18 TEMPERA TURE AS A VARIABLE IN PHARMACEUTICAL APPLICATIONS Roger M. Smith 18.1 THE INFLUENCE OF TEMPERATURE ON CHROMATOGRAPHY For much of the early development of liquid chromatography, separations were carried out at ambient temperature and many laboratories did not attempt to regulate or control the temperature of the column. Frequently, the column would be mounted on the side of the pump or detector and thus would be subjected to changes in the room temperature or changes due to external factors, such as sunlight. However, the influence of temperature on the reten- tion times of analytes was well known and had been studied by a number of groups—in particular, Melander et al. [1]. They demonstrated that for most analytes there was a linear relationship between the retention factor of an analyte and the inverse of the absolute column temperature (see Chapter 1). However, for a few samples there has been an increase in retention with increasing temperature usually attributed to entropy effects. In the case of polyethylene glycol oligomers, the optimum separation was achieved with a negative temperature gradient [2]. The retention of leucine-phenylalanine at low pH and high % acetonitrile also increased with increasing tempera- ture [3]. As a result, temperature can play an important role in pharmaceutical analysis. The precise and accurate control of temperature can improve repro- ducibility and method transferability (Section 18.2). In recent years, the use of 811 HPLC for Pharmaceutical Scientists, Edited by Yuri Kazakevich and Rosario LoBrutto Copyright © 2007 by John Wiley & Sons, Inc. elevated or unconventional temperatures have been examined as methods to alter selectivity and column efficienc y, either with conventional mobile phases (Section 18.3) or with solvent free systems, such as superheated water (Section 18.4). Although normally the interest has been in elevated temperatures, sub- ambient chromatography has provided a number of interesting separations (Section 18.5). 18.2 EFFECTS ON METHOD TRANSFERABILITY AND REPRODUCIBILITY As pharmaceutical analysis developed and the need for long-term repro- ducibility became more important, instrument manufacturers recognized the need for temperature stability and by the early 1990s started to include column ovens as an integral part of their instruments. In most cases the temperatures were controlled near or just above ambient because the aim was to ensure a reproducible result rather than to employ temperature as a method variable. However, even now, many chromatographers carry out separations at ambient temperature, partly on the assumption that the conditions in a heated or air- conditioned building are constant. The reality is often different and the tem- perature around a column can alter quite markedly. Sunlight can shine on a column, draughts can blow on the column, or the air-conditioning can be pro- grammed to lower the laboratory temperature overnight or at weekends as a cost-saving exercise.The result is that the retention of analyte compounds can move outside predefined retention windows and the system can show daily or long-term variations and poor reproducibility. Of particular concern is that methods that have been developed and tested in one laboratory are often transferred to another laboratory in the same or a different company as the drug product moves from discovery through toxi- cology, stability studies, formulation, scale-up, and eventually to manufactur- ing quality control. Frequently, it is found that at each transfer, a new method optimization and a revalidation are required, each taking time and money.Sur- prisingly, relatively little research had addressed this problem. There are only few reports of interlaboratory collaborative studies where the target has been to assess the transferability of retention or resolution. In contrast, the trans- fer of quantitation has been repeatedly examined, but this is based on relative peaks areas to an internal or external standard measured under the same con- ditions. This usually compensates for differences in retention time. Typically, interlaboratory studies produce retention time reproducibility, which is much worse than intralaboratory measurements. A comparison of the analysis of forensic drugs in different UK labs [4] and in an international study [5] showed wide variation in relative and absolute retention times even through the mobile phases were closely specified and all the columns were from a single batch of packing material. Within a single laboratory, the situation could be improved markedly by employing temperature control of the column, with an oven or water bath [6]. 812 TEMPERATURE AS A VARIABLE IN PHARMACEUTICAL APPLICATIONS The values were then quite usable for quality control and identification espe- cially when the system was calibrated with standards at frequent intervals .The main residual source of variation was then batch-to-batch differences between columns, although these differences have been reduced in recent years, and uncertainties in the preparation of the mobile phase, which can be reduced by close control of the protocols used. As part of a project to develop a certified reference material for high- performance liquid chromatography [7], it was necessary to demonstrate that the proposed method would yield identical results in different laboratories and on different equipment. However, initial results using a specified temperature and columns from a single batch of packing material gave poor interlabora- tory results, and temperature variations were suspected as a cause [8]. It was found that although the ovens in each laboratory were set to the same nominal temperature, different oven types, air, fan air, convection/conduction, and water bath gave significantly different results, the worst results coming from heaters where the column rested against a heated block. The effective tem- perature could be up to 6°C lower than the set value, and this could be mon- itored by using the changes in shape selectivity and hydrophobicity of a test mixture [9]. Similar observations of oven variability were made by Paesen and Hoogmartens [10]. The protocol for the CRM was then tightened so all the laboratories used a water bath or circulating water jacket and specified lengths of eluent preheating tubing. This gave interlaboratory and intralaboratory variations that were comparable and within the acceptable range [11].Thus to achieve good transferability of a method, not only the obvious factors, such as column make and mobile phase, need to be defined, but also the method of maintaining a constant temperature needs to be specified. Part of the cause of the problem is attributable to differences in the dissi- pation of the fictional heating generated by the movement of the mobile phase through the stationary phase. In a liquid bath, this heat is readily lost to the bath as the external temperature of the column is constant along its length, whereas in a noncontrolled or static air system the mobile phase elutes from the column at a higher temperature (2–3°C) than the inlet [12]. There is also an axial temperature gradient in each case. The effect of different tempera- ture control was also examined by Welsch et al. [13], who found differences between air oven and water baths on normal-phase separations and also studied the effect of inlet temperatures. These effects were later studied for reversed-phase separations by Wolcott et al. [14], who suggested a number of temperature-related reasons for poor method transferability and suggested how different effects changed the temperature profile within the column. 18.3 ELEVATED TEMPERATURE AND PHARMACEUTICAL SEPARATIONS Although temperature has been used for many years to alter separation prop- erties,especially selectivity and efficiency,the operating range has usually been ELEVATED TEMPERATURE AND PHARMACEUTICAL SEPARATIONS 813 modest, typically up to 40–60°C.The principal aim has been to establish a con- trolled system sufficiently above ambient temperature so that day-to-day changes in the laboratory conditions have no effect on the separation. Although the use of non-ambient temperatures might offer advantages in bio- analytical methods , it has been noted that the selection often seems arbitrary and without specific justification [15]. However, Brinkman et al. [16] com- mented that temperature was a variable that should be considered in method development. This comment was echoed in a recent review of the use of mod- erate temperature changes for drug assays by Zhu et al. [17], who noted that “temperature should be considered as a useful variable to control resolution only when components in a mixture are of different types.” For many years, analysts have been deterred from applying significantly ele- vated temperatures because of concern about the volatility of mobile phases and the stability of stationary phases and analytes. More recently as a spin-off from work with supercritical fluid chromatography, many laboratories learned how to handle separations in pressurized columns (up to 300 bar), and hard- ware with pressure-resistant detector flow cells became available. As a con- sequence, the expertise and equipment were commercially available, which could control mobile phases above their boiling point. This has enabled the examination of separations under conditions up to 250°C based on either con- ventional mobile phases or less common solvents, such as superheated water [18]. Temperatures above 80°C, where pressure has to be applied to prevent the mobile phase from boiling, are usually termed either pressurized, super- heated, or subcritical conditions, the latter two terms being more frequently applied to separations with just water as the mobile phase. Either the separa- tions can be isothermal or a temperature gradient can be employed, which generates an effect similar to gradient elution, speeding up the later compo- nents [19, 20]. However, concern is often expressed that the mass of a packed HPLC column might cause the internal temperature to lag behind the oven setting but as long as the internal temperature is reproducible, a valid method can be developed. A number of early studies employed packed capillary columns with a low thermal mass [19]. Three main aims have driven these studies: the use of temperature as a variable to optimize separations, an interest in improved efficiency, and the potential for “green” separations methods, such as superheated water chromatography, which can eliminate the organic solvent from the mobile phase. 18.3.1 Effect of Temperature on Selectivity Although temperature has been proposed as a variable in altering selectivity, it has not been widely used, because the majority of analytes show very similar changes on changing temperature (especially over the limited conventional temperature range). Significant differences may be observed if temperature can cause ionization changes or if analytes with very different functional 814 TEMPERATURE AS A VARIABLE IN PHARMACEUTICAL APPLICATIONS groups are present. However, care must be taken in these situations because relative retention changes with column temperature could result in a lack of method robustness , especially if caused by ionization changes. A few application and studies have examined temperature effects, such as the selectivity dependence of the carotenoids [21] on different columns from 25–45°C. Studies of the prediction of the influence of temperature and solvent strength on the separation of 47 basic acidic and neutral drugs compounds were reported by Zhu et al. [22] in an interlaboratory collaborative study. More recently the influence on temperature on selectivity has been reviewed by Dolan [23, 24]. The changes in retention and selectivity can also be exploited in the thermally tuned tandem column concept by Mao and Carr [25], in which the temperatures of two sequentially linked columns containing different sta- tionary phases can be altered to provide the optimum separation. The tech- nique was applied to the separation of barbiturates, phenylthiohydantoin amino acids [26], and selected basic pharmaceuticals, such as antihistamines (Figure 18-1) [27]. Berthod et al. [28] examined the effect of temperature on chiral separations between 5°C and 45°C using four macrocyclic glycopeptides phases; and although the efficiencies increased with temperature, in 83% of cases the chiral selectivity decreased. 18.3.2 Effect of Temperature on Separation Efficiency One of the reported advantages of raising the temperature of a chromato- graphic separation is an increase in peak efficiency. This is usually attributed ELEVATED TEMPERATURE AND PHARMACEUTICAL SEPARATIONS 815 Figure 18-1. Separation of antihistamines on linked columns with different tempera- tures. ODS at 40°C and PBD-ZrO 2 at 35°C. Mobile phase 40:60 methanol/pH 7 buffer. Solutes: 1, pheniramine; 2, chlorpheniramine; 3, thenyldiamine; 4, bromopheniramine; 5, cyclizane; 6, pyrrobutamine; 7, chlorcyclizane; 8, thonylamine; 9, meclizane. (Reprinted from reference 28, with permission. Copyright 2001, American Chemical Society.) to a reduction in the viscosity of the eluent and an increase in the diffusion rate of the analyte as the temperature is increased. A higher diffusion rate should reduces the mass transfer term effect (the C term) in the van Deemter equation but can also worsen the influence of longitudinal molecular diffusion in the column (the B term) [29]. The improvement in efficiency is normally regarded as most significant for larger analytes, such as biological and synthetic macromolecules, whose size reduces their mobility [30]. For smaller molecules the effects are relatively small, and often an increase in efficiency can be attributed to a reduction in the retention factor on raising the temperature. A second factor, which influences peak shape and apparent efficiency, is the temperature of the incoming mobile phase relative to the column tempera- ture. The presence of this underestimated factor may have obscured or con- fused previous studies of efficiency. Frequently, it has been claimed that the mobile phase in a high-temperature separation must be heated to the same temperature as the column; otherwise, peaks distortion and broadening are observed [18, 31]. This was demonstrated by Thompson and coworkers [32, 33], who reported the band-broadening effect of a thermal mismatch and advocated the use of narrow bore columns to reduce the effects. Guillarme et al. [34] also demonstrated the need for some preheating of the mobile phase and considered the length of tubing required for effective preheating (within 5°C of the oven) particularly with high flow rates. However, even early studies including those by Cooke et al. [35] and by Poppe and Kraak [36] demonstrated that using a mobile phase slightly cooler than the column temperature can improve column efficiency. Usually, differ- ences of 10–20°C gave the highest efficiency. For example, Mayr and Welsch [12], who found the highest efficiency for the separation of five hormone steroids was obtained with the incoming mobile phase at 10°C and a column at 30°C (Figure 18-2). Spearman et al. [9] reported that in one case reducing the inlet to 37°C below the column temperature optimized the results. The effect is thought to have two origins. First, a cooler mobile phase and by implication cooler sample causes an initial sample focusing at the head of the column. Second, the cooler eluent flow reduces the analyte mobility at the center of the column, thus balancing the enhanced temperature and hence mobility in the center of the column caused by friction heating (Figure 18-3) [14]. At its most serious, not only efficiency but also peak distortion has been observed caused apparently by a temperature in-balance. The selection of the optimum column inlet temperature is not totally clear, and this is an area of ongoing research. In such a situation the internal diameter of the column might also effect the equilibration process but Molander et al. [37] found that even using a tem- perature gradient, the differences were minimal for columns narrower than 4.6mm internal diameter. A recent study has found that elevated tempera- tures, up to 70°C, markedly improved the efficiency and peak shapes of bases with intermediate pH eluents [38]. 816 TEMPERATURE AS A VARIABLE IN PHARMACEUTICAL APPLICATIONS 18.3.3 Other Temperature Effects A side effect of a lack of temperature control is that changes can alter the refractive index of the mobile phase , causing baseline disturbances and reduc- ing sensitivity The problem is principally with refractive index detection [39], but it can also influence spectroscopic detectors and their light path can be distorted.Temperature has also been reported to alter the nature of some sta- tionary phases. For example, it caused a change in the chiral selectivity of the resolution of dihydropyrimidone acid and its methyl ester on amylose and cel- lulose stationary phases [40]. 18.3.4 Applications of Elevated Temperatures Almost all the high-temperature work on pharmaceutical compounds has employed reversed-phase separations. In a series of studies since the 1990s, Greibrokk and co-workers [41, 42] have examined the role of elevated tem- perature and temperature gradients. Many were devoted to polymer separa- tions where the use of a temperature gradient speeded up the larger oligomers and provided clear advantages because of the complexity of the sample. As part of the optimization of the conditions for a separation on a packed ELEVATED TEMPERATURE AND PHARMACEUTICAL SEPARATIONS 817 Figure 18-2. Separation of hormones on ChromSpher UOP column at 30°C at differ- ent eluent inlet temperatures: a, 5°C; b, 10°C; c, 22°C; d, 30°C. Compounds: 1, thiourea; 2, hydrocortisone; 3, nortestosterone; 4, dehydro-17a-methyltestosterone; 5, testos- terone; 6, 17a-methyltestosterone. (Reprinted from reference 12, with permission from Elsevier.) capillary, Tran et al. [3] included the effect of temperature on a range of compounds , including naphthalene, acenaphthene, ibuprofen, butylparaben, diethyl phthalate, monoethyl phthalate, amitriptyline, propranolol, ampheta- mine, all-trans-retinol, 13-cis-retinol, and dl-leucine-dl-phenylalanine. A few applications have employed conventional packed columns, although recent developments in new thermally stable stationary-phase materials have generated a renewed interest and the temperature stability of the different stationary-phase materials has been reviewed by Claessens and van Straten [43]. The new materials have included stable metal oxide materials, based on zirconia (Figures 18-4 and 18-5) and titania [44, 45] and hybrid phases com- bining silica and methylene or ethyl bridges [46]. These have been applied in a number of applications to pharmaceutical compounds (Table 18-1). One of the most interesting thermally stable groups of stationary phase materials has been the polybutadiene, carbon and phenyl-coated zirconia 818 TEMPERATURE AS A VARIABLE IN PHARMACEUTICAL APPLICATIONS Figure 18-3. Band-broadening due to thermal effects.(a) Ideal case, no thermal effects; (b) effect of incoming mobile phase that is at a lower temperature than the column; (c) effect of frictional heating; (d) combined effect of cold incoming mobile phase and frictional heating. An oven temperature of 70°C is assumed. (Reproduced from refer- ence 14, by permission from Elsevier.) phases developed by Carr and colleagues [48, 49]. They reported that at high temperatures , the zirconia material offered a much higher stability than silica- based columns [50]. Under these conditions the reduced solvent viscosity gave advantages as flow rates as high at 5ml/min were feasible [51]. The PBD zir- conia column has been used for the separation of tricyclic antidepressants [50] and lidocaine, quinidine, norephrine, tryptamine, amitriptyline, and nor- triptyline. Some selectivity changes with temperature were noted.The low vis- cosity at 100°C also enabled very small 1-µm particles to be used for the separation of benzdiazepines (Figure 18-6) [52]. Guillarme applied these techniques to the separation of a series of caffeine derivatives, including ELEVATED TEMPERATURE AND PHARMACEUTICAL SEPARATIONS 819 Figure 18-4. The separation of barbiturates on (A) ODS at 30°C, (B) C-ZrO 2 at 30°C, (C) ODS at 60°C, (D) C-ZrO 2 at 60°C. Mobile-phase 20/80 acetonitrile. (Reproduced from reference 27, with permission. Copyright 2001, American Chemical Society.) theobromine, theophylline, and caffeine when separation on a PBD zirconia column at 150°C took place in less than 1 minute compared to 7 minutes at 40°C on a Hypercarb column (F igure 18-7) [35]. In a recent study, Marin et al. [53] used a set of test compounds including amitriptyline, salicylic acid, and ibuprofen to compare the temperature stabil- ity of six stationary phases at temperatures up to 150°C. 820 TEMPERATURE AS A VARIABLE IN PHARMACEUTICAL APPLICATIONS Figure 18-5. The separation of therapeutic tricyclic antidepressants on PBD-coated zirconia at different temperatures. Solutes: 1, lidnocaine; 2, quinidine; 3, norephedrine; 4, tryptamine; 5, amitriptyline; 6, nortriptyline. (Reproduced from reference 52, with permission. Copyright 1997, American Chemical Society.) [...]... detectors (UV spectroscopy, FT-IR, 1HNMR, and MS) were applied to model pharmaceuticals [70] and ecdysteroids in plant extracts [71] 18.4.3 Pharmaceutical Applications of Superheated Water Chromatography One attraction of SHW is that it can be used for reversed-phase separations and is therefore readily applicable to a wide range of pharmaceutical compounds including barbiturates, sulfonamides, analgesics... water—A clean eluent for reversedphase high performance liquid chromatography, Anal Commun 33 (1996), 327–329 57 R M Smith and R J Burgess, Superheated water as an eluent for reversed-phase high performance liquid chromatography, J Chromatogr A 785 (1997), 49–55 58 R M Smith, R J Burgess, O Chienthavorn, and J R Stuttard, Superheated water: a new look at chromatographic eluents for reversed-phase liquid... Antia and C Horávth, High-performance liquid chromatography at elevated temperature: examination of conditions for the rapid separation of large molecules J Chromatogr 435 (1988), 1–15 31 M S McCown, D Southern, B E Morrision, and D Garteiz, Solvent properties and their effect on gradient elution high-performance liquid chromatography II Temperature gradients in high performance liquid chromatography... Smith and S Dube, Certified Reference material for HPLC, Chromatographia 61 (2005), 325–332 12 G Mayr and T Welsch, Influence of viscous heat dissipation on efficiency in highspeed high-performance liquid chromatography, J Chromatogr A 845 (1999), 155–163 13 T Welsch, M Schmid, J Kutter, and A Kalman, Temperature of the eluent: A neglected tool in high performance liquid chromatography, J Chromatogr A... TEMPERATURE AS A VARIABLE IN PHARMACEUTICAL APPLICATIONS 60 I D Wilson, Investigation of a range of stationary phases for the separation of model drugs by HPLC using superheated water as the mobile phase, Chromatographia 52 (Suppl) (2000), S28–S34 61 S F Fields, C Q Ye, D D Zhang, B R Branch, X J Zhang, and N Okafo, Superheated water as eluent in high-temperature high-performance liquid chromatographic... Sinclair, E Lenz, and I D Wilson, High temperature reversed-phase HPLC using deuterium oxide as a mobile phase for the separation of model pharmaceuticals with multiple on-line spectroscopic analysis (UV, IR, 1H-NMR and MS), Analyst 126 (2001), 1625–1629 71 D Louden, A Handley, R Lafont, S Taylor, I Sinclair, E Lenz, T Orton, and I D Wilson, HPLC analysis of ecdysteroids in plant extracts using superheated... conditions: Opinions and supplementary observation from Bioanalytical Forum participants, Chromatographia 52 (Suppl) (2000), S57–S59 17 C Zhu, D M Goodall, and S A C Wren, Elevated temperature HPLC: Principles and applications to small molecules and biomolecules, LC-GC Eur 17 (2004), 530–540 18 B Ooms, Temperature control in high performance liquid chromatography, LC-GC Internat 9 (1996), 574–585 19... 61 75–185°C 74 etoposide (Figure 18-11) [75] The method could be applied to even relatively nonpolar pharmaceutical compounds, such as the steroids [62] In a related study, Tajuddin and Smith demonstrated the on-line coupling of SHW extraction with SHW chromatography for the separation of a series of pharmaceuticals [76] The drugs could be sequentially released from the extraction by stepwise temperature... and M R Euerby, Rationalisation of unusual changes in efficiency and retention with temperature shown for bases in reversedphase high-performance liquid chromatography at intermediate pH, J Chromatogr A 1060 (2004), 117–126 39 K Choiket and G, Rozing, The physicochemical causes of baseline disturbance in HPLC- Part II: Column temperature and refractive index detection, LC-GC Eur 16 (2003), 811–818 40 F... Claessens and M A van Straten, Review on the chemical and thermal stability of stationary phases for reversed-phase liquid chromatography, J Chromatogr A 1060 (2004), 23–41 REFERENCES 833 44 J Nawrocki, C Dunlap, A McCormick, and P W Carr, Part I Chromatography using ultra-stable metal oxide-based stationary phases for HPLC, J Chromatogr A 1028 (2004), 1–30 45 J Nawrocki, C Dunlap, J Li, J Zhao, C V McNeff, . recent years, the use of 811 HPLC for Pharmaceutical Scientists, Edited by Yuri Kazakevich and Rosario LoBrutto Copyright © 2007 by John Wiley & Sons,. REPRODUCIBILITY As pharmaceutical analysis developed and the need for long-term repro- ducibility became more important, instrument manufacturers recognized the need for