Hadley F. Myers

Hall effect measurements on Bridgman-grown CuInSe2 with sodium

The presence of sodium, either in the substrate or as a co-evaporant during absorber deposition, has been shown to improve the performance of polycrystalline photovoltaic devices made with Cu(In, Ga)Se(2), as well as the ternary CuInSe(2). Investigations have shown Na or Na compounds deposited on the grain boundaries, but none have been found within intact crystal grains, leading to suggestions that grain boundaries may play a role in the improved performance of the cells. Therefore, in this study, ingots containing large monocrystals of CuInSe(2) have been grown, using a vertical-Bridgman method, from melts that also include a varying quantity of sodium. In order to simulate the conditions under which cells are constructed, a proportion of Se above stoichiometry has been added to some of the melts. Resistivity and Hall effect measurements were then performed on the material after growth. The results show no large change in either resistivity or majority hole concentration in either set of samples, although a slight decrease in the latter value was apparent in the excess Se samples with 0.2 and 0.3 at.% Na additions. No clear trend in hole mobility could be discerned, although an increase was seen with 0.2 at.% Na addition for both samples.

Reaction between sodium and selenium in Bridgman-grown CuInSe 2

In the fabrication of CIGS-based solar cells, the incorporation of sodium to increase performance is now standard practice. Because grain boundaries may play a role in the Na action, in this laboratory, studies have been made of the effects of introducing Na into a melt of composition CuInSe2+x to obtain Bridgman-grown monocrystals. It was found that if the added Na exceeds a critical value, [Na]crit, the resulting material changes from p- to n-type. The quantity [Na]crit is almost proportional to x, the excess of Se over stoichiometry according to the formula [Na]crit= (2x+δ)/(1+2x+δ). This can be explained by the strong affinity of Na for Se to form Na2Se-like compounds, starving the ternary of Se to produce a Se-deficient and therefore n-type chalcopyrite. This conclusion is based on measurements of transport properties, along with XRD, SEM/EDX measurements. On the monocrystal surfaces, CuIn3Se5 was detected by XPS.

Sodium — Excess selenium interaction in Bridgman-grown CuInSe 2

In otherwise stoichiometric Bridgman-grown CuInSe2, it has been found, in this laboratory, that the conductivity type changes from p- to n-type when more than 0.25 at. % of elemental sodium is added to the melt. For melts with excess selenium over stoichiometry, x, corresponding to the formula CuInSe2+x + Na, no conversion was observed up to at least 3 at. % of added Na for x = 0.2. However, for smaller x-values (x ≤ 0.05), the inversion continued to take place, where the percentage of Na causing the inversion ([Na]crit) increased with x. From a relationship emerging between [Na]crit and x, it appears that one Se atom reacts with two Na atoms in the inversion process, suggesting the formation of the binary Na2Se. While this compound has not been detected by XRD in the residue from growth runs with a few percent of added sodium, it was detected in a special run where the element addition was greatly increased. XRD has also confirmed that the Na-added n-type material remains chalcopyrite, even with large Na additions of 3 at. %.

P- to n-type conversion with sodium addition in bridgman-grown CuInSe 2

Ingots containing monocrystals of CuInSe2 have been fabricated, using a Bridgman-growth procedure, in quartz ampoules, to which varying concentrations of sodium and excess selenium have been added. The CuInSe2 ingots were nominally p-type, but could be made n-type by adding sufficient sodium to the melts before growth. Small additions of excess Se required that greater additions of Na be introduced for the type conversion to occur. The occurrence of deposits found within the ampoules after growth was strongly indicative of conductivity type. Hall coefficient measurements were carried out on some samples, as were compositional analyses by SEM/EDX, revealing the grown material to be deficient in Cu, regardless of type. Additionally, analysis of the interior of an ingot uncovered no Na within the bulk material, but a significant quantity was found on the exterior surface. This marks the first time sodium was shown to reside only on surfaces, rather than within the bulk crystals, in melt-grown material where sodium was added before compound synthesis.

Investigations of sodium in bridgman-grown CuInSe 2

Measurements of thermoelectric power (α) and electrical resistivity (ρ) were made at room temperature on filamentary samples of CuInSe² (stoichiometric) and CuInSe<inf>2.2</inf> (excess Se), Bridgman-grown with added elemental sodium in the melt in the amounts of 0, 0.1, 0.2, 0.5, 1, 2 and 3 at. %. In the CuInSe<inf>2</inf> + Na samples, conversion from p- to n-type occurred between 0.2 and 0.5 at. % Na, yielding electron concentrations of the order of 10¹⁶ cm⁻³. By contrast, the CuInSe<inf>2.2</inf> + Na samples remained p-type at all the sodium addition levels, yielding hole concentrations of the order of 10¹⁸ cm⁻³. In the important range 0 to 0.5 at. % Na, where sample brittleness is less of a problem than for higher additions, no major changes occurred in hole mobility or resistivity but a small increase in μ<inf>p</inf> and a small decrease in ρ was apparent at 0.1 at. % Na.

Bridgman-grown CuInSe 2 with added Na 2 Se and Na

Ingots containing single crystals were grown from melts containing Cu, In and Se in either stoichiometric proportions (CuInSe<inf>2</inf>) or with an excess of Se (CuInSe<inf>2.2</inf>). In addition, elemental sodium Na⁰ was introduced to both sets of compositions in concentrations ranging from 0 to 3 at. %. Analysis of white deposits on the ampoule walls after growth revealed the presence of Na, however, none was found within the crystals. Further, a red deposit was also analyzed, and was revealed to contain sodium and various forms of the other elements. Thermoelectric power (α) measurements showed little change in the p-type conductivity of CuInSe<inf>2.2</inf> ingots with increases of Na⁰, but a dramatic type change from p to n was observed for stoichiometric CuInSe<inf>2</inf> with Na⁰ concentrations above at least 0.1 at. %. This is in stark contrast to previous work using Na<inf>2</inf>Se, for which α was seen to decrease for both CuInSe<inf>2</inf> and CuInSe<inf>2.2</inf>, implying a hole concentration increase of more than an order of magnitude.

Optical absorption measurements in the wavelength range 1.3-1.6 microns in Bridgman CuInSe 2 grown from melts with added sodium

Measurements of optical absorption coefficient (α) were made on Bridgman-grown CuInSe2 samples, over the wavelength range 1300 to 1600 nm, at room temperature. The samples were obtained from multi-crystalline ingots, grown from melts of composition CuInSe2+x, plus y atomic % of elemental sodium. Here, x, the excess of Se over stoichiometry, ranged from 0 to 0.4 and y, the percentage of Na in the melt, ranged from 0 to 11 at %. It was found that α was considerably increased by the original Na in the melt and also by a stoichiometric deficiency of Se in the melt; that is with a melt formula corresponding to CuInSe1.8, (or x = - 0.2). A smaller increase of α was observed with increase of x, the Se excess. With the addition of Na2Se in the melt, little change was obtained compared with the large increase with elemental Na.

Relation between sodium addition and excess selenium in Bridgman-grown CuInSe2

Ingots containing cm-sized monocrystals of CuInSe2 have been grown in quartz ampoules by a vertical-Bridgman method. Included also in the ampoules were varying quantities of elemental Na, as well as excess Se above stoichiometry. Deposits were seen within the ampoules after growth, including a white deposit formed through a reaction of the Na with the quartz, and a reddish-brown deposit mostly containing Na and Se. Thermoelectric power measurements indicate that stoichiometric material grown with no excess Se and no Na is p-type. With the addition of Na above a critical amount to the ampoule, the resulting ingots were always n-type. This p to n conductivity type change is inhibited when excess Se is also present, and more Na is required for the change to occur. The critical amount of Na was found to be approximately twice the atomic amount of excess Se. XRD indicated that the chalcopyrite structure of the material was maintained with additions of up to 3 at. % Na, even as the material was changed from p- to n-type, and no sodium was detected in the bulk. It is suggested that a reaction between the Na and the Se occurs, resulting in the formation of Na2Se. The Se used in this reaction is therefore unable to be a part of the CuInSe2 lattice, thereby increasing the number of donor-like Se vacancies, rendering it n-type. In this model, the addition of excess Se replaces those Se atoms which react with the Na, so that the material remains p-type.

Room temperature transport measurements on Bridgman-grown CuInSe2 with added sodium

Thermoelectric power, Hall coefficient and resistivity measurements were carried out on material cut from monocrystalline ingots of CuInSe2. The ingots were grown using a vertical-Bridgman procedure, whereby Cu, In and Se were melted and directionally cooled within a sealed quartz ampoule. When stoichiometric proportions of the starting elements were used, the material was always p-type, with hole concentrations of the order of 1017-1018 cm-3. However, the incorporation of a sufficient amount of sodium (0.3 at. %) into the melt was seen to result in n-type material, with electron concentrations of the order of 1016. This conversion from p to n was hindered by the inclusion Se above stoichiometry into the ampoule, and more Na was required for the material to change type. The mobility of the p-type samples, with low sodium additions (0-0.2 at. %), was on average 17 cm2V-1s-1, and was seen to be higher for material grown from melts containing excess Se, corresponding to the chemical formula CuInSe2.2, than from stoichiometry. SEM/EDX (Scanning electron microscope / Energy-dispersive X-ray spectroscopy) analysis of the ingots after growth indicated no sodium residing within the interior of the bulk material, but a significant amount was found on the exterior surface in the case of an ingot grown with CuInSe2.05 and 5 at. % Na. Various deposits found within the ampoules after growth were analyzed, including a copper-rich precipitate found only in ampoules which included a high concentration of Na and a low concentration of excess Se, resulting in n-type material.

Contact resistance to Al-doped ZnO thin films

Polycrystalline ZnO thin films doped with Al are used as transparent conducting oxide (TCO) for solar cells. For such applications, it is necessary to have control on the contact resistance with metals. In this project, we observed the variation of contact resistance of Au/Ti to Al-doped ZnO thin films with sheet resistance.