|
The availability of millijoule-pulse-energy kHz repetition rate femtosecond solid state Ti:Sapphire lasers enables fs transient absorption (TA) to be a standard probe of ultrafast electronic and vibrational processes.
The Center for Photoinduced Charge Transfer houses a SpectraPhysics Tsunami Mode-Locked Ti:Sapphire Laser. It currently delivers a range of wavelengths from 720nm to 850nm, although with propoerly chosen optics it is capable of 690-1080nm. It also gives a range of pulse lengths continuously variable from 80ps to <50fs.
The Tsunami uses an acousto-optic modulator (AOM) to ensure an 82MHz mode-locked operation without the dropouts or shut-downs associated with standard passive mode-locking systems.
The Tsunami is currently pumped by the 5Watt TEM00 output of a solid state Millenia Diode-Pumped Nd:YAG laser. The output of the Tsunami with this pumping is optimized at 450mW average power.
This output is directed into a SpectraPhysics (Positive Light) Spitfire Regenerative Amplifier which is pumped by a Merlin Q-Switched Neodymium:YLF (Nd:YLF) laser of 10W average output and a repetition rate of 1KHz.
The Tsunami output (the "780nm") is stretched initially in the Spitfire in 16passes of a grating-mirror system such that the ultrashort pulses are rendered incapable of damage to the regenerative amplifier oscillator. A high voltage is applied to a pockels cell timed to allow the injection of the 780nm at the same time as a pulse (microseconds) from the Merlin pumps a second Titanium:Sapphire rod. The lifetime of the Ti:Sapphire excited state is long and the gain induced is depleted by several round trips of the cavity by the injected 780nm light. Upon each round trip this 780nm light is amplified giving an increase in the observed pulse energy until the majority of the gain is depleted and subsequent losses in the cavity result in pulses of later round trips having a lower observed pulse energy. Subsequent to a time of maximum depletion of gain from the rod and close to maximum pulse energy a second pockels cell is activated by application of a second high voltage and the amplified stretched pulse is ejected from the cavity. This pulse passes through a pulse compressor to regain a pulsewidth of ca. 100fs and is directed towards an Optical Parametric Amplifier (OPA). The Pulse energy at this point has recently been obtained as 700microjoules per pulse although a fully optimized 1mj of pulse energy is achievable.
The 780nm Regen output is directed towards the double pass OPA. It passes through a beam splitter which directs some 7% of the light towards a sapphire plate where white light is generated and collimated towards a BBO (beta Barium Borate) crystal. The 780nm light is split again such that a further 10% is sent to join the white light in going through the BBO crystal. These beams are aligned spatially (colinearly) and temporally to generate the white-light seeded signal and idler, the result of the optical parametric effect, the wavelengths of these are determined by energy conservation and momentum conservation (through phase matching) in the crystal. The signal is blocked and the idler is retroreflected back into the BBO to become colinear with the 83% of the 780nm light that remains. Again these beams are aligned spatially and temporally, and the idler in this case acts as the seed for the optical parametric effect yielding several hundred microjoules of signal and idler determined by the equation
1 / fundamental (780nm) = 1 / signal + 1 / idler.
These outputs, in the near to mid IR are then either i) frequency doubled of ii) sum-frequency mixed with the residual fundamental (780nm) to yield wavelengths in the visible. Pulse energy varies.
Before frequency mixing some of the residual 780nm light is split off in order to generate the probe beam. The 780nm light passes round a delay stage driven by a computer and stepper motor before being focused into a sapphire plate, again to generate a white light continuum via self-phase modulation. The energy per pulse for this process is moderated by a further beam splitting wedge and a wave-plate and polarizer combination to achieve, as far as possible, a single filament temporally and spatially stable continuum. Stability does fundamentally drift in and out as a function of wavelength but wavelengths available for experimentation range from ~400-900nm determined by previous detection. Wavelengths beyond this are accessible but current PMT detectors need to be swapped out for more IR sensitive photodiodes.
The OPA sum-frequency mixed output is invariably the pump beam for the pump-probe TA experiment but doubled fundamental (output from Regen doubled to give 390nm, high energy per pulse) or difference-frequency-mixed far-IR can be used. We have the equipment to generate such wavelengths.
The pump beam passes through a chopper set to half the frequency of the repetition rate of the Regen output (500Hz). This provides subsequent "pump-on" and "pump-off" conditions in the sample. As the white light passes through the excited region (where spatial overlap between pump and probe occurs) the white light will experience an absorption dependent on the presence of the pump. A monochromator selects the white light wavelength for observation.
For example in the case of wavelength selection in a range where;
1) the sample's ground state absorbs, the high flux pump will cause a bleach in the ground state. With "pump-on" the depleted ground state will look to be absorbing less light. Therefore more probe light will be transmitted and the measured "Delta-T-over-T" (the change in transmission as a fraction of transmission) will be positive.
2) Fluorescence is usually observed from an excited sample, the pump will cause a high concentration of excited state species which the probe will stimulate to emit. This will show up as if more light is being transmitted through the sample on account of the gain yielding correlated photons. This, too, will give rise to a positive "Delta-T-over-T" (change in transmission over transmission).
3) The excited state itself absorbs, (photoinduced absorption), the probe beam will be attenuated further by the high concentration of excited states thus yielding a negative "Delta-T-over-T".
The current set-up has a delay line of about a meter length which, in principal, gives a temporal delay range for the probe beam of 6ns. In practice, since time-zero invariably occurs in the first nanosecond and as a result of hard to remove spatial misalignments from 0-6ns delays an experiment may only yield accurate "Delta-T-over-T" results in the first 2 or 3 ns. If one is prepared to go through a long and painstaking alignment procedure of the delay line retro-reflector this range can be increased.
Further range can be achieved by delaying the probe beam relative to the pump by sending it through an optical fiber of known length.
Data processing occurs by way of three SRS SR250 gated integrator/boxcar averagers and an SR235 Analog Processor.
The SR250 consists of a gate generator, a fast gated integrator, and exponential averaging circuitry. The gate generator, triggered externally by the 500Hz Chopper frequency, provides an adjustable delay from a few nanoseconds to 100 milliseconds before it generates a continuously adjustable gate with a width between 2 ns to 15 us.
The fast gated integrator integrates the input signal during the gate. The output from the integrator is then normalized by the gate width to provide a voltage proportional to the average of the input signal during the sampling gate. This signal is further amplified and sampled by a low droop sample and hold amplifier, and output via a front panel BNC connector. The last sample output allows for a shot-by-shot analysis of the signal.
Signal Input.
The sensitivity of the instrument (volts out / volts in) may be set from 1V/1V to 1V/5mV. The input has a 1 MegOhm input impedance. An input filter can reject unwanted signals before the input is sampled by the integrator, however the filter is normally left in the "off" position.
Gate Timing.
The delay of the sample gate from the trigger is set by the delay multiplier and scale. The delay scale is multiplied by the setting on the 10-turn multiplier dial, allowing continuously adjustable delays from a few nanoseconds to 100 milliseconds.
The width of the sampling gate may be continuously adjusted from 2 ns to 15 ms over 8 width ranges. A simple modification of the unit allows gate widths of up to 150 ms. The front panel gate output provides a representation of the gate that can be overlayed with the signal on an oscilloscope to provide a precise display of the gate timing.
Signal Outputs
A moving exponential average of 1 to 10,000 samples can be selected from the front panel. This traditional averaging technique is useful for pulling small signals from noisy backgrounds. In the case of a random white noise background, the signal-to-noise ratio increases as the square root of the number of samples in the average. This allows a S/N improvement of up to a factor of 100 using this technique alone. If no averaging is desired, or if averaging is to be performed on an external computer, the last sample output provides a voltage proportional to the average value of the input signal during the last gate period.
The reset button sets the average output to zero.
Polarity Control and Active Baseline Subtraction
The polarity of the last sample and averaged outputs is controlled by rear panel toggle switches. Positive outputs can be selected for negative signals, and vice versa, allowing easy interfacing with unipolar analog to digital conversion systems. In addition to the traditional averaging modes, the SR250 possesses a unique Active Baseline Subtraction mode which allows you to actively cancel baseline drift. In the Active Baseline Subtraction mode the SR250 is triggered at twice the source repetition rate. On alternate triggers, when the signal is not present only the baseline is sampled and the SR250 inverts the polarity of the last sample output before it is added to the moving average. Thus, any baseline drift not associated with the source will be subtracted out.
Additional Outputs
The signal input is passed on to the signal output by a length of coaxial cable for termination and for gate timing. It is delayed exactly 3.5 ns from the input and can be terminated to optimize either signal gain or response time. The gate output provides a pulse synchronized with the internal gate signal. The gate output is timed so that it can be overlayed with the signal output for precise adjustment of gate timing. The busy output provides a TTL timing pulse which is high while the unit is integrating, and goes low when the SR250 is ready to accept another trigger. These outputs help simplify experimental setup and troubleshooting.
|
SR250 # |
1 |
2 |
3 |
|
Trigger |
External 500Hz |
External 500Hz |
External 500Hz |
|
Delay |
100ns |
100ns |
100ns |
|
Delay multiplier |
8.61 |
8.60 |
10.02 (MAX) |
|
Width |
30ns |
30ns |
3ms |
|
Width Multiplier |
5 |
5 |
5 |
|
Signal Sensitivity |
0.2V |
0.1V |
0.2V |
|
Filter |
10Hz AC |
10Hz AC |
10Hz AC |
|
Input Offset Used |
Yes |
Yes |
Yes |
|
Averaged Output? |
No |
To PC, 300 Samples |
To PC, 300Samples |
|
Last Sample Output? |
To SR235, Channel A |
To SR235, Channel B |
No |
Analog Processor
The SR235 Analog Processor provides a variety of convenient signal processing functions on up to two inputs. For example, background subtraction, ratioing, and logarithmic compression are a few of the functions that can be implemented.
The SR235 outputs a voltage proportional to a function of an argument formed from its two inputs (A and B). Allowable arguments are: A, B, SQRT(A2 + B2), A-B, A´B/10, and 10A/|B|. The functions that can be selected are: x, x2, SQRT x, ln|x|, -dx/dt, and -(dx/dt)/100. In our case the 10A/|B| is used as the argument where A is the "last sample" output of boxcar 1 and B is the "last sample" output of boxcar 2. The function ln|x| is used to act upon the argument. Filtering can be performed on the argument with time constants from 0.3 ms to 30 s. The final function output can be amplified with gains of 0.1 to 20 in a 1-2-5 sequence within the SR235's linear output range of ±10 V.
Absorption =
taking account of the switch from base 10 to base e, the switch from moles to molecules and the use of units in cm (sigma being the cross-section given in cm2 and N being number of molecules per cm3).
This is Beer's law and is usually written for the absorption of a ground state to an excited state where I0 is the transmitted light with no ground state sample present in the beam and IT the transmitted light with ground state sample in the beam.
For the photoexcited states, I0 becomes the transmitted light through the sample with no pump; i.e. there are no photoexcited states present and I0 becomes It,nopump. Correspondingly IT becomes It,pump as if with the photoexcitation we are suddenly throwing a sample into the absorption spectrometer. Therefore the photoinduced absorption becomes
Photoinduced Absorption = 
[1]
Beer's Law: 
[2]
Therefore: 
[3]