanu, J. Hajdu, R. Akre, E. Bong, P. Krejcik, J. Arthur, S. Brennan, K. Luening, and J. Hastings, Science 308, 392 (2005). [CrossRef]   [PubMed]  

13. , Appl. Phys. A 100, 105 (2010).

14. K. Sokolowski-Tinten, C. Blome, J. Blurns, A. Cavalleri, C. Dietrich, A. Tarasevitch, I. Uschmann, E. Forster, M. Kammler, M. Horn-von Hoegen, and D. von der Linde, Nature 422, 287 (2003). [CrossRef]   [PubMed]  

15. M. Hase, K. Mizoguchi, H. Harima, S. Nakashima, M. Tani, K. Sakai, and M. Hangyo, Appl. Phys. Lett. 69, 2474 (1996). [CrossRef]  

16. C. von Korff Schmising, M. Bargheer, M. Kiel, N. Zhavoronkov, M. Woerner, T. Elsaesser, I. Vrejoiu, D. Hesse, and M. Alexe, Appl. Phys. B 88, 1 (2007). [CrossRef]  

17. C. von Korff Schmising, A. Harpoeth, N. Zhavoronkov, Z. Ansari, C. Aku-Leh, M. Woerner, T. Elsaesser, M. Bargheer, M. Schmidbauer, I. Vrejoiu, D. Hesse, and M. Alexe, Phys. Rev. B 78, 060404 (2008). [CrossRef]