Crystallization kinetics of organic-inorganic trihalide perovskites and ...
Crystallization kinetics of organic-inorganic trihalide perovskites and the role of the lead anion in crystal growth David T. Moore, Hiroaki Sai, Kwan W. Tan, Detlef‐M. Smilgies, Wei Zhang, Henry J. Snaith, Ulrich Wiesner, and Lara A. Estroff
Figure S1. 1D radial plots of scattering images from the chloride system integrated over all azimuthal angles. Traces are taken at ~10 min. intervals and are shown prior to background subtraction to demonstrate the consistency of the background signal. Black dashed line is the background signal fit for the trace at t=50 min. superimposed on top of all traces for comparison.
S1
PbI2
PbCl2
S2
PbAc2
Pb(NO3)2
Figure S2. 1D and 2D scattering data for the precursor and perovskite material for the iodide (a,b), chloride (c,d), acetate (e,f) and nitrate (g,h) systems. (a,c,e,g) show the integrated plots of the precursor (top) and perovskite (bottom) along with the calculated peak locations for the tetragonal iodide perovskite (black stick markers). (b,d,f,h) are the raw 2D plots from which the respective 1D plots were generated. The iodide and acetate systems were made on TiO2 substrates and the peak locations for the TiO2 are marked with red “X”s on both the 1D and 2D plots. S3
Figure S3. UV‐Vis absorption for optimally annealed films made from all four lead sources, PbI2 (blue), PbCl2 (black), PbAc2 (red), and Pb(NO3)2 (green); traces are offset on the y‐axis for clarity, graph is adapted from reference 1. The data for Pb(NO3)2 was taken on a Cary 5000 UV‐Vis‐NIR spectrometer and was added subsequent to the publication of ref. 1; the Pb(NO3)2 films were much thinner (optically) and the spectra has been normalized to the absorbance at 750 nm for easy comparison.
S4
Figure S4. TGA plot showing residual solvent prior to the precursor‐to‐perovskite transition for a thin film using PbCl2 as the lead source. Black line (left axis) is the relative mass of the film only (TGA was tared with the substrate), red line (right axis) is the temperature. The point at which the transition of precursor to perovskite begins is approximately 130 minutes and is marked with a vertical, dotted line. To account for solvent loss during the transfer into the TGA, 10 films were spincast and immediately weighed, the average mass was used as the starting mass for the TGA samples. The salt only weight is the total mass of the PbCl2 and MAI (horizontal dashed line at 0.45 Rel. Mass) and is calculated by assuming the difference between the starting mass, and the first mass reported by the TGA, is due to solvent loss only, and the salt concentration calculated accordingly. The difference between the solid black line (TGA trace) and the dotted horizontal line is taken to be residual solvent that remains in the film.
Calculation of pathway process bars used to create Figure 6 in the main text.
Time [au]
S5
Evap Solvent
From TGA we know how long it takes DMF to evaporate, we know DMF isn’t in the lattice so we set the start of this to t=0 and the length of time to the value derived from TGA data; we assume the lead salt doesn’t change the evaporation rate of the solvent.
Diffuse (MA)Sp
This process is the removal of the spectator from the precursor and is the precursor‐to‐perovskite transition. Both the lag time and the total time for this process are calculated from the WAXS data and the kinetic parameters.
Sublime (MA)Sp
This is the removal of the “amorphous” spectator salt from the film (not the lattice), and has recently been confirmed to occur by sublimation.2 The start time is set based on the assumption that sublimation begins as soon as there is amorphous (MA)Sp to be removed, so it lags the diffusion by a short time in all cases. This process can’t end until the diffusion is complete. Once diffusion is complete, in all cases we have a perovskite film with some (MA)Sp laying around, so we set the time between completion of diffusion and the end of sublimation based on the volatility of (MA)Sp. The volatility of (MA)I, (MA)Cl, and (MA)Ac was measured by TGA and reported in reference 1, Figure 3, the relative position of (MA)NO3 is taken by comparison of the melting temperatures of (MA)NO3 and (MA)Ac which are 110 °C and 80 °C, respectively.
Diffuse (MA)I
This is the removal of (MA)I from the perovskite leading to the decomposition. The case of all iodide is easiest to consider as (MA)Sp=(MA)I; therefore, we assume that the diffusion of (MA)I from the lattice (precursor or perovskite) is continuous. The delineation between the desirable process and the decomposition is just a matter of counting the number of (MA)I units removed from the lattice; thus, the start of decomposition coincides with the end of the Diffuse (MA)Sp step. The case of non‐iodide (MA)Sp is the most difficult to assess. For chloride it is based on experimental observation. For acetate and nitrate, the timing of the first three processes suggests that all three are complete before decomposition begins. Therefore we select a lag time for the start of decomposition at 100 °C and set it the same for both systems. Note: the length of this bar on the chart is random, it continues until only PbI2 is left.
Growth
Time available for perovskite grain growth, which begins when the transformation begins and ends when the (MA)Sp is completely removed from the lattice OR when the solvent is removed, whichever comes first. i.e. growth only occurs while both precursor and solvent still exist.
Coarsen
Time available for coarsening. This begins when (MA)Sp is done diffusing from the lattice and ends when decomposition begins.
S6
References 1. Zhang, W.; Saliba, M.; Moore, D.T.; Pathak, S.K.; Hörantner, M.T.; Stergiopoulos, T.; Stranks, S.; Eperon, G.E.; Webber, J.A.; Abate, A.; Sadhanala, A.; Yao, S.; Chen, Y.; Friend, R.H.; Estroff, L.A.; Wiesner, U.; Snaith, H.J. Ultra‐smooth organic‐inorganic perovskite thin‐film formation and crystallization for efficient planar heterojunction solar cells. Nat. Commun. 2014. doi: 10.1038/ncomms7142 2. Unger, E. L.; Bowring, A. R.; Tassone, C. J.; Pool, V.; Gold‐Parker, A.; Cheacharoen, R.; Stone, K. H.; Hoke, E. T.; Toney, M. F.; McGehee, M. D. Chem. Mater. 2014, doi: 10.1021/cm503828b
S7
Figure S1. 1D radial plots of scattering images from the chloride system integrated over all azimuthal angles. Traces are taken at ~10 min. intervals and are shown prior to background subtraction to demonstrate the consistency of the background signal. Black dashed line is the background signal fit for the trace at t=50 min. superimposed on top of all traces for comparison.
S1
PbI2
PbCl2
S2
PbAc2
Pb(NO3)2
Figure S2. 1D and 2D scattering data for the precursor and perovskite material for the iodide (a,b), chloride (c,d), acetate (e,f) and nitrate (g,h) systems. (a,c,e,g) show the integrated plots of the precursor (top) and perovskite (bottom) along with the calculated peak locations for the tetragonal iodide perovskite (black stick markers). (b,d,f,h) are the raw 2D plots from which the respective 1D plots were generated. The iodide and acetate systems were made on TiO2 substrates and the peak locations for the TiO2 are marked with red “X”s on both the 1D and 2D plots. S3
Figure S3. UV‐Vis absorption for optimally annealed films made from all four lead sources, PbI2 (blue), PbCl2 (black), PbAc2 (red), and Pb(NO3)2 (green); traces are offset on the y‐axis for clarity, graph is adapted from reference 1. The data for Pb(NO3)2 was taken on a Cary 5000 UV‐Vis‐NIR spectrometer and was added subsequent to the publication of ref. 1; the Pb(NO3)2 films were much thinner (optically) and the spectra has been normalized to the absorbance at 750 nm for easy comparison.
S4
Figure S4. TGA plot showing residual solvent prior to the precursor‐to‐perovskite transition for a thin film using PbCl2 as the lead source. Black line (left axis) is the relative mass of the film only (TGA was tared with the substrate), red line (right axis) is the temperature. The point at which the transition of precursor to perovskite begins is approximately 130 minutes and is marked with a vertical, dotted line. To account for solvent loss during the transfer into the TGA, 10 films were spincast and immediately weighed, the average mass was used as the starting mass for the TGA samples. The salt only weight is the total mass of the PbCl2 and MAI (horizontal dashed line at 0.45 Rel. Mass) and is calculated by assuming the difference between the starting mass, and the first mass reported by the TGA, is due to solvent loss only, and the salt concentration calculated accordingly. The difference between the solid black line (TGA trace) and the dotted horizontal line is taken to be residual solvent that remains in the film.
Calculation of pathway process bars used to create Figure 6 in the main text.
Time [au]
S5
Evap Solvent
From TGA we know how long it takes DMF to evaporate, we know DMF isn’t in the lattice so we set the start of this to t=0 and the length of time to the value derived from TGA data; we assume the lead salt doesn’t change the evaporation rate of the solvent.
Diffuse (MA)Sp
This process is the removal of the spectator from the precursor and is the precursor‐to‐perovskite transition. Both the lag time and the total time for this process are calculated from the WAXS data and the kinetic parameters.
Sublime (MA)Sp
This is the removal of the “amorphous” spectator salt from the film (not the lattice), and has recently been confirmed to occur by sublimation.2 The start time is set based on the assumption that sublimation begins as soon as there is amorphous (MA)Sp to be removed, so it lags the diffusion by a short time in all cases. This process can’t end until the diffusion is complete. Once diffusion is complete, in all cases we have a perovskite film with some (MA)Sp laying around, so we set the time between completion of diffusion and the end of sublimation based on the volatility of (MA)Sp. The volatility of (MA)I, (MA)Cl, and (MA)Ac was measured by TGA and reported in reference 1, Figure 3, the relative position of (MA)NO3 is taken by comparison of the melting temperatures of (MA)NO3 and (MA)Ac which are 110 °C and 80 °C, respectively.
Diffuse (MA)I
This is the removal of (MA)I from the perovskite leading to the decomposition. The case of all iodide is easiest to consider as (MA)Sp=(MA)I; therefore, we assume that the diffusion of (MA)I from the lattice (precursor or perovskite) is continuous. The delineation between the desirable process and the decomposition is just a matter of counting the number of (MA)I units removed from the lattice; thus, the start of decomposition coincides with the end of the Diffuse (MA)Sp step. The case of non‐iodide (MA)Sp is the most difficult to assess. For chloride it is based on experimental observation. For acetate and nitrate, the timing of the first three processes suggests that all three are complete before decomposition begins. Therefore we select a lag time for the start of decomposition at 100 °C and set it the same for both systems. Note: the length of this bar on the chart is random, it continues until only PbI2 is left.
Growth
Time available for perovskite grain growth, which begins when the transformation begins and ends when the (MA)Sp is completely removed from the lattice OR when the solvent is removed, whichever comes first. i.e. growth only occurs while both precursor and solvent still exist.
Coarsen
Time available for coarsening. This begins when (MA)Sp is done diffusing from the lattice and ends when decomposition begins.
S6
References 1. Zhang, W.; Saliba, M.; Moore, D.T.; Pathak, S.K.; Hörantner, M.T.; Stergiopoulos, T.; Stranks, S.; Eperon, G.E.; Webber, J.A.; Abate, A.; Sadhanala, A.; Yao, S.; Chen, Y.; Friend, R.H.; Estroff, L.A.; Wiesner, U.; Snaith, H.J. Ultra‐smooth organic‐inorganic perovskite thin‐film formation and crystallization for efficient planar heterojunction solar cells. Nat. Commun. 2014. doi: 10.1038/ncomms7142 2. Unger, E. L.; Bowring, A. R.; Tassone, C. J.; Pool, V.; Gold‐Parker, A.; Cheacharoen, R.; Stone, K. H.; Hoke, E. T.; Toney, M. F.; McGehee, M. D. Chem. Mater. 2014, doi: 10.1021/cm503828b
S7
